English by Sam Ashworth

Home , Cuyo (Argentina), Geography of Argentina, Jujuy Province, La Quiaca, Provinces of Argentina, Purmamarca

TERRITORIAL DEVELOPMENT IN ARGENTINA: DIAGNOSING KEY BOTTLENECKS | 1

Territorial development in Argentina: diagnosing key bottlenecks as the first step toward effective policy

VOLUME I

ARGENTINA

Territorial development in Argentina: diagnosing key bottlenecks as the first step toward effective policy © 2020 International Bank for Reconstruction and Development / The World Bank 1818 H Street NW Washington DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org

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. Rights and Permissions The material in this work is subject to copyright. Because The World Bank encour- ages dissemination of its knowledge, this work may be reproduced, in whole or in part, for noncommercial purposes as long as full attribution to this work is given.

Attribution— World Bank. 2020. Territorial development in Argentina: diagnosing key bottlenecks as the first step toward effective policy. Washington, DC: World Bank.

Any queries on rights and licenses, including subsidiary rights, should be addressed to World Bank Publications, The World Bank Group, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2625; e-mail: [email protected]

Cover photo: © Isabel de Lara. Further permission required for reuse. Cover design: © Isabel de Lara. Further permission required for reuse. TERRITORIAL DEVELOPMENT IN ARGENTINA: DIAGNOSING KEY BOTTLENECKS | V

CONTENTS

TERRITORIAL DEVELOPMENT IN ARGENTINA: diagnosing key bottlenecks as the first step toward effective policy I

ACKNOWLEDGEMENTS VIII

ABBREVIATIONS IX EXECUTIVE SUMMARY 1

REFERENCES 12

CHAPTER I A framework to analyze territorial development: scale, specialization, and convergence 13

REFERENCES 20 CHAPTER 2 Scale, specialization, and convergence: where does Argentina stand? 21

LOW SCALE IN ARGENTINA 22 low firm density, with little growth of firms LOW SPECIALIZATION IN ARGENTINA 33 lack of integration between local markets LACK OF CONVERGENCE IN ARGENTINA 49 some progress is visible, but gaps still exist

REFERENCES 56 CHAPTER 3 Scale, Specialization, and Convergence: a closer look at the Provinces of Salta and Jujuy 58

LOW SCALE IN SALTA AND JUJUY 59 firms’ low growth and productivity LIMITED SPECIALIZATION IN SALTA AND JUJUY 67 public sector services dominance and weak connectivity LITTLE CONVERGENCE IN SALTA AND JUJUY 78 low-quality and inaccessible basic services, and low human capital

REFERENCES 83 VI | CONTENTS

CHAPTER 4 Summary: scale, specialization, and convergence in a nutshell 86

ANNEXES ANNEX A. Proxy indicators 91 ANNEX B. Important drivers of recent and future employment (by economic sector) 98

VIII |

ACKNOWLEDGEMENTS

The Territorial development in Argentina: diagnosing key bottlenecks as the first step toward effective policy report was prepared by a team led by Nancy Lozano-Gracia (Senior Economist GSU10– co-Task Team Leader) and Horacio Cristian Terraza (Lead Urban Specialist GSU10, co-Task Team Leader), and consisting of Beatriz Eraso Puig (Urban Development Specialist, GSU10), Veronica I. Raffo (Senior Infrastructure Specialist, GTD04), Somik Lall (Global Lead, Territorial Development), Nathalie Picarelli (Young Professional), Jessica Grisanti (Consultant, GSU10), Diana Cristina Tello Medina (Consultant, GSU10), Sally Murray (Consultant, GSU19), and Sebastian Anapolsky (Consultant). Marcela Nandllely Gonzalez (Program Assistant, GSU10) provided overall administrative assistance.

The team is grateful for peer review input from Ellen Hamilton (Lead Urban Specialist, GSU11), Ayah Mahgoub (Senior Urban Development Specialist, GSU11), and Javier Sanchez-Reaza (Senior Economist, GPSJB).

