The Innovation Imperative for Developing East Asia

Home , Four Asian Tigers

The Innovation Imperative for Developing East Asia The Innovation Imperative fter a half century of transformative economic progress that moved hundreds of millions of people out of poverty, countries in developing for Developing East Asia AEast Asia are facing an array of challenges to their future development. Slowed productivity growth, increased fragility of the global trading system, and rapid changes in technology are all threatening export-oriented, labor- intensive manufacturing—the region’s engine of growth. Significant global challenges—such as climate change and the COVID-19 pandemic—are exacerbating economic vulnerability. These developments raise questions about whether the region’s past model of development can continue to deliver rapid growth and poverty reduction.

Against this background, The Innovation Imperative in Developing East Asia aims to deepen understanding of the role of innovation in future development. The report examines the state of innovation in the region and analyzes the main constraints that firms and countries face to innovating. It assesses current policies and institutions, and lays out an agenda for action to spur more innovation-led growth.

A key finding of the report is that countries’ current innovation policies are not aligned with their capabilities and needs. Policies need to strengthen the capacity of firms to innovate and support technological diffusion rather than just invention. Xavier Cirera Policy makers also need to eliminate policy biases against innovation in services, Andrew D. Mason a sector that is growing in economic importance. Moreover, countries need to strengthen key complementary factors for innovation, including firms’ managerial Francesca de Nicola quality, workers’ skills, and finance for innovation. Smita Kuriakose Davide S. Mare Countries in developing East Asia would also do well to deepen their tradition of international openness, which could foster openness in other parts of the world. Trang Thu Tran Doing so would help sustain the flows of ideas, trade, investment, and people that facilitate the creation and diffusion of knowledge for innovation.

ISBN 978-1-4648-1606-2 WORLD BANK EAST ASIA AND PACIFIC 90000 REGIONAL REPORT

9781464 816062 SKU 211606

The Innovation Imperative for Developing East Asia WORLD BANK EAST ASIA AND PACIFIC REGIONAL REPORTS

Known for their economic success and dynamism, countries in the East Asia and Pacific region must tackle an increasingly complex set of challenges to continue on a path of sustainable development. Learning from others within the region and beyond can help identify what works, what doesn’t, and why, in the search for practical solutions to these challenges. This regional flagship series presents analyses of issues relevant to the region, drawing on the global knowledge and experience of the World Bank and its partners. The series aims to inform public discussion, policy formulation, and development practitioners’ actions to turn challenges into opportunities.

TITLES IN THE SERIES

The Innovation Imperative for Developing East Asia

A Resurgent East Asia: Navigating a Changing World

Growing Smarter: Learning and Equitable Development in East Asia and Pacific

Riding the Wave: An East Asian Miracle for the 21st Century

Live Long and Prosper: Aging in East Asia and Pacific

East Asia Pacific at Work: Employment, Enterprise, and Well-Being

Toward Gender Equality in East Asia and the Pacific: A Companion to the World Development Report

Putting Higher Education to Work: Skills and Research for Growth in East Asia

All books in this series are available for free at https://openknowledge.worldbank.org /handle/10986/2147. World Bank East Asia and Pacific Regional Report

The Innovation Imperative for Developing East Asia

Xavier Cirera Andrew D. Mason Francesca de Nicola Smita Kuriakose Davide S. Mare Trang Thu Tran © 2021 International Bank for Reconstruction and Development / The World Bank 1818 H Street NW, Washington, DC 20433 Telephone: 202-473-1000; Internet: www.worldbank.org

Some rights reserved 1 2 3 4 24 23 22 21 This work is a product of the staff of The World Bank with external contributions. The findings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of The World Bank, its Board of Executive Directors, or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Nothing herein shall constitute or be considered to be a limitation upon or waiver of the privileges and immunities of The World Bank, all of which are specifically reserved.

Rights and Permissions

This work is available under the Creative Commons Attribution 3.0 IGO license (CC BY 3.0 IGO) http:// creativecommons.org/licenses/by/3.0/igo. Under the Creative Commons Attribution license, you are free to copy, distribute, transmit, and adapt this work, including for commercial purposes, under the following conditions: Attribution—Please cite the work as follows: Cirera, Xavier, Andrew D. Mason, Francesca de Nicola, Smita Kuriakose, Davide S. Mare, and Trang Thu Tran. 2021. The Innovation Imperative for Devel- oping East Asia. World Bank East Asia and Pacific Regional Report. Washington, DC: World Bank. doi:10.1596/978-1-4648-1606-2. License: Creative Commons Attribution CC BY 3.0 IGO. Translations—If you create a translation of this work, please add the following disclaimer along with the attribution: This translation was not created by The World Bank and should not be considered an official World Bank translation. The World Bank shall not be liable for any content or error in this translation. Adaptations—If you create an adaptation of this work, please add the following disclaimer along with the attribution: This is an adaptation of an original work by The World Bank. Views and opinions expressed in the adaptation are the sole responsibility of the author or authors of the adaptation and are not endorsed by The World Bank. Third-party content—The World Bank does not necessarily own each component of the content contained within the work. The World Bank therefore does not warrant that the use of any third-party-owned individual component or part contained in the work will not infringe on the rights of those third parties. The risk of claims resulting from such infringement rests solely with you. If you wish to reuse a component of the work, it is your responsibility to determine whether permission is needed for that reuse and to obtain permission from the copyright owner. Examples of components can include, but are not limited to, tables, figures, or images. All queries on rights and licenses should be addressed to World Bank Publications, The World Bank Group, 1818 H Street NW, Washington, DC 20433, USA; e-mail: [email protected].

ISBN (paper): 978-1-4648-1606-2 ISBN (electronic): 978-1-4648-1656-7 DOI: 10.1596/978-1-4648-1606-2

Cover and interior design: Bill Pragluski, Critical Stages, LLC.

Library of Congress Control Number: 2021931560 Contents

Foreword...... xiii

Acknowledgments...... xv

About the Authors ...... xvii

Abbreviations...... xix

Overview...... 1 Introduction...... 1 The innovation imperative for developing East Asia...... 3 The state of innovation in developing East Asia ...... 6 Heterogeneity of innovation capabilities within countries, sectors, and firms...... 9 Why diffusion matters...... 15 What inhibits innovation?...... 16 Spurring innovation in developing East Asia: Directions for policy...... 23 Final remarks...... 29 Notes...... 30 References...... 30

1 The State of Innovation in Developing East Asia...... 33 Introduction...... 33 Defining innovation...... 34 The innovation imperative for developing East Asia...... 36 Innovation performance in developing East Asia ...... 42 Road map for this report ...... 51 Notes ...... 52 References...... 53

v vi CONTENTS

2 Conceptual Framework and Stylized Facts...... 57 Introduction...... 57 Key concepts...... 57 Importance of innovation and diffusion of technology in addressing the region’s challenges...... 62 Core elements for designing policies that accelerate technology adoption and diffusion . .70 Conclusions...... 75 Notes ...... 76 References...... 77

3 Technology Adoption and Diffusion: A Firm-Level Perspective...... 81 Introduction...... 81 Is East Asia converging with, or diverging from, the technological frontier? Why diffusion matters ...... 81 Heterogeneity in the pattern and diffusion of innovation across space, sectors, and firms. . .88 What inhibits innovation?...... 94 Conclusions...... 105 Annex 3A The Firm-level Adoption of Technologies survey...... 107 Annex 3B Supplementary tables...... 108 Notes ...... 112 References...... 112

4 Skills and Finance for Innovation...... 117 Introduction...... 117 Skills for innovation...... 117 Finance for innovation...... 130 Conclusions...... 136 Notes ...... 137 References...... 139

5 Innovation Policies and Institutions in the Region: An Assessment ...... 145 Introduction...... 145 Are policies coherent with the objective of supporting diffusion as well as invention?. . 146 Agencies to support innovation...... 154 Assessing public research institution performance in facilitating innovation and technology transfer...... 156 Annex 5A What is the role for innovation agencies?...... 167 Annex 5B Sampling and implementation of the World Bank survey of public research in stitutions ...... 170 Notes ...... 171 References...... 172

6 Action for Innovation: A Policy Agenda...... 175 Introduction...... 175 Addressing the innovation policy mismatch and building firms’ capabilities...... 175 Strengthening key complementary factors: Skills and finance...... 188 Reforming innovation institutions and agencies and building their capacity ...... 193 Final remarks...... 195 Notes ...... 195 References...... 196 CONTENTS vii

Boxes O.1 Defining innovation...... 2 1.1 Defining firm-level innovation...... 35 1.2 COVID-19: Changing the innovation landscape...... 40 1.3 Mitigating effects of climate change through effective technology adoption. . . . .43 2.1 Industry 4.0: An ongoing technological revolution ...... 62 2.2 Understanding adoption and diffusion: S-shaped diffusion curves ...... 65 2.3 Organizational changes associated with Industry 4.0 ...... 69 3.1 Republic of Korea’s use of technology to implement contact tracing and testing strategies in response to COVID-19...... 86 3.2 Innovation in manufacturing versus services...... 91 3.3 International joint ventures and technology transfer in China...... 106 4.1 Employer-employee linked survey data from China and Vietnam on skills for innovation ...... 120 4.2 The effects of technology adoption and innovation on employment ...... 125 5.1 Policy Effectiveness Reviews in developing East Asia...... 147 6.1 Innovation financing done right: Lessons from Israel and China...... 192

Figures O.1 Total factor productivity in most developing East Asian countries is below what would be predicted based on their GDP per capita ...... 4 O.2 Labor productivity and TFP growth have declined in developing East Asia since the Global Financial Crisis...... 4 O.3 Several developing East Asian countries are significant participants in the global value chains for high-tech products...... 6 O.4 The share of firms with international certification is low in much of developing East Asia, and in half of the countries, fewer firms acquire licenses to foreign technology than expected given their countries’ per capita incomes ...... 8 O.5 Most countries in developing East Asia spend less on R&D and produce fewer patents than would be predicted by their per capita incomes...... 8 O.6 Technology adoption lags in developing East Asia are converging with those of OECD countries, but the intensity of technology use is diverging...... 10 O.7 Developing East Asian countries vary widely in firm-level innovation activity . . . .11 O.8 Manufacturing and services firms differ in rates of innovation, especially of new products or processes...... 11 O.9 The intensive use of cutting-edge technology for manufacturing, retail, and agriculture remains limited in Vietnam...... 12 O.10 Radar diagrams show substantial heterogeneity in technological sophistication within firms...... 13 O.11 There is significant duality in firm-level R&D investment...... 14 O.12 Technology adoption brings labor productivity gains to Vietnamese firms. . . . . 16 O.13 Lack of demand and uncertainty is the top self-reported barrier to technology adoption among Vietnamese firms of all sizes...... 17 O.14 Firms in developing East Asian countries score lower on management capabilities than firms at the global frontier...... 18 O.15 Employees in more-innovative firms in China and Vietnam have jobs that are more intensive in nonroutine cognitive analytical and interpersonal tasks...... 19 O.16 Most innovative firms in developing East Asia report difficulties in hiring workers with adequate skills...... 20 viii CONTENTS

O.17 Banks remain the dominant source of finance to firms in most of developing East Asia (except China) ...... 21 O.18 Appropriate policy instruments to foster innovation differ depending on the level of innovation capabilities...... 24 O.19 Developing East Asian countries occupy three distinct clusters with respect to innovation capabilities...... 25 1.1 Although productivity in the major economies of developing East Asia has been rising, it remains well below the frontier...... 37 1.2 Total factor productivity in most developing East Asian countries is below what would be predicted based on their GDP per capita ...... 38 1.3 Labor productivity and TFP growth have declined in developing East Asia since the Global Financial Crisis...... 38 1.4 COVID-19 has induced greater use of and investment in digital solutions in developing East Asia since the beginning of the pandemic...... 40 1.5 Several developing East Asian countries are significant participants in the global value chains for high-tech products...... 45 1.6 The share of firms with international certification is low in much of developing East Asia, and in half of the countries, fewer firms acquire licenses to foreign technology than expected given their countries’ per capita incomes ...... 46 1.7 Most countries in developing East Asia spend less on R&D and produce fewer patents than would be predicted by their per capita incomes...... 47 1.8 With some noteworthy exceptions, developing East Asian countries still face challenges concerning innovation in services, including in the digital space . . . . .48 1.9 Countries in developing East Asia show considerable variation in the incidence and type of firm-level innovation...... 49 1.10 In developing East Asia, a higher share of firms reports innovating in manufacturing than in services ...... 50 1.11 Firms in developing East Asia report a mix of innovations that are new to the firm and new to the domestic market ...... 51 2.1 The innovation knowledge function begins with investments and inputs that, once transformed into innovation outputs and outcomes, can yield improved firm performance...... 58 2.2 Sophistication of firm-level innovation correlates strongly with countries’ economic development...... 59 2.3 Innovation quality correlates closely with countries’ per capita income...... 60 2.4 The relationship between returns to innovation and distance to the frontier follows an inverted U-shape...... 61 B2.1.1 The four stages of industrial revolution have repeatedly transformed manufacturing. . 63 B2.2.1 S-shaped diffusion curves in the adoption of hybrid seed corn in selected US states, 1933–56 ...... 66 2.5 Lags in the adoption of new technologies are now similar between Western countries and the rest of the world, but gaps in the intensity of technology use have widened. .67 2.6 Accumulating knowledge to reach the innovation frontier: A firm-centered process. . .72 2.7 Developing East Asian countries occupy three distinct clusters with respect to innovation capabilities...... 73 2.8 Appropriate policy instruments to foster innovation differ depending on the level of innovation capabilities...... 74 3.1 Technology adoption lags in developing East Asia are converging with those of OECD countries, but intensity of technology use is diverging...... 83 CONTENTS ix

3.2 Data on internet and robot use in East Asia tell two different tales of diffusion and convergence with the frontier...... 84 3.3 The intensive use of cutting-edge technology for manufacturing, retail, and agriculture remains limited in Vietnam...... 85 3.4 Digital readiness indexes show widespread internet access among Vietnamese firms, but few are fully equipped to use digitalized processes for primary business functions. . 87 3.5 Technology adoption brings labor productivity gains to Vietnamese firms. . . . . 89 3.6 Developing East Asian countries vary widely in firm-level innovation activity. . . . .90 B3.1.1 Manufacturing and services firms differ in rates of innovation, especially of new products or processes...... 91 3.7 Firm-level innovation in East Asia also differs across sectors...... 93 3.8 Radar diagrams show substantial heterogeneity in technological sophistication within firms...... 94 3.9 There is significant duality in firm-level R&D investment...... 96 3.10 Inventions and foreign technology adoption have increased over time, by country, 2000–19...... 97 3.11 Firms in developing East Asian countries score lower on management capabilities than firms at the global frontier...... 99 3.12 Managers of Vietnamese firms express overconfidence about their firms’ technological capabilities...... 100 3.13 Lack of demand and uncertainty is the top self-reported barrier to technology adoption among Vietnamese firms of all sizes...... 101 4.1 More-innovative firms in China and Vietnam have more highly educated employees and a higher share with STEM degrees...... 122 4.2 Stronger cognitive and socioemotional skills are associated with greater innovation intensity among firms in Vietnam, 2019 ...... 122 4.3 High-innovation firms in China exhibit high demand for workers with strong socioemotional skills...... 123 4.4 Employees in more-innovative firms in China and Vietnam have jobs that are more intensive in nonroutine cognitive analytical and interpersonal tasks . . . . .124 4.5 Managers’ and technicians’ skills are particularly important to firm innovation in Vietnam, 2019...... 124 4.6 Most innovative firms in developing East Asia report difficulties in hiring workers with adequate skills ...... 128 4.7 Several developing East Asian countries underperform in basic skills formation, creating challenges for the development of more-advanced skills...... 129 4.8 Although developing East Asian countries often produce a higher share of STEM graduates than do countries with similar income levels, most have lower tertiary enrollment rates...... 129 4.9 Banks remain the dominant source of finance to firms in most of developing East Asia ...... 132 4.10 In developing East Asia, most firms—both innovative and non-innovative— finance fixed assets using internal funds...... 133 4.11 The relationship between firms’ external financing of fixed assets and firm-level innovation is stronger in some developing East Asian countries than in others. . . 134 4.12 Growth in risk capital financing has varied considerably across developing East Asia. . 136 4.13 The factors that affect the attractiveness of venture capital vary by country across the region ...... 136 5.1 Despite numerous innovation policy instruments in Indonesia, the Philippines, and Vietnam, public spending concentrates on only a few...... 149 x CONTENTS

5.2 The distribution of public resources for innovation, by firm type, varies considerably across Indonesia, the Philippines, and Vietnam...... 150 5.3 Indonesia, the Philippines, and Vietnam allocate few, if any, public resources to innovation policy instruments whose potential beneficiaries include firms in services sectors...... 151 5.4 In all three studied countries, numerous innovation policy instruments have expenditures below US$100,000, indicating a potential lack of scale for significant impact...... 152 5.5 Innovation policy instruments in the Philippines and Vietnam remain far from the design, implementation, and governance frontiers...... 153 5.6 The distribution of R&D spending in East Asia, by sector, reflects variations in R&D intensity related to countries’ levels of income and institutional development. . 157 5.7 Across several East Asian countries, scientific output has increased dramatically, but scientific productivity trails that of high-income countries ...... 158 5.8 Research institutions in Malaysia, the Philippines, and Vietnam that were autonomous and included industry stakeholders on their boards engaged in greater collaboration with industry in 2017–18...... 159 5.9 Research institutions with good governance exhibit a greater intensity of technology transfer ...... 160 5.10 Research organizations with higher shares of industry-funded R&D undertake more research collaboration...... 161 5.11 The survey showed which technology transfer activities are most important and attract the most participation from research organizations...... 162 5.12 PROs are more likely than RCs to report collaboration and knowledge transfer links with industry...... 164 5.13 Research institutions report a variety of financial and nonfinancial incentives for technology transfer ...... 165 6.1 Republic of Korea’s policy choices have enabled significant technological catch-up. . 184 6.2 A range of policy instruments can be used to promote technology adoption, diffusion, and invention...... 185

Maps 1.1 East Asian countries are among those most at risk from climate change...... 42 1.2 Developing East Asia is home to several of the world’s top 100 innovation clusters . . 44 3.1 The spatial distribution of innovative start-ups in developing East Asia shows a concentration in urban hubs, excluding large areas of all countries...... 95 3.2 Global innovation networks with developing East Asia have grown denser in recent decades ...... 104 4.1 Countries in developing East Asia attracted relatively large amounts of risk capital financing, as a percentage of GDP, in 2018 ...... 135

Tables 3.1 Firms that innovate in developing East Asia tend to be larger, exporters, and located in larger cities ...... 98 3B.1 Management scores of selected East Asian countries relative to the United States, by quantile...... 108 3B.2 Import penetration is positively associated with innovation in developing East Asia. . 108 3B.3 The discouraging effect for laggard industries in developing East Asia...... 109 3B.4 Effects of WTO accession: Innovation and import-export activity in Vietnam. . . 109 CONTENTS xi

3B.5 Co-inventions gravity model for developing East Asia countries...... 110 3B.6 FDI spillover effects on innovation in developing East Asia...... 111 3B.7 Effect of firms’ distance from the frontier on the strength of spillovers in developing East Asia ...... 112 4.1 A range of workforce skills are important for enabling innovation...... 118 4.2 Socioemotional skills reflect personal attributes that are essential to workplace performance...... 119 4.3 Market frictions have distinct effects on different types of firm innovation . . . . 131 5.1 Policy effectiveness reviews in Indonesia, the Philippines, and Vietnam assess the portfolio of innovation policy instruments and related spending...... 148 5.2 In all three countries, a significant share of innovation policy instruments have overlapping scope...... 152 5A.1 Overview of main support mechanisms used by leading innovation agencies . . . .167 5A.2 Examples of innovation agencies, by type...... 170 6.1 The most appropriate mix of policy instruments to facilitate innovation and diffusion depends on countries’, sectors’, and firms’ innovation capabilities. . . . 177 6.2 Developing East Asian countries have varying innovation policy priorities, based on the capabilities escalator...... 180 6.3 Republic of Korea’s innovation policies continually evolved to reflect its accumulation of capabilities...... 181 6.4 Japan, Republic of Korea, and Singapore have integrated innovation skills into their curricula...... 189

Foreword

he remarkable economic rise of East Developing East Asia—will deepen policy Asia over the past 50 years is well makers’ understanding of the critical role for Tknown. The track records of these innovation in the region’s future growth and economies in sustaining rapid growth and development and will lay out a policy agenda poverty reduction were well documented in for innovation-led growth. two preceding volumes of this regional series, For middle-income countries with widely Riding the Wave: An East Asian Miracle for varying institutional, technological, and the 21st Century (2018) and A Resurgent firm-level capacities, “innovation” entails East Asia: Navigating a Changing World not only the invention of new products and (2019). processes at the Industry 4.0 frontier but also Middle-income countries in the region the dif fusion and adoption of existing tech- aspire still higher but now face unprec- nologies or practices. And these more basic edented challenges to their economic suc- forms of innovation can have big payoffs. cesses. Productivity growth has declined since This is where The Innovation Imperative the Global Financial Crisis. Changes in the makes its most practical and meaningful global trade environment and technological contribution: as a guide to tailoring innova- advances are challenging the region’s tradition-enabling reforms in country-specific, tional growth engine: export-oriented manu- capacity-specific ways. facturing. Climate change and, more recently, Such guidance is sorely needed. Despite the COVID-19 pandemic have increased the the significant promise of innovation—and urgency of transitioning to new modes of some high-profile success stories—most coun- production that are more environmentally tries in developing East Asia (except China) friendly, more flexible, more digitally inte- underperform on several key indicators of grated, and more connected to consumers. innovation, including spending on research Can the region’s past model for success and development (R&D) and patents, as well continue to deliver rapid future gains? Not as the adoption and use of new technologies. without the advances and rewards of inno- In fact, only a small share of the region’s firms vation. To that end, this latest volume in currently engage in any type of innovation the series—The Innovation Imperative for activity.

xiii xiv Foreword

Why is this? New survey evidence pre- So what will it take to spur greater inno- sented in this report reveals that the region’s vation in developing East Asia? As this report firms face several barriers to adopting new makes clear, the region’s policy makers must technologies that could transform their pro- reorient innovation policy to focus on incen- ductivity. They lack adequate information on tivizing a large mass of firms to simply start these technologies and face high uncertainty innovating. It is the broad adoption and diffu- about the returns to their use. They often sion of existing technologies, not only inven- lack the capacity to innovate—due to inad- tion of new ones, that will determine the pace equate management capabilities and work- of economic and productivity growth in the force skills. And they frequently lack access region. to external financing for technology adoption By fostering greater innovation in this way, or broader innovation projects. Importantly, countries in developing East Asia will be able countries’ innovation policies and institutions to surmount the array of challenges they now are also often misaligned with firms’ capabili- face and reinvigorate growth and develop- ties and needs. ment for a better future.

Victoria Kwakwa Regional Vice President East Asia and Pacific Region The World Bank Acknowledgments

his report is a joint product of the Rakheja, Miguel Sarzosa, Haris Tiew, Erik Office of the Chief Economist, East van der Marel, and Wenshi Xuan. Cecile TAsia and Pacific Region, and the Wodon provided invaluable administrative Finance, Competitiveness, and Innovation and logistical assistance. (FCI) Global Practice of the World Bank. The team benefited from helpful sug- It was authored by Xavier Cirera, Andrew gestions from colleagues both within and D. Mason, Francesca de Nicola, Smita outside the World Bank Group. Particular Kuriakose, Davide S. Mare, and Trang Thu thanks to Asya Akhlaque, Irina Astrakhan, Tran, under the supervision of Aaditya Mary Hallward-Driemeier, Keun Lee (Seoul Mattoo, chief economist, East Asia and National University), Toby Linden, Rajah Pacific; and Denis Medvedev, practice man- Rasiah (University of Malaya), and Cecile ager, FCI; with the guidance of Victoria Thioro Niang for serving as peer review- Kwakwa, regional vice president, East Asia ers. Helpful comments and inputs were and Pacific. also received from Zubair Khurshid Bhatti, The report builds on the work of a Alejandro Cedeno, Paulo Correa, Souleymane larger team that included Pinyi Chen, Sarah Coulibaly, Philippe de Meneval, Ousmane Hebous, Ergys Islamaj, Sachiko Kataoka, Dione, Pierre Graftieaux, Caren Grown, Birgit Faruk Miguel Liriano, Fabiola Saavedra, Dea Hansl, Vera Kehayova, Michel Kerf, Martin Tusha, and Pluvia Zuñiga. Valuable contri- Melecky, Martin Raiser, Siddharth Sharma, butions were also made by Anna Alejo, Cha Sudhir Shetty, Cao Thu Anh, Georgia Wallen, Crisostomo, Martina Ferracane, Jaime Frias, and Xiaoqing Yu. The team also appreciates Jin Lee, Kevin Macdonald, Koji Miyamoto, the feedback provided by the participants in a Son Nguyen Hoai, Albert Park, Prerna December 2019 Authors’ Workshop.

About the Authors

Xavier Cirera is a senior economist in the protection, labor, skills, and gender equal- Finance, Competitiveness, and Innovation ity. He is coauthor of several World Bank (FCI) Global Practice of the World Bank. flagship reports, including A Resurgent His work focuses on innovation and entre- East Asia: Navigating a Changing World; preneurship. He has led the evaluation of Toward Gender Equality in East Asia innovation policies, including through the and the Pacific; and Informality: Exit development of public expenditure reviews and Exclusion; as well as the World Bank in science, technology, and innovation imple- Policy Research Report, Engendering mented in Brazil, Chile, Colombia, Ukraine, Development: Through Gender Equality and Vietnam. He is the coauthor of The in Rights, Resources, and Voice. He has Innovation Paradox: Developing-Country also been an affiliated professor at the Capabilities and the Unrealized Promise of Georgetown Public Policy Institute of Technological Catch-Up and A Practitioner’s Georgetown University. He holds a doc- Guide to Innovation Policy: Instruments torate in applied economics from Stanford to Build Firm Capabilities and Accelerate University and a master’s degree in public Technological Catch-Up in Developing policy from Harvard University. Countries. His most recent work focuses on the measurement and impact of technology Francesca de Nicola is a senior economist in adoption and diffusion. Before joining the the Office of the Chief Economist, East Asia World Bank, he served as a research fellow and Pacific Region, of the World Bank. She at the Institute of Development Studies at the has led operations and analytical work to University of Sussex. He holds a doctorate in support the development of the private sector economics from the University of Sussex. and financial markets in Africa and Eastern Europe. Her research interests span various Andrew D . Mason is lead economist in the areas of development economics, including Office of the Chief Economist, East Asia productivity, the role of political connections, and Pacific Region, of the World Bank. and weather insurance. She has published He has carried out policy research on a in leading journals, including the Journal range of issues, including poverty, social of Development Economics, Quantitative

xvii xviii About the Authors

Economics, and Energy Economics. She holds Journal of Operational Research and a doctorate in economics from the Johns the Journal of Financial Stability, as well Hopkins University. as in the World Bank’s Global Financial Development Report 2019/2020: Bank Smita Kuriakose is a senior economist in Regulation and Supervision a Decade after the FCI Global Practice of the World Bank. the Global Financial Crisis. Before joining Currently based in Kuala Lumpur, she man- the World Bank, he was a lecturer (assis- ages the World Bank’s engagement on private tant professor) in business economics at the sector development, innovation, and entre- University of Edinburgh Business School and a preneurship policy in Malaysia. Her work financial consultant for large European banks. has covered more than 20 countries across He holds a doctorate in banking and finance Africa, East Asia and Pacific, and Europe and from the University of Rome Tor Vergata. Central Asia. She has led regional and country flagship publications on innovation and Trang Thu Tran is a senior economist in entrepreneurship issues in addition to leading the FCI Global Practice of the World Bank. multisectoral operations. She holds gradu- She is interested in research on firms, with ate degrees in economics from the University experience working on topics spanning firm of Maryland and from the Delhi School of dynamics, entrepreneurship and innovation Economics. policy, international investment, and business regulations. Her most recent coauthored Davide S . Mare is a researcher at the World work includes the World Bank Group’s flag- Bank and honorary lecturer at the University ship reports, Making It Big: Why Developing of Edinburgh Business School. His main Countries Need More Large Firms and the research interests lie in banking, focusing Global Investment Competitiveness Report on bankruptcy prediction, credit risk mea- 2019/2020: Rebuilding Investor Confidence surement, competition, and firm finance. in Times of Uncertainty. She is an applied His work has appeared in leading inter- microeconomist with a doctorate from the national journals, including the European University of Maryland. Abbreviations

A*STAR Agency for Science, Technology, and Research AI artificial intelligence AIM Agensi Inovasi Malaysia ALL Adult Literacy and Life Skills CEES China Employer-Employee Survey DARPA Defense Advanced Research Projects Agency EISS Epidemic Investigation Support System ESIS Enterprise Survey on Innovation and Skills EU European Union FAT Firm-level Adoption of Technology (survey) FDI foreign direct investment fintech financial technology FTT forced technology transfer GBF general business functions GDP gross domestic product GII Global Innovation Index GIN global innovation network GPT general purpose technology GVC global value chain HEI higher education institution HR human resources IALS International Adult Literacy and Life Skills ICT information and communication technology IJV international joint venture IoT internet of things IP intellectual property IPO initial public offering IPR intellectual property right(s) IRAP Industrial Research Assistance Program IT information technology ITA industrial technology adviser

xix xx Abbreviations

M&E monitoring and evaluation MNE multinational enterprise MSMEs micro, small, and medium enterprises NIS national innovation system NIST National Institute of Standards and Technology NQI national quality infrastructure NRF National Research Foundation OECD Organisation for Economic Co-operation and Development PA performance-based agreement PE performance evaluation PER policy effectiveness review PIAAC Programme for the International Assessment of Adult Competencies PISA Programme for International Student Assessment PRO public research organization R&D research and development RC research center S&T science and technology SBIR Small Business Innovation Research SMEs small and medium enterprises SOE state-owned enterprise STEM science, technology, engineering, and mathematics STEP Skills Toward Employability and Productivity STI science, technology, and innovation TES technology extension services TFP total factor productivity TiVA Trade in Value Added TRIPs Agreement Trade-Related Aspects of Intellectual Property Rights TTO technology transfer office TVET technical and vocational education and training VC venture capital(ist) WBES World Bank Enterprise Survey WMS World Management Survey WTO World Trade Organization Overview

Introduction on the region’s growth prospects, narrowing the opportunities for reaping demographic Countries in developing East Asia dividends. Second, the slowing of global have undergone significant economic goods trade, uncertainty about the future of transformation, but the region now the global trading system, and rapid changes faces an array of challenges in in technology are all challenging a key engine sustaining growth of growth in the region: export-oriented East Asia’s economic success over the past manufacturing. Third, the COVID-19 pan- 50 years has been transformative. High rates demic, together with ongoing climate change, of growth have propelled countries in the are increasing economic vulnerability and region from low-income to middle-income, highlighting a pressing need for new modes of and even in a few cases, to high-income sta- production in the region. tus. An approach that has become known These forces, alone and together, raise as the “East Asian development model”— questions about whether the model that has a combination of policies that fostered driven the region’s economic success in the outward-oriented, labor-intensive sectors past can continue to deliver rapid growth and growth; investments in basic human capital; development in the future. and sound economic governance—has been instrumental in moving hundreds of millions Innovation is increasingly important to of people out of poverty and into economic future growth security. Despite their past successes, the region’s Recent studies have highlighted the critical role middle-income countries now face an array that innovation must play in developing East of challenges as they strive to continue their Asia if the region’s countries are to maintain economic progress: First, productivity growth or increase productivity growth in a rapidly has declined since the 2008–09 Global changing and highly uncertain global eco- Financial Crisis. This, and rapid population nomic environment (Mason and Shetty 2019; aging in several countries, is putting pressure World Bank and DRC 2019). Reinforcing 2 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

the case for more innovation-led growth is a innovation and lays out an agenda for action significant global literature showing strong aimed at spurring innovation-led growth in links between innovation and productivity at the region. the macro- and microeconomic levels (Cirera The report emphasizes the importance and Maloney 2017; Comin and Hobijn 2010; for the region of effectively using tech- Griliches 1998; Hall 2011; Mohnen and Hall nologies that are already available in high- 2013; Solow 1957). income economies as a means of raising Against this background, this report seeks productivity and addressing economic and to deepen policy makers’ understanding of societal challenges. For this reason, the the critical role for innovation in the future report adopts a broad definition of inno- growth and development of developing East vation that encompasses both innovation Asia.1 To achieve this, the report examines the as “invention” of new products and pro- region’s key innovation challenges, assesses cesses at the knowledge frontier and as its state of innovation, and analyzes the main “diffusion and adoption” of existing tech- constraints firms face in effectively pursuing nologies and practices that enable firms to innovation. The report then examines the pol- undertake new and more effective modes of icies and institutions needed to enable greater production (box O.1).

BOX O.1 Defining innovation

The report adopts a broad view of innovation as the to existing functions or user utility. “Relevant accumulation of knowledge and implementation of functional characteristics include quality, techni- new ideas. Specifically, following the Oslo Manual cal specifications, reliability, durability, economic 2018, a “business innovation” is defined as a “new efficiency during use, affordability, convenience, or improved product or business process (or com- usability, and user friendliness” (OECD and bination thereof) that differs significantly from the Eurostat 2018, 71). firm’s previous products or business processes and • A business process innovation is “a new or that has been introduced on the market or brought improved business process for one or more busi- into use by the firm” (OECD and Eurostat 2018, ness functions that differs significantly from the 20). The report considers innovation defined both firm’s previous business processes and that has as “invention” or “discovery” (that is, those devel- been brought into use by the firm.” The Oslo opments that push the technological frontier) and Manual 2018 lists the six functional categories as “diffusion” or “adoption” of existing technolo- to identify and distinguish between types of busi- gies and practices that lead firms to novel ways of ness process innovations (OECD and Eurostat producing or acting. The latter definition is perti- 2018, 73): nent to most of the firms operating in developing  Innovative methods for manufacturing East Asia. products or offering services An innovation may be either technological or nontechnological. Specifically, the Oslo Manual  Innovations in distribution and logistics 2018 defines the following two main types of inno-  Innovations in marketing and sales activities vations (OECD and Eurostat 2018, 21):  Innovations in the provision and maintenance of information and communication systems • A product innovation is “a new or improved good  I nnovations in administration and management or service that differs significantly from the firm’s  Innovations in product and business process previous goods or services and that has been development. introduced on the market.” This includes the addition of either new functions or improvements Source: Adapted from OECD and Eurostat 2018. Overview 3

The innovation imperative for Changes in global trade and developing East Asia technologies are challenging the region’s main engine of growth: export- Several economic forces are driving an imper- oriented manufacturing ative for a more innovation-led growth model in developing East Asia. The slowing of global goods trade and ambiguity about the future of the global trading system pose risks to a development model Productivity remains relatively low in that has effectively used trade, foreign direct developing East Asia—and productivity investment (FDI), and integration into global growth has declined since the Global value chains (GVCs) as critical channels for Financial Crisis growth. Furthermore, a new technological Despite their remarkable growth perfor- revolution—Industry 4.0—poses a risk of mance, countries in developing East Asia disrupting existing production structures as still face important productivity challenges. it moves toward more flexible manufacturing Productivity—whether measured in terms and customization and increases the impor- of labor productivity (output per worker) tance of proximity to customers. These tech- or as total factor productivity (TFP, a mea- nological advances could potentially shorten sure of economic efficiency)—has been GVCs or result in the reshoring of produc- rising over time, although it remains well tion systems that have been central in fueling below the productivity frontier, defined as growth in developing East Asian countries. the productivity level in the United States. Even in Malaysia, whose productivity is The COVID-19 pandemic and other the highest in developing East Asia, labor shocks, including climate change, are productivity was only about 42 percent, accelerating the need for new modes and TFP about 62 percent, of levels in the of production United States in 2017. Although productivity generally increases as countries develop, The COVID-19 pandemic TFP in most developing East Asian coun- The COVID-19 pandemic has underscored the tries is below what would be predicted on importance of innovation as policy makers and the basis of their gross domestic product private firms have rushed to adopt or develop (GDP) per capita (figure O.1). technologies to address both the health and the Productivity growth has slowed world- economic effects of the outbreak. This effort wide since the Global Financial Crisis, and has included, among other things, the appli- developing East Asia has not been immune. cation of digital mobile technologies to pro- Indeed, the region has experienced the vide real-time information about the spread second steepest slowdown in labor proof the virus and support social distancing; ductivity growth of all emerging market drone technologies for such applications as and developing regions since the Global aerial disinfection, contactless transportation Financial Crisis (World Bank 2020). While of medical supplies, and consumer deliveries; labor productivity growth has declined and advanced biomedical technologies and across the region, the decline has been par- artificial intelligence (AI) to develop testing, ticularly pronounced in China (figure O.2). vaccines, and treatments for the virus.2 A decomposition of labor productivity Importantly, the COVID-19 pandemic growth shows that the slowdown largely is a shock to GDP not seen for decades in reflects weaker TFP growth. the region—one that may have long-lasting 4 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

FIGURE O.1 Total factor productivity in most developing East Asian countries is below what would be predicted based on their GDP per capita

1.2

1.0 United States

0.8 Singapore Japan 0.6 Philippines Myanmar Korea, Rep. Thailand Malaysia Lao PDR Indonesia 0.4 China Cambodia Mongolia

TFP level at current PPP (US = 1) Vietnam 0.2

0 7 8 9 10 11 12 13 Log GDP per capita, PPP (constant 2011 international $)

Source: World Bank elaboration, based on Penn World Table (PWT) version 9.1 and Asian Productivity Organization (APO) Database 2019, version 2 (https://www.apo-tokyo.org/wedo/productivity-measurement/). Note: Data are from 2017. Countries in light blue designate the 10 “developing East Asia” countries studied in this report. Total factor productivity (TFP) is measured in purchasing power parity (PPP) terms relative to the United States (1.0). TFP series are calculated by the PWT team, except for Cambodia, Myanmar, and Vietnam, which are estimated through the methodology of Feenstra, Inklaar, and Timmer (2015), using data from PWT (version 9.1) and labor share estimates from the APO Database 2019 (version 2). GDP = gross domestic product; PPP = purchasing power parity; TFP = total factor productivity.

FIGURE O.2 Labor productivity and TFP growth have declined in developing East Asia since the Global Financial Crisis

b. Decomposition of productivity growth, a. Decomposition of productivity growth, China East Asia and Paci c, excluding Chinaa

12 6

10 5

8 owth (%) 4

6 3

4 2 erage annual growth (%) erage annual growth 2 erage annual gr 1 Av Av 0 0 1991–96 2003–08 2013–18 1991–96 2003–08 2013–18 Labor productivity growth Capital deepening TFP Human capital growth

Sources: Conference Board (CB), Asian Productivity Organization (APO) Database 2019, version 2 (https://www.apo-tokyo.org/wedo/productivity-measurement/); and Penn World Table (PWT) version 9.1 data; van der Eng (2009); World Development Indicators database; and World Bank calculations. Note: Labor productivity is defined as GDP per worker. PWT data were used as the baseline. When PWT (version 9.1) data were not available, the APO Database 2019 (version 2) was used. Conference Board data were used for 2018. GDP = gross domestic product; TFP = total factor productivity. a. Panel b shows weighted averages calculated using GDP weights at 2010 prices. Countries included are Cambodia, Indonesia, Lao PDR, Malaysia, Mongolia, Myanmar, the Philippines, Thailand, and Vietnam. For Indonesia, TFP growth was calculated by the World Bank, extending data from van der Eng (2009). Overview 5

effects. So large a shock, affecting both According to the Global Climate Risk demand and supply, has highlighted the need Index 2020, four Southeast Asian coun- for more flexible management and produc- tries—Myanmar, the Philippines, Thailand, tion processes, both to accommodate restric- and Vietnam—were among the 10 countries tions due to social-distance measures and to most affected by extreme weather events prepare for what may be very different econo- between 1999 and 2018 (Eckstein et al. 2019). mies in the post–COVID-19 era. Production Moreover, the continued reliance of Southeast processes will be more automated, digitally Asian countries on agriculture and the con- integrated, and connected to consumers. centration of populations in coastal regions One challenge for policy makers, how- exacerbate their vulnerability. Many major ever, is that the pandemic may have conflict- coastal cities are seriously imperiled, including effects on each of the two dimensions ing Shanghai and Tianjin, China; Jakarta, of innovation: invention and diffusion. Indonesia; Ho Chi Minh City, Vietnam; and Regarding invention, the pandemic is boost- Bangkok, Thailand. These changes demand ing research and development (R&D) on urgent technological solutions, whether to tests, vaccines, and treatment to combat ensure that agricultural production is sustain- the disease. This is likely to have positive able or to enable safe and productive factory spillovers for broader scientific and medical environments at higher temperatures. research in areas such as biotechnology. At the same time, the social distancing needed To sustain high economic performance to contain the disease has impeded scientific in the face of these challenges, the research not related to COVID-19, by shut- region’s countries must move toward ting down laboratories and durably disrupt- a more innovation-led growth model ing experiments. As for diffusion, adapting to social To address all these challenges will demand distancing has boosted firms’ and households’ that societies become more innovative. demand for technologies supporting digital Countries in developing East Asia must find communication, conveyance, and commerce new and more effective ways to increase pro- that will likely be used well beyond the pan- ductivity growth as they seek to build on past demic. However, the crisis-induced economic economic success and move progressively from contraction and uncertainty are inhibiting middle- to high-income status. Indeed, their investments in both invention and diffusion high-income neighbors—Japan, the Republic in a variety of other areas by cutting resources of Korea, and Singapore—have all used inno- and dampening expected returns. Policy mak- vation as a vehicle to improve efficiency and ers will thus need to find ways to accelerate boost their incomes with great success. the technological transformation of their The narrowing of productivity and techno- economies while managing these tensions. logical gaps with high-income economies could help developing East Asian countries to address Climate change trade challenges, including threats of reshoring, Similarly, climate change is challenging tradi- by increasing their competitiveness and upgrad- tional approaches to production and growth. ing their participation in GVCs. Similarly, an Regarding mitigation, it is imperative to have effective response to the COVID-19 pandemic cleaner, more energy-efficient production that and the risk of other health shocks requires reduces carbon emissions. As for adapta- strong research and innovation fundamentals tion, temperatures will increase significantly to address and monitor health impacts, as well in developing East Asia. Warming is already as more innovative, automated, and digitally causing severe weather events more fre- integrated business models. Finally, the risks quently: heat waves, droughts, flooding, and costs of climate change for the region’s wildfires, and hurricanes. East and Southeast economies and societies demand more innova- Asian countries are among those likely to be tion and adoption of new technologies for both the hardest hit as the climate warms. adaptation and mitigation. 6 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

The state of innovation in The region has experienced some developing East Asia important innovation-related successes Interest in innovation among the region’s Data suggest that developing East Asia has policy makers has peaked recently with the registered some important successes with rise of digital technologies. Indeed, high- respect to innovation. Recent data on the spa- profile accomplishments by private sector tial density of patents filed under the Patent actors—in e-commerce, digital financial tech- Cooperation Treaty (PCT) of the World nology (fintech), ridesharing, and mobile Intellectual Property Organization (WIPO) app-enabled service delivery—have captured indicate a growing number of innovation the imaginations of policy makers, the media, clusters in the region, most notably in China and citizens alike. Enterprises in the digital (Bergquist, Fink, and Raffo 2017; Dutta, space, like the Chinese multinational technol- Lanvin, and Wunsch-Vincent 2019). ogy company Alibaba and the ride-hailing Looking more broadly across the region, services Grab and Go-Jek in Southeast Asia, data indicate that the region’s export-oriented have become household names. growth model has enabled most countries Although the achievements of high- to participate in more sophisticated forms performing “unicorns” are important and of manufacturing trade over time. Cross- noteworthy, realizing the economic promise country data show, for example, that most of innovation will require a broad swath of developing East Asian countries perform at firms across different sectors of the region’s or above what would be predicted from their economies to engage in innovation activities. per capita income levels with respect to both But just how well are developing East Asian high-tech imports (figure O.3, panel a) and countries performing overall on innovation? high-tech exports (figure O.3, panel b).

FIGURE O.3 Several developing East Asian countries are significant participants in the global value chains for high-tech products

a. High-tech importsa b. High-tech exportsb

China Korea, Rep. 100 Vietnam Malaysia 100 Vietnam Philippines Malaysia Singapore Philippines China 80 Singapore 80 Thailand Japan United States 60 60 Thailand Korea, Rep. Japan 40 Indonesia 40 High-tech exports (score) High-tech imports (score) Indonesia 20 20 Mongolia Mongolia Cambodia 0 Cambodia 0 6810 12 6810 12 Log GDP per capita (PPP constant 2011 international $) Log GDP per capita (PPP constant 2011 international $)

Source: World Bank elaboration, using the Global Innovation Index database (https://www.globalinnovationindex.org/analysis-indicator). Note: High-tech export and import indicators include technical products with high research and development (R&D) intensity, as defined and classified by Eurostat, the statistical office of the European Union. “Developing East Asia” refers to the 10 middle-income countries covered in this study (designated in light blue): Cambodia, China, Indonesia, Lao PDR, Malaysia, Mongolia, Myanmar, the Philippines, Thailand, and Vietnam. The figure excludes Lao PDR and Myanmar, for which no recent data exist. GDP = gross domestic product; PPP = purchasing power parity. a. The high-tech imports indicator measures high-tech imports as a percentage of total trade. b. The high-tech exports indicator is defined by high-tech exports minus re-exports as a percentage of total trade. Overview 7

Although much of the region’s participa- Licensing of foreign technologies— tion in this trade began with less-sophisticated another important input for the diffusion components and assembly, these measures and adoption of new technologies—is asso- reflect the increased adoption of global tech- ciated with higher innovation output among nologies and production processes over time firms in developing East Asian countries through FDI, creation of joint ventures, and (Iootty 2019). The region’s performance participation in trade and GVCs. For exam- regarding foreign technology licensing ple, between 2000 and 2008, the share of the is more mixed: in half of the countries, a domestic content of exports in electronics smaller share of firms obtain licenses to for- grew significantly in Malaysia and Thailand, eign technologies than would be expected as well as in industrial machinery in Indonesia given their countries’ per capita incomes and the Philippines (WTO and IDE-JETRO (figure O.4, panel b). 2011), probably as a result of FDI to pro- Data on the main input of discovery of duce locally and the participation of local new products and technologies, research suppliers. Central to the region’s outward- and development (R&D), and one key oriented manufacturing and growth strategy, proxy of invention, patents, show simi- these forms of international engagement have lar patterns. Most countries in the region represented important opportunities for tech- spend less on R&D than would be expected nology transfer and knowledge diffusion over given their per capita incomes (figure O.5, the past half century. panel a). Only three countries (China, Malaysia, and Vietnam) spend at or above expected levels. Most countries in the region perform Similarly, most developing East Asian below predicted levels on several key countries produce fewer patents than would indicators of innovation, however be expected given their per capita incomes Despite the great promise of innovation in the (figure O.5, panel b). Again, Malaysia, region—and some high-profile successes— Vietnam, and in this case, Mongolia, perform analysis of a range of key innovation indi- at or near the predicted levels. China is note- cators suggests that countries in developing worthy in that it performs significantly above East Asia still face important challenges to expectations regarding both R&D spending fostering innovation-led growth. Most of and patents granted. these countries appear to underperform on Similar patterns are seen with respect several standard indicators of innovation for to the region’s other innovation inputs and both diffusion (the adoption of existing tech- outputs, including several key areas related nologies) and discovery (the invention of new to innovation in services (despite some high- products, processes, and technologies). profile successes). One critical input for more-basic forms of innovation, such as improving the quality of Developing East Asia is converging in products and processes, is international cer- adoption lags but diverging in intensity tification, which gives firms access to other of technology use countries’ markets. International certification has been found to contribute to firm-level Despite countries’ increasing participation in productivity in several middle-income coun- high-tech value chains, new technologies do tries, including China and four Southeast not appear to be penetrating as deeply in devel- Asian countries (Cirera and Maloney 2017; oping East Asia’s economies as they could. Escribano and Guasch 2005). However, Analysis of the Cross-Country Historical all countries in developing East Asia except Adoption of Technology (CHAT) dataset on China perform below their predicted values adoption and use of primarily general pur- with respect to international certification pose technologies (Comin and Mestieri 2018) (figure O.4, panel a). indicates, on the one hand, that technology 8 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

FIGURE O.4 The share of firms with international certification is low in much of developing East Asia, and in half of the countries, fewer firms acquire licenses to foreign technology than expected given their countries’ per capita incomes

a. International certi cationa b. Foreign technology licensingb 60 40

China

30 40 Indonesia Malaysia 20 China Cambodia Mongolia Philippines 20 share of firms (%) Philippines Mongolia 10 Malaysia Vietnam Vietnam Lao PDR Cambodia Thailand Myanmar Thailand

International certification, share of firms (%) Indonesia

Myanmar Using technology licensed from a foreign firm, 0 Lao PDR 0 6 8 10 12 6 81012 Log GDP per capita, PPP (constant 2011 international $) Log GDP per capita, PPP (constant 2011 international $)

Source: World Bank elaboration, using World Bank Enterprise Survey data (most recent available years). Note: “Developing East Asia” refers to the 10 middle-income countries covered in this study (designated in light blue): Cambodia, China, Indonesia, Lao PDR, Malaysia, Mongolia, Myanmar, the Philippines, Thailand, and Vietnam. GDP = gross domestic product; PPP = purchasing power parity. a. International certification provides independent assurance that products or services comply with certain mutually recognized standards. b. Foreign technology licensing includes purchase or licensing of both patented and nonpatented technologies by firms as part of their production or organizational processes.

FIGURE O.5 Most countries in developing East Asia spend less on R&D and produce fewer patents than would be predicted by their per capita incomes

a. R&D expenditure b. Domestic patents granted

5 Korea, Rep. Korea, Rep. Japan 0 United States 4 China Singapore

Japan Malaysia 3 United States Mongolia Singapore –5 China Vietnam 2 Thailand Indonesia Malaysia Philippines

R&D expenditure, % of GDP Thailand 1 Indonesia Philippines Vietnam Cambodia –10

Log of patents per 1,000 working-age population Myanmar 0 Cambodia Lao PDR Mongolia 6 8 Myanmar 10 12 6 8 10 12 Log GDP per capita, PPP (constant 2011 international $) Log GDP per capita, PPP (constant 2011 international $)

Sources: World Bank elaboration, based on 2018 data from the World Intellectual Property Organization (WIPO) Statistics Database (https://www3.wipo.int/ipstats/index.htm) and the World Development Indicators database. Note: “Developing East Asia” refers to the 10 middle-income countries covered in this study (designated in light blue): Cambodia, China, Indonesia, Lao PDR, Malaysia, Mongolia, Myanmar, the Philippines, Thailand, and Vietnam. GDP = gross domestic product; PPP = purchasing power parity; R&D = research and development. Panel b of the figure excludes Lao PDR, for which no 2019 patent data exist. Overview 9

adoption lags between developing East Asia of innovation-related activity (figure O.7). and Organisation for Economic Co-operation The data also show China as a positive outlier and Development (OECD) countries—that in the region. Close to 60 percent of Chinese is, the time between introduction of a new firms report having a product or service inno- technology and when it is first adopted—has vation, and 20 percent license foreign tech- narrowed over time (figure O.6, panel a). nology. At the other end of the spectrum, less On the other hand, however, differences than 15 percent of firms in Myanmar and between the region and the OECD in the Thailand report having a product or service “intensity of use” of new technologies—that innovation, and a mere 5 percent have any is, how widely new technologies have been technology licensed from foreign companies. used—have increased over time (figure O.6, panel b). Innovation activity also varies widely across sectors, with less innovation in Heterogeneity of innovation services capabilities within countries, sectors, and firms Although the most salient innovations portrayed in the region’s popular press are The aggregate performance figures presented examples from services sector companies (for above mask significant heterogeneity. What example, Grab, Go-Jek, Alibaba, or Tencent), matters most for a country’s growth and pro- data from statistical sample surveys tell a ductivity performance is how rapidly technol- different story. Measured as having imple- ogy and innovation diffuse across enterprises mented a product or process innovation, within a country. Without positive spillovers services sector firms in developing East Asia from sectors that perform well relative to the (and elsewhere) appear to be significantly less rest of the economy, the contribution of inno- innovative than manufacturing firms (figure vation to overall growth is limited. Within O.8). Innovation in services is key to enabling sectors, the productivity and technological new business models and services required in divide between the leading and lagging firms the transition to Industry 4.0, but the region in developing East Asia reflects the slow diffu- is lagging behind. sion of technology. Indeed, the region shows substantial heterogeneity in the pattern of technology adoption and innovation across Most firms remain far from the and within countries, sectors, and in some technological frontier; even within cases, even within firms. This heterogeneity, firms, they vary in their use of if persistent over time, will significantly con- technology across business functions strain growth in the region. Micro evidence from the Firm-level Adoption of Technology (FAT) survey in Vietnam shows Countries across the region show that most firms remain far from the frontier significant differences in innovation in their adoption and use of new technolo- performance gies. Figure O.9 shows the most frequently used technology for different business func- Firm-level measures of innovation, based on tions, by sector, with the most sophisticated World Bank Enterprise Survey data, reveal sig- technology on top and the least sophisticated nificant heterogeneity in performance across on the bottom, as follows: the region’s countries. Firms in Indonesia, the Lao People’s Democratic Republic, Myanmar, • In manufacturing, for fabrication, and Thailand report little innovation activ- most Vietnamese firms (70 percent) ity; well under half of all firms in those coun- use operator-controlled machines, only tries indicate that they engage in any form 9 percent use computer-controlled 10 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

FIGURE O.6 Technology adoption lags in developing East Asia are converging with those of OECD countries, but the intensity of technology use is diverging

a. Adoption lag, 1700s–2000

200

150 Steel (Bessemer, open hearth)

100 Railway passengers Telegraph Tractor Cars Electric arc furnace Adoption lags in years Electricity Spindles Mail Fertilizer Telephone 50 Aviation freight Harvester Ships Synthetic ber Blast oxygen furnace Cellphones Internet Railway freight PCs 0 Trucks Aviation passengers 1800 1850 1900 1950 2000 Year of technology invention

b. Intensity of use, 1700s–2000

0 Ships

–0.5 Railway passengers Spindles Mail

–1.0 Aviation freight PCs Electricity Telegraph Railway freight Fertilizer –1.5 Electric arc furnace Cellphones Trucks Internet Tractor log intensive margin (US = 0) Aviation passengers –2.0 Steel (Bessemer, open hearth) Telephone Harvester Cars Synthetic ber

–2.5 Blast oxygen furnace 1800 1850 1900 1950 2000 Year of technology invention Fitted line, OECD Fitted line, developing East Asia

Source: World Bank, using country technology-level estimates from Comin and Mestieri 2018. Note: Adoption lag (the number of years for a technology to arrive to a country after invention) and the intensive margin, or usage intensity (how widely new technologies are adopted), are both country-specific model parameters estimated structurally using the Cross-Country Historical Adoption of Technology (CHAT) database developed by Comin and Hobijn (2004). The blue and orange lines are fitted, respectively, to Organisation for Economic Co-operation and Development (OECD) and developing East Asian countries (the sample here including Cambodia, China, Indonesia, Malaysia, Mongolia, the Philippines, Thailand, and Vietnam). The bars show the median adoption lags (panel a) or intensive margins (panel b) of the two country groups for each labeled technology. PCs = personal computers. Overview 11

FIGURE O.7 Developing East Asian countries vary widely in firm-level innovation activity

40 Share of rms (%) 20

0 China Philippines Cambodia Mongolia Vietnam Malaysia Myanmar Lao PDR Indonesia Thailand Firms that introduced a process innovation Firms that introduced a new product or service Firms that spend on R&D Firms using technology licensed from foreign companies Average innovation score

Source: World Bank calculations using latest World Bank Enterprise Survey data. Note: The innovation score captures both innovation outputs and inputs. It is calculated as the average of the likelihood that firms have a product innovation, a process innovation, positive research and development (R&D) spending, or license technology from foreign companies. FIGURE O.8 Manufacturing and services firms machines, and less than 1 percent use differ in rates of innovation, especially of new more-advanced technologies like robots, products or processes 3-D printers, or additive manufacturing (figure O.9, panel a). 60 • In retail, for inventory management, 63 percent of firms use computer databases with manual updates, 25 percent use ware- 40 house management systems with specialized software, and only 1 percent use advanced technologies such as automated storage and 20 retrieval systems (figure O.9, panel b). • In agriculture, for weeding and pest

Share of rms (%) reportingShare control, almost one-third of firms still rely innovation in past three years in past three innovation largely on manual techniques, and another 0 Manufacturing Services one-third use mechanical techniques, whereas the use of automated precision Product Process Organization Marketing agricultural techniques is almost nonexistent (figure O.9, panel c). Source: World Bank Enterprise Survey data, latest round. Note: The figure shows the average share and 95 percent confidence interval in the pooled sample, accounting for country fixed effects. The radar diagrams in figure O.10 reinforce Data include manufacturing and services sector firms covered by the the substantial heterogeneity in the use and Enterprise Surveys in all 10 middle-income countries covered in this study: Cambodia, China, Indonesia, Lao PDR, Malaysia, Mongolia, Myanmar, the sophistication of technology across firms Philippines, Thailand, and Vietnam. but also highlight the often considerable 12 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

FIGURE O.9 The intensive use of cutting-edge technology for manufacturing, retail, and agriculture remains limited in Vietnam

a. Manufacturing: use of fabrication technology Other advanced manufacturing processes (e.g. as laser, plasma sputtering, high-speed machine, e-beam, micromachining)

Additive manufacturing including rapid prototyping and 3-D printers

Robots Computer numerical controlled machine or other machines controlled by computers Machines controlled by operators without computers

Manual processes

0 20 40 60 80 Share of rms adopting technology, by type of use (%)

b. Retail: use of inventory technology

Automated storage and retrieval systems (AS/RS) Automated inventory control (CAO), vendor-managed inventory, and/ or radio-frequency identi cation Warehouse management system with specialized software

Computer databases with manual updates

Handwritten record keeping

0 10 20 30 40 50 60 70 Share of rms adopting technology, by type of use (%)

c. Agriculture: use of weeding and pest control technology Drone application in combination with remote sensing or on-site sensors (advanced precision agriculture) Fully automated variable rate application (VRA) tools in combination with soil and plant sensors (precision agriculture) Biological methods of fertilizing, weeding, or pest control

Mechanical application of herbicide, fertilizer, or pesticide

Manual application of herbicide, fertilizer, or pesticide

0 10 20 30 40 Share of rms adopting technology, by type of use (%)

Source: World Bank estimates, using the 2020 Firm-level Adoption of Technology (FAT) survey of Vietnamese firms (Cirera, Comin, Cruz, Lee, and Martins-Neto 2020). Overview 13

heterogeneity within firms. Using information Only a small share of firms engage in on a typical large firm (Firm 1) and a medium- more sophisticated innovation activities size firm (Firm 2) in the food processing sector, such as R&D the figure shows that, for the same business There is similarly considerable heterogene- functions, these two firms can be very different ity between firms when it comes to the more in the extensive margin of technology use (fig- sophisticated forms of innovation that could ure O.10, panel a), although the gap is smaller result in invention at the frontier, as evidenced when use intensity (the intensive margin) is by the high level of concentration in firms’ considered (figure O.10, panel b). It is note- R&D investments. Figure O.11 shows the worthy that the same firm (for example, Firm distribution of R&D intensity (measured by 1) can be near the technology frontier in its use R&D expenditure per full-time employee) of food storage technology but far from the in Cambodia, China, Malaysia, and the frontier in its use of input testing technology Philippines, using Israel as a benchmark. (figure O.10, panel d).

FIGURE O.10 Radar diagrams show substantial heterogeneity in technological sophistication within firms

a. General business functions, b. General business functions, food processing: extensive margin food processing: intensive margin Business administration Business administration Production Production Quality control planning Quality control planning

1 1 Payment 2 Sourcing Payment 2 Sourcing 3 3 4 4 5 5 Sales Marketing Sales Marketing Firm 1 Firm 2 Firm 1 Firm 2

c. Sector-speci c functions, d. Sector-speci c functions, food processing: extensive margin food processing: intensive margin Input testing Input testing

Food Mixing/blending/ Food Mixing/blending/ storage cooking storage cooking 1 1 2 2 3 3 4 4 Packaging 5 Antibacterial Packaging 5 Antibacterial Firm 1 Firm 2 Firm 1 Firm 2

Source: Cirera, Comin, Cruz, Lee, and Martins-Neto 2020. Note: In each radar diagram, the values 1–5 indicate relative distance from the frontier in a firm’s use of technology for a given business function (1 being the most distant and 5 representing the frontier). Firm 1 and Firm 2 are Vietnamese food-processing firms that provided data for the Firm-level Adoption of Technology (FAT) survey. 14 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

In all countries shown, the vast major- countries (figure O.11, panel b). The dis- ity of firms perform no R&D whatsoever tribution of R&D intensity among firms in (figure O.11, panel a). Only a relatively China most closely resembles the distribu- small share engage intensively in R&D tion found in Israel, whereas performance activities. Notably, among firms that among the most R&D-intensive firms in do invest in R&D, the distribution of Cambodia, Malaysia, and the Philippines R&D intensity differs significantly across still falls well below Israeli levels.