The team is thankful for the additional contributions provided by Pablo Fajgelbaum (Associate Professor of Economics, UCLA), Olivia D’Aoust (Urban Economist, GSU19), Alice Duhaut (Economist, DECIE), Gilles Duranton (Dean’s Chair in Real Estate Professor, Wharton Real Estate Department), Steven Farji Weiss (Economist, GFCAE), Daniel Gomez Gaviria (Senior Economist, GFCLC), Katie L. McWilliams (IT Officer, Data and Information Management, ITSOP), Diana Van Patten (Research Assistant, UCLA), and Anthony Venables (BP Professor of Economics, Oxford).

The team benefited from the guidance from Jesko Hentschel (former Country Director, LCC7C) and Carole Megevand (former Program Leader, LCC7C). Valuable comments and guidance were provided by Emily Sinnott (Program Leader, LCC7C), Cristian Quijada Torres (Senior Private Sector Specialist, GTC04), and Catherine Lynch (Senior Urban Specialist, GSU10). The report was prepared under the overall guidance of Ming Zhang (Practice Manager, GSU10).

The report was edited in English by Sam Ashworth. The book’s design is credited to Isabel De Lara. The policy discussion in this report has benefited from discussions during 2017-2019, with high- level government officials, including representatives from: Ministry of the Interior, Public Works and Housing of Argentina, Ministry of Production of Argentina, Undersecretary of Municipal Relations of Argentina, Secretariat of Productive Integration, Province of Salta, Province of Jujuy, Municipality of Salta, Municipality of Jujuy, Observatory of Employment and Business Dynamics (Ministry of Labor), Argentine Government social insurance agency (ANSES-Jujuy), and the Economic Social Council of the province of Salta. TERRITORIAL DEVELOPMENT IN ARGENTINA: DIAGNOSING KEY BOTTLENECKS | IX

ABBREVIATIONS

AMBA Buenos Aires Metropolitan Area CABA Autonomous City of Buenos Aires CNPV National Population and Housing Census (Censo Nacional de Población y Vivienda) DB World Bank’s Doing Business Report EPH Permanent Household Survey (Encuesta Permanente de Hogares) GCI World Economic Forum’s Global Competitiveness Index GDP Gross Domestic Product GGP Gross Geographical Product GHSL Global Human Settlement Layer GoA Government of Argentina GVA Gross Value Added IADB Inter-American Development Bank ICT Information and Communication Technologies INDEC National Institute of Statistics and Censuses KSI Krugman’s Specialization Index LISA Local Indicator of Spatial Association AELs Local Economic Areas LQ Location Quotients MAS Metropolitan Area of Salta NBI Unmet Basic Needs (Necesidades Básicas Insatisfechas) NEA Northeast region NOA Northwest region OECD Organisation for Economic Co-operation and Development OEDE Observatory of Employment and Business Dynamics (Observatorio de Empleo y Dinamica Empresarial) ONDAT National Observatory of Transport Data (Observatorio Nacional de Datos de Transporte) OSM Open Street Map PBA Province of Buenos Aires PET Strategic Territorial Plan (Plan Estratégico Territorial) SAME CABA’s Emergency Medical Services SMEs Small and Medium-Sized Enterprises

TERRITORIAL DEVELOPMENT IN ARGENTINA: DIAGNOSING KEY BOTTLENECKS | 1 EXECUTIVE SUMMARY

CONO DE ARITA IN SALTA, ARGENTINA Sebastian del Val © 2 | EXECUTIVE SUMMARY

EXECUTIVE SUMMARY

Argentina’s population and economic activity geographical disparities remain; for example, is highly concentrated in few places, similar provinces in the North of Argentina continue to to global trends. Globally, economic develop- struggle with extreme poverty and considerable ment has historically been concentrated in a few gaps in living standards when compared to the places: In fact, only 1.5 percent of the planet’s rest of the country. While overall poverty rates land is home to half of its production. With a have been falling since 2002, poverty incidence in population of 40.12 million, Argentina’s popula- the Northwest and Northeast regions are at 25.4 tion is highly concentrated in its capital (INDEC, and 33.2 percent respectively and are the highest 2010). According to the 2010 Census, 66 percent in Argentina. Although regional gaps in access to of the population is concentrated in Buenos Aires water and primary education have been impro- Metropolitan Area and the nearby Pampeana ving, the Northwest still lags in terms of access to Region. Further, 91 percent of the country’s popu- sanitation, secondary education, and human deve- lation lives in urban areas,1 and 70 percent of the lopment outcomes. total urban population are concentrated in just 31 Given that Argentina had a primary deficit of cities. Recent work suggests that those high levels 4.2 percent of GDP in 2017 and inflation of 24.8 of concentration have persisted, with the United percent at end-2017, prioritizing investments Nations estimating a 91.8 urban population rate will be necessary also in the context of territorial in the country in 2015. Economic activities in development. How can the government reduce Argentina are also geographically concentrated in development gaps across the national terri- the central and coastal areas, with two-thirds of tory while at the same time supporting growth the national GDP being produced in two regions: opportunities within a context of national Buenos Aires Metropolitan Area (AMBA, per its fiscal deficit? Using a territorial development name in Spanish) and Pampeana (Muzzini et al.), lens that allows the identification of challenges 2016). and opportunities at the sub-national level, this But unlike countries like South Korea, the report provides a framework and diagnostics to concentration of economic activity has not understand Argentina through three dimensions: been balanced by successful efforts to improve Scale, Specialization, and Convergence. Chapter 1 living standards across the country. Despite explains the territorial development framework a sustainable overall poverty reduction, today used for the analysis. Chapter 2 provides an over- 40 percent of Argentina’s population is still view of the economic geography of Argentina and vulnerable to falling into poverty, and large the challenges the country faces along these three dimensions. Chapter 3 presents a closer look at two 1 All localities with a population of 2,000 or more are considered urban in Argentina. TERRITORIAL DEVELOPMENT IN ARGENTINA: DIAGNOSING KEY BOTTLENECKS | 3

provinces, Salta and Jujuy, and puts them under the But economic theory and international expe- same lens. Chapter 4 summarizes the key messages rience suggest that spatial concentration is of the report, providing benchmarking to compare inevitable and desirable -successful places have Argentina to other countries around the world in managed to extend the benefits brought by such Scale, Specialization, and Convergence. economic concentration to all regions throu- Scale, Specialization And ghout a country. Further, when places are well connected, people can reach more economic Convergence: A Framework For opportunities, and firms can reach larger Territorial Development markets for inputs and outputs. When connections are lacking, capital and labor markets Territorial development is development that is are segmented, economies of scale limited, aware of where things happen. Economic acti- and opportunities truncated. The benefits of vity, people, jobs, and living standards are often concentration can be spatially spread when not equally distributed across a country. Hence, places are well connected, and connected places when thinking about how to reach every corner can specialize in the tasks and product lines in of a country, every citizen within its boundaries, which they are most productive. But concentra- policy makers must understand the landscape of tion and specialization do not need to lead to territorial development along three dimensions: disparities in living standards across regions. • Scale, because people and firms often They can instead be leveraged to achieve spatially concentrate in just a few places, and convergence across a territory. An analysis that this is evidence that proximity is valuable. looks at the economic geography of a country • Specialization, because the benefits of along these three dimensions can therefore concentration can be spatially spread when help identify challenges and define priorities in places are well connected and connected a territorial development strategy. places can either specialize in their most For each territorial development dimension, productive sectors, or diversify and keep analysis was steered by key guiding questions innovating. and proxy indicators chosen to answer such • Convergence, because spatial concentra- questions. It is worth noting that no single tion of people and firms can be harnessed indicator can answer the key questions around to improve living conditions everywhere, Scale, Specialization, and Convergence, and enabling the convergence of living standards. hence several indicators are used in this anal- Scale, Specialization, and Convergence – the ysis. Scale, for example, is represented by indi- key dimensions of territorial development. cators on concentration of people, economic People and firms are often concentrated in activity, and firms. For Specialization, indica- a few places within a country, reflecting the tors around the distribution of such economic value of proximity. However, spatial concentra- activity and employment were used, coupled tion of economic activity is often a concern for with measures on connectivity and an enabling governments that would like to see economic environment. Convergence was measured growth evenly spread out across a territory. through indicators on living standards. 4 | EXECUTIVE SUMMARY