FIGURE O.11 There is significant duality in firm-level R&D investment

a. Distribution of rm-level R&D investment, selected countriesa

0.20

0.15

0.10 Probability density

0.05

0 −20 −10 0 10 20 Log R&D expenditure ($) per FT employee

b. Concentration of high R&D activities, selected countriesb 0.3

0.2

0.1 Probability density

0 5 10 15 20 Log R&D expenditure ($) per FT employee

Israel China Cambodia Malaysia Philippines

Source: World Bank Enterprise Survey data. Note: In both panels, Israel (dark solid line) represents the benchmark. FT = full-time; R&D = research and development. a. Panel a illustrates the distribution of R&D investment across all firms, by country. b. Panel b (an enlarged portion of panel a) illustrates the distribution of R&D intensity (R&D expenditure per FT employee) among firms that do invest in R&D. Overview 15

Why diffusion matters East Asian countries. Consistent with the broader literature, innovation among firms Although inventions and new technologies in the region is associated with both higher offer the possibility for large increases in pro- labor productivity and higher revenue TFP ductivity, it is the diffusion of these new tech- (de Nicola 2019). The positive relationship nologies rather than invention that ultimately between innovation and productivity holds determines the pace of economic and pro- both for firms that introduce new products ductivity growth (Hall and Khan 2003). The and those that introduce new processes. considerable heterogeneity observed in developing East Asia suggests that diffusion is not occurring at the pace and level that would be Although much of the literature focuses desirable. The approach of accelerating inno- on invention, more-basic forms of vation through the acquisition of technolo- innovation—increased diffusion and gies embedded in imports and FDI, while an adoption—also pay off important part of the region’s growth model Much of the literature has focused on inno- to date, has not induced broad diffusion of vation defined as invention, especially using new technologies and processes beyond patenting data. But it is important to empha- export-linked firms. And it will be insufficient size that innovation in the form of imitation for propelling future productivity growth. of products and processes, adoption of new In light of the significant heterogeneity in technologies, or increased product quality is firm innovation performance in the region, a also important for productivity and growth. broader innovation-based model is needed— And policies that successfully encourage one that still tries to maximize the absorp- innovations that are new to the firm or new tion from FDI and participation in GVCs but to the domestic market can have significant that also supports a critical mass of firms in returns. adopting new technologies and undertak- This advantage is clear from growth mod- ing innovation. Although it remains impor- els, as in Madsen, Islam, and Ang (2010), but tant to enable more-sophisticated firms to also from new microeconomic evidence. For undertake R&D projects and potentially to example, in one of the few studies that specif- invent at the technological frontier, it is criti- ically examines the impact of innovations that cal to support a large mass of firms to start are only new to the firm or the local market innovating. (that is, that have been invented elsewhere), Fazliog˘lu, Dalgiç, and Yereli (2019) find A large empirical literature links all positive returns, based on a panel of Turkish types of innovation to higher firm firms. This less-sophisticated type of innova- productivity tion, which helps firms to remain competitive, also pays off. The empirical literature that examines Data from the recent FAT survey in the relationship between innovation and Vietnam also indicate a positive relation- productivity—focused mostly on high-income ship between technology adoption and labor countries but also on China—indicates that productivity at the firm level. Figure O.12 innovation generally increases firm-level pro- shows the results of regressing the logarithm ductivity. (For surveys of the evidence, see of value added per worker on a technology Hall [2011] and Mohnen and Hall [2013].) index and sector dummies to control for Productivity impacts tend to be strongest for different production functions, by sector. product innovations, although this may par- Firms that use more-sophisticated technolo- tially reflect challenges associated with mea- gies for general business functions (GBF) suring other forms of innovation. such as human resource management, sup- Recent analysis of Enterprise Survey data ply chain management, and sales (the exten- provides further evidence for developing sive margin, shown in panel a) and that use 16 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

FIGURE O.12 Technology adoption brings labor productivity gains to Vietnamese firms

a. Correlation between extensive margin b. Correlation between intensive margin index and labor productivitya index and labor productivityb

10.5 12.0

10.0 11.0

9.5 10.0 9.0

9.0 8.5 Log of value added per worker Log of value added per worker 8.0 8.0 1 2 3 4 5 1 2 3 4 5 Extensive GBF tech index Intensive GBF tech index

Source: Cirera, Comin, Cruz, and Lee 2020. Note: The figures show the conditional predictions (solid lines) and 95 percent confidence levels (dashed lines) from regressing the log of labor value added per worker on a technology index and sector dummy variables. The index (from 1 to 5) measures the sophistication of technology—1 being the least sophisticated and 5 the most sophisticated technology for a set of general business functions (GBF) including business administration (human resources, accounting, and so on); production or service operations planning; sourcing, procurement, and supply chain management; marketing and product development; sales; payments; and quality control. The indexes were developed using data from the Firm-level Adoption of Technology (FAT) survey of Vietnamese firms in manufacturing, retail, and agriculture. a. The “extensive” index refers to the most sophisticated use—here, as the average among all GBFs of the most sophisticated technology used in each business function, even if marginally used. b. The “intensive” index refers to the most sophisticated use within the technologies that are used more intensively—here, as the average among all GBFs of the most sophisticated technology used more intensively for each business function (that is, the main technology used by the firm).

them more intensively (the intensive margin, What inhibits innovation? shown in panel b) tend to have higher value added per worker. Technology adoption and diffusion are Some caution is advised in interpreting determined not only by relative prices these findings in terms of causality and the but also by factors such as differential overall impact on TFP, because it is hard to returns to innovation, uncertainty distinguish empirically between the effects about demand, and differences in firm- of technology adoption and capital deepen- level capabilities ing. Nonetheless, the analysis suggests that If returns to technology adoption, and to technology upgrading is likely associated innovation more broadly, tend to be posi- with higher productivity. This is consistent tive, what is constraining firm-level diffu- with a broader micro literature that has sion, adoption, and invention in the region? found positive impacts from the adoption One prominent view of technology adop- of information and communication technol- tion and diffusion from a macro perspective ogy (ICT) (Bloom, Sadun, and Van Reenan focuses on the role of relative prices and fac- 2012; Brynjolfsson and Hitt 2003) and tor abundance. The idea is that less-developed other technologies (Kwon and Stoneman countries, because of differences in economic 1995; Mcguckin, Streitwieser, and Doms conditions and factor prices—including bar- 1998). The observed relationship is stron- riers to technology transfer, abundant labor, ger for the intensive margin (the most and a scarcity of skilled labor—will use differ- intensively used technology), as expected. ent technologies than high-income countries Overview 17

(Acemoglu and Zilibotti 2001). The implica- technology are major obstacles to technology tion is that countries in developing East Asia adoption (figure O.13). would be expected to use more labor-intensive In addition, more than 50 percent of firms, and less capital-intensive technologies than regardless of their size, report that lack of do high-income countries. The extent of this knowledge or related capabilities are a key will depend on how complementary technol- barrier to investment in new technologies. ogy and labor are (Acemoglu 2010) and how Moreover, nearly half of small and medium- successful firm-level R&D is in generating size firms and roughly one-third of large new technologies. firms report difficulty in obtaining financing Although it is unclear empirically how as a major barrier to technology adoption.3 strongly some of these macro patterns of tech- Costs of government regulations and lack of nology adoption based on factor prices affect adequate infrastructure (such as electricity or diffusion, micro evidence suggests that other internet) are also cited by firms as barriers to elements are at play in impeding successful technology adoption, albeit to a lesser extent. technology adoption, diffusion, and invention in the region. These include uncertainty and Second, firms’ innovation capabilities, as lack of information, weak firm capabilities, reflected in management quality, is weak inadequate workforce skills, lack of external Innovation requires a range of capabilities financing, and weak or misaligned country that enable firms to respond to market con- innovation policies and institutions—each of ditions, identify new technological opportu- which are discussed here, in turn. nities, develop a plan to exploit them, and

First, firms face considerable uncertainty regarding investments in technology FIGURE O.13 Lack of demand and uncertainty is the top self- Innovation is an inherently risky endeavor. reported barrier to technology adoption among Vietnamese firms Indeed, the process of technology adop- of all sizes tion is often characterized by significant uncertainty—as to the future path of the technology and its benefits—and by limited infor- Lack of demand mation about the benefits, costs, and even the and uncertainty very existence of the technology’s viability (Hall 2004). Lack of capabilities Uncertainty about demand for new products or the efficiency of new technologies Lack of nance can lead to low initial adoption of new technologies among firms (as explained theoreti- Government cally by Atkin et al. [2017]). Evidence of an regulations increased investment in quality upgrading in response to new export demand provides Other empirical support for this argument (Atkin, Khandelwal, and Osman 2017). Poor infrastructure Data from the FAT survey in Vietnam confirm that firms consider uncertainty to 0 25 50 75 100 be an important factor in the decision to Share of rms (%) adopt new technology. Over 75 percent of Small Medium Large small and medium-size firms and two-thirds of large firms surveyed indicated that uncer- Source: Cirera, Comin, Cruz, Lee, and Martins-Neto 2020. tainty about demand and doubts about Note: Firm sizes small (5–19 employees), medium (20–99), and large (100+). Data were gathered the economic benefit of investing in a new from the Firm-level Adoption of Technology (FAT) survey of Vietnamese firms in manufacturing, retail, and agriculture. 18 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

FIGURE O.14 Firms in developing East Asian countries score lower having a product, process innovation, R&D on management capabilities than firms at the global frontier project, or patent) is significantly associated with the firm’s management quality (Park and a. Average management scores in relation to per capita GDP Xuan 2020). 3.5 Data from the World Management Survey United States (WMS) indicate that firms’ management qual- Japan Germany Sweden ity in selected East Asian countries is roughly Canada France on par with what would be expected, given 3.0 Great Britain Australia 4 Mexico Poland Italy Singapore their per capita income levels. However, New Zealand Malaysia Northern Ireland Chile their management quality remains far from Turkey Ireland China Spain the global frontier (proxied by the United Vietnam Brazil Kenya Greece 2.5 India Colombia States). Compared with US firms, firms in Nigeria Nicaragua Portugal Myanmar Argentina developing East Asia are less well-managed, Ethiopia Zambia Tanzania Ghana on average and across the whole distribution

Average management practices (score) (figure O.14). 2.0 Mozambique Moreover, poor overall performance is 6 7 8 9 10 11 Log 10-year average GDP per capita, PPP driven by quality gaps that are generally 5 (current international $, billions) larger for the best firms. That is, the frontier firms in developing East Asia perform dispro- Africa East Asia South Asia Latin America Australasia Europe North America portionally worse than the frontier firms in the United States. This gap in management capabilities likely contributes to the innova- b. Distribution of management practices scores, selected countries tion gaps between the region and the global 0.8 frontier while also helping to explain some of the firm heterogeneity within countries described earlier. 0.6 Third, inadequate workforce skills impede 0.4 innovation in the region Density A range of advanced skills are important in 0.2 enabling innovation at the firm and country levels, including advanced cognitive, socio- 0 emotional, and technical skills. Such advanced 1 2 3 4 5 skills become increasingly important as firms Overall management score move from diffusion and technology adop- China Myanmar Vietnam United States tion toward the technological frontier. New analysis carried out for this study high- Source: World Management Survey (WMS), Centre for Economic Performance. lights that employees in highly innovative firms Note: The WMS scores (collected over several years) range from 1 (worst practice) to 5 (best practice) across key management practices used by organizations in different sectors. These carry out more nonroutine cognitive and inter- practices are grouped into five areas: Operations Management, Performance Monitoring, Target personal tasks and fewer manual tasks than Setting, Leadership Management, and Talent Management. GDP = gross domestic product; PPP = purchasing power parity. those in less-innovative firms (figure O.15). For this reason, highly innovative firms in then cultivate the necessary resources to do so. developing East Asia employ more (college) The acquisition or the lack of these capa- educated employees, with advanced technical bilities—and specifically, managerial capa- training, greater cognitive abilities, and stron- bilities—is fundamental to the process of ger socioemotional skills. upgrading (Sutton 2012). Firms in the region consistently report New survey evidence from China supports skills gaps as serious impediments to their this view. The degree of a firm’s innovative- operations. This is true of firms whether ness (as measured by the firm’s incidence of or not they innovate. Nonetheless, the Overview 19

FIGURE O.15 Employees in more-innovative firms in China and Vietnam have jobs that are more intensive in nonroutine cognitive analytical and interpersonal tasks

a. Task intensity in China, by type, 2018a b. Task intensity in Vietnam, by type, 2019b

0.3 0.3

0.2 0.2

0.1 0.1

0 0 efficient Index

–0.1 Co –0.1

–0.2 –0.2

–0.3 –0.3 Nonroutine Nonroutine Routine task Nonroutine Nonroutine Nonroutine Routine cognitive cognitive intensity cognitive cognitive manual manual analytical interpersonal analytical interpersonal

Low-innovation firms Medium-innovation firms High-innovation firms

Sources: World Bank elaboration, based on Park and Xuan 2020 and Miyamoto and Sarzosa 2020, using, respectively, the 2018 China Employer-Employee Survey (CEES) and the 2019 Enterprise Survey on Innovation and Skills (ESIS) for Vietnam. Note: Firms are categorized by “innovation intensity,” measured by the number of innovation activities undertaken, as captured in the respective surveys. Scaled from 0–5, low-, medium-, and high-innovation are defined, respectively, as those undertaking 0–1, 2–3, and 4–5 innovation activities. The Vietnam analysis does not include an aggregated measure of “routine task intensity”; therefore, panel b shows instead an individual measure of “routine manual” tasks. No information was included in either panel on routine cognitive tasks because none of the related regression coefficients was statistically significant. a. The CEES collected responses of 2,001 manufacturing firms and 16,379 workers from five Chinese provinces: Guangdong, Jiangsu, Jilin, Hubei, and Sichuan. b. The ESIS collected responses from 201 manufacturing and information and communication technology services firms and 849 staff in five Vietnamese provinces: Hanoi, Bac Ninh, Da Nang, Ho Chi Minh City, and Binh Duong. challenges that innovative firms face in Assessment, or PISA)—considerably below finding suitably skilled staff are consider- what would be predicted based on their coun- able. Over 50 percent of innovating firms tries’ income levels. in Indonesia, Malaysia, Myanmar, the Weak development of reading, math, Philippines, Thailand, and Vietnam cite and science capabilities—the “foundational a lack of managerial and leadership skills skills”—represents an important impediment as a challenge when hiring new workers to the development of the more advanced (figure O.16). And more than half of all cognitive, technical, and socioemotional skills innovative firms in at least three of those six needed to support innovation. These skills- countries cite the scarcity of interpersonal related constraints are relevant to the process and communication skills, foreign language of diffusion and adoption of existing technol- skills, computer and information technologies, but they are all the more severe when it ogy (IT) skills, or technical (non-IT) skills as comes to invention. critical challenges when it comes to hiring. A fundamental challenge in this context Fourth, a lack of diversified sources is that most of the region’s countries are of external finance constrains firm still struggling to ensure that their students innovation develop basic reading, math, and science Access to external finance—and to a suitable skills. Indeed, students in several countries range of financial instruments—is critical to (Indonesia, Malaysia, the Philippines, and enabling firm innovation. How firms finance Thailand) score poorly on international edu- their investments and operations influences cation assessment tests given to eighth grad- both the decision to innovate and the quality ers (the Programme for International Student of that innovation. 20 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

FIGURE O.16 Most innovative firms in developing East Asia report difficulties in hiring workers with adequate skills

a. Indonesia, 2015 b. Malaysia, 2015

Foreign language skills Foreign language skills

Managerial and leadership skills Managerial and leadership skills Technical (other than IT), Writing skills vocational, or job-speci c skills Technical (other than IT), Computer or general IT skills vocational, or job-speci c skills Work ethic and commitment Computer or general IT skills Interpersonal and communication Writing skills rms (%) citing diculty, by skill rms (%) citing diculty, rms (%) citing diculty, by skill rms (%) citing diculty, skills Fi Fi Interpersonal and communication Work ethic and commitment skills 0 20 40 60 80 100 020406080100 Percent Percent

c. Myanmar, 2016 d. Philippines, 2015

Managerial and leadership skills Managerial and leadership skills

Foreign language skills Work ethic and commitment Interpersonal and communication Computer or general IT skills skills Interpersonal and communication Technical (other than IT), skills vocational, or job-speci c skills Writing skills Foreign language skills Technical (other than IT), rms (%) citing diculty, by skill rms (%) citing diculty, rms (%) citing diculty, by skill rms (%) citing diculty, Computer or general IT skills Fi Fi vocational, or job-speci c skills Work ethic and commitment Writing skills

0 20 40 60 80 100 0204060 80 100 Percent Percent

e. Thailand, 2016 f. Vietnam, 2015

Foreign language skills Managerial and leadership skills

Managerial and leadership skills Work ethic and commitment Technical (other than IT), Interpersonal and communication vocational, or job-speci c skills skills Technical (other than IT), Computer or general IT skills vocational, or job-speci c skills Work ethic and commitment Foreign language skills Interpersonaland communication Computer or general IT skills rms (%) citing diculty, by skill rms (%) citing diculty, skills by skill rms (%) citing diculty, Fi Fi Writing skills Writing skills

0204060 80 100 0204060 80 100 Percent Percent

Source: World Bank calculations, based on World Bank Enterprise Surveys. Note: “Innovative” firms are defined as those that introduced new or significantly improved products or services (product), or adopted new production methods (process), during the past three fiscal years. IT = information technology. Overview 21

As noted earlier, about 16 percent of characteristics regarding maturity, cost, and Vietnamese firms surveyed report difficulty ancillary services. And these characteristics, in obtaining loans as the primary barrier to in turn, help reduce market frictions associ- new technology adoption. More broadly, an ated with asymmetric information, cash flow analysis of Enterprise Survey data from devel- uncertainty, and an extended time lag between oping East Asian countries shows that firms investment and returns. Such frictions are that make use of external finance, other than especially serious among firms undertaking from banks, are more likely to innovate and long-gestation R&D projects commonly asso- to engage in more innovation activities (Mare, ciated with invention. de Nicola, and Liriano forthcoming). A recent Although financial sectors in develop- study of innovation in China also shows that ing East Asia have become more diversified financial constraints affect innovation quality. in recent years, most remain heavily bank- Firms that are financially constrained are less ing based, having neither the depth nor the likely to invest in large innovation projects breadth to effectively support innovation-led with the potential to transform productivity, growth (figure O.17). Moreover, financial focusing rather on making marginal improve- markets remain accessible mostly to large ments to existing products (Cao 2020). firms (Abraham, Cortina, and Schmukler Well-developed, deep financial markets 2019). Key challenges, therefore, involve allow firms to take advantage of diverse the continued deepening of countries’ finan- financial instruments, supporting both cial markets to ensure greater diversity of increased quantity and quality of innovation. sources of finance for innovation—especially A diversity of sources is important because invention—and the enabling of greater access different financial instruments have different to small and medium-size enterprises.

FIGURE O.17 Banks remain the dominant source of finance to firms in most of developing East Asia (except China)

140

120

100

Share of GDP (%) Share 40

0 2006–11 2012–17 2000–05 2006–11 2012–17 2000–05 2006–11 2012–17 2000–05 2006–11 2012–17 2000–05 2006–11 2012–17 2000–05 2006–11 2012–17 2000–05 2006–11 2012–17 2000–05 China Indonesia Malaysia Mongolia Philippines Thailand Vietnam

Banks Equity Corporate bonds

Sources: The World Bank’s Global Financial Development and FinDebt databases. Note: The graph reports averages over the three periods. “Equity” refers to stock market capitalization, “corporate bonds” to the amount outstanding of domestic bonds issued by private entities in industries other than finance, and “banks” to the outstanding amount of private credit granted by domestic banks. “Developing East Asia” refers to the 10 middle-income countries covered in this study: Cambodia, China, Indonesia, Lao PDR, Malaysia, Mongolia, Myanmar, the Philippines, Thailand, and Vietnam. The figure excludes Cambodia, Lao PDR, and Myanmar because of unavailability of data. For Mongolia, data on the corporate bond issuance are not available before 2011. 22 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

Fifth, countries’ innovation policies and numerous shortcomings in the design and institutions are often not fit for purpose implementation of innovation policies. Among Policies and agencies in developing East the most important design shortcomings are: Asia are not well positioned to enable the • A lack of adequate economic justification process of increasing innovation and tech- for public policy; nological catch-up. Among the factors that • The absence of a logical framework to dull the impact of policies on innovation, guide the design and implementation of countries’ current policy mixes are not well policy interventions; and oriented toward building firms’ capabilities • A lack of monitoring and evaluation for innovation or accelerating technology (M&E) mechanisms for most policy diffusion and adoption; nor are they set up instruments. to enable innovation in services. In addition, weak governance and institutional capac- The outcome of this lack of good prac- ity among innovation agencies and public tices is that both the design and imple- research organizations (PROs) often impede mentation of policies fall below potential. their ability to address countries’ most press- The overall quality of the policy mix was ing innovation challenges. also found to be hampered by a lack of coordination across government agencies Innovation policies in the region neither involved in innovation. Such coordination focus on key bottlenecks nor prioritize is critical because innovation needs and technology adoption or building challenges cut across sectors, ministries, innovation capabilities and agencies. Given the lagging innovation performance of Innovation policy making is also hampered most countries in the region, it would make by weak institutions sense for their policies to emphasize building Agencies supporting innovation in the firms’ basic capabilities and to prioritize sup- region use outdated governance models, and port for technology adoption and diffusion. together with the lack of coordination, this An in-depth review of the mix of innovation undermines the coherence of policies across policies in Indonesia, the Philippines, and countries’ innovation systems. Some salient Vietnam suggests that their policies do not features of high-income countries’ experience support these objectives, however. To differ- underline the importance of (a) having a clear ent degrees across these countries, innova- strategy addressing market failures, (b) hiring tion policies are poorly aligned with the most capable staff, (c) instituting effective gover- pressing innovation challenges. nance structures, and (d) instituting robust Two important policy gaps stand out: M&E frameworks. (a) the lack of support for technology adop- Adopting these good practices in inno- tion and diffusion, and (b) a virtual absence vation agencies in the region would help of support for innovation in services sectors. improve both policy design and imple- This misalignment of policies suggests, among mentation and should be a top priority. other things, that many countries in the Creating and empowering a dedicated region have been poorly equipped to respond innovation agency to take a high-level quickly to the technological challenges that view of policy and to coordinate could the COVID-19 crisis has created, whether in be one way of improving and profession- terms of digitalization or the creation of more alizing policy making—although without flexible production systems. due care and an appropriate mandate, there is a risk that such an agency could Most countries’ capacity to implement fail in its mission, resulting in continued innovation policies remains relatively weak lack of coordination along with addi- Detailed benchmarking exercises carried out in tional fragmentation and competition Indonesia, the Philippines, and Vietnam reveal for resources. Overview 23

Inadequate governance structures and Spurring innovation in a lack of mission orientation constrain developing East Asia: the contribution of public research to Directions for policy innovation A new survey of PROs and research centers To spur innovation more effectively in devel- in Malaysia, the Philippines, and Vietnam oping East Asia—both diffusion of existing shows that those governments have strength- technologies and invention at the frontier— ened their national research capacity, increas- and to better keep pace with the wave of new ing their investments in supporting PROs technologies, policy makers in the region need to create opportunities for new knowledge to invest in building firms’ innovation capa- generation and innovation-based competen- bilities. This approach was effectively used cies. In Malaysia and to a lesser extent in by now high-income countries in East Asia, the Philippines and Vietnam, the number of such as Japan, Korea, and Singapore, which researchers in the public sector has grown accomplished rapid technological conver- rapidly during the past decade. This has been gence by focusing on policies that addressed accompanied by a significant increase in their innovation capabilities gaps (Cirera and publication activity, especially in Malaysia, Maloney 2017). and an upsurge in patenting by universities But what does such an approach look and PROs. like in practice? How should policy mak- However, the results of these efforts and ers deal with the substantial heterogeneity the impact of PROs on innovation and the in their countries’ innovation capabilities? economy are still far from clear. The surveys To strengthen innovation policies and spur showed that, with a few noteworthy excep- innovation-led growth by addressing the tions, PROs and university research depart- innovation inhibitors described above, coun- ments develop few industry-science links tries can take several key steps: (including knowledge links and human capi- • Reorient policy objectives in a graduated tal interactions). Indeed, technology transfer manner (the “capabilities escalator”) to activities are still embryonic and mainly con- reduce uncertainty and information prob- centrated among a few organizations. lems by removing current biases against The impact of PROs and research centers diffusion; building management and inno- could be leveraged by addressing governance vation capabilities; and including a focus and funding issues, along with inconsistencies on services sector innovation in national regulatory frameworks governing • Strengthen complementary factors—skills public research systems. Specifically, inad- and finance—for innovation equate governance is related to: • Reform innovation institutions and agen- • Low levels of autonomy; cies and strengthen their capacity. • A lack of links between institutional funding policies and performance measure- ments; and Reorient policy objectives and remove • Inadequacy of academic incentives, which policy biases against adoption and deters technology transfer activities, with services sector innovation often unclear mechanisms for sharing Effectively fostering innovation, both intellectual property. diffusion and invention, requires a The main factor that dissuades research- graduated approach of moving firms ers from engaging in technology transfer toward the frontier—one that recognizes and entrepreneurial activities, however, is heterogeneity in the capacity to innovate the heavy weight that scientific publication Cirera and Maloney (2017) propose assess- (that is, the number of published articles) still ing the adequacy of policies and institu- receives among the criteria for researchers’ tions through the lens of the “capabilities career advancement and salary enhancement. escalator”—to reflect the capacity of firms 24 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

and country systems to absorb and use the goal of policy should be to enable inven- knowledge (figure O.18). tion by supporting more-complex, long-term Production capabilities. On the lowest R&D projects. At this stage, countries also step of the escalator, firms have productive, require adequate intellectual property protec- but few technological, capabilities. Policies tion and will benefit from significant collabo- should focus on building technological capa- ration between industry and universities or bilities by addressing the uncertainty and crit- other knowledge providers. ical information required for adoption (for Figure O.18 illustrates several sets of pol- example, through management extension and icy instruments, each corresponding to a dif- national quality infrastructure), improving ferent level of the capabilities escalator. skills, and supporting improvements in management quality. Where the business climate Addressing heterogeneity in innovation and competition are weak, policies should capabilities requires that governments focus on creating an environment condu- support diffusion and adoption as well cive to investment and knowledge diffusion, as invention, prioritizing policies and where firms compete, have access to com- allocating resources consistent with petitive inputs, and can maximize knowledge capabilities spillovers through FDI and trade. The framework presented here does not imply Technology adoption capabilities. In coun- that only one type of policy applies to each tries where some firms have technological country. It does mean, however, that the poli- capabilities, but few have R&D and invention cies and public resources for innovation should capabilities—the next step on the escalator— be well aligned with the capacities and needs policies should focus on expanding and of the private sector. Thus, countries with rela- strengthening technological capabilities while tively low innovation capabilities—typically also supporting more firms in implementing the region’s lower-middle-income countries— R&D projects oriented toward invention. are best advised to prioritize the adoption and Invention capabilities. In countries where diffusion of existing technologies. As innova- firms have more sophisticated capabilities, tion capabilities rise, the mix of policies can

FIGURE O.18 Appropriate policy instruments to foster innovation differ depending on the level of innovation capabilities

Long-term R&D programs Direct and indirect support to R&D Invention capabilities Collaborative projects Precommercial procurement Risk nance

Technology extension and technology centers R&D grants Technology adoption capabilities Grants to industry-university collaboration Accelerators and other infrastructure Upgrading and export quality support

Management extension Vouchers to collaborate Production capabilities Skills development National quality infrastructure (NQI) Incubation

Source: Adapted from Cirera and Maloney 2017. ©World Bank. Further permission required for reuse. Note: R&D = research and development. Overview 25

shift, progressively focusing on the more tech- outputs index that captures the quality of nically advanced needs of leading firms. knowledge, technology, and creative outputs Notably, even East Asia’s high-income, of the economy. high-capacity countries such as Japan, Korea, As expected, the relationship is positive. and Singapore—as well as Canada and the Most countries in the upper-right portion of United States—offer support for technology the figure, with the exception of China, are adoption as well as invention, with different high-income countries, whereas the coun- sets of policies to encourage both dimensions tries in the lower-left portion of the figure of innovation. At any level of capabilities, the are low- and middle-income countries. The point is not to focus policies only on either figure shows three clusters that, with a few adoption or invention but rather to allocate exceptions, correspond to the levels of capa- more resources in a way commensurate with bilities depicted in the “capabilities escalator” innovation capabilities. (figure O.18). Figure O.19 provides an approximation of where the countries of developing East Asia Policy priorities require adjustment may be in terms of innovation capabilities. over time as innovation capabilities are The scatterplot uses Global Innovation Index developed data on innovation inputs (measuring infra- Climbing the capabilities escalator is a structure, institutions, R&D, research, and dynamic process and hence requires adjust- human capital quality) and an innovation ing priorities over time. High-innovation

FIGURE O.19 Developing East Asian countries occupy three distinct clusters with respect to innovation capabilities

60 Korea, Rep.

China Japan Vietnam Singapore 40 Thailand Mongolia Malaysia

Indonesia Cambodia

Innovation output subindex score 20

Philippines

0 20 40 60 80 Innovation input subindex score

Cluster 1 Cluster 2 Cluster 3

Source: World Bank elaboration from Global Innovation Index (GII) data (https://www.globalinnovationindex.org/). Note: The “innovation input” subindex scores aspects such as infrastructure, institutions, research and development (R&D), and human capital quality. The “innovation output” subindex scores the quality of knowledge, technology, and creative outputs of the economy. Among the 10 developing East Asia countries in this study (Cambodia, China, Indonesia, Lao PDR, Malaysia, Mongolia, Myanmar, the Philippines, Thailand, and Vietnam), Lao PDR and Myanmar are excluded for lack of data. 26 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

countries in East Asia, such as Korea and Removing this bias requires action on two Singapore, periodically adjusted their policy fronts. First, it is important to go beyond mixes over time to achieve convergence to the instruments that traditionally support services technological frontier. firms, such as accelerators, and to reach out Korea’s journey has two important les- to services and retail firms with other innova- sons for developing East Asia: First, the tion instruments such as matching grants for country has pursued an overarching objec- innovation projects or digitalization. Second, tive and focus throughout the period on it is necessary to expand the scope of innova- the importance of developing technological tion activities to include design—a significant capabilities. Second, it has had an evolv- component of R&D (Kox and Rubalcaba ing prioritization of policies grounded in its 2007) in manufacturing but also in services— evolving innovation and technological capa- and to strengthen firms’ digital capabilities. bilities. Policy priorities were updated over Services sectors are extremely diverse, and time, reflecting changing challenges—from innovation takes different forms with differ- prioritizing the building of basic innovation ent priorities across subsectors. For example, capabilities in the 1960s and 1970s; to maxi- digital and AI elements are more important mizing links to GVCs, FDI, and entry into in routine services, whereas design, business export markets in the 1980s; to a significant models, and delivery are more important focus on R&D and patenting in the 2000s; in knowledge-intensive services (Salter and and to technological leadership in selected Tether 2006). sectors in the 2000s. Recognizing these differences and designing adequate policies will thus be critical A current priority includes removing policy to enabling innovation in this increasingly biases against services important sector of the economy. This can Traditionally, innovation and technology be seen in the United Kingdom, for example, development have been seen as primar- where the government’s innovation policy has ily processes driven by the manufacturing supported the growth of the creative indus- and agriculture sectors. Networks of PROs try, a sector that contributes ₤90 billion and performing R&D have been established 2 million jobs to the UK economy. throughout the region and globally in narrowly defined areas of manufacturing and Develop key complementary factors: agriculture. skills and finance The reality, however, is that innovation in services is increasingly important for The ability of firms to innovate depends on competitiveness in manufacturing, for the multiple factors that fall outside the realm of strengthening of GVCs, and also for ser- innovation policy, strictly defined. These fac- vices themselves, which employ the larg- tors include the availability of a sufficiently est share of people in all countries. For skilled workforce and adequate financing example, business internationalization to support firms’—often risky—innovation depends on transport, logistics, and com- activities. munication networks. Innovations in these services are thus critical to facilitating inte- Building strong workforce skills is critical gration of local firms into global networks. to fostering innovation Improvements in digital infrastructure, As noted earlier, policy makers face a dual digital networks, and platforms are also challenge of ensuring that their populations enabling the proliferation of innovative develop the necessary foundational skills services firms in the region. Yet, innova- while also building the types of advanced tion policy rarely supports innovation in skills needed to enable innovation. To meet services. this dual challenge, it will be important for Overview 27

policy makers to act on several fronts, as education in much of the region. Establishing described below. opportunities for continuous skills develop- Strengthen students’ foundational skills ment will support skills upgrading in the cur- by improving basic education . Lessons rent workforce. from high-performing education systems in Continuous skills development—or East Asia and beyond suggest that building lifelong learning—systems for adult workers stronger basic skills will require strengthen- are necessary to support skills upgrading in ing the conditions for learning; improving the face of rapid technological change. There teacher preparation and the quality of teach- may be scope for incentivizing on-the-job ing; ensuring adequate public spending for training, which, evidence suggests, contributes basic education; increasing children’s readi- to firm-level innovation activity. Singapore, ness to learn, including through early child- for example, has instituted a promising hood education and development services; system of individual training accounts to pro- and undertaking regular learning assessments mote upgrading of people’s workforce skills to diagnose challenges and inform improve (Kataoka and Alejo 2019). Technical and ments (World Bank 2018). vocational education and training (TVET) Lay the foundation for advanced cognitive programs can also play a role—although, and socioemotional skills early in the educa- to be effective, such programs must closely tion life cycle . Even where countries have rec- reflect private sector demand to ensure their ognized the importance of cultivating more market relevance. critical thinking, creativity, problem solving, and the ability to work effectively in teams, Strengthening finance can enable there remains a need to institutionalize the innovation development of advanced cognitive and socio- As discussed above, access to both exter- emotional skills into standard school curricula nal finance and a suitable range of finan- and extracurricular programs. Strong innova- cial instruments are critical to enabling tion performers in the region—Japan, Korea, innovation at the firm level. To strengthen and Singapore—have all revised their curri- finance for innovation, the region’s coun- cula to include an emphasis on higher-order tries should implement policies in three dif- cognitive and socioemotional skills develop- ferent areas: (a) developing well-functioning ment (Kataoka and Alejo 2019). Recent stud- capital markets, (b) promoting venture capi- ies of skills formation emphasize that building tal markets, and (c) broadening the range of strong cognitive and socioemotional skills financial instruments available to innovating is best begun early in the learning life cycle firms through the banking sector. (Arias, Evans, and Santos 2019; Cunningham Developing deep, well-functioning capi- and Villaseñor 2016). tal markets . The development of deep capi- Strengthen technical skills through tal markets is critical to ensuring alternative improved access to and quality of science, sources of external capital to innovative firms technology, engineering, and mathematics at different stages of a firm’s life cycle. Some (STEM) education . The demand for technical countries in the region have already moved in skills to innovate is quite diverse across firms. this direction by introducing capital market For firms focused on diffusion and adoption reforms targeted to increasing the investor of existing technologies, basic digital literacy base; improving financial market infrastruc- and the capacity to use general purpose tech- ture (for example, introducing a capital mar- nologies and existing software applications ket data warehouse system); and enhancing may be sufficient. As firms move toward the investor protection (Abraham, Cortina, and technical frontier, more sophisticated techni- Schmukler 2019). Where countries have cal skills are required. Efforts are needed to made progress in deepening capital markets, improve access to and the quality of technical the main beneficiaries to date have been 28 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

relatively larger firms. Innovative instruments Reform innovation institutions and are still required to improve capital market agencies and strengthen their capacity access to small and medium-size firms, however (Mason and Shetty 2019). Investing in institutional capacity is critical Promoting the development of venture for more effective innovation policies To date, discussions of innovation and tech- capital markets . Three broad, complemen- nology policies have commonly ignored tary sets of measures are important for the countries’ capacities to effectively design development of successful venture capital and implement innovation policy, but these markets: First, the supply side of the market capacities are critical for the effectiveness of can be enhanced by enabling domestic invest- interventions. An initial review of innovation ment and attracting foreign capital, such as agencies indicates that some developing East through clear bankruptcy rules and trans- Asian countries lag in the use of best practices parent accounting standards. Second, the in public management for innovation policy. demand for risk finance can be stimulated by Going forward, it will be critical to invest fostering an active entrepreneurial and inno- more in policy-making capacity. Countries vation ecosystem. And third, governments need to recruit capable staff and provide ade- can provide support for all market players by quate training on innovation policy and pub- strengthening the institutional and regulatory lic management, and to ensure that managers framework for venture capital, as well as by have adequate digital infrastructure to moni- investing directly or through public-private tor beneficiaries and register innovation proj- partnerships (for example, the successful ect outcomes. Yozma program in Israel). Careful design of public intervention is important to ensure that programs are effectively addressing mar- More professionalized innovation agencies ket failures and do not crowd out investment and increased interagency coordination are from the private sector. essential Agencies supporting innovation policy in Leveraging existing bank-firm relation- the region use outdated governance models ships . Governments in the region could also and lack coordination across entities, which channel financing to firms through the bank- undermines policy alignment. Innovation ing sector, exploiting existing bank-firm rela- policy requires coordination among agen- tionships. Well-designed government lending cies or ministries because of its cross-cutting programs can help to align debtors’ and nature. The current lack of coordination creditors’ incentives, lessening moral hazard results in significant fragmentation in effort, problems (Cirera, Frias, Hill, and Li 2020; along with policies that are poorly designed SQW 2019). This approach is not without and executed. risk, however, because identifying, targeting, In addition to improving the innovation and monitoring potential innovation projects agencies’ public management capabilities, may be difficult and costly for government governments in the region need to (a) ensure agencies. Credit guarantee schemes may be better coordination of the different ministries more efficient because they make use of exist- and institutions in charge of innovation policy, ing lending products, allowing banks to select and (b) adopt new agency models that enable projects and maintain the incentives to moni- recruitment of sufficient talent and profes- tor borrowers’ behavior. An example of such sionalized services. There is no single model a scheme is the Korea Technology Finance for coordination, and each country needs to Corporation (KOTEC), which has been suc- find its own approach. Nonetheless, coordi- cessfully providing loan guarantees to small nation is needed to ensure more integrated, and young firms in high-tech sectors. focused, and effective innovation policy. Overview 29

Strengthening the governance and incentive This report has argued that the region’s structures of PROs and research centers growth performance is under threat if coun- would help to maximize their contribution tries do not transform their development to innovation and technology diffusion model to one in which innovation is at The impacts of government efforts to the forefront. The impacts of the COVID- strengthen the national research capacity and 19 pandemic have been severe, and as the of investments in science and technology in region focuses on the recovery, it is an oppor- public research institutions are still unclear. tune moment to accelerate pending reforms Public institutions and university research to accelerate the process of technological departments engage little in industry-science catch-up. collaboration (including knowledge links and Although this transformation would personnel exchanges). Moreover, technol- be important under all circumstances, the ogy transfer activities remain embryonic and COVID-19 pandemic has served to highlight concentrated among a small number of orga- the urgency of reform. The pandemic is likely nizations. Maximizing the contributions of to tighten the constraints on innovation iden- these research institutions to innovation will tified in this report: limited capabilities of require reforms in four key areas: firms, scarcity of human capital and finance, and uncertainty about demand and returns to • Improving governance, for example, innovation. Moreover, the pandemic is raising increasing autonomy, clarifying legal uncertainty and may deepen the international mandates for technology transfer, and divisions that were already simmering before strengthening links between institutional the outbreak. Restrictions on trade, invest- funding and performance; and disseminat- ment, and the mobility of people can hurt ing good public management practices and not just the flow of existing knowledge but strategic planning also the creation of new knowledge through • Improving academic incentives for international collaboration. Similarly, mutual research-industry collaboration and tech- suspicion can divide the digital infrastructure nology transfer and curb the digital flows that are vital for all • Adopting mission-oriented policies to creative activity today. address specific innovation challenges fac- To accelerate progress in the face of ing the region (such as COVID-19 and cli- challenges, countries in developing East mate change) Asia must update their objectives and give • Incentivizing PROs to enhance their greater priority to better innovation poli- impact on firm innovation and productiv- cies. It is critical to focus on technological ity through technology extension; upgrad- diffusion and to incentivize more firms to ing of support services (including for undertake innovation. This process requires new technology-based entrepreneurship); stronger regulatory frameworks as well as licensing of new technologies to small and policies that are aligned with the techno- medium enterprises (SMEs); and support logical capabilities of the private sector in to start-ups. each country. Beyond domestic policy, East Asia must continue to deepen its tradition Final remarks of international openness, which could also induce openness in other parts of the world, Developing East Asia has achieved unprec- and help sustain the flows of ideas, trade, edented growth in the past several decades investment, and people that facilitate the that has raised incomes broadly and lifted creation and diffusion of knowledge. The hundreds of millions of people out of poverty. time for action is now. 30 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

Notes and Adaptability. Africa Development Forum series. Washington, DC: World Bank. 1. The terms “East Asia” and “developing East Atkin, David, Azam Chaudhry, Shamyla Chaudry, Asia” will be used throughout the report. Amit K. Khandelwal, and Eric Verhoogen. For convenience, unless otherwise specified, 2017. “Organizational Barriers to Technology these terms refer to the 10 middle-income Adoption: Evidence from Soccer-Ball Producers countries covered in this study: Cambodia, in Pakistan.” Quarterly Journal of Economics China, Indonesia, the Lao People’s 132 (3): 1101–64. Democratic Republic, Malaysia, Mongolia, Atkin, David, Amit K. Khandelwal, and Myanmar, the Philippines, Thailand, and Adam Osman. 2017. “Exporting and Firm Vietnam. Performance: Evidence from a Randomized 2. For more information on innovative responses Experiment.” Quarterly Journal of Economics to COVID-19, see the Coronavirus Innovation 132 (2): 551–615. Map (https://coronavirus.startupblink.com/); Bergquist, Kyle, Carsten Fink, and Julio Raffo. the World Economic Forum COVID Action 2017. “Identifying and Ranking the World’s Platform (for example, Chandran [2020], at Largest Clusters of Inventive Activity.” https://www.weforum.org/agenda/2020/03 Economic Research Working Paper No. 34, /asia-technology-coronavirus-covid19 World Intellectual Property Organization -solutions/); and Huang, Sun, and Sui (2020). (WIPO), Geneva. 3. Challenges associated with obtaining external Bloom, Nicholas, Raffaella Sadun, and John Van finance for innovation are discussed in more Reenen. 2012. “Americans Do IT Better: detail later in this Overview. US Multinationals and the Productivity 4. The WMS (https://worldmanagementsurvey Miracle.” American Economic Review 102 .org/), operated by the Centre for Economic (1): 167–201. Performance of the London School of Brynjolfsson, Erik, and Lorin M. Hitt. 2003 Economics and Political Science, is conducted “Computing Productivity: Firm-Level through in-depth interviews of over 20,000 Evidence.” Review of Economics and Statistics firms in 35 countries. Management practices, 85 (4): 793–808. as measured in the WMS, capture several Cao, Yu. 2020. “Financial Constraints, Innovation dimensions, including firms’ practices in target Quality, and Growth.” Doctoral dissertation, setting, monitoring, and human resource University of Southern California, Los Angeles. management. The WMS is not specific to Chandran, Rina. 2020. “Here’s How Asia Is innovation, but it is a proxy for firms’ overall Using Tech to Tackle COVID-19.” Reuters capabilities. article on the World Economic Forum COVID 5. These differences are statistically significant Action Platform, March 18. https://www at 95% confidence level and hold true for .weforum.org/agenda/2020/03/asia-technology different quantiles of the overall WMS -coronavirus-covid19-solutions/. management score distribution. Cirera, Xavier, Diego A. Comin, Marcio Cruz, and Kyung Min Lee. 2020. “Technology Within References and Across Firms.” NBER Working Paper No. 28080, National Bureau of Economic Abraham, Facundo, Juan J. Cortina, and Sergio Research, Cambridge, MA. L. Schmukler. 2019. “The Rise of Domestic Cirera, Xavier, Diego A. Comin, Marcio Cruz, Capital Markets for Corporate Financing.” Kyung Min Lee, and Antonio Soares Martins- Policy Research Working Paper 8844, Neto. 2020. “Firm-Level Technology Adoption World Bank, Washington, DC. in Vietnam.” Unpublished manuscript, Acemoglu, Daron. 2010. “When Does Labor World Bank, Washington, DC. Scarcity Encourage Innovation?” Journal of Cirera, Xavier, Jaime Frias, Justin Hill, and Political Economy 118 (6): 1037–78. Yanchao Li. 2020. A Practitioner’s Guide to Acemoglu, Daron, and Fabrizio Zilibotti. 2001. Innovation Policy: Instruments to Build Firm “Productivity Differences.” Quarterly Journal Capabilities and Accelerate Technological of Economics 116 (2): 563–606. Catch-Up in Developing Countries. Arias, Omar, David Evans, and Indhira Santos. 2019. Washington, DC: World Bank. The Skills Balancing Act in Sub-Saharan Africa: Cirera, Xavier, and William F. Maloney. 2017. Investing in Skills for Productivity, Inclusivity, The Innovation Paradox: Developing-Country Overview 31

Capabilities and the Unrealized Promise of Griliches, Zvi. 1998. R&D and Productivity: Technological Catch-Up. Washington, DC: The Econometric Evidence. Chicago: Chicago World Bank. University Press. Comin, Diego, and Bart Hobijn. 2004. “Cross- Hall, Bronwyn H. 2004. “Innovation and Country Technology Adoption: Making the Diffusion.” NBER Working Paper No. 10212, Theories Face the Facts.” Journal of Monetary National Bureau of Economic Research, Economics, 51 (1): 39–83. Cambridge, MA. Comin, Diego, and Bart Hobijn. 2010. “An Hall, Bronwyn H. 2011. “Innovation and Exploration of Technology Diffusion.” Productivity.” NBER Working Paper No. American Economic Review 100 (5): 2031–59. 17178, National Bureau of Economic Comin, Diego, and Martí Mestieri. 2018. “If Research, Cambridge, MA. Technology Has Arrived Everywhere, Why Hall, Bronwyn H., and Beethika Khan. 2003. Has Income Diverged?” American Economic “Adoption of New Technology.” NBER Journal: Macroeconomics 10 (3): 137–78. Working Paper No. 9730, National Bureau of Cunningham, Wendy V., and Paula Villaseñor. Economic Research, Cambridge, MA. 2016. “Employer Voices, Employer Demands, Huang, Yasheng, Meicen Sun, and Yuze Sui. 2020. and Implications for Public Skills Development “How Digital Contact Tracing Slowed Covid-19 Policy Connecting the Labor and Education in East Asia.” Harvard Business Review, April Sectors.” World Bank Research Observer 31 15. https://hbr.org/2020/04/how-digital-contact (1):102–34. -tracing-slowed-covid-19-in-east-asia. de Nicola, Francesca. 2019. “Assessing How Iootty, Mariana. 2019. “Assessing Innovation Returns to Innovation Vary Depending Patterns and Constraints in Developing East on the Business Environment and Firms’ Asia: An Introductory Analysis.” Policy Characteristics.” Background paper for this Research Working Paper 8706, World Bank, study, World Bank, Washington, DC. Washington, DC. Dutta, Soumitra, Bruno Lanvin, and Sacha Kataoka, Sachiko, and Ana Alejo. 2019. “Skills for Wunsch-Vincent, eds. 2019. Global Innovation Innovation.” Background paper for this report, Index 2019: Creating Healthy Lives—The World Bank, Manila, Philippines. Future of Medical Innovation. 12th ed. Ithaca, Kox, Henk, and Luis Rubalcaba. 2007. “The NY: Cornell University; Fontainebleau, France: Contribution of Business Services to European INSEAD; and Geneva: World Intellectual Economic Growth.” In Business Services in Property Organization (WIPO). European Economic Growth, edited by Luis Eckstein, David, Vera Künzel, Laura Schäfer, Rubalcaba and Henk Kox, 74–94. London: and Maik Winges. 2019. Global Climate MacMillan/Palgrave. Risk Index 2020: Who Suffers Most from Kwon, Myung Joong, and Paul Stoneman. 1995. Extreme Weather Events? Weather-Related “The Impact of Technology Adoption on Firm Loss Events in 2018 and 1999 to 2018. Berlin: Productivity.” Economics of Innovation and Germanwatch eV. New Technology 3 (3–4): 219–34. Escribano, Alvaro, and J. Luis Guasch. 2005. Madsen, Jakob B., Md. Rabiul Islam, and “Assessing the Impact of the Investment James B. Ang. 2010. “Catching Up to the Climate on Productivity Using Firm-Level Technology Frontier: The Dichotomy between Data: Methodology and the Cases of Innovation and Imitation.” Canadian Journal Guatemala, Honduras, and Nicaragua.” Policy of Economics 43 (4): 1389–411. Research Working Paper 3621, World Bank, Mare, Davide S., Francesca de Nicola, and Faruk Washington, DC. Miguel Liriano. Forthcoming. “Financial Fazlıog˘lu, Burcu, Bas¸ak Dalgıç, Ahmet Burçin Structure and Firm Innovation: Cross-Country Yereli. 2019. “The Effect of Innovation Evidence,” Background paper for this report, on Productivity: Evidence from Turkish World Bank, Washington, DC. Manufacturing Firms.” Industry and Mason, Andrew D., and Sudhir Shetty. 2019. Innovation 26 (4): 439–60. A Resurgent East Asia: Navigating a Changing Feenstra, Robert C., Robert Inklaar, and Marcel World. East Asia and Pacific Regional Report. P. Timmer. 2015. “The Next Generation of Washington, DC: World Bank. the Penn World Table.” American Economic Mcguckin, Robert H., Mary L. Streitwieser, and Review 105 (10): 3150–82. Mark Doms. 1998. “The Effect of Technology 32 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

Use on Productivity Growth.” Economics of SQW. 2019. “Evaluation of Innovation Loans: Innovation and New Technology 7 (1): 1–26. Final Interim Report to Innovate UK.” Miyamoto, Koji, and Miguel Sarzosa. 2020. Evaluation report, SQW, Oxford, UK. “Workforce Skills and Firm Innovation: Sutton, John. 2012. Competing in Capabilities: Evidence from an Employer-Employee Linked The Globalization Process. Oxford: Oxford Survey Data in Vietnam.” Unpublished University Press. manuscript, World Bank, Washington, D.C. van der Eng, Pierre. 2009. “Capital Mohnen, Pierre, and Bronwyn H. Hall. 2013. Formation and Capital Stock in Indonesia, “Innovation and Productivity: An Update.” 1950–2008.” Bulletin of Indonesian Eurasian Business Review 3 (1): 47–65. Economic Studies 45 (3): 345–71. OECD and Eurostat (Organisation for Economic doi:10.1080/00074910903301662. Co-operation and Development and the World Bank. 2018. Growing Smarter: Learning Statistical Office of the European Union). 2018. and Equitable Development in East Asia Oslo Manual 2018: Guidelines for Collecting, and Pacific. East Asia and Pacific Regional Reporting and Using Data on Innovation. 4th Report. Washington, DC: World Bank. ed. Paris: OECD; Luxembourg: Eurostat. World Bank. 2020. Global Economic Prospects, Park, Albert, and Wenshi Xuan. 2020. “Skills for January 2020: Slow Growth, Policy Innovation in China.” Background paper for Challenges. Washington, DC: World Bank. this report, Hong Kong University of Science World Bank and DRC (Development Research and Technology, Hong Kong SAR, China. Center of the State Council, The People’s Salter, Ammon J., and Bruce S. Tether. 2006. Republic of China). 2019. Innovative China: “Innovation in Services: Through the New Drivers of Growth. Washington, DC: Looking Glass of Innovation Studies.” World Bank. Background paper for the Advanced Institute WTO and IDE-JETRO (World Trade of Management (AIM) Grand Challenge on Organization and the Institute of Developing Service Science, April 7, Oxford University. Economies). 2011. Trade Patterns and Solow, Robert M. 1957. “Technical Change and Global Value Chains in East Asia: From the Aggregate Production Function.” Review Trade in Goods to Trade in Tasks. Geneva: of Economics and Statistics 39 (3): 312–20. WTO. The State of Innovation in Developing East Asia 1

Introduction export-oriented manufacturing. Third, the COVID-19 pandemic, together with ongoing East Asia has undergone a significant climate change, is increasing economic vul- economic transformation over the past nerability and highlighting the need for new half century. High rates of growth have modes of production in the region. propelled countries in the region from low- Such forces, alone and together, raise ques- to middle-income, and even in a few cases, tions about whether the model that has driven to high-income status. An approach that has the region’s economic success in the past can become known as the “East Asian develop- continue to deliver rapid growth and development model”—a combination of policies that ment in the future. fostered outward-oriented, labor-intensive Innovation will be essential to overcom- growth; investments in basic human capital; ing these challenges. In fact, it has already and sound economic governance—has been played a critical role in the economic trans- instrumental in moving hundreds of millions formation of developing East Asian countries of people out of poverty and into economic over the past half century, beginning with security. the diffusion and adoption of high-yielding Despite their past economic successes, rice varieties, modern fertilizers, and other the region’s middle-income countries face an agricultural technologies (for example, pes- array of challenges as they strive to continue ticides and machinery)—a set of advances in their economic progress. First, productiv- agricultural production commonly known ity growth has declined since the 2008–09 as the Green Revolution. High-yielding rice Global Financial Crisis. This, coupled with varieties developed initially by international rapid population aging in several countries, agricultural research centers (such as the is putting pressure on the region’s growth International Rice Research Institute in the prospects, narrowing the opportunities for Philippines) in partnership with national reaping demographic dividends. Second, the agricultural research systems were not only slowing of global goods trade, uncertainty instrumental in raising agricultural produc- about the future of the global trading system, tivity across the region but also helped lay and rapid changes in technology are all chal- the foundation for the structural transforma- lenging a key engine of growth in the region: tion that fueled development in the region. 34 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

Innovation will play a central role in enabling and development of East Asia as well as to the countries of developing East Asia to move provide guidance on the types of policy and from lower-middle-income to upper-middle- institutional reforms needed to enable more income and, ultimately, to high-income levels. innovation-led growth in the region.2 To Interest in innovation among the region’s achieve this, the report examines the region’s policy makers has peaked recently with key innovation challenges, assesses its state of the rise of digital technologies. Indeed, innovation, and analyzes the main constraints high-profile accomplishments by private firms face in effectively pursuing innova- sector actors—in e-commerce, digital finan- tion. The report then analyzes the policies cial technology (fintech), ridesharing, and and institutions currently in place to support mobile app-enabled service delivery—have innovation in developing East Asia and lays captured the imaginations of policy mak- out an agenda for action aimed at spurring ers, the media, and citizens alike. Enterprises innovation-led growth. in the digital space, like the Chinese multi- This chapter provides an overview of the national technology company Alibaba and state of innovation in the region. It begins the ride-hailing services Grab and Go-Jek by providing the broad definition of innova- in Southeast Asia, have become household tion that will be used throughout the report. names. In addition, most countries are devel- The chapter then examines the productivity oping strategies to prepare for the next wave and related challenges facing developing East of advanced technologies. These technolo- Asian countries as they seek to sustain their gies (including advanced robotics, artificial high economic performance. To help contex- intelligence, 3-D printing, and the internet of tualize the analysis presented in the remain- things) are often referred to collectively as the der of the report, the chapter then reviews “fourth industrial revolution” (or “Industry the region’s innovation accomplishments 4.0”) technologies.1 and challenges, placing developing East Asian Despite the great promise of innovation— countries in a global context. The chapter and some high-profile successes—analysis concludes with a road map for the remainder of a range of key innovation indicators sug- of the report. gests that countries in developing East Asia still face important challenges to fostering Defining innovation innovation-led growth. They are not unique in this regard. Recent global research on It is useful to begin by defining innova- innovation shows that despite potentially tion explicitly because innovation means high returns to investment in innovation, different things to different people. This low- and middle-income countries around report adopts a broad definition of inno- the world commonly underinvest and vation based on the Oslo Manual (OECD underperform in innovation for several rea- and Eurostat 2018), which uses the term to sons: weak capacity of firms to adopt, use, refer to a new or improved product or busi- or invent new technologies; misalignment ness process that differs significantly from between public innovation policies and insti- the firm’s previous products or business tutions and countries’ needs; unconducive processes and that has been introduced to economic policy environments; and missing the firm and on the market (box 1.1). This complementary factors, such as the neces- definition relates to various functions of sary skills to innovate or inadequate instru- the firm (not only its products) and, more ments to finance innovation (Cirera and importantly, requires only that the “innova- Maloney 2017). tion” represent a significant improvement The objective of this report is to deepen to the firm, which may include upgrading policy makers’ understanding of innova- of processes and imitation of other existing tion and its critical role in the future growth products in the market. The State of Innovation in Developing East Asia 35

BOX 1.1 Defining firm-level innovation

The report adopts a broad view of innovation as specifications, reliability, durability, economic the accumulation of knowledge and implementa- efficiency during use, affordability, convenience, tion of new ideas. Specifically, the Oslo Manual usability, and user friendliness” (OECD and 2018 defines “business innovation” as a “new or Eurostat 2018, 71). In this context, traditional improved product or business process (or combina- surveys have used three metrics to capture the tion thereof) that differs significantly from the firm’s complexity or novelty of the innovation: previous products or business processes and that has  New products to the firm been introduced on the market or brought into use  New products to the market by the firm” (OECD and Eurostat 2018, 20). This  New products to the international market. report considers innovation defined both as “inven- • A business process innovation is “a new or tion” or “discovery” (that is, those developments improved business process for one or more busi- that push the technological frontier) and as “diffu- ness functions that differs significantly from sion” or “adoption” of existing technologies and the firm’s previous business processes and that practices that lead firms to novel ways of producing has been brought into use by the firm.” The or acting. The latter definition is pertinent to most of Oslo Manual 2018 lists six functional catego- the firms operating in developing East Asia. ries to identify and distinguish between types Several observations can be made regarding this of business process innovations, based on con- definition: temporary management research, that capture innovations in both core and supporting busi- • “An innovation must be novel or a significant ness functions of the firm (OECD and Eurostat improvement, at least to the firm and possibly to 2018, 73): the market (or other higher levels)” and must be  Innovative methods for manufacturing prod- implemented (that is, introduced inside the firm ucts or offering services or commercialized). There is no requirement for  Innovations in distribution and logistics, the innovation to be successful, however. including transportation and delivery of • The general definition does not mention intention inputs, products, or services; warehousing; or objective, but implicit is the idea that inno- and order processing vation aims at improving the firm’s competitive  Innovations in marketing and sales activities, position and is associated with uncertainty. which cover innovative marketing methods, • The broad definition makes no reference to tech- pricing strategies, and sales and after-sales nology, and the current definition (based on the activities (including help desks and customer fourth edition of the Oslo Manual [OECD and support) Eurostat 2018]) considers nontechnological forms  Innovations in the provision and maintenance of innovation, thereby encompassing a broader of information and communication systems, range of knowledge types than strictly scientific including hardware, software, databases, or technical ones. repairs, web hosting, and so on The Oslo Manual 2018 defines the following two  Innovations in administration and manage- main subtypes of innovations (OECD and Eurostat ment, including strategic and general business 2018, 21): management; corporate governance; accounting, auditing, and other financial activities; • A product innovation is “a new or improved good management of human resources; procure- or service that differs significantly from the firm’s ment; and management of external relations previous goods or services and that has been with suppliers, alliances, and so on introduced on the market.” This includes the  Innovations in product and business process addition of new functions or improvements to development, which covers activities to scope, existing functions or user utility. “Relevant func- identify, develop, or adapt products or a firm’s tional characteristics include quality, technical business processes.

box continues next page 36 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

BOX 1.1 Defining firm-level innovation (continued)

It is important to note that although this report To highlight as clearly as possible the multiple adopts a broad and multifaceted definition of inno- dimensions across which firms innovate, the analy- vation, existing data do not always enable analysis sis of Enterprise Survey data presented in this report across all the dimensions identified here. For will occasionally distinguish between four differ- example, data on firm-level innovation in the World ent types of innovation defined in the 2005 Oslo Bank Enterprise Surveys are based on the defini- Manual. A complete mapping of the differences and tion used in the third edition of the Oslo Manual correspondence of activities between the 2005 and (OECD and Eurostat 2005) because the surveys were 2018 definitions can be found in the Oslo Manual designed before the fourth edition was released in 2018 (OECD and Eurostat 2018, 75). Hence, 2018. The Enterprise Surveys thus include two addi- although data availability makes it impossible to tional categories of innovation: organizational and examine all dimensions empirically, the analysis will marketing. These categories have been subsumed be cognizant of the full range of innovation activi- under business process innovations in the 2018 ties defined here. edition and are captured by the six newly introduced functional categories. Source: Adapted from OECD and Eurostat 2018.

The term “significantly improved”—and The innovation imperative for the fact that some advances are new to the developing East Asia firm or to the local market while others are new to the world—reflects the broad Several economic forces are driving the imper- literature on innovation, which often uses ative for a more innovation-led growth model different terms to convey similar concepts. in developing East Asia. Improvements in products or processes Productivity remains relatively low in that are new to the firm or to the domes- developing East Asia—and productivity tic market are alternatively referred to as growth has declined since the Global “imitation” or “incremental innovation.” Financial Crisis Improvements that are new to the world are alternatively referred to as “invention,” Despite their remarkable growth perfor- “discovery,” or “radical innovation,” terms mance over time, countries in developing intended to indicate the newness or unique- East Asia still face important productiv- ness of the innovation. ity challenges. Although labor productivity Another important distinction is between has grown rapidly in recent years (averag- innovation and technology. Innovation ing 6.3 percent per year between 2013 and often involves the adoption of a new 2018), the region’s output per worker technology because implementation of a remains well below the productivity fron- product or process innovation may require tier, defined as the productivity level in the the introduction and use of a specific tech- United States (figure 1.1, panel a). Even in nology. Thus “diffusion” involves the pro- Malaysia, whose output per worker is the cess through which both past innovations highest in the region, labor productivity (of some firms) and new technologies are was only about 42 percent of the US level spread, adopted, and used by other firms. in 2017. In the region’s next tier of coun- The diffusion process can occur across or tries (China, Indonesia, the Philippines, and within countries, sectors, firms, and loca- Thailand), labor productivity only ranged tions—and it is mirrored by the process of between 18 percent and 24 percent of the “adoption” of new technologies by firms. US level in 2017. The State of Innovation in Developing East Asia 37

FIGURE 1.1 Although productivity in the major economies of developing East Asia has been rising, it remains well below the frontier

a. Labor productivity relative to b. Total factor productivity relative to United States, 1990–2017a United States, 1990–2017b

1.0 1.0

0.8 0.8

0.6 0.6

0.4 0.4

0.2 0.2 Labor productivity (US = 1)

0 Total factor productivity (US = 1) 0

1990 1993 1996 1999 2002 2005 2008 2011 2014 2017 1990 1993 1996 1999 2002 2005 2008 2011 2014 2017

China Indonesia Malaysia Philippines Thailand Vietnam

Sources: World Bank elaboration, based on Asian Productivity Organization (APO) Database 2019, version 2 (https://www.apo-tokyo.org/wedo /productivity-measurement/); and Penn World Table version 9.1 data. Note: The developing East Asian countries included in these figures—China and the five largest emerging economies of the Association of Southeast Asian Nations (ASEAN)—are those for which the most reliable data series are available. a. Labor productivity is measured as gross domestic product (GDP) per worker. b. Total factor productivity (TFP) is measured in purchasing power parity (PPP) terms. TFP series are calculated by the Penn World Table (PWT) team, except for Vietnam. For Vietnam, TFP is estimated using the methodology in Feenstra, Inklaar, and Timmer (2015), capital stock data from PWT (version 9.1) and labor share estimates from the APO Productivity Database 2019 (version 2).

Similarly, total factor productivity (TFP) Recent studies have highlighted the key levels in most of developing East Asia remain role that innovation must play in develop- far below the frontier (figure 1.1, panel b). ing East Asia if its countries are to maintain Again, Malaysia’s TFP, while higher than or increase productivity growth in a rapidly elsewhere in the region, was still only about changing and highly uncertain global eco- 62 percent of US TFP in 2017. Although nomic environment (Mason and Shetty 2019; productivity generally increases as countries World Bank and DRC 2019). Reinforcing develop, TFP in most developing East Asian the case for innovation-led growth is empiri- countries is below what would be predicted cal evidence from high-income economies on the basis of their gross domestic product as well as the region, showing strong links (GDP) per capita (figure 1.2). between technology adoption, innovation, Productivity growth has slowed world- and higher productivity (Cirera and Maloney wide since the Global Financial Crisis, and 2017; Comin and Hobijn 2010; de Nicola developing East Asia has not been immune. 2019; Hall 2011; Mohnen and Hall 2013). Indeed, the region experienced the second Andrews, Criscuolo, and Gal (2016) fur- steepest slowdown in labor productivity ther suggest that a critical component of growth of all emerging market and develop- the global productivity slowdown since the ing regions since the Global Financial Crisis late 1990s is the technological divergence (World Bank 2020b). Although labor produc- between leading and lagging firms. For that tivity growth declined across the region, the reason, promoting broad-based innovation— decline has been particularly pronounced in oriented not only to leading firms capable of China (figure 1.3). A decomposition of labor invention but also to other firms with the productivity growth shows that the slow- potential to adopt existing technologies—is down largely reflects weaker TFP growth. warranted. 38 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

FIGURE 1.2 Total factor productivity in most developing East Asian countries is below what would be predicted based on their GDP per capita

1.2

1.0 United States

0.8 Singapore Japan 0.6 Philippines Myanmar Korea, Rep. Thailand Malaysia Lao PDR Indonesia 0.4 China Cambodia Mongolia

TFP level at current PPP (US = 1) Vietnam 0.2

0 7 8 9 10 11 12 13 Log GDP per capita, PPP (constant 2011 international $)

Sources: World Bank elaboration, based on Penn World Table (PWT) version 9.1 and Asian Productivity Organization (APO) Database 2019, version 2 (https://www.apo-tokyo.org/wedo/productivity-measurement/). Note: Data are from 2017. Countries in light blue designate the 10 “developing East Asia” countries studied in this report. Total factor productivity (TFP) is measured in purchasing power parity (PPP) terms relative to the United States (1.0). TFP series are calculated by the PWT team, except for Cambodia, Myanmar, and Vietnam, which are estimated through the methodology of Feenstra, Inklaar, and Timmer (2015), using data from PWT (version 9.1) and labor share estimates from the APO Database 2019 (version 2). GDP = gross domestic product; PPP = purchasing power parity; TFP = total factor productivity.

FIGURE 1.3 Labor productivity and TFP growth have declined in developing East Asia since the Global Financial Crisis

b. Decomposition of productivity growth, a. Decomposition of productivity growth, China East Asia and Paci c, excluding Chinaa

12 6

10 5

8 owth (%) 4

6 3

Sources: Conference Board; Asian Productivity Organization (APO) Database 2019, version 2 (https://www.apo-tokyo.org/wedo/productivity-measurement/); and Penn World Table (PWT) version 9.1 data; van der Eng 2009; World Bank calculations. Note: Labor productivity is measured as GDP per worker. PWT data were used as the baseline. When PWT (version 9.1) data were not available, the APO Database 2019 (version 2) was used. Conference Board data were used for 2018. GDP = gross domestic product; TFP = total factor productivity. a. Panel b shows weighted averages calculated using GDP weights at 2010 prices. Countries included are Cambodia, Indonesia, Lao PDR, Malaysia, Mongolia, Myanmar, the Philippines, Thailand, and Vietnam. For Indonesia, capital stock was calculated by World Bank staff, extending data from van der Eng (2009). The State of Innovation in Developing East Asia 39

Spurring productivity growth through the productivity and technological gaps greater innovation will thus be important with high-income economies and preserve to the economic fortunes of developing East countries’ competitiveness in the global Asia in the medium term. Innovation can economy. enable the production of new and better goods and services as well as increase produc- The COVID-19 pandemic and climate tivity through organizational or marketing change are accelerating the need for improvements. Innovation can enhance pro- digitalization and new production ductivity growth largely by raising TFP—that methods is, enabling higher levels of output with the same quantity of resources. Indeed, innova- The COVID-19 pandemic tion has been an important element in the The COVID-19 pandemic has underscored rise of the region’s high-income countries. the importance of innovation as policy mak- Japan, the Republic of Korea, and Singapore ers and private firms have rushed to adopt have all used innovation to improve efficiency or develop technologies to address both the and boost their incomes, with great success health and the economic effects of the out- (Hobday 1995). break. This effort has included, among other things, the application of digital technologies to provide real-time information about the Changes in global trade and rapid spread of the virus; the use of drone tech- technological advances are challenging nologies for such applications as aerial dis- the region’s main engine of growth: infection, noncontact transport of medical export-oriented manufacturing supplies, and consumer deliveries; and the use The slowing of the global goods trade and of advanced biomedical technologies and arti- ambiguity about the future of the global ficial intelligence to develop testing, vaccines, trading system in general pose risks to a and an eventual treatment for the virus.3 development model that has effectively used Firms in the region have made greater use of trade, foreign direct investment (FDI), and digital platforms and invested in digital solu- integration into global value chains (GVCs) tions to maintain their businesses in the face as critical channels for growth. Furthermore, of mobility restrictions and social distancing a new technological revolution—Industry (figure 1.4). 4.0—poses a risk of disrupting existing pro- As importantly, the COVID-19 pandemic duction structures as it moves toward more has been a shock to GDP not seen for decades flexible manufacturing and customization in the region—one that can have long-lasting and increases the importance of proximity to effects. So large a shock, affecting both customers. These technological advances, dis- demand and supply, has highlighted the need cussed further in chapter 2, could result in the for more flexible management, more auto- shortening of GVCs or the reshoring of pro- mated and digitally integrated production duction systems that have helped fuel growth processes, and more digitally enabled delivery in developing East Asian countries. of services to offset some of the costs imposed Although the extent to which these by social-distance measures (see, for example, changes will occur is still uncertain, new Hatayama, Viollaz, and Winkler 2020). technology-driven production processes This shift has already begun in the region, could put pressure on the region’s labor- although the shock is also altering research intensive, export-oriented production and innovation priorities, with potentially model, potentially unwinding gains asso- important trade-offs (box 1.2). Policy makers ciated with countries’ participation in the will need to find ways to accelerate the tech- global trading system. Meeting these chal- nological transformation of their economies, lenges will require innovation to narrow while managing these trade-offs, to pave the 40 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

FIGURE 1.4 COVID-19 has induced greater use of and investment way for what will likely be very different in digital solutions in developing East Asia since the beginning of economies in the post–COVID-19 era. the pandemic Indeed, many of the changes induced by the pandemic are likely to be long lasting, 60 with changes in consumer preference requir- 53 52 48 ing many traditional brick-and-mortar com- 50 panies to adopt business models that increase 40 40 36 their presence on digital platforms. The closure of physical retail stores dictated by 30 30 28 COVID-19 containment measures already 23 20 19 appears to have encouraged the diffusion of 20 17 Share of rms (%) 14 e-commerce across the region (box 1.2). And the growth of e-commerce provides a poten- 10 5 tial avenue for new business and job creation 0 by generating opportunities for start-ups Indonesia Philippines Vietnam Cambodia Myanmar offering value-added services. Accelerating this digital transforma- Increased use of digital platforms tion in firms is imperative to recovery given Increased digital sales continuing unknowns about the trajec- New digital investment tory of COVID-19 and the high likelihood

Source: World Bank 2020a, based on COVID-19 Business Pulse Survey (BPS) data. of other, future pandemics. Firms’ digital Note: Data on increased digital sales and new digital investment since the pandemic are not transformation in the region will thus deter- available for Myanmar. The survey was conducted in May 2020 for Myanmar; in June 2020 for Cambodia, Indonesia, and Vietnam; and in July 2020 for the Philippines. The BPS is a telephone mine not only their ability to survive this survey implemented in 51 countries in six regions of the world using the same questionnaire. The crisis but also their competitiveness and resil- figure shows results for the countries included in the BPS from developing East Asia. Sample size 4 varies by country but is representative for small, medium, and large firms. ience in the face of unforeseen crises.

BOX 1.2 COVID-19: Changing the innovation landscape

The spread of COVID-19 and the policy response of technology or improved their product mix in may have conflicting effects on each of the two dimen- response to the pandemic. sions of innovation: (a) invention, and (b) diffusion However, the crisis-induced economic contrac- and adoption. Regarding invention, the pandemic has tion and uncertainty are inhibiting investments in boosted research and development (R&D) on tests, both invention and adoption in a variety of other vaccines, and treatment to combat the disease. This areas by cutting resources and dampening expected will likely have positive spillovers for broader scien- returns. To mitigate COVID-19’s pernicious impact, tific and medical research in areas such as biotechnol- significant funds and resources have been allocated ogy. At the same time, the social distancing needed toward medical and pharmaceutical R&D to develop to contain the disease has hurt scientific research not diagnostics, treatment, and vaccines. But research related to COVID-19, by shutting down laboratories on other topics and in other fields faces a differ- and durably disrupting experiments. ent reality. Containment measures have resulted in As for diffusion, adapting to social distancing has the closure of research labs that are not working boosted firms’ and households’ adoption of tech- on COVID-19, thus delaying, if not compromis- nologies for digital communication, conveyance, and ing, the realization of technological improvements. commerce that will likely be used beyond the pan- For example, bacterial colonies were slid into freez- demic. Apedo-Amah et al. (2020), using a sample of ers for an indefinite rest, clinical trial participants 51 countries (mainly low- and middle-income coun- were told to stay home, and populations of lab mice tries, including 5 developing East Asia countries),a were reduced to the bare minimum, while research- find that 49 percent of firms made greater use ers shifted their focus to writing grant proposals box continues next page The State of Innovation in Developing East Asia 41

BOX 1.2 COVID-19: Changing the innovation landscape (continued)

and drafting manuscripts, being unable to continue of physical retail stores dictated by COVID-19 con- hands-on research (Servick et al. 2020). tainment measures has encouraged the diffusion of Even when labs and firms remain operative, e-commerce as well. Shopee, one of Southeast Asia’s heightened uncertainty due to the pandemic shock largest e-commerce platforms by web traffic, dou- is hampering economic activity. Focusing on the bled its year-on-year orders during the first quarter US economy, Baker et al. (2020) show that about of 2020 (Loh 2020). 60 percent of output contraction is because of Moreover, the growth in e-commerce in the COVID-induced uncertainty. Moreover, Apedo- region corresponded to a change in the composi- Amah et al. (2020) find that firms in low- and tion of goods sold online. Before the pandemic, middle-income countries are three times more pes- e-commerce marketplaces focused on fashion; simistic and four times more uncertain than firms in cosmetics; and computer, communication, and the United States—and these more-pessimistic firms consumer electronics. The pandemic has prompted tend to fire more workers. an increase in online sales of grocery and fresh Furthermore, beyond these current effects, the produce. Brick-and-mortar supermarkets and fresh uncertainty likely will have longer-term negative con- produce specialty stores have striven to increase sequences because of the postponing and downscal- their presence on platforms such as Shopee, ing of irreversible expenses, such as investments in Lazada, and Tokopedia when COVID-19 contain- intangible capital and R&D expenditures (Barrero, ment measures forced them to close their physical Bloom, and Wright 2017). These cuts to innovation stores. inputs may happen in firms of all sizes: recent sur- Finally, the pandemic is also catalyzing greater veys underscore that pandemic-induced uncertainty use of fintech. Digital financial services have become is a key concern for firms both small (Apedo-Amah key to (a) facilitating the disbursement of emergency et al. 2020; Bloom et al. 2020) and large (Hassan liquidity to businesses and cash transfers to house- et al. 2020). holds, as well as (b) keeping people safe by promoting COVID-19 is inducing diffusion of technolo- contactless payment systems during this time of social gies in several ways that may permanently change distancing (Pazarbasioglu and Mora 2020). This dif- the economic landscape in the region. First, the fusion of fintech may, in turn, yield productivity gains, use of digital platforms has increased everywhere. by (a) promoting competition in the financial sector Apedo-Amah et al. (2020) estimate that 34 percent (Ruehl and Kynge 2019), and (b) increasing financial of firms have increased or started using digital plat- inclusion, because information gathered via fintech forms and social media. E-commerce grew substan- enables people and enterprises to build credit scores, tially in China, in response to the restrictions put in which in turn strengthens their access to finance. place because of the 2003 severe acute respiratory syndrome (SARS) outbreak, and COVID-19 is hav- a. The countries are Cambodia, Indonesia, Mongolia, the Philippines, and ing a similar effect (Economist 2020). The closure Vietnam.