Where does Argentina stand?

Scale Specialization Economic activity is concentrated in the Employment specialization is weak in Province of Buenos Aires but its bene- most provinces of the country fits do not extend to all regions

Economic Concentration: Nighttime lights Krugman Specialization Index (KSI) at the province Source: data from NOAA, 2015. level Source: Data from OEDE, authors’ computation.

Convergence Argentina has seen some convergence in living standards, but gaps remain in some key indicators

Percentages of the total number of poor people living in each district Source: INDEC (2010) TERRITORIAL DEVELOPMENT IN ARGENTINA: DIAGNOSING KEY BOTTLENECKS | 5

LOW SCALE IN ARGENTINA – capital. 71 percent of passengers in 2016 used an low firm density, with little growth of firms airport in the Pampeana region, of which 61.4 4 Concentration of people and economic activities is percent had Buenos Aires city as their origin or high in Argentina, but firm density is low, as most return destination. In terms of cargo, 91 percent firms remain micro and small, implying a lack of of the total airport freight passes through Ezeiza scale. International Airport in Buenos Aires. Finally, the greatest flows of cargo and passengers are in the Argentina aligns with global trends, with popu- corridor connecting Buenos Aires with Rosario lation and economic activity being concentrated and Cordoba –where much of the country’s popula- in a few places. The Province of Buenos Aires hosts tion and main economic activity are concentrated. close to 40 percent of the population and contributes to 1/3 of the national Gross domestic product But despite concentration of people and economic (GDP). The capital city is nearly 11 times as large activity, lack of scale is reflected in mainly firms as the second largest city, Cordoba, reflecting the that are mainly local and small, with overall low fact that Argentina has one of South America’s firm density across different areas. Most firms highest urban primacy rates with 44.4 percent, just are concentrated in the Pampeana region, hosting below Paraguay with 58.6 percent and Peru with 77.5 percent of the country’s firms. In 2015, 67 48 percent. Brazil, on the contrary, has the lowest percent of firms were microenterprises, 22 percent primacy rate in South America with 13.1 percent. small, 6 percent medium, and 5 percent large firms. Moreover, 91 percent of the country’s population While most of the firms in Argentina are micro- lives in urban areas2, and only 31 cities contain 70 enterprises, 51 percent of the private employment percent of the total urban population. Northern is generated by large firms. This pattern is not Argentina is second in terms of concentration of unique though, international experience suggests population, followed by Cuyo and Patagonia, in that in most countries the clear majority of enter- that order. However, despite experiencing rapid prises (between 70% and 95%) are microenter- urbanization in the past two decades, the North’s prises. But, firm density in Argentina is two to contribution to national GDP is still well below its four times lower than in neighboring Chile and population share. The North is home to 21 percent Brazil. Argentina has only 13 firms per 1,000 habi- of the population but contributes only 10 percent tants, while Chile and Brazil have 48 and 25 firms to national GDP3 and 7 percent of total exports per 1,000 habitants respectively. The Pampeana (INDEC, 2016). Cuyo and Patagonia, on the other and Northern regions have low density with just hand, host 8 percent and 6 percent, respectively, of 16 firms and six firms per 1,000 inhabitants. national population, and contribute 7 percent and Lack of scale is also seen in the dominance of 8 percent of GDP (Ministry of Production, 2015). microenterprises that do not graduate to higher Argentina displays concentration in other size groups. Firms in Argentina stagnate in their ways: Buenos Aires acts as Argentina’s trans- size group over time, with only a few graduating port hub, with most air transportation of pass- to higher size classes. For example, 57 percent engers and road freight flowing through the of microenterprises remained a microenterprise after 5 years, and 40 percent closed. In the case of 2 All localities with a population of 2,000 or more are considered urban 4 Share of the passengers that used the two airports serving AMBA: in Argentina’s official definition. Aeroparque Jorge Newbury and Ezeiza International Airport. 3 GDP in constant prices of 2004; Ministry of Production (2015). 6 | EXECUTIVE SUMMARY

medium firms, only 6 percent grew to become large all years analyzed, revealing that specialization firms. The fast-growing firms—those that manage at the province level between 2006 and 2016 was to increase employment quickly and sustainably— weak, and grew weaker over time. At a smaller only represented less than 1 percent of the total scale, some small- to medium-sized cities show number of firms between 2010-2016 (Ministry of high levels of employment specialization, partic- Production, 2018). ularly in the Northern provinces. But less than 1 percent of the cities in Argentina have employ- LOW SPECIALIZATION IN ARGENTINA – ment specialization. lack of integration between local markets Local markets in Argentina are not well-integrated Weak specialization levels point at connectivity due to poor connectivity and institutional barriers. challenges. There are significant spatial disparities in terms of access to connective infrastruc- Services dominate GDP and employment, and ture across Argentina, which largely align with there is high contribution of non-tradeables economic under-performance. While provinces in to employment in Argentina. In 2017, services central Argentina are well connected, reflecting contributed to 50.4 percent of national GDP. In a dense transportation network between cities, all regions, employment has also been mainly hard connective infrastructure is weak particu- 5 concentrated in the services sector , about 43 larly in the North and Patagonia. The North of percent nationally in 1996 and 48 percent in Argentina displays lower road density than the 2016. Regionally, Pampeana’s services sector national average, Pampeana, or Cuyo. Except for contributes 78 percent of local GDP and consti- Jujuy, road density in the NOA region was 8 km tutes almost half of total employment, followed and 5.9 km every 100 km2 for national and provin- by manufacturing and commerce. In Northern cial roads respectively in 2015. Moreover, areas in Argentina, services represent about 39 percent of Patagonia together with areas East and West of employment, but their prominence has grown over the Jujuy-Salta-Cordoba corridor have the lowest time. Finally, Metropolitan Buenos Aires and the levels of road accessibility and relatively higher top five agglomerations have lower employment in travel times. To put this into perspective, Tanzania tradeable sectors compared with the average for and Namibia, in Africa, have a road density of 9.1 both direct and best-practice comparator cities. km and 5.5 km every 100 km2 of land area (FAO, Specialization at the regional and city levels is 2011), respectively. These connectivity challenges weak and becoming weaker over time. To better translate to high logistics cost, particularly for understand the levels of specialization of employ- the northern provinces. Logistics cost for the ment across locations in Argentina, a Krugman northern regions has been estimated to be up to 50 Specialization Index (KSI) was built. In general, percent higher than in the neighboring Center and highly specialized localities have indices higher Cuyo regions (World Bank, 2016). Also, in the 2016 than 0.75, whereas localities with an index Logistics Performance Index ranking, Argentina below 0.35 can be considered as not special- ranked 66 – after Mexico, Brazil (54, 55) Uganda ized (Blankespoor et al., 2017). In Argentina, the (58); Kenya (42) and India (35). Furthermore, ‘Soft’ mean and median KSI are very close to 0.35 for connective infrastructure is weak in the country. Overall, northern Argentina is the most disad- 5 OEDE data. Total number of employed in the province is obtained from the total number of employees declared in that province by each firm. vantaged in terms of internet access. In 2016, TERRITORIAL DEVELOPMENT IN ARGENTINA: DIAGNOSING KEY BOTTLENECKS | 7