Climate change According to the Global Climate Risk Climate change is also challenging tradi- Index 2020, 4 Southeast Asian countries— tional approaches to production and growth. Myanmar, the Philippines, Thailand, and Temperatures will increase significantly Vietnam—were among the 10 countries most in developing East Asia, and warming is affected between 1999 and 2018 (Eckstein already causing severe weather events more et al. 2019). Moreover, the continued reli- frequently: heat waves, droughts, flooding, ance of Southeast Asian countries on agricul- wildfires, and hurricanes. East and Southeast ture and the concentration of populations in Asian countries are among those likely to coastal regions exacerbates their vulnerabil- be hardest hit as the climate warms further ity. Many major coastal cities are seriously (map 1.1). imperiled, including Shanghai and Tianjin, 42 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

MAP 1.1 East Asian countries are among those most at risk from climate change

Countries and territories most aﬀected by extreme weather events (1999–2018) 1 Puerto Rico 2 Myanmar 3 Haiti 4 Philippines 5 Pakistan 6 Vietnam 7 Bangladesh 8 Thailand 9 Nepal 10 Dominica Italics: Countries and territories where more than 90 percent of losses or deaths occurred in one year or event. IBRD 45475 | Climate risk index: ranking 1999–2018 1 – 10 11 – 20 21 – 50 51 – 100 > 100 No data JANUARY 2021

Source: Adapted from Eckstein et al. 2019. ©Germanwatch. Reproduced, with permission, from Germanwatch; further permission required for reuse.

China; Jakarta, Indonesia; Ho Chi Minh City, high-profile, high-tech firms operating and Vietnam; and Bangkok, Thailand. innovating in the region, particularly those in Significant technological changes are the digital space. Although the achievements required to address climate change effectively of high-performing “unicorns” are impor- (box 1.3). From the perspective of mitigation, tant and noteworthy, realizing the economic it is imperative to have “cleaner” and more promise of innovation will require a broad energy-efficient production that reduces car- swath of firms across different sectors of bon emissions. From the perspective of adap- the region’s economies to engage in innova- tation, new and sustainable technological tion activities. But how well are developing solutions are necessary to ensure that agricul- East Asian countries performing overall on tural production is sustainable and to enable innovation? And how do the region’s econo- safe and productive factory environments at mies measure up to others at similar levels of higher temperatures. These responses can development? only be carried out if policy objectives shift toward prioritizing more innovation and the The region has experienced some adoption of new technologies. important innovation-related successes Innovation performance in There are several examples of strong inno- developing East Asia vation performance in developing East Asia (map 1.2). Beijing; Shanghai; and the region of Much attention has been paid in the popu- Shenzhen and Hong Kong SAR, China, are all lar press recently to a growing number of among the top 20 global innovation clusters The State of Innovation in Developing East Asia 43

BOX 1.3 Mitigating effects of climate change through effective technology adoption

Low- and middle-income economies’ long-term Crops are also affected by disease and pests, disap- growth is dependent on technology advancement pearance of pollinators, water stress, reduction in that drives productivity and economic performance food quality, increased risk of spoilage, and infesta- (Jones 2016). New technologies that are embodied tion by mycotoxins (Mbow et al. 2019). Agricultural in capital equipment and that generate a stream of output losses due to warming in Asia could innovations are among the principal drivers of pro- contribute to greater malnutrition and stunting. ductivity and arbiters of economic performance. Mitigation remedies include biogenetic, nano- However, research and development (R&D) invest- chemical, mechanical, and digital technologies to ment is below 1 percent of gross domestic product improve crop variety and resistance-maximizing (GDP) in most Southeast Asian countries. Increasing yields. Countries that lack the physical and human global temperatures stand to hit East and Southeast capital for frontier technologies can use adaptive Asian countries hard, affecting food security, water innovations to minimize output loss and avoid acute availability, mortality, labor productivity, student food crisis. In fact, a number of existing technolo- learning, migration patterns, physical infrastruc- gies, if adopted, hold considerable promise for rais- tures, conflict incidence, and localized violence. ing productivity and improving sustainability of Hence, technology must be borrowed and adapted agriculture in developing East Asia (Rajalahti 2021). while domestic innovation capabilities are being built through R&D and investments in science, tech- Ensuring water availability nology, engineering, and mathematics (STEM) fields. Freshwater resources underpin agricultural processes and determine the survival, growth, and quality of life. Buttressing labor productivity Domestic and industrial demands intensify as warm- Labor productivity, export and industrial value ing causes renewable water availability to decline and added, service output, and GDP decrease as the people’s exposures to floods to rise. For each degree temperature rises (Flouris et al. 2018). The poor of warming, roughly 7 percent of the world’s popula- are most affected because their activities are often tion is projected to experience a decrease in renew- concentrated in heat-prone areas. Air conditioning able water resources of at least 20 percent; whereas on a broad scale is one solution, but it is costly and by the end of the century, the number of people energy intensive. exposed annually to 100-year river floods could Technological advances in three areas could bring triple (Jiménez Cisneros et al. 2014). Water scarcity climate control within reach of low- and middle- in Asian countries is the result of drought and water income countries while curtailing carbon emissions: politics, water pollution, and excessive groundwater (a) development of low-carbon, cost-effective cool- extraction. These forces, combined with warming, ing technologies;a (b) electrification using renewable could mean that 40 percent of Asia’s population lack energy technologies; and (c) use of distributed gen- access to drinking water by 2030. eration, smart grid technology, and utility-scale stor- Technology must complement efficient age devices. use of water resources to reduce the energy Minimizing exposure to extreme heat in farming, and costs of recycling and desalinating water. manufacturing, and construction calls for greater Because 70 percent of water globally is used for mechanization and automation. So, countries need to agriculture,b research in low- and middle-income increase R&D and innovation to ensure user-friendly, countries (publicly or privately financed) must cost-effective technology including customized support development of cost-effective technolo- fabrics for extreme heat. gies for the desalination of brackish water and seawater. However, current technologies of reverse Maintaining agricultural yields osmosis and flash distillation are capital and Feeding the world’s population will be an increasing energy intensive. Although researchers are devel- challenge in a warming environment. Four commonly oping new techniques to reduce and recapture consumed crops (wheat, rice, maize, and soybean) energy, technologies for contaminant removal decrease in yield as the global temperatures rise. remain costly.

box continues next page 44 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

BOX 1.3 Mitigating effects of climate change through effective technology adoption (continued)

A range of technologies can support countries in Source: Adapted from Yusuf 2020. a. Recent technology competitions, such as the Global Cooling Prize, have controlling carbon emissions and coping with the spotlighted several promising cooling solutions. These include “smart hybrid threats to labor productivity, food and water avail- technology to optimize on efficiency and handle temperature and humid- ability, and human welfare posed by climate change. ity separately; no or low-GWP [global warming potential] climate-friendly refrigerants; reusing system-generated waste heat and water; smart controls, Countries in developing East Asia will need to sensors, and automation to optimize hybrid operation based on outdoor and increase spending on systems of innovation to cre- indoor conditions; or integration of a small solar panel on the outdoor unit to ate or adapt the technologies needed to address their significantly reduce the overall climate impact” (Lalit 2020). b. A higher percentage of water is needed for agriculture in low-income coun- current and emerging needs, however. Indeed, inno- tries (over 80 percent in Africa and Asia) and significantly lower percentages are vations to support greener growth in the region is an needed in high-income countries (41 percent on average, 21 percent in Europe, increasingly urgent priority. and 51 percent in North America).

MAP 1.2 Developing East Asia is home to several of the world’s top 100 innovation clusters

Source: Bergquist, Fink, and Raffo 2017. ©World Intellectual Property Organization (WIPO). Reproduced, with permission, from WIPO; further permission required for reuse. Note: The innovation clusters are identified based on the density of patent filings in a city or set of neighboring cities, using the geocoded addresses of inventors listed in patents published between 2010 and 2015 under WIPO’s Patent Cooperation Treaty (PCT). Cluster size is determined by the number of PCT applications associated with the inventors present in a given cluster. “Developing East Asia” refers to the 10 middle-income countries covered in this study: Cambodia, China, Indonesia, Lao PDR, Malaysia, Mongolia, Myanmar, the Philippines, Thailand, and Vietnam.

according to data on the spatial density of well as Kuala Lumpur, Malaysia—all three of patents filed under the World Intellectual which specialize in computer technology. Property Organization’s Patent Cooperation Other evidence indicates that the region’s Treaty (Bergquist, Fink, and Raffo 2017). All export-oriented growth model has enabled of these clusters specialize in innovation in countries to participate in more sophisti- digital communication. In addition, the top cated forms of manufacturing trade over 100 global innovation clusters include two time. Cross-country data show, for example, Chinese cities (Guangzhou and Hangzhou) as that most of the region’s countries perform The State of Innovation in Developing East Asia 45

at or above what would be predicted based Although much of the region’s participation on their per capita income levels with respect in this trade began with less-sophisticated com- to high-tech imports, high-tech and medium- ponents and assembly, these measures reflect high-tech outputs, and high-tech exports the increased adoption of global technologies (figure 1.5). and production processes over time through

FIGURE 1.5 Several developing East Asian countries are significant participants in the global value chains for high-tech products

a. High-tech importsa b. High-tech and medium-high-tech outputsb

100 Vietnam Malaysia 100 Singapore

Philippines China 80 Singapore 80 Korea, Rep. United States Japan United States 60 60 China Malaysia Thailand Korea, Rep. Philippines Japan Vietnam Thailand

40 Indonesia 40 Indonesia

High-tech imports (score) 20 20 Mongolia

0 Cambodia High-tech and medium-high-tech output (score) 0 Mongolia 6 81012 6 810 12 Log GDP per capita (PPP constant 2011 international $) Log GDP per capita (PPP constant 2011 international $)

c. High-tech exportsc

China Korea, Rep. 100 Vietnam Philippines Malaysia Singapore

80 Thailand Japan 60

Indonesia High-tech exports (score) 20 Mongolia Cambodia 0

6810 12 Log GDP per capita (PPP constant 2011 international $)

Source: World Bank elaboration, using data reported in Global Innovation Index 2019 (Dutta, Lanvin, and Wunsch-Vincent 2019). Note: High-tech export and import indicators include technical products with high research and development (R&D) intensity, as defined and classified by Eurostat, the statistical office of the European Union. “Developing East Asia” refers to the 10 middle-income countries covered in this study (designated in light blue): Cambodia, China, Indonesia, Lao PDR, Malaysia, Mongolia, Myanmar, the Philippines, Thailand, and Vietnam. The figure excludes Lao PDR and Myanmar, for which no recent data exist. GDP = gross domestic product; PPP = purchasing power parity. a. The high-tech imports indicator measures high-tech imports as a percentage of total trade. b. The high-tech and medium-high-tech output indicators measure high-tech and medium-high-tech output as a percentage of total manufacturing output, based on the classification of technology intensity defined by the Organisation for Economic Co-operation and Development (OECD). High-tech and medium-high-tech output indicators are missing for Cambodia. c. The high-tech exports indicator is defined as high-tech exports minus re-exports as a percentage of total trade. 46 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

FDI, creation of joint ventures, and participa- these countries appear to underperform on tion in trade and GVCs. For example, between several standard indicators of innovation for 2000 and 2008, the share of the domestic con- both diffusion (the adoption of existing tech- tent of exports in electronics grew significantly nologies) and discovery (the invention of new in Malaysia and Thailand as well as in indus- products, processes, and technologies, as dis- trial machinery in Indonesia and the Philippines cussed in box 1.1).6 (WTO and IDE-JETRO 2011), probably as a One critical input for more-basic forms of result of FDI to produce locally and increased innovation, such as improving the quality of participation of local suppliers. Central to the products and processes, is international cer- region’s outward-oriented manufacturing and tification, which gives firms access to other growth strategy, these forms of international countries’ markets. International certification engagement have represented important oppor- has contributed to firm-level productivity in tunities for technology transfer and knowledge several middle-income countries, including diffusion over the past half century.5 China and four Southeast Asian countries (Cirera and Maloney 2017; Escribano and Guasch 2005). However, all countries in Most countries in the region developing East Asia except China perform underperform on several key indicators below their predicted values regarding inter- of innovation, however national certification (figure 1.6, panel a). Despite the great promise of innovation in the Licensing of foreign technologies—another region, analysis of a range of key innovation important input for the diffusion and adop- indicators reveals that countries in develop- tion of new technologies—is also associated ing East Asia still face important challenges with higher innovation output among firms to fostering innovation-led growth. Most of in developing East Asia (Iootty 2019). Across

FIGURE 1.6 The share of firms with international certification is low in much of developing East Asia, and in half of the countries, fewer firms acquire licenses to foreign technology than expected given their countries’ per capita incomes

a. International certi cationa b. Foreign technology licensingb 60 40

China

30 40 Indonesia Malaysia 20 China Cambodia Mongolia Philippines 20 share of firms (%) Philippines Mongolia 10 Malaysia Vietnam Vietnam Lao PDR Cambodia Thailand Myanmar Thailand

International certification, share of firms (%) Indonesia

Myanmar Using technology licensed from a foreign firm, 0 Lao PDR 0 6 8 10 12 6 81012 Log GDP per capita, PPP (constant 2011 international $) Log GDP per capita, PPP (constant 2011 international $)

the region, performance regarding foreign China is noteworthy in that it performs sig- technology licensing is more mixed: in half of nificantly above expectations regarding both the countries, a smaller share of firms obtain R&D spending and patents granted. Similar licenses to foreign technologies than would patterns are seen in China’s other innovation be expected given their countries’ per capita inputs and outputs. incomes (figure 1.6, panel b). Data on the main input of discovery of Despite some high-profile successes, new products and technologies (R&D) and countries face ongoing challenges one key proxy of invention (patents) high- with services sector innovations, light similar challenges. Most countries in including in the digital space the region spend less on R&D, a key innovation input, than would be expected given Services are playing an increasingly important their per capita incomes (figure 1.7, panel a). role in the region’s economies—in manufac- Only three countries (China, Malaysia, and turing, trade, and in their own right—as the Vietnam) spend at or above expected levels. demand for services grows. Nonetheless, Similarly, most developing East Asian developing East Asian countries still face countries produce fewer patents—a com- challenges concerning innovation in services, monly used measure of innovation output including in the digital space. As figure 1.8 that can result in invention—than would shows, most of the region’s countries perform be expected given their per capita incomes below expectations, given per capita GDP, in (figure 1.7, panel b). Again, Malaysia, mobile app creation, information and com- Vietnam, and in in this case, Mongolia, per- munication technology (ICT) services exports, form at or near the predicted levels. and the export of “cultural and creative

FIGURE 1.7 Most countries in developing East Asia spend less on R&D and produce fewer patents than would be predicted by their per capita incomes

a. R&D expenditure b. Domestic patents granted

5 Korea, Rep. Korea, Rep. Japan 0 United States 4 China Singapore

Japan Malaysia 3 United States Mongolia Singapore –5 China Vietnam 2 Thailand Indonesia Malaysia Philippines

R&D expenditure, % of GDP Thailand 1 Indonesia Philippines Vietnam Cambodia –10

Source: World Bank elaboration, based on 2018 data from the World Intellectual Property Organization (WIPO) Statistics Database (https://www3.wipo.int/ipstats/index.htm) and the World Development Indicators database. Note: “Developing East Asia” refers to the 10 middle-income countries covered in this study (designated in light blue): Cambodia, China, Indonesia, Lao PDR, Malaysia, Mongolia, Myanmar, the Philippines, Thailand, and Vietnam. GDP = gross domestic product; PPP = purchasing power parity; R&D = research and development. 48 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

FIGURE 1.8 With some noteworthy exceptions, developing East Asian countries still face challenges concerning innovation in services, including in the digital space

a. Mobile app creationa b. ICT services exportsb

100 100

80 80

60 60 Philippines Singapore Vietnam Korea, Rep. 40 40 United States Mongolia Indonesia Singapore 20 United 20 ICT services exports (score) Malaysia Mobile app creation (score) Indonesia Thailand Philippines Japan Cambodia Vietnam Japan States Cambodia Malaysia China Korea, Rep. 0 0 Thailand Mongolia 6 8 10 12 6 8 10 12 Log GDP per capita (PPP constant 2011 international $) Log GDP per capita (PPP constant 2011 international $)

c. Cultural and creative services exportsc

100

United States 60 Singapore

China Korea, Rep. 20 Mongolia Japan Philippines Cambodia Malaysia

Cultural and creative services exports (score) Thailand 0 6 8 Indonesia 10 12 Log GDP per capita (PPP constant 2011 international $)

Source: World Bank elaboration, using data reported in Global Innovation Index 2019 (Dutta, Lanvin, and Wunsch-Vincent 2019). Note: “Developing East Asia” refers to the 10 middle-income countries covered in this study (labeled in light blue): Cambodia, China, Indonesia, Lao PDR, Malaysia, Mongolia, Myanmar, the Philippines, Thailand, and Vietnam. The figure excludes Lao PDR and Myanmar, for which no recent data exist. GDP = gross domestic product; ICT = information and communication technology; PPP = purchasing power parity. a. The mobile app creation indicator measures global downloads of mobile apps, by origin of the headquarters of the developer or firm, scaled by GDP (in PPP international $, billions). Global downloads are compiled by App Annie Intelligence, public data sources, and the company’s proprietary forecast model based on data from the Google Play store and iOS App store in each country between January 1, 2018, and December 31, 2018. Since data for China are not available for Google Play store and only for iOS App store, data from China are treated as missing and considered “n/a” (Dutta, Lanvin, and Wunsch-Vincent 2019, 365). b. The information and communication technology (ICT) services exports indicator represents telecommunications, computer, and information services exports (percentage of total trade), based on the 2010 Extended Balance of Payments Services (EBOPS 2010) classification (UNSD 2012). c. The cultural and creative services exports indicator measures creative services exports (percentage of total exports), covering information services; advertising, market research, and public opinion polling services; audiovisual and related services; and other personal cultural and recreational services, based on EBOPS 2010 (UNSD 2012). The cultural and creative services exports indicator is missing for Vietnam.

services”—defined to include information ser- related services; and other personal cultural vices; advertising, market research, and pub- and recreational services (Dutta, Lanvin, and lic opinion polling services; audiovisual and Wunsch-Vincent 2019). The State of Innovation in Developing East Asia 49

However, the data show a few clear exam- FIGURE 1.9 Countries in developing East Asia show considerable ples of better-than-expected performance. variation in the incidence and type of firm-level innovation Vietnam stands out as a high performer in mobile app creation (figure 1.8, panel a) and 70 the Philippines as a high performer in ICT 60 services exports (figure 1.8, panel b), most 50 likely reflecting that country’s global role in 40 providing business process outsourcing (BPO) 30 services (World Bank 2019).7 20 Share of rms (%) 10 Firms in the region engage in product, 0 process, marketing, and organizational China Vietnam Indonesia Lao PDR Malaysia Mongolia Myanmar Thailand innovation—although the extent varies Cambodia Philippines considerably across countries Introduced a new product or service Data from the World Bank’s Enterprise Introduced a process innovation Surveys provide a more granular view of Introduced marketing innovation countries’ innovation outputs. Consistent Introduced organizational innovation with the Oslo Manual definitions (discussed Source: World Bank estimates from Enterprise Survey data (latest available year). in box 1.1), the surveyed firms report Note: No data on marketing or organizational innovations are reported in the China or Myanmar whether they have introduced new or surveys. significantly improved (a) products (goods or services), (b) processes, (c) marketing innovation than in any other country in the methods, and (d) organizational practices. region. For example, nearly 60 percent of Figure 1.9 presents the share of firms in Chinese firms report introducing a process developing East Asia that reported innova- innovation, while around 45 percent report tions in each of those four domains. introducing a product innovation. Despite considerable variability across Also noteworthy is the relatively high share countries in the share of firms that report of firms reporting innovation activities in sev- innovations, a few patterns emerge: eral fast-growing lower-middle-income countries. Around one-third of firms in Cambodia • In 7 out of 10 countries in the region, a and Mongolia report undertaking process higher share of firms report undertaking innovations, while in Vietnam, the share is process innovations than other forms of close to 40 percent. These outcomes likely innovation. reflect the adoption and diffusion of new pro- • In all but one country (Lao PDR), a higher duction and business processes among firms share of firms report undertaking process integrated into the international economy innovations than product innovations. through trade and GVCs. As figure 1.9 also • A relatively large share of firms (20 percent shows, a much lower share of firms in sev- or more in 6 out of 10 countries) also eral other countries—Indonesia, Lao PDR, report undertaking marketing innovations. Myanmar, and to a lesser extent, Thailand— Although not shown in figure 1.9, the report engaging in innovation. Enterprise Survey data indicate that when Deeper analysis of the same Enterprise firms in developing East Asia innovate, Survey data indicates that firm innovation they are likely to engage in multiple inno- in developing East Asia is positively associ- vation activities across the four innovation ated with firm productivity—whether mea- domains. sured as labor productivity or as revenue Consistent with the cross-country indica- TFP. Specifically, firms that innovate gen- tors discussed above, a substantially larger erate 20 percent more output per worker share of firms in China report undertaking than those that do not, and TFP can be up 50 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

to 35 percent higher among innovative firms While many firms in the region still than non-innovative firms (de Nicola 2019). do not innovate, innovating firms The positive relationship between innova- commonly engage in the adoption of tion and productivity holds for firms in the existing technologies and practices region that introduce either new products Innovating firms in the region also report or new processes. This evidence on prod- whether the innovations they introduce are uct innovation is consistent with a broader new to the firm (only) or new to their main empirical literature, largely focused on high- market. Innovations that are new only to the income countries but also on China, that firm can be interpreted as resulting from the shows positive impacts of product innovation diffusion, adoption, and use of already exist- on revenue productivity. For surveys of that ing technologies and practices, whereas some evidence, see Hall (2011) and Mohnen and share of innovations that are new to the mar- Hall (2013). ket are likely to be the result of invention. In the Enterprise Surveys for developing East Firms in the region are more likely to Asia, it is not possible to distinguish between innovate in manufacturing than in new-to-market innovations that are new only services to the local market and those that are new to international markets. In reality, many of Enterprise Survey data also indicate that those innovations reported as new to the mar- higher shares of firms undertake innovation in ket likely represent diffusion and adoption manufacturing than in services (figure 1.10). applied to the local market, especially in the Only in Cambodia are the shares of firms lower-middle-income countries. reporting innovation in manufacturing and Innovating firms in developing East Asia in services roughly equivalent. In China and report both new-to-firm and new-to-market Malaysia, the differences in the shares of innovations (figure 1.11). The visible excep- firms reporting innovations in manufacturing tion is China, for which no data on new-to- and in services are particularly large. market innovations are available. But separate As chapter 2 will discuss further, innova- firm-level data from five Chinese provinces tion in services is important not only in its reinforce the finding that firms introduce own right but also to ensure that firms in the both new-to-firm and new-to-market innova- region can maintain their competitiveness tions, depending on their capabilities (Park in manufacturing and global trade. This, in and Xuan 2020). turn, has implications for innovation policy.8 This distinction between innovation defined as diffusion and adoption of existing FIGURE 1.10 In developing East Asia, a higher share of firms technologies and innovation defined as inven- reports innovating in manufacturing than in services tion is important in the context of develop- 80 ing East Asia. As the subsequent chapters will 70 discuss in greater detail, there is tremendous 60 heterogeneity in countries’ and firms’ capaci- 50 ties to innovate across the region. Especially 40 among the lower-middle-income countries 30 in the region (such as Cambodia, Lao PDR,

innovations (%) 20 and Myanmar) and among most small and 10

Share of rms introducing medium enterprises (SMEs), adoption and 0 diffusion of existing but new-to-firm tech-

China nologies and practices may be the only fea- Lao PDR Vietnam Indonesia Malaysia MongoliaMyanmar Thailand Cambodia Philippines sible route to innovation in the short term and thus of critical importance to increasing Manufacturing Services firms’ productivity and aggregate productiv- Source: World Bank estimates, based on Enterprise Survey data (latest available year). ity growth. The State of Innovation in Developing East Asia 51

FIGURE 1.11 Firms in developing East Asia report a mix of innovations that are new to the firm and new to the domestic market

100

40 Share of rms (%)

0 Cambodia China Indonesia Lao PDR Malaysia Mongolia Myanmar Philippines Thailand Vietnam New-to-market innovators New-to- rm only innovators Non-innovators

Source: World Bank calculations, using Enterprise Survey data (latest available year). Note: Enterprise Survey data from developing East Asia distinguish between innovations that are new to the firm and new to the market, but do not distinguish between innovations that are new to the local versus international market. The Enterprise Survey for China did not collect data on new-to- market innovation.

In fact, historical evidence suggests that be explained by foreign innovations embod- the development accomplishments of Japan ied in imports; in OECD countries, this and then the Four Asian Tigers (Hong Kong figure is only 35.5 percent. SAR, China; Korea; Singapore; and Taiwan, China) all involved substantial reduction in technology adoption lags relative to the Road map for this report Organisation for Economic Co-operation So, what will it take for developing East Asian and Development (OECD) country aver- countries to successfully transition to an age (Comin and Hobijn 2010). Comin innovation-led growth model, one in which and Hobijn (2011) estimate that the post– non-innovating firms begin to adopt new World War II growth experience of Western technologies and undertake innovation activi- Europe and Japan can be explained, in large ties, while more-advanced firms progressively part, by the rapid adoption of existing for- engage in R&D and innovation at the tech- eign technologies. Moreover, US economic nological frontier? In seeking to answer this aid and technical assistance (for example, question, the chapters of this report are orga- through the Marshall Plan) is significantly nized as follows: associated with reduction in adoption lags across these countries. • Chapter 2, Conceptual Framework and More-recent estimates by Santacreu Stylized Facts, begins by describing some (2015), using imports of intermediate key concepts for thinking about innova- goods, suggest that technology diffusion is tion and the adoption and diffusion of an especially important source of growth technology. It presents a framework for for less-developed countries. The author thinking about a graduated set of policies estimates that between 1996 and 2007, through which countries with different 65 percent of labor productivity growth can levels of innovation capability can more 52 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

effectively promote diffusion and adop- in developing East Asia effectively foster tion as well as invention. innovation-led growth. • Chapter 3, Technology Adoption and Diffusion: A Firm-Level Perspective, then focuses more specifically on the dif- Notes fusion of innovation and technology. It describes where the region stands in 1. For further discussion of Industry 4.0, see terms of technology adoption and docu- chapter 2 (particularly box 2.1). ments its tremendous heterogeneity across 2. The terms “East Asia” and “developing East countries, sectors, and firms with respect Asia” will be used throughout the report. For convenience, unless otherwise specified, these to innovation performance. The chapter terms refer to the 10 middle-income countries also examines the factors that affect firms’ covered in this study: Cambodia, China, abilities to innovate, including access Indonesia, the Lao People’s Democratic to information and management qual- Republic, Malaysia, Mongolia, Myanmar, the ity, as well as several important channels Philippines, Thailand, and Vietnam. for diffusion of knowledge for innova- 3. For more information on innovative responses tion that are external to the firm, such as to COVID-19, see the Coronavirus Innovation FDI, trade, and countries’ participation in Map (https://coronavirus.startupblink.com/); GVCs. the World Economic Forum COVID Action • Chapter 4, Skills and Finance for Platform (for example, Chandran [2020], at Innovation, examines two key comple- https://www.weforum.org/agenda/2020/03 /asia-technology-coronavirus-covid19 mentary factors associated with innova- -solutions/); and Huang, Sun, and Sui (2020). tion in developing East Asia: skills and 4. As part of its economic response to the finance for innovation. Specifically, the pandemic, the government of Korea has chapter discusses the roles of workforce announced the “K-New Deal,” which has skills and different sources of finance in two pillars—a digital new deal and a green enabling innovation, the status of skills economy new deal—to accelerate recovery development and innovation finance in the from the pandemic and to address climate region, and the challenges that develop- shocks. These are considered key priorities to ing East Asian countries face in ensuring increase employment in the country (MOEF adequate supplies of skills and risk-capital 2020). To support firms’ digital adoption, the finance to support innovation. government has undertaken special measures to ensure that micro, small, and medium • Chapter 5, Policies and Institutions in enterprises (MSMEs), not only large firms, the Region: An Assessment, analyzes the can implement the digital transition. adequacy of the region’s current policies 5. The roles of FDI, joint ventures, trade, and and institutions in addressing the barri- GVC participation in enabling innovation in ers that firms face concerning innovation. developing East Asian countries are discussed To that end, it assesses the mix of poli- at greater length in chapter 3. cies designed to promote innovation. It 6. For a more detailed discussion of key also examines the effectiveness of institu- innovation concepts, see chapter 2. tions and polices designed to facilitate the 7. No data are available for China on mobile transfer of knowledge essential to innova- app creation. However, given China’s role as tion. Special attention is paid to the role a global innovation cluster leader on digital communications, it is possible that it performs and efficacy of country research institu- above predicted levels in this area as well. tions that aim to facilitate innovation and 8. See the chapter 2 discussion of “servicification,” knowledge transfer. the increasing importance of service inputs into • Chapter 6, Action for Innovation: A Policy manufactures. Chapter 5 examines existing Agenda, builds on the analysis presented biases against services in innovation policy, in the first five chapters by outlining sev- while chapter 6 examines the implications for eral directions for policy to help countries innovation policy. The State of Innovation in Developing East Asia 53

References Comin, Diego, and Bart Hobijn. 2011. “Technology Diffusion and Postwar Andrews, Dan, Chiara Criscuolo, and Peter N. Growth.” NBER Macroeconomics Annual Gal. 2016. “The Best versus the Rest: The 25 (1): 209–46. Global Productivity Slowdown, Divergence de Nicola, Francesca. 2019. “Assessing How across Firms and the Role of Public Policy.” Returns to Innovation Vary Depending OECD Productivity Working Papers No. 5, on the Business Environment and Firms’ Organisation for Economic Co-operation and Characteristics.” Background paper for this Development, Paris. study, World Bank, Washington, DC. Apedo-Amah, Marie Christine, Besart Avdiu, Dutta, Soumitra, Bruno Lanvin, and Sacha Xavier Cirera, Marcio Cruz, Elwin Davies, Wunsch-Vincent, eds. 2019. Global Innovation Arti Grover, Leonardo Iacavone, et al. 2020. Index 2019: Creating Healthy Lives—The “Unmasking the Impact of COVID-19 on Future of Medical Innovation. 12th ed. Ithaca, Businesses: Firm Level Evidence from Across NY: Cornell University; Fontainebleau, France: the World.” Policy Research Working Paper INSEAD; and Geneva: World Intellectual 9434, World Bank, Washington, DC. Property Organization (WIPO). Baker, Scott, Nicholas Bloom, Steven Davis, Eckstein, David, Vera Künzel, Laura Schäfer, and Stephen Terry. 2020. “COVID- and Maik Winges. 2019. Global Climate Induced Economic Uncertainty and Its Risk Index 2020: Who Suffers Most from Consequences.” NBER Working Paper Extreme Weather Events? Weather-Related No. 26983, National Bureau of Economic Loss Events in 2018 and 1999 to 2018. Berlin: Research, Cambridge, MA. Germanwatch eV. Barrero, Jose Maria, Nicholas Bloom, and Economist. 2020. “Less Globalisation, More Ian Wright. 2017. “Short and Long Run Tech: The Changes Covid-19 Is Forcing on Uncertainty.” NBER Working Paper No. to Business.” Economist, April 11. https:// 23676, National Bureau of Economic www.economist.com/briefing/2020/04/11/the Research, Cambridge, MA. -changes-covid-19-is-forcing-on-to-business. Bergquist, Kyle, Carsten Fink, and Julio Raffo. Escribano, Alvaro, and J. Luis Guasch. 2005. 2017. “Identifying and Ranking the World’s “Assessing the Impact of the Investment Largest Clusters of Inventive Activity.” Climate on Productivity Using Firm-Level Economic Research Working Paper No. 34, Data: Methodology and the Cases of World Intellectual Property Organization Guatemala, Honduras, and Nicaragua.” Policy (WIPO), Geneva. Research Working Paper 3621, World Bank, Bloom, Nicholas, Philip Bunn, Scarlet Chen, Washington, DC. Paul Mizen, and Pawel Smietanka. 2020. Feenstra, Robert C., Robert Inklaar, and Marcel “The Economic Impact of Coronavirus on P. Timmer. 2015. “The Next Generation of UK Businesses: Early Evidence from the the Penn World Table.” American Economic Decision Maker Panel.” VoxEU/CEPR Policy Review 105 (10): 3150–82. Portal, March 27, https://voxeu.org/article Flouris, Andreas D., Petros C. Dinas, Leonidas G. /economic-impact-coronavirus-uk-businesses. Ioannou, Lars Nybo, George Havenith, Glen P. Chandran, Rina. 2020. “Here’s How Asia Is Kenny, and Tord Kjellstrom. 2018. “Workers’ Using Tech to Tackle COVID-19.” Reuters Health and Productivity under Occupational article on the World Economic Forum COVID Heat Strain: A Systematic Review and Meta- Action Platform, March 18. https://www Analysis.” The Lancet Planetary Health 2 (12): .weforum.org/agenda/2020/03/asia-technology e521–e531. -coronavirus-covid19-solutions/. Hall, Bronwyn H. 2011. “Innovation and Cirera, Xavier, and William F. Maloney. 2017. Productivity.” NBER Working Paper No. The Innovation Paradox: Developing-Country 17178, National Bureau of Economic Capabilities and the Unrealized Promise of Research, Cambridge, MA. Technological Catch-Up. Washington, DC: Hassan, Tarek Alexander, Stephan Hollander, World Bank. Laurence van Lent, Markus Schwedeler, and Comin, Diego, and Bart Hobijn. 2010. “An Ahmed Tahoun. 2020. “Firm-Level Exposure Exploration of Technology Diffusion.” to Epidemic Diseases: COVID-19, SARS, and American Economic Review 100 (5): 2031–59. H1N1.” Working Paper No. 26971, National 54 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

Bureau of Economic Research, Cambridge, Mbow, Cheikh, Cynthia Rosenzweig, Luis G. MA. Barioni, Tim G. Benton, Mario Herrero, Hatayama, Maho, Mariana Viollaz, and Hernan Murukesan Krishnapillai, Emma Liwenga, et al. Winkler. 2020. “Jobs’ Amenability to Working 2019. “Food Security.” In Climate Change and from Home: Evidence from Skills Surveys for Land: An IPCC Special Report, Supplementary 53 Countries.” Policy Research Working Paper Material, 5SM-1–17. https://www.ipcc.ch/site 9241, World Bank, Washington, DC. /assets/uploads/sites/4/2020/06/IPCCJ7230 Hobday, Michael. 1995. Innovation in East Asia: -Land_SM5_200226.pdf. The Challenge to Japan. Cheltenham, UK: MOEF (Ministry of Economy and Finance, Edward Elgar. Republic of Korea). 2020. “Korean New Deal: Huang, Yasheng, Meicen Sun, and Yuze Sui. 2020. National Strategy for a Great Transformation.” “How Digital Contact Tracing Slowed Covid- National strategic plan, MOEF, Seoul. 19 in East Asia.” Harvard Business Review, Mohnen, Pierre, and Bronwyn H. Hall. 2013. April 15. https://hbr.org/2020/04/how-digital “Innovation and Productivity: An Update.” -contact-tracing-slowed-covid-19-in-east-asia. Eurasian Business Review 3 (1): 47–65. Iootty, Mariana. 2019. “Assessing Innovation OECD and Eurostat (Organisation for Economic Patterns and Constraints in Developing East Co-operation and Development and the Asia: An Introductory Analysis.” Policy statistical office of the European Union). 2005. Research Working Paper 8706, World Bank, Oslo Manual: Guidelines for Collecting and Washington, DC. Interpreting Innovation. 3rd ed. Paris: OECD; Jiménez Cisneros, Blanca E., Taikan Oki, Nigel Luxembourg: Eurostat. W. Arnell, Gerardo Benito, J. Graham OECD and Eurostat. 2018. Oslo Manual 2018: Cogley, Petra Döll, Tong Jiang, and Shadrack Guidelines for Collecting, Reporting and Using S. Mwakalila. 2014. “Freshwater Resources.” Data on Innovation. 4th ed. Paris: OECD; In Climate Change 2014: Impacts, Adaptation, Luxembourg: Eurostat. and Vulnerability. Part A: Global and Sectoral Park, Albert, and Wenshi Xuan. 2020. “Skills for Aspects. Working Group II Contribution Innovation in China.” Background paper for to the Fifth Assessment Report of the this report, Hong Kong University of Science Intergovernmental Panel on Climate Change, and Technology, Hong Kong SAR, China. edited by Christopher B. Field, Vicente R. Pazarbasioglu, Ceyla, and Alfonso Garcia Mora. Barros, David Jon Dokken, Katharine J. Mach, 2020. “Expanding Digital Financial Services Michael D. Mastrandrea, et al., 229–69. Can Help Developing Economies Cope with Cambridge and New York: Cambridge Crisis Now and Boost Growth Later.” Voices University Press. (blog), World Bank, April 29. https://blogs. Jones, Charles I. 2016. “The Facts of Economic worldbank.org/voices/expanding-digital Growth.” Handbook of Macroeconomics, -financial-services-can-help-developing Vol. 2A, edited by John B. Taylor and Harald -economies-cope-crisis-now-and-boost-growth Uhlig, 3–69. Amsterdam: Elsevier. -later. Lalit, Radhika. 2020. “These Innovations Could Rajalahti, Riikka. 2021. “Agricultural Innovation Keep Us Cool without Warming the Planet.” in Developing East Asia: Productivity, Safety, Article for the World Economic Forum Annual and Sustainability.” Background paper for this Meeting, January 21–24, Davos, Switzerland. study, World Bank, Washington, DC. https://www.weforum.org/agenda/2020/01 Ruehl, Mercedes, and James Kynge. 2019. /these-innovations-could-keep-us-cool “Fintech: The Rise of the Asian ‘Super App.’” -without-heating-us-up/. Financial Times, December 12. https://on.ft Loh, Dylan. 2020. “Coronavirus Pandemic Fuels .com/36F7wd5. Asia E-Commerce Boom.” Nikkei Asia, Santacreu, Ana Maria. 2015. “Innovation, May 31. https://asia.nikkei.com/Business Diffusion, and Trade: Theory and /Retail/Coronavirus-pandemic-fuels-Asia-e Measurement.” Journal of Monetary Economics -commerce-boom. 75: 1–20. Mason, Andrew D., and Sudhir Shetty. 2019. Servick, Kelly, Adrian Cho, Jennifer Couzin- A Resurgent East Asia: Navigating a Changing Frankel, and Giorgia Guglielmi. 2020. World. East Asia and Pacific Regional Report. “Coronavirus Disruptions Reverberate through Washington, DC: World Bank. Research.” Science 367 (6484): 1289–90. The State of Innovation in Developing East Asia 55

UNSD (United Nations Statistics Division). World Bank. 2020b. Global Economic Prospects, 2012. Manual on Statistics of International January 2020: Slow Growth, Policy Trade in Services 2010. New York: United Challenges. Washington, DC: World Bank. Nations. World Bank and DRC (Development Research van der Eng, Pierre. 2009. “Capital Center of the State Council, The People’s Formation and Capital Stock in Indonesia, Republic of China). 2019. Innovative China: 1950–2008.” Bulletin of Indonesian New Drivers of Growth. Washington, DC: Economic Studies 45 (3): 345–71. World Bank. doi:10.1080/00074910903301662. WTO and IDE-JETRO (World Trade Organization World Bank. 2019. “Systematic Country and the Institute of Developing Economies). Diagnostic of the Philippines: Realizing 2011. Trade Patterns and Global Value Chains the Filipino Dream for 2040.” Report No. in East Asia: From Trade in Goods to Trade in 143419-PH, World Bank, Washington, DC. Tasks. Geneva: WTO. World Bank. 2020a. From Containment to Yusuf, Shahid. 2020. “Technological Priorities for Recovery. World Bank East Asia and Pacific Developing Countries in a Warming World.” Economic Update, October 2020. Washington, Unpublished manuscript, World Bank, DC: World Bank. Washington, DC.

Conceptual Framework and Stylized Facts 2

Introduction support greater invention among frontier firms. Developing East Asia faces a set of challenges that is putting a strain on the region’s economic development model, as Key concepts discussed in chapter 1. The transition to Innovation is a broad concept that includes an innovation-led growth model is even many types of knowledge activities and more urgent in the current context of rapid levels of complexity technological change, further accelerated by the COVID-19 pandemic. The questions As discussed in chapter 1, this report adopts for policy makers are how to speed up this a broad definition of “innovation” based transition and what policies and institutions on the Oslo Manual (OECD and Eurostat are required to facilitate it. 2018). The term refers to a product, process, Successful responses to these questions technology, business model, organizational require identifying key actors, institutions, structure, or marketing strategy that is “new and policies that facilitate innovation or significantly improved” and that is effec- and address the bottlenecks that impede tively introduced in the firm or the market. It progress. To that end, this chapter starts relates to different functions of the firm, not by defining some key concepts. It then just products, and more importantly requires examines the rising importance of support- only that the “innovation” represent a sig- ing technology adoption by firms in light nificant improvement to the firm, which may of rapid technological advances around the include upgrading of processes or the imita- world. The chapter then presents a frame- tion of other products already in the market. work linking innovation performance with The concepts related to innovation and the (1) a set of policies and institutions that diffusion of innovation and technology are can facilitate the diffusion and adoption broad, comprising activities of very different of technologies, and (2) policies that can levels of sophistication and complexity. 58 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

Innovation is a process of accumulating literature is that “invention” and other and using knowledge more-complex types of innovation, often proxied by patents, require more- All types of innovation require significant intensive R&D investments.1 Other investment in what are called “innovation forms of innovation may require other inputs,” which vary according to the inputs more intensively. For example, type and complexity of the innovation adopting lean manufacturing processes implemented. Figure 2.1 introduces the may require training, organizational concept of the “innovation knowledge changes, and investments in software function.” The idea is that firms invest in and technology. and accumulate knowledge (inputs) that • Different firms have different “produc- are transformed into innovation outputs tivity” or ability to transform knowledge (products, services, or processes) or out- activities or inputs into innovation outcomes, such as patents. These investments puts. A burgeoning literature has identi- vary in type—from knowledge embedded fied the heterogeneous capacity of firms in technology and technology licenses, to innovate as central to explaining persis- use of intellectual property, investment in tent productivity differences across firms human capital and training, and develop- (Acemoglu et al. 2018). ment of managerial and organizational • Knowledge activities are not always quality, to the most commonly known developed in-house but often imple- innovation input: research and develop- mented in collaboration with other firms ment (R&D). and knowledge providers—for example, In interpreting the knowledge function, from partner firms in joint innovation three dimensions are critical: projects, through contracting of public • Different combinations of knowledge research organizations (PROs) and uni- activities are possible, depending on the versities for R&D projects, or through type of innovation introduced. A stylized transfer of knowledge from buyers and fact drawn from the knowledge function suppliers.

FIGURE 2.1 The innovation knowledge function begins with investments and inputs that, once transformed into innovation outputs and outcomes, can yield improved firm performance

Innovation inputs and Innovation outputs knowledge activities and outcomes Impact

• Technology Firm growth • Equipment • New or improved (new demand or increased market share products and services • R&D due to enhanced quality or cost advantage) • New or improved • Intellectual property use business processes • Human capital • New business models • Training Productivity growth • New or improved (improved business processes • Engineering and design organizational and and technology) • Software and databases managerial practices • Managerial and • Patents and other organizational capital intellectual property Economic diversi cation and practices

Source: Cirera and Maloney 2017. ©World Bank. Further permission required for reuse. Note: R&D = research and development. Conceptual Framework and Stylized Facts 59

Imitation and adoption are more FIGURE 2.2 Sophistication of firm-level prevalent than invention—especially in innovation correlates strongly with countries’ low- and middle-income countries economic development

The reality in most low- and middle-income 20 economies is that innovation consists primarily of (a) imitating existing products and pro- 15 cesses, and (b) facilitating the diffusion and adoption of existing technologies. Science and R&D are important, but the capabilities of 10 firms to successfully implement R&D proj-

ects, patenting, and invention are commonly Share of all rms (%) 5 limited. The ability to undertake more-complex forms of innovation increase as firms 0 build their innovation capabilities. LIC LMIC UMIC HIC In more-advanced economies, sophis- Firms innovating Firms innovating ticated firms able to patent and invent new products new to rm products new to products and technology coexist with many or local market international market less sophisticated firms, for which innova- Source: World Bank calculations, using Enterprise Survey data (latest tion is still about imitating and improving available years). Note: Enterprise Survey data used are from 44 selected countries outside existing products and processes. Figure 2.2 of developing East Asia (whose data do not distinguish between shows that the extent of imitation—new innovations that are new to the local market versus the international market). HIC = high-income countries; LIC = low-income countries; products that are new only to the firm or LMIC = lower-middle-income countries; UMIC = upper-middle-income the local market—is greater in lower-income countries. Country income classifications are calculated using the World countries, while more “radical” or sophis- Bank Atlas method (https://datahelpdesk.worldbank.org/). ticated forms of innovation that are new to the international market are more prevalent of higher education and academia than those in higher-income countries.2 This mix of new- of industry). to-firm and new-to-market innovations is Following this linear view of innovation— also seen specifically in developing East Asia from science to R&D to innovation—has (chapter 1, figure 1.11). often translated into misdirected and ineffi- cient use of public resources to support R&D and PROs. Specifically, it has led to a focus on A linear view of innovation—from generating new technologies rather than on science and R&D to innovation—is not firm adoption and diffusion of existing tech- aligned with the reality of most low- nologies and undertaking the basic upgrading and middle-income countries and can that better suits firms’ capabilities. As firms harm policy effectiveness increase their capacity to innovate over time, Innovation has traditionally been associ- technological upgrading can become progres- ated in popular media outlets and policy sively more sophisticated and unique. circles with the discovery and invention of new products, processes, and technologies. The process of technological catch-up As such, innovation processes have generally requires moving more firms from been characterized as the result of science and imitation to discovery R&D efforts, and this conceptualization of innovation has significantly biased innovation The process of technological catch-up to the policies. In many countries, the responsibility frontier requires accelerating technologi- for innovation policies lies with ministries of cal diffusion and adoption and, ultimately, science and technology or sometimes even transitioning more firms from imitation to with ministries of education (the latter being increased R&D investments and invention more concerned with addressing the problems (Madsen, Islam, and Ang 2010). This process 60 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

is captured visually in figure 2.3, which shows between high-income countries (right of the the relationship between countries’ proxim- orange line in figure 2.3), which are relatively ity to the technological frontier (proxied by strong performers on innovation, and most their income per capita) and innovation qual- countries in developing East Asia (left of the ity (proxied by their score in the 2019 Global orange line). Policies in developing East Asia Innovation Index [GII], issued by the World must therefore focus on narrowing technolog- Intellectual Property Organization [WIPO] ical and innovation gaps, helping more firms and its partners). to become more innovative and to use more The GII is a composite index that incor- sophisticated technologies. porates data on the quality of innovation It is difficult to define precisely how this outputs (such as patents) and the quality of process of technological catch-up occurs and innovation inputs (such as R&D). Although when countries have successfully transitioned the GII is a crude indicator of innovation, to an innovation-led growth model. However, it includes multiple dimensions of innova- East Asia has examples of such transitions in tion and, as such, captures the level of inno- countries like Japan, the Republic of Korea, vation complexity across countries. The and Singapore, and a large literature has figure shows a clear and positive correlation described their experiences (see, for example, between per capita income and the GII, with Hobday 1995), highlighting the importance of higher-income countries demonstrating more outward orientation and learning. However, sophisticated innovation capacity. as chapter 1 discussed and chapter 3 will Within the East Asia region, there are clear further elaborate, technology embedded in differences in overall innovation performance foreign direct investment (FDI) and trade

FIGURE 2.3 Innovation quality correlates closely with countries’ per capita income

60 Korea, Rep. Singapore China Hong Kong SAR, China

Malaysia Philippines 40 Vietnam Thailand

Mongolia GII overall score Cambodia

20 Indonesia Lao PDR Myanmar

0 5 6 7 8 9 10 11 12 ln GNI per capita, Atlas method

Sources: World Bank elaboration, using Global Innovation Index (GII) 2019 data (https://www.globalinnovationindex.org/) and the World Development Indicators database. Note: The teal vertical line separates low- from middle-income countries, and the orange vertical line, middle- from high-income countries. No data are available for Lao PDR and Myanmar. The World Bank Atlas method estimates of the size of economies based on GNI converted to current US dollars, smoothing exchange rate fluctuations by using a three-year moving average, price-adjusted conversion factor. GNI = gross national income; ln = natural logarithm. Conceptual Framework and Stylized Facts 61

in developing East Asia has not diffused At the same time, the extent to which inno- widely beyond the sectors directly affected. vation increases productivity depends not Broadening and accelerating the diffusion of only on the firm’s knowledge investments but technologies and innovation is thus critical to also on other (complementary) factors, such developing an innovation-led growth model. as management capabilities, availability of Indeed, the evidence suggests that without technical skills and finance, and on having an adequate diffusion, innovation and new tech- enabling business environment and competi- nologies have little impact on productivity tive framework that incentivizes and facili- and growth (Hall and Khan 2003). tates innovation and diffusion (Cirera and Maloney 2017). Without adequate workers’ skills, firms cannot effectively use new tech- Several complementary factors are nologies. Without adequate finance, it is dif- critical to enabling innovation ficult for firms to purchase new machinery or If innovation is likely to boost productiv- finance R&D for innovation projects. And ity growth, the question becomes why lag- without good management, it is difficult for ging firms and countries fail to invest more firms to identify and effectively implement in innovation inputs and knowledge activities innovation projects. (as listed in figure 2.1). Chapter 3 explores Using country-level panel data, Goñi and this issue in more detail for developing East Maloney (2017) find that the relationship Asia, but, in general, the incentives to invest between returns to R&D and country gross in innovation depend on the benefits (returns) domestic product (GDP) per capita follows an that innovation brings to firms in terms of inverted-U shape (figure 2.4). A likely expla- increased productivity or profitability. nation of this inverted-U shape is that the

FIGURE 2.4 The relationship between returns to innovation and distance to the frontier follows an inverted U-shape

4.0 TUR(1971–1975) 3.5 BRA(1976–1980) ARG(2006–2010) TUR(2006–2010) ZAF (2006–2010) CHL(2006–2010) CRI(1976–1980) 3.0 MEX(2006–2010) CHL(1981–1985) LTU(2006–2010) CRI(2006–2010) SVK(2006–2010) MEX(1971–1975) ARG(1966–1970) SVK(1996–2000) BRA(2006–2010) SAU(2006–2010) CZE(2006–2010) PER(1976–1980) POL(1991–1995) CZE(1996–2000) 2.5 URY(1971–1975) RUS(2006–2010) CYP(1981–1985) BGR(2006–2010) VEN(2001–2005) ZAF(1986–1990) ISR(1966–1970) BGR(1991–1995) RUS(1991–1995) URY(2006–2010) 2.0 KOR(1966–1970) POL(2006–2010) TTO(2001–2005 COL(2006–2010) GTM(1971–1975) PAN(1991–1995) MLT(1986–1990) KOR(2006–2010) TTO(2006–2010) ECU(2006–2010) PAN(2006–2010) VEN(1971–1975)

1.5 IRN(2006–2010) GRC(1966–1970) SGP(1981–1985) JOR(1976–1980) COL(1976–1980) JOR(2006–2010) GRC(2006–2010) GTM(2006–2010)

1.0 ECU(1971–1975) MAR(2006–2010) TUN(2006–2010) PRT(2006–2010) HND(2001–2005) CYP(2006–2010) CHN(2006–2010) BOL(1996–2000) MAR(2001–2005) PRY(2006–2010) ESP(2006–2010) HKG(2006–2010) PHL(2006–2010) HKG(1996–2000) UKR(1996–2000) JPN(1966–1970) ISR(2006–2010) 0.5 SGP(2006–2010) PER(2001–2005) NOR(2006–2010) CHE(1986–1990) CAN(2006–2010) DEU(1976–1980) ISL(2006–2010) LUX(2006–2010) DNK(1976–1980) DNK(2006–2010) LUX(2001–2005) CAN(1966–1970) CHE(2006–2010) TUN(1966–1970) IRL(2006–2010) BOL(2006–2010) 0 PHL(1966–1970) IDN(2006–2010) ESP(1966–1970) UKR(2006–2010) EGY(2006–2010) KEN(1966–1970) IRL(1976–1980) NZL(2006–2010) SEN(1996–2000) LKA(2006–2010) –0.5 NZL(1991–1995) DEU(2006–2010) GBR(1976–1980) JPN(2006–2010) ISL(1971–1975) GBR(2006–2010) NOR(1966–1970) USA(2006–2010) USA(1966–1970) Percentage increase in output per unit of R&D investment –1.0 –4.5 –4.0 –3.5 –3.0 –2.5 –2.0 –1.5 –1.0 –0.5 0

Source: Goñi and Maloney 2017. Note: “Distance from the technological frontier” is measured as the difference from the highest percentile of income per capita, with a 0 value when the country is, during that period (shown within parentheses), at the higher income-per-capita level. Countries are designated by ISO alpha-3 code. R&D = research and development. 62 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

potential gains from technological catch-up changes affecting the global economy. This increase as the distance to the technological ongoing wave of change, sometimes referred frontier increases (as denoted by higher nega- to as the fourth industrial revolution (Schwab tive values on the x-axis of the figure). At some 2016) or “Industry 4.0,” comprises several point, however, these returns are reduced by interrelated strands—physical, biological, and the lack of complementary factors. Below a digital—each of which benefits from the oth- certain level of development, firms lack these ers as new discoveries are made and progress necessary factors, such as capital markets achieved (box 2.1). that would enable them to buy the necessary Changes in physical technologies emerge machinery, managers who know how to take from advances in molecular-level engineer- new ideas to market, and skilled workers to ing and high-tech manufacturing, and they manage innovation projects, all leading to a include advanced materials, advanced decrease in returns to innovation activity. manufacturing (including advanced robot- Two of these key complementary factors, ics), and additive manufacturing (or 3-D finance and skills for innovation, are analyzed printing). New biological technologies are in more detail in chapter 4. More broadly, enabling the manipulation of genes and however, identifying these missing comple- deoxyribonucleic acid (DNA) sequences mentary factors is essential to designing and to influence medical outcomes and correct implementing policies to increase returns genetic defects. And digital technological to innovation—the expectation of which change includes transformations in such is the main incentive for firms to decide to fields as financial technology (fintech), the innovate.3 internet of things (IoT), and artificial intelligence (AI). Importance of innovation These technological advances have the and diffusion of technology potential to lower costs but, even more funda- in addressing the region’s mentally, to change patterns of comparative challenges advantage by altering the relative importance of labor and capital. In doing so, adoption of A new technology revolution is under Industry 4.0 technologies by companies in the way that threatens current patterns of high-income economies—or even in China— trade and production could disrupt developing East Asia’s low-cost The importance of innovation to developing labor advantage. East Asia’s economic future is all the more For instance, cheaper and more user- urgent in the face of the rapid technological friendly robots offer the prospect of

BOX 2.1 Industry 4.0: An ongoing technological revolution

The term “fourth industrial revolution,” or Industry • The first industrial revolution, which began at 4.0, refers to a current wave of rapid technological the end of the 18th century, used water and steam advances that stand to dramatically change the global power to mechanize production. economy. The term, popularized around 2011, first • The second industrial revolution, beginning at arose from a project promoting the computerization the start of the 20th century, followed the intro- of manufacturing, led by the German government. duction of electric power, which enabled mass Industry 4.0 follows three earlier phases of tech- industrial production. nological change that have revolutionized produc- • The third industrial revolution, which began tion since the late 1700s (figure B2.1.1): in the latter part of the 20th century, used

box continues next page Conceptual Framework and Stylized Facts 63

BOX 2.1 Industry 4.0: An ongoing technological revolution (continued)

electronics and information technology (IT) to The fourth industrial revolution goes beyond the use further automate manufacturing. of technology per se, going hand in hand with novel • The fourth industrial revolution, which builds on production and management processes, including the third, is characterized by a fusion of technolo- user-centered design or flexible production, custom- gies based on cyber-physical systems. ization, and data management. Industry 4.0 is also significant because of its The transition between the third and the fourth potential impacts on employment and income dis- industrial revolutions has many analogies to the tribution (Autor and Dorn 2013; Frey and Osborne transition from the first to the second. While elec- 2017). This is already becoming apparent, with tricity networks and the railroad established the net- advanced robotics and AI enabling the reshoring of works for the second industrial revolution, advanced some manufacturing and services to high-income electronics and digital systems are providing a criti- economies, often with fewer jobs that require more- cal foundation for Industry 4.0. sophisticated skill sets (Acemoglu and Restrepo Industry 4.0 is characterized by the adoption of 2018; Ford 2015). This raises important challenges cyber-physical systems, including advanced robotics for developing East Asia because traditional labor and drones, 3-D printing, artificial intelligence (AI), cost advantages become less relevant with advanced and machine learning, whose effects are being felt automation and manufacturing that could poten- across all economic sectors. These technologies are tially be reshored to more advanced economies. reshaping not only how manufacturing is done or services are provided but also where they are located. Source: DFKI 2011.

FIGURE B2.1.1 The four stages of industrial revolution have repeatedly transformed manufacturing

FOURTH INDUSTRIAL REVOLUTION Cyber-physical systems THIRD INDUSTRIAL REVOLUTION Electronics and information technology SECOND INDUSTRIAL REVOLUTION Electric power enables FIRST INDUSTRIAL mass production REVOLUTION Water and steam power mechanize manufacturing facilities

18th century Early 20th century 1970s Today

Source: Adapted from DFKI 2011. 64 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

productivity increases while substituting developing East Asia. Together, these tech- for low-skilled labor. This prospect, in turn, nologies have three important implications could encourage labor-scarce high-income for innovation policy. Specifically, there is a and upper-middle-income economies to con- need to sider “reshoring” their production, marking • Accelerate innovation and technology a move away from the offshoring and the adoption, based on a better understanding fragmentation of production that has charac- of the factors that constrain adoption and terized much of the export-oriented manufac- diffusion; turing in the region. • Adopt new organizational and business Similarly, 3-D printing allows for the cus- models that are better aligned with more- tomization of production closer to consum- advanced technologies; and ers, which could reduce the importance of • Eliminate policy biases against innova- scale economies in production and disrupt tion in services sectors, which are becom- existing supply networks. Advances in AI and ing increasingly important to the region’s machine learning will also drastically change economies. the way knowledge tasks are organized and reinforce the incentives to automate. These three implications are now discussed As Industry 4.0 technologies continue to in turn. develop, there are concerns that reshoring of production to high-income economies will The policy imperative is to accelerate disrupt existing production structures and technology adoption and diffusion value chains in the region. Indeed, a rapid reshoring of production to Europe or North Rapid technological change, along with America could put pressure on the region’s changing global economic forces, is likely to export-oriented production model and transform the structure of production, and unwind many of the gains associated with developing East Asia must accelerate its tech- countries’ participation in global value chains nological catch-up. To do so, the region must (GVCs). first recognize that reshoring decisions will Evidence regarding the impacts of Industry depend on a set of factors that can be summa- 4.0 technologies on investment, produc- rized by productivity differentials, costs and tion, and exports in developing East Asia is customization, and closeness to customers still relatively scarce. And, to the extent that and supply chains. In this regard, the region such evidence exists, the findings are mixed.4 must try to minimize these productivity dif- Nevertheless, as the costs of new technologies ferentials by accelerating its own path toward continue to fall, there is a risk that invest- the technological frontier. In addition, the ments in such technologies could erode much increasing demand for customized products of the region’s comparative advance and will occur in the countries with an increas- shorten the value chains that have been so ingly demanding middle class, implying that important to developing East Asia’s devel- firms need technologies to supply goods and opment. Moreover, other external forces services with greater agility to meet rising (discussed in chapter 1), including slowing local demand. global trade, increased protectionist senti- Both these factors—minimizing the differ- ment in some economies, and the COVID-19 entials and meeting rising demand—require pandemic, are accelerating the adoption of that the region accelerate the diffusion of advanced technologies. new technologies. As discussed in chapter 1, In short, rapidly advancing technologies, global challenges arising from pandemics or especially when viewed together with other climate change also demand rapid adoption key global economic forces, reinforce the of advanced technologies. But what deter- importance of policies to spur innovation in mines technology diffusion? Conceptual Framework and Stylized Facts 65

Supply and demand factors are more and innovation outcomes. A rich microeco- important than relative prices in nomic literature has explained the diffusion determining technology adoption and of technology as depending on a series of diffusion supply and demand factors (Stoneman and A widespread view of technology adoption Battisti 2010). Supply factors determine the and diffusion at a more macro level focuses complexity and price of the technology, which on relative prices and factor abundance. The affect the firms’ decision to adopt it. Demand idea is that poor countries typically have factors determine the firms’ ability and will- barriers to technology transfer, abundant ingness to either adopt a given technology low-skilled labor, and a scarcity of skilled now or wait for later. Hall (2006) summarizes labor, and thus they will use different tech- some of these key demand factors: (a) benefits nologies than wealthy countries (Acemoglu that users can obtain from technology, (b) and Zilibotti 2001). Factor prices and context network effects from several users in a clus- play a critical role in selecting technologies, ter, (c) costs of implementing the technology, and there is an appropriate set of technologies (d) information available and the uncertainty for each country (Caselli and Coleman 2006). associated with implementing technology, Ultimately, however, the choice depends on and (e) market structure. how complementary technology and labor All these factors influence the adoption are (Acemoglu 2010) and also how force- of technology and explain some of the styl- fully R&D in the country can generate new ized facts that arise from the empirical work technologies. on diffusion, specifically (a) the S-shape Although it is unclear exactly how factor diffusion curve (box 2.2), and (b) the faster prices affect diffusion, micro evidence sug- diffusion of technologies across countries gests that other elements may be at play in (the extensive margin) than within countries explaining differences in technology adoption (the intensive margin).

BOX 2.2 Understanding adoption and diffusion: S-shaped diffusion curves

A common pattern observed throughout the tech- to differences in the rate of farmers’ adoption of nology diffusion literature is that the process of hybrids in states for which the technology was diffusion across regions fits an S-shaped curve already available; and (b) the availability problem, (that is, a logistic function). Evidence on this which refers to the lag in the development of began with the seminal work of Griliches (1957), hybrid corn technologies that were suitable for who analyzed the technological gap across regions specific areas. of the United States in the use of hybrid seed corn Following Griliches’s work, several other studies (figure B2.2.1). According to the author, hybrid supported S-shaped curves as a good fit for tradi- corn was a new method of breeding superior corn tional measures of technology diffusion. Mansfield for specific locations. But it was not immediately (1961) analyzes the factors determining the speed adopted everywhere. of technology diffusion across firms.b In addition The differences in S-shaped curvesa across to observing heterogeneity across industries, his certain states in the United States reflected two findings suggest that the growth over time of firms different problems associated with technology having introduced an innovation conforms to a adoption: (a) the acceptance problem, which refers logistic function (S-shaped). The author finds that

box continues next page 66 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

BOX 2.2 Understanding adoption and diffusion: S-shaped diffusion curves (continued)

FIGURE B2.2.1 S-shaped diffusion curves in the adoption of hybrid seed corn in selected US states, 1933–56

Percent of total corn acreage planted with hybrid seed

100

Iowa 80 Wisconsin Kentucky 60

Texas Percent 40 Alabama

0 1932 ‘34 ‘36 ‘38 1940 ‘42 ‘44 ‘46 ‘48 1950 ‘52 ‘54 ‘56

Source: Griliches 1957.

the probability of a firm introducing a new technique Most of the studies supporting S-shaped diffusion is an increasing function of the proportion of firms curves do not take into consideration the intensity in already using it and the profitability of doing so and, the use of technologies, however. conversely, a decreasing function of the size of the investment required. Other studies supporting this pat- a. The S-curve is partially the result of the metric of analysis, which is bounded between 0 and 1. tern include Skinner and Staiger (2007), Alm and Cox b. Mansfield (1961) focuses on 12 innovations across four industries in the (1996), and Gort and Klepper (1982), among others. United States.

Evidence suggests that technological in technologies invented after 1925 than in catch-up is not accelerating in low- and technologies invented earlier.5 Indeed, using middle-income countries data from the Cross-Country Historical Although the diffusion of technology between Adoption of Technology (CHAT) dataset,6 countries (extensive margin) may be accel- Comin and Mestieri (2018), show that the erating, the speed of internal diffusion adoption lag (defined as the time for the tech- (intensive margin) and intensity of its use nology to arrive to a country following its in low- and middle-income countries is lag- invention) between “Western countries” and ging behind that of high-income countries. the rest of the world has narrowed consid- Comin, Hobijn, and Rovito (2006) find that erably over time (figure 2.5). However, the the speed of convergence is three times faster authors also show that the gap in the intensity Conceptual Framework and Stylized Facts 67

FIGURE 2.5 Lags in the adoption of new technologies are now similar between Western countries and the rest of the world, but gaps in the intensity of technology use have widened

a. Adoption lags, 1800–2000 200 Ships Spindles

150

100 Railway passengers and freight Steel (Bessemer, open hearth) Tractor Electric arc furnace Adoption lags in years Telephone Electricity Fertilizer Aviation passengers and freight Harvester

50 Cars and trucks Synthetic ber Liver transplant Heart surgery Blast oxygen furnace Kidney transplant Cell phones Mail Internet Telegraph

0 PCs

1800 1850 1900 1950 2000 Year of technology invention

b. Intensity of use, 1800–2000

0 Ships Telegraph

–0.5 Spindles Steel (Bessemer, open hearth) Mail Electricity Aviation passengers and freight PCs –1.0 Electricity Internet Electric arc furnace Kidney transplant Cell phones Log intensive margin Fertilizer –1.5 Telephone Tractor Synthetic ber Railway passengers and freight Cars and trucks Heart surgery Harvester Liver transplant

–2.0 Blast oxygen furnace

1800 1850 1900 1950 2000 Year of technology invention Fit for Western countries Fit for non-Western countries

Source: Comin and Mestieri 2018. Note: Bars show the median margins of adoption of various technologies for Western versus non-Western countries. The adoption lag (the number of years it takes for a technology to arrive to a country since invention) and the intensive margin (the intensity of use in a country) are estimated structurally using country-specific model parameters derived from the Cross-Country Historical Adoption of Technology (CHAT) database. The lines are fitted lines for “Western countries” and the rest of the world. The bars show the median adoption lags or intensive margins of the two country groups for each technology. “Western countries” are defined to include Australia, Austria, Belgium, Canada, Denmark, Finland, France, Germany, Italy, Japan, the Netherlands, New Zealand, Norway, Sweden, Switzerland, United Kingdom, and United States. The technologies are coded numerically as follows (here listed in order of invention): 1. spindles; 2. ships; 34. railway passengers and freight; 5. telegraph; 6. mail; 7. steel (Bessemer, open hearth); 8. telephone; 9. electricity; 101. cars and trucks; 12. tractors; 134. aviation passengers and freight; 15. electric arc furnaces; 16. fertilizer; 17. harvester; 18. synthetic fiber; 19. blast oxygen furnaces; 20. kidney transplant; 21. liver transplant; 22. heart surgery; 23. personal computers (PCs); 24. cell phones; and 25. internet. 68 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

of technology use is widening. Facilitating traditional modes of production. Value addi- technological convergence when the frontier tion is likely to shift from production to is constantly moving requires accelerating design, customization, and maintenance ser- the diffusion and more-intensive use of new vices associated with the products sold. technologies. Weking et al. (2019) studied 32 case studies of organizational innovation related to Industry 4.0 implementation. The authors There is a need for organizational identify two types of new business mod- innovation and new business models els related to integration with value chains: that are more aligned with advanced (a) “servicification” associated with the com- technologies bined production of goods with services, and Although it is still early to know exactly (b) expertise, which includes the provision of how the nature of production will evolve consulting services. in response to the new technologies, some Policies focusing on innovation have tra- important changes in production methods are ditionally focused on product and process already occurring. For example, all the lead- innovation, not innovation in organizational ing sports footwear producers already offer structures or business models. As technolo- the possibility of customizing some of their gies rapidly advance, it will be important to products. take a broader view of innovation policies to Previous experience during technologi- support changes in organizations as well as in cal transitions suggests that new technolo- products and processes. gies come hand in hand with organizational changes and new business models. (See Emerging “servicification” warrants Bresnahan [2010] for the implementation of a reduction in policy biases against information and communication technology innovation in services [ICT].) In the case of the new technological transition, the move from fixed and mass As suggested earlier, one characteris- production to more flexible and custom- tic of Industry 4.0 is the potential for ized production will require different sets of “servicification”—the ability to produce skills but also stronger organizational and services that complement manufacturing managerial practices to process and use data production. For example, Weking et al. integration, manage leaner inventories and (2019) emphasize that the use of IoT and logistics systems, and coordinate production the integration of sensors into products processes. Similarly, more-flexible produc- to enable the provision of services (for tion systems require additional capabilities, example, remote monitoring and predictive such as IT engineers, as well as new business maintenance services) transforms manufac- models for the firm, to offer new services turing firms into de facto service providers. associated with products and develop more These new production structures require a integrated customer relationships. new set of knowledge-intensive inputs and Box 2.3 describes some of the key organi- outputs that complement more-traditional zational changes likely to be required for the manufacturing sector inputs and outputs. successful implementation of Industry 4.0 The trend toward servicification challenges technologies. As chapter 1 discussed, these policy makers’ longstanding focus on inno- will be further accelerated given the increase vation in manufacturing, which has biased in digitalization and automation in response public policy against services (as discussed to the COVID-19 pandemic. These new tech- in chapter 5). nologies will require firms to adapt and cre- Innovation has been seen traditionally in ate new business models that are often very terms of product and process inventions and different from those associated with more improvements, especially in manufacturing. Conceptual Framework and Stylized Facts 69

BOX 2.3 Organizational changes associated with Industry 4.0

Key to understanding the impact of Industry 4.0 is IoT (Internet of Things) machines will enable more understanding the types of organizational changes active, autonomous, and self-organizing production needed to adapt to these technologies. Industry 4.0 based on small units. An increasing number of prod- will lead to the emergence of dynamic, real-time ucts will belong to a global digital chain. optimized, self-organizing value chains. This will Top-floor to shop-floor integration . Factories will likely have a profound impact on firms’ organiza- adapt automatically to changes because of more tional structures and processes, which will need to be autonomous decision making. Autonomous manu- redesigned to enable adoption of these technologies. facturing units—coupling robotics with highly skilled These changes will offer new opportunities but also workers—will adapt to continuous customer-driven several challenges, particularly for countries that are product changes, enabling a single production line to lagging behind in terms of digital infrastructure and create different product types without reengineering human capital. Here are some examples of organiza- the production process. tional changes that are likely to become more prevalent as firms adopt Industry 4.0 technologies. Real-time, value-added networks . Digitalization and pervasive connectivity will enable real-time analysis Evolving business models . Mass production will of all business activities. Cost structures can be simu- likely yield to greater customization, leading to lated to support decision making. Market changes customer-specific products integrated with new ser- can be anticipated and business ideas implemented vice offerings. Manufacturers will increasingly shift more quickly. their revenue from products to services, creating new value-added services embedded in products. Borders Enhanced work environments . As customized between companies within industries are already production becomes the norm, workers can be becoming blurred. Companies can create value assigned as needed to coordinate automated pro- within business networks, for example, by offering duction processes and intervene when machines unused production capacity in a marketplace to com- call for action. Workers will have new demands panies that temporarily need more capacity. in managing complexity, problem solving, and self-organization. End-to-end digital engineering . Rigid preplan- ning processes will likely disappear. Smart Source: Cirera et al. 2017.

This is partly explained by the nature of of productive development, have translated services, which makes service innovation into a large innovation policy bias against the more difficult to define. As Tether (2003, services sector. Even though services account 483) describes, “Because services tend not for the largest share of GDP and employment to have an independent physical existence, in many economies, services innovation has service innovations can be invisible, and been neglected in innovation policy mixes because services are interactive, often being in both high-income and low- and middle- coproduced by the provider and user acting income economies. As chapter 5 will discuss, together (with simultaneous production and most policies and PROs focus on manufac- consumption), the authorship of any innova- turing or agriculture, and public programs to tion is often unclear. Furthermore, because support innovation tend to have fewer ser- service events are often unique, it is often dif- vices firms as beneficiaries. Eliminating this ficult to differentiate between service varia- policy bias against services will be important tions and innovations.” to facilitating faster diffusion and adoption of This difficulty in defining innovation in technologies and enabling greater innovation- services, and a manufacturing-centric view led growth. 70 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

Core elements for designing Innovation also depends on building policies that accelerate absorptive capacity and removing technology adoption and barriers to knowledge accumulation diffusion Several issues must be considered as policy Given the technological imperative that the makers seek to encourage innovation in region faces, a key question for innovation developing East Asia. policy is how to accelerate diffusion of new Matching of knowledge supply and technologies. Central to this process is build- demand . Knowledge-supply capabilities ing firms’ innovation capabilities and remov- (the research, knowledge, and technologies ing barriers to firms’ accumulation of the created by research institutions, universi- knowledge needed to adopt and productively ties, and other firms) must be matched by use new technologies. This section outlines firms’ demand for that knowledge (that is, several key principles for designing policies adequate absorptive capacity). This match- to enable greater technology adoption and ing of supply and demand for knowledge diffusion. is often overlooked, with policies and institutions traditionally focused on encouraging public R&D that may be disconnected Innovation depends on a constellation from firms’ capabilities and demand. It is of actors, policies, and frameworks impotant that policies focus on the accumu- A first entry point to understanding which lation of capabilities in both the supply and actors, policies, and institutions matter for demand for knowledge. innovation and technology diffusion can be In cases where the quality of knowledge found in a country’s national innovation sys- created—whether by firms, PROs, or other tem (NIS). Although different authors char- domestic knowledge providers—does not acterize such systems in different ways,7 for suit firms’ capabilities and needs, obtain- practical purposes a country’s NIS can be ing knowledge generated from abroad thought of as comprising firms; a country’s is critical. This can be done by incentiv- science and technology institutions, includ- izing knowledge spillovers from multi- ing PROs, universities, and private research national enterprises (MNEs), diffusing centers; government ministries and agencies existing technologies via imports, and in whose actions can either enable or impede some cases by working with international innovation; the broader business and regula- research organizations or private sector tory environment; industry associations; and consortia. other institutions that build a country’s skills Legal and financial barriers . When encour- base and support finance for innovation. aging innovation, one must be mindful of a Operationalizing this framework requires range of barriers that impede knowledge defining the boundaries of actors and insti- accumulation. For example, labor laws that tutions that affect innovation. There are dif- discourage the recruitment of skills needed ferent ways of doing this conceptualization for innovation or barriers to the financing and setting these boundaries. A key com- of innovation projects and ventures directly mon element across these NIS frameworks undermine innovation. By affecting the is the dependence of firm-level innovation broader economic environment, factors such on a large set of factors and institutions and as regulatory distortions, barriers to firm their systemic interactions. Acting upon one entry, and weak rule of law raise the cost of pillar only—for example, skills for innova- doing business and thus affect investment, tion—does not guarantee that innovation including investment in knowledge for inno- activity will be increased or technology diffu- vation. Such impediments include regulatory sion facilitated. Policies need to be systemic, bottlenecks that affect the deployment and use affecting incentives and mobilizing the range of new technologies. The IoT, for example, of relevant actors.8 largely depends not only on infrastructure but Conceptual Framework and Stylized Facts 71

also on regulations on its use and access to and generate new technologies often entails the internet network. initially investing in imitation of existing These more-general barriers have received products and processes, adopting existing relatively little attention in the innovation technologies, and eventually investing in for- literature because, among other reasons, mal and more sophisticated forms of R&D. they often lie outside the policy space occu- This knowledge accumulation and learning pied by innovation agencies and ministries. process depends on a firm’s internal efforts Nevertheless, these are critical elements but also on its ability to learn and partner affecting innovation decisions. with external sources of learning—for exam- Factors affecting firm incentives and ple, knowledge acquired from other firms capabilities . Another important set of issues or MNEs, knowledge embedded in imports, relate to factors that directly affect firms’ research produced in universities or PROs, demand for innovation. These factors may and knowledge services purchased from pub- be either external or internal to the firm. lic or private sources. This learning process is External factors affect the incentives and affected by several types of policies, ranging use of inputs for innovation, including trade from regulatory frameworks that affect the regimes that can encourage or discourage cost of doing business or market competi- the import of technology; competition and tion to regulations on trade and investment. market structure; and macroeconomic poli- The intellectual property regime, policies to cies. A critical set of internal factors relate incentivize scientific enterprise and university- to firms’ own capabilities. This includes industry collaboration, as well as incentives employees’ skills and management quality, to carry out R&D activities, also affect this as well as the decisions and processes firms learning process. use to accumulate, use, and create knowledge and bring it to market. Policy makers must prioritize the An enabling environment for innova- policies that build innovation tion . Finally, a critical consideration relates capabilities to complementary factors that affect all the Policy makers need to prioritize certain poli- interactions needed for innovation. When cies over others based on the level of techno- creating an enabling environment for inno- logical capabilities of the country’s private vation, it is important for policy makers to sector, which itself tends to be heterogenous, ensure the flow of knowledge from its cre- differing tremendously between large, inter- ators to the firms absorbing it; an adequate nationally linked firms and small domestic supply of skills, finance, and good manage- enterprises. The question is how to make such ment practices that can support the imple- a prioritization. mentation of innovative projects; and the Figure 2.7 approximates where the coun- appropriate regulatory and policy frame- tries in developing East Asia are in terms of works that enable innovation. These are innovation capabilities. The scatterplot uses key characteristics of a well-developed NIS GII data on innovation inputs (measuring and are needed to accelerate the adoption infrastructure, institutions, R&D, and human of new technologies. capital quality) and an innovation outputs index that captures the quality of knowledge, technology, and creative outputs of the Innovation requires knowledge economy. accumulation—with the firm at the As expected, the relationship is positive center of the process and linear: except for China, most of the These considerations surrounding the NIS countries on the right of the graph—the high- can be summarized from the firm’s perspec- est scorers on both inputs and outputs—are tive (figure 2.6). The ability to patent inven- high-income countries, whereas the countries tions, introduce new products, and invent on the left side of the graph are low- and 72 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

FIGURE 2.6 Accumulating knowledge to reach the innovation frontier: A firm-centered process

External sources of knowledge Firms Policies and frameworks for innovation (a ect the incentives to invest in innovation) Pushing the frontier Science (patenting and generation of technologies)

Finance Property rights and PROs R&D incentives

R&D Technology licensing investments Skills, educational, and and MNE spillovers research policies

Technology support Trade, competition, and Adoption of management infrastructure and services Adoption of investment policies practices, new processes, Complementary factors existing technologies quality upgrading Regulatory framework Imports Skills and cost of doing business

Business advisory Imitation organizations

Source: Original table for this publication. Note: MNE = multinational enterprise; PROs = public research organizations; R&D = research and development.

middle-income countries. The figure shows capture the overall location of these countries three clusters that, with some few exceptions, on the continuum of innovation capabilities. resemble the levels of capabilities depicted And, as described below, these different lev- earlier in figure 2.3: els of capabilities imply that countries in the region need to take differentiated approaches • Cluster 1 includes the high-income coun- to innovation policy. (Directions for policy are tries in the region, such as Japan, Korea, discussed further in chapter 6.) and Singapore, which are clustered at the frontier. Driven primarily by a large increase in the number of patents and sci- Effectively fostering innovation—both entific production, China has also joined diffusion and invention—requires a this first group with more-advanced tech- graduated approach of moving firms nological capabilities. toward the frontier, recognizing their • Cluster 2 includes Malaysia, Vietnam, heterogeneous capacities to innovate Thailand, and Mongolia, with only An NIS comprises a multiplicity of actors, Malaysia being close to the leading group. institutions, and policies, which poses • Cluster 3 includes the Philippines, Indonesia, significant challenges when identifying and Cambodia. weaknesses in innovation policy frameworks Although the choice of three clusters is and, especially, when deciding where to focus somewhat arbitrary, figure 2.7 does reasonably public policy efforts to encourage innovation. Conceptual Framework and Stylized Facts 73

FIGURE 2.7 Developing East Asian countries occupy three distinct clusters with respect to innovation capabilities

60 Korea, Rep.

China Japan Vietnam Singapore 40 Thailand Mongolia Malaysia

Indonesia Cambodia

Innovation output subindex score 20

Philippines

0 20 40 60 80 Innovation input subindex score

Cluster 1 Cluster 2 Cluster 3

Source: World Bank elaboration from Global Innovation Index (GII) data (https://www.globalinnovationindex.org/). Note: The “innovation input” subindex scores aspects such as infrastructure, institutions, research and development, and human capital quality. The “innovation output” subindex scores the quality of knowledge, technology, and creative outputs of the economy. Among the 10 developing East Asia countries in this study (Cambodia, China, Indonesia, Lao PDR, Malaysia, Mongolia, Myanmar, the Philippines, Thailand, and Vietnam), Lao PDR and Myanmar are excluded for lack of data.

This is especially the case for most low- and achieved rapid technological convergence middle-income economies, where multiple based on policies that addressed their inno- elements of the NIS are underdeveloped or vation capabilities gaps (Cirera and Maloney have significant weaknesses, including a lack 2017). (See chapter 6, table 6.3, for the evo- of well-functioning agencies and institutions, lution of Korean policies.) For firms to reach mistargeted policies, and a lack of enabling the technological frontier, policies and insti- conditions or complementary factors. tutions must focus on building the capabili- Innovation policy thus becomes complex, ties of the private sector and developing the and the critical question is, therefore, how innovation system elements that are absent, to prioritize policies that will be effective in not functional, or poorly aligned with need. enabling innovation given the country’s capa- Cirera and Maloney (2017) propose bilities, which in turn depend on the abilities assessing the adequacy of policies and insti- of firms to innovate and of agencies to design tutions through the lens of a “capabilities and implement appropriate policies. escalator”—to reflect the capacity of firms Based on the experiences of successful and country systems to absorb and use East Asian economies such as Japan, Korea, knowledge (figure 2.8). At the lower level and Singapore, an effective way of address- of capabilities, firms have mostly producing this problem and guiding policies is to tive, but few technological, capabilities; use a graduated approach. These economies hence the policy objective should be to build 74 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

those capabilities. On this lowest step of the Figure 2.8 presents an illustrative set of escalator, policies should focus on develop- policy instruments, corresponding to the dif- ing management quality, skills, and national ferent levels of the capabilities escalator. quality infrastructure. Where the business climate is weak, policies should focus on creat- Countries should support diffusion and ing an environment conducive to investment adoption as well as invention, prioritizing and knowledge diffusion, including through policies and allocating resources FDI and trade. consistent with existing capabilities On the middle step of the escalator, some firms have technological capabilities, but few The framework presented here does not imply have R&D and invention capabilities. Policies that only one type of policy applies to each in those countries should therefore focus on country. It does imply, however, that the expanding and strengthening technological policies and public resources for innovation capabilities while also supporting more firms should be well aligned with the capacities and in implementing R&D projects oriented the needs of the private sector. Thus, countries toward invention. with relatively low innovation capabilities— On the highest step, in countries where typically the region’s lower-middle-income firms have more sophisticated capabilities, the countries—are best advised to prioritize the goal of policy should be to enable invention adoption and diffusion of existing technolo- by supporting more complex, longer-term gies. As innovation capabilities rise, policy R&D projects. At this stage, countries also priorities should also shift, progressively require adequate intellectual property protec- focusing on the more technically advanced tion and will benefit from significant collabo- needs of leading firms. ration between industry and universities or Notably, even the region’s high-income, other knowledge providers. high-capacity countries such as Japan, Korea,

FIGURE 2.8 Appropriate policy instruments to foster innovation differ depending on the level of innovation capabilities

Long-term R&D programs Direct and indirect support to R&D Invention capabilities Collaborative projects Precommercial procurement Risk nance

Management extension Vouchers to collaborate Production capabilities Skills development National quality infrastructure (NQI) Incubation

Source: Adapted from Cirera and Maloney 2017. ©World Bank. Further permission required for reuse. Note: R&D = research and development. Conceptual Framework and Stylized Facts 75 and Singapore—as well as the United States Conclusions and Canada—offer support for technology adoption as well as invention, with different Rapid technological change, along with the sets of policies to encourage both dimensions diverse shocks that developing East Asia is of innovation. At any level of capabilities, the experiencing, means that technology adoption point is not to focus policies only on either for higher productivity needs to happen faster adoption or invention but rather to allocate and more broadly than in the past. Climbing more resources in a way commensurate with the capabilities escalator requires efforts to innovation capabilities. develop firms’ and countries’ innovation Taking a graduated approach ensures that capabilities and to progressively address the the different parts of the NIS can effectively key barriers on the supply of knowledge and support the development of the needed capa- firm absorption. bilities. The capacities of public agencies to Developing and implementing policies design and implement effective policies also that are comprehensive yet proportional to evolve with increasing capabilities in the local capabilities can pay large dividends in private sector. addressing existing constraints and enhancing innovation and technological development. It is important, however, that policy makers Supporting basic forms of innovation be realistic about the current strength of their also pays off countries’ institutions and policy-making Finally, it is important to emphasize that, capacity. If there is a mismatch between coun- contrary to some popular beliefs, more-basic tries’ capabilities and the types of policies forms of innovation, including imitation of and institutional reforms implemented, such products and processes, adoption of new reforms can be both costly and ineffective technologies, or increases in product qual- (Cirera and Maloney 2017). ity are important for productivity growth. This chapter has highlighted some key ele- Thus, policies that successfully encourage ments for understanding how policy makers innovations that are new to the firm or new can accelerate innovation in developing East to the domestic market can have significant Asia: returns. This is clear from some growth mod- • Innovation needs to be understood in its els such as presented in Madsen, Islam, and broad sense, ranging from basic upgrad- Ang (2010) but also from microeconomic ing to invention, and from imitation evidence. and adoption of new technologies to Studies looking at the relationship discovery. between innovation and productivity rarely • Innovation and technology are not static distinguish between types of innovation concepts; they change over time and and largely find a positive impact of inno- have accelerated during certain historical vation on productivity. (See Mohnen and periods through the invention of general Hall [2013] for a survey of the evidence purpose technologies that enable the mul- from OECD countries.) In one of the few tiplication and use of new technologies studies looking at this more directly, using that change production processes. The a panel of Turkish firms, Fazliog˘lu, Dalgiç, world is currently undergoing one such and Yereli (2019) find positive returns to period—the fourth industrial revolution, innovations that are new to the firm or new or Industry 4.0. to the local market only. This less sophis- • Supporting innovation and technology ticated type of innovation, in addition to adoption requires understanding the sys- allowing firms to remain competitive, also temic nature of innovation and building generates positive returns. the private sector’s innovation capabilities, 76 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

as demonstrated by the rapid technological 2. Figure 2.2 excludes the 10 developing East catch-up in Japan, Korea, and Singapore. Asia countries in this study (Cambodia, This process is often gradual but needs to China, Indonesia, the Lao People’s Democratic be accelerated. Republic, Malaysia, Mongolia, Myanmar, • Designing policies and institutions the Philippines, Thailand, and Vietnam) because, although their Enterprise Survey to achieve this transition to a more data distinguish between innovations that are innovation-based growth model requires new to the firm and new to the market, they keeping in mind the capabilities of the pri- do not distinguish between innovations that vate sector. Most developing East Asian are new to the local market and those that are countries are clustered in two main groups new to the international market. on the capabilities escalator—those on the 3. Although the preceding discussion focuses bottom step and those in the middle. These on countries as a unit, there is significant groups likely require a different mix of heterogeneity within countries regarding policy instruments and associated alloca- innovation activity. Even in countries such tions of public resources. as Korea that have reached the technological • Policies should support both diffusion and frontier (based on several metrics, such as R&D, robotization, patents, and income per invention, but for most of developing East capita), there is considerable heterogeneity Asia, the priority should be the diffusion in firms’ innovation performance. This of existing technologies and the building heterogeneity and its relevance to promoting of managerial practices while maintain- innovation-led growth in developing East Asia ing some instruments to support invention is examined in more detail in chapter 3. among those few firms capable of pushing 4. Hallward-Driemeier and Nayyar (2019), the frontier. for example, find that the intensity of robot • There is no evidence that convergence use in high-income countries has a positive to the technological frontier can hap- impact on FDI growth in low- and middle- pen quickly. On the contrary, there are income countries up to some threshold, after reasons to expect a greater divergence if which increased robotization negatively affects FDI in these countries. Most countries governments do not (a) act fast, (b) pri- in their study are still below the inflection oritize diffusion, (c) remove policy biases point, however. Artuc, Bastos, and Rijkers against services, and (d) support innova- (2018) find that greater use of robots in tion in business models and organizational Organisation for Economic Co-operation and changes, especially given the risk of tech- Development (OECD) economies increases nological disruptions in some markets. demand for imports from low- and middle- income countries. Dachs, Kinkel, and Jäger (2019) find a positive relationship between Notes Industry 4.0 technologies and reshoring of production for a sample of European 1. An important empirical literature in manufacturing firms. De Backer et al. (2016) knowledge functions estimates the elasticity also find some evidence, albeit weak, of of patenting with respect to R&D. reshoring affecting capital investment. For Trajtenberg (2001) and Griliches (1990) an overview of the literature on reshoring offer as a stylized fact that, in cross-section, and its impact on investment, see Wan et al. patents are probably roughly proportional (2019). to R&D, implying an elasticity of unity. 5. Comin and Mestieri (2018) define “Western The elasticities decrease when using panel countries” to include Australia, Austria, data estimation and for low- and middle- Belgium, Canada, Denmark, Finland, France, income countries (see Bosch, Lederman, and Germany, Italy, Japan, the Netherlands, New Maloney 2005), suggesting some decreasing Zealand, Norway, Sweden, Switzerland, the returns to scale to knowledge. United Kingdom, and the United States. Conceptual Framework and Stylized Facts 77

6. The CHAT dataset (http://www.nber.org Alm, Richard, and W. Michael Cox. 1996. “The /data/chat) is an unbalanced panel dataset Economy at Light Speed: Technology and with information on the adoption of over 100 Growth in the Information Age and Beyond.” technologies in more than 150 countries since In Annual Report 2016, Federal Reserve Bank 1800. of Dallas, 2–17. 7. Nelson and Rosenberg (1993) define the Artuc, Erhan, Paulo Bastos, and Bob Rijkers. NIS narrowly as “a set of institutions 2018. “Robots, Tasks and Trade.” Policy whose interactions determine the innovative Research Working Paper 8674, World Bank, performance of national firms,” especially Washington, DC. those supporting R&D. Other authors, such Autor, David H., and David Dorn. 2013. “The as Lundvall (1992), include a broader set Growth of Low-Skill Service Jobs and the of institutions such as national education Polarization of the U.S. Labor Market.” systems, labor markets, financial markets, American Economic Review 103 (5): 1553–97. intellectual property rights policies, Bosch, Mariano, Daniel Lederman, and William F. competitive product markets, and welfare Maloney. 2005. Patenting and Research and regimes. A challenge with applying some Development: A Global View. Washington, of the variations of these frameworks is DC: World Bank. that they can be too broad and lead to Bresnahan, Timothy. 2010. “General Purpose different interpretations of what matters Technologies.” In Handbook of the Economics for innovation, including the key actors and of Innovation, vol. 2, edited by Bronwyn policies. This makes analyzing policy and H. Hall and Nathan Rosenberg, 761–91. formulating clear recommendations difficult. Amsterdam: Elsevier. 8. Maloney (2017) offers a way of conceptualizing Caselli, Francesco, and Wilbur John Coleman II. the NIS that has two major advantages: (a) it is 2006. “The World Technology Frontier.” more specific about which elements to include American Economic Review 96 (3): 499–522. and assess (where to set the boundaries), and Cirera, Xavier, Marcio Cruz, Stefan Beisswenger, (b) it includes some important considerations and Gregor Schueler. 2017. “Technology related to the costs of doing business and Adoption in Developing Countries in the Age investment barriers, which can be important of Industry 4.0.” Unpublished manuscript, deterrents of innovation and have often been World Bank, Washington, DC. omitted from previous analyses. Cirera, Xavier, and William F. Maloney. 2017. The Innovation Paradox: Developing-Country Capabilities and the Unrealized Promise of References Technological Catch-Up. Washington, DC: World Bank. Acemoglu, Daron. 2010. “When Does Labor Comin, Diego A., and Bart Hobijn. 2009. “The Scarcity Encourage Innovation?” Journal of CHAT Dataset.” NBER Working Paper Political Economy 118 (6): 1037–78. No. 15139, National Bureau of Economic Acemoglu, Daron, Ufuk Akcigit, Harun Alp, Research, Cambridge, MA. Nicholas Bloom, and William Kerr. 2018. Comin, Diego, Bart Hobijn, and Emilie Rovito. “Innovation, Reallocation, and Growth.” 2006. “Five Facts You Need to Know About American Economic Review 108 (11): Technology Diffusion.” NBER Working Paper 3450–91. No. 11928, National Bureau of Economic Acemoglu, Daron, and Pascual Restrepo. 2018. Research, Cambridge, MA. “The Race between Man and Machine: Comin, Diego, and Martí Mestieri. 2018. “If Implications of Technology for Growth, Technology Has Arrived Everywhere, Why Factor Shares, and Employment.” American Has Income Diverged?” American Economic Economic Review 108 (6): 1488–542. Journal: Macroeconomics 10 (3): 137–78. Acemoglu, Daron, and Fabrizio Zilibotti. 2001. Dachs, Bernhard, Steffen Kinkel, and Angela “Productivity Differences.” Quarterly Journal Jäger. 2019. “Bringing It All Back Home? of Economics 116 (2): 563–606. Backshoring of Manufacturing Activities and 78 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

the Adoption of Industry 4.0 Technologies.” Hallward-Driemeier, Mary C., and Gaurav Journal of World Business 54 (6): 101017. Nayyar. 2019. “Have Robots Grounded the De Backer, Koen, Carlo Menon, Isabelle Flying Geese? Evidence from Greenfield FDI Desnoyers-James, and Laurent Moussiegt. in Manufacturing.” Policy Research Working 2016. “Reshoring: Myth or Reality?” Science, Paper 9097, World Bank, Washington, DC. Technology and Industry Policy Papers No. Hobday, Michael. 1995. “East Asian Latecomer 27, Organisation for Economic Co-operation Firms: Learning the Technology of Electronics.” and Development, Paris. World Development 23 (7): 1171–93. DFKI (German Research Centre for Artificial Lundvall, Bengt-Åke, ed. 1992. National Systems Intelligence). 2011. “Industrie 4.0: Mit of Innovation: Towards a Theory of Innovation dem Internet der Dinge auf dem Weg zur 4. and Interactive Learning. London: Pinter. Industriellen Revolution.” Madsen, Jakob B., Md. Rabiul Islam, and Fazlıog˘lu, Burcu, Bas¸ak Dalgıç, Ahmet Burçin James B. Ang. 2010. “Catching Up to the Yereli. 2019. “The Effect of Innovation Technology Frontier: The Dichotomy between on Productivity: Evidence from Turkish Innovation and Imitation.” Canadian Journal Manufacturing Firms.” Industry and of Economics 43 (4): 1389–411. Innovation 26 (4): 439–60. Maloney, William. 2017. “Revisiting the National Ford, Martin. 2015. The Rise of the Robots: Innovation System in Developing Countries.” Technology and the Threat of a Jobless Future. Policy Research Working Paper 8219, World New York: Basic Books. Bank, Washington, DC. Frey, Carl Benedikt, and Michael A. Osborne. 2017. Mansfield, Edwin. 1961. “Technical Change and “The Future of Employment: How Susceptible the Rate of Imitation.” Econometrica 29 (4): Are Jobs to Computerisation?” Technological 741–66. Forecasting and Social Change 114: 254–80. Mohnen, Pierre, and Bronwyn H. Hall. 2013. Goñi, Edwin, and William F. Maloney. 2017. “Innovation and Productivity: An Update.” “Why Don’t Poor Countries Do R&D? Eurasian Business Review 3 (1): 47–65. Varying Rates of Factor Returns across the Nelson, Richard R., and Nathan Rosenberg. Development Process.” European Economic 1993. “Technical Innovation and National Review 94: 126–47. Systems.” In National Innovation Systems: A Gort, Michael, and Steven Klepper. 1982. “Time Comparative Analysis, edited by Richard R. Paths in the Diffusion of Product Innovations.” Nelson, 3–21. Oxford and New York: Oxford Economic Journal 92 (367): 630–53. University Press. Griliches, Zvi. 1957. “Hybrid Corn: An OECD and Eurostat. 2018. Oslo Manual 2018: Exploration in the Economics of Technological Guidelines for Collecting, Reporting and Using Change.” Econometrica 25 (4): 501–22. Data on Innovation. 4th ed. Paris: OECD; Griliches, Zvi, 1990. “Patent Statistics as Economic Luxembourg: Eurostat. Indicators: A Survey.” Journal of Economic Schwab, Klaus. 2016. The Fourth Industrial Literature 28 (4): 1661–707. Revolution. Geneva: World Economic Forum. Hall, Bronwyn H. 2006. “Innovation and Skinner, Jonathan, and Douglas Staiger. 2007. Diffusion.” In The Oxford Handbook of “Technology Adoption from Hybrid Corn to Innovation, edited by Jan Fagerberg, David Beta-Blockers.” In Hard-to-Measure Goods C. Mowery, and Richard R. Nelson, 459–84. and Services: Essays in Honor of Zvi Griliches, Oxford and New York: Oxford University edited by Ernst R. Berndt and Charles R. Press. Hulten, 545–70. Chicago: University of Hall, Bronwyn H., and Beethika Khan. 2003. Chicago Press. “Adoption of New Technology.” NBER Stoneman, Paul, and Giuliana Battisti. 2010. “The Working Paper No. 9730, National Bureau of Diffusion of New Technology.” In Handbook Economic Research, Cambridge, MA. of the Economics of Innovation, vol. 2, edited Hallward-Driemeier, Mary, and Gaurav Nayyar. by Bronwyn H. Hall and Nathan Rosenberg, 2018. Trouble in the Making? The Future 733–60. Amsterdam: Elsevier. of Manufacturing-Led Development. Tether, Bruce S. 2003. “The Sources and Aims of Washington, DC: World Bank. Innovation in Services: Variety between and Conceptual Framework and Stylized Facts 79

within Sectors.” Economics of Innovation and Country Matter?” Journal of Purchasing and New Technology 12 (6): 481–505. Supply Management 25 (4): 100551. Trajtenberg, Manuel. 2001. “Government Support Weking, Jörg, Maria Stöcker, Marek Kowalkiewicz, of Commercial R&D: Lessons from the Israeli Markus Böhm, and Helmut Krcmar. 2019. Experience.” Working Papers from Tel Aviv “Leveraging Industry 4.0—A Business Model No. 2001–8, Tel Aviv University, Israel. Pattern Framework.” International Journal of Wan, Li, Guido Orzes, Marco Sartor, and Guido Production Economics 225: 107588. Nassimbeni. 2019. “Reshoring: Does Home

Technology Adoption and Diffusion: A Firm-Level 3 Perspective

Introduction adoption and the reasons why firms in developing East Asia may not be innovating and As emphasized in previous chapters, develop- adopting new technologies. Understanding ing East Asia is facing great challenges associ- the extent of technology adoption and dif- ated with changing production structures and fusion, along with the heterogeneity across global shocks that require the region to accel- firms, is critical to determining the right com- erate its diffusion of technologies and its rate bination of policies that can help accelerate and sophistication of innovation. Chapter 2 technology diffusion and adoption as well as argued that moving to a more innovation- invention in the region. led growth model requires a focus on facilitating the diffusion of technology as well as on building the capabilities within firms to Is East Asia converging enable more sophisticated forms of innova- with, or diverging from, the tion and discovery. technological frontier? Why Building on the benchmarking exercise pre- diffusion matters sented in chapter 1, this chapter looks more deeply at where the region stands in terms of Although inventions and new technologies innovation and technology. It first discusses push the technological frontier and offer the the extent of technology adoption and diffu- possibility for large increases in productivity, sion and analyzes how ready the region is to it is the diffusion of these innovations and meet the new challenges ahead. The chapter new technologies that ultimately determines then examines the considerable heterogeneity the pace of income and productivity growth in innovation performance across the region’s (Hall and Khan 2003). Historical estimates firms. Given their capabilities, firms’ innova- by Comin and Mestieri (2018) suggest that tion in the region is primarily about adoption differences in the evolution of technology and imitation, but there are also frontier firms diffusion since the industrial revolution have and locations engaged in discovery and gen- generated an annual difference in growth erating new technologies. Finally, the chapter between countries of 0.75 percentage points— describes some of the potential drivers of responsible for 80 percent of the Great 82 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

Divergence (also known as the “European older technologies such as steamships (invented miracle”) starting in the 19th century as in the later 1700s) was almost 170 years in European countries moved ahead of other developing East Asia, compared with 60 years high-income countries. The post–World War in OECD countries. In contrast, the gap in the II growth experience of Western Europe and median adoption lag for the internet (invented Japan can be explained, in large part, by the in the 1980s) was merely five years. rapid adoption of existing foreign technolo- At the same time, gaps in intensity of tech- gies (Comin and Hobijn 2010). nology use—how widely new technologies Similarly, estimates by Santacreu (2015), have been adopted—have widened over time using imports of intermediate goods, sug- (figure 3.1, panel b). Comin and Mestieri gest that technology diffusion is an especially (2018) suggest this gap is a main driver of important source of growth for less-developed the divergence in income between Western countries. She estimates that between 1996 countries and the rest of the world in the past and 2007, 65 percent of overall labor pro- 200 years. For developing East Asia, the esti- ductivity growth in low- and middle-income mated intensity of technology use has declined countries can be explained by foreign inno- at an annual rate of 0.50 percent relative to vations embodied in imports, whereas for OECD countries, implying a divergence in the Organisation for Economic Co-operation and intensity of use of new technologies during Development (OECD) countries, this figure the same period. is 35.5 percent. This process of technological catch-up is also true for the advanced econo- Different technologies diffuse mies in East Asia: historical evidence suggests at different speeds that the development miracles of Japan and then the Four Asian Tigers—Hong Kong SAR, A more nuanced view of technological con- China; the Republic of Korea; Singapore; and vergence highlights heterogeneity across Taiwan, China—over the past 50 years all technologies. Figure 3.2 shows the diffusion involved substantial reduction in technology in East Asia and the United States of two adoption lags relative to the OECD (Comin different technologies: the internet and indus- and Hobijn 2011). trial robots. The internet—a general purpose technology (GPT) that enables many other tech- Developing East Asia is converging in nologies—has diffused rapidly, becoming adoption lags but diverging in intensity ubiquitous in all countries in recent years. of use However, the diffusion path for robots in the Recent evidence suggests, however, that while region is different. At the forefront, Japan and new technology is arriving in developing East Korea started adopting robots at the same Asia at an accelerating pace, the region’s gaps time and now use robots in industry at higher with high-income economies in the intensity intensity—that is, more widely—than in the of technology use may be increasing. United States. China’s adoption of industrial As discussed in chapter 2, Comin and robots has been the fastest in the region, Mestieri (2018) suggest that the technology growing exponentially in the past 20 years. adoption lag—defined as the number of years However, many other countries in the region it takes for a technology to arrive to a coun- seem unlikely to converge with the fron- try after invention—has narrowed significantly tier anytime soon, probably because few of between high-income and lower-income coun- them are currently specialized in sectors that tries. Using their estimates (from chapter 2, can use robots intensively. This suggests that figure 2.5), the same pattern appears to hold when thinking about specific technologies, true for developing East Asia (figure 3.1, panel one also needs to understand their uses within a). For example, the median adoption lag for the firm. Technology Adoption and Diffusion: A Firm-Level Perspective 83

FIGURE 3.1 Technology adoption lags in developing East Asia are converging with those of OECD countries, but intensity of technology use is diverging

a. Adoption lag, 1700s–2000

200

150 Steel (Bessemer, open hearth)

b. Intensity of use, 1700s–2000

0 Ships

–0.5 Railway passengers Spindles Mail

–2.5 Blast oxygen furnace 1800 1850 1900 1950 2000 Year of technology invention Fitted line, OECD Fitted line, developing East Asia

FIGURE 3.2 Data on internet and robot use in East Asia tell two different tales of diffusion and convergence with the frontier

a. Internet, 1990–2018 b. Robots, 1990s–2018 0 0

–5 –5

–10 –10 ln of ratio internet users to population ln of ratio industrial robots to employment –15 –15 1990 2000 2010 2020 1990 2000 2010 2020 Year Year United States China Hong Kong SAR, China Indonesia Japan Cambodia Korea, Rep. Myanmar Mongolia Malaysia Philippines Thailand Vietnam

Source: Internet data from the Cross-Country Historical Adoption of Technology (CHAT) database (Comin and Hobjin 2010), updated using the World Development Indicators (WDI) database for more recent years; robot data from the International Federation of Robotics; employment data from the WDI database. Note: ln = natural logarithm.

A firm-level view of diffusion suggests by business function. (For more details on the a ladder of complexity in the use of FAT survey, see annex 3A.) different technologies Moving to more sophisticated technologies may require capabilities across a range of To fully unpack this heterogeneity and business functions that are not immediately understand diffusion, a firm’s perspective of available and need to be developed within the technology adoption is necessary. A focus firm. The complexity of implementing more on whether a technology is being used, for sophisticated technologies translates into low what task, and how intensively, is needed if intensity of use, as the Vietnam FAT results one wants to measure whether technology show: brings better performance. Indeed, the question is not only whether a firm uses internet • In manufacturing, for fabrication, most or blockchain but also for what purpose and Vietnamese firms (70 percent) use oper- with what intensity. ator-controlled machines, only 9 percent For each task, a firm can use different use computer-controlled machines, and types of technologies, which can range from less than 1 percent use more-advanced less to more sophisticated. This can be seen technologies like robots, 3D-printers, or using data from the Firm-level Adoption of additive manufacturing (figure 3.3). Technology (FAT) survey of Vietnamese firms • In retail services, for inventory manage- (Cirera, Comin, Cruz, and Lee 2020), which ment, 63 percent of firms use computer shows the most frequently used technology, databases with manual updates, Technology Adoption and Diffusion: A Firm-Level Perspective 85

FIGURE 3.3 The intensive use of cutting-edge technology for manufacturing, retail, and agriculture remains limited in Vietnam

a. Manufacturing: use of fabrication technology Other advanced manufacturing processes (e.g. as laser, plasma sputtering, high-speed machine, e-beam, micromachining)

Additive manufacturing including rapid prototyping and 3-D printers

Robots Computer numerical controlled machine or other machines controlled by computers Machines controlled by operators without computers

Manual processes

0 20 40 60 80 Share of rms adopting technology, by type of use (%)

b. Retail: use of inventory technology

Computer databases with manual updates

Handwritten record keeping

0 10 20 30 40 50 60 70 Share of rms adopting technology, by type of use (%)

Mechanical application of herbicide, fertilizer, or pesticide

Manual application of herbicide, fertilizer, or pesticide

0 10 20 30 40 Share of rms adopting technology, by type of use (%)

Source: World Bank estimates, using the 2020 Firm-level Adoption of Technology (FAT) survey of Vietnamese firms (Cirera, Comin, Cruz, Lee, and Martins-Neto 2020).

25 percent use warehouse management • In agriculture, almost one-third of firms still systems with specialized software, while use manual weeding and pest techniques as only 1 percent use advanced technologies the predominant technology, whereas the such as automated storage and retrieval use of automated precision agricultural systems. techniques is almost nonexistent. 86 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

These results illustrate the quality ladder in of the virus, as Korea has demonstrated technology use and indicate that most firms (box 3.1). Social-distance measures have also are at a significant distance from the techno- led to a need for more flexibility in the imple- logical frontier. mentation of business functions and tasks: firms are facing challenges associated with moving certain functions to home-based Firms are also far from the intensive use work, increasing “contactless” sales, and of digital technologies and were largely enforcing social distancing on the manufac- unprepared to face new challenges from the COVID-19 crisis turing shop floor. Technologies to solve or minimize these The need for accelerated technology diffusion challenges include digital e-commerce has become more apparent with the recent platforms and more flexible manufactur- COVID-19 pandemic. Technology can play ing processes. Despite the early evidence of an important role in containing transmission a sharp increase in the use of online sales and platforms as a response to the pandemic

BOX 3.1 Republic of Korea’s use of technology to implement contact tracing and testing strategies in response to COVID-19

In its fight against COVID-19, the Republic of Korea and (c) efficient management of access to personal has made effective use of technology and innovation. information (MOEF 2020). Using the country’s information and communication Complementary to the EISS are the mobile Self- technology (ICT) infrastructure and accumulated out- Diagnosis App and Self-Quarantine Safety Protection puts from national research and development (R&D) App, which the government developed to support the programs, the Korean government acted swiftly to monitoring of inbound travelers and those under use advanced technologies and innovative approaches self-quarantine, respectively (Republic of Korea in its response measures (Oh et al. 2020). In particu- 2020). These digital tools support government offi- lar, its contact tracing and aggressive testing strategies cials and researchers throughout the four stages of are considered to have significantly contributed to contact tracing: investigation, risk assessment, con- flattening the coronavirus curve in Korea. tact classification, and contacts management. At the core of Korea’s contact tracing strategy is Korea’s COVID-19 testing strategy also benefited the Epidemic Investigation Support System (EISS). from technology and innovation. For instance, domes- The Korean government introduced this data sys- tic biotech firms used artificial intelligence (AI)-based tem in March 2020 to support rapid, accurate epi- big data systems and high-performance computing to demiological investigations in close coordination dramatically shorten development of a coronavirus with the Korean National Police Agency, Credit diagnostic kit (from several months to around two Finance Association of Korea, telecommunications weeks), which was aided by the government’s swift companies, and credit card companies. Information emergency use authorization. This rapid development and data collected from these information partners of test kits enabled universal testing in the country become available on the COVID-19 data platform from the early phases of the COVID-19 pandemic. to enable epidemiological researchers and health Furthermore, Korea was one of the first countries that officials to identify transmission routes and con- established walk-through and drive-through screening duct “big data” analyses for the prediction of areas stations. This innovative approach to testing, along where people may be vulnerable to infection (Park with the rapid development of diagnostic kits, helped et al. 2020). Additionally, the EISS enables (a) auto- the country meet its extensive testing needs while mated analysis of the movement of confirmed cases minimizing the risks involved. within 10 minutes, (b) close interinstitutional coordination through real-time information exchanges, Source: Frias, Lee, and Shin 2020. Technology Adoption and Diffusion: A Firm-Level Perspective 87

(Apedo-Amah et al. 2020), a key question is 30 percent of small firms. Access to online how well firms in developing East Asia are platforms is also relatively low relative to its equipped to use them and to adapt to both potential, especially for small and medium- preexisting and newly emerging challenges. size firms. Moreover, only 7 percent of firms To better assess firms’ readiness to adopt have adopted cloud computing for their busi- technology, a “digital readiness index” was ness processes. developed, using the FAT survey for Vietnam As for the use of digital technologies in gen- (Cirera, Comin, Cruz, Lee, and Martins- eral business functions (figure 3.4, panel b), Neto 2020). The index measures the extent the indexes—percentages of firms that use of firms’ digitalization by business function.1 a digital technology as the main one for the Figure 3.4 shows the results of the indexes specified business function—similarly suggest (ranging between 0 and 100) for Vietnam. substantial room for increased digitalization. Panel a focuses on use of digital infrastructure On average across different business func- and platforms, while panel b focuses on the tions, only 20 percent of Vietnamese firms use extent of digitalization of general business fully digitalized processes to perform general functions. business functions ranging from marketing, Regarding digital infrastructure use, payment methods, and production planning figure 3.4, panel a, shows that almost all to sales and supply chain management.2 firms have access to the internet, but far fewer The results show considerable room for firms have their own websites or use social improvement in digitalization across different media for marketing or sales. Large firms firm functions and tasks. For example, most have a 60 percent probability of having their firms in Vietnam do not use digital technolo- own website, in contrast to slightly more than gies as their main tool for sales. Although

FIGURE 3.4 Digital readiness indexes show widespread internet access among Vietnamese firms, but few are fully equipped to use digitalized processes for primary business functions

a. Use of digital infrastructure, by rm sizea b. Digital readiness, by general business functionb 100 Business administration

80 Marketing

Payment methods 60 Production planning 40 Quality control Predicted probability (%) probability Predicted 20 Sales

Sourcing and procurement 0 Internet Website Social media 0 10 20 30 40 50 Small Medium Large Intensive margin digital index

Source: Cirera, Comin, Cruz, Lee, and Martins-Neto 2020. Note: The indexes were developed using data from the Firm-level Adoption of Technology (FAT) survey of Vietnamese firms in manufacturing, retail, and agriculture. For more details about the FAT data, see annex 3A. a. Panel a shows the predicted probability of firms with internet, own website, and social media on size—small (5–19 employees); medium (20–99); and large (100+)—from the Probit regressions, while controlling for sector and region. All estimates are weighted by sampling and country weights. Vertical lines show confidence intervals. b. Panel b shows the percentages of firms that use a digital technology as the main one for the specified business function. 88 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

around 30 percent of the firms can do some from regressing the logarithm of value sales online, only 2.5 percent of those firms added per worker on a technology index that sell directly to consumers use some digi- and sector dummies to control for differ- tal sales method—either social platforms or ent production functions, by sector. Firms (most often) their own websites—more inten- that use more sophisticated technologies in sively than other methods. general business functions such as human Nearly half (47 percent) of Vietnamese resource (HR) management, supply chain firms use digital technologies for market- management, or sales (the extensive mar- ing and advertising. The current use of digi- gin, shown in panel a) and use them more tal technologies in business administration intensively (the intensive margin, shown in remains low (31 percent), however, implying panel b) tend to have higher value added low ability to provide the flexibility to enable per worker. As expected, the more inten- workers to cope with potential disruptions sively the sophisticated technology is used created by COVID-19 or other shocks. (panel b), the stronger the relationship.

Technology adoption generates a Heterogeneity in the pattern and positive productivity dividend diffusion of innovation across space, sectors, and firms To make the most of existing and emerging technologies, it will be important for firms in As argued earlier, what matters most for developing East Asia to continue accumulat- a country’s growth performance is how ing innovation capabilities—first by upgrad- rapidly technology and innovation diffuse ing their processes using digital technologies, across enterprises in a country. Indeed, and then by adopting more sophisticated the productivity and technological divide Industry 4.0 tools. between the leading and lagging firms in The dividends from doing so, in terms developing East Asia is likely the conse- of productivity and competitiveness gains, quence of slow diffusion within countries. can be large. As discussed in previous chap- And as Andrews, Criscuolo, and Gal (2016) ters, the empirical literature that examines suggest, this technological divide is a likely the relationship between innovation and cause of the global productivity slowdown productivity indicates that innovation of since 2000. This section shows that there all types generally has a positive impact on is substantial heterogeneity in the pattern firm-level productivity. (For surveys of the of technology adoption and innovation evidence, see Hall [2011] and Mohnen and across and within countries, sectors, and Hall [2013].) in some cases, even within the same firms. Recent analysis using World Bank This heterogeneity merits the concerns of Enterprise Survey (WBES) data produces policy makers because, if persistent, it could similar findings for developing East Asian further slow the region’s productivity and countries. Product and process innovation growth. among the region’s firms, including both new-to-market and new-to-firm innovation, Countries across developing East Asia is associated with higher labor productivity show significant heterogeneity in firm- and higher revenue total factor productivity level innovation performance (TFP), controlling for firm-level and market characteristics (de Nicola 2019). Firm-level measures of innovation activities, Data from the FAT survey in Vietnam using WBES data, highlight considerable reinforce the positive relationship between heterogeneity in firm-level innovation activ- technology adoption—a type of diffu- ity across the region. Close to 60 percent sion—and productivity at the firm level. of Chinese firms introduce a product or Figure 3.5 shows the conditional prediction service innovation, and 20 percent have a Technology Adoption and Diffusion: A Firm-Level Perspective 89

FIGURE 3.5 Technology adoption brings labor productivity gains to Vietnamese firms

a. Correlation between extensive margin b. Correlation between intensive margin index and labor productivitya index and labor productivityb

10.5 12.0

10.0 11.0

9.5 10.0 9.0

9.0 8.5 Log of value added per worker Log of value added per worker 8.0 8.0 1 2 3 4 5 1 2 3 4 5 Extensive GBF tech index Intensive GBF tech index

Source: Cirera, Comin, Cruz, Lee, and Martins-Neto 2020. Note: The figures show the conditional predictions (solid lines) and 95 percent confidence levels (dashed lines) from regressing the log of labor value added per worker on a technology index and sector dummy variables. The index (from 1 to 5) measures the sophistication of technology—1 being the least sophisticated and 5 the most sophisticated technology for a set of general business functions (GBF) including business administration (human resources, accounting, and so on); production or service operations planning; sourcing, procurement, and supply chain management; marketing and product development; sales; payments; and quality control. The indexes were developed using data from the Firm-level Adoption of Technology (FAT) survey of Vietnamese firms in manufacturing, retail, and agriculture. For more details about the FAT data, see annex 3A. a. The “extensive” index refers to the average among all GBFs of the most sophisticated technology used in each business function, even if marginally used. b. The “intensive” index refers to average among all GBFs of the most sophisticated technology used more intensively for each business function—that is, the main technology used by the firm.

foreign-licensed technology (figure 3.6). services sector to undertake a product or At the other end of the spectrum, less than process innovation. 15 percent of firms in countries such as However, these findings may paint a biased Myanmar and Thailand report having a picture of the true extent of innovation in product or service innovation, and a mere 5 manufacturing and services, being partly percent have any technology licensed from driven by the different nature of innovation foreign companies. As discussed in chapter 2 in these sectors (box 3.2). Among more nar- (figure 2.7), with the exception of China, rowly defined sectors, the share of firms that countries in the region are located in the undertook a product or process innovation lower two (out of three) clusters of innova- is highest in computer and related activities tion capabilities. (ICT) and other high-tech industries such as machinery and equipment, electronics, and chemicals (figure 3.7). Innovation activity also varies widely across sectors Firms in developing East Asia also differ Firms are not technology savvy in substantially in innovation outcomes across everything sectors—as measured by the incidence of firms implementing or adopting a new Finally, even within firms, there can be sig- product, process, or organizational or mar- nificant heterogeneity in the process of keting approach. Evidence from the WBES technology adoption. Evidence from the data suggests that manufacturing firms are FAT survey in Vietnam offers new insights significantly more likely than firms in the about the complexities of firms’ technology 90 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

FIGURE 3.6 Developing East Asian countries vary widely in firm-level innovation activity

40 Share of rms (%) 20

adoption decisions. The data suggest that the show, in fact, that within-firm variation in the use of sophisticated technologies can differ use of technology (figure 3.8, panels b and d) is significantly across business functions even larger than the cross-firm variation (figure 3.8, within a given firm. panels a and c). The radar diagrams in figure 3.8 illustrate both between- and within-firm heterogeneity in The pattern of heterogeneity in technology adoption, using data from a large innovation activity is also spatial, and a medium-size firm in the food process- demanding different policies across ing sector. They show that, for the same busi- geographical areas ness administration functions, these two firms can be very different in the extensive margin The unevenness of technological discovery of technology use (figure 3.8, panel a), but and diffusion across the region’s countries is the gap is minimal when use intensity is con- mirrored by innovation performance within sidered (figur e 3.8, panel b). Moreover, the these countries. Using data from Crunchbase, same firm (for example, Firm 1) can be near an investment and funding platform, map 3.1 the frontier in its use of food storage technol- shows the spatial distribution of companies ogy but far from the frontier in the use of input that have received early-stage equity between testing technology (figure 3.8, panel d). Cirera, 2007 and 2019—a proxy for the incidence Comin, Cruz, Lee, and Martins-Neto (2020) of innovative start-ups—across cities in Technology Adoption and Diffusion: A Firm-Level Perspective 91

BOX 3.2 Innovation in manufacturing versus services

There are major differences in the observed innova- activities suggests, for example, that knowledge- tion rates between manufacturing and services firms. intensive business services are just as likely as the Measured by having implemented a product or pro- high-technology manufacturing subsectors to have a cess innovation, services firms in developing East product or process innovation. (See Pires, Sarkar, and Asia (and elsewhere) appear to be significantly less Carvalho [2008] and Iacovone, Mattoo, and Zahler innovative than manufacturing firms (figure B3.1.1). [2013] for evidence from Portugal and Chile.) This At the same time, the most salient innovations por- likelihood of innovation also holds true for the com- trayed in the region’s popular press are examples puter and related activities (ICT) services sector in from service companies such as those providing developing East Asia (figure 3.7). However, because ride-sharing services (Grab, Go-Jek); e-commerce this sector accounts for only a small share (around (Alibaba); or gaming and payment platforms 3 percent) of firms in the services sector, the overall (Tencent). rate of innovation in services is driven by more traditional subsectors such as wholesale and retail. Reasons for this innovation gap between Second, innovations in the services and manu- manufacturing and services firms facturing sectors are inherently different from each What can explain this apparent discrepancy? One other. Ettlie and Rosenthal (2011) argue that key reason is the huge underlying heterogeneity in differences concern the alternative ways in which services. Services subsectors are very different from services sector enterprises implement the innovative each other in their input-to-output process and the process (often with less-formal research and develop- levels of technology of the inputs used (Pires, Sarkar, ment (R&D) processes or piloting and testing); the and Carvalho 2008). Detailed examination of service unique way in which service providers test customer concepts (given that services are often unique to the specific customer); and the more pronounced role of FIGURE B3.1.1 Manufacturing and services firms informal sourcing of innovation ideas by enterprise differ in rates of innovation, especially of new managers. products or processes Difficulty of measuring services innovation Service activities have a range of characteristics that 60 render measurement extremely difficult (Pires, Sarkar, and Carvalho 2008): the intangible nature of most services, the overlap of the moment of production and consumption, nonstorability, and the strong user- 40 producer links. Intangibility, in particular, means that service products often cannot be displayed in advance, and their qualities are not easily explained to the 20 customer. These same characteristics hinder efforts toward standardization. Therefore, intangible services

Share of rms (%) reportingShare lack the ability to create a temporary monopoly with innovation in past three years in past three innovation the help of some sort of patent protection to redeem 0 the innovation annuities. This lack of potential for Manufacturing Services protection may reduce the incentive for innovation Product Process Organization Marketing activities in the services sector (Hipp and Grupp 2005). Source: World Bank Enterprise Survey data, latest round. Because of the differences between manufacturing Note: The figure shows the average share and 95 percent confidence and services innovation, traditional measures to cap- interval in the pooled sample, accounting for country fixed effects. Data include manufacturing and services sector firms surveyed by the ture technological innovations have a manufacturing Enterprise Surveys in all 10 middle-income countries covered in this study: bias and do not adequately capture the full extent of Cambodia, China, Indonesia, Lao PDR, Malaysia, Mongolia, Myanmar, the Philippines, Thailand, and Vietnam. innovations in services. Innovation in services has been

box continues next page 92 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

BOX 3.2 Innovation in manufacturing versus services (continued)

characterized as “soft” or nontechnological, more likely has become increasingly blurred to the extent that to be incremental, and more on the organizational side. innovation in manufacturing assumes the form of “ser- (For a review, see Doloreux, Shearmur, and Rodriguez vicification.” At least 20 percent of Spanish manufac- [2016].) As a result, measures such as R&D and patents turing firms have introduced services in the recent past tend to systematically overlook innovation activities in (Santamaría, Nieto, and Miles 2012). And the vast services sectors (Hall 2011; Hipp and Grupp 2005).a majority of French manufacturers sell services in addi- When looking at other measures of organizational and tion to producing goods (Crozet and Milet 2017). In a marketing innovations, however, services firms in East contrasting phenomenon in the United States, sourc- Asia are not significantly different from those in the ing and design activities are now performed by “fac- manufacturing sector (figure B3.1.1). toryless goods producers” whose activities were once There appear to be some differences in the returns done within manufacturing (Bernard and Fort 2017). from different types of innovation in services rela- In addition, innovation of services has been found to tive to manufacturing, but the evidence has been increase productivity for manufacturing activities, both very mixed. In general, empirical evidence so far has within the firms and for downstream sectors using ser- suggested that the returns to innovation are similar vices inputs (Crozet and Milet 2017). between these two sectors (Audretsch et al. 2018; Musolesi and Huiban 2010). a. A firm’s accounting practices can also play a role. Perhaps the most extreme example of how reported R&D can underestimate innovation is in the case of In both the nature and returns to innovation, the Amazon. Amazon reports only “technology and content” investment, with no line between manufacturing and services innovations separate reporting of R&D efforts (Hernández et al. 2020).

developing East Asia. Across all countries and There is a high concentration of the sectors, innovative companies are concen- most-inventive activities trated in a few urban hubs but remain absent Another form of within-country heterogene- in large areas of all countries. The incidence ity is the duality of firm-level R&D invest- of innovative start-ups in China, in particular, ment and concentration of the most-inventive appears consistent with the pattern of patent activities. Figure 3.9 shows the distribution of filings, which suggest that technological and R&D intensity for selected East Asian coun- design capabilities are concentrated in a few tries and uses Israel as a benchmark coun- locations in Eastern China (Prud’homme and try. Each country has a separate sample year Zhang 2019). between 2012 and 2016. Across all countries, These patterns affirm that despite a level of the vast majority of firms perform no R&D technological achievement in major cities that at all (figure 3.9, panel a). Notably, of those might rival that of high-income countries, low firms that do invest in R&D, the distribu- levels of technological advancement in lag- tion of R&D intensity (as measured by R&D ging areas mean that, in aggregate, the coun- expenditure per employee) differs signifi- tries in the region are not as technologically cantly across countries (figure 3.9 panel b). advanced as high-income economies. With the exception of China, it appears that Although innovation, especially in tech the share of firms investing in R&D as well companies, tends to concentrate in specific as R&D intensity among these firms tends to tech hubs, spatial inequality suggests the need increase with a country’s income level, both for countries to have a differentiated develop- on average and across the whole distribution. ment strategy for firms outside these hubs and This result is not surprising: an empirical to build on preexisting endowments in other regularity in any economy is that the most- subregions within developing East Asia. inventive activities—those resulting from Technology Adoption and Diffusion: A Firm-Level Perspective 93

FIGURE 3.7 Firm-level innovation in East Asia also differs across sectors

100

40 Share of rms (%) 20

Recycling (n = 44) Computeractivities and related (n = 175) Construction (n = 432)

Precision instruments (n = 37) Hotel andOther restaurants manufacturing (n = 505) (n = 132) Machinery and equipment (nMetal = 322) and metal products (n = 742) Wood and wood products (n = 194) Wholesale and retail trade (n = 2,127) NonmetallicTextile, mineral clothing, products andRubber leather(n = 523)and (n plastics = 1,382) products (n = 668) Food, beverages, and tobacco (n = 1,108) Chemicals and chemical products (n = 519) Electrical and electronic equipment (n = 522) Transport, storage, and communications (n = 287)

Motor vehicles and other transport equipment (n = 177) Coke, re ned petroleum products, and nuclear fuel (n = 20)

Publishing, printing, and reproduction of recorded media (n = 134) New product or process New product New process

Source: World Bank Enterprise Survey data. Note: Data include manufacturing and services sector firms surveyed by the Enterprise Surveys in all 10 middle-income countries covered in this study: Cambodia, China, Indonesia, Lao PDR, Malaysia, Mongolia, Myanmar, the Philippines, Thailand, and Vietnam. formal R&D projects—are carried out by a above 0.9) and income per capita (correlation handful of frontier firms. For example, calcu- above 0.7). The country saw an explosion in lations using data from the European Union the number of patent applications as well as (EU) Industrial R&D Investment Scoreboard investment in the adoption of foreign tech- suggest that the top 10 public companies nologies, as measured by aggregate payments account for just under 20 percent of aggre- for technology licensing fees (figure 3.10). gate private sector R&D spending in the In fact, in 2011, China overtook the United United States and close to 10 percent in China States as the country with the most patent (Hernández et al. 2020). filings. The explosion of patents mentioned ear- Moreover, tracking the growth in different lier for China’s manufacturing sector, simi- types of patents and foreign citations suggests larly, can be accounted for by a tiny, highly that Chinese patent quality also exhibits a select group of Chinese companies in the ICT real and robust improvement over time (Wei, equipment industry (Eberhardt, Helmers, and Xie, and Zhang 2017). The extent of innova- Yu 2011). The companies’ growth in patent tion and improvements over time have been applications closely mirrors the countries’ considerably lower and slower in other devel- growth in R&D expenditures (correlation is oping East Asian countries, however. 94 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

FIGURE 3.8 Radar diagrams show substantial heterogeneity in technological sophistication within firms

1 1 Payment 2 Sourcing Payment 2 Sourcing 3 3 4 4 5 5 Sales Marketing Sales Marketing Firm 1 Firm 2 Firm 1 Firm 2

c. Sector-speci c functions, d. Sector-speci c functions, food processing: extensive margin food processing: intensive margin Input testing Input testing

Food Mixing/blending/ Food Mixing/blending/ storage cooking storage cooking 1 1 2 2 3 3 4 4 Packaging 5 Antibacterial Packaging 5 Antibacterial Firm 1 Firm 2 Firm 1 Firm 2

Overall, this large heterogeneity in inno- adoption in developing East Asia is not only vation activities and the high concentration diverse across countries but also uneven of invention and more-complex innovation within countries. To facilitate the diffusion activities in only a few locations and firms of technology and innovation, countries need suggests that diffusion is not occurring (or at more nuanced policies that can target key least not occurring fast or broadly enough to areas of underperformance. The challenge support higher productivity growth), which for policy makers is to understand the driv- is slowing the creation of higher-quality jobs ers of this heterogeneity. In other words, what that could significantly boost incomes in the inhibits firms from adopting technology and region. A policy priority should be to promote innovating more? the diffusion of knowledge and technology from frontier firms, domestic or foreign, to Who innovates? the rest of the economy. The first element in understanding what What inhibits innovation? inhibits innovation is the profile of innovators. Firms that innovate look consistently different Evidence from the previous section has high- from non-innovators in several dimensions. lighted that the process of invention and WBES data suggest that, on average, East Technology Adoption and Diffusion: A Firm-Level Perspective 95

MAP 3.1 The spatial distribution of innovative start-ups in developing East Asia shows a concentration in urban hubs, excluding large areas of all countries

a. Total VC-funded companies b. VC-funded agriculture companies

Number of VC-funded Number of VC-funded companies (all sectors) agriculture companies 400 400

100 100 50 50 10 10

c. VC-funded manufacturing companies d. VC-funded service companies

Number of VC-funded Number of VC-funded manufacturing companies service companies 400 400

100 100 50 50 10 10

IBRD 45415 | NOVEMBER 2020

Source: World Bank elaboration using Crunchbase data (https://www.crunchbase.com/). Note: The maps show distribution of innovative start-ups—defined as currently operating companies that received early-stage equity deals between 2007 and 2019—across cities in developing East Asia (Cambodia, China, Indonesia, Lao PDR, Malaysia, Mongolia, Myanmar, the Philippines, Thailand, and Vietnam). The size of the circles is proportional to the number of companies. City names were extracted from companies’ addresses and merged into a geographic information system (GIS) dataset using a fuzzy match algorithm; hence a subsample of companies is not included because of nonmatches. VC = venture capital. 96 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

FIGURE 3.9 There is significant duality in firm-level R&D investment

a. Distribution of rm-level R&D investment, selected countriesa

0.20

0.15

0.10 Probability density

0.05

0 −20 −10 0 10 20 Log R&D expenditure ($) per FT employee

b. Concentration of high R&D activities, selected countriesb 0.3

0.2

0.1 Probability density

0 5 10 15 20 Log R&D expenditure ($) per FT employee

Israel China Cambodia Malaysia Philippines

Source: World Bank Enterprise Survey data. Note: In both panels, Israel (dark solid line) represents the benchmark. FT = full-time; R&D = research and development. a. Panel a illustrates the distribution of R&D investment across all firms, by country. b. Panel b (an enlarged portion of panel a) illustrates the distribution of R&D intensity (R&D expenditure per employee) among firms that do invest in R&D, by country.