21 percent of people had residential broadband The Northern provinces are home to a high share connections, well below the national average of of Argentina’s poor population, and they encom- 41 percent. The same differences are observed in pass Argentina’s poorest provinces as well. Even other telecommunications: there were on average today, the provinces in the Northwest of Argentina 146 mobile lines for every 100 people in Argentina struggle with extreme poverty and considerable in 2017, but only 119 in the Northern provinces gaps in living standards when compared to the compared for instance to 152 in Patagonia. rest of the country. Some localities in the Northern provinces have more than 35 percent of house- Weaknesses in institutional barriers also holds with unmet basic needs. The two munici- dampen integration. Quality of institutions in palities that fare the worst on unsatisfied basic Argentina is a key constraint for further invest- needs are in Formosa and Salta, with 68 and 49 ments and private sector expansion. Analysis percent of households lacking basic needs. The using information from the Argentina Enterprise North also dominates in terms of absolute poverty. Surveys for 2010 and 2017 suggests that firms in Most districts with a share of the overall house- both the informal and formal sector highlight holds with unsatisfied needs at the upper end of institutional barriers as the key obstacles to their the distribution, are in Northern provinces and regular operation. Buenos Aires. In fact, three districts in Buenos LACK OF CONVERGENCE IN ARGENTINA – Aires have the largest proportion of total house- some progress, but gaps still exist holds with unmet basic needs (i.e. Distrito Federal, Despite convergence in living standards in some La Matanza and General Sarmiento). areas, large disparities between regions in Argentina Those regions with high levels of poverty and gaps remain. in living standards are also less effectively equip- Argentina has seen some convergence in living ping their populations with skills for produc- standards, but gaps remain. While access to tive work. As with access to water, there has water has converged over the years, the Northern been convergence in primary education coverage. region still lags behind the rest of the country in According to the last census (2010) the national terms of access to sanitation and basic services. average was at 92.3 percent, and the range at 75 percent of households the Northeastern prov- provincial level from 88.5 to 97.4 percent. However, inces (Chaco, Formosa, Santiago del Estero) have the North of Argentina still lags in human capital access to water, which is the lowest percentage development, with provinces in the North system- in the country, but it is still not that far from the atically ranking among the bottom provinces proportion in the city of Buenos Aires (99 percent) in education indicators. All the provinces in the (World Bank, 2018). In the case of sanitation, only North have secondary completion rates below or 18.01 percent of the population in the province of at 50 percent. Of all provinces below the median Misiones in NEA has access to sewage, while in (45.8 percent), seven are in the North, with the Chaco and Formosa, less than 35 percent have. In remaining three in Patagonia, one in Pampeana, contrast, more than 80 percent of the population in and one in Cuyo. Similarly, the illiteracy rate in the provinces of Tierra del Fuego, Santa Cruz, and northern Argentina (at 3.7 percent) is twice as high Chubut in Patagonia has access to sewage. as the national average. 8 | EXECUTIVE SUMMARY