Asian firms that engage in all or at least one dimensions of innovation. Conditional on dimension of innovation—product, process, country, sector, and other firm characteris- organization, or marketing—tend to be larger, tics, exporters are 3.6 times more likely than more likely to export, and more likely to be nonexporters to have some innovation in all located in larger cities (table 3.1). four dimensions. In addition, firms in cities Market size appears to be particularly with populations exceeding 1 million are four important for firms engaged in multiple times more likely than those in cities with Technology Adoption and Diffusion: A Firm-Level Perspective 97

FIGURE 3.10 Inventions and foreign technology adoption have increased over time, by country, 2000–19

5 ln number of patent applications

0 10 20 30 ln payments (current US$) for use of foreign intellectual property

China Indonesia Cambodia Mongolia Malaysia Philippines Thailand Other countries

Source: World Development Indicators database. Note: “Inventions” reflects the number of patent applications. “Foreign technology adoption” reflects aggregate fees (in current US$) paid to other countries for technology licenses. The darker the shade, the more recent the year of data. ln = natural logarithm. populations below 50,000 to engage in all Case study research about the adoption of four dimensions of innovation. ICT in the United States, for example, has Considering other characteristics, young pointed out that the adoption of many GPTs and foreign-owned firms do not tend to often involves major reorganization and innovate more than other firms. However, co-investment in new products or business firm size is highly correlated with innovation models (Bresnahan and Trajtenberg 1995; propensity, particularly innovations in mul- Brynjolfsson, Hitt, and Yang 2002). tiple dimensions. Compared with small and These requirements are likely even more medium-size firms, large firms are eight times intense for Industry 4.0 technologies, as and three times more likely, respectively, to described in chapter 2. The acquisition or the participate in all four types of innovation. lack of these capabilities—and specifically, managerial and organizational practices— is fundamental to the process of upgrading Factors internal to the firm: firm (Sutton 2012) and other basic forms of capabilities and uncertainty innovation. One potentially important inhibitor of innovation is a lack of absorptive capabilities. Poor management quality is a deterrent to Innovation requires a range of capabilities that innovation enable firms to respond to market conditions, Evidence from a detailed matched employer- identify new technological opportunities, employee survey for China supports this develop a plan to exploit them, and then view: the degree of a firm’s innovativeness— cultivate the necessary resources to do so. as measured by a firm’s incidence of having 98 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

TABLE 3.1 Firms that innovate in developing East Asia tend to be larger, exporters, and located in larger cities

Firm characteristic Innovate in all four dimensionsa Innovate in at least one dimensiona Firm age 0.992 0.988 (0.0132) (0.00820) Share (%) of foreign ownership 0.989* 1.004 (0.00638) (0.00318) Exporter 3.595*** 1.535** (1.509) (0.294) Firm size: Medium (20–99 employees) 2.699** 1.852*** (1.136) (0.285) Firm size: Large (≥ 100 employees) 8.046*** 2.533*** (3.628) (0.444) Location size: > 1 million population 4.434** 3.617*** (2.567) (1.450) 2.559 2.838** Location size: 250,000–1 million population (1.881) (1.164) Location size: 50,000–250,000 population 1.422 6.354*** (1.196) (2.743) N 6,344 9,555

Source: World Bank Enterprise Survey data. Note: Odds ratio from logistic regression of firm innovation on firm characteristics, conditional on country (Cambodia, China, Indonesia, Lao PDR, Malaysia, Mongolia, Myanmar, Philippines, Thailand, Vietnam) and sector (manufacturing, trade, construction, transport, other services) fixed effects. The base for firm size is “Small (< 20 employees)” and for location size, “less than 50,000” population. Robust standard error in parentheses. a. Firms can introduce an innovation in all or any of the following four dimensions: product/service, process, organization, and marketing. *** p < 0.01 ** p < 0.05 * p < 0.1

a product or process innovation, R&D proj- Informational constraints and uncertainty ect, or patent—is positively associated with are barriers to innovation and technology the firm’s management quality and share adoption of highly (college) educated workers (Park Innovation is an inherently risky endeavor. and Xuan 2020). Data from the World The process of technology adoption, for Management Survey (WMS) show, however, example, is often characterized by significant that compared with US firms—a proxy for uncertainty (as to the future path of the the global frontier—firms in selected devel- technology and its benefits) and by limited oping East Asian countries are significantly information about the benefits, costs, and less-well-managed on average and along the even the technology’s viability (Hall 2004). whole distribution (figure 3.11).3 Informational constraints . In this environ- Moreover, poor overall innovation per- ment, informational failures may constitute a formance appears to be driven by manage- significant barrier to technology adoption and ment quality gaps that are generally larger innovation. A study of large Indian textile for the top firms.4 That is, the frontier firms plants, for example, argues that “informa- in developing East Asia perform dispropor- tional constraints” are an important factor tionally worse than the frontier firms in the leading firms not to adopt simple, appar- United States. This gap in management capa- ently beneficial practices that are widespread bilities likely contributes to the innovation elsewhere (Bloom et al. 2013). The work of gaps both (a) between the region and the Atkin et al. (2017) on barriers to technol- global frontier, and (b) between firms within ogy adoption provides a possible microfoun- the same countries. dation for such informational constraints. Technology Adoption and Diffusion: A Firm-Level Perspective 99

Their evidence suggests that misalignment Informational failures can also mani- of incentives between different types of fest in misperceptions about firms’ own workers within firms can impede the flow technological sophistication. Evidence of information on the benefits of a new from outside the region suggests that firms technology between the shop-floor workers significantly overestimate their own man- and managers. agement capabilities, and those with worse

FIGURE 3.11 Firms in developing East Asian countries score lower on management capabilities than firms at the global frontier

a. Average management scores in relation to per capita GDP 3.5 United States Japan Germany Sweden Canada France 3.0 Great Britain Australia Mexico Poland Italy Singapore New Zealand Malaysia Northern Ireland Turkey Chile China Spain Ireland Vietnam Brazil Kenya Greece 2.5 India Colombia Nigeria Nicaragua Portugal Myanmar Argentina Ethiopia Zambia Tanzania Ghana

Average management practices (score) 2.0 Mozambique 6 7 8 9 10 11 Log 10-year average GDP per capita, PPP (current international $, billions) Africa East Asia South Asia Latin America Australasia Europe North America

b. Distribution of management practices scores, selected countries

0.8

0.6

0.4 Density

0.2

0 1 2 3 4 5 Overall management score China Myanmar Vietnam United States

Source: World Management Survey (WMS), Centre for Economic Performance. Note: The WMS scores (collected over several years) range from 1 (worst practice) to 5 (best practice) across key management practices used by organizations in different sectors. These practices are grouped into five areas: Operations Management, Performance Monitoring, Target Setting, Leadership Management, and Talent Management. GDP = gross domestic product; PPP = purchasing power parity. 100 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

management capabilities are more likely to be an important factor in their technology be overconfident (Del Carpio and Taskin adoption decisions. Well over 50 percent 2016). Firms in developing East Asia are of surveyed firms, regardless of size, indi- no exception. The FAT survey in Vietnam cated that they are deterred by uncer- suggests that most firm managers are over- tainty of demand and high cost along with confident about their firms’ technological doubts about value of the new technolo- capabilities: they routinely assess their own gies (figure 3.13). More than half also cited companies’ relative technology levels to be lack of capabilities in terms of skills and higher than what’s revealed in the bench- available information as a critical constraint marking data (figure 3.12). to adoption. Uncertainty about returns from technology . The often-high fixed cost of innova- External drivers of innovation: trade, tion and technology adoption also means global value chains, and foreign direct that uncertainty about returns—for exam- investment ple, because of uncertainty about demand for new products or the efficiency of new Firms’ decisions to adopt technologies and technologies—will lead to low initial adop- innovate are not made in a vacuum. Their tion of new technologies (as explained in external environment shapes the returns and Atkin et al. 2017). Evidence of increased their incentives to innovate, depending on mar- investment in quality upgrading in response ket size, prices of and access to complementary to new export demand offers empirical sup- inputs, competitive pressure with both domes- port for this argument (Atkin, Khandelwal, tic and foreign firms, the flows of information and Osman 2017). and opportunities for learning-by-doing, and Data from the FAT survey in Vietnam the quality of regulations, including intel- confirms that firms consider uncertainty to lectual property protection. For extensive

FIGURE 3.12 Managers of Vietnamese firms express overconfidence about their firms’ technological capabilities

2 Self-assessment of technology

1 1 2 3 4 5 Intensive GBF Tech Index

Source: Cirera, Comin, Cruz, Lee, and Martins-Neto 2020. Note: The blue line shows the quadratic fit with a 95 percent confidence interval (dotted red lines). The black 45-degree line shows where expectations and index are equal. The technology index is regressed on the firms’ self-assessment of sophistication relative to other firms in the country. The technology index applied here covers the most-used technologies in general business functions (GBF), which include business processes, production planning, supply chain management, marketing, sales, payment methods, and quality control. The index was developed using data from the Firm-level Adoption of Technology (FAT) survey of Vietnamese firms in manufacturing, retail, and agriculture. For more details about the FAT data, see annex 3A. Technology Adoption and Diffusion: A Firm-Level Perspective 101

FIGURE 3.13 Lack of demand and uncertainty is the top self-reported barrier to technology adoption among Vietnamese firms of all sizes

Lack of demand and uncertainty

Lack of capabilities

Lack of nance

Government regulations

Other

Poor infrastructure

0 25 50 75 100 Share of rms (%) Small Medium Large

Source: Cirera, Comin, Cruz, and Lee 2020. Note: Firm sizes small (5–19 employees), medium (20–99), and large (100+). Data were gathered from the Firm-level Adoption of Technology (FAT) survey of Vietnamese firms in manufacturing, retail, and agriculture. For more details about the FAT data, see annex 3A.

discussions, see Syverson (2011), Verhoogen strategy, which in turn has facilitated a dif- (2020), and Zanello et al. (2016). fusion of innovation to the region that often The cost of capital relative to labor complements low labor cost. At the same (especially low-skilled labor) also plays a time, diffusion within countries remains slow, role in determining the direction of inno- consistent with the heterogeneity described vation toward more or less labor savings. earlier. This strategy will likely have decreas- Evidence from China suggests that minimum- ing returns without serious efforts to address wage increases lead firms to upgrade to firms’ capability constraints—efforts that are more labor-saving, capital-intensive tech- essential to reaping the spillover benefits from nologies (Hau, Huang, and Wang 2020). At trade and investment flows. the same time, in urban areas that receive a higher influx of low-skilled workers, firms Trade and global value chains can enable tend to use less human-capital-intensive tech- innovation, but benefits are concentrated nologies and produce fewer patents (Imbert in the most-productive firms et al. 2018). Theory and evidence suggest trade can be a It is not possible to empirically assess major channel for knowledge diffusion and the extent to which lack of innovation a driver of domestic innovation. Exports in developing East Asia is driven by rela- can create incentives to innovate through tive factor prices. (Chapter 4 looks at the increased market size—by spreading the supply of human capital and finance in fixed cost of innovation across markets—and more detail.) Nevertheless, for many coun- enable innovation through learning-by-doing tries in the region, low labor cost has sup- (Aghion et al. 2018; Aghion et al. 2019; Atkin ported an outward-oriented development et al. 2017). 102 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

Meanwhile, lowering import barriers can with both innovation outputs and R&D create competitive pressure, which induces investment. The positive association between firms to innovate to “escape competition” import penetration and innovation is appar- (Aghion et al. 2005) and gain market share ent for firms in mature industries but more (Raith 2003). Imports are also a source of muted for firms in high-tech industries. access to foreign knowledge, some of which Because mature industries arguably enjoy a is embedded in the goods traded (Acharya narrower technology gap with high-income and Keller 2009). Empirically, evidence of countries, the evidence suggests that laggard increasing domestic innovation and qual- firms further away from the technologi- ity upgrading following trade liberaliza- cal frontier are discouraged from innovat- tion episodes supports the positive impact ing when pre-innovation rents are low and of import competition on innovation increased competition primarily reduces (Amiti and Khandelwal 2013; Bustos 2011; post-innovation rents. This is consistent with Gorodnichenko, Svejnar, and Terrell 2010). evidence from both high-income and low- to Empirical evidence from East Asia has middle-income countries, which suggests that been somewhat mixed, however. Exporting firms must have adequate capabilities for has been found to be positively correlated these positive impacts from imports to materi- with R&D and innovation among manu- alize (Zanello et al. 2016). Evidence in China facturing firms in Indonesia and Malaysia is consistent with the international evidence: (Lee 2004; Yang and Chen 2012). Increased only the most-productive firms increased import penetration also positively affects their patenting rates and R&D expenditures innovation in China, making firms more after the country’s accession to the WTO likely to engage in incremental innovation (Bombardini, Li, and Wang 2017). (Lu and Ng 2012). Similarly, Chen, Zhang, and Zheng (2017) report a surge in R&D Global value chains can help in increasing intensity when Chinese firms import from innovation capabilities, but spillovers do high-income sources or when private high- not happen automatically tech firms use these imports. In contrast, Global value chains (GVCs) represent reductions in input tariffs after China’s acces- another important channel for the diffusion sion to the World Trade Organization (WTO) of innovation. The relational aspect of GVCs, have reduced firms’ patent applications where firms form longer-term relationships because high-quality imports have substituted and share an interest in specializing in specific for internal innovation (Liu and Qiu 2016). tasks, means that trading firms have even stronger incentives to exchange knowledge Import penetration correlates with (World Bank 2020). Indeed, survey evidence innovation, but host firms need some from Vietnam suggests that domestic firms minimum innovation capabilities with stronger links to foreign buyers (that is, New analysis for developing East Asia car- those with exclusive selling relationships) are ried out for this study indicates that import more likely to receive assistance from lead penetration is positively associated with inno- firms, especially through sharing knowledge vation. (For results, see annex 3B, tables 3B.2, and know-how. Case study research of IKEA’s 3B.3, and 3B.4.) For example, among suppliers in China and Southeast Asia shows Vietnamese firms engaged in international similarly that many suppliers take advantage trade, the reduction in tariffs brought about of this relationship to improve their “adap- by accession to the WTO is associated with tive” capabilities by learning about new prod- a significant increase in the number of R&D ucts and their process requirements (Ivarsson projects and research topics implemented and Alvstam 2011). (annex 3B, table 3B.4). The relational nature of GVCs does not For the region, on average, increased automatically result in technology transfer, import penetration is positively associated however, but depends critically on the Technology Adoption and Diffusion: A Firm-Level Perspective 103

governance of value chains and local absorp- co-invention is proportional to the innovative tive capacity. New capabilities may be activities of the country and its partners and is especially difficult to gain when lead firms in inversely related to the technological distance GVCs tightly control their technology (World between the partner countries. Bank 2020) or when technology gaps are too large (De Marchi, Giuliani, and Rabellotti FDI can also facilitate innovation, but 2018; Nocke and Yeaple 2008). spillover benefits are uneven Experience from the electronics industry in A large body of theoretical and empiri- Vietnam illustrates these barriers. Although cal evidence has documented how foreign Vietnam has been remarkably successful in direct investment (FDI) can facilitate technol- gaining global market shares for electronic ogy spillovers. (For a review, see Aghion and equipment, it has struggled to increase its Tirole 1994; Görg and Greenaway 2004; and domestic value-added share, because local sup- Keller 2010.) The literature identifies two pliers have been slow to climb the quality lad- types of spillovers: horizontal (or intraindus- der. Driving this outcome is a mix of inadequate try) and vertical (or interindustry). The latter domestic capabilities and difficulties in breaking can be further differentiated into backward into global supply chains with hyperspecialized (depending on upstream suppliers) or for- parts suppliers (World Bank 2017). ward (depending on downstream customers) spillovers. Trade flows and GVC activities increase the Horizontal spillovers may arise when probability of co-invention the multinational affiliate generates techno- Beyond technological spillovers, GVCs may logical learning spillovers to other firms in also promote innovation through co-inven- the industry through its business operation. tion (De Backer and Flaig 2017). By trading Similarly, (backward) vertical spillovers in goods and services, firms can build a rap- may arise when the multinational affili- port that may later induce collaboration on ate provides technology to its suppliers at R&D and ultimately co-invention. New anal- below-market prices. The evidence on these ysis from the OECD’s Trade in Value Added spillovers is mixed, however. There is firm- (TiVA) database reveals interesting, stylized level evidence of positive spillovers for some facts that support the interplay between high-income countries, as shown by Keller GVCs and global innovation networks and Yeaple (2009) for the United States; (GINs)—globally organized networks of Javorcik (2004) for Lithuania; or Haskel, firms and nonfirm organizations that develop Pereira, and Slaughter (2007) for the United and diffuse innovations. Kingdom. The evidence appears less likely to First, GINs have grown denser in develop- be positive in less-advanced economies, how- ing East Asia over the past 30 years, suggest- ever. The impact of FDI on technological ing an improvement in the region’s ability to change in Chinese domestic firms could also collaborate and innovate (map 3.2). Second, be negative (Fu and Gong 2011). This result although GINs are still not as developed as could depend on crowding-out effects in GVC networks, GVC hubs also tend to be product and human-capital markets as well key nodes in GINs. Econometric evidence as on the technology gap between foreign from bilateral trade and patent data further and domestic firms. suggests that co-invention increases in the Brahmbhatt and Hu (2010) provide an presence of a trading relationship (see annex overview of the diverse diffusion patterns of 3B, table 3B.5), in line with the evidence from FDI in East Asia. On one end of the spec- OECD countries documented in De Backer trum are economies like Singapore’s, whose and Flaig (2017). The results illustrate that development has largely been driven by FDI. the benefits of GVCs for co-invention also On the other end of the spectrum, Korea (and depend on the capacity of local firms rela- to a lesser extent, Taiwan, China) tended tive to their trading partners: the incidence of to restrict FDI and favored the use of other 104 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

MAP 3.2 Global innovation networks with developing East Asia have grown denser in recent decades

Source: World Bank elaboration using the REGPAT database 2018 of the Organisation for Economic Co-operation and Development (OECD). Note: The flows of global innovation networks correspond to the number of international co-inventions by country pairs during the selected period. The OECD REGPAT database presents patent data that have been linked to regions according to the addresses of applicants and inventors. REGPAT data were available for the following countries in developing East Asia: Cambodia, China, Indonesia, Malaysia, the Philippines, Thailand, and Vietnam. Plots created using flowmap.blue. Technology Adoption and Diffusion: A Firm-Level Perspective 105

instruments to spread innovation—such as the case of R&D investments, suggests that through the licensing of foreign technology the increased presence of foreign firms in the and upgrading of domestic technological downstream sectors provides an incentive capabilities (including through domestic R&D for domestic firms to innovate. Increased and strengthening of technical education and competition for foreign customers or the labor force skills). Between these extremes ability to exploit economies of scale may are Indonesia, Malaysia, the Philippines, and raise incentives for domestic suppliers to Thailand, which have also drawn substantial improve the quality of their products. FDI inflows since the 1980s. As in the case of import penetration, the The diverse country experience in East evidence supports the idea that local firms’ Asia suggests that the business environment ability to take advantage of FDI spillovers as well as firm-specific characteristics are depends on their absorptive capacity and likely to affect the formation and diffusion more precisely on their level of technology of FDI spillovers. An enabling business envi- efficiency. Firms that are closer to the techno- ronment is more likely to attract foreign logical frontier are more likely to benefit from investors. But to ensure that FDI supports the presence of foreign firms (see results in diffusion and technology adoption, adequate annex 3B, table 3B.7). absorptive capacity on the part of local firms is also needed. Conclusions To target spillovers, some countries have considered conditions for FDI entry—so- This chapter has argued that although devel- called forced technology transfer policies. oping East Asia has been home to some lead- China provides a noteworthy, if unique, ing innovative companies that might rival example of how governments use policies those in high-income countries, there remains such as mandated formation of international a large technological divide between and joint ventures (IJVs) (box 3.3) or quid pro within the region’s countries. By and large, quo requirements for foreign firms to trans- firms are far from the global frontier and fer technology domestically in return for are underprepared for rapid technological access to China’s domestic market. Recent changes such as the shift to digital and flex- evidence suggests that such policies have led ible manufacturing technologies. to significant gains in innovation for China The slow process of innovation diffusion but might impose significant costs on its and the presence of lagging firms is a styl- FDI partners (Girma et al. 2015; Holmes, ized fact that is not unique to the region. McGrattan, and Prescott 2015; Jiang et al. Nevertheless, it highlights the role of innova- 2018). Given China’s market size, there is tion policy to not only support firms pushing also a question about the replicability of the frontier but also those firms investing in this approach for most countries in the catching up. Given that the aggregate con- region, given fierce global competition for tribution made by innovation and new tech- FDI, especially in the current context of nologies to productivity, growth, and welfare potential reshoring of production. is largely determined by how well and how New evidence from the most recent World rapidly innovations diffuse within countries, Bank Enterprise Surveys for developing East policy needs to actively promote diffusion by Asia points to negative effects of horizontal nurturing and developing domestic innova- linkages on innovation, possibly because of tion capacity. competition effects, across all the dimen- Understanding the drivers of firms’ sions of innovation considered. (See further innovation decisions is key to inform pol- details in annex 3B, table 3B.7.) Evidence icy. A potential explanation for low diffu- of positive backward linkages, especially in sion is a lack of firm capabilities—a set of 106 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

BOX 3.3 International joint ventures and technology transfer in China

Inward foreign direct investment (FDI) may take the • Further, this positive effect is present through form of joint ventures—legal partnerships between both intrafirm (the joint venture) and interfirm a foreign investor and a domestic firm to create a (the domestic partner firm and other firms in new operation in the domestic market. China’s gov- the industry) channels. The IJV partner seems ernment mandates the formation of international to matter, however. Among different FDI source joint ventures (IJVs) for foreign investors in certain countries, IJVs with US firms appear to have the industries that seek to invest domestically. Despite its strongest spillover effects. policy relevance, robust empirical evidence about its • Moreover, the companies chosen to be IJV part- effectiveness has emerged only recently. ners are systematically different: they are larger, Girma et al. (2015) find that increased foreign more connected to government officials, and tend acquisitions following FDI liberalization episodes to be more innovative and productive. in China significantly raise the research and development (R&D) activities of domestic firms, driven These results might suggest that the same policy primarily by joint ventures. Jiang et al. (2018) pro- will not automatically yield the same impact else- vide the most comprehensive assessment of the role where, depending on local markets and the existing of IJVs in China to date. They rely on administra- stock of domestic firms. Even in China, new research tive data on all IJVs from 1998 to 2007, with direct by Chen and Lawell (2020) warns that more nuanced information on firm-to-firm linkages, and document results can be found in the automobile industry: several interesting findings: whether positive technology transfer for firms in IJVs • First, IJVs have a positive significant effect on firm materializes depends on the country of origin of the innovation (as measured by patents) and productivity. international partner.

“know-hows” that enable firms to respond to uncertainties may constitute major barriers to market conditions, identify relevant new tech- firms’ investments in technological upgrading. nologies, develop plans to exploit them, and Traditionally, East Asia’s openness to then acquire the necessary resources to do so. trade and foreign investments has helped These capabilities are often linked to manage- reduce these barriers by facilitating the flow ment quality; however, the data show substan- of knowledge and expanding firms’ access tial gaps between the region’s firms and the to both input and output markets. However, global frontier as well as tremendous hetero- the extent of spillover benefits has been geneity within countries. New survey data on somewhat limited and will continue to depend firm-level technologies also offer a glimpse into on the absorptive capacity of domestic firms. the complexities of the technology adoption With the restructuring of global GVCs, there decisions: multiple technologies exist for dif- is enormous uncertainty about the potential ferent business functions, and the same firms future gains from the region’s growth strat- can be technologically advanced or laggards in egy. Uncertainty has only intensified because different functions. Navigating these decisions of the COVID-19 pandemic. In this con- requires substantial knowledge and entails text, continued investment in building firm a great deal of uncertainty. Evidence sug- capabilities for innovation represents a critical gests that these informational constraints and no-regrets policy option. Technology Adoption and Diffusion: A Firm-Level Perspective 107

Annex 3A The Firm-level The FAT survey comprises five modules: Adoption of Technologies • Module A: Collects general informa- survey tion about the characteristics of the The goal of the World Bank’s Firm-level establishment Adoption of Technologies (FAT) survey is • Module B: Covers the technologies used in to collect comprehensive information about eight generic business functions the technologies used in each of the key busi- • Module C: Covers the use of technolo- ness functions for a representative sample of gies for functions that are specific to each companies. The survey is comprehensive in of 10 industry and services functions three different dimensions: • Module D: Includes questions about the drivers of and barriers to technology • It covers a large number of business adoption functions. • Module E: Collects information on • For each business function, the survey employment, balance sheet, and perfor- includes a wide range of technologies, mance, which enables computation of requiring different levels of sophistication labor productivity and other measures at to accomplish the tasks needed to fulfill the company level. the function. • It collects sufficient information to Modules B and C collect the informa- measure the various dimensions of the tion to measure technology adoption, while technology adoption process (for exam- the other modules collect information on ple, range of technologies, most-used firm characteristics, performance, and other technology, and adoption lag). variables that can provide information on the barriers to and drivers of technology The survey is stratified to ensure the adoption. representativeness of the sample by sector, In addition to Vietnam, the survey has firm size, and region. This feature enables been carried out in Bangladesh, the state of construction of representative statistics about Ceará in Brazil, Malawi, and Senegal, and technology adoption and use along those it is currently being implemented in India, three dimensions as well as at the country- Kenya, Korea, and Poland. and firm-levels. 108 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

Annex 3B Supplementary tables

TABLE 3B.1 Management scores of selected East Asian countries relative to the United States, by quantile

Differences relative to the United States at Kolmogorov- Country Smirnov Test Mean q5 q10 q50 q90 q95 China 0.426 –0.577*** –0.472*** –0.444*** –0.583*** –0.694*** –0.651*** (0) (0.0272) (0.0755) (0.0594) (0.0375) (0.0544) (0.0589) Vietnam 0.418 –0.616*** –0.500*** –0.611*** –0.611*** –0.750*** –0.833*** (0) (0.0519) (0.105) (0.117) (0.0578) (0.0563) (0.0732) Myanmar 0.601 –1.175*** –1.111*** –1.167*** –1.333*** –1.083*** –0.722** (0) (0.0918) (0.103) (0.0973) (0.128) (0.234) (0.289)

Source: Worldwide Management Survey (WMS), Centre for Economic Performance of the London School of Economics and Political Science. Note: WMS sample of firms with at least 100 workers. Results from the Kolmogorov-Smirnov distribution test, ordinary least squares (OLS), and quantile (q) regressions (at q5, q10, q50, q90, q95) of the management score on country dummies, with United States as the baseline. *** 1%; ** 5%, and * 10% confidence level.

TABLE 3B.2 Import penetration is positively associated with innovation in developing East Asia

New or New or improved Innovation Innovation Innovation improved product/ R&D Use of foreign a b c Independent score (1) score (2) score (3) process service expenses technology variables OLS LPM 0.475*** 0.496*** 0.461*** 0.284 0.793*** 0.903*** 0.0972 Import penetration (3.53) (3.69) (3.29) (1.28) (3.65) (3.65) (0.41) Capital stock Yes Yes Yes Yes Yes Yes Yes Firm characteristics Yes Yes Yes Yes Yes Yes Yes City FE Yes Yes Yes Yes Yes Yes Yes Country FE Yes Yes Yes Yes Yes Yes Yes Survey FE Yes Yes Yes Yes Yes Yes Yes Observations 5,620 5,628 5,348 4,031 4,098 4,078 5,516 R2 0.249 0.251 0.255 0.152 0.137 0.145 0.121

Source: World Bank Enterprise Survey data for East Asian countries, latest round. Note: “Developing East Asia” refers to the 10 middle-income countries covered in this study: Cambodia, China, Indonesia, Lao PDR, Malaysia, Mongolia, Myanmar, the Philippines, Thailand, and Vietnam. Firms are asked four questions. One question asks whether a firm has introduced a product innovation; a second question asks if the firm has introduced a process innovation; a third question asks if the firm has introduced an organizational innovation; and a fourth question asks if the firm has introduced a marketing innovation. The firm may answer “yes” or “no” to each of these questions individually, and the score(s) depend on how many of the firm responses are a “yes.” Results from instrumental variable (IV) regression where import penetration is instrumented for with the average of import penetration among firms belonging to the same industry but located in other cities within the same country. Standard errors are robust and clustered at the sample stratification level. The sample is stratified by industry, size of firm, and region. t statistics are reported in parentheses. FE = fixed effects; LPM = linear probability model; OLS = ordinary least squares; R&D = research and development. a. For innovation score (1), the numerator is given by the number of “yes” responses to the four questions listed above and the denominator is four. b. For innovation score (2), where the numerator is given by the number of “yes” responses to the four questions listed above and the denominator is the number of questions answered by the firm. c. For innovation score (3), the numerator is given by the number of “yes” responses to the four questions listed above, and the denominator is the number of nonmissing answers from each firm. * p < 0.10 ** p < 0.05 *** p < 0.01 Technology Adoption and Diffusion: A Firm-Level Perspective 109

TABLE 3B.3 The discouraging effect for laggard industries in developing East Asia

a. Mature industries New or New or improved Innovation Innovation Innovation improved product/ Use of foreign a b c Independent score (1) score (2) score (3) process service R&D expenses technology variables OLS LPM 0.247*** 0.247*** 0.250*** 0.358*** 0.361*** 0.332*** 0.00154 Import penetration (3.55) (3.53) (3.54) (4.03) (3.98) (3.71) (0.01) Firm characteristics Yes Yes Yes Yes Yes Yes Yes Country FE Yes Yes Yes Yes Yes Yes Yes Survey FE Yes Yes Yes Yes Yes Yes Yes Observations 4,124 4,140 3,961 2,902 2,945 2,933 4,062 R2 0.259 0.260 0.259 0.136 0.110 0.167 0.137 b. High-tech industries New or New or improved Independent Innovation Innovation Innovation improved product/ Use of foreign variables score (1)a score (2)b score (3)c process service R&D expenses technology OLS LPM 0.484 0.484 0.482 1.558** 1.426 1.897** 0.119 Import penetration (1.50) (1.50) (1.50) (2.02) (1.57) (2.38) (0.59) Firm characteristics Yes Yes Yes Yes Yes Yes Yes

Country FE Yes Yes Yes Yes Yes Yes Yes Survey FE Yes Yes Yes Yes Yes Yes Yes Observations 1,364 1,366 1,305 1,033 1,053 1,043 1,341 R2 0.232 0.232 0.232 0.202 0.209 0.140 0.125

Source: World Bank Enterprise Survey data for East Asian countries, latest round. Note: “Developing East Asia” refers to the 10 middle-income countries covered in this study: Cambodia, China, Indonesia, Lao PDR, Malaysia, Mongolia, Myanmar, the Philippines, Thailand, and Vietnam. Firms are asked four questions. One question asks whether a firm has introduced a product innovation; a second question asks if the firm has introduced a process innovation; a third question asks if the firm has introduced an organizational innovation; and a fourth question asks if the firm has introduced a marketing innovation. The firm may answer “yes” or “no” to each of these questions individually, and the score(s) depend on how many of the firm responses are a “yes.” Standard errors are robust and clustered at the sample stratification level. The sample is stratified by industry, size of firm, and region. t statistics are reported in parentheses. FE = fixed effects; LPM = linear probability model; OLS = ordinary least squares; R&D = research and development. a. For innovation score (1), the numerator is given by the number of “yes” responses to the four questions listed above and the denominator is four. b. For innovation score (2), where the numerator is given by the number of “yes” responses to the four questions listed above and the denominator is the number of questions answered by the firm. c. For innovation score (3), the numerator is given by the number of “yes” responses to the four questions listed above and the denominator is the number of nonmissing answers from each firm. * p < 0.10 ** p < 0.05 *** p < 0.01

TABLE 3B.4 Effects of WTO accession: Innovation and import-export activity in Vietnam

R&D projects R&D projects R&D expense implemented R&D expense implemented Independent variables Revenue per workers TFPR (Levinsohn Petrin) 3391261.6 86.29 3396545.1 87.38* Imports-exports × after 2007 (1.06) (1.66) (1.06) (1.68) –397631.3 –265.6* –395125.5 –266.6* Imports-exports (–1.08) (–1.87) (–1.07) (–1.89) –388899934.4 –44351.4 –440548810.5 –40509.8 After 2007 (–0.30) (–1.13) (–0.35) (–1.03) table continues next page 110 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

TABLE 3B.4 Effects of WTO accession: innovation and import-export activity in Vietnam (continued)

R&D projects R&D projects R&D expense implemented R&D expense implemented Independent variables Revenue per workers TFPR (Levinsohn Petrin) 540.9 –1.104 177039888.2 –46610.0 Productivity (0.08) (–0.81) (0.63) (–0.96) 1375.0 0.0669 1346.8 0.0709 Total investment (0.97) (0.92) (0.97) (0.99) 462395.0 275.2* 448843.9 276.7* Number of workers (1.14) (1.85) (1.11) (1.86) ISIC FE Yes Yes Yes Yes Year FE Yes Yes Yes Yes Observations 20,738 11,454 26,183 15,171 R2 0.203 0.536 0.202 0.530

Source: World Bank Enterprise Survey, Vietnam, 2004–16. Note: Ordinary least squares (OLS) model. Standard errors are robust and clustered at the sample stratification level. The sample is stratified by industry, size of firm, and region. t statistics are reported in parentheses. The sample excludes firms reporting negative values for (a) total assets, at either the beginning or the end of the year; (b) fixed assets, at either the beginning or the end of the year; (c) revenues; (d) total payroll payable; (e) total investment; (f) cost of goods and services; or (g) number of employees. Firms reporting zero or missing values for the total assets are also dropped. ISIC = International Standard Industrial Classification; R&D = research and development; TFPR = revenue total factor productivity; WTO = World Trade Organization. * p < 0.10 ** p < 0.05 *** p < 0.01

TABLE 3B.5 Co-inventions gravity model for developing East Asia countries

Independent variables OLS PPML 0.13*** 0.12*** 0.13*** 0.12*** 0.597*** 0.779*** 0.600*** 0.778*** Log(Patent stock country) (0.01) (0.00) (0.01) (0.00) (0.039) (0.040) (0.039) (0.040) 0.12*** 0.11*** 0.12*** 0.11*** 0.562*** 0.674*** 0.569*** 0.676*** Log(Patent stock partner) (0.00) (0.00) (0.00) (0.00) (0.045) (0.057) (0.043) (0.056) –0.13*** –0.13*** –0.13*** –0.13*** –0.090*** –0.078*** –0.098*** –0.088*** Log(Technological distance) (0.01) (0.01) (0.01) (0.01) (0.020) (0.020) (0.019) (0.019) 0.22*** 0.22*** 0.22*** 0.22*** –0.004*** –0.004*** –0.004*** –0.005*** Bilateral trade (0.01) (0.01) (0.01) (0.01) (0.001) (0.001) (0.001) (0.001) –0.09*** –0.12*** –0.09*** –0.11*** –6.074*** –5.984*** –6.353*** –6.273*** No bilateral trade (0.01) (0.01) (0.01) (0.01) (0.311) (0.308) (0.309) (0.308) 0.42*** 0.41*** 0.42*** 0.41*** 0.301*** 0.305*** 0.353*** 0.357*** Contiguity? (0.07) (0.07) (0.07) (0.07) (0.111) (0.112) (0.103) (0.104) 0.15*** 0.15*** 0.15*** 0.15*** 1.084*** 1.076*** 1.088*** 1.080*** Common language? (0.03) (0.03) (0.03) (0.03) (0.085) (0.081) (0.082) (0.078) Pair ever in colonial –0.12*** –0.12*** –0.12*** –0.12*** –0.914*** –0.918*** –0.953*** –0.957*** relationship? (0.04) (0.04) (0.04) (0.04) (0.098) (0.096) (0.096) (0.095) –0.00*** –0.00*** –0.00*** –0.00*** –0.000*** –0.000*** –0.000*** –0.000*** Physical distance (0.00) (0.00) (0.00) (0.00) (0.000) (0.000) (0.000) (0.000) 0.04*** 0.04*** 0.04*** 0.04*** 0.008 0.007 –0.041** –0.042*** Time difference (0.00) (0.00) (0.00) (0.00) (0.015) (0.014) (0.016) (0.016) 0.06*** 0.07*** 0.07*** 0.07*** 0.405*** 0.398*** 0.427*** 0.418*** No time difference (0.01) (0.01) (0.01) (0.01) (0.110) (0.107) (0.105) (0.103) 0.01* 0.01* –0.266*** –0.264*** WTO deep agreement (0.01) (0.01) (0.036) (0.035) Year FE Yes Yes Yes Yes Yes Yes Yes Yes Origin FE Yes Yes Yes Yes Yes Yes Yes Yes table continues next page Technology Adoption and Diffusion: A Firm-Level Perspective 111

TABLE 3B.5 Co-inventions gravity model for developing East Asia countries (continued)

Independent variables OLS PPML Destination FE Yes Yes Yes Yes Yes Yes Yes Yes Sector FE No Yes No Yes No Yes No Yes Mean dep. var. 0.13 0.13 0.13 0.13 0.16 0.16 0.16 0.16 R2 0.05 0.05 0.05 0.05 0.47 0.49 0.49 0.51 Adjusted-within-R2 0.03 0.03 0.03 0.03 Observations 765,952 765,952 765,952 765,952 650,947 650,947 650,947 650,947

Sources: Trade in Value Added (TiVA) and REGPAT databases, 2018, Organisation for Economic Co-operation and Development. Note: Sample includes 63 economies. All pairs include one of 10 middle-income “developing East Asia” countries: Cambodia, China, Indonesia, Lao PDR, Malaysia, Mongolia, Myanmar, the Philippines, Thailand, and Vietnam. Robust standard errors in parentheses. The dependent variable is the number of international co-inventions for the origin- destination pair in a given sector and year. Distance is the absolute value of the difference between the stock of patents in countries. FE = fixed effects; OLS = ordinary least squares; PPML = Poisson pseudo maximum likelihood; R&D = research and development; WTO = World Trade Organization. * p < 0.10 ** p < 0.05 *** p < 0.01

TABLE 3B.6 FDI spillover effects on innovation in developing East Asia

New or improved New or improved product R&D Use of foreign Independent variables process or service expenses technology 0.500 –0.600 –0.328 –0.00904 Forward linkage (0.88) (–1.39) (–0.74) (–0.03) 0.233 1.056* 2.188*** 0.501 Backward linkage (0.36) (1.96) (3.08) (1.32) –0.325*** –0.441*** –0.310*** –0.213** Horizontal linkage (–2.75) (–5.90) (–2.67) (–2.32) 0.000971* 0.00318*** 0.00112* 0.0000914 % Material inputs imported (1.71) (5.42) (1.88) (0.22) 0.00222*** 0.00351*** 0.00293*** 0.000877 % Sales exported indirectly (2.79) (3.62) (3.56) (1.10) 0.00111* 0.000831 0.000748 0.00170*** % Sales exported directly (1.79) (1.32) (1.19) (3.51) 0.182*** 0.0737** 0.0590** 0.0721*** Medium (20–99) (5.61) (2.42) (2.26) (3.29) 0.253*** 0.166*** 0.187*** 0.203*** Large (≥ 100) (7.29) (5.02) (5.66) (7.05) –0.00656*** –0.00204 –0.000544 0.000702 Age (–4.04) (–1.31) (–0.33) (0.55) 0.0537 –0.0526 –0.0209 –0.177*** 100% domestic ownership (1.23) (–1.14) (–0.45) (–4.23) –0.00183 –0.0663** 0.0339 0.00396 No female owners (–0.07) (–2.32) (1.01) (0.19) Access to finance: minor 0.138*** 0.180*** 0.00530 0.0459** obstacle (4.53) (5.83) (0.20) (2.14) Access to finance: moderate 0.117*** 0.155*** 0.137*** 0.0282 obstacle (2.76) (3.97) (3.65) (0.93) Access to finance: major 0.0826 0.0353 0.173*** 0.0219 obstacle (1.36) (0.67) (3.21) (0.46) Country FE Yes Yes Yes Yes Survey FE Yes Yes Yes Yes Observations 5,620 5,704 5,675 7,808 R2 0.149 0.149 0.132 0.106

Source: World Bank Enterprise Survey data for East Asian countries, latest round. Note: Linear probability model. Standard errors are robust and clustered at the sample stratification level. The sample is stratified by industry, size of firm, and region. t statistics are reported in parentheses. “Developing East Asia” refers to the 10 middle-income countries covered in this study: Cambodia, China, Indonesia, Lao PDR, Malaysia, Mongolia, Myanmar, the Philippines, Thailand, and Vietnam. FDI = foreign direct investment; FE = fixed effects; R&D = research and development. * p < 0.10 ** p < 0.05 *** p < 0.01 112 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

TABLE 3B.7 Effect of firms’ distance from the frontier on the strength of spillovers in developing East Asia

Independent Innovation Innovation Innovation Improved New Foreign variables score (1)a score (2)b score (3)c process product/service R&D technology 0.791 0.849 0.562 0.957 1.149 –0.0284 1.028 Forward linkage (1.24) (1.31) (0.92) (0.88) (1.47) (–0.04) (1.35) –0.750 –0.710 –0.493 –1.211 –2.249** –1.195 1.703 Backward linkage (–1.06) (–1.00) (–0.71) (–0.85) (–2.30) (–1.05) (1.51) –0.451*** –0.454*** –0.504*** –0.690*** –0.328* –0.569*** –0.238 Horizontal linkage (–4.40) (–4.38) (–5.04) (–3.18) (–1.88) (–3.85) (–1.21) Forward linkage × –0.856* –0.909** –0.729 –0.569 –1.292** –0.990* –0.586 distance (–1.89) (–1.99) (–1.63) (–0.80) (–2.24) (–1.84) (–1.07) Backward linkage × 0.568 0.588 0.421 0.324 1.421** 0.574 –0.0247 distance (1.18) (1.22) (0.88) (0.39) (2.10) (0.79) (–0.04) Horizontal linkage × –0.158** –0.158** –0.132* –0.123 –0.237** –0.207** –0.0520 distance (–2.24) (–2.23) (–1.87) (–0.98) (–2.11) (–2.06) (–0.44) Distance 0.0438* 0.0461* 0.0399 0.0204 0.0558 0.0482 0.0488 (1.69) (1.77) (1.49) (0.51) (1.40) (1.18) (1.20) Firm characteristics Yes Yes Yes Yes Yes Yes Yes N 2,904 2,910 2,775 2,850 2,880 2,873 2,859 R2 0.208 0.204 0.207 0.171 0.145 0.139 0.099

Source: World Bank Enterprise Survey data for East Asian countries, latest round. Note: “Developing East Asia” refers to the 10 middle-income countries covered in this study: Cambodia, China, Indonesia, Lao PDR, Malaysia, Mongolia, Myanmar, the Philippines, Thailand, and Vietnam. Standard errors are robust and clustered at the sample stratification level. The sample is stratified by industry, size of firm, and region. t statistics are reported in parentheses. Firm controls include percentage of material inputs imported, percentage of sales exported indirectly, percentage of sales exported directly, size of the firm, age of the firm, dummy for 100 percent domestic ownership, dummy for no female owners, and a categorical variable for access to finance. Distance with the frontier for local firms is a Mahalanobis distance between local and foreign firms’ productivity. R&D = research and development. a. For innovation score (1), the numerator is given by the number of “yes” responses to the four questions listed above and the denominator is four. b. For innovation score (2), where the numerator is given by the number of “yes” responses to the four questions listed above and the denominator is the number of questions answered by the firm. c. For innovation score (3), the numerator is given by the number of “yes” responses to the four questions, one each for (whether they have introduced (1) a product innovation, (2) a process innovation, (3) an organizational innovation, and/or (4) a marketing innovation. The denominator is the number of nonmissing answers from each firm. * p < 0.10 ** p < 0.05 *** p < 0.01

Notes firms in 35 countries. The WMS captures several aspects of management, including 1. A firm can be digital in several tasks. For firms’ practices in target setting, monitoring, example, it can process sales online but and human resource management. It is not also manage the supply chain digitally. The specific to innovation, but is a proxy for business functions included in the FAT survey firms’ overall capabilities. capture the extent of digitalization not only 4. These differences are statistically significant in a firm’s ability to process sales online but and hold true for different quantiles of the also to perform other management tasks in overall management score distribution. an integrated system, including remotely. See additional results in annex 3B, table 3B.1. The “digital readiness index” focuses on the intensive margin. Thus, for example, a References value of 100 in sales corresponds to a firm processing all sales on online platforms or on Acharya, Ram C., and Wolfgang Keller. the firm’s website. 2009. “Technology Transfer through 2. A comparable index in the state of Ceará in Imports.” Canadian Journal of Economics Brazil is 42 percent (Cirera, Comin, Cruz, and 42 (4): 1411–48. Lee 2020). Aghion, Philippe, Antonin Bergeaud, Timothee 3. The WMS (https://worldmanagementsurvey. Gigout, Matthieu Lequien, and Marc Melitz. org/), operated by the Centre for Economic 2019. “Spreading Knowledge across the Performance of the London School of World: Innovation Spillover through Trade Economics and Political Science, is conducted Expansion.” Manuscript, Harvard University, through in-depth interviews of over 20,000 Cambridge, MA. Technology Adoption and Diffusion: A Firm-Level Perspective 113

Aghion, Philippe, Antonin Bergeaud, Matthieu by Lionel Fontagné and Ann Harrison, 136–68. Lequien, and Marc J. Melitz. 2018. “The Oxford: Oxford University Press. Heterogeneous Impact of Market Size on Bloom, Nicholas, Benn Eifert, Aprajit Mahajan, Innovation: Evidence from French Firm-Level David McKenzie, and John Roberts. 2013. Exports.” NBER Working Paper No. 24600, “Does Management Matter? Evidence from National Bureau of Economic Research, India.” Quarterly Journal of Economics 128 Cambridge, MA. (1): 1–51. Aghion, Philippe, Nick Bloom, Richard Bombardini, Matilde, Bingjing Li, and Ruoying Blundell, Rachel Griffith, and Peter Wang. 2017. “Import Competition and Howitt. 2005. “Competition and Innovation: Innovation: Evidence from China.” Unpublished. An Inverted-U Relationship.” Quarterly Journal Brahmbhatt, Milan, and Albert Hu. 2010. “Ideas of Economics 120 (2): 701–28. and Innovation in East Asia.” The World Bank Aghion, Philippe, and Jean Tirole. 1994. “The Research Observer 25 (2): 177–207. Management of Innovation.” Quarterly Bresnahan, Timothy F., and Manuel Trajtenberg. Journal of Economics 109 (4): 1185–209. 1995. “General Purpose Technologies ‘Engines Amiti, Mary, and Amit K. Khandelwal. of Growth’?” Journal of Econometrics 65 (1): 2013. “Import Competition and Quality 83–108. Upgrading.” Review of Economics and Brynjolfsson, Erik, Lorin M. Hitt, and Shinkyu Statistics 95 (2): 476–90. Yang. 2002. “Intangible Assets: Computers and Andrews, Dan, Chiara Criscuolo, and Peter N. Organizational Capital.” Brookings Papers on Gal. 2016. “The Best versus the Rest: The Economic Activity 2002 (1): 137–81. Global Productivity Slowdown, Divergence Bustos, Paula. 2011. “Trade Liberalization, across Firms and the Role of Public Policy.” Exports, and Technology Upgrading: Evidence Productivity Working Papers No. 5, on the Impact of MERCOSUR on Argentinian Organisation for Economic Co-operation and Firms.” American Economic Review 101 (1): Development, Paris. 304–40. Apedo-Amah, Marie Christine, Besart Avdiu, Chen, Yuan, and C.-Y. Cynthia Lin Lawell. Xavier Cirera, Marcio Cruz, Elwin Davies, 2020. “Supply and Demand in the Chinese Arti Grover, Leonardo Iacavone, et al. Automobile Market: A Random Coefficients 2020. “Unmasking the Impact of COVID- Mixed Oligopolistic Differentiated Products 19 on Businesses: Firm Level Evidence Model.” Working paper, Cornell University, from Across the World.” Policy Research Ithaca, NY. Working Paper 9434, World Bank, Chen, Zhiyuan, Jie Zhang, and Wenping Zheng. Washington, DC. 2017. “Import and Innovation: Evidence Atkin, David, Azam Chaudhry, Shamyla Chaudry, from Chinese Firms.” European Economic Amit K. Khandelwal, and Eric Verhoogen. Review 94: 205–20. 2017. “Organizational Barriers to Technology Cirera, Xavier, Diego A. Comin, Marcio Cruz, and Adoption: Evidence from Soccer-Ball Kyung Min Lee. 2020. “Technology Within Producers in Pakistan.” Quarterly Journal of and Across Firms.” NBER Working Paper Economics 132 (3): 1101–64. No. 28080, National Bureau of Economic Atkin, David, Amit K. Khandelwal, and Research, Cambridge, MA. Adam Osman. 2017. “Exporting and Firm Cirera, Xavier, Diego A. Comin, Marcio Cruz, Performance: Evidence from a Randomized Kyung Min Lee, and Antonio Soares Martins- Experiment.” Quarterly Journal of Neto. 2020. “Firm-Level Technology Adoption Economics 132 (2): 551–615. in Vietnam.” Unpublished manuscript, World Audretsch, David B., Marian Hafenstein, Bank, Washington, DC. Alexander S. Kritikos, and Alexander Schiersch Comin, Diego, and Bart Hobijn, 2004. “Cross- 2018. “Firm Size and Innovation in the Service Country Technology Adoption: Making the Sector.” Discussion Paper No. 12035, Institute Theories Face the Facts.” Journal of Monetary of Labor Economics (IZA), Bonn. Economics, 51 (1): 39–83. Bernard, Andrew B., and Teresa C. Fort. 2017. Comin, Diego, and Bart Hobijn. 2010. “An “Factoryless Goods Producers in the USA.” Exploration of Technology Diffusion.” In The Factory-Free Economy: Outsourcing, American Economic Review 100 (5): Servitization, and the Future of Industry, edited 2031–59. 114 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

Comin, Diego, and Bart Hobijn. 2011. “Technology Innovation Policies in Emerging Economies.” Diffusion and Postwar Growth.” NBER World Bank, Washington, DC. Macroeconomics Annual 25 (1): 209–46. Fu, Xiaolan, and Yundan Gong. 2011. “Indigenous Comin, Diego, and Martí Mestieri. 2018. “If and Foreign Innovation Efforts and Drivers Technology Has Arrived Everywhere, Why of Technological Upgrading: Evidence from Has Income Diverged?” American Economic China.” World Development 39 (7): 1213–25. Journal: Macroeconomics 10 (3): 137–78. Girma, Sourafel, Yundan Gong, Holger Görg, Crozet, Matthieu, and Emmanuel Milet. 2017. and Sandra Lancheros. 2015. “Estimating “Should Everybody Be in Services? The Direct and Indirect Effects of Foreign Direct Effect of Servitization on Manufacturing Investment on Firm Productivity in the Presence Firm Performance.” Journal of Economics & of Interactions between Firms.” Journal of Management Strategy 26 (4): 820–41. International Economics 95 (1): 157–69. De Backer, Koen, and Dorothee Flaig. Görg, Holger, and David Greenaway. 2004. 2017. “The Future of Global Value Chains: “Much Ado about Nothing? Do Domestic Business as Usual or ‘A New Normal’?” Firms Really Benefit from Foreign Direct Science, Technology and Innovation Policy Investment?” The World Bank Research Papers No. 41, OECD Publishing, Paris. Observer 19 (2): 171–-97. De Marchi, Valentina, Elisa Giuliani, and Roberta Gorodnichenko, Yuriy, Jan Svejnar, and Katherine Rabellotti. 2018. “Do Global Value Chains Terrell. 2010. “Globalization and Innovation Offer Developing Countries Learning and in Emerging Markets.” American Economic Innovation Opportunities?” European Journal Journal: Macroeconomics 2 (2): 194–226. of Development Research 30 (3): 389–407. Hall, Bronwyn H. 2004. “Innovation and de Nicola, Francesca. 2019. “Assessing How Diffusion.” NBER Working Paper No. 10212, Returns to Innovation Vary Depending National Bureau of Economic Research, on the Business Environment and Firms’ Cambridge, MA. Characteristics.” Background paper for this Hall, Bronwyn H. 2011. “Innovation and study, World Bank, Washington, DC. Productivity.” NBER Working Paper No. Del Carpio, Ximena, and Temel Taskin. 2016. 17178, National Bureau of Economic “Management Quality Matters: Measuring Research, Cambridge, MA. and Benchmarking the Quality of Firms’ Hall, Bronwyn H., and Beethika Khan. 2003. Management in Turkey.” Private Sector “Adoption of New Technology.” NBER Development Blog, November 14. https:// Working Paper No. 9730, National Bureau of blogs.worldbank.org/psd/management-quality Economic Research, Cambridge, MA. -matters-measuring-and-benchmarking Haskel, Jonathan E., Sonia C. Pereira, and -quality-firms-management-turkey. Matthew J. Slaughter. 2007. “Does Inward Doloreux, David, Richard Shearmur, and Foreign Direct Investment Boost the Mercedes Rodriguez. 2016. “Determinants of Productivity of Domestic Firms?” Review of R&D in Knowledge-Intensive Business Services Economics and Statistics 89 (3): 482–96. Firms.” Economics of Innovation and New Hau, Harald, Yi Huang, and Gewei Wang. 2020. Technology 25 (4): 391–405. “Firm Response to Competitive Shocks: Eberhardt, Markus, Christian Helmers, and Evidence from China’s Minimum Wage Zhihong Yu. 2011. “Is the Dragon Learning Policy.” Review of Economic Studies 87 (6): to Fly? An Analysis of the Chinese Patent 2639–71. Explosion.” Research Paper 11/16, Centre Hernández, Hector, Nicola Grassano, Alexander for Research on Globalisation and Economic Tübke, Sara Amoroso, Zoltan Csefalvay, and Policy, University of Nottingham, UK. Petros Gkotsis. 2020. The 2019 EU Industrial Ettlie, John E., and Stephen R. Rosenthal. R&D Investment Scoreboard. Luxembourg: 2011. “Service versus Manufacturing European Union. Innovation.” Journal of Product Innovation Hipp, Christiane, and Hariolf Grupp. Management 28 (2): 285–99. 2005. “Innovation in the Service Sector: Frias, Jaime, Jin Lee, and Kyeyoung Shin. The Demand for Service-Specific Innovation 2020. “Innovation Policy Learning from Measurement Concepts and Typologies.” Korea: Lessons for Design and Execution of Research Policy 34 (4): 517–35. Technology Adoption and Diffusion: A Firm-Level Perspective 115

Holmes, Thomas J., Ellen R. McGrattan, and MOEF (Ministry of Economy and Finance, Edward C. Prescott. 2015. “Quid Pro Quo: Republic of Korea). 2020. “Safeguarding Technology Capital Transfers for Market Economic Resilience: In Responding to the Access in China.” Review of Economic Economic Impact of COVID-19.” Report, Studies 82 (3): 1154–93. MOEF, Seoul. Iacovone, Leonardo, Aaditya Mattoo, and Mohnen, Pierre, and Bronwyn H. Hall. 2013. Andrés Zahler. 2013. “Trade and Innovation “Innovation and Productivity: An Update.” in Services: Evidence from a Developing Eurasian Business Review 3 (1): 47–65. Economy.” Policy Research Working Paper Musolesi, Antonio, and Jean-Pierre Huiban. 2010. 6520, World Bank, Washington, DC. “Innovation and Productivity in Knowledge Imbert, Clément, Marlon Seror, Yifan Zhang, Intensive Business Services.” Journal of and Yanos Zylberberg. 2018. “Migrants and Productivity Analysis 34 (1): 63–81. Firms: Evidence from China.” Working Paper Nocke, Volker, and Stephen Yeaple. 2008. No. 7440, CESifo, Munich. “An Assignment Theory of Foreign Direct Ivarsson, Inge, and Claes Göran Alvstam. 2011. Investment.” Review of Economic Studies 75 “Upgrading in Global Value-Chains: A (2) 529–57. Case Study of Technology-Learning among Oh, Juhwan, Jong-Koo Lee, Dan Schwarz, IKEA-Suppliers in China and Southeast Hannah L. Ratcliffe, Jeffrey F. Markuns, and Asia.” Journal of Economic Geography 11 (4): Lisa R. Hirschhorn. 2020. “National Response 731–52. to COVID-19 in the Republic of Korea and Javorcik, Beata Smarzynska. 2004. “Does Foreign Lessons Learned for Other Countries.” Health Direct Investment Increase the Productivity Systems & Reform 6 (1): e1753464. of Domestic Firms? In Search of Spillovers Park, Albert, and Wenshi Xuan. 2020. “Skills for through Backward Linkages.” American Innovation in China.” Background paper for Economic Review 94 (3): 605–27. this report, Hong Kong University of Science Jiang, Kun, Wolfgang Keller, Larry D. Qiu, and and Technology, Hong Kong SAR, China. William Ridley. 2018. “International Joint Park, Young Joon, Sang Yun Cho, Jin Lee, Ikjin Lee, Ventures and Internal vs. External Technology Won-Ho Park, Seungmyeong Jeong, Seongyun Transfer: Evidence from China.” NBER Kim, Seokjun Lee, Jaeho Kim, and Ok Park. Working Paper No. 24455, National Bureau of 2020. “Development and Utilization of a Rapid Economic Research, Cambridge, MA. and Accurate Epidemic Investigation Support Keller, Wolfgang. 2010. “International Trade, System for COVID-19.” Osong Public Health Foreign Direct Investment, and Technology and Research Perspectives 11 (3): 118–27. Spillovers.” In Handbook of the Economics of Pires, Cesaltina Pacheco, Soumodip Sarkar, and Innovation, vol. 2, edited by Bronwyn H. Hall Luísa Carvalho. 2008. “Innovation in Services– and Nathan Rosenberg, 793–829. Amsterdam: How Different from Manufacturing?” The Elsevier. Service Industries Journal 28 (10) : 1339–56. Keller, Wolfgang, and Stephen R. Yeaple. 2009. Prud’homme, Dan, and Taolue Zhang. “Multinational Enterprises, International 2019. China’s Intellectual Property Regime for Trade, and Productivity Growth: Firm Level Innovation: Risks to Business and National Evidence from the United States.” Review of Development. Cham, Switzerland: Springer. Economics and Statistics 91 (4): 821–31. Raith, Michael. 2003. “Competition, Risk, and Lee, Cassey. 2004. “The Determinants of Managerial Incentives.” American Economic Innovation in the Malaysian Manufacturing Review 93 (4): 1425–36. Sector: An Econometric Analysis at the Firm Republic of Korea. 2020. “Preventing the Inflow Level.” ASEAN Economic Bulletin 21 (3): and Spread of the Virus.” COVID-19 response 319–29. information, Coronavirus-19 website of the Liu, Qing, and Larry D. Qiu. 2016. “Intermediate Republic of Korea. http://ncov.mohw.go.kr/en Input Imports and Innovations: Evidence from /baroView.do?brdId=11&brdGubun=111. Chinese Firms’ Patent Filings.” Journal of Santacreu, Ana Maria. 2015. “Innovation, Diffusion, International Economics 103: 166–83. and Trade: Theory and Measurement.” Journal Lu, Yi, and Travis Ng. 2012. “Do Imports Spur of Monetary Economics 75: 1–20. Incremental Innovation in the South?” China Santamaría, Lluís, María Jesus Nieto, and Economic Review 23 (4): 819–32. Ian Miles. 2012. “Service Innovation in 116 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

Manufacturing Firms: Evidence from World Bank. 2017. “Vietnam: Enhancing Spain.” Technovation 32 (2): 144–55. Enterprise Competitiveness and SME Sutton, John. 2012. Competing in Capabilities: Linkages: Lessons from International and The Globalization Process. Oxford: Oxford National Experience.” Working paper, World University Press. Bank, Washington, DC. Syverson, Chad. 2011. “What Determines World Bank. 2020. World Development Report Productivity?” Journal of Economic 2020: Trading for Development in the Age Literature 49 (2): 326–65. of Global Value Chains. Washington, DC: Verhoogen, Eric. 2020. “Firm-Level Upgrading World Bank. in Developing Countries.” Working paper Yang, Chih-Hai, and Ying-Hui Chen. 2012. No. 83, Center for Development Economics “R&D, Productivity, and Exports: Plant- and Policy and Center on Global Economic Level Evidence from Indonesia.” Economic Governance, School of International Modelling 29 (2): 208–16. and Public Affairs, Columbia University, Zanello, Giacomo, Xiaolan Fu, Pierr Mohnen, New York. and Marc Ventresca. 2016. “The Creation Wei, Shang-Jin, Zhuan Xie, and Xiaobo and Diffusion of Innovation in Developing Zhang. 2017. “From ‘Made in China’ to Countries: A Systematic Literature ‘Innovated in China’: Necessity, Prospect, Review.” Journal of Economic Surveys 30 (5): and Challenges.” Journal of Economic 884–912. Perspectives 31 (1): 49–70. Skills and Finance for Innovation 4

Introduction Several studies find a significant positive relationship between workers’ human capital The ability of firms to innovate depends on and firms’ innovation in high-income econo- numerous factors that fall outside the realm mies.1 Analysis of World Bank Enterprise of innovation policy, strictly defined. As dis- Survey data from across developing East Asia cussed in previous chapters, these include also finds that employees’ human capital— policy and institutional factors that establish both education and firm-level training—is business and regulatory environments condu- positively correlated with the intensity of their cive to firms investing in innovation. Other research and development (R&D) investment key complementary factors include the avail- (ADB 2020; de Nicola 2019). ability of a sufficiently skilled workforce and Moreover, the effects of workers’ human adequate financing to support firms’—often capital on innovation appear to be sizable. risky—innovation activities. This chapter A recent report by the Asian Development focuses on the roles of, and challenges asso- Bank finds, for example, that a 1 percent ciated with, the adequate supply of skills and increase in worker education, as proxied by finance in enabling innovation. secondary school enrollment, is associated with a 2 percent increase in innovation as Skills for innovation measured by patent flows (ADB 2020). While the positive effect of education Strong workforce skills are critical to and skills on innovation appears to cut enabling innovation among firms across categories of workers, one strand A significant literature highlights the impor- of the literature focuses on the impact of tance of workers’ human capital and skills in managers’ human capital. Maloney and enabling innovation. This is not surprising, Sarrias (2017) find that managers’ educa- as greater education and skill levels provide tion is a key driver of differences in mana- workers with the cognitive abilities needed gerial quality—an important determinant to absorb knowledge, develop and interact of innovation—across countries (Cirera, with new technologies, and create new prod- Maloney, and Sarrias 2020).2 Having highly ucts and processes at the frontier (Acemoglu, educated managers is also directly associated Aghion, and Zilibotti 2006; Toner 2011). with greater firm-level innovation (Ayyagari, 118 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

Demirgüç-Kunt, and Maksimovic 2011). These include technical skills, higher-order Interestingly, a recent study of Japanese start- cognitive skills, noncognitive skills, and ups finds that a firm founder’s level of human managerial and entrepreneurial skills, capital is, likewise, positively associated as summarized in table 4.1 and further with the firm’s innovation outcomes (Kato, described below. With rapid technologi- Okamuro, and Honjo 2015). cal change, there is a growing demand for the higher-order cognitive skills required to carry out nonroutine tasks (Mason and A range of workforce skill types are Shetty 2019; World Bank 2016). In addi- important to fostering innovation tion, across the literature, there is a con- Which types of human capital—and skills, sensus that noncognitive skills, also called in particular—are important for innovation? socioemotional skills, are of increasing Until recently, analysis of skills for innovation importance to enabling innovation.3 has focused predominantly on the role and Technical skills . Several types of techni- importance of science, technology, engineer- cal skills can be important in fostering inno- ing, and mathematics (STEM) skills (see, for vation. At a basic level, digital literacy is example, World Bank 2012). STEM skills are increasingly critical to innovation-related still considered important. Over time, how- tasks. Along with the knowledge and ability ever, there has been a growing understanding to use basic digital business tools and applica- that a broader set of skills are needed to fos- tions, there is a growing need for special ized ter innovation, including not only technical skills to develop, operate, and maintain knowledge but also advanced problem solv- information and communication technology ing; creative decision making; and the abilities (ICT) systems. Demand for these digital and to manage complexity, to communicate effec- technical capabilities is high in developing tively, and to work well in teams. East Asia, yet workers with such skills remain Recent studies of the skills needed to in short supply; indeed, employers across the support innovation and to enable successful region report difficulty in finding employees adoption of new technologies thus recog- with the required technical skills (Mason, nize the importance of a wider set of skills Kehayova, and Yang 2018). and disciplines beyond those related to Higher-order cognitive skills . Skills such STEM (Deming and Khan 2018; Green, as problem solving, critical thinking, creative Jones, and Miles 2007; OECD 2011, 2016). thinking, and logical thinking are becoming

TABLE 4.1 A range of workforce skills are important for enabling innovation

Skill type Description Technical These are specific skills needed in an occupation, are often discipline-specific, and may include knowledge of certain tools or processes such as common software (for example, Microsoft Excel) and specific software (for example, Java).

Higher-order cognitive These include skills such as problem solving, critical thinking, learning to learn, and the ability to manage complexity. These more-advanced cognitive skills must be built on a strong foundation of basic cognitive skills (that is, literacy and numeracy).

Noncognitive (socioemotional) These noncognitive skills include the ability to work and interact effectively in teams, communicate, motivate oneself, take initiative, and be able to read and manage one’s own emotions and reactions to others’ behaviors. Managerial and entrepreneurial These competencies are needed to implement innovative ideas and enable organizations to adapt and respond in competitive environments. They also include leadership skills like team building, negotiation, motivation, and coordination.

Source: Adapted from Kataoka and Alejo 2019. Note: STEM = science, technology, engineering, and mathematics. Skills and Finance for Innovation 119

increasingly valuable relative to basic cogni- organizational citizenship behavior, and tive skills like literacy and numeracy. The negatively correlated with counterproduc- higher-order cognitive skills needed for tive work behavior (OECD 2017). With jobs innovation also include the ability to learn becoming increasingly intensive in nonrou- and adapt. Technology, for instance, is only tine interpersonal tasks, the demand by firms useful to the extent that people can use it. for strong socioemotional skills, includ- Individuals also need the ability to accumu- ing communication and leadership, is also late knowledge and skills to operate new pro- expected to increase (Bughin et al. 2018).4 cesses, create new knowledge, and improve Managerial and entrepreneurial skills . adopted innovations. Such “trainable” indi- The importance of managerial skills to inno- viduals who can learn are valuable to innovation is well emphasized in the literature. vative firms. Higher-order cognitive skills are Even after controlling for inputs such as often not included in traditional school cur- R&D, for instance, managerial and organi- ricula, however. zational practices are important predictors of Socioemotional skills . Numerous personal innovation across countries, firm sizes, and attributes are essential to one’s performance income levels (Cirera and Maloney 2017). and effectiveness in the workplace and in Managerial skills are an essential input of social situations. These attributes encom- management quality. Moreover, managerial pass a wide range of nonroutine social and talent—the ability to build teams, commu- behavioral skills (table 4.2). Socioemotional nicate, motivate people, identify talent, and skills, particularly “conscientiousness,” are strategize—influences innovation because it positively correlated with job performance enables firms to identify productive opportu- dimensions such as task orientation and nities, evaluate their feasibility, and allocate human resources effectively. Entrepreneurship TABLE 4.2 Socioemotional skills reflect skills are similarly valuable to innovation, personal attributes that are essential to workplace particularly in the ability to manage risks. performance

Domain Skills or characteristics Evidence on skills for innovation Conscientiousness (task · Achievement focus Evidence from high-income economies performance) · Self-discipline suggests that advanced cognitive, Responsibility · socioemotional, and technical skills are all · Perseverance important to fostering innovation Extraversion (interpersonal · Sociability To date, most evidence on the skills that engagement) · Assertiveness enable innovation comes from high-income · Energy countries and reinforces the idea that a range · Enthusiasm of advanced skills are important to foster- Agreeableness (collaboration) · Empathy ing innovation. A survey of employees in 20 · Trust Organisation for Economic Co-operation · Cooperation and Development (OECD) economies finds, · Straightforwardness for example, that those working in firms that Openness (open-mindedness) · Curiosity innovate use more nonroutine cognitive skills · Imagination than employees in firms that do not innovate · Creativity (Avvisati, Jacotin, and Vincent-Lancrin 2013). · Tolerance The skills found to most distinguish those Emotional stability (emotion · Stress resistance working in innovating firms include “coming regulation) · Optimism up with new ideas and solutions” (creativ- · Control of one’s ity); “willingness to question ideas” (critical emotions thinking); and “ability to present new ideas or Source: Adapted from Kataoka and Alejo 2020. products to an audience” (communication). 120 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

Firm-level studies from Ireland and Norway New evidence from developing East Asia underscore the importance of higher-order also indicates that innovative firms demand technical skills—those associated with greater workers with a range of advanced skills R&D capabilities—for fostering innovation New evidence, based on data from recent at the technological frontier (Doran and Ryan employer-employee linked surveys in China 2014; Engen and Holen 2014). and Vietnam, sheds further light on how Recent research also highlights the key role skills differ across more- and less-innovative of socioemotional skills in fostering innova- firms in developing East Asia. The analysis of tion. One study of innovation in the biotech these surveys, carried out for this study, ben- industry in Taiwan, China, finds that greater efits from unusually rich data on (a) firms’ conscientiousness (task performance), open- characteristics and innovation activities; (b) ness (open-mindedness), extraversion (inter- employees’ education, training, cognitive personal engagement), and emotional stability achievement (in Vietnam), and socioemo- (emotion regulation) are all associated with tional skills; and (c) employees’ job-related firms’ overall innovation performance, while tasks (box 4.1). conscientiousness, agreeableness (collabora- Both employer-employee surveys include tion), and extraversion are associated specifi- data on multiple dimensions of firm innova- cally with technological innovation (Hsieh, tion, such as whether a firm carries out prod- Hsieh, and Wang 2011). Moreover, a study on uct and process innovation, whether it has an innovation in the marine tourism industry in R&D department, and whether it has gen- Taiwan, China, finds that greater agreeable- erated new patents. Based on these multiple ness, extraversion, and openness are all sig- indicators of firm-level innovation activity, an nificantly correlated with three key innovative innovation score—or “innovation intensity skills: “idea generation,” “idea promotion,” index”—is generated to help clarify the types and “idea implementation” (Chen, Wu, and and levels of skills demanded by highly inno- Chen 2010). vative versus less or non-innovative firms.

BOX 4.1 Employer-employee linked survey data from China and Vietnam on skills for innovation

The analysis of skills for innovation presented in inces: Guangdong, Jiangsu, Jilin, Hubei, and Sichuan. this chapter draws on two recent employer-employee The survey samples are designed to be representative linked surveys from China and Vietnam. These surveys at the province level. contain detailed firm-level data, including on innova- The CEES 2018 data include multiple measures tion, drawing on lessons from the World Bank Enter- of firm innovation, including measures of prod- prise Surveys. They also provide unusually rich data uct and process innovation, whether a firm has an on employee skills and task allocation, building on the R&D department, whether it has generated inven- lessons from other recent surveys on workers’ skills, tion patents in the past three years, and whether it is such as the Organisation for Economic Co-operation a high-tech firm. These multiple measures facilitate and Development’s (OECD) Programme for the Inter- the creation of an innovation score—or “innova- national Assessment of Adult Competencies (PIAAC) tion intensity index”—for each firm, against which and the World Bank’s Employer Skills Toward it is possible to analyze the skills and characteristics Employability and Productivity (STEP) surveys. demanded of workers in more-innovative firms compared with those in less-innovative firms. China Employer-Employee Survey The survey also includes multiple measures of In 2018, the China Employer-Employee Survey workers’ human capital, including education lev- (CEES) collected responses of 2,001 manufacturing els, fields of study (among those with vocational firms and 16,379 workers from five Chinese prov- and higher education), training, and socioemotional box continues next page Skills and Finance for Innovation 121

BOX 4.1 Employer-employee linked survey data from China and Vietnam on skills for innovation (continued)

skills, as well as detailed measures related to workers’ 849 staff in four occupational categories: managers, tasks on the job (Park and Xuan 2020). professionals, technicians, and clerks. The survey was designed to capture firms’ engagement in the Enterprise Survey on Innovation and Skills, Vietnam four main types of innovation defined in the third The 2019 Enterprise Survey on Innovation and Skills edition of the Oslo Manual: product, process, orga- (ESIS) data for Vietnam was implemented in five nizational, and marketing innovations (OECD and Vietnamese provinces: Hanoi, Bac Ninh, Da Nang, Eurostat 2005). Ho Chi Minh City, and Binh Duong. This survey was The survey collected data on a range of firm- designed to be representative of firms in the manufac- level variables, including innovation activities and turing and information and communication technol- managers’ perceptions about factors that hamper ogy (ICT) services sectors and included state-owned innovation. It also collected detailed data on work- enterprises. From four strategic sectors—high-skilled ers’ human capital and skills, including through an innovator, medium-skilled innovator, labor intensive, assessment of employees’ literacy and cognitive skills. and ICT services—201 firms were randomly selected The literacy assessment measured cognitive achieve- from the Vietnam Enterprise Registry 2017 data. ment using internationally recognized proficiency Given the sampling design, it is likely that the levels, as in the International Adult Literacy and firms surveyed are more engaged in innovation activ- Life Skills (IALS) and Adult Literacy and Life (ALL) ities than the average Vietnamese firm. ESIS collected skills surveys. ESIS also implemented an assessment information on firm characteristics from 201 man- of workers’ socioemotional skills (Miyamoto and agers and on employee background and skills from Sarzosa 2020).

Using the innovation intensity index, firms education), and more workers with STEM- in both the China Employer-Employee Survey related degrees (figure 4.1).5 Indeed, average (CEES) and Vietnam’s Enterprise Survey on employee education levels rise with firms’ lev- Innovation and Skills (ESIS) were divided into els of innovation intensity. three categories: Cognitive abilities . In Vietnam, the survey measured employees’ cognitive skills • Low-innovation firms: those that have using direct assessment of literacy linked to carried out only one or no innovation two internationally recognized proficiency activities identified in the survey scales: the International Assessment of • Medium-innovation firms: those that Adult Competencies (PIAAC) and the Adult have carried out two or three innovation Literacy and Life Skills Survey (ALL). Scaled activities from 0 to 5, a score of 2 captures basic pro- • High-innovation firms: those that have ficiency, whereas a proficiency level 3 denotes carried out four or five innovation the minimum level required for individuals to activities. autonomously perform nonroutine tasks— Education in innovative firms . Using this cat- often taken as the level of literacy required egorization, consistent patterns were found to function effectively in 21st-century work- regarding the education and skills profiles of places (Miyamoto and Sarzosa 2020).6 high-innovation firms relative to medium- or Consistent with the educational composi- low-innovative firms in China and Vietnam. tion of the workforce found in more-innova- In both countries, high-innovation firms tend tive firms, an analysis of employees’ cognitive to employ younger workers, more-educated skills indicates that advanced cognitive skills— workers (specifically, those with a college problem solving ability, capacity for critical 122 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

FIGURE 4.1 More-innovative firms in China and Vietnam have more highly educated employees and a higher share with STEM degrees

a. China, 2018a b. Vietnam, 2019b 35 80 30 70 25 60 50 20 40 15 30 10 20 5 Share of employees (%) of employees Share Share of employees (%) of employees Share 10 0 0 High school College STEM Management, High school College STEM Management, educated educated majors economics, educated educated majors economics, or law majors or law majors

Low-innovation rms Medium-innovation rms High-innovation rms

FIGURE 4.2 Stronger cognitive and socioemotional skills are thinking, ability to learn, and ability to man- associated with greater innovation intensity among firms in age complexity—are associated with higher Vietnam, 2019 innovation intensity among Vietnamese firms (figure 4.2). Although one must be careful 2.0 about asserting a causal relationship from 1.8 this analysis, the strength of the association 1.6 between employees’ cognitive skills and firm- 1.4 1.2 level innovation is noteworthy. Firms whose 1.0 employees have advanced cognitive skills— 0.8 averaging proficiency level 3 or above—score 0.6 1.5 points higher on an “innovation inten- 0.4 sity index” (on a 0–5 scale) than firms whose (0–5 scale) relative to baseline (0–5 scale) relative Change in innovation intensity 0.2 employees have an average proficiency level 0 1 or below. Further analysis of the Vietnam Literacy Literacy Interpersonal Emotion data suggests, moreover, that having employ- Level 2 (basic Level 3+ engagement regulation ees with advanced cognitive skills is particu- pro ciency) (advanced larly important to firms engaged in invention cognitive skills) (Miyamoto and Sarzosa 2020). CognitiveSocioemotional Socioemotional skills . As figure 4.2 also Source: Miyamoto and Sarzosa 2020, using 2019 Enterprise Survey on Innovation and Skills (ESIS) shows, greater socioemotional skills in the data for Vietnam. form of greater interpersonal engagement Note: ESIS collected responses from 201 manufacturing and information and communication technology (ICT) services firms and 849 staff in five Vietnamese provinces: Hanoi, Bac Ninh, Da Nang, Ho Chi Minh (“extraversion”) and emotion regulation City, and Binh Duong. The figure compares the difference in “innovation intensity” between firms (“emotional stability”) are also associated with based on average employee literacy proficiency and scores from socioemotional skills assessments. For cognitive skills, changes in innovation intensity are measured against a baseline of a literacy proficiency greater innovation intensity among firms. The level of 1 and below (omitted category). For socioemotional skills, changes are measured as the estimated association between interpersonal engagement effect of a 1 standard deviation increase in the average assessment score. The “innovation intensity index” measures firms by the number of innovation activities undertaken—on a 0–5 scale, where 0 skills and innovation appears to be especially equals no firm-level innovation activities (least innovation intensive) and 5 equals all the measured strong with respect to firms engaged in tech- innovation activities (most innovation intensive) as defined in the ESIS. The darker bars indicate estimates are statistically significant at the 5 percent level or above. Literacy Levels 2 and 3+ capture increasing nology adoption and diffusion (Miyamoto and cognitive skill levels, where Literacy Level 3 represents the minimum proficiency level required for Sarzosa 2020). individuals to autonomously perform nonroutine tasks. Skills and Finance for Innovation 123

These findings on cognitive and socioemo- FIGURE 4.3 High-innovation firms in China exhibit high demand tional skills and innovation in Vietnam are for workers with strong socioemotional skills consistent with earlier analysis for Vietnam using the 2013 Skills Toward Employability 0.12 and Productivity (STEP) Employer Survey 0.10 (Macdonald 2019). Specifically, managers and professionals working in innovating firms 0.08 (defined as firms having R&D units) report engaging in more activities requiring both rou- 0.06 tine and nonroutine cognitive skills (reading, 0.04 math, and complex problem solving) than do managers and professionals working in non- 0.02 innovating firms. Managers and professionals 0 in innovating firms also report using a greater range of socioemotional skills—engaging regu- –0.02 Coe cient (relative to low-innovation rms) (relative Coe cient larly in persuading clients or coworkers and –0.04 interacting with teams—than do their counterparts in non-innovating firms. Stronger socioemotional skills are significantly associated with greater innovation inten- Collaboration Task performance sity in China as well (figure 4.3). Specifically, Open-mindednessEmotion regulation

high-innovation firms demonstrate high Interpersonal engagement demand for employees who exhibit strong task Medium-innovation rms High-innovation rms performance (“conscientiousness”), interpersonal skills (“extraversion”), and an ability to Source: World Bank elaboration, based on Park and Xuan 2020, using the 2018 China Employer- Employee Survey (CEES). collaborate and work in teams (“agreeable- Note: The CEES collected responses of 2,001 manufacturing firms and 16,379 workers from five ness”). This higher demand for strong socio- Chinese provinces: Guangdong, Jiangsu, Jilin, Hubei, and Sichuan. Firms are categorized based on “innovation intensity,” measured by the number of innovation activities undertaken, scaled emotional skills holds even after controlling for from 0–5. Low-, medium-, and high-innovation are defined, respectively, as those undertaking firm and individual worker characteristics (Park 0–1, 2–3, and 4–5 innovation activities as defined in the CEES. Regression coefficients relative to 7 low-innovation firms (the omitted category). Regressions control for firm and individual worker and Xuan 2020). characteristics. Darker bars are statistically significant at the 5 percent level (or above).

Employees in highly innovative firms Managers’ skill sets are particularly perform more nonroutine cognitive tasks important Consistent with higher education and greater The data also reinforce the importance of cognitive skills, workers in highly innovative managers’ skills to innovation. Analysis firms in China and Vietnam perform more of skills across categories of workers in nonroutine analytical and interpersonal cogni- Vietnam—managers, professionals, techni- tive tasks than their counterparts in medium- cians, and clerks—indicates that managers’ or low-innovation firms (figure 4.4). In China, capabilities are important contributors to the difference between workers in highly firms’ innovation performance. innovative and less innovative firms is particu- Specifically, advanced cognitive skills larly striking regarding nonroutine analytical among managers and technicians (as mea- tasks (figure 4.4, panel a). In both China and sured by literacy proficiency levels of 3, 4, or Vietnam, employees in highly innovative firms higher), along with managers’ interpersonal carry out significantly less routine work than skills, are all positively associated with higher those in less-innovative firms. In Vietnam, the innovation intensity among Vietnamese firms difference in the extent of routine manual tasks (figure 4.5). The association between manag- between employees in high-innovation firms ers’—and especially technicians’—cognitive and those in low- and medium-innovation abilities and firm innovation performance is firms is especially large (figure 4.4, panel b). particularly strong. 124 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

FIGURE 4.4 Employees in more-innovative firms in China and Vietnam have jobs that are more intensive in nonroutine cognitive analytical and interpersonal tasks

a. Task intensity in China, by type, 2018a b. Task intensity in Vietnam, by type, 2019b

0.3 0.3

0.2 0.2

0.1 0.1 x 0 0 Inde

–0.1 Coefficient –0.1

–0.2 –0.2

Low-innovation firms Medium-innovation firms High-innovation firms

Sources: World Bank elaboration based on Park and Xuan 2020 and Miyamoto and Sarzosa 2020, using, respectively, the 2018 China Employer-Employee Survey (CEES) and the 2019 Enterprise Survey on Innovation and Skills (ESIS) for Vietnam. Note: Firms are categorized by “innovation intensity,” measured by the number of innovation activities undertaken, as captured in the respective surveys. Scaled from 0–5, low-, medium-, and high-innovation are defined, respectively, as those undertaking 0–1, 2–3, and 4–5 innovation activities. The Vietnam analysis does not include an aggregated measure of “routine task intensity;” therefore, panel b shows instead an individual measure of “routine manual” tasks. No information was included in either panel on routine cognitive tasks because none of the related regression coefficients was statistically significant. a. The CEES collected responses of 2,001 manufacturing firms and 16,379 workers from five Chinese provinces: Guangdong, Jiangsu, Jilin, Hubei, and Sichuan. b. ESIS collected responses from 201 manufacturing and information and communication technology (ICT) services firms and 849 staff in five Vietnamese provinces: Hanoi, Bac Ninh, Da Nang, Ho Chi Minh City, and Binh Duong.

FIGURE 4.5 Managers’ and technicians’ skills are particularly important to firm innovation in Vietnam, 2019

2.0 y 1.5

1.0

0.5

0.0 index (scale 0–5) –0.5 Change in innovation intensit –1.0 Managers ProfessionalsTechnicians Clerks

Literacy Level 2 (basic proficiency) Literacy Level 3+ (advanced cognitive skills) Literacy Level 4+ (advanced cognitive skills) Interpersonal engagement Emotion regulation

Source: Miyamoto and Sarzosa 2020, using the 2019 Enterprise Survey on Innovation and Skills (ESIS) data for Vietnam. Note: ESIS collected responses from 201 manufacturing and information and communication technology (ICT) services firms and 849 staff in five Vietnamese provinces: Hanoi, Bac Ninh, Da Nang, Ho Chi Minh City, and Binh Duong. The figure compares the difference in “innovation intensity” between firms based on average literacy proficiency and on socioemotional skills assessment scores of staff in each occupation. For cognitive skills, changes in innovation intensity are measured against a baseline defined as literacy proficiency levels of 1 and below (omitted category). For socioemotional skills, changes are measured as the estimated effect of a 1 standard deviation increase in the average assessment score. The firm innovation intensity index measures the number of innovation activities a firm undertakes on a 0–5 scale, where 0 equals no firm-level innovation activities (least innovation intensive) and 5 equals all measured innovation activities (most innovation intensive). Darker bars indicate estimates that are statistically significant at the 5 percent level or above, except for “interpersonal engagement,” which is significant at the 10 percent level. Literacy levels 2, 3, and 4 capture increasing cognitive skill levels, where level 3 represents the minimum proficiency level required for individuals to autonomously perform nonroutine tasks. Skills and Finance for Innovation 125

Returns to education and advanced skills high-skilled workers remains an open ques- are especially high in high-innovation firms tion. Recent studies in the United Kingdom Emerging evidence suggests that innovation (Aghion et al. 2019) and Brazil (Cirera and skills are rewarded in developing East Asia’s Soares Martins Neto 2020) find positive and labor markets. In China, for example, work- significant wage premia associated with inno- ers in more-innovative firms receive higher vation, after controlling for education and wages. The wage gap between low-innovation other worker characteristics. Although the firms and high-innovation firms is 14–67 study of the United Kingdom finds that the percent, depending on the occupation (Park wage premium is larger for low-skill occupa- and Xuan 2020). Analysis of the CEES data tions, the evidence on the distribution of the indicates that returns to college education are wage premium in Brazil is inconclusive. higher in high-innovation firms than in low- The positive returns associated with more- or medium-innovation firms, as are returns advanced innovation skills have also raised to such socioemotional skills as interpersonal questions about whether technology adop- engagement (extraversion) and task perfor- tion and innovation represent an opportu- mance (conscientiousness), even after control- nity or a threat to workers and employment ling for workers’ education and experience. more broadly. These questions are of particu- These higher returns appear to be driven lar interest given evidence from high-income by higher wages among the most-skilled economies of job polarization resulting workers (such as managers, technical workers, from technology adoption and automation. and salesmen) rather than by wages of lower- Although analysis of the overall employment skilled workers (such as office and production effects of innovation in low- and middle- workers). Separate analyses from China and income economies is still relatively scarce, Vietnam also find positive returns to perfor- the available evidence suggests that innova- mance of the nonroutine analytical tasks that tion raises employment—among high-skilled are critical to innovation, again controlling for workers and low- to medium-skilled workers workers’ education and experience (Bodewig alike—when the productivity gains associated et al. 2014; Du and Park 2017).8 Whether with diffusion or invention enable firms to wage premia in innovative firms always favor expand their output and to grow (box 4.2).

BOX 4.2 The effects of technology adoption and innovation on employment

One question in the minds of policy makers is how of middle- or low-skilled workers (Autor, Dorn, and adoption of new technologies and innovation more Hanson 2015; Graetz and Michaels 2018). broadly affect employment, especially among low- At least one recent study on automation also found and middle-skilled workers. Concerns about the significant negative effects of robot use on aggregate potential employment effects of innovation have employment. Examining the effects of robot use on been amplified in recent years as a result of research local labor markets in the United States from 1990 on high-income economies (for example, in Europe to 2007, the study found that each additional robot and the United States) showing that technology per 1,000 workers reduced employment on the order adoption—whether in the form of computers, infor- of 3.0–5.6 jobs per robot (Acemoglu and Restrepo mation and communication technologies (ICTs) 2017). Recent research on developing and emerging or industrial robots—has resulted in labor market economies has thus far found little evidence of job polarization (the relative growth of employment polarization, however (Das and Hilgenstock 2018; in high-skill and low-skill jobs accompanied by a Maloney and Molina 2016). decline in middle-skill jobs) or, alternatively, a shift Although most research on technology, skills, in demand toward high-skilled labor at the expense and jobs has focused on high-income economies,

box continues next page 126 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

BOX 4.2 The effects of technology adoption and innovation on employment (continued)

a growing number of studies are examining these innovation on employment, with an estimated elas- issues in low- and middle-income countries. This ticity of jobs created close to unity (Harrison et al. evidence generally finds that technology adoption 2014). Analysis of a 20-year panel of firms in raises relative demand for more-skilled workers Germany similarly finds a significant positive rela- (Mason and Shetty 2019). The evidence on employ- tionship between innovation and employment, in ment often paints a more positive picture, however. the cases of both product and process innovation A recent study of technology adoption among firms (Lachenmaier and Rottmann 2011). And an analy- in Argentina, Brazil, Chile, Colombia, and Mexico sis of longitudinal data on Italian manufacturing finds, for example, that investments in ICT raised firms finds a significant positive relationship, albeit firm productivity, enabling firms to expand their relatively small, between innovation and employ- output, which in turn led to net employment growth ment (Piva and Vivarelli 2005). for both low-skilled and high-skilled workers (Dutz, A smaller but growing literature on low- and Almeida, and Packard 2018). middle-income countries likewise finds that product A study of technology adoption (and trade) in innovation is associated with increased employment Vietnam comes to similar conclusions. Whereas at the firm level. A recent study of firm-level innova- computer use tends to be “labor saving” and raises tion in Argentina, Chile, Costa Rica, and Uruguay relative demand for higher-skilled workers, over- finds, for example, that both process and product all employment has increased with rising demand innovation have positive effects on employment for Vietnamese production and exports (Poole et al. (Crespi, Tacsir, and Pereira 2019). 2017). Cross-country evidence (based on World Bank These country-level findings are reinforced by a Enterprise Survey data from 82 low- and middle- recent study that examines the employment effects income countries) is consistent with the country-level of innovation using Enterprise Survey data from findings: increased adoption of digital technologies is 53 low- and middle-income countries (Cirera and associated with greater firm-level demand for labor, Sabetti 2019). This cross-country analysis indicates controlling for other factors (Cusolito, Lederman, and that product innovation, when successful in raising Peña 2020). firm-level sales, also has a positive direct impact on As with the literature on technology adoption, employment. The evidence suggests, moreover, that most research on the effect of innovation on employ- positive employment effects tend to be larger in low- ment to date has focused on high-income economies and middle-income countries, where innovations are and finds evidence of “skill-biased technological commonly more incremental in nature, than in high- change”—the relative increase in demand for more- income countries. In contrast to product innovation, skilled labor resulting from the introduction of a new however, Cirera and Sabetti (2019) find no evidence technology or technology-driven change in produc- that either process or organizational innovation have tion methods or business processes. Nonetheless, any effect on employment. the empirical literature that focuses specifically on The relationship between technology adoption, innovation more consistently finds a positive rela- innovation, and employment is complex. Nonetheless, tionship between product innovation and firm-level despite evidence of labor polarization in high-income employment, often related to expanded demand for economies, emerging evidence for low- and middle- firm output. income countries suggests that technology adoption For example, a study of innovation and and innovation have the potential to increase employ- employment in France, Germany, Spain, and the ment, particularly when increased productivity and United Kingdom finds a positive effect of product profitability results in firm expansion. Skills and Finance for Innovation 127

Ongoing skills challenges in developing workers mirrors the range of cognitive, tech- East Asia nical, and socioemotional skills found to be important among high-innovation firms. Innovative firms cite inadequate skills as a major constraint to their operations Many countries still struggle to build Despite the importance of human capital sufficient basic skills among students—the in fostering innovation, most countries in foundation for more-advanced skills among developing East Asia still face critical skills workers challenges. These challenges are reflected in The concerns that firms express about numerous human capital and skills-related human capital and skills gaps are also indicators for the region. Firms throughout reflected in the results of the most recent the region report a lack of adequate skills as a Programme for International Student crucial constraint to their operations. In most Assessment (PISA) tests, administered to of the region’s countries, firms that innovate 15-year-olds in several developing East report a lack of adequate skills as a greater Asian countries in 2018.11 Although PISA constraint to their operations than do non- scores in four major cities in China are well innovating firms. above those that would be predicted from More than one-quarter of innovating China’s gross domestic product (GDP) per firms in Cambodia and Indonesia, and nearly capita, assessment scores on reading, math, one-third of innovating firms in Mongolia, and science in Indonesia, Malaysia, the report that inadequate skills are either a Philippines, and Thailand are considerably 9 major or severe obstacle to their operations. below what would be expected given their Inadequate skills appear to be an even greater income levels (figure 4.7).12 Although the problem when firms try to hire new work- figure presents only reading scores, similar ers. Indeed, a significant majority of inno- patterns are seen in the 2018 PISA scores vating firms in Indonesia, the Philippines, for math and science.13 and Vietnam (76 percent, 79 percent, and Together, these test results suggest that 64 percent, respectively) report inadequate many of the region’s countries face continu- skills as an obstacle when trying to hire new ing challenges in building the “foundational employees. skills” upon which the more-advanced skills Moreover, innovating firms across the for innovation must be built. That most coun- region report a range of skills deficits when tries’ PISA scores have not improved over trying to recruit new employees. The specific time is also a matter of concern; indeed, in the skills deficits, as well as the level of severity, case of reading, some countries’ performance differ from country to country, but the mag- has even deteriorated (World Bank 2020). nitude of the problem is often reported to be large. In Indonesia, Malaysia, Myanmar, The region’s countries must also keep the Philippines, Thailand, and Vietnam, for building their technical skills bases example, over 50 percent of all innovative On top of the challenges developing East firms cite lack of managerial and leadership Asian countries face in building their pop- skills as a challenge when hiring new work- ulations’ foundational skills, they need ers (figure 4.6). And more than half of all to continue building their technical skills innovative firms in at least three of those six bases, especially if they are to spur innova- countries also cite a scarcity of interpersonal tion at the technological frontier. Although and communication skills, foreign language the region’s share of tertiary students skills, computer and information technology graduating with degrees in STEM fields (IT) skills, and technical (non-IT) skills as is relatively high, tertiary enrollment rates critical challenges when it comes to hiring.10 in many countries are lower than would Interestingly, the range of skills challenges be predicted given their per capita income that innovating firms report facing in hiring levels (figure 4.8). 128 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

FIGURE 4.6 Most innovative firms in developing East Asia report difficulties in hiring workers with adequate skills

a. Indonesia, 2015 b. Malaysia, 2015

Foreign language skills Foreign language skills

c. Myanmar, 2016 d. Philippines, 2015

Managerial and leadership skills Managerial and leadership skills

0 20 40 60 80 100 0204060 80 100 Percent Percent

e. Thailand, 2016 f. Vietnam, 2015

Foreign language skills Managerial and leadership skills

0204060 80 100 0204060 80 100 Percent Percent

The quality of STEM training remains FIGURE 4.7 Several developing East Asian countries underperform an issue, however, given the relatively weak in basic skills formation, creating challenges for the development of development of foundational skills in most more-advanced skills of the region’s countries. Even in China, where PISA scores are high, rapid expansion 600 in the number of tertiary-level students over 550 B-S-J-Z China Singapore the past two decades has posed significant Korea, Rep. challenges in ensuring quality across the sys- 500 Japan tem. Between 1999 and 2016, college enroll- Vietnam ments nearly quintupled (from 1.6 million 450 to 7.5 million) in China, while the number Malaysia of tertiary institutions increased by roughly 400 Thailand 2.5 times (from 1,071 to 2,596). Although Indonesia China has several world-class universities, 350 there remains significant variation in educa- PISA 2018 mean score in reading Philippines tional quality across institutions, with a siz- 300 able share of graduates getting low returns 8.5 9.0 9.5 10.0 10.5 11.0 11.5 12.0 (World Bank and DRC 2019). Log GDP per capita (PPP constant 2011 internatioanl $) Ensuring that countries’ education systems support the development of both adequate Sources: World Bank elaboration, based on Programme for International Student Assessment (PISA) data and the World Development Indicators database. foundational skills and more-advanced cogni- Note: The light blue points designate 6 of the 10 “developing East Asia” countries studied in this tive, socioemotional, and technical skills will report. The others (Cambodia, Lao PDR, Mongolia, and Myanmar) did not participate in PISA in either 2018 or 2015. “B-S-J-Z China” scores are not nationally representative but capture test results be critical to fostering innovation, whether from urban Beijing, Shanghai, Jiangsu, and Zhejiang only. Vietnam data are from 2015; no PISA defined as adoption of existing technologies results were reported for 2018. GDP = gross domestic product; PPP = purchasing power parity. or as invention at the technological frontier.

FIGURE 4.8 Although developing East Asian countries often produce a higher share of STEM graduates than do countries with similar income levels, most have lower tertiary enrollment rates

a. Share of STEM graduates in recent tertiary cohorts, circa 2017 b. Gross tertiary enrollment rates, circa 2017

50 150

Malaysia 40 China Myanmar Singapore 100 United States Korea, Rep. Singapore Philippines Mongolia Thailand 30 Korea, Rep. Vietnam Mongolia Thailand 50 20 Indonesia Philippines Lao PDR China Malaysia Vietnam Indonesia United States 10 Cambodia Lao PDR 0 Myanmar

Gross enrollment in tertiary education (%) Cambodia

Share of STEM graduates in tertiary education (%) 0 6 8 10 12 6 8 10 12 Log GDP per capita (PPP constant 2011 international $) Log GDP per capita (PPP constant 2011 international $)

Sources: World Bank elaboration, based on United Nations Educational, Scientific and Cultural Organization (UNESCO), World Development Indicators, and World Economic Forum data. Note: GDP = gross domestic product; PPP = purchasing power parity; STEM = science, technology, engineering, and mathematics. 130 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

Finance for innovation to commercialization is often long. Key frictions characteristic to the financing of firm Financial markets are an enabling factor invention include the nonrival nature of for innovation knowledge (that is, development by one firm Deep, well-functioning financial markets does not prevent the use by another) and promote growth through innovation by effi- asymmetric information between firms and ciently allocating capital to the firms with the prospective financiers (Hall and Lerner 2010). most promising projects (Levine 2005). Firms Nonrival knowledge . Knowledge spill- with new and better ideas are provided with overs to other firms may make the private the means to finance their projects, gaining rate of return to innovation lower than the competitive advantages through increases socially optimal return to innovation. Lack in productivity (Beck, Levine, and Loayza of full appropriability can reduce investments 2000) and eventually becoming more profit- in innovation. Innovative firms may also be able. Firms that fail to innovate may exit the reluctant to share the outcomes of invention market, allowing for an efficient allocation of in early stages of development with investors (often scarce) financial resources in the econ- for fear of losing the intellectual property omy. In turn, this efficient allocation leads to right of their innovation to competitors. This higher productivity and, ultimately, growth. reluctance increases information asymmetries Financial markets also perform important in the market, which, together with the high- screening and monitoring functions—col- risk nature of invention, may make investors lecting and processing information that can reluctant to invest, again resulting in under- reduce asymmetric information problems.14 investment. Stronger intellectual property New innovative projects are evaluated and rights, accounting standards, and disclosure financed, facilitating development and com- requirements should help to alleviate this mercialization. Finance is key at all stages of problem.15 Nevertheless, these frictions may the innovation process, as the expected mon- lead to underinvestment in invention. etary returns from innovation are realized Asymmetric information . Innovative proj- with a lag. Specifically, monetary outflows ects exacerbate asymmetric information (expenses) for innovation projects precede between the entrepreneur and the financier. monetary inflows (revenues), creating a finan- Ex ante, evaluating invention is problem- cial gap. The time lag and the amount of the atic because returns are highly uncertain and financial gap depend on the type of innova- skewed. The time between the development of tion. Firms can plug the financial gap using the business concept and commercialization is either internal or external sources of finance. also often long, making for an illiquid investment. Although the innovator and financier face the same uncertainty, the innovator has Different types of innovation face better technological and business expertise, different types of financing challenges, and could also be overoptimistic about the which may lead to underinvestment chances of success. Thus, a contingent con- Innovation, as defined in this report, entails tract is hard to achieve, and this may lead to both diffusion and adoption of new technolo- moral hazard on the inventor’s part,16 which gies as well as invention. These two distinct in turn may lead to underinvestment. types of innovation commonly have different Relative to invention, the diffusion and project timelines and cash-flow patterns and adoption of new technologies is safer (in face different financial frictions and agency terms of probability of success), its time- costs. Thus, they often face different financ- line is shorter, and firms’ loss exposure is ing challenges. more contained. Agency problems are also Invention is difficult to finance in freely more limited because the project’s novelty is competitive markets because cash flows are more limited, which makes it easier to col- uncertain and the time line from development lect information to appraise the investment. Skills and Finance for Innovation 131

TABLE 4.3 Market frictions have distinct effects on different types of firm innovation

Likelihood of Innovation Asymmetric Innovation type Returns success Loss exposure project time line information Moral hazard Diffusion or adoption Moderate Moderate Moderate Short Moderate Moderate Invention Large but highly Low Very high Long High High uncertain Source: Original table for this publication. Note: This table provides a synthetic representation for illustrative purposes.

Technology adoption may nonetheless be Financial constraints may also hamper the costly, depending on the tacit know-how quality of firms’ innovation by limiting their involved as well as country- and firm-level abilities to adopt new technologies in pro- characteristics. If firms cannot absorb this duction processes and to incorporate them in cost with their available resources and can- final products and services. Indeed, financial not find external resources, they may decide constraints make firms less likely to invest to delay or abandon the adoption of the new in risky, exploratory R&D that could lead technology altogether. Table 4.3 summarizes to productivity-enhancing discovery inno- the key market frictions faced by innovating vations and more likely to invest in R&D firms, both those inventing and those under- that leads to low-quality patented innova- taking diffusion and adoption. tions. This appears to be the case in China, where recent research finds that financially Financial constraints impede firm-level constrained firms orient their R&D toward innovation and may hamper project quality smaller changes that can raise firms’ profits In light of the challenges associated with the in the short term rather than toward more- financing of firm innovation, both theory significant innovations that could increase and empirical evidence support the view that firms’ longer-term productivity (Cao 2020). financial constraints and the type of funding affect firms’ decisions to innovate as well as Financial structure is also relevant for the extent to which they innovate.17 Financial firms’ innovation processes constraints affect firms’ decisions to innovate The financial structure of firms also influ- and adopt new technologies because the pro- ences both the decision to innovate and the pensity of firms to undertake innovative proj- quality of innovation. Internal financing— ects depends on their ability to satisfy current specifically, retained earnings and new equity capital expenditures and to borrow in the from existing shareholders—is the main future to meet potentially large adjustment source of funding for most innovation proj- costs during difficult times (Hall and Lerner ects (Czarnitzki and Hottenrott 2011). These 2010).18 There is also the need to accumulate sources of funding are especially relevant for capital over time to finance a new business small and medium-size enterprises and start- or firm innovation unless external sources ups, which cannot rely on banks or the finan- of finance help smooth investment over time cial market because of the lack of reputation (Berger, Molyneux, and Wilson 2020). or lack of the collateral needed to benefit In the absence of external finance, finan- from external sources of finance. cially constrained firms could slow or delay Relying on operating profits for innova- investment in innovative projects or could tion projects is far from ideal, however. Cash prematurely stop or abandon such projects flows are volatile sources of finance (Brown, altogether. Financial constraints can thus Fazzari, and Petersen 2009), and raising new impede the generation of new (novel) inno- equity can be costly and sometimes unwar- vation and, ultimately, the aggregate level of ranted (Hottenrott and Peters 2012). As a invention in an industry or economy.19 result, innovative projects with high initial 132 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

costs may be delayed, postponed, or aban- including management skills and expansion doned for lack of external finance. This issue of firms’ business networks.21 is likely to be most prominent among smaller, younger firms with greater constraints to Firms in developing East Asia face accessing external finance. finance-related obstacles to innovation External sources of finance not only plug the financial gap of innovative firms but also Financial systems in much of the region can affect a firm’s managerial incentives, gov- remain bank based, limiting firms’ options ernance, and risk-taking behavior, because for financing innovation projects of the presence of external monitoring by Differences in firms’ access to finance depend, financial intermediaries.20 In addition, diverse in part, on the level of financial development sources of finance have distinct characteris- of each country. Most developing East Asian tics that can help to reduce market frictions economies are characterized by bank-based in different manners. Regarding informa- financial systems that lack the necessary tion, for example, capital markets provide diversity to foster firm innovation. Selected price signals and encourage the borrowing indicators of financial market development firms to publicly release hard data about their in developing East Asia show that banks enterprises. Banks also fill information gaps are the dominant source of finance in most by collecting soft information that is mostly of the region’s countries except for China held privately. As a result, banks are better (figure 4.9).22 equipped to operate in more opaque and risky In parts of developing East Asia where environments. Moreover, different forms of capital markets are less developed, undertak- funding often bring additional benefits to ing invention-oriented innovation may be firms beyond the simple provision of finance, precluded. Once again, these constraints may

FIGURE 4.9 Banks remain the dominant source of finance to firms in most of developing East Asia (except China)

140

120

100

Share of GDP (%) Share 40

0 2000–05 2006–11 2012–17 2000–05 2006–11 2012–17 2000–05 2006–11 2012–17 2000–05 2006–11 2012–17 2000–05 2006–11 2012–17 2000–05 2006–11 2012–17 2000–05 2006–11 2012–17 China Indonesia Malaysia Mongolia Philippines Thailand Vietnam

Banks Equity Corporate bonds

be particularly severe for younger and smaller both the quantity and quality of innovation firms. Not only do such firms have more lim- projects pursued. Interestingly, few coun- ited access to equity financing (because of the tries in the region exhibit significant differ- small market size), but they also face greater ences between innovative and non-innovative challenges in raising debt financing because firms in the financing of fixed assets through asymmetric information problems are com- external funding sources. Although the lack monly worse for smaller and younger firms of external financing may constrain both than for larger and more established firms. innovative and non-innovative firms, it Younger and smaller firms are thus more may represent a greater constraint on inno- likely than larger firms to allocate investments vative firms—especially those focused on away from intangible assets and to delay or invention—given the potential costs, com- abstain from innovation projects. plexity, and uncertainty of investments associated with such projects. Most firms in the region still rely on internal financing, which can constrain Improving the availability of external their abilities to finance innovation projects financing would increase the likelihood of By far, most of the innovative firms in develop- firm-level innovation ing East Asia still use internal funds to finance Low use of external finance may reduce the purchase of fixed assets (figure 4.10).23 innovation. Indeed, an analysis of a global And this lack of external finance may affect sample of firms finds that firms with more

FIGURE 4.10 In developing East Asia, most firms—both innovative and non-innovative—finance fixed assets using internal funds

a. Firms engaged in adoption or di usion b. Firms that spend on research and development (R&D)

Cambodia Cambodia

China China

Indonesia Indonesia

Lao PDR Lao PDR

Malaysia Malaysia***

Mongolia Mongolia

Myanmar Myanmar

Philippines Philippines***

Thailand* Thailand

Vietnam Vietnam**

010203040 50 60 708090 100 021004 3006 5008 709010 00 Mean percentage of xed assets nanced internally Mean percentage of xed assets nanced internally

Non-innovative rms Innovative rms

Sources: World Bank elaboration using World Bank Enterprise Survey data. Note: The graph reports the mean percentage of fixed assets financed through internal funds for firms that either adopt a new product, service, or process (panel a) or spend on research and development (R&D) (panel b). “Developing East Asia” refers to the 10 middle-income countries covered in this study: Cambodia, China, Indonesia, Lao PDR, Malaysia, Mongolia, Myanmar, the Philippines, Thailand, and Vietnam. Statistically significant difference in the mean value for each regional group in the two reference years, using a two-tailed test: *** p < 0.01 ** p < 0.05 * p < 0.1 134 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

diversified financial structures (that is, using Risk finance is growing in the region, and external funding instead of retained earnings this could spur firm innovation or other internal resources) are more likely Several alternative sources of finance to to engage in innovation (Mare, de Nicola, internal funds and bank lending, commonly and Liriano 2020). Moreover, firms that use known as risk finance (for example, private external sources of funding undertake more equity, venture capital, or angel investing), innovation activities. This appears to be the are often associated with firm invention case for at least a subset of developing East activities because this form of financing spe- Asian countries, where funding alternatives cifically targets high-risk, high-return invest- to internal funds and bank financing are sig- ments.24 The amount of risk capital financing nificantly associated with firms’ decisions to remains limited around the world, however, innovate as well as the extent of firm innova- especially in low- and middle-income coun- tion activity. tries. As noted above, most external financ- The association between financial struc- ing of firms in developing East Asia still ture and firm-level innovation is stronger in comes from banks. some countries than in others (figure 4.11). Nonetheless, countries in developing This reinforces the idea that innovation East Asia perform relatively better than activities and outcomes depend on several those in other regions in attracting risk country-specific factors—including, from the capital investment, which in 2018 was perspective of innovation finance, the pres- equivalent to 0.34 percent of GDP on ence of well-developed equity markets (Hsu, average (totaling approximately US$139 Tian, and Xu 2014). billion) (map 4.1). Moreover, in some

FIGURE 4.11 The relationship between firms’ external financing of fixed assets and firm-level innovation is stronger in some developing East Asian countries than in others

0.60

0.40

0.20

–0.20 innovation (and 90% condence interval) E ect of external nancing xed assets on rm –0.40

China Vietnam Indonesia Lao PDR Malaysia Mongolia Myanmar Thailand Cambodia Philippines

Source: Adapted from Mare, de Nicola, and Liriano forthcoming. Note: This figure shows the marginal effect (at means, marked with a diamond) of the different sources of working capital funding and fixed assets funding on a country’s firm-level innovation score, ranging from 0 (no innovation) to 4 (both adoption and research and development [R&D] expenditures undertaken). The confidence intervals for those coefficients are computed at the 10 percent significance level. The country regressions relate the innovation score for firms in the manufacturing sector to their funding sources, controlling for size, labor productivity, age, and domestic ownership. Skills and Finance for Innovation 135

MAP 4.1 Countries in developing East Asia attracted relatively large amounts of risk capital financing, as a percentage of GDP, in 2018

Share of GDP (%) 0.823 – 3.615 0.266 – 0.823 0.096 – 0.266 0.015 – 0.096 0.001 – 0.015 0.000 – 0.001 No data

IBRD 45519 | JANUARY 2021

Source: World Bank elaboration, using Crunchbase data (https://www.crunchbase.com/), Note: The map reports the amount of risk capital financing (for example, angel, venture, private, or equity crowdfunding) as a percentage of a country’s GDP in 2018. “Developing East Asia” refers to the 10 middle-income countries covered in this study: Cambodia, China, Indonesia, Lao PDR, Malaysia, Mongolia, Myanmar, the Philippines, Thailand, and Vietnam. GDP = gross domestic product.

countries, there are fast-growing private enterprises, whereas the growth of risk capital enterprises already valued at more than financing in Thailand has been more limited. US$1 billion (also known as “unicorns”), The amount of risk financing in Malaysia and such as in China (for example, Alibaba the Philippines actually declined recently rela- Group’s global parcel tracking platform, tive to previous years. Cainiao) and Indonesia (for example, ride- Several factors appear to have contributed hailing service Go-Jek). to these differential growth patterns, including country-specific differences in the ease The growth in risk finance is uneven across of starting a business, investor protection, the region, suggesting a role for financial and the insolvency framework (figure 4.13). policy Indeed, developing East Asian countries The growth of risk capital finance is uneven appear to lag behind other regions on several across developing East Asian countries of the World Bank’s Doing Business indexes,26 (figure 4.12). Chinese companies have raised especially in terms of enforcing contracts and a relatively high amount of risk capital since resolving insolvency, which may hamper the 2007, enabling China to lead the world in its development of risk finance and, thus, impede number of unicorns as of June 2019.25 Before firm innovation. the COVID-19 pandemic began, Indonesia Country-specific conditions, such as inves- and Vietnam also demonstrated upward tor preferences and government policies, may trends in the amount of capital raised by also help to explain the observed patterns of 136 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

risk capital financing. Interviews with ven- investors and scale (fund sizes) as two of the ture capitalists in the Philippines, Thailand, major challenges to investing in these coun- and Vietnam identified inexperience of tries (Scheela et al. 2015). Similarly, Nor (2015) reports that lack of funding for ven- FIGURE 4.12 Growth in risk capital financing has varied ture capital (VC) firms in Malaysia is one of considerably across developing East Asia the main reasons why the country has not seen faster growth of risk financing. Because 5,000 150,000 most of Malaysia’s funding comes from the government in the form of loans that firms 4,000 120,000 must repay rather than as equity capital, VC investors do not find it profitable to invest in 3,000 90,000 these companies due to low expected returns (Vivekarajah 2018). Besides investor inex- China

luding China 2,000 60,000 perience and the risk profile of the available xc e investment opportunities, the Philippines has 1,000 30,000 faced the additional challenge of uncertainty associated with political instability, which Risk capital nancing (US$, millions), Risk capital nancing (US$, millions), 0 0 erodes confidence especially for foreign inves-

2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 tors and further reduces incentives to invest in high-risk projects (Lopez 2019). Thailand Malaysia Vietnam Philippines Indonesia China Conclusions Source: World Bank elaboration, using Crunchbase data (https://www.crunchbase.com/). Note: The graph reports the amount of risk capital financing (for example, angel, venture, private, The availability of adequately skilled labor or equity crowdfunding) in US$, millions. “Developing East Asia” refers to the 10 middle-income countries covered in this study: Cambodia, China, Indonesia, Lao PDR, Malaysia, Mongolia, Myanmar, and risk-capital finance, while not strictly fall- the Philippines, Thailand, and Vietnam.” The figure excludes Cambodia, Lao PDR, Mongolia, and ing under the purview of innovation policy, Myanmar due to lack of data. For all countries except China, the scale is the left y-axis.

FIGURE 4.13 The factors that affect the attractiveness of venture capital vary by country across the region

100 90 80 70 60 50 40 30

Doing Business score (0–100) 20 10 0 Starting a Registering Protecting minority Enforcing Resolving business property investors contracts insolvency China Thailand Myanmar Mongolia Vietnam Malaysia Indonesia Philippines Lao PDR Cambodia

Source: World Bank elaboration from 2020 Doing Business data. Note: For more detailed explanations of the variables, see the Doing Business website: https://www.doingbusiness.org/en/doingbusiness. A score of 100 indicates the best performance within a specific topic. Lao PDR lacks a “resolving insolvency” score because data on that variable were unavailable. Skills and Finance for Innovation 137

is critical to enabling innovation at the firm As for financing innovation, several addi- and country levels. Understanding the roles tional messages emerge: of these complementary factors is particularly important because the countries of devel- • Both the availability and the type of financ- oping East Asia face ongoing challenges in ing matters for innovation at the firm level. developing an adequate skills base as well as In most of developing East Asia, how- sufficiently deep and diverse financial sectors ever, financial sectors remain heavily bank to foster innovation effectively. The specific based, having neither the depth nor the skills and financing needs and challenges dif- breadth to support innovation-led growth. fer by country and depend on whether one • Firm-level financial constraints—while not is focusing on innovation defined as diffu- as severe in developing East Asia as in other sion and adoption of existing technologies developing regions—do affect firms’ deci- or as invention—the latter requiring more sion to innovate, the extent to which they advanced skills and more sophisticated risk innovate, and the quality of innovation. financing instruments. • Diverse sources of finance can help to Regarding skills development, several key enable greater and higher-quality inno- messages emerge from the analysis in this vation by helping to reduce problems of chapter: asymmetric information between the financier and the borrower. Such information • Higher-order cognitive, socioemotional, asymmetries are more severe for projects and technical skills are critical enabling that are innovative than for those that are factors for all types of innovation— not. Information asymmetries are particu- although these skills become increasingly larly severe in the context of invention. important as firms move toward the tech- • Continued financial deepening is needed nological frontier. in developing East Asia to support greater • Inadequate provision of foundational innovation in all its forms. This becomes skills remains a fundamental challenge fac- particularly important as firms and coun- ing most developing East Asian countries. tries move up the “capabilities escalator” Their national education systems are not from a focus on technological catch-up to providing students with the basic reading, innovation at the technological frontier. math, and science skills upon which more- Meeting the challenges associated with advanced skills must be built. ensuring adequate skills and financing for • A dual challenge for the region’s policy innovation in developing East Asia will take makers, consequently, is that of (a) ensur- concerted public action. Specific directions ing that their education systems deliver for policy to strengthen these critical comple- the necessary foundational skills to their mentary factors are discussed in chapter 6. populations, while (b) strengthening the ability of education and training systems to support the types of advanced skills Notes development needed to enable innovation- led growth. 1. A study of the effects of research and • Rapidly changing technologies make con- development (R&D) expenditures and human fronting this dual challenge all the more capital on innovation behavior in Spain, urgent because they are quickly ratcheting for example, finds a positive relationship between the share of workers with university up the skills needed to innovate, whether degrees and the number of patents that firms defined as diffusion and adoption or inven- produce (Gumbau-Albert and Maudos 2009). tion. The emerging economic and social Similarly, a study of French industrial firms challenges associated with climate change finds a robust positive relationship between and the COVID-19 pandemic just add to worker training and firm-level patenting this urgency. activity (Gallié and Legros 2012). An analysis 138 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

of Finnish manufacturing firms finds, achieve one’s goals, and to develop one’s moreover, that innovative firms with more- knowledge and potential” (OECD 2013, educated employees, higher technical skills, table 2.1). The literacy proficiency levels and greater research skills are more profitable identified in the assessment, scaled from 0 to than non-innovative firms (Leiponen 2000). 5, characterize individuals’ range of cognitive 2. As discussed in chapter 3, managerial quality abilities from basic to advanced. Proficiency is critical to enabling innovation among firms. levels 1 and 2 capture individuals’ capacity to It increases innovation directly as well as retrieve information and apply it in routine and through more efficient use of R&D (Cirera, predictable ways, whereas proficiency levels Maloney, and Sarrias 2020). 3 and above capture the capacity to analyze, 3. For a review of the literature on skills for evaluate, and create new information, which innovation, see Kataoka and Alejo (2019). are precursors to fluid problem solving, critical 4. In addition to the five categories listed in thinking, and creativity. Proficiency level 3 table 4.2, the Organisation for Economic can be considered the minimum proficiency Co-operation and Development (OECD) level required for individuals to function includes a separate domain of socioemotional autonomously in carrying out nonroutine skills called “compound skills,” which tasks (Miyamoto and Sarzosa 2020). represent a combination of two or more 7. Park and Xuan (2020) also find that open- individual skills. One such compound skill mindedness (“openness to experience”) is is self-efficacy (strongly believing in one’s strongly positive and statistically significant own ability to execute tasks and achieve at the 5 percent level when controlling for goals), which represents a combination of firm-level characteristics. It is no longer conscientiousness, emotional stability, and significant when controlling for individual extraversion (OECD 2017). Bassi et al. (2012) characteristics, however. The authors find the find that greater self-efficacy is associated same regarding a measure of “risk-seeking” with better labor market outcomes among orientation, which can be interpreted as a workers in South America. proxy for entrepreneurial spirit. 5. Although China’s CEES collected data only 8. Analysis of data from Indonesia, Mongolia, on manufacturing firms, the 2016 China the Philippines, and Vietnam further suggests Urban Labor Survey (CULS), collected in that returns to nonroutine analytical tasks six large Chinese cities, includes many of the tend to be increasing over time (Macdonald same individual-level variables as the CEES, 2018). which enables some comparison of how skills 9. These findings are from World Bank demands differ across the manufacturing and calculations based on Enterprise Survey data. services sectors. As with more-innovative 10. Similar analysis was carried out for non- manufacturing firms in China, employees in innovating firms. A significant share of non- high-skill services sectors tend to be relatively innovating firms also reported skills-related young and well educated. According to the challenges when hiring. In many but not all 2016 CULS data, 72 percent of workers in cases, the reported skills deficits were larger high-skill services sectors are college educated, for innovating than for non-innovating firms. compared with 42 percent in manufacturing Differences in reported skills deficits between and 26 percent in low-skill services (Park and innovating and non-innovating firms reflect Xuan 2020). In addition, just over one-third differences in skills demand across the firm of tertiary graduates in high-skill services types, which also appear to differ across firms have STEM degrees, comparable to the countries. share among high-innovation firms surveyed 11. For more information and the 2018 PISA in the CEES and slightly higher than the data, see the PISA website: https://www.oecd share in all manufacturing firms surveyed in .org/pisa/. the CULS. 12. Data on learning outcomes comparable to 6. The assessment of cognitive skills in the PISA are not available outside China’s major Vietnam ESIS provides a holistic perspective urban centers. Nevertheless, a recent study on literacy, capturing a person’s “ability to carried out jointly by China’s Development understand, evaluate, use and engage with Research Center and the World Bank written texts to participate in society, to highlights continuing learning challenges Skills and Finance for Innovation 139

faced by students in China’s rural areas on the role of financing received from family, (World Bank and DRC 2019). friends, and nonfinancial entities. 13. PISA data for 2018 are from the PISA 21. For a detailed explanation of how different website: https://www.oecd.org/pisa/. sources of finance address diverse market Although several lower-middle-income frictions in different manners, see Mare, de countries in the region have not participated Nicola, and Liriano (forthcoming) and the in the PISA exams, Early Grade Reading references therein. Assessments (EGRAs) in Cambodia, Lao 22. Abraham, Cortina, and Schmukler (2019) PDR, and Myanmar provide further document a growth in capital market evidence of the challenges that countries financing in East Asia Pacific since the early in the region face in building adequate 2000s. The authors note, however, that capital foundational skills among their populations. market access is still limited to a relatively Results from the 2012 EGRAs in Cambodia small share of large corporations. and Lao PDR showed, for example, that 23. Fixed assets represent long-term investments nearly one-third of second-grade students in of firms; such assets can proxy for those countries could not read a single word investment in innovation, because innovative (World Bank 2018). characteristics could be incorporated in new 14. Asymmetric information arises when the capital equipment, for example. parties involved in a transaction have 24. A key factor enabling risk finance for different information sets (for example, one innovation projects is the possibility that party has more accurate information than investors can liquidate their investments when the other). conditions are favorable. Such liquidation 15. Stronger disclosure requirements may is possible where capital markets are well be beneficial for investors. For example, developed. In such settings, private equity more stringent disclosure requirements on investors can liquidate their investment in an alternative investment funds (including innovative firm in one of several ways: (a) the venture capital and private equity firms) firm can be acquired by another firm; (b) it enable a higher supply of capital and business can be sold to another investor or bought by creation (Cumming and Knill 2012). those in the firm itself; (c) it can go public via 16. Moral hazard occurs after a contract is an initial public offering (IPO); or (d) in the signed and a borrower changes behavior—for worst-case scenario, if the firm defaults, the instance, by taking on more risk than a lender private equity investor can file for bankruptcy. would have envisioned at the time of signing 25. See the Hurun Research Institute’s Global the contract. Unicorn List 2019, which ranks the world’s 17. For a theoretical and quantitative illustration billion-dollar tech start-ups founded in the on how financial frictions affect the quality of 2000s and not yet listed on a public exchange: firm innovation, see Cole, Greenwood, and http://www.hurun.net. Sanchez (2016). For empirical evidence, see 26. For more information on these indexes, Ayyagari, Demirgüç-Kunt, and Maksimovic including more detailed explanations of (2011); Gorodnichenko and Schnitzer (2013); the variables, see the World Bank’s Doing and Cao (2020). Business website: https://www.doingbusiness 18. Financing and investment decisions are closely .org/en/doingbusiness. related, with the available mix of financing affecting a firm’s choice of the source of finance. References 19. Aghion, Howitt, and Mayer-Foulkes (2005) Abraham, Facundo, Juan J. Cortina, and Sergio formulate a theory where the “advantage L. Schmukler. 2019. “The Rise of Domestic of backwardness” argued by Gerschenkron Capital Markets for Corporate Financing.” (1952)—that is, the advantage to technologically Policy Research Working Paper 8844, World lagging countries of adopting frontier Bank, Washington, DC. technology developed elsewhere—is not fully Acemoglu, Daron, Philippe Aghion, and Fabrizio exploited because of financial constraints. Zilibotti. 2006. “Distance to Frontier, Selection, 20. The statement applies to formal sources of and Economic Growth.” Journal of the finance. There is limited theory and evidence European Economic Association 4 (1): 37–74. 140 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

Acemoglu, Daron, and Pascual Restrepo. 2017. Brown, James R., Steven M. Fazzari, and Bruce “Robots and Jobs: Evidence from US Labor C. Petersen. 2009. “Financing Innovation and Markets.” NBER Working Paper No. 23285, Growth: Cash Flow, External Equity, and the National Bureau of Economic Research, 1990s R&D Boom.” Journal of Finance 64 (1): Cambridge, MA. 151–85. ADB (Asian Development Bank). 2020. Asian Bughin, Jacques, Eric Hazan, Susan Lund, Peter Development Outlook 2020: What Drives Dahlström, Anna Wiesinger, and Amresh Innovation in Asia? Semiannual report, April Subramaniam. 2018. “Skill Shift: Automation 2020. Manila: ADB. and the Future of the Workforce.” Discussion Aghion, Philippe, Antonin Bergeaud, Richard paper, McKinsey Global Institute, McKinsey & Blundell, and Rachel Griffith. 2019. “The Co., New York. Innovation Premium to Soft Skills in Low- Cao, Yu. 2020. “Financial Constraints, Innovation Skilled Occupations.” Discussion Paper No. Quality, and Growth.” Doctoral dissertation, 1665, Centre for Economic Performance, University of Southern California, Los Angeles. London School of Economics. Chen, Su-Chang, Ming-Chung Wu, and Chun- Aghion, Philippe, Peter Howitt, and David Mayer- Hung Chen. 2010. “Employee’s Personality Foulkes. 2005. “The Effect of Financial Traits, Work Motivation and Innovative Development on Convergence: Theory and Behavior in Marine Tourism Industry.” Journal Evidence.” Quarterly Journal of Economics of Service Science and Management 3 (2): 120 (1): 173–222. 198–205. Autor, David H., David Dorn, and Gordon Cirera, Xavier, and William F. Maloney. 2017. H. Hanson. 2015. “Untangling Trade and The Innovation Paradox: Developing-Country Technology: Evidence from Local Labor Capabilities and the Unrealized Promise of Markets.” Economic Journal 125 (584): Technological Catch-Up. Washington, DC: 621–46. World Bank. Avvisati, Francesco, Gwenaël Jacotin, and Stéphan Cirera, Xavier, William F. Maloney, and Mauricio Vincent-Lancrin. 2013. “Educating Higher Sarrias. 2020. “Management Quality as an Education Students for Innovative Economies: Input for Innovation.” Unpublished, World What International Data Tell Us.” Tuning Bank, Washington, DC. Journal for Higher Education 1 (1): 223–40. Cirera, Xavier, and Leonard Sabetti. 2019. Ayyagari, Meghana, Asli Demirgüç-Kunt, and “The Effects of Innovation on Employment Vojislav Maksimovic. 2011. “Firm Innovation in Developing Countries: Evidence from in Emerging Markets: The Role of Finance, Enterprise Surveys.” Industrial and Corporate Governance, and Competition.” Journal of Change 28 (1): 1–16. Financial and Quantitative Analysis 46 (6): Cirera, Xavier, and Antonio Soares Martins- 1545–80. doi:10.1017/S0022109011000378. Neto. 2020. “Do Innovative Firms Pay Higher Bassi, Marina, Matías Busso, Sergio Urzúa, and Wages? Micro-Level Evidence from Brazil.” Jaime Vargas. 2012. Disconnected: Skills, Policy Research Working Paper 9442, World Education and Employment in Latin America. Bank, Washington, DC. Washington, DC: Inter-American Development Cole, Harold L., Jeremy Greenwood, and Juan Bank. M. Sanchez. 2016. “Why Doesn't Technology Beck, Thorsten, Ross Levine, and Norman Loayza. Flow from Rich to Poor Countries?” 2000. “Finance and the Sources of Growth.” Econometrica 84 (4): 1477–1521. Journal of Financial Economics 58 (1–2): Crespi, Gustavo, Ezequiel Tacsir, and Mariano 261–300. Pereira. 2019. “Effects of Innovation on Berger, Allen N., Phil Molyneux, and John O. S. Employment in Latin America.” Industrial and Wilson. 2020. “Banks and the Real Economy: Corporate Change 28 (1): 139–59. An Assessment of the Research.” Journal of Cumming, Douglas, and April Knill. 2012. Corporate Finance 62: 101513. “Disclosure, Venture Capital and Entrepreneurial Bodewig, Christian, Reena Badiani-Magnusson, Spawning.” Journal of International Business with Kevin Macdonald, David Newhouse, and Studies 43 (6): 563–90. Jan Rutkowski. 2014. Skilling Up Vietnam: Cusolito, Ana Paula, Daniel Lederman, and Jorge Preparing the Workforce for a Modern Market Peña. 2020. “The Effects of Digital-Technology Economy. Directions in Development Series. Adoption on Productivity and Factor Demand: Washington, DC: World Bank. Firm-Level Evidence from Developing Skills and Finance for Innovation 141

Countries.” Policy Research Working Paper Graetz, Georg, and Guy Michaels. 2018. “Robots 9333, World Bank, Washington, DC. at Work.” Review of Economics and Statistics Czarnitzki, Dirk, and Hanna Hottenrott. 2011. 100 (5): 753–68. “R&D Investment and Financing Constraints Green, Lawrence, Barbara Jones, and Ian Miles. of Small and Medium-Sized Firms.” Small 2007. “Mini Study 02 – Skills for Innovation.” Business Economics 36 (1): 65–83. Paper for the INNO-Grips (Global Review Das, Mitali, and Benjamin Hilgenstock. 2018. of Innovation Intelligence and Policy Studies) “The Exposure to Routinization: Labor INNO-Views Policy Workshop, “Skills for Market Implications for Developed and Innovation: Fostering Skills to Improve Developing Economies.” Working Paper European Innovation Performance,” Glasgow, No. 18/135, International Monetary Fund, September 27–28. Washington, DC. Gumbau-Albert, Mercedes, and Joaquin Maudos. Deming, David, and Lisa B. Khan. 2018. “Skill 2009. “Patents, Technological Inputs Requirements Across Firms and Labor and Spillovers among Regions.” Applied Markets: Evidence from Job Postings for Economics 41 (12): 1473–86. Professionals.” Journal of Labor Economics Hall, Bronwyn H., and Josh Lerner. 2010. “The 36 (S1): S337–S369. Financing of R&D and Innovation.” In de Nicola, Francesca. 2019. “Assessing How Handbook of the Economics of Innovation, Returns to Innovation Vary Depending vol. 2, edited by Bronwyn H. Hall and Nathan on the Business Environment and Firms’ Rosenberg, 609–39. Amsterdam: Elsevier. Characteristics.” Background paper for this Harrison, Rupert, Jordi Jaumandreu, Jacques study, World Bank, Washington, DC. Mairesse, and Bettina Peters. 2014. “Does Doran, Justin, and Geraldine Ryan. 2014. “Firms’ Innovation Stimulate Employment? A Firm- Skills as Drivers of Radical and Incremental Level Analysis Using Comparable Micro-Data Innovation.” Economics Letters 125 (1): 107–09. from Four European Countries.” International Du, Yang, and Albert Park. 2017. “Changing Journal of Industrial Organization 35: 29–43. Demand for Tasks and Skills in China.” Hottenrott, Hanna, and Bettina Peters. 2012. Unpublished manuscript, Chinese Academy “Innovative Capability and Financing of Social Science, Beijing, and Hong Kong Constraints for Innovation: More Money, University of Science and Technology. More Innovation?” Review of Economics and Dutz, Mark A., Rita K. Almeida, and Truman Statistics 94 (4): 1126–42. G. Packard. 2018. The Jobs of Tomorrow: Hsieh, Hsiow-Ling, Jia-Ru Hsieh, and I-Ling Wang. Technology, Productivity, and Prosperity in 2011. “Linking Personality and Innovation: Latin America and the Caribbean. Directions The Role of Knowledge Management.” World in Development Series. Washington, DC: Transactions on Engineering and Technology World Bank. Education 9 (1): 38–44. Engen, Marit, and Inger Elisabeth Holen. 2014. Hsu, Po-Hsuan, Xuan Tian, and Yan Xu. 2014. “Radical versus Incremental Innovation: The “Financial Development and Innovation: Importance of Key Competences in Service Cross-Country Evidence.” Journal of Financial Firms.” Technology Innovation Management Economics 112 (1): 116–35. Review 4 (4): 15–25. Kataoka, Sachiko, and Ana Alejo. 2019. “Skills for Gallié, Emilie-Pauline, and Diègo Legros. 2012. Innovation.” Background paper for this report, “Firms’ Human Capital, R&D and Innovation: World Bank, Manila, Philippines. A Study on French Firms.” Empirical Kato, Masatoshi, Hiroyuki Okamuro, and Yuji Economics 43 (2): 581–96. Honjo. 2015. “Does Founders’ Human Gerschenkron, Alexander. 1952. “Economic Capital Matter for Innovation? Evidence from Backwardness in Historical Perspective.” In Japanese Start-Ups.” Journal of Small Business The Progress of Underdeveloped Areas, edited Management 53 (1): 114–28. by Bert F. Hoselitz. Chicago: University of Lachenmaier, Stefan, and Horst Rottmann. 2011. Chicago Press. “Effects of Innovation on Employment: Gorodnichenko, Yuriy, and Monika Schnitzer. A Dynamic Panel Analysis.” International Journal 2013. “Financial Constraints and Innovation: of Industrial Organization 29 (2): 210–20. Why Poor Countries Don’t Catch Up.” Journal Leiponen, Ada. 2000. “Competencies, Innovation of the European Economic Association 11 (5): and Profitability of Firms.” Economics of 1115–52. Innovation and New Technology 9 (1): 1–24. 142 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

Levine, Ross. 2005. “Finance and Growth: Theory OECD (Organisation for Economic Co-operation and Evidence.” In Handbook of Economic and Development). 2011. Skills for Innovation Growth, vol. 1A, edited by Philippe Aghion and Research. Paris: OECD Publishing. and Steven N. Durlauf, 865–934. Amsterdam: OECD (Organisation for Economic Co-operation North-Holland Elsevier. and Development). 2013. OECD Skills Lopez, Elyssa. 2019. “No Country for Unicorns: Outlook 2013: First Results from the Survey of Why Philippine Start-Ups Are Struggling.” Adult Skills. Paris: OECD Publishing. South China Morning Post, September 16. OECD (Organisation for Economic Co-operation https://www.scmp.com/week-asia/economics and Development). 2016. Innovating Education /article/3027171/no-country-unicorns-why and Educating for Innovation: The Power of -philippine-start-ups-are-struggling. Digital Technologies and Skills. Paris: OECD Macdonald, Kevin. 2018. “Task Composition Publishing. doi:10.1787/9789264265097. Globally and their Returns in Select East Asian OECD (Organisation for Economic Co-operation Countries.” Background paper for A Resurgent and Development). 2017. “Social and East Asia: Navigating a Changing World. Emotional Skills: Well-Being, Connectedness Washington, DC: World Bank. and Success.” Report on the Study on Social Macdonald. 2019. “How Do Innovating and Emotional Skills, OECD, Paris. Employers’ Value and Use of Skills Differ OECD and Eurostat (Organisation for Economic from Non-Innovating Employers in Vietnam?” Co-operation and Development and the Background paper for this study, World Bank, statistical office of the European Union). 2005. Washington, DC. Oslo Manual: Guidelines for Collecting and Maloney, William F., and Carlos Molina. 2016. Interpreting Innovation. 3rd ed. Paris: OECD; “Are Automation and Trade Polarizing Luxembourg: Eurostat. Developing Country Labor Markets, Too?” Park, Albert, and Wenshi Xuan. 2020. “Skills for Policy Research Working Paper 7922, World Innovation in China.” Background paper for Bank, Washington, DC. this report, Hong Kong University of Science Maloney, William F., and Mauricio Sarrias. 2017. and Technology, Hong Kong SAR, China. “Convergence to the Managerial Frontier.” Piva, Mariacristina, and Marco Vivarelli. 2005. Journal of Economic Behavior & Organization “Innovation and Employment: Evidence from 134: 284–306. Italian Microdata.” Journal of Economics Mare, Davide S., Francesca de Nicola, and Faruk 86 (1): 65–83. Miguel Liriano. Forthcoming. “Financial Poole, Jennifer P., Amelia U. Santos-Paulino, Structure and Firm Innovation: Cross-Country Maria V. Sokolova, and Alisa DiCaprio. 2017. Evidence.” Policy Research Working Paper, “The Impact of Trade and Technology on Skills World Bank, Washington, DC. in Viet Nam.” Working Paper No. 770, Asian Mason, Andrew D., Vera Kehayova, and Judy Development Bank Institute, Tokyo. Yang. 2018. “Trade, Technology, Skills Scheela, William, Edmundo Isidro, Thawatchai and Jobs: Exploring the Road Ahead for Jittrapanun, and Nguyen Thi Thu Trang. Developing East Asia.” Background paper for 2015. “Formal and Informal Venture Capital A Resurgent East Asia: Navigating a Changing Investing in Emerging Economies in Southeast World. Washington, DC: World Bank. Asia.” Asia Pacific Journal of Management Mason, Andrew D., and Sudhir Shetty. 2019. 32 (3): 597–617. A Resurgent East Asia: Navigating a Changing Toner, Phillip. 2011. “Workforce Skills and World. East Asia and Pacific Regional Report. Innovation: An Overview of Major Themes Washington, DC: World Bank. in the Literature.” Science, Technology and Miyamoto, Koji, and Miguel Sarzosa. 2020. Industry (STI) Working Paper, Organisation “Workforce Skills and Firm Innovation: for Economic Co-operation and Development, Evidence from an Employer-Employee Linked Paris. Survey Data in Vietnam.” Unpublished Vivekarajah, Sivapalan. 2018. “What Ails Venture manuscript, World Bank, Washington, D.C. Capital in Malaysia and Why Consolidation Is Nor, Eliza. 2015. “Venture Capitalists in Malaysia: Not the Answer.” Digital News Asia, October 9. Challenges and Future Directions.” Journal https://www.digitalnewsasia.com/insights of Business and Management Sciences /what-ails-venture-capital-malaysia-and-why 3 (4):124–29. -consolidation-not-answer. Skills and Finance for Innovation 143

World Bank. 2012. Putting Higher Education World Bank. 2020. “PISA 2018: East Asia to Work: Skills and Research for Growth and Pacific Regional Brief.” Unpublished in East Asia. East Asia and Pacific Regional manuscript, Education Global Practice, World Report. Washington, DC: World Bank. Bank, Washington, DC. World Bank. 2016. World Development Report World Bank and DRC (Development Research 2016: Digital Dividends. Washington, DC: Center of the State Council, The People’s World Bank. Republic of China). 2019. Innovative China: World Bank. 2018. Growing Smarter: Learning New Drivers of Growth. Washington, DC: and Equitable Development in East Asia and World Bank. Pacific. East Asia and Pacific Regional Report. Washington, DC: World Bank.

Innovation Policies and Institutions in the Region: 5 An Assessment

Introduction forging effective linkages between different This chapter assesses the adequacy of policies actors in the innovation system. This requires and institutions in addressing the constraints adequate capacity on the part of public agen- firms face regarding diffusion, technology cies to implement such policies—capacity that adoption, and invention in developing East is often limited due to scarce human and finan- Asia. It pays special attention to the effective- cial resources. ness of knowledge creation and its transfer to Policy makers should follow several steps firms as well as the adequacy of countries’ insti- when thinking about innovation and technol- tutions and policies in facilitating this trans- ogy policies. First, the policy mix should effec- fer and generating innovation. The chapter tively address the constraints on the private first identifies the gaps in the policy mix that sector’s ability to build innovation capabili- support innovation in the region and highlights ties. Second, it is essential to implement robust the key principles to improve the efficiency and policy-making processes, using good practices effectiveness of implementing agencies. It then in public management. Just as for firms, good evaluates the public research institutions that quality management practices also matter for aim to facilitate innovation and knowledge public policy (Rasul and Rogger 2017). Third, transfer. innovation agencies should have the incen- As described in chapter 2, innovation policy tives and governance structures to perform must use multiple policy instruments simul- good diagnostics of the main market failures taneously to address the market failures that that constrain innovation and diffusion, as discourage innovation (Cirera et al. 2020). well as the resources to deliver effective poli- This introduces some complexity in its design cies. And policy makers need to understand because of complementarities that exist across not only the market failures but also the risks policy instruments. It also raises the premium of government failure when innovation poli- associated with identifying and prioritiz- cies are not well designed and implemented, ing those policies that are most effective in or when institutional fragmentation and lack enabling innovation and facilitating diffusion. of coordination result in undesired program For firms to reach the technological frontier, overlaps, inefficiencies, or mistargeting of policies and institutions must focus on build- beneficiaries. This chapter examines each of ing the capabilities of the private sector and these elements. 146 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

The central actors of the national innova- Therefore, a key question is whether the tion system (NIS) include institutions such combination of policies supporting innova- as universities and public research organization is adequate to building the basic innovations (PROs), whose main role is to generate tion capabilities that most firms require. Or, knowledge that can become an innovation rather, are policies biased toward higher-end when brought to the market or implemented R&D projects that benefit a much narrower in society. In more mature innovation sys- set of enterprises? As discussed in chapters 2 tems, the flow or transfer of knowledge from and 3, given innovation capabilities, policies these knowledge-creating institutions to firms and public resources in developing East Asia is fluid. However, in many low- and middle- should focus more on diffusion and adoption income economies, the diffusion of knowl- than on invention. edge to firms is often limited by the lack of adequate policies to support commercial Importance of good practices in policy research as well as the lack of well-functioning design and implementation industry-research linkages, resulting in low innovative activity and little technology Another key issue for innovation policy, transfer. Often, these knowledge institutions in addition to policy mix, is the quality of neither facilitate diffusion nor perform their policy design and implementation. Cirera mission of finding the technological solutions and Maloney (2017) describe the need for needed to address key societal challenges, innovation policies to respond to clearly such as COVID-19 or climate change. identified innovation problems and market Understanding the quality and strengths failures. Doing so requires the adoption of of these knowledge institutions, along with good practices in public management, moni- their governance structures, is critical to toring, and evaluation. When governments maximizing their contributions to innovation. lack the capacities to design and implement The chapter discusses these issues using the sound innovation policies, the risks of policy results of new surveys implemented to assess capture, inefficiency, and ineffectiveness esca- the quality of these institutions in three coun- late—potentially creating policy distortions tries in developing East Asia: Malaysia, the that can undermine innovation. As a result, it Philippines, and Vietnam (Cirera, Kuriakose, is important that countries in developing East and Zuñiga 2021). Asia invest in the adoption of these good practices to maximize the impact of their policies. Are policies coherent with the This section discusses the policy mixes used objective of supporting diffusion to support innovation and technology adop- as well as invention? tion in three countries in the region based on novel data collected by World Bank policy A key characteristic of innovation policies in effectiveness reviews (PERs) in Indonesia, many low- and middle-income countries is the Philippines, and Vietnam. PERs collect fragmentation, with little coordination, result- data at the national level on policy instru- ing in unnecessary overlaps in interventions. ments, including information on objectives, Also, paradoxically, the use of public resources beneficiary types, and budgets.1 These data is often determined by research and develop- enable a granular characterization of coun- ment (R&D) considerations, focusing on tries’ policy mixes, and although the ultimate instruments that do not necessarily support effectiveness of individual policy instruments more basic innovation projects (such as qual- can only be determined via rigorous impact ity upgrading or the adoption of new technolo- evaluations, these reviews facilitate a detailed gies) that would benefit most firms. Rather, qualitative assessment of the extent to which large amounts of resources are focused on sup- resources are being directed toward areas porting R&D projects carried out by PROs or where previous analysis suggests they can through tax incentives benefiting firms already have the greatest impact. (See box 5.1 for a doing R&D. detailed description of the methodology.) Innovation Policies and Institutions in the Region: An Assessment 147

BOX 5.1 Policy effectiveness reviews in developing East Asia

The World Bank has implemented policy effectiveness Functional review reviews (PERs) on science, technology, and innovation A functional review is the second part of a PER, (STI) in several countries in Latin America (Argentina, which assesses specific policy instruments for the Brazil, Chile, and Colombia) and developing East Asia use of good practices in design, implementation, (Indonesia, the Philippines, and Vietnam, and currently and coordination. The functional review provides in Malaysia). The goals of the PERs are (a) to assess the analytical foundation to make recommendations whether the government’s allocation of resources has to strengthen the design, delivery, and effectiveness an impact on STI outcomes, and (b) to support policy of policy instruments under implementation as well makers in increasing their capacity for, and the quality as to inform the design of new instruments. Based of, their STI policies. on models of good public management, the analysis PERs include an analysis of the quality of the assesses whether good practices are implemented in policy mix and a functional review (assessment of key elements of policy making, such as the following: the quality of design, implementation, and gov- • The market failure is identified, and it justifies ernance of the different innovation policy instru- design and intervention. ments). Specifically, they answer the following • The policy instrument is originated by a sound questions: identification of a problem, and it is not ad hoc. • How much is being spent on STI policy? For • The policy instrument features clearly identified what? And by whom? objectives that are measurable. • How coherent is the policy mix with the country’s • The policy instrument has an explicit and realistic greatest innovation challenges? logical framework. • How well does this spending align with the coun- • The selection of beneficiaries is appropriate. try’s strategic aspirations? • The policy instrument builds upon sufficient • What are the opportunities to improve the human, financial, and organizational resources, capacity and coordination of implementing and it features good managerial practices. agencies? • The choice of instrument is evidence based, and consideration of costs and alternatives is well Quality of the policy mix documented. The analysis of the policy mix relies on examin- • The policy instrument employs monitoring and ing the patterns of public spending on STI and evaluation (M&E) frameworks, with sound indi- the way resources get allocated. This component cators and good measurement. Consideration of provides the basis for analysis of policy coherence impact evaluation is well documented. and consistency. PERs focus on the policy “instru- • There is a formal system to adopt lessons and ment” as the unit of analysis. An instrument is a learning to make the policy instrument more effi- mechanism or intervention by which the govern- cient over time. ment uses public expenditures, laws, or regulations to achieve specific objectives. Examples of The methodology of the functional review involves an innovation policy instrument include credit using semistructured interviews for a sample of guarantees for innovation projects or tax incen- policy instruments previously identified to evaluate tives for R&D. the quality of design, implementation, coordination PERs start with a policy mapping exercise, among instruments and among institutions, and collecting data to map the portfolio of innovation- their relation to international best practices. A scor- supporting policy instruments. This provides the ing matrix assigns values to instruments from 1 to 5 basis for generating descriptive analytics and for based on best practices, with 1 being the lowest score profiling the portfolio of interventions that provides and 5 the highest. Scores are based on a previously support to firms for innovation. developed matrix of good practices. 148 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

Assessing the innovation policy mix 146, 70, and 127 policy instruments, respectively, with corresponding total expenditures A first step in analyzing the quality of the of approximately US$1.12 billion, US$455 different policy mixes in Indonesia, the million, and US$1.16 billion (table 5.1). Philippines, and Vietnam involves mapping The relatively small number of instruments all of the relevant policy instruments (as and total expenditures for the Philippines described in box 5.1). This requires combin- is partly due to the government’s interest in ing data from secondary sources, including instruments that support small and medium budget reports and specialized documents enterprises (SMEs), which may drive some of from key government agencies. Expenditure the differences with the other two countries data used in the mapping exercise include both discussed below. Vietnam’s median expen- budget expenses and forgone revenue from diture on policy instruments is significantly tax incentives. The data on public spending lower than in the other two countries, but underestimate total expenditure on innova- its total portfolio is the largest because of tion, as they exclude block funding to PROs the high value of its top two tax incentive and other institutions. The data also exclude programs. regulatory instruments that support innovation Analyzing the policy mixes in Indonesia, but do not receive budgetary funding, such as the Philippines, and Vietnam reveals a lack regulatory measures for equity investments for of coherence in how resources are allocated, start-ups. especially relative to the objectives out- The scope of the analysis includes policy lined in the countries’ innovation strategies. instruments that use public expenditure and These strategies emphasize outcomes such as whose resources are usually competitively allo- greater innovation activity and faster pro- cated—covering instruments that extend credit ductivity and employment growth. However, to firms; offer tax incentives to firms that the resources to support innovation are invest in innovation; and deliver direct support highly concentrated in a few programs in the form of grants, technical assistance, and that largely support research, as follows other services. (figure 5.1):2 Innovation policies in developing East • Vietnam’s strategy emphasizes innova- Asia tend to be fragmented, with budget tion activity and productivity growth, but allocations concentrated on a few resources are highly concentrated in pro- instruments, some of which focus on large grams dedicated to supporting R&D-based R&D projects innovation in firms via foreign direct invest- The policy portfolio analyzed in Indonesia, ment (FDI) spillovers. The 10 largest policy the Philippines, and Vietnam includes instruments in terms of spending account

TABLE 5.1 Policy effectiveness reviews in Indonesia, the Philippines, and Vietnam assess the portfolio of innovation policy instruments and related spending

Policy instrument level of expenditure (US$, thousands) Policy instruments Total portfolio Country (number) (US$, millions) Minimum Average Median Maximum Indonesia 146 1,129 12 7,735 289 959,476 Indonesia (excluding top 145 170 12 1,172 287 52,473 policy instrument) Philippines 70 456 8 6,507 788 72,257 Vietnam 127 1,155 1 9,097 52 939,605 Vietnam (excluding top two 125 71 1 567 52 11,384 policy instruments)

Source: World Bank elaboration based on country-level policy effectiveness reviews. Note: Expenditure data are from 2018 for Indonesia and from 2017 for the Philippines and Vietnam. Innovation Policies and Institutions in the Region: An Assessment 149

FIGURE 5.1 Despite numerous innovation policy instruments in Indonesia, the Philippines, and Vietnam, public spending concentrates on only a few

Indonesia Vietnam Indonesia Philippines Vietnam (excluding top program) (excluding top two programs)

3.1 27.0 4.8 89.0 23.6

0.8 6.7 1.7 7.3 16.6

4.0 35.4 20.0 0.3 4.1

0.6 5.1 5.8 0.9 14.2

0.6 5.7 6.2 0.5 7.6

88.8 2.0 6.1 0.0 0.4

0.3 2.9 1.2 0.1 0.9

1.0 8.5 36.9 1.1 17.5

0.6 5.3 4.4 0.5 7.4

0.1 0.7 1.2 0.2 4.0

0.1 1.0 11.6 0.2 3.7

0 20 40 60 80 100 0 20 40 60 80 100 0 20 40 60 80 100 0 20 40 60 80 100 0 20 40 60 80 100 Share of innovation spending (%), by policy instrument objective

Market access (domestic) Improving business environment Entrepreneurship Skills formation Export promotion Access to nance Management pratices Tech transfer, science-industry collaboration Research excellence Non-R&D innovation, tech adoption/diusion R&D, R&D-based innovation

Source: World Bank elaboration based on country-level policy effectiveness reviews. Note: Spending percentages refer to the allocation of resources either directly (via subsidies or technical assistance) or indirectly (via forgone revenue to beneficiaries that can be firms, research organizations, or entrepreneurs). The percentages exclude direct funding to public research organizations (PROs) and universities. R&D = research and development.

for 97 percent of the total portfolio; and if almost six times that of the rest of the pol- two large policy instruments that provide icy instruments combined) and research tax exemptions for high-tech enterprises are excellence. excluded, R&D still accounts for 24 percent of total spending and the largest single share Importantly, few resources are allocated of all instruments. in any of the three countries to technology • The Philippines’ strategy emphasizes the diffusion, adoption, or non-R&D-related growth of micro, small, and medium enter- efforts to build firms’ capabilities (figure 5.1). prises (MSMEs) and employment, but a Similarly, relatively little funding is allocated large share of policy resources is devoted toward technology transfer, entrepreneur- to skills formation and research excellence. ship or improving market access, especially • Indonesia’s strategy emphasizes a mix of for export promotion. As discussed in chap- the two other countries’ objectives, but the ter 3, most firms in these countries are a sig- top 10 policy instruments, which account nificant distance from the technology frontier for 93 percent of the total portfolio, are and could most efficiently increase innovation largely devoted to an interest subsidy by adopting existing technology from more for MSMEs (whose total expenditure is advanced countries. 150 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

FIGURE 5.2 The distribution of public resources for innovation, by The three selected countries target innovative firm type, varies considerably across Indonesia, the Philippines, and firms and beneficiaries very differently Vietnam The distribution of resources by firm size and innovation potential is rather different across a. By rm life-cycle stage countries (figure 5.2). Vietnam’s portfolio has a distinctively heavier focus on large firms 100 as well as R&D-intensive and technology- 30.0 31.5 80 48.3 intensive firms, driven partly but not entirely by the top two tax incentives policy instru- 60 36.3 30.0 ments (figure 5.2, panels b and c). In addition, 40 resources are highly concentrated toward 48.9 firms at the scale-up and mature stages rather Share of rms (%) Share 20 35.6 31.3 than at the start-up stage (figure 5.2, panel a). 2.4 2.9 0.5 2.3 0 Innovation policy instruments in the Indonesia Philippines Vietnam Philippines and Indonesia appear to be more Ideas/concept stage Young/start-ups Scale-up Mature inclusive, targeting different types and sizes of firms more evenly. These two countries spend more resources on MSMEs (figure 5.2, panel b. By rm size b) and on incentivizing the less-sophisticated 100 firms (“potential innovators”) to start inno- 0.6 vating (figure 5.2, panel c). 80 33.1 33.0

60 19.0 The countries also exhibit a strong policy 34.9 95.3 40 bias against supporting innovation in 20.1 services Share of rms (%) Share 20 One interesting element is the bias against 31.4 27.0 1.8 1.7 0 1.1 firm beneficiaries in the services sectors. Indonesia Philippines Vietnam Figure 5.3 shows the allocation of resources Micro Small Medium Large between policy instruments that target ex ante (as objectives of the program) firms in c. By rm’s innovation propensity services sectors and those that do not include firms in services as potential beneficiaries. 100 18.0 10.0 The results are striking, showing a large 8.6 35.4 bias of innovation policies toward manu- 80 8.5 18.6 facturing firms. This strong bias ignores the 60 18.3 innovation activities among firms in the ser- 46.1 34.1 40 vices sector and that innovation in services 26.0 (including “servicification,” as discussed in Share of rms (%) Share 20 26.8 27.7 chapter 2) is an increasingly important part 19.2 0.4 2.4 0 of innovation and adoption of technologies Indonesia Philippines Vietnam in manufacturing. Non-innovator Potential innovator HG potential R&D intensive Technology intensive Analyzing innovation policy coherence

Source: World Bank elaboration based on country-level policy effectiveness reviews. The coherence of a particular policy mix Note: Graphs show the percentage of innovation policy instruments that target each category in the can be analyzed along two dimensions: three countries. HG = high growth; R&D = research and development. internal and external coherence. Internal coherence refers to avoiding overlaps and ensuring that the scale of policy instruments is sufficient to achieve the expected results. Innovation Policies and Institutions in the Region: An Assessment 151

External coherence refers to the extent to FIGURE 5.3 Indonesia, the Philippines, and Vietnam allocate few, which the main policy instruments aim to if any, public resources to innovation policy instruments whose address the key challenges and market fail- potential beneficiaries include firms in services sectors ures faced by the private sector with a view to minimizing gaps in key areas of support. 150,000 Assessing the coherence of existing policy instruments provides important information about the alignment (or lack thereof) of exist- 100,000 ing policies, highlighting some potential risks to their effectiveness. 50,000 Fragmentation of policy instruments often Total budget (US$, thousands) Total results in the lack of scale for large impact 0 One problem related to excessive fragmenta- No services Services No services Services No services Services tion is the lack of meaningful scale of policy targeted targeted targeted targeted targeted targeted instruments to achieve significant impact. Indonesia Philippines Vietnam For example, a policy instrument in one of the countries analyzed aimed to increase Source: World Bank elaboration based on country-level policy effectiveness reviews. Note: The budget allocations shown include those made directly (through grants, subsidies, productivity in a sector by 6 percent, but it or technical assistance) or indirectly (through tax incentives) for innovation policy instruments was supporting only small innovative proj- targeting two beneficiary groups: those instruments that exclude services firms as potential beneficiaries (“No services targeted”) and those that include services firms (“Services targeted”). ects in seven firms in that sector. There was clearly a mismatch between the objective There are important overlaps in policy and the instrument: because the instrument instruments across agencies in the region supported less than 1 percent of all firms in Duplication of innovation instruments across the sector, extremely large productivity gains agencies can lead to further inefficiencies would have been needed to diffuse across in countries’ national innovation systems. the majority of supported firms to have a Table 5.2 shows the incidence of innovation measurable impact on sector productivity. In policy instruments that overlap in scope in addition, resources allocated to the instru- each country—with respect to objectives, tar- ment were too small in scale to achieve any get beneficiaries, and support mechanisms. meaningful impact. In all countries, a nonnegligible share of Overall, policy instruments are relatively policy instruments have an identical scope larger in the Philippines and thus present less with at least one other policy instrument. The of a scale problem. Nevertheless, the lack of incidence of overlap appears particularly high scale appears to be an issue for at least some in Indonesia, with 80 such policy instruments policy instruments in all three countries. In the (more than half of the current portfolio) hav- Philippines, there are 10 policy instruments ing similar objectives. In one instance, 14 pol- (around 14 percent of instruments) and in icy instruments run by a single agency pursue Indonesia, 27 policy instruments (about 19 the same objective of improving research qual- percent), with annual budgetary resources of ity through provision of research infrastructure less than US$100,000—an amount that could to researchers and other government agencies. support very few beneficiaries (figure 5.4). In Vietnam, there are 70 such policy instru- Assessing the quality of policy design ments, representing more than half of all policy and implementation instruments in the portfolio. At face value, these policy instruments operate at a very For innovation policies to be effective, the low scale, raising questions about whether quality of policy design and implementation they have a viable minimum scale and, if not, is critical. Although empirical evidence on the whether they should either be merged with impact of policy design and implementation is similar programs or discontinued. scarce, the literature is starting to evaluate how 152 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

FIGURE 5.4 In all three studied countries, numerous innovation policy instruments have expenditures below US$100,000, indicating a potential lack of scale for significant impact

100

20 Instrument’s budget (current US$, thousands)

0 1 10 27 70 Policy instruments with annual budgets below US$100,000

Indonesia Philippines Vietnam

Source: World Bank elaboration based on country-level policy effectiveness reviews. Note: The figure ranks each country’s innovation policy instruments with budgets below US$100,000, from lowest to highest budget.

TABLE 5.2 In all three countries, a significant share of innovation policy instruments have overlapping scope

Policy instruments for which at least one other Share of policy instruments for which at least one other Country instrument has an identical objective (number) instrument has an identical objective (%) Indonesia 80 55 Philippines 13 19 Vietnam 43 34

Source: World Bank elaboration based on country-level policy effectiveness reviews. Note: Scope is defined as the combination of policy instrument objectives, supporting mechanisms, and target beneficiaries.

robust public management processes affect This functional review examined key fea- policy outcomes. In the case of Nigeria, for tures of functionality for a sample of policy example, good project management practices instruments (15 instruments in the Philippines affect the quality of the outcomes achieved and 13 in Vietnam) based on the quality of in infrastructure projects (Rasul and Rogger design and implementation processes as well 2017). However, to date, discussions of inno- as complementarities between instruments vation and technology policies have commonly within and across government institutions. ignored these critical issues. This section sum- Scores were assigned to 31 key processes marizes the results from an evaluation of the related to policy instrument design, imple- design, implementation, and interinstitutional mentation, and governance or coordination. coordination for selected policy instruments in Scores range from 1 (poor practice) to 5 the Philippines and Vietnam. (best practice). Innovation Policies and Institutions in the Region: An Assessment 153

In design, implementation, and governance, of innovation support policy instruments is innovation policies in the region are far still a significant distance from the frontier. from the public management frontier The analysis reveals major weaknesses in The analysis reveals ample scope to design and implementation that are prevalent improve policy design and implementation across policy instruments in both countries. The most important design shortcomings Figure 5.5 shows each country’s average score relate to the lack of adequate economic jus- for each metric under the three dimensions— tification, the absence of a logical framework design, implementation, and governance or used to develop the intervention, and a lack coordination. The average scores for the two of M&E mechanisms for most policy instru- countries, across agencies and across dimen- ments. Often, the origin of policy instru- sions of public policy, are 3.0 and 2.8 for the ments is not based on well-identified market Philippines and Vietnam, respectively, imply- failures, and the process of introducing new ing that there is room for substantial improve- instruments is in some cases ad hoc. The ment in practices and that the implementation lack of economic justification for the policy

FIGURE 5.5 Innovation policy instruments in the Philippines and Vietnam remain far from the design, implementation, and governance frontiers

a. Design frontier b. Governance frontier

Alignment with Origin other programs 5 Implementation Justi cation 5 4 4 Evaluation 3 Alignment 3 2 Selection 2 1 Objectives 1 0 Outside Alignment with 0 Impact Logframe coordination other institutions

Instrument Inputs Stakeholders Activities Bene ciaries Outputs Jurisdiction

c. Implementation frontier

Products 5 Process monitoring Eligibility 4 3 M&E 2 Time frames 1 Sta incentives 0 Databases

Training Closure

Autonomy Resources Management

Vietnam Philippines

Source: World Bank elaboration based on country-level policy effectiveness reviews. Note: The radar diagrams show the average score for each dimension between 1 (poor practice) and 5 (best practice). The sample includes 15 and 13 policy instruments in the Philippines and Vietnam, respectively. M&E = monitoring and evaluation. 154 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

instruments may result in design problems innovation capabilities and the diffusion of (poor quality of entry) and, more importantly, technology—objectives that are especially in inefficiencies in the choice of instruments important to countries that are placed low to address an identified problem. on the capabilities escalator (see chapter 2, In general, the design of most policy figure 2.8). Countries in the region need instruments is also not based on a logical to reduce this fragmentation and existing framework. The objectives tend to be unre- overlaps by better clarifying the objectives alistic—with no clear theory of change from of innovation policy and developing stron- inputs to activities to outputs to impacts— ger coordinating mechanisms. and without due consideration of external Moreover, it is critical for these countries conditions that affect impact. to adopt good practices in public manage- As for implementation, M&E frame- ment. This will require better mechanisms to works are generally absent, and lax report- recruit qualified staff as well as training to ing requirements with discretionary formats strengthen staff knowledge and capacity. prevent systematic collection of information that could provide real-time feedback to Agencies to support innovation improve policy instruments and decisions on future resource allocation. Moreover, As discussed in the previous section, inno- data on applications for support and pro- vation policies are not implemented in a gram beneficiaries are fragmented and rarely vacuum. They are implemented by agencies shared within and across institutions, even and line ministries that establish the priori- within government agencies. There is also ties, set up the processes, and execute the a lack of clarity on how beneficiaries are different policy programs. Evaluating the selected in practice, which can make the cap- adequacy of policies to promote innova- ture of support by a small number of firms tion and the diffusion of technologies thus more likely. requires understanding how these agencies Beyond the issues of design and imple- work. This section briefly discusses some mentation of specific policy instruments, the important features required by such agencies overall quality of the policy mix is hampered to implement effective innovation policies in by a lack of coordination across the relevant the region. government agencies. Governments have taken various approaches to creating innovation agencies; no one ideal type of agency is appro- Summary of findings on coherence of priate in all circumstances. High-income innovation policy instruments countries have different models of innova- The mix of policies reviewed in developing tion agencies, generally related to the dif- East Asian countries are not well oriented ferent types of innovation being promoted toward building firms’ capabilities for inno- (see annex 5A). For example, agencies that vation or accelerating technology diffusion focus on upgrading differ from agencies and adoption. Assessment of the policy mix that aim at more sophisticated innovation in Indonesia, the Philippines, and Vietnam and discovery. highlights several shortcomings that under- Although agency models from high-income mine the effectiveness of innovation policy. countries can be instructive, it is important The most important bottlenecks relate that institutional structures and practices be to (a) significant fragmentation of policy adapted to the country contexts in developing instruments, with many lacking minimum East Asia. Trying to mimic the design of high- viable scale; (b) unnecessary overlaps income countries’ agencies would likely result across instruments in some countries; in what Andrews, Pritchett, and Woolcock and (c) an allocation of resources that (2012) call “isomorphic mimicry”—that is, is inadequate to support the building of when copying other countries’ successful Innovation Policies and Institutions in the Region: An Assessment 155

designs and practices may not address the There are different possible models for local context and hence result in a failure innovation agencies, but existence of an of functionality. Determining which kind of implementing agency is not sufficient to innovation agency is best for a given context address institutional and coordination is difficult given a scarcity of relevant data, weaknesses but recent experience in the region provides Innovation agencies operate within a given some guidance. NIS, and their role is mostly focused on the execution of innovation policy instruments. They do not play a significant role in formu- Some innovation agencies in the region lating innovation policies, because national have overlapping mandates and are strategies are usually designed by ministries. not aligned with the country’s key Some good practices in the region can be innovation challenges found in Malaysia and Singapore, where the Although most developing East Asian coun- Agensi Inovasi Malaysia (AIM, the Malaysian tries have acknowledged the importance of Innovation Agency) and the National innovation and have taken laudable steps to Research Foundation (NRF), respectively, are improve their NISs and establish a related under the authority of the prime minister’s innovation agency, most efforts to create a office to coordinate innovation policy across dedicated agency charged with promoting ministries and agencies. innovation are rather nascent (OECD 2013). Having a dedicated innovation agency, In general, the sophistication and efficiency however, will not fix the inherent weaknesses of NISs in developing East Asian coun- in the innovation policy framework in devel- tries, while uneven, broadly correlate with oping East Asian countries. The success of the countries’ levels of economic development. Malaysian and Singaporean agencies relies on Nonetheless, Indonesia, Malaysia, Thailand, their professional management and sufficient and Vietnam all have innovation institu- convening power to coordinate innovation- tions that appear fairly ad hoc in nature, related initiatives across the range of relevant characterized by multiple layers of decision actors. Without a strong mandate and com- and policy making and, often, redundant or petent, professional management, the estab- overlapping interventions. lishment of agencies with similar functions in A dedicated innovation agency should other East Asian countries could potentially design and execute instruments that respond result in greater fragmentation of innovation to the country’s economic profile and inno- policy. vation needs. With the exception perhaps of Innovation agencies ought to help bring China, developing East Asian countries are about greater industry orientation to sci- not at the innovation frontier and are still net ence, technology, and human resources in importers of technology. FDI is also critical for the country. Furthermore, where possible, these countries not only in terms of capital but they should steer public research to align also for knowledge inflows. Innovation agen- more closely with industry needs. Although cies in the region should thus focus on instru- countries often aspire toward more radical, ments that would bring the following results: cutting-edge types of innovation (as the next • Greater commercialization of public section will discuss), given weak public R&D research systems and PROs, many countries should • Greater innovation by the private sector prioritize more incremental improvements— • Better links between SMEs and multina- including through greater diffusion and tech- tional enterprises (MNEs) nology adoption. Among the categories of • Higher FDI and associated knowledge innovation agencies described in annex 5A, transfer the “directed upgraders” and “productivity • Stronger entrepreneurial culture with facilitators” are perhaps most apt for devel- greater focus on business models. oping East Asian countries. These agency 156 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

models focus primarily on supporting small- private sectors (Cirera, Kuriakose, and Zuñiga scale, incremental product and process inno- 2021). Today, Malaysia has about as many vations. Such an approach would help steer researchers as Korea (more than 60,000), public sector research toward industry needs . although the number of researchers per person Of the various governance models for an in the labor force in Malaysia is still lower than innovation agency, a government agency with the average in high-income countries. high autonomy is perhaps the most appro- A common trend in national policy agen- priate and feasible in the region. Although a das has been the increasing recognition of the public-private partnership might be ideal, such importance of STI policies as potential engines an approach does not appear feasible given the of growth and development. Differences in political economy of many developing East levels of income and institutional develop- Asian countries. More importantly, by stay- ment across the region result in different levels ing within the government ambit, the agency of R&D spending relative to gross domestic would have a greater chance of informing sci- product (GDP), sometimes referred to as R&D ence, technology, and innovation (STI) policies. intensity. China and Singapore spend about 2 percent of GDP on R&D, for example, while Assessing public research Malaysia spends 1.4 percent and Indonesia, institution performance in 0.2 percent. Countries with the lowest rates facilitating innovation and of R&D intensity also report the highest par- technology transfer ticipation of the public sector in the funding of R&D (figure 5.6). Knowledge-creating public institutions— As countries’ levels of development rise, the namely public research organizations (PROs) share of R&D financed by the private busi- and universities—are critical actors in the ness sector also increases, which, combined NIS, central to innovation and technology with high rates of public investment, pulls total adoption. This section first discusses research R&D intensity upward. Thus, China, Korea, performance in the region. It then describes Malaysia, Singapore, and Thailand have high the results of a survey of PROs and university rates of overall R&D intensity, with the busi- research centers in Malaysia, the Philippines, ness sector performing a high share of R&D and Vietnam with the objective of assess- (near or exceeding 60 percent of total spending). ing the effectiveness of these institutions The enhanced resources invested in pub- and identifying performance bottlenecks. A lic research, along with a growing supply of key message from these surveys is that these human resources in several countries, has institutions must adopt changes if they are enabled scientiﬁc output to increase dramati- to support governments’ expressed goals of cally since the mid-1990s (figure 5.7, panel a). promoting the successful adoption and diffu- According to 2020 bibliometric data from the sion of existing technologies in local contexts SCImago Journal & Country Rank portal, and strengthening firms’ capabilities. between 2008 and 2018, the number of peer- reviewed publications in Indonesia increased Research capabilities and national 20 times and quadrupled in Malaysia. Peer- strategies reviewed publications in the Philippines, Singapore, and Thailand doubled, or nearly Research effort in the region has increased doubled, over the period.3 in recent years Over the past 10 years, the number of research- Research efficiency remains low in the ers in PROs and higher education institutions region, however (HEIs) has grown by 8–9 times in Malaysia, Despite improvements in some countries, the by about 10 times in Indonesia, and has more quality of research—as reflected in the citation than doubled in Thailand, while China and the impact index of scientific publications (the Republic of Korea have continued to expand H-index)4—remains low in most of developing their base of researchers in both the public and East Asia compared with that of high-income Innovation Policies and Institutions in the Region: An Assessment 157

FIGURE 5.6 The distribution of R&D spending in East Asia, by sector, reflects variations in R&D intensity related to countries’ levels of income and institutional development

100 90 80 70 60 50 40 30 20 10 0 Distribution of gross R&D spending by source (%) source R&D spending by Distribution of gross

China Mongolia Indonesia Malaysia Thailand Myanmar Cambodia Philippines Singapore Korea, Rep.

High-income (OECD) Business sector Government Higher education institutions Private nonprofit and other

Source: Cirera, Kuriakose, and Zuñiga 2021. Note: Gross research and development (R&D) spending data are from 2017 or the latest year available. OECD = Organisation for Economic Co-operation and Development. economies (figure 5.7, panel b). Furthermore, transfer; and the state of technology transfer as in most low- and middle-income coun- links between scientific institutions, industry, tries, the use of domestic scientific knowledge and government.5 (reflected in the citation intensity of domesti- The combined sample from the three coun- cally authored publications) also remains low. tries includes 80 institutions, covering a range This highlights the continuing importance of institutions by scientific field.6 All the insti- of learning from foreign sources of scientific tutions surveyed are involved in R&D. The knowledge in most countries; it also likely RCs in the sample allocate roughly 46 percent reflects a limited impact of domestic research of their R&D expenditures to applied research that is building on local knowledge. and 43 percent to basic research. This greatly contrasts with the PROs, whose main R&D Measuring technology transfer activities activities are in applied research (52 percent of R&D expenditure) and experimental develop- Surveys of PROs and university research ment and preproduction activities (34 percent centers (RCs) in Malaysia, the Philippines, of R&D expenditure). and Vietnam, developed specifically for this report, were implemented between November Assessing governance and performance 2019 and February 2020. The purpose of the monitoring surveys is to deepen policy makers’ understanding of how knowledge institutions in Good governance of RCs and PROs is key the region function and to gauge their ability to making public research systems more to transfer knowledge to industry and soci- accountable, more mission oriented, and ety. The surveys collected information about more impact driven, with connections to pro- institutions’ research capacity; the prevailing ductive systems and to society more broadly. types of institutional governance; the use of The governance issues explored by the survey public policies for research and technology include institutions’ level of autonomy; links 158 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

FIGURE 5.7 Across several East Asian countries, scientific output has increased dramatically, but scientific productivity trails that of high-income countries

a. Scienti c output trends, 2008–18 35,000

30,000

25,000

20,000

15,000

10,000

5,000 Number of peer-reviewed publications 0

2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018

Singapore Thailand Philippines Lao PDR Malaysia Indonesia

b. Quality of research (H-index) and average citations per document, 1996–2018 1,600 25

1,400 20 1,200

1,000 15 800 H-index 600 10

400 5

200 Average citations per document

0 0

Chile India China Japan Mexico Brazil FranceCanada Lao PDR UruguayVietnam Malaysia Myanmar Cambodia Indonesia Colombia Thailand Singapore Korea, Rep. Germany Philippines South Africa United Kingdom

H index Citations per document

Source: SCImago Journal & Country Rank bibliometric indicators (https://www.scimagojr.com). a. Scientific output is measured by the number of peer-reviewed publications per country per year. b. Scientific productivity is measured by the average number of citations per document of national authors (right axis) and by the H-index (left axis). The H-index measures how many national researchers’ publications have at least h number of citations each. For example, an H-index of 5 indicates that the average scientist has five articles with five citations each. As such, the H-index is a measure of the number of highly impactful papers that a country’s researchers have published.

between funding and performance; and man- key features of high-performing PROs and agement practices, including M&E. universities (Cruz-Castro and Sanz-Menéndez 2018). These governance structures include Limited autonomy can impair the effectiveness recruitment procedures, criteria for career of national research institutions promotion, rules regarding the creation of Institutional autonomy and governing struc- intellectual property rights (IPRs) and tech- tures that engage external stakeholders are nology commercialization, and support Innovation Policies and Institutions in the Region: An Assessment 159

programs for knowledge transfer and com- civil servants. The areas where PROs enjoy mercialization. Autonomy should not lead to the strongest decision-making powers are strategies or rules that are disconnected from in day-to-day operational management and national policies for research and innovation, revenue generation strategies, with half of all however. Compliance with national strate- PROs in the sample reporting they have full gies, while preserving autonomy, can be pro- autonomy in these areas. moted through funding mechanisms (at both RCs report having less autonomy than institutional and project levels) and through PROs. This is not surprising, because RCs institutional performance assessments. are generally subordinated to the university Only 24 out of the 80 institutions in the management and rectorates and must comply sample are legally autonomous. Moreover, with institutional directives regarding bud- the level of autonomy of PROs varies across get allocation, infrastructure management, institutions depending on their legal status commercialization, and IPRs. Consequently, and their relationship with the corresponding the only area where RCs report having full ministries or departments. In Malaysia, for autonomy is in the setting of research objec- example, a group of research organizations tives, with about 70 percent of them declaring have private company status and enjoy a high autonomous decision making in this area. level of autonomy and independence. For Autonomy is positively associated with most PROs, however, autonomy is limited by the intensity of collaborative links between statute and by their financial dependence on PROs and RCs and industry, including the sectoral ministries or departments. intensity of technology contracting with Within PROs, the weakest areas of auton- SMEs (figure 5.8, panel a). On average, omy relate to institutional policies and sal- autonomous institutions (both PROs and ary setting because these are often defined by RCs combined) demonstrate a larger number broader laws governing the employment of of collaborative links (per researcher) than

FIGURE 5.8 Research institutions in Malaysia, the Philippines, and Vietnam that were autonomous and included industry stakeholders on their boards engaged in greater collaboration with industry in 2017–18

a. Intensity of technology transfer links, b. Intensity of technology transfer links, by autonomy status by industry participation on boards

1.5 1.0

0.8

1.0 0.6

0.4 0.5 0.2 Technology transfer links per researcher Technology transfer links per researcher 0 0 Entities with Autonomous Entities with no Entities with industry no autonomy entities industry participation participation

Source: Cirera, Kuriakose, and Zuñiga 2021. Note: In both panels, “technology transfer links” refers to all collaborative agreements and contracts with industry, including personal exchanges and training services, collaborative research, contract research, technology services contracts, and licensing of intellectual property rights. The average reported number of collaborative links (per researcher) covers all contracting activity over the period 2017–18 for all researchers in 2018. The combined sample from the three countries included 80 institutions, covering a range of public research organizations and university research centers by scientific field. 160 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

institutions that are not autonomous, with systems—demonstrate greater collaboration a median value of 0.4 links (agreements and with industry and better technology transfer contracts) per researcher. This compares with performance (figure 5.9) . Despite some evi- a median value of close to zero for nonauto- dence of outliers in the data, a positive cor- nomous research institutions. The variance in relation exists between the extent of good outcomes is also higher among the group of governance, as measured by an index of gover- autonomous institutions, however. In addi- nance indicators, and the extent of collaboration, institutions whose boards of trustees tion between national research institutions and include industry stakeholders tend to have industry. a greater number of collaborative links with industry per researcher than those institu- Private sector funding of R&D increases tions without industry participation on their commercialization boards of trustees (figure 5.8, panel b). Another way to incentivize greater technol- Overall, institutions having more good ogy transfer from research organizations to governance practices in place—for exam- industry is through performance-based fund- ple, boards of trustees, autonomy, indus- ing schemes. Performance-based agreements try representation on steering committees (PAs) are a growing mechanism for funding or boards of trustees, performance evalu- among HEIs in high-income countries.7 These ations, and performance-based funding contracts set performance goals and, in most cases, bind a share of their block funding allocation to reaching those targets. FIGURE 5.9 Research institutions with good governance exhibit a Such contracts can stimulate knowledge greater intensity of technology transfer transfer by including not only traditional targets related to teaching and research 6 but also targets associated with the level of engagement with firms or the commercialization of research results. Indeed, the share of research funding coming from the pri- 4 vate sector could be indicative of how much research institutions work with industry and other innovation actors through commercialization activities. The survey data reveal, for example, that PROs with more 2 private sector funding engage in more technology transfer and collaborative projects

Governance index (number of points) (figure 5.10).

Institutions need to continually strengthen 0 0.5 1.0 1.5 research management practices and M&E Collaborative links per researcher Monitoring and evaluation (M&E) is essential to measure the impact of resources invested Governance index Fitted values as well as to guide future policies.8 Across

Source: Cirera, Kuriakose, and Zuñiga 2021. Malaysia, the Philippines, and Vietnam, Note: The figure displays the number of technology transfer links (average per researcher over 65 percent of the institutions surveyed report 2017–18) in relation to an index of good governance. This index is the sum of points (one point per feature) attributed to each of the following features of governance: (a) having autonomy; that authorities conduct performance evalua- (b) having a board of trustees or steering committee; (c) having industry representation on tion of their research and technology transfer boards of trustees; (d) receiving funding through performance-based funding systems; and (e) being subject to institutional performance evaluation for research and technology transfer activities. However, most entities lack imple- activities. Outliers were excluded (entities with more than 100 collaborative links and more mentation plans to achieve evaluation and than 1.5 agreements per researcher on average over 2017–18). The combined sample included 80 institutions from Malaysia, the Philippines, and Vietnam, covering a range of public research technology transfer goals. Only Malaysia organizations and university research centers by scientific field. implements performance-based mechanisms Innovation Policies and Institutions in the Region: An Assessment 161

and has steps to link institutional funding FIGURE 5.10 Research organizations with higher shares of to performance and governance reforms industry-funded R&D undertake more research collaboration (OECD 2016). Even there, performance- based funding is not the main budget source. 70 So, questions remain as to how significantly that approach affects research institutions’ 60 activities. The survey results show that some impor- 50 tant research management practices and use of 40 performance evaluation (PE) systems (periodi- Correlation: 0.31*** cal evaluations, internal rewards, and so on) 30 are well-established practices in Malaysian and Philippine institutions. However, Vietnamese 20 = 1.03182 + 5.0586 + 5.7547 entities (both PROs and RCs) appear signifi- 2 Share of industry-funded R&D (%) = 0.1436 cantly behind in most managerial and good 10 governance practices. Despite the increased use in Vietnam of good managerial and policy 0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 frameworks, their effects on technology trans- Average industry-related projects per researcher fer performance and researchers’ engagement on innovation projects with the private sector Source: World Bank calculations from survey data on public research organizations and university remain unclear. research centers. Note: The intensity in collaborative research with industry is the average number of collaborative In all three countries, publications remain research projects with firms per scientist (in 2017–18) including joint and contract research, doctoral the major determinant in research evalua- projects, and technology services with small and medium enterprises (SMEs). The combined sample included 80 institutions from Malaysia, the Philippines, and Vietnam, covering a range of public tions, while technology transfer activity is not research organizations and university research centers by scientific field. The figure excludes entities considered adequately in performance frame- reporting a share of industry-funded research and development (R&D) exceeding 60 percent as well works or decisions about researchers’ career as outliers. *** designates a 99 percent confidence level. advancement. This absence of institutional incentives could, in turn, be a major deterrent activities across the NIS (Mazzucato 2015). to researchers engaging in technology transfer Malaysia and the Philippines have recently activities. increased their emphasis on mission-oriented research, and their national research strategies have also defined research priorities for Strategies for applied research and innovation and competitiveness to improve technology transfer alignment with industry and societal needs and Commercialization of innovations and trans- challenges. Nevertheless, the survey shows that fer of technology in developing East Asia is despite having a legal mandate to serve private limited by a lack of mission-oriented institu- sector needs, half of the PROs in those two tional planning, limited services to support countries have neither strategic plans for tech- commercialization and technology transfer, nology transfer activities nor permanent con- little emphasis given to technology exten- sultation mechanisms with industry. Except for sion services, and weak technology transfer a few star performers (in agriculture mostly), offices. PROs remain weak in deploying strategies and implementing efforts to achieve their mission- The surveyed public research organizations oriented commitments. lack mission orientation The use of mission-oriented research or inno- Services to support commercialization and vation policies can bring enormous benefits technology transfer are limited by accelerating research and innovation solu- The survey also measures how much of a pri- tions and breaking up institutional silos and ority PROs and RCs give to commercializa- barriers to knowledge and technology transfer tion and technology transfer activities, as well 162 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

as whether PROs and RCs are currently ben- funding for entrepreneurial support services eficiaries of public programs in those areas. In are considered the least important priori- this context, PROs indicated that far greater ties for RCs, while start-up support as well public resources are needed in translational as TTOs and science and technology parks (or applied) research and product develop- appear the least important for PROs. This is ment, whereas RCs called for greater funding not surprising given that regulations restrict for proof of concept, technology validation, public servants working at PROs from engag- and packaging (figure 5.11). ing in entrepreneurial activities—which, in The policy areas in which these institu- practice, critically limit PROs’ technology tions participate least are those related to transfer to industry as well as the creation of start-ups and nurturing services, business spin-off enterprises. services support programs, and technology transfer offices (TTOs). To address the latter The region lacks technology extension issue, public funding support for technology services to support last-mile transfer of transfer activities has been expanded consid- technology to industry erably in Malaysia and the Philippines. In the The provision of technology extension ser- Philippines, a stronger policy agenda has just vices (TES) is an important channel for recently been established for start-ups and technology transfer, and in high-income spin-offs to facilitate research commercializa- countries. TES represents a core area of tion and entrepreneurship.9 PROs’ missions. However, in developing In line with these findings, programs for East Asia, survey respondents from both start-ups (including incubation services) and PROs and RCs report that this type of policy

FIGURE 5.11 The survey showed which technology transfer activities are most important and attract the most participation from research organizations

100 90 80 70 60 50 40 30 availability or priority (%) 20

Share of respondents reporting policy 10 0 Proof of concept, Translational Technology Start-ups Start-up TTOs and science technology research and extension support (business planning, infrastructure and and technology validation, and product programs mentoring, business services parks technology development incubation) support programs packaging Areas of technology transfer program activity

University RCs - Policy exists University RCs - Top priority PROs - Policy exists PROs - Top priority University RCs - Current bene ciary PROs - Current bene ciary

Source: Cirera, Kuriakose, and Zuñiga 2021. Note: The figure displays percentages of public research organizations (PROs) and university research centers (RCs) for which the specified areas are a priority as well as the percentages of PROs and RCs that participate in some government program in those areas. The combined sample included 80 institutions from Malaysia, the Philippines, and Vietnam, covering a range of public research organizations and university research centers by scientific field. TTOs = technology transfer offices. Innovation Policies and Institutions in the Region: An Assessment 163

intervention is only considered of moderate However, TTOs cannot provide all the services importance (figure 5.11)—with the possible needed to support technology transfer and exception of support for agriculture and commercialization processes. Seventy percent electronics in Malaysia. Moreover, the sur- or more of the surveyed RCs report that most vey results indicate that TES for manufac- technology transfer services are available turing systems and industrial upgrading are internally, but the surveys also indicate that mostly weak or missing in countries’ NISs. many of these services (for example, assistance with searching for partners, networking, and There is an ongoing need for clear strategies assessing the value of new technologies) are for technology transfer and intellectual inadequate to address current needs. property revenue sharing Among PROs, access to funding for tech- Institutional policies for IPR ownership and nology transfer activities and assistance in commercialization of intellectual assets are intellectual property protection and manage- expected to provide certainty and clarity in ment, as well as in the creation of spin-offs the legal framework (especially if no national and start-ups, are provided in only about half policy or law on IPRs exists), thereby reduc- of the institutions surveyed. Assessing the ing transaction costs related to asymmetric value of new technologies also remains a criti- information and agency problems (Macho- cal task, not easily accessed or insufficiently Stadler, Martinez-Giralt, and Perez-Castrillo provided by the TTOs. 1996; Siegel, Veugelers, and Wright 2007). It is now common practice to vest the rights Industry links and technology transfer of intellectual property ownership to the activities research institutions themselves (instead of to inventors or national councils) and to pro- There are few links between public research vide revenue participation rights for stake- institutions and industry holders (such as researchers and university There are still relatively few links for knowl- faculties as well as TTOs). edge transfer between PROs or RCs and In Malaysia and the Philippines, IPR and industry. Only a handful of institutions are technology transfer laws have been enacted engaged in a large variety of such links, with to promote the management and commer- most of them undertaking only two or three cialization of research output at PROs and types of knowledge transfer or collaborative public universities. Although institutions activities. Traditional means of knowledge are the main owners of IPRs, there is some transfer—such as human capital exchanges degree of flexibility, and other university (for example, research staff taking on tem- stakeholders can also be entitled with owner- porary assignments in industry) or doctoral ship rights. As for revenue-sharing incentives projects in industry—appear underdeveloped. and the rights of researchers to profit from The most frequent technology transfer the revenues derived from IPR commercial- activities are collaborative research proj- ization, there is a great deal of heterogene- ects with industry and with government ity across the countries. Vietnam lacks a (figure 5.12). University RCs are more clear policy framework at the national level, active than PROs in these projects, report- whereas national laws in Malaysia and the ing a higher average number of collaborative Philippines provide inventors with equity projects (per researcher) with both industry participation rights in addition to other and government as well as a greater number financial incentives (for example, for inven- (albeit still low) of human capital exchanges. tion disclosures and patenting in Malaysia).10 Institutions’ engagement in Research entities also lack sufficient commercialization activities and technology transfer services technology services remains weak Most of the entities surveyed report the pres- Engagement by both RCs and PROs in com- ence of a TTO or an industry liaison unit. mercialization activities involving technology 164 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

FIGURE 5.12 PROs are more likely than RCs to report collaboration and knowledge transfer links with industry

90 80 70 60 50 40 30 Share of respondents (%) 20 10 0 PhD projects in Research sta Sta with joint Collaborative Industry sta in Collaborative Industry sta industry in industry positions research PROs or research received (leaves, sabbaticals, (industry and projects w/ PROs universities projects with training secondment) academia/PRO) or government industry

Type of collaboration or knowledge transfer activity PROs RCs

Source: Cirera, Kuriakose, and Zuñiga 2021. Note: The figure shows the percentages of surveyed research entities that reported undertaking activities in 2017–18. The combined sample included 80 institutions from Malaysia, the Philippines, and Vietnam, covering a range of public research organizations and university research centers by scientific field. PROs = public research organizations; RCs = university research centers.

services (not involving IPRs), licensing of testing services, technology commercializa- IPRs, and entrepreneurship also remains tion activities are absent or marginal in both weak. As in other low- and middle-income PROs and RCs. Entrepreneurial activities by countries—despite a significant increase in start-ups using new technologies (licensing to patenting in universities (and in some PROs) start-ups) and spin-off ventures are also mar- and other forms of IPRs—technology com- ginal in both PROs and RCs. mercialization through licensing of IPRs and start-up creation remains embryonic in The role of academic incentives Malaysia, the Philippines, and Vietnam. PROs are significantly more involved Both international experience and academic than RCs (in terms of both participation and research underscore the importance of pro- intensity) in technology services and train- viding appropriate incentives to those who ing because of PROs’ institutional mandates participate in and manage the technology to provide technical expertise and training transfer process (Siegel, Waldman, and Link to industry, especially to MSMEs. University 2003; Siegel, Veugelers, and Wright 2007). RCs appear less inclined than PROs to engage An appropriate policy framework should in collaborative research with industry. This recognize the role of researchers and encour- pattern illustrates the lack of common inter- age them to engage in technology transfer est and difficulties on the part of academia in activities. Incentives may include (a) inventor engaging with the business sector to conduct royalty compensation and economic partici- joint research projects. pation in revenues from technology services; Except for research contract and tech- (b) awards; (c) recognition in curricula (for nology services (not involving IPRs) such as example, credits for tenure); and (d) equity technical assistance, engineering, and product participation in spin-offs, among others. Innovation Policies and Institutions in the Region: An Assessment 165

The use of incentives is widespread, and such as incubators and accelerators, and on average, 72–90 percent of survey respon- these types of services are still uncommon dents reported having individual performance at public institutions (figure 5.13, panel a). evaluation and career incentives for technol- Further, researchers consider launching start- ogy transfer as well as public recognition ups or spin-offs to be a difficult process, with rewards and assistance in IPR protection and complex procedures. In many cases, PROs management (figure 5.13, panel a). However, are less inclined than RCs to undertake while more than half of the institutions sur- start-up or spin-off development because of veyed report having performance evaluation regulatory constraints that prevail for civil criteria that take technology transfer activi- servants and public institutions under minis- ties into account, universities still generally terial authority. consider publication activity as the main indicator of scientific performance (figure 5.13, Summary of findings on the role of panel b) and thus the key requirement for research institutions in facilitating career advancement. Most institutions have technology transfer some type of financial incentives in place, but the survey indicates that only a few institutions The findings presented in this chapter sug- implement them. In practice, financial rewards gest that national research policies still from commercialization are weak or insuffi- support a model of innovation and tech- ciently implemented. nology that is heavily supply driven. With Equity participation and funding for start- little industry participation, the relevance ups and spin-offs are the least developed types of PROs’ and RCs’ outputs to firms and the of incentives employed by the institutions broader society remains weak. Government surveyed. Only a few large organizations are mandates for technology commercialization able to launch entrepreneurial funding and through start-ups and licensing have had assistance support through specialized units little impact to date.

FIGURE 5.13 Research institutions report a variety of financial and nonfinancial incentives for technology transfer

a. Type of incentive b. Type of performance evaluation 100 100

80 80

60 60

40 40 evaluations (%) out performance

Share of institutions 20 20 Share of institutions carrying having incentives in place (%) 0 0

and careers Funding spin-o s Financial rewards Equity participation Publication focused Funding for IPRFunding costs researchand TT management Not only publications Performance evaluation EntrepreneurshipSecondment assistance opportunities Assistance in IPRRewards, protection public recongition

Total RCs PROs

Source: Cirera, Kuriakose, and Zuñiga 2021. Note: The combined sample included 80 institutions from Malaysia, the Philippines, and Vietnam, covering a range of public research organizations and university research centers by scientific field. IPR = intellectual property right; PROs = public research organizations; RCs = research centers; TT = technology transfer. 166 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

Key conclusions from the surveys include the ability to assess what has worked well, the following: what has not, and how to enhance institu- • Efforts to strengthen research and edu- tions’ impacts on innovation and technol- cation systems have lacked (except in ogy adoption and diffusion. Malaysia) a link between institutional • There remains considerable scope for funding policies and performance mea- strengthening incentives for researchers to surement, including with respect to pro- engage in technology transfer from pub- moting technology transfer. lic research institutions to private sector • The lack of autonomy in many universi- enterprises. University-industry partner- ties and PROs and the lack of good man- ships, technology transfer initiatives, and agement practices in many public research other entrepreneurial activities are inad- institutions in the region further impede equately rewarded in the performance the effectiveness of these institutions in evaluations and careers of scientists, with generating and transferring technologies a much heavier weight placed on scien- that promote private sector innovation. tific publications. This, again, inhibits the • The use of monitoring and performance impact that research institutions could evaluation practices is limited, impeding have in supporting innovation-led growth in the region. Innovation Policies and Institutions in the Region: An Assessment 167

Annex 5A What is the role for States has several mission-oriented innova- innovation agencies? tion agencies, such as the National Institute of Standards and Technology (NIST), the Innovation depends on the availability of Defense Advanced Research Projects Agency national and sectoral factors, such as capital (DARPA), and the Small Business Innovation availability, knowledge, entrepreneurial cul- Research (SBIR) program. ture, a well-educated workforce, and a well- Despite differences in structure and eco- functioning intellectual property regime. The nomic and political context, innovation efficacy of these factors relies on institutions’ agencies have many similarities in the way interconnectivity to collectively promote they operate and the challenges they face. innovation and on government’s role to cre- Table 5A.1 provides an overview of the main ate conditions for innovation to flourish. support methods provided by some of the Innovation agencies help governments best-known agencies globally. to design and implement innovation policy. Many countries have established dedicated agencies to develop and implement strate- Key elements in designing an effective gies to address challenges within their NISs. innovation agency Some of these, in addition to executing inno- The small number of studies analyzing innovation interventions, are also responsible for vation agencies have drawn only a few con- formulating government innovation policy. clusions about how successful innovation However, agencies vary in purpose and struc- agencies are designed and managed. The ture along with the array of policy interven- challenge of this literature is the lack of well- tions reflecting each country’s unique set defined performance indicators that allow of NIS needs. In low- and middle-income one to assess success across different models countries, innovation agencies play a specific and agencies. Nevertheless, these studies sug- and limited role, providing more narrowly gest a few design features that some of the defined interventions than do innovation agencies in the most innovative countries pos- agencies operating at a subnational level in sess that could be correlated with effective- high-income nations. For instance, the United ness (Glennie and Bound 2016):

TABLE 5A.1 Overview of main support mechanisms used by leading innovation agencies

a Direct financial support Nonfinancial Support for Connecting and In-house R&D Agency Grants Loans Other assistanceb intermediariesc institution buildingd projectse FFG, Austria x x x x x x Finep, Brazil x x x x x Corfo, Chile x x x x x Tekes, Finland x x x x x x OCS, Israel x x x x x VINNOVA, Sweden x x x x CTI, Switzerland x x x x ITRI, Taiwan, China x x x x Innovate UK, United Kingdom x x x DARPA, United States x x x x x

Source: Glennie and Bound 2016. Note: Corfo = Production Development Corporation; CTI = Swiss Commission for Technology and Innovation; DARPA = Defense Advanced Research Projects Agency; FFG = Austrian Research Promotion Agency; Finep = Funding Authority for Studies and Projects; ITRI = Industrial Technology Research Institute; OCS = Israel Innovation Authority (formerly Office of the Chief Scientist); Tekes = Finnish Funding Agency for Technology and Innovation; VINNOVA = Sweden’s innovation agency. a. The type of financial support the agency provides for innovation. b. Whether the agency provides nonfinancial assistance such as technical assistance and expertise. c. Whether the agency supports intermediary institutions for innovation such as accelerators or technology centers. d. Whether the agency provides support by connecting institutions. e. Whether the agency and its staff implement research and development (R&D) projects. 168 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

• Innovation agencies should operate with NIS (Aridi and Kapil 2019); and (b) necessary considerable autonomy to be effective. oversight and accountability for a public or • Innovation agencies cannot be entirely quasi-public organization. separate from political processes. Innovation agencies commonly have one • Measuring and evidencing the long-term of four types of governance structures: min- or systemic impact of their portfolio is isterial unit, government agency, government challenging for innovation agencies. agency with a high degree of autonomy, and • It is hard to identify a single model of a nonprofit public-private partnership. “successful” innovation agency. A ministerial unit is an office within a ministry that reports to the minister and Despite a wide range of innovation agencies receives allocations from the ministry’s bud- in low- and middle-income countries, certain get. Although it is controlled by and aligned parameters have emerged from studying dif- with the ministry’s strategy (therefore making ferent agencies. The four observed key steps it more accountable), it may be hampered by involved in the design and management of bureaucracy and political pressures constrain- an effective innovation agency are as follows ing resources as well as by weak firm or entre- (Glennie and Bound 2016): preneurial know-how. • Identifying the right organizational mis- A government agency, being a formal gov- sion depending on markets and identified ernmental bureaucracy under direct political challenges control, generally has stronger long-term plan- • Choosing effective management struc- ning ability but is resource dependent, prone tures and staffs with relevant skills and to political pressure, and has less flexibility experience to hire and retain talent. However, a govern- • Selecting the most appropriate set of meth- ment agency with high autonomy is an agency ods and tools, such as diagnostic-based owned by the government but with a degree interventions to address NIS gaps (Aridi of independence and flexibility to derive its and Kapil 2019) funding from multiple sources, such as from • Establishing a set of metrics and measure- donors or the private sector. However, its deci- ments that will help the agency to under- sion making is still prone to political influence. stand and achieve impact through internal Finally, a nonprofit public-private partner- or external capabilities. ship is a separate legal entity, often a nonprofit organization or a foundation. It combines Almost all parameters relate to an organiza- resources from public and private sources and tion’s strategic and operational capabilities, is led by the private sector to establish cred- complement each other, and adjust to the ibility in the market and to develop relevant evolving needs of the NIS. Strategic capa- services to firms. Nevertheless, this governance bilities clearly define and adjust the agen- structure may drift from public sector priorities cy’s mission according to market needs to as activities are adapted to “follow the money.” enhance their relevance and impact (Aridi and Kapil 2019). Managerial capabilities allow staff and teams to grow, learn, and Funding of innovation agencies establish conducive partnerships. Sustainable funding is essential to avoiding funding fluctuations that jeopardize programs and agencies. Reliable funding, often from gov- Governance of innovation agencies ernment sources, is critical for hiring and train- Effective governance and management ing staff, designing policies, and establishing structures involve balancing two competing M&E protocols. Relying too heavily on gov- goals: (a) autonomy to make decisions based ernment funding, however, may mean fluctua- on their professional judgment given the tions in funding, which in turn can jeopardize strategic positioning of the agency within the the program and even the agency’s existence. Innovation Policies and Institutions in the Region: An Assessment 169

Proactive agencies cultivate multiple Program (IRAP) is another example of a funding sources to enhance program offer- productivity facilitator. Its mission is to ings and diversify their portfolios, which is assist SMEs with technological innovation critical to ensure the sustainability of financ- and diffusion to promote firm growth. ing. Alternative funding sources include IRAP’s embeddedness enables it to (a) international partners from high-income effectively address barriers to innovation. countries offering economic and social assis- Moreover, rather than solving particular tance; (b) multinational or international issues by developing technological solutions organizations as part of their overall eco- themselves, the organization’s industrial nomic development program; (c) charitable technology advisers (ITAs) use their net- foundations that contribute to social goals works to locate other organizations that can within low- and middle-income countries; assist with the necessary R&D. In doing so, (d) fee-based services; and (e) investment they construct a framework for technologi- income from direct equity investments. cal diffusion and continuous, incremental innovation. Directed upgraders . Another way to Innovation agencies and public organize an innovation agency is through sector R&D reliance on the public sector, rather than the Rapid technological and economic changes private sector, to steer technological devel- place a premium on the effective design of opment. Singapore’s Agency for Science, innovation agencies as it relates to public Technology and Research (A*STAR) sector R&D. In this context, four different and Chile’s Production Development agency categories have been defined in Corporation (Corfo) are examples of more developed-country contexts: “directed “directed upgraders” (Breznitz, Ornston, upgraders,” “productivity facilitators,” and Samford 2018). Singapore’s A*STAR “state-led disruptors,” and “transformation is a centrally situated, classic developmental enablers” (Breznitz, Ornston, and Samford agency that is connected to private industry 2018). These categories differ in (a) the level while also having close ties to other public of public sector R&D involvement, (b) the sector actors. Corfo, another example of a agencies’ positioning within the public sec- directed upgrader, maintains close relation- tor, and (c) the degree of embedding within ships to the private sector—both individ- private industry. Table 5A.2 shows the four ual firms and industry organizations—as a innovation agency categories with examples means of understanding the kinds of mar- from around the world that vary in the spec- ket failures that prevent Chilean firms from ificity and structure of their missions. upgrading and what technologies might be Productivity facilitators . The Danish imported from abroad to assist the local GTS institutes (Research and Technology economy. Organizations) represent almost ideal State-led disrupters . “State-led disrup- examples of “productivity facilitators” tors” (Breznitz, Ornston, and Samford 2018) (Breznitz, Ornston, and Samford 2018). These also take a primary role in the performance institutions vary in size and specialization, but of research and technology development; all work closely with private sector partners however, their position at the periphery of to identify and solve technological challenges. the public sector and autonomy from estab- Even though the government views them as an lished industries enable them to develop important policy instrument and funds roughly radical, novel innovations. The Industrial 10 percent of their activities, they operate at Technology Research Institute (ITRI) of the periphery of the public service, with almost Taiwan, China, and DARPA of the United all their budget coming from private industry. States are state-led disruptors and sit at the Like the GTS institutes in Denmark, periphery of the public sector. Further, both Canada’s Industrial Research Assistance are actively involved in network creation 170 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

TABLE 5A.2 Examples of innovation agencies, by type

Category Definition Organizational features Examples Productivity facilitator Introduces small-scale, incremental product Locus of R&D: Private GTS institutes (Denmark) and process innovations across a wide range of IRAP (Canada) established industries Position in public sector: Peripheral Relation to established industry: Embedded Directed upgrader Specializes in incremental innovation but Locus of R&D: Public A*STAR (Singapore) mobilizes resources around a relatively narrow Corfo (Chile) range of industries and activities, facilitating Position in public sector: Core large-scale change Relation to established industry: Embedded State-led disruptor Excels at radically innovative technological Locus of R&D: Public ITRI (Taiwan, China) breakthroughs DARPA (United States) Position in public sector: Peripheral Relation to established industry: Autonomous Transformation enabler Radically innovative but characterized by many Locus of R&D: Private Sitra (Finland) small-scale experiments rather than a narrow, OCS (Israel) focused approach Position in public sector: Peripheral Relation to established industry: Autonomous

Source: Breznitz, Ornston, and Samford 2018. Note: A*STAR = Agency for Science, Technology and Research; Corfo = Production Development Corporation; DARPA = Defense Advanced Research Projects Agency; GTS = Research and Technology Organizations; IRAP = Industrial Research Assistance Program; ITRI = Industrial Technology Research Institute; OCS = Israel Innovation Authority (formerly Office of the Chief Scientist); R&D = research and development; Sitra = Finnish Innovation Fund.

with an ever-changing set of private firms Annex 5B Sampling and where the agency sits at the nodal point implementation of the World charting out technology trajectories up to Bank survey of public research the level of specific early-stage products. The institutions main difference between these two agencies is that while ITRI is actively engaged A proportional-stratified random sampling in R&D, DARPA contracts out the R&D was implemented in three countries— to external researchers, private firms, and Malaysia, the Philippines, and Vietnam— universities. based on the type of institution and location, Transformation enablers . While DARPA and with the aim of capturing research and ITRI actively shape and steer techno- engagement and sufficient heterogeneity logical change, “transformation enablers” in research capabilities across institutions. (Breznitz, Ornston, and Samford 2018) such In the case of universities, the focus was as the Finnish Innovation Fund (Sitra) and research-engaged institutions. Institutions Israel’s Israel Innovation Authority (OCS, involved in the social sciences, management, formerly Office of the Chief Scientist) oper- and arts and humanities were excluded. The ate more indirectly. DARPA, for example, is surveys were implemented from November a classic mission-oriented agency intended to 2019 to February 2020 through face-to-face identify and develop cutting-edge technolo- interviews with directors of PROs and uni- gies for defense. Sitra, on the other hand, was versity RCs. Interviewees received question- established in the 1960s to promote private naires one to two weeks in advance. In most industry during a period that was dominated cases, half of the survey was covered during by state-owned enterprises and heavily regu- the interview, while the rest of the question- lated industries. It did so by prioritizing tech- naire was completed and submitted after the nological innovation. meetings. Innovation Policies and Institutions in the Region: An Assessment 171

In Malaysia, the survey was administered For Vietnam, which has more than 1,000 to 10 PROs and 16 university RCs across four universities, the list of research universities public research universities. Both PROs and was produced based on the top 20 universi- RCs were randomly selected. The PROs sur- ties whose publication activity in 2016 and veyed are under different ministerial bod- 2017 is recorded in the Web of Science cita- ies covering different technical fields and tion indexing database (of the Institute for industries. The list of public research insti- Scientific Information, or ISI). From this list, tutes compiled by the Malaysia Science and the selection focused only on two cities: Hanoi Technology Information Centre (MASTIC) and Ho Chi Minh City (HCMC). Next, for under the Ministry of Science, Technology and each university, a comprehensive list of RCs Innovation (MOSTI) was used as a sampling was produced and analyzed by accessing the frame, supplemented by a list of PROs com- university websites and selecting RCs using the piled by World Bank staff. Random samples following criteria: an RC unit must (a) con- were drawn from this list, with PROs repre- duct research, (b) have a research director, and senting various ministries in Malaysia. For (c) have equipment or a laboratory for doing universities, the team used the national listing research. In Vietnam, for both PROs and RCs, of “research universities” officially defined by the target sample for PROs was 20. The final the government. Before a random sample was sample contained 16 RCs and 16 PROs. drawn, the research centers were classified into three categories: (a) Higher Institution Centers of Excellence (HICoEs), which are Notes specified by the Ministry of Education; (b) 1. Policy instruments usually have a defined Centers of Excellence (CoEs), classified by set of objectives, group of beneficiaries, and the universities; and (c) research centers that logical framework. They often involve one are neither HICoEs nor CoEs. Each university mechanism of intervention—for example, has at least one of each type of research cen- a grant, scholarship, or tax incentive. They ter within the sample to ensure accurate and differ from government programs in that robust representation. RCs were randomly government programs often include several selected, but an effort was made to ensure that policy instruments and much broader the team surveyed at least one HICoE, one objectives. For example, a national program CoE, and one research center per university. to support start-ups can include a range of For the Philippines, an initial strati- instruments, such as incubators, accelerators, fied sample of 14 RCs out of a total of 47 hackathons, and equity finance policy instruments. research centers (engaged and producing 2. Although there is no blueprint for how STI outputs) and 14 out of 17 PROs was expenditure for innovation should be defined. Roughly 59 percent of the 17 PROs distributed, the concentration of resources are under the Department of Science and for innovation in just a few programs Technology (DoST), followed by 29 percent in these countries is higher than what is under the Department of Agriculture (DA), observed in most Organisation for Economic while the others belong to other line agencies. Co-operation and Development (OECD) To identify RCs, the Commission on Higher countries. Education (CHED) provided the team with a 3. The region’s lower-middle-income countries list of HEIs, which identified 882 universities (for example, Lao PDR) have also increased conducting R&D in the Philippines. Out of their publication activity, but their total numbers remain low relative to other these, 243 conduct research activities associ- countries in the region. ated with science, technology, and innova- 4. The H-index is a measure of the number of tion—42 percent of which are in the National highly impactful papers that researchers have Capital Region (Manila), Region III, and published in the country. The more influential Region IVA. The final sample contained 15 are publications, as measured by citations, the RCs and 8 PROs. higher the H-index. 172 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

5. For additional details on the implementation Industrial and Corporate Change 27 (5): of the survey, see annex 5B. 883–96. 6. For 35 percent of the entities, their main Cirera, Xavier, Jaime Frias, Justin Hill, and research field is engineering (environmental Yanchao Li. 2020. A Practitioner’s Guide to engineering, chemical engineering, transport Innovation Policy: Instruments to Build Firm engineering, and so on); 21 percent specialize Capabilities and Accelerate Technological in electronics and telecommunications Catch-Up in Developing Countries. research; 14 percent in medical sciences; 7.5 Washington, DC: World Bank. percent in life sciences and biotechnology; Cirera, Xavier, Smita Kuriakose, and Pluvia and the remainder in other fields including Zuñiga. 2021. “Assessing the Effectiveness of agricultural research. Public Research Institutions in the Transfer of 7 . The OECD (2016) estimates that about 60 Knowledge and Technology to Industry and percent of universities in OECD countries Society: Evidence from a New Survey in Three have adopted this funding mechanism. Developing East Asia Countries.” Unpublished, 8. In several OECD countries, evaluation World Bank, Washington, DC. is mandatory for the renewal of research Cirera, Xavier, and William F. Maloney. 2017. funding agreements through the use of The Innovation Paradox: Developing-Country multiannual performance-based contract- Capabilities and the Unrealized Promise of agreements, which require specific results Technological Catch-Up. Washington, DC: to be accomplished. M&E systems are World Bank. also being adopted to assess the progress Cruz-Castro, Laura, and Luis Sanz-Menéndez. of national research and innovation plans 2018. “Autonomy and Authority in Public (OECD 2016). Research Organisations: Structure and 9. In the Philippines, the Startup Innovation Funding Factors.” Minerva 56 (2): 135–60. Act (passed in 2019) provides incentives for doi:10.1007/s11024-018-9349-1. innovative start-up creation and growth, Glennie, Alex, and Kirsten Bound. 2016. “How including subsidies for firm creation, use of Innovation Agencies Work: International facilities, training on entrepreneurship and Lessons to Inspire and Inform National intellectual property, and financial assistance Strategies.” Report, Nesta, London. to start-ups. The act solidifies mechanisms Macho-Stadler, Ines, Xavier Martinez-Giralt, and in providing financial assistance through the David Perez-Castrillo Jr. 1996. “The Role of creation of the Startup Grant Fund and the Information in Licensing Contract Design.” Startup Venture Fund. Research Policy 25 (1): 43–57. 10. Invention disclosures take the form of a Mazzucato, Mariana. 2015. “Beyond Market report that represents the first recording of Failures: Shaping and Creating Markets for the invention. The disclosure describes the Innovation-Led Growth.” In Mission-Oriented invention’s novelty and establishes the date Finance for Innovation: New Ideas for and scope. Such disclosures are needed to Investment-Led Growth, edited by Mariana ensure that inventors do not lose patent rights Mazzucato, and Caetano C. R. Penna, 147–59. associated with their inventions. London: Rowman & Littlefield. OECD (Organisation for Economic Co-operation and Development). 2013. Innovation in References Southeast Asia. OECD Reviews of Innovation Policy Series. Paris: OECD Publishing. Andrews, Matt, Lant Pritchett, and Michael OECD (Organisation for Economic Co-operation Woolcock. 2012. “Escaping Capability Traps and Development). 2016. “Enhancing through Problem-Driven Iterative Adaptation Research Performance through Evaluation, (PDIA).” Working Paper No. 299, Center for Impact Assessment and Priority Setting.” Global Development, Washington, DC. Report, Directorate for Science, Technology Aridi, Anwar, and Natasha Kapil. 2019. “Innovation and Innovation, OECD, Paris. Agencies: Cases from Developing Economies.” Rasul, Imran, and Daniel Rogger. 2017. Report, World Bank, Washington, DC. “Management of Bureaucrats and Public Breznitz, Dan, Darius Ornston, and Steven Service Delivery: Evidence from the Nigerian Samford. 2018. “Mission Critical: The Ends, Civil Service.” Economic Journal 128 (608): Means, and Design of Innovation Agencies.” 413–46. doi:10.1111/ecoj.12418. Innovation Policies and Institutions in the Region: An Assessment 173

Siegel, Donald S., Reinhilde Veugelers, and Siegel, Donald S., David Waldman, and Albert Mike Wright. 2007. “Technology Transfer Link. 2003. “Assessing the Impact of Offices and Commercialization of University Organizational Practices on the Relative Intellectual Property: Performance and Policy Productivity of University Technology Transfer Implications.” Oxford Review of Economic Offices: An Exploratory Study.” Research Policy 23 (4): 640–60. Policy 32 (1): 27–48.

Action for Innovation: A Policy Agenda 6

Introduction Chapter 5 argued that the region’s policies and agencies are not well positioned to pro- This report started by making the case for a mote increased innovation and technological change in the growth model for developing catch-up. The main factors that undermine East Asia. The productivity growth slow- the impact of policies on innovation include down, exacerbated by the numerous chal- a mismatch between policy objectives, inno- lenges ahead—ongoing trade tensions, rapid vation capabilities, and resource allocation, technological advances that threaten exist- as well as weak governance and institutional ing production methods and trade patterns, capacity among innovation agencies and large economic shocks from the COVID-19 public research organizations (PROs). pandemic, and increasing climate change This final chapter discusses several sets challenges—demand swift policy actions to of policy actions that can help policy makers accelerate innovation and the adoption of address these three key bottlenecks—policies, new technologies. Such actions are critical if complementary factors, and institutions—to the region’s countries are to transition success- facilitate technological catch-up and promote fully to high-income status with rapid poverty greater innovation-led growth in the region. reduction and increased shared prosperity. Chapter 3 provided a diagnostic of some Addressing the innovation policy of the main bottlenecks limiting innovation mismatch and building firms’ in the region. It highlighted the need for more capabilities firms to adopt new technologies and engage in innovation, while those firms already To spur innovation more effectively—both undertaking innovation activities should diffusion of existing technologies and inven- focus on building their capacities to imple- tion at the frontier—and to better keep ment more sophisticated innovation proj- pace with the wave of new technologies, ects and invent new technologies. Chapter 4 the region’s policy makers must better align highlighted additional bottlenecks to inno- country policies, innovation capabilities, vation in the region, specifically the lack of and resources. They must also work to build adequate skills and finance, which are impor- firms’ capacities to innovate. But what does tant complementary factors for innovation. this look like in practice? How should policy 176 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

makers deal with the substantial heterogene- more-basic innovation activities—that is, to ity in firms’ innovation capabilities? diffusion and adoption. Several sets of policy actions are critical But which policy instruments can most to strengthening national innovation poli- effectively promote innovation in the cies: (a) reorienting policy objectives in a region? The answer depends on the spe- graduated manner—guided by the capabili- cific innovation-related objective as well ties escalator—to build innovation capabili- as where countries, sectors, and firms are ties; (b) eliminating biases against diffusion situated on the capabilities escalator. Cirera and adoption relative to invention and et al. (2020) provide a detailed description against services sector innovation relative to of the main policy instruments available to manufacturing; and (c) choosing the most support innovation, the conditions needed appropriate policy instruments to support for their successful implementation, and the countries’ objectives. existing evidence on impact. Table 6.1 summarizes some of the key innovation policy objectives (column 1), the minimum state Adjusting the innovation policy mix to of capabilities needed (column 2), the chal- enable diffusion and adoption lenges firms face and the types of firms that The capabilities escalator can guide should be targeted (column 3), and the pol- innovation policy choices icy instruments that can support each objec- Innovation policy requires clear objectives tive (column 4). and sustained commitment. At the same Objectives 1 and 2 in table 6.1—“build time, the appropriate combination of policies, basic innovation capabilities” and “acceler- emphasizing specific instruments, is likely to ate technology diffusion and adoption”— change as innovation capabilities evolve. are critical to building the foundation for Chapter 2 presented the idea of the greater innovation, specifically adoption and “capabilities escalator” to help guide the diffusion. These objectives, and the related choice of innovation policy instruments. policy instruments, are broadly relevant in As the chapter illustrated (in figure 2.8), developing East Asia but are particularly different types of policies are required important in countries that have relatively at different stages of development of undeveloped (“incipient”) national innova- countries’ national innovation systems. An tion systems (NISs) and where most firms approximation of where the countries in the do not engage in innovative activities and region are located with respect to innova- use outdated technologies. The main objec- tion capabilities was presented in figure 2.7, tives in these contexts should be in building which shows three clusters of countries that basic innovation capabilities and supporting largely correspond to the capability levels more firms in undertaking more innovation shown in figure 2.8. activities, especially adopting new technolo- At the lower levels of innovation capabil- gies. Although these objectives should be at ity, where most countries in the region (except the center of innovation policies in relatively China) can be found, the emphasis should be low-capability, lower-middle-income coun- on supporting more firms to initiate innova- tries such as Cambodia, the Lao People’s tion projects and adopt new technologies. For Democratic Republic, and Myanmar, they firms that are already carrying out research continue to apply broadly to countries and and development (R&D), however, the firms across the region. emphasis should be on enabling them to per- As countries’ and firms’ innovation capa- form more complex R&D projects, patent- bilities increase, other objectives begin to take ing, and invention. Thus, given the region’s greater precedence. These include (a) taking current innovation capabilities, one should advantage of the positive spillovers from col- expect more public resources to be allo- laborating with multinational enterprises cated to policy instruments that encourage (MNEs) and participating in global value Action for Innovation: A Policy Agenda 177 table continues next page to find to innovative solutions for MNEs a innovation objective (4) objective innovation Key policyKey instruments to support er- or network-baseder- policies facilitate to diffusion of n innovation programs chers for innovation incentivize to collaboration between plier development programs build to suppliers’ capabilities to ity finance and angel investment elerators help to initiate and scale up start-ups ns to purchase to ns new equipment andtechnology finance adoption iness advisory services improve managerial to capabilities iness advisory services improve management to processes ional quality infrastructure (NQI) services support to quality hnology extension services provide to information and know- hnology extension services provide to information and know- hnology for centers sector-specifictechnical assistance and ubators provide to business advisory services for transforming ence and technology parks attract to MNEs establish to R&D centers andcenters collaborate with local universities and industry and large firms Sci Inc innovative ventures ideas into Acc Equ Bus Tec Nat Clust Vou firms and institutions and knowledge providers Bus and support organizational changes implement required new to technologies Tec Tec Loa Sup absorb technology and innovate Ope how for adopting new technologies standards and upgrading sectors within technologies how for adopting new technologies support in adapting existing technologies

• • • • • • • • • • • • • • • eneral innovation activity due oung innovative ventures ort manufacturing and other f old, outdated technology; very firms to be targeted(3) echnological and innovation vices sectors that provide s operating in the digital space s operating in traditional s operating in traditional s operating in t firmscountries in all in the t firmscountries in all in the ential for enterprise spin-offs from Exp GVCs into integrated sectors Ser services GVCs to

∙ ∙ Few y Pot Firm and in services sectors universities manufacturing and services, nonexport sectors low technology adoption manufacturing and services, nonexport sectors Low g lowto capabilities; firms many not adoption or innovation in engaged Mos region Firm o Use Mos region Firm Few t spillovers from MNEs Firm Challenges firms face and types of

• • • • • • • • • • • Minimum state of roduction capabilities roduction capabilities nd medium technological nd medium technological e research and technology ium production capabilities ium production capabilities ally middle-income countries countries middle-income ally ally middle-income countries countries middle-income ally ally low-ally and lower-middle- low-ally and lower-middle- innovation capabilities (2) innovation capabilities y low technological capabilities y low technological capabilities te mix of policy instruments to facilitate innovation and diffusion depends on countries’, sectors’, and firms’ innovation capabilities innovation and firms’ sectors’, and diffusion depends on countries’, innovation mix of policyte facilitate to instruments Usu Med a Low Som capabilities transfer capabilities income countries income countries income Usu Low p Ver Usu Low p Ver Usu Med a Low capabilities

• • • • • • • • • • • • • The most appropria The

6.1 4. Increase the number of innovative start-ups 1. Build basic innovation1. capabilities technology2. Accelerate adoption and diffusion 3. Maximize spillovers from MNEs and GVCs TABLE Innovation policyInnovation objective (1) 178 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA b innovation objective (4) objective innovation Key policyKey instruments to support nance reforms universities at and incentivize PROs to centives for R&D projects centives for R&D projects g-term grants for R&D projects and collaboration perform to g-term finance for innovation projects ns to purchase to ns R&D equipment andtechnology finance adoption nts for individual and collaborative R&D projects between firms hnology transfer offices to enable transfer of new technologies hnology facilitate to centers the diffusion of technologiesnew to hnology extension services support to diffusionto industry laboration grants to support collaboration university-industry to grants laboration commercial innovation procurement ence and technology parks attract to MNEs establish to R&D Gover Tec Tec Tec Col in Tax Gra and between firms and knowledge providers Loa Sci andcenters collaborate with local universities and industry in Tax Lon joint and complex R&D projects Lon Pre collaboration with industry, technology generation, and and generation, technology industry, with collaboration commercialization from universities and industry PROs to industry and provide to links PROs and to universities

• • • • • • • • • • • • • s of medium and large firms ffectivetechnology transfer e of technological start-ups firms to be targeted(3) s operating in lead sectors and/ k technology generation and optimal investment in R&D h-risk, long-term R&D and h R&D capabilities and strong ential for enterprise spin-offs from ge and leading firms or connected GVCs to technological projects taking place collaborative projects R&D Wea commercialization Pot universities Few e channels and limited collaboration between universities/PROs and industry Firm Sub A mas conducting R&D in different sectors, sectors technology including A bas Hig Hig Lar Challenges firms face and types of

• • • • • • • • • • Minimum state of nd medium technological ium production capabilities ium research and technology ium and high production ium and high technological ium and high research and ium and high production ium and high technological ium and high research and ally middle-income countries countries middle-income ally ally middle- and high-income ally high-income countries innovation capabilities (2) innovation capabilities te mix of policy instruments to facilitate innovation and diffusion depends on countries’, sectors’, and firms’ innovation capabilities innovation and firms’ sectors’, and diffusion depends on countries’, innovation mix of policyte facilitate to instruments transfer capabilities capabilities capabilities technology transfer capabilities capabilities capabilities technology transfer capabilities Usu Med a Low capabilities Med Usu countries Med Med Med Usu Med Med Med

• • • • • • • • • • • • The most appropria The

6.1 World Bank elaboration, adapted from Cirera et al. 2020. GVC = global value chain; MNE = multinational enterprise; PRO = public research organization; R&D = research and development. 5. Strengthen5. collaboration for technology generation, and transfer, commercialization Support frontier7. innovation via high-risk and complex R&D projects 6. Increase quantity and quality of R&D investments TABLE Innovation policyInnovation objective (1) Source: (continued) Note: a. In open innovation programs, firms typically identify problems in their production and management processes and search for solutions,see Cireraoften et al. working (2020). with other firms, universities, and knowledge providers. For further details, b. Precommercial innovation procurement refers to support to firms’ innovation through the purchase of R&D services that include the delivery “productof a further prototype” details, see or some Cirera innovation et al. (2020). project before it commercialized.is For Action for Innovation: A Policy Agenda 179

chains (GVCs) (table 6.1, objective 3); and encouraging non-innovating firms to innovate (b) expanding the base of innovative start- by building basic innovation capabilities (policy ups in the country (table 6.1, objective 4). objective 1) and on accelerating technology Strengthening the foundations for technology adoption and diffusion (policy objective 2). As generation, transfer, and commercialization— discussed throughout this report, most firms in by promoting greater collaboration between the region are not innovating and continue to research entities and industry—also becomes use old or outdated technologies. a priority as countries move up the capabilities Second, most countries in the region escalator (table 6.1, objective 5). Promoting should also focus on realizing positive spill- the generation and effective transfer of new overs from MNE investments and GVC technologies requires more advanced techno- participation (policy objective 3). Especially logical capabilities, as are generally found in for countries with relatively low to moder- upper-middle-income countries such as China ate innovation capabilities, scarce public and Malaysia. resources in these countries are better spent While countries often strive to do more and on adoption of existing technologies embod- better R&D (table 6.1, objective 6), effective ied in foreign direct investment (FDI) and par- R&D projects require that countries and firms ticipation in GVCs than on efforts to generate have adequate capabilities to carry out such cutting-edge technologies directly. initiatives. These capabilities tend to be more Third, as countries’ innovation capabilities common in more-developed economies and, increase, they will benefit from progressively within them, among larger firms and firms in adjusting their policy focus to include support more technology-intensive sectors. Finally, suc- to innovative start-ups (policy objective 4); cessfully supporting frontier innovation proj- improving collaboration between research ects (invention) through high-risk and complex entities and industry to increase technology R&D projects (table 6.1, objective 7) requires generation, transfer, and commercializa- high innovation capabilities (along with sig- tion (policy objective 5); strengthening of nificant coordination between research enti- domestic R&D (policy objective 6); and, ulti- ties, universities, and firms) that are difficult to mately, pursuing complex and high-risk R&D find in countries with less-mature NISs. China projects at the frontier (policy objective 7). is an exception in the region—with a number of firms already engaged in frontier R&D—as Countries should adjust their policy are Malaysia and Thailand, to a lesser extent. priorities as they develop their innovation Building on the policy objectives, innova- capabilities tion challenges, and policy instruments out- Climbing the capabilities escalator is a dynamic lined in table 6.1 as well as the concept of the process and hence requires adjustment of inno- capabilities escalator illustrated in chapter 2 vation policy priorities over time. This can be (figure 2.8), table 6.2 illustrates a set of policy seen in the experience of highly innovative priorities for the countries in developing East countries in the region, such as the Republic Asia. Country priorities shown in the table of Korea and Singapore, that have periodically reflect the analysis in this report along with updated their policy mixes to achieve conver- other World Bank country studies on innova- gence with the technological frontier. tion (Garcia et al. 2020 [on the Philippines]; Table 6.3 summarizes Korea’s innovation Kuriakose and Tiew 2020 [on Malaysia]; policy journey since the 1960s. While the World Bank 2020 [on Vietnam]; World Bank journey reflects the country’s specific charac- and DRC 2019 [on China]). The darker cells teristics—for example, the presence of large indicate higher levels of priority considering business conglomerates (“chaebols”)—it countries’ and firms’ innovation capabilities. has also two important lessons for develop- A few aspects of the table are noteworthy: ing East Asia: First, the country has pursued First, although countries in the region have a consistent and overarching objective of different levels of innovation capabilities, it developing technological capabilities over the is still important for all of them to focus on entire period. Second, Korea has regularly 180 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA China Malaysia ure 2.8 (ranging from lower capabilities on the left to Vietnam Thailand Mongolia Philippines Indonesia reflects their location on the capabilities escalator as shown in chapter 2, fig Cambodia Myanmar ountries have varying innovation policy based on the capabilities varying escalator priorities, innovation have ountries Lao PDR Developing East Asian c Developing

6.2 Original table for this publication. Darker cells denote higher policy priorities. The order of the countries presented in this table 1. Build basic innovation1. capabilities technology2. Accelerate diffusion adoption encourage and 3. Maximize spillovers from MNEs and GVCs 4. Increase the number of innovative start-ups Strengthen5. collaboration for technology generation, transfer, and commercialization 6. Increase quantity and quality of R&D investments Support frontier7. innovation via high-risk, complex R&D projects TABLE Innovation policyInnovation objective Source: Note: higher capabilities on the right). GVCs = global value chains; MNEs = multinational enterprises; R&D = research and development. Action for Innovation: A Policy Agenda 181

TABLE 6.3 Republic of Korea’s innovation policies continually evolved to reflect its accumulation of capabilities

Policy aspect Innovation capabilities, objectives, and priorities 1960s–1970s: An incipient national innovation system (NIS) Innovation capacity In the early 1960s, Korea was largely a closed agrarian economy with limited natural resources, a small and enabling domestic market, and cheap labor. conditions Firms: Low awareness and lack of technological literacy and managerial capabilities to assimilate foreign technologies or conduct R&D Markets and institutions: Low export orientation, FDI intensity, and associated knowledge flow; state-owned commercial banks that provided policy loans; weak intellectual property rights (IPR) framework Research: Newly established PROs and low R&D capacity in universities; incipient innovation infrastructure with low availability of laboratories and testing facilities Skills: Basic STEM education and postsecondary technical programs Policy objectives • Build institutions and human resources to strengthen the country’s science and technology (S&T) capabilities • Develop domestic technologies as well as adopt, absorb, and assimilate imported technologies • Promote the growth of heavy and chemical industries led by large firms (chaebols) to increase exports and promote import substitution Policy mix • Management extension programs • Grants and loans for business innovation and productivity improvements in strategic industries • Loan guarantees and accompanying TA • Tax incentives for technology adoption and R&D • Te chnology licensing • Standards and basic NQI infrastructure • Export promotion policies • Shared R&D facilities within industrial parks and cooperatives • Specialized PROs in strategic industries • Expanded domestic STEM education, overseas training scholarships, and repatriation of experts 1980s: A maturing NIS Innovation capacity By the 1980s, the economy had grown rapidly, led by chaebols in strategic industries and supported and enabling by interventionist policies. conditions Firms: Links to GVCs and export markets; growing investment in in-house R&D (surpassing the government’s R&D expenditure); improvements in large firms’ product standards Markets and institutions: Increasing export orientation and FDI intensity, leading to some knowledge spillovers; expanded access to finance from rapidly growing nonbanking financial institutions owned by chaebols; improved IPR framework Research: Emerging clusters of applied research in PROs and universities; strong links and collaboration with strategic industries but not with other sectors; availability of competitive scientific research funding Skills: Increased availability of STEM-skilled workers

table continues next page 182 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

TABLE 6.3 Republic of Korea’s innovation policies continually evolved to reflect its accumulation of capabilities (continued)

Policy aspect Innovation capabilities, objectives, and priorities Policy objectives • Obtain and supply advanced technologies and improve productivity of industries to secure competitiveness internationally • Enhance productivity, managerial capabilities, and technological literacy of SMEs • Strengthen R&D capacity of large firms and SMEs • Promote research cooperation among PROs, industry, and academia Changes in policy mix • Reduced policy loans for chaebols in strategic industries • Expanded and/or adjusted innovation policy instruments to better target SMEs and start-ups (such as tax incentives, grants, matching grants, loans, loan guarantees, business advisory services, and technology extension programs) • Accelerated import and FDI liberalization policies to introduce foreign competition and pressure domestic firms to enhance productivity • Public procurement for innovation (during commercial stages)a • Restructured PROs and competitively distributed R&D grants 1990s: A mature NIS Innovation capacity With increasing globalization, Korea deregulated and opened its economy during the 1990s. and enabling Firms: Significant R&D activities at three to four times the level of government R&D expenditure; focus conditions on increased competencies, especially for SMEs Markets and institutions: High export orientation, FDI intensity, and knowledge spillover; strong consumer protection mechanisms in place Research: Improved research quality in academia; modern R&D infrastructure; well-developed quality and standards infrastructure Skills: Readily available STEM- and other high-skilled workers Policy objectives • Catch up with high-income economies by developing high-tech industries • Shift government policies to support basic research and R&D activities in long-term projects and emerging technologies Changes in policy mix • National R&D projects targeting long-term public R&D needs • Creation of centers of excellence • Government funding to establish and operate corporate-affiliated research institutes • Deepening of export and investment promotion, including through promotion of outward FDI for firms seeking improved market access abroad 2000s–2010s: A mature NIS Innovation capacity Korea pursued a knowledge-based economy and equitable growth, recognizing availability of and enabling advanced technologies as well as rising income disparities. conditions Firms: Enhanced R&D capacity across firm size; emerging innovation clusters and increased collaboration in innovation projects across firms Markets and institutions: Increased access to early-stage finance; improvements needed in IPRs and developing standards for advanced technology companies Research: Weak links and cooperation among PROs, industry, academia, and international research institutes and firms Skills: Creativity and socioemotional skills desired among STEM- and other high-skilled workers Policy objectives • Transition from catch-up model to generation of domestic technological innovation through development of basic and original technologies • Expand support for SMEs, promote entrepreneurship, and create good-quality jobs

table continues next page Action for Innovation: A Policy Agenda 183

TABLE 6.3 Republic of Korea’s innovation policies continually evolved to reflect its accumulation of capabilities (continued)

Policy aspect Innovation capabilities, objectives, and priorities Changes in policy mix Increased focus on • Targeted support for firms with high growth potential, such as technology-extensive firms, start- ups, and medium-size enterprises • Government-backed venture capital (VC) to provide equity finance and early-stage capital • Business incubators and accelerators • Precommercial public procurement of R&D • Clusters and networking for R&D cooperation • Op en innovation

Source: Frias, Lee, and Shin 2020. Note: FDI = foreign direct investment; GVC = global value chain; IPRs = intellectual property rights; NIS = national innovation system; NQI = national quality infrastructure; PROs = public research organizations; R&D = research and development; S&T = science and technology; Skills = skills and human resources; SMEs = small and medium enterprises; STEM = science, technology, engineering, and mathematics; TA = technical assistance; VC = venture capital. a. Public procurement for innovation involves support to innovation through the purchase of R&D services and innovative products that have already reached the commercialization stage. For further details, see Cirera et al. (2020). revised its policies as its innovation and the allocation of public resources, which technological capabilities evolved. It updated do not align well with the need to support its policy priorities to reflect changing greater adoption and diffusion among firms challenges—from focusing on the building of in Indonesia, the Philippines, and Vietnam. basic innovation capabilities in the 1960s and Although all three countries have set a de jure 1970s; to maximizing links to GVCs, FDI, objective of promoting innovation and devel- and entry into export markets in the 1980s; oping new technologies, the de facto reality— to a significant focus on R&D and patenting expressed through resource allocation and the in the 2000s; and to technological leadership implementation of policy instruments—does in some sectors in the 2000s. not match some of the key objectives (1 to 4) These policy choices have paid off, as Korea discussed in table 6.1. For example, none of has converged to the technological frontier in these countries has well-developed manage- many sectors. Significant technological catch- ment and technology extension services that up—proxied by patents and R&D spending—is can assist firms in adopting new technologies. observed from 1990, which is then consoli- Moreover, PROs are performing their role dated in the 2000s (fig ure 6.1). Investment in poorly—often more focused on bringing their R&D, especially in the private sector, acceler- own-developed technologies to industry than ated in the early 1990s (figure 6.1, panel a) and on addressing firms’ specific technology needs.1 then again in the 2000s, at which point Korea, A more careful but long-term commit- like Israel, became one of the most R&D- ment to fostering innovation and technology intensive countries in the world. A similar pic- adoption is needed in the region, one that pro- ture is seen with patents (figure 6.1, panel b), vides services to build the needed innovation which accelerated in 1995 and then converged capabilities and, as discussed in chapter 2, to with the intensity of the United States in the create a business environment for innovation. 2000s, as Korea surpassed the Organisation for Economic Co-operation and Development Eliminating policy biases against (OECD) average. innovation in services In developing East Asia, however, there is a mismatch between overarching policy The analysis in previous chapters showed that objectives and capabilities. As highlighted although services are becoming increasingly in chapter 5, this is especially the case in important in the wave of new technologies, a 184 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

FIGURE 6.1 Republic of Korea’s policy choices have enabled significant technological catch-up

a. R&D investment, by source, 1980–2015 b. Number of patents, 1985–2015 50,000 0.08 45,000 40,000 0.06 35,000 30,000 25,000 0.04 20,000 15,000 0.02 10,000 Patents per 1,000 population Patents R&D spending (US$, millions) 5,000 0 0 1980 1985 1990 1995 2000 2005 2010 2015 1985 1990 1995 2000 2005 2010 2015 Year Year Public R&D Private R&D Foreign source R&D Korea, Rep. United States OECD

Source: Organisation for Economic Co-operation and Development (OECD) Innovation and Technology database (https://data.oecd.org/innovation-and-technology.htm). Note: R&D = research and development.

strong bias persists against services in innova- the implementation of innovation projects tion policy support. Traditionally, innovation such as the digitization of management and and technology development have been seen delivery processes or the introduction of new as primarily processes driven by the manufac- services. Second, it is necessary to expand turing and agriculture sectors. Networks of the scope of innovation activities that are PROs performing R&D activities have been financed to include activities such as design— set up throughout the region as well as glob- a significant component of R&D (Cox 1990) ally in narrowly defined areas of manufac- in manufacturing but also in services—and to turing and agriculture. The reality, however, strengthen firms’ digital capabilities. is that innovation in services is increasingly Services sectors are extremely diverse, and important for competitiveness in manufactur- innovation takes different forms across ser- ing and for the strengthening of GVCs as well vices subsectors. For example, digital and as for services themselves, which employ the artificial intelligence (AI) elements are more largest share of people in all the countries. important in routine services, whereas design, For example, international business transac- business models, and delivery are more tions depend on transport, logistics, and com- important in knowledge-intensive services munication networks. Innovations in these (Salter and Tether 2006). Recognizing these services are thus key to facilitating integration differences and designing policies that are of local firms into global networks. Moreover, aligned with different capabilities and needs improvements in digital infrastructure and dig- will be critical to enabling innovation in this ital networks and platforms are enabling the increasingly important sector of the economy. proliferation of innovative services firms. Yet innovation policy is still not focusing enough Choosing the right policy instruments on promoting innovation in services. to accelerate technology adoption and Eliminating this bias requires actions on diffusion two fronts: First, it is important to reach out to services and retail firms with more tradi- To support the choice of appropriate policy tional innovation policy instruments (for instruments to accelerate technology adop- example, matching grants) that can finance tion and diffusion, this section provides Action for Innovation: A Policy Agenda 185

a brief overview of five policy areas that and transfer), and others focus more on firms’ are critical for countries in the region: (a) capability to generate technologies, as follows enabling technology adoption and diffusion; (Cirera et al. 2020): (b) maximizing spillovers from MNE and trade; (c) developing appropriate intellectual • Business advisory services focus on property (IP) policies; (d) strengthening the building firms’ absorptive capacity for enabling environment by increasing market technology adoption. competition; and (e) building the infrastruc- • Technology extension services focus on ture for innovation and technology adoption. helping small and medium enterprises (SMEs) adopt technologies and develop Enabling technology adoption and related capabilities. These policy instru- diffusion ments are critical to supporting the diffu- A range of instruments can promote adop- sion of technologies among firms. tion and diffusion (left side of figure 6.2) as • Technology centers support both the adop- well as the generation, commercialization, tion and the generation of new technologies. and transfer of technology (right side of • Technology transfer offices support the figure 6.2). Several instruments focus more generation and commercialization of tech- directly on equipping firms with the capabili- nologies developed at universities and ties to use technologies (technology adoption PROs. In some cases, these offices also

FIGURE 6.2 A range of policy instruments can be used to promote technology adoption, diffusion, and invention

Building Supporting tech generation, absorptive capacity commercialization, and transfer

Technology Technology Technology adoption transfer generation

Universities Firms and research institutions

Technology Science and Technology Business Technology and technology transfer advisory extension services R&D centers parks o ces

Grants for Loans for (process) innovation Open innovation innovation (equipment)

Source: Cirera et al. 2020. ©World Bank. Further permission required for reuse. Note: The vertical dotted line separates the policy instruments supporting the demand (from firms) and supply (by universities and research institutions) of technologies and knowledge. Green boxes designate overarching objectives of technology policy, while the yellow boxes represent the three aspects of technology that the policy instruments (in dark blue boxes) support. R&D = research and development. 186 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

support researchers and entrepreneurs in However, it is unclear whether countries in addressing knowledge gaps in the com- developing East Asia have been able to maxi- mercialization process. mize the absorption of technology and capabilities from MNEs and GVCs. The evidence Other instruments that can facilitate adop- indicates that technology and knowledge tion and diffusion include science and technol- have not diffused widely across sectors and ogy (S&T) parks and government support for firms—suggesting that FDI and trade still innovation finance. In some cases, S&T parks remain large untapped sources of technologi- can attract technology-intensive firms with the cal capabilities in the region. An important objective of generating spillovers with local barrier to capturing the positive spillovers universities and industries. As for government- associated with MNEs and GVCs is local supported finance, external finance for equip- firms’ continued lack of capacity to adopt ment can greatly facilitate the purchase of new technologies and knowledge. technology, especially in places where there are Most countries in the region have opted for significant financial market inefficiencies. “soft” or “friendly” policies to support the A critical element of the model shown transfer of technologies from FDI and GVCs. in figure 6.2 is the set of complementari- These have focused primarily on incentives to ties across different policy instruments. The MNEs to establish R&D centers and, more effective transfer of technologies requires generally, to attract companies to locate in that firms have enough absorptive capac- host countries through tax incentives and ity to adopt new technologies. Similarly, the establishment of S&T parks. Some countries ability to discover and transfer new technolo- may also be tempted to adopt more-stringent gies depends on the capacity of some firms to forced technology transfer (FTT) measures, implement complex R&D projects as well as as China has done.2 But the reality of devel- on the quality and incentives in universities to oping East Asia is that countries are com- perform high-quality applied research. peting for the location of MNEs and R&D One challenge in the region is a frequent centers, and this competition could increase expectation that the transfer of technologies further if rapid technological change contin- from MNEs and PROs to domestic firms will ues to reduce low labor cost advantages in the happen more or less automatically. However, region.3 Introducing FTT measures, therefore, such technology transfer is unlikely to happen could undermine the attractiveness of MNE when insufficient instruments are in place to investments and thus eliminate the related build firms’ capabilities. Another widespread opportunities for technology transfer. expectation is that PROs will develop technologies that are useful for firms or indus- Developing appropriate intellectual tries even when the incentives facing research property policies institutions and enterprises are not aligned or A key factor in technology transfer is the when the quality of research is not sufficient to intellectual property (IP) framework. meet firms’ technology needs. Building capa- Historically, however, patent frameworks bilities among private firms, on one hand, and have been developed primarily to protect improving the quality and incentives within inventors and not to facilitate technol- the research sector, on the other, must therefore ogy transfer (although, clearly, having a be priorities for developing East Asian coun- functional IP system can facilitate transfer tries—all while facilitating invention among a through FDI and trade). A recent literature small share of more-advanced firms with their review finds that FDI and trade are both own research capabilities. positively correlated with the strength of intellectual property rights (IPR) enforce- Maximizing spillovers from MNEs and trade ment (Hall 2014). Less clear is the impact FDI and trade have been critical channels of a strong IP system on domestic innova- for the region’s growth and development. tion (Branstetter 2004). Action for Innovation: A Policy Agenda 187

An older literature, which examined the efficiency or through spillovers of technology successful cases of Japan and Korea, sug- and expertise. gested that weak IP protection supported the Evidence from OECD countries on the accumulation of technological capabilities relationship between market competition and as countries pursued technological catch- innovation paints a more nuanced picture, up (Kim 2003; Kumar 2003). This is also however. Aghion et al. (2005) observe, for reflected in the so-called “patent puzzle”: the example, that greater product market com- lack of correlation between patent activity petition between incumbent firms could have and productivity (Boldrin and Levine 2013). different effects that both discourage and Indeed, a recent review of innovation policies promote innovation. Their study finds strong concludes that the impact of IP policies on evidence of an “inverted-U curve” in multi- technology generation in high-income econ- industry panel data for UK firms. In indus- omies is unclear (Bloom, Van Reenen, and tries where competition is low and firms have Williams 2019). similar technological capability, more compe- In short, patents and the IP framework tition may promote innovation by giving the appear to be of relatively low importance for innovating firm a competitive advantage— inducing invention in middle-income coun- although in industries where there is already tries (Hall 2020), including in developing high product market competition and one East Asian countries. An IP framework that firm has a large technological lead over oth- facilitates FDI and complies with the World ers, greater competition may discourage inno- Trade Organization (WTO) Agreement vation among lagging firms. A study using on Trade-Related Aspects of Intellectual firm-level data from Chile finds a positive Property Rights (TRIPS Agreement) is needed relationship between markups and innova- for any hope of some transfer of technology tion in lagging firms, although this relation- (Branstetter, Fisman, and Foley 2006). To be ship is not significant in more-advanced effective, however, this framework needs to be firms (Cusolito, Garcia-Marin, and Maloney implemented jointly with the complementary 2018). The authors find, moreover, that the technology transfer policies discussed above. precise relationship between competition and innovation depends on the sector and type of Strengthening the enabling environment by firm. increasing market competition On balance, the literature suggests that Market competition is a key driver of innova- greater market competition can help enable tion. Firms are driven to be more efficient and innovation, especially in cases where competi- productive and to offer new and improved tion is low to start with. Increasing compe- products and services to customers in more tition could be particularly beneficial for the open and competitive markets. Thus, there services sector in developing East Asia, where is a role for competition policies to encour- market restrictions are greater than in OECD age innovation. In some developing East Asia countries (Constantinescu, Mattoo, and Ruta countries, this implies that the state should 2018). Without competitive pressure, firms move away from engaging directly in pro- have little incentive to innovate. Reducing ductive activities via state-owned enterprises restrictions in the services sector would help (SOEs) and focus on playing a more market- to promote greater competition and growth supportive role. A study by Brahmbhatt and of services. Hu (2010) shows a strong negative associa- More importantly, opening up the services tion between state ownership and innovation. sector to both foreign and domestic private The study also finds a positive association investment would also facilitate the “servici- between innovation and the extent of compe- fication” of manufacturing (that is, the inte- tition faced by firms. FDI, as well as domes- gration of services and manufacturing), which tic competition, may enhance productivity in is currently lower in developing East Asia the rest of the economy by increasing firms’ than in higher-income countries. In doing 188 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

so, opening up the services sector would also innovation (as discussed in chapter 4). serve to strengthen countries’ manufacturing Such advanced skills become increas- competitiveness. ingly important as firms move from diffusion and technology adoption toward Building the infrastructure for innovation the technological frontier. However, most and technology adoption education systems in the region still face A final important element is the need to important challenges in generating ade- develop adequate infrastructure to enable quate learning outcomes, even at the basic innovation. This includes infrastructure to education level. The region’s policy makers support universal digitization (for example, thus face a dual challenge of providing the through broadly available, affordable fixed necessary foundational skills to their pop- and mobile broadband) as well as infrastruc- ulations while also developing the types ture to ensure quality standards, such as a of advanced workforce skills needed to national quality infrastructure (NQI) for enable innovation-led growth. metrology and testing. Although countries Successfully building skills for innovation in the region have made progress in build- will require action on several fronts, including both digital and quality infrastructure, ing strengthening basic educational quality; further investments are needed. If digital updating national curricula to increase focus infrastructure is poor, achieving universal on such innovation skills as creative problem digitization will not be possible; similarly, solving and socioemotional skills; enhancing if firms lack access to testing labs and qual- the quality and relevance of technical train- ity certification, quality upgrading will be ing and tertiary education; and increasing impeded. opportunities for continuous skills upgrading among adult workers. Strengthening key complementary factors: Skills Building foundational skills will require and finance improvements in basic education For most countries in the region, build- The ability of firms to innovate depends on ing the advanced skills needed to enable multiple factors that often fall outside the innovation-led growth will require sustained realm of innovation policy, strictly defined. efforts to strengthen people’s foundational These factors include the availability of a skills. Building those skills will require rais- sufficiently skilled workforce and adequate ing education quality, starting at the basic financing to support firms’—often risky— level. innovation activities. This section focuses The experience of high-performing edu- on policy approaches to help build these cation systems in East Asia (for example, in complementary factors to promote greater Japan, Korea, and Singapore) and beyond innovation, whether defined as diffusion and suggests that building stronger founda- technology adoption or as invention. tional skills requires that school systems (a) strengthen the conditions for learning, including availability of educational mate- Developing skills for innovation rials; (b) improve teacher preparation and Policy makers in much of the region the quality of teaching; (c) ensure adequate face a dual challenge of strengthening public spending for basic education; (d) basic learning outcomes and fostering increase children’s readiness to learn, includ- advanced skills to support innovation ing through early childhood education Evidence from the region and from and development services; and (e) under- high-income economies highlights the take regular learning assessments to diag- importance of advanced cognitive, socio- nose challenges and inform improvements emotional, and technical skills in enabling (World Bank 2018). Action for Innovation: A Policy Agenda 189

Development of advanced cognitive Developing advanced technical skills for and socioemotional skills needs to innovation will require addressing both start early access and quality issues Recent studies on skills formation empha- Developing increasingly advanced technical size that building strong cognitive and socio- skills in the labor force is also a critical part of emotional skills is best begun early (Arias, strengthening workers’ skills for innovation. Evans, and Santos 2019; Cunningham and This is exemplified by the importance that Villaseñor 2016). Efforts to build such skills highly innovative firms in the region assign are still underdeveloped in much of the to hiring employees with science, technol- region, however. Even where developing East ogy, engineering, and mathematics (STEM) Asian countries have recognized the impor- education (chapter 4). Technical skills are tance of better cultivating critical thinking, often sector or discipline specific, and require creativity, problem solving, and the ability knowledge of specific tools or processes. And to work effectively in teams, there remains the focus on technical skills development a need to institutionalize the development of typically begins later, around the secondary advanced cognitive and socioemotional skills school level. into standard school curricula and extracur- Given the region’s significant heterogene- ricular programs. ity in firm capabilities and technology adop- Several high-income innovators in the tion, the needs for technical skills across firms region—Japan, Korea, and Singapore—have are also quite diverse. For firms focused on already done this, revising their curricula to diffusion and adoption of existing technolo- include emphasis on higher-order cognitive gies, basic digital literacy and the capacity to and socioemotional skills development. These use general purpose technologies and exist- adaptations to school curricula are part of a ing software applications may be sufficient. broader set of cross-cutting measures in these As firms move toward the technical frontier, countries to foster innovation skills among however, more sophisticated technical skills their workforces (table 6.4). are required.

TABLE 6.4 Japan, Republic of Korea, and Singapore have integrated innovation skills into their curricula

Country Strategies for building innovation skills Japan • Revision of national curriculum to focus on active learning strategies to strengthen problem solving, creativity, and critical thinking and to strengthen motivation to learn • Provision of online learning opportunities for tertiary education (for example, Cyber University, accredited by the Ministry of Education, Culture, Sports, Science and Technology) • Provision of technical and vocational education and training (TVET) through TVET institutions, upper secondary schools, and higher education institutions Korea, Rep. • Revision of national curriculum to focus on higher-order cognitive skills (such as information processing skills and creative thinking) • Creation of a mid- and long-term Master Plan for Vocational Education to address Industry 4.0-related challenges • Establishment of employment-oriented high schools and the Meister Vocational High School Program • Provision of online learning opportunities for tertiary education (for example, Open Cyber University, authorized by the Ministry of Education) and vocational training (for example, the Online Lifelong Education Institute, funded by the Korean Ministry of Employment and Labor) Singapore • Shift from traditional focus on academic performance to emphasis on socioemotional skills through “positive education” model • Provision of TVET through polytechnic schools offering work immersions (often incentivizing on-the-job training in small and medium enterprises) • Establishment of an independent training authority, the Institute of Technical Education, with close links to industry and a recognized certification system Sources: Kataoka and Alejo 2019, drawing from Banerji et al. 2010; Moore and Kearsley 2012; OECD 2018; World Bank 2015; and Yian and Park 2017. 190 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

In developing the technical skills needed for in the face of rapid technological change. innovation, most countries in the region still To do so, the region’s policy makers should face challenges related to both educational focus on developing continuous skills devel- access and quality. Tertiary enrollment rates opment—or lifelong learning—systems for are relatively low in much of the region (chap- adult workers. ter 4, figure 4.8), and the quality of techni- TVET can play an important role as part cal education remains highly variable among of countries’ broader skills development and tertiary institutions. Efforts to broaden access lifelong learning strategies. To be effective, to and raise the quality of TVET and tertiary however, programs should be designed and education will thus be increasingly impor- implemented in close coordination with the tant as countries seek to build the technical private sector, including via partnerships in skills necessary to innovate and as firms seek which leaders from private sector firms help to move progressively toward the technical design the curricula as well as deliver training. frontier. Indeed, studies show that returns to TVET Japan, Korea, and Singapore have all been and related trainings are higher when they working to strengthen their TVET systems are demand driven and well adapted to labor as part of their strategies to build innova- market needs (Kluve 2016). tion skills (as summarized in table 6.3). In Several recent studies also show that on- Japan, TVET is provided by upper secondary the-job training can contribute to greater and higher education institutions as well as firm-level innovation activity (ADB 2020; TVET institutions, and it commonly includes Iootty 2019; Miyamoto and Sarzosa 2020). both on-the-job and off-the-job training con- Although on-the-job training appears to be ducted within private enterprises (World Bank relatively rare among firms in the region, 2015). In Korea, the government has devel- there may be scope for governments to incen- oped a Master Plan for Vocational Education tivize training as a way to build greater skills to address lifelong learning challenges aris- innovation. One promising approach to ing from Industry 4.0. The plan emphasizes incentivizing skills development can be seen in skills like creativity, problem solving, and Singapore, where the government has created learning agility that are essential in the 21st- the SkillsFuture Council to promote lifelong century workplace (Chung 2019). Korea has learning. Under this initiative, Singaporeans also established the Meister Vocational High aged 25 and above receive an opening School Program, which offers training in spe- SkillsFuture Credit of US$500 that they can cialized skills needed by local industries, with use to take training courses. The credit does curricula developed in collaboration with not expire, is periodically topped up by the industry partners (Yian and Park 2017). And government, and can be accumulated over in Singapore, polytechnic schools offer a vari- time. ety of courses as well as “work attachments” with firms to equip learners with practi- Strengthening finance for innovation cal skills (Yian and Park 2017). To support TVET’s role in lifelong learning, independent Financing firm innovation requires that training authorities have been established countries have vibrant, well-diversified with close links to the private sector (Banerji financial markets that offer a broad range et al. 2010). of financial instruments to enterprises As discussed in chapter 4, both access to Lifelong learning systems will be important external finance and to a suitable range of to build innovation skills in the current financial instruments are important to enable workforce the financing of firm innovation. However, To strengthen firms’ abilities to innovate, it most developing East Asian countries still rely is also important to regularly upgrade the heavily on the banking system for external skills of the current workforce, especially finance. In addition, in countries with relatively Action for Innovation: A Policy Agenda 191

well-developed financial markets, such as in gradual steps, will depend on the level of China and Malaysia, it appears to be mostly development of a country’s financial sector. large firms that have benefited from the deep- Deep private markets require the creation ening of domestic capital markets (Abraham, of an ecosystem of investors (both private Cortina, and Schmukler 2019). and public) and of fledgling businesses with To ensure that their financial sectors viable projects (Lerner 2010). Stimulating can better support innovation, the region’s cooperation between industry and universi- countries will need to implement poli- ties helps to build a more robust stream of cies in three important areas: developing a innovative ideas that attract private investors. well-functioning capital market; promoting Moreover, it is important to attract experi- venture capital (VC) markets; and improving enced international fund managers, as seen in and broadening the range of available finan- the development of the VC market in Israel cial instruments through the banking sector. (box 6.1). In East Asia, Singapore has fostered innovation by encouraging research Well-developed capital markets provide and entrepreneurial activity in universities, firms with access to a diverse set of courting global venture capital funds to financial instruments, eventually increasing establish connections, and welcoming foreign the quantity and quality of firm innovation entrepreneurs. Entrepreneurial efforts have The development of deep capital markets is also been rewarded and recognized through paramount to offering alternative sources of national competitions and awards. China’s external capital to innovative firms at different multipronged policy approach has also been stages of a firm’s life cycle. Some countries in successful in spurring development of the the region have already made progress in this country’s VC market (box 6.1). area by introducing capital market reforms tar- The success of private capital market geted to increasing the investor base; improving development depends heavily on policy financial market infrastructure (for example, design. Governments can play an important introducing a capital market data warehouse role by investing directly or by participat- system); and enhancing investor protections ing in public-private partnerships. Successful (Abraham, Cortina, and Schmukler 2019). examples of cooperation between the public Nonetheless, access to these markets is still and private sectors are the Yozma program mostly available to relatively large firms. In in Israel (which, among other things, estab- many countries, financial markets still lack the lished limited partnership companies with a necessary depth to cover the financial needs 40 percent investment from the government)4 of a wide range of firms. Continued efforts to and the Small Business Innovation Research develop and deepen countries’ financial mar- (SBIR) program in the United States, which kets are thus necessary. involves partnership between SMEs and nonprofit research institutions (Audretsch, Link, Financial policies can enable firm and Scott 2019). There are less successful innovation by promoting VC funding examples, however, such as Canada’s creation There are three broad and complementary in the 1980s of labor-sponsored venture capi- dimensions for the development of successful tal corporations (LSVCCs) to invest primar- VC markets: (a) enhancing the VC supply by ily in SMEs; in later decades, public funding enabling domestic investment and attracting crowded out private investment (Cumming foreign capital; (b) stimulating demand by and MacIntosh 2006). building an active entrepreneurial and inno- A successful design is one in which pub- vative ecosystem; and (c) supporting all mar- lic support empowers private investors in ket players by strengthening the institutional channeling resources to innovative ventures. and regulatory framework (Owen and Mason Hence, a hybrid model where governments 2019). Whether the development of these play the role of arm’s-length investor, while conditions happens simultaneously, or occurs allowing private fund managers to make 192 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

BOX 6.1 Innovation financing done right: Lessons from Israel and China

Israel’s phased approach to building total capital, the number of foreign investment banks VC foundations increased from 1 to 26, and the firms backed by VC funds totaled more than 2,000. Moreover, the pro- Israel has fostered venture capital (VC) funding over gram led to the growth of the VC market beyond its time. The background conditions phase (1970–89) scope, as it attracted many follow-on funds that were and the preemergence phase (1989–92) demonstrate not funded by the program itself. the importance of developing solid underlying structures that can support and facilitate the later stages. China’s three-dimensional approach to During the background conditions phase, Israel VC market development tried to shift research and design (R&D) from the China’s gradual, three-dimensional approach exem- military sector toward the academic and business sec- plifies a successful VC market development policy tors. To achieve this goal, it relied heavily on links with from developing East Asia (Lin 2017). The govern- American experts and researchers who had the most ment introduced policy reforms to attract capital, experience with the development of private capital established a new investment vehicle (the limited markets to finance high-tech projects. The government partnership), and fostered entrepreneurship. In terms undertook policies to ensure stabilization and liberaliza- of capital, China has introduced market liberalization of capital markets. To attract foreign investment, tion reforms that have allowed more institutional policies were put in place allowing full repatriation of investors to enter the VC market. The government earnings from financial investments. These policies cre- has also taken a market approach to public fund- ated rife opportunities for innovation projects, gener- ing, implementing matching programs between ating demand for capital and a sprawling pipeline of government and private funds. And recognizing projects for the nascent VC industry to invest in. the importance of foreign capital and knowledge, During the preemergence phase, policy was more it has introduced capital liberalization and foreign- directly aimed at the creation of a VC market to investor-friendly business environment reforms. address financing and management failures. The On the fiscal side, China provided tax exemptions first formal attempt, a government-backed insurance for limited partnerships and tax deductions to firms company for publicly traded funds called Inbal, was operating in science parks and incubators. Regarding mired by bureaucracy and micromanagement, how- financial markets, the government has invested in the ever, and was not successful. development of stock markets (NEQ and Chinext) The emergence phase (1993–2000) saw the intro- and lowered listing requirements to allow VC-backed duction of Yozma, a government program that com- firms to exit successfully. To build a strong ecosys- bined several important aspects of a successful VC tem of VCs and innovative firms, it has introduced market design. It served as a fund of funds—that is, reforms to streamline business operations. it provided funds to VC firms, which in turn pro- China’s approach has been clearly policy led and vided financing to domestic firms with a technologi- is a good example of the balance that the govern- cal base. Program funds had to be matched by the ment needs to strike—providing support but not entrepreneur, and local VC firms had to attract and overreaching its scope. Hence, there are lessons to be pair with a foreign investor. Interactions with foreign learned from its approach. In fact, other countries in capital investors were important sources of learn- developing East Asia are already introducing similar ing and expertise while also serving as a check on reforms. For example, Vietnam is introducing a com- the performance of local VCs. The program was a pany law reform to make business incorporation and great success: VC firms raised about US$10 billion in operations easier (Lin 2017).

investment decisions, can be successful. participation ensures that funds are allocated This type of intervention need not be overly to the most promising ventures. Regular costly, because private investors provide a impact evaluations and monitoring are share of the funding. Government funds important to making sure that programs are can identify and support market segments effectively and efficiently tackling the associ- with financing gaps, while private sector ated market failures. Action for Innovation: A Policy Agenda 193

Government interventions can leverage Investing in institutional capacity bank-firm relationships and enable firm To date, discussions of innovation and tech- innovation by targeting underserved or nology policies have commonly ignored the strategic segments As noted earlier, bank loans remain the main capacity of countries to effectively design source of external finance for innovative and implement innovation policy, but this is firms in the region. Governments can take critical for the effectiveness of interventions. advantage of this by channeling financing to Good project management also affects the innovative firms through the banking sector, quality of the outcomes. Nonetheless, several thereby increasing availability and lower- countries in developing East Asia still do not ing the cost of risk financing. One benefit of apply best practice in public management to this approach is that it can take advantage of innovation policy. existing bank-firm relationships. Some of the most significant shortcomings An alternative to direct loans for innovation relate to the lack of an adequate economic is credit guarantee schemes.5 This instrument justification, the absence of a logical frame- may be more efficient than direct subsidized work to develop interventions, and a lack of loans because it makes use of existing lend- monitoring and evaluation (M&E) mecha- ing products, allowing banks to select projects nisms. Recall the example in chapter 5 of a while maintaining the incentives to monitor program aiming at sectorwide productivity borrowers’ behavior. Loan guarantees also growth but supporting only a handful of firms have less of an impact on government bud- with few resources. A clear logical framework gets because credit schemes are funded by the would have helped policy makers see that the intermediaries. As with loans, however, guar- program was too small to achieve a meaning- antees should be targeted at credit-constrained ful impact. Weaknesses in the identification firms in later innovation stages and should not and selection of innovation policies are also be considered a substitute for private capital often reinforced during their implementation. markets. Moreover, careful implementation For most policy instruments analyzed for this is critical to ensuring that by providing guar- report, no resources were allocated to evalu- antees, the government does not incentivize ate whether programs’ intended outcomes banks’ moral hazard behavior or disincentivize were achieved. rigorous screening and monitoring procedures These capacity bottlenecks need to be (Cirera et al. 2020). addressed, and to do so it will be critical to invest in capacity for policy making. Agencies in the region need to recruit capable Reforming innovation staff, as well as provide adequate training for institutions and agencies and staff working on innovation policy. In addition building their capacity to on-site training in public management, it is As highlighted in chapter 5, innovation agen- important to ensure that managers have ade- cies and knowledge-creation institutions in quate digital infrastructure to perform neces- the region experience important governance sary management tasks, such as monitoring gaps and commonly lack incentives to per- beneficiaries and registering project outcomes. form their role in fostering firm-level innovation. Moreover, many of these institutions Building more professionalized have large gaps in their capacity to design and innovation agencies and increasing implement innovation policies. It will therefore interagency coordination be important for countries to address critical governance challenges in these institutions and Agencies supporting innovation policy in invest in their capacity to make policy if devel- the region use outdated governance models oping East Asian countries are to transition to and lack coordination across entities, which more innovation-led growth models. undermines policy alignment. Innovation 194 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

policy, because of its cross-cutting nature, by a small number of firms and to ensure the requires coordination among agencies or alignment of objectives and the use of effi- ministries. The current lack of coordination cient solutions. For example, private sector results in significantly fragmented efforts participation on boards of trustees of research along with policies that are poorly designed institutions can help guide PROs toward and executed. Innovation agencies in the greater industry collaboration. More impor- region need to (a) ensure better coordination tantly, some market failures in innovation are across ministries and institutions responsible a function of coordination failures that can for innovation policy, and (b) adopt new be addressed with broader mobilization of agency models that enable recruitment of suf- the private sector (rather than through more ficient talent and professionalized services. extensive use of public resources). There are several models for coordination. Some coordination mechanisms have Strengthening the governance and been elevated to prime ministers’ offices to incentive structures of PROs and ensure mobilization of resources and actors. research centers Other models have been integrated into or led by economics ministries that oversee all eco- The region’s governments have been working nomic areas;6 still others are located in edu- to strengthen their national research capac- cation or science ministries. A weakness of ity and have increased S&T investments in the latter model is that education and science public research institutions to create new ministries tend to experience more difficulties knowledge. The results of these efforts—and in coordinating other ministries and agencies the impact of PROs on innovation—remain because they are less connected to the relevant unclear, however. budgetary processes. With this in mind, each As discussed in chapter 5, survey findings country needs to assess which coordination for three countries in the region (Indonesia, the model best fits its institutional realities. Even Philippines, and Vietnam) suggest that PROs if carried out by ad hoc coordination working and university research departments have few groups, effective coordination is a necessary links with industry (including knowledge links condition for more integrated, focused, and and human capital interactions). Moreover, effective innovation policy. technology transfer activities are still embryonic and are largely concentrated among a small number of research organizations. Increasing the private sector’s National research policies, despite being con- contribution to innovation policy tinuously updated, still follow a model of inno- making vation that is heavily supply oriented. This is reflected in limited interaction with industry— Innovation policy aims to foster innovation in both governance and research activities— and the diffusion of technology mainly in and in few technology transfer activities. the private sector. However, private sector To maximize the contributions of these stakeholders are severely underrepresented research institutions to innovation, govern- in the identification, design, governance, and ments must undertake reforms in four critical monitoring of such policies in the region. This areas: often results in a misalignment between policies and private sector needs, as shown in the • Improving governance conditions, public earlier discussion of PROs (chapter 5). management practices, and strategic plan- Innovation support programs are often ning capacity (in education, research, and captured by a small number of firms—and technology transfer), including through often the same ones time and again. It is thus measures to increase the autonomy of critical to have broad private sector represen- research institutions, disseminate good tation in the formulation and implementation practices in research management, create of innovation policies to avoid policy capture clear legal mandates to support technology Action for Innovation: A Policy Agenda 195

transfer, and strengthen the links between 40 years, and China is now achieving techno- institutional funding policies and the per- logical leadership in some sectors. formance of institutions To accomplish this transformation, devel- • Improving academic incentives as engines oping East Asian countries must update their of change by increasing the recognition of objectives, giving greater priority to innova- technology transfer activities in the perfor- tion policies; focus on technological diffu- mance evaluation and career development sion; and incentivize more firms to innovate. of researchers, and by clarifying policy This effort requires conducive regulatory incentive frameworks and IPR regulations frameworks but also the right policies— to strengthen researchers’ engagement in those aligned with the innovation capabili- technology transfer ties of their private sectors. The COVID-19 • Adopting mission-oriented policies though pandemic has highlighted more than ever the budget allocations to address key societal urgency of achieving this transformation. challenges, such as those associated with COVID-19 and climate change Notes • Strengthening the impact of public research by (a) incentivizing PROs to provide tech- 1. Although Vietnam is dedicating significant fiscal nology extension and upgrading support resources to facilitate R&D spillovers from services to firms (such as by offering pilot- MNEs, there is no evidence that such spillovers ing and testing services for new technolo- are being realized (World Bank 2017). gies); and (b) enhancing the links between 2. FTT policies aim to increase foreign-to- domestic technology transfer from FDI firms PROs and universities or research centers that tend to weaken appropriability of foreign to support new technology-based entrepre- innovations and related intellectual property neurship, licensing of new technologies to (Prud’homme et al. 2018). Such policies are SMEs, and start-up creation. often implemented through requirements directed to investors and can include transfer of patented technologies to local companies or the forced location of R&D centers. Final remarks 3. Prud’homme et al. (2018) analyze more Over the past several decades, developing East formally the conditions required for effective Asia has achieved unprecedented growth that technology transfer under FTT policies. has lifted millions of people out of pov erty They establish seven preconditions based on and created unprecedented levels of economic complementary policies along with monitoring and enforcement abilities that countries in security. Nevertheless, this report has argued developing East Asia are unlikely to possess. that the region now faces an array of pressing 4. For a detailed explanation of Israel’s Yozma challenges and that its extraordinary develop- program, see Avnimelech, Schwartz, and ment performance is under threat if countries Bar-El (2007). do not transform their growth model to place 5. An example of a successful credit guarantee innovation at the forefront. scheme is the Korea Technology Finance The impact of the COVID-19 pandemic Corporation (KOTEC), which provides has been severe, and as the region’s policy guarantees for small and young firms in makers turn their attention to recovery, there high-tech sectors. An impact evaluation is an extraordinary opportunity to move for- study found that the program had been ward with pending reforms to accelerate the effective in increasing R&D investment for participants, especially for young high-tech process of technological catch-up and move firms (Heshmati 2013). Moreover, it led to toward a more innovation-led growth model. higher sales and productivity. Taking advantage of this opportunity is not 6. The best-known case is the Israel Innovation without precedent in the East Asia region; Authority (formerly Office of the Chief Japan, Korea, and later Singapore have Scientist in Israel), which is part of the become technological leaders over the past Ministry of Economy. 196 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

References of International Economic Law 7 (2): 359–70. Abraham, Facundo, Juan J. Cortina, and Sergio Branstetter, Lee G., Raymond Fisman, and C. L. Schmukler. 2019. “The Rise of Domestic Fritz Foley. 2006. “Do Stronger Intellectual Capital Markets for Corporate Financing.” Property Rights Increase International Policy Research Working Paper 8844, World Technology Transfer? Empirical Evidence Bank, Washington, DC. from U.S. Firm-Level Panel Data.” ADB (Asian Development Bank). 2020. Asian Quarterly Journal of Economics 121 (1): Development Outlook 2020: What Drives 321–49. Innovation in Asia? Semiannual report, April Chung, Jisun. 2019. “Lifelong Learning 2020. Manila: ADB. and Skills Development: South Korea’s Aghion, Philippe, Nick Bloom, Richard Blundell, Response to the 4th Industrial Revolution.” Rachel Griffith, and Peter Howitt. 2005. Presentation at the World Bank Group “Competition and Innovation: An Inverted-U Korea Office’s webinar and brown bag Relationship.” Quarterly Journal of lunch (BBL), “Bridging Skills Gap by Economics 120 (2): 701–28. Lifelong Learning, Advanced Technology Arias, Omar, David K. Evans, and Indhira Santos. and Skills,” September 10. https://olc 2019. The Skills Balancing Act in Sub-Saharan .worldbank.org/content/5-korea-office Africa: Investing in Skills for Productivity, -bbl-bridging-skills-gap-lifelong-learning Inclusivity, and Adaptability. Africa -advanced-technology-and-skills. Development Forum Series. Washington, DC: Cirera, Xavier, Jaime Frias, Justin Hill, and World Bank. Yanchao Li. 2020. A Practitioner’s Guide to Audretsch, David B., Albert N. Link, and John Innovation Policy: Instruments to Build Firm T. Scott. 2019. “Public/Private Technology Capabilities and Accelerate Technological Partnerships: Evaluating SBIR-Supported Catch-Up in Developing Countries. Research.” In The Social Value of New Washington, DC: World Bank. Technology, edited by Albert N. Link and John Constantinescu, Ileana Cristina, Aaditya Mattoo, T. Scott, 264–78. Cheltenham, UK: Edward and Michele Ruta. 2018. “Trade in Developing Elgar. East Asia: How It Has Changed and Why It Avnimelech, Gil, Dafna Schwartz, and Raphael Matters.” East Asian Economic Review 22 (4): Bar-El. 2007. “Entrepreneurial High-Tech 427–65. Cluster Development: Israel’s Experience Cox, Peter J. 1990. “Research and Development— with Venture Capital and Technological or Research Design and Development?” Incubators.” European Planning Studies International Journal of Project Management 15 (9): 1181–98. 8 (3): 144–50. Banerji, Arup, Wendy Cunningham, Ariel Cumming, Douglas J., and Jeffrey G. MacIntosh. Fiszbein, Elizabeth King, Harry Patrinos, 2006. “Crowding Out Private Equity: David Robalino, and Jee-Peng Tan. 2010. Canadian Evidence.” Journal of Business “Stepping Up Skills for More Jobs and Higher Venturing 21 (5): 569–609. doi:10.1016/j Productivity.” Report No. 55566, World Bank, .jbusvent.2005.06.002. Washington, DC. Cunningham, Wendy V., and Paula Villaseñor. Bloom, Nicholas, John Van Reenen, and Heidi 2016. “Employer Voices, Employer Demands, Williams. 2019. “A Toolkit of Policies to and Implications for Public Skills Development Promote Innovation.” Journal of Economic Policy Connecting the Labor and Education Perspectives 33 (3): 163–84. Sectors.” World Bank Research Observer Boldrin, Michele, and David K. Levine. 2013. “The 31 (1): 102–34. Case against Patents.” Journal of Economic Cusolito, Ana Paula, Alvaro Garcia- Perspectives 27 (1): 3–22. Marin, and William F. Maloney. 2018. Brahmbhatt, Milan, and Albert Hu. 2010. “Ideas “Competition, Innovation and Within- and Innovation in East Asia.” The World Bank Plant Productivity: Evidence from Chilean Research Observer 25 (2): 177–207. Plants.” Background paper for Productivity Branstetter, Lee G. 2004. “Do Stronger Patents Revisited: Shifting Paradigms in Analysis and Induce More Local Innovation?” Journal Policy. Washington, DC: World Bank. Action for Innovation: A Policy Agenda 197

Frias, Jaime, Jin Lee, and Kyeyoung Shin. Lerner, Josh. 2010. “The Future of Public Efforts 2020. “Innovation Policy Learning from to Boost Entrepreneurship and Venture Korea: Lessons for Design and Execution of Capital.” Small Business Economics 35 (3): Innovation Policies in Emerging Economies.” 255–64. doi:10.1007/s11187-010-9298-z. World Bank, Washington, DC. Lin, Lin. 2017. “Engineering a Venture Capital Garcia, Andres F., Asya Akhlaque, Xavier Market: Lessons from China.” Columbia Cirera, Jaime Frias, Cha Crisostomo, Anne Journal of Asian Law 30 (1): 160–220. Lopez, Kalyah Alaina Ford, Reinaluz Ona, doi:10.2139/ssrn.2643311. and Yvette Camba, and 2019. “Philippines: Miyamoto, Koji, and Miguel Sarzosa. 2020. Assessing the Effectiveness of MSME “Workforce Skills and Firm Innovation: and Entrepreneurship Support”. Report, Evidence from an Employer-Employee Linked World Bank, Washington, DC. Survey Data in Vietnam.” Unpublished Hall, Bronwyn H. 2014. “Does Patent Protection manuscript. World Bank, Washington, DC. Help or Hinder Technology Transfer?” Moore, Michael, and Greg Kearsley. 2012. In Intellectual Property for Economic Distance Education: A Systems View of Online Development, edited by Sanghoon Ahn, Learning. 3rd ed. Belmont, CA: Wadsworth Bronwyn H. Hall, and Keun Lee, 11–32. Cengage Learning. Cheltenham, UK: Edward Elgar. OECD (Organisation for Economic Co-operation Hall, Bronwyn H. 2020. “Patents, Innovation, and Development). 2018. Education Policy and Development.” NBER Working Paper in Japan: Building Bridges Towards 2030. No. 27203, National Bureau of Economic Reviews of National Policies for Education, Research, Cambridge, MA. Paris: OECD. https://www.oecd.org/education Heshmati, Almas. 2013. “The Effect of Credit /education-policy-in-japan-9789264302402 Guarantees on R&D Investment of SMEs in -en.htm. Korea.” Discussion Papers 7851, Institute of Owen, Robyn, and Colin Mason. 2019. “Emerging Labor Economics (IZA), Bonn. Trends in Government Venture Capital Policies Iootty, Mariana. 2019. “Assessing Innovation in Smaller Peripheral Economies: Lessons from Patterns and Constraints in Developing East Finland, New Zealand, and Estonia.” Strategic Asia: An Introductory Analysis.” Policy Change 28 (1): 83–93. Research Working Paper 8706, World Bank, Prud’homme, Dan, Max von Zedtwitz, Joachim Washington, DC. Jan Thraen, and Martin Bader. 2018. “‘Forced Kataoka, Sachiko, and Ana Alejo. 2019. “Skills for Technology Transfer’ Policies: Workings Innovation.” Background paper for this report, in China and Strategic Implications.” World Bank, Manila, Philippines. Technological Forecasting and Social Change Kim, Linsu. 2003. “Technology Transfer and 134: 150–68. Intellectual Property Rights: The Korean Salter, Ammon J., and Bruce S. Tether. 2006. Experience.” UNCTAD-ICTSD Issue Paper No. 2, International Centre for Trade and “Innovation in Services: Through the Sustainable Development, Geneva. Looking Glass of Innovation Studies.” Kluve, Jochen. 2016. “A Review of the Effectiveness Background paper for the Advanced Institute of Active Labour Market Programmes with a of Management (AIM) Grand Challenge on Focus on Latin America and the Caribbean.” Service Science, April 7, Oxford University. Working Paper No. 9, Research Department, World Bank. 2015. “Japan—National International Labour Office, Geneva. Qualifications Framework Summary.” Kumar, Nagesh. 2003. “Intellectual Property Rights, Working paper, Report No. 120585, World Technology and Economic Development: Bank, Washington, DC. Experiences of Asian Countries.” Economic and World Bank. 2017. “Vietnam: Enhancing Enterprise Political Weekly 38 (3): 209–26. Competitiveness and SME Linkages: Lessons Kuriakose, Smita, and Haris Tiew. 2020. from International and National Experience.” “Assessing the Effectiveness of Public Research Working paper, World Bank, Washington, DC. Institutions in Fostering Knowledge Linkages World Bank. 2018. Growing Smarter: Learning & and Transferring Technology in Malaysia.” Equitable Development in East Asia Pacific. Report, Malaysian Development Experience East Asia and Pacific Regional Reports. Series, World Bank, Washington, DC. Washington, DC: World Bank. 198 THE INNOVATION IMPERATIVE FOR DEVELOPING EAST ASIA

World Bank. 2020. “Vietnam: Science, Technology New Drivers of Growth. Washington, DC: and Innovation Report: Embracing World Bank. Development Opportunities through Yian, Theresa Thang Tze, and Jonghwi Park. 2017. Innovation and Technology Diffusion.” World Beyond Access: ICT-Enhanced Innovative Bank, Washington, DC. Pedagogy in TVET in the Asia-Pacific. Paris: World Bank and DRC (Development Research United Nations Educational, Scientific and Center of the State Council, The People’s Cultural Organization (UNESCO) and Republic of China). 2019. Innovative China: UNESCO Bangkok Office.

ECO-AUDIT Environmental Benefits Statement

The World Bank Group is committed to reducing its environmental footprint. In support of this commitment, we leverage electronic publishing options and print-on-demand technology, which is located in regional hubs worldwide. Together, these initiatives enable print runs to be lowered and shipping distances decreased, resulting in reduced paper consumption, chemical use, greenhouse gas emissions, and waste. We follow the recommended standards for paper use set by the Green Press Initiative. The majority of our books are printed on Forest Stewardship Council (FSC)–certified paper, with nearly all containing 50–100 percent recycled content. The recycled fiber in our book paper is either unbleached or bleached using totally chlorine-free (TCF), processed chlorine–free (PCF), or enhanced elemental chlorine–free (EECF) processes. More information about the Bank’s environmental philosophy can be found at http://www.worldbank.org/corporateresponsibility. The Innovation Imperative for Developing East Asia The Innovation Imperative fter a half century of transformative economic progress that moved hundreds of millions of people out of poverty, countries in developing for Developing East Asia AEast Asia are facing an array of challenges to their future development. Slowed productivity growth, increased fragility of the global trading system, and rapid changes in technology are all threatening export-oriented, labor- intensive manufacturing—the region’s engine of growth. Significant global challenges—such as climate change and the COVID-19 pandemic—are exacerbating economic vulnerability. These developments raise questions about whether the region’s past model of development can continue to deliver rapid growth and poverty reduction.

ISBN 978-1-4648-1606-2 WORLD BANK EAST ASIA AND PACIFIC 90000 REGIONAL REPORT

9781464 816062 SKU 211606