Where Do The Provinces Of Salta public accounts. Moreover, hard and soft connec- And Jujuy Stand? tive infrastructure are weak within and between Salta and Jujuy. Salta and Jujuy’s provincial-level Salta and Jujuy have high concentration of people roads are particularly low quality, with only 12.5 in northern Argentina, but lack of SCALE is and 14.8 percent paved respectively, suggesting evidenced by firms’ low growth and producti- that they are relatively poorly connected inter- vity. Salta and Jujuy are important provinces nally to facilitate local-level trade, scale, and within Argentina’s Northwest (NOA). They specialization. However, the Northwestern Road have three international borders with Chile to Development Corridor Project has the potential the West, Bolivia to the North, and Paraguay to to facilitate trade between Jujuy’s main cities, and the Northeast. Within the Northwestern region with the city of Salta. This may facilitate develop- (NOA), they represent 41.2 percent of the popula- ment of regional value chains in key sectors such tion, and 42 percent of the economic activity (26.9 as mining, agriculture, and tourism. and 15.3 percent of Geographical Gross Product- GGP-, respectively). However, the provinces of Low access to and quality of basic services, low Salta and Jujuy have lower firm density than the human capital, and lack of public infrastruc- national average, with 6 and 5 firms per 1,000 ture perpetuate the gaps in livings standards people, respectively. Moreover, in the capital city and opportunities in Salta and Jujuy, making of Salta, the share of informal firms is 45.1 percent, CONVERGENCE a challenge. Poverty rates in the highest in the country. Informality is typically Salta and Jujuy rank highest among all urban associated with low-scale firms that are poorly agglomerations nationally. Salta and Jujuy have a specialized, entailing low productivity, invest- poverty incidence of 24.8 and 24.2 percent respec- ment, and survival. tively, topped only by Santiago del Estero (38.3 percent). Regarding households with unmet basic Public services’ dominance and weak connectivity needs, Salta has the highest proportion (19.4) and limit opportunity and economic integration for Jujuy (15.5) the third highest in the NOA region. Salta and Jujuy, hindering SPECIALIZATION. Additionally, Salta and Jujuy have qualitative A large proportion of Salta and Jujuy’s services housing deficits of 25.7 and 24.0 percent respec- sectors involve public services. Public services tively. In terms of human capital, illiteracy rates in account for almost a quarter of total GGP in Salta the provinces Salta and Jujuy are low (3 percent), province. 3 6 . 5 percent of the cit y of Salta’s residents but much higher than the national average (1.9 were employed by the public sector in 2015. This percent). Regarding public infrastructure, only share distinguishes the city of Salta from inter- 3.6 and 4.8 percent of the houses in Salta and San national cities like Bloemfontein in South Africa, Salvador de Jujuy have streetlights outside their which is the judicial capital of the country, but still houses. Nevertheless, the municipalities of Salta has around 9.6 percentage points lower employ- and San Salvador de Jujuy have higher access to ment in public services than Salta. International basic services relative to the rest of the cities in experience suggests that continued reliance on their provinces and lower shares of houses whose the government for productive activities indi- conditions are below habitability standards (16.8 cates low levels of productivity in a territory, percent for Salta and 16.6 percent for San Salvador while also presenting an unsustainable burden on Jujuy). TERRITORIAL DEVELOPMENT IN ARGENTINA: DIAGNOSING KEY BOTTLENECKS | 9

Where Does Argentina Stand In providing some benchmarking to assess how Argentina compares to other countries around the The World? world in Scale, Specialization, and Convergence. Key The tables below summarize the key messages messages for policy makers on where to go from of the report, pointing at the key findings and here are also included for each dimension.

TERRITORIAL DEVELOPMENT IN ARGENTINA

Scale can mean concentration of people Argentina vs the World

of the population is concentrated in the Argentina does not differ from 37% Buenos Aires Metropolitan Area global trends with population and economic activity being 91% of the country’s population lives in urban areas highly concentrated in a few of this urban population is concentrated in just places. Santiago (Chile) accounts 70% 31 cities for 40% of the country’s urban population.

Scale can mean concentration of economic activity Concentration around the world is similar, where Tokyo and Paris of GDP is produces in two regions: Buenos 66% concentrate over 40% and 30% Aires Metropolitan Region and Pampeana. of their own nation’s economic The economy of AMBA is nearly 11 times activity in less than 4 and 2% of as large as that of the second largest city, the country’s land respectively. Cordoba. But these places have managed brought But AMBA has not managed to extend these to distribute the benefits by this economic concentration benefits to all regions. throughout all regions.

Scale can mean concentration of firms Countries in the region such as Chile and Brazil have firm Lack of scale is also indicated by a high share of densities 2 to 4 times as high. local and micro firms with overalllow firm density,

Scale Australia and South Korea have a and a small share of fast-growing firms. firm density of 88 and 66 firms per Argentina has only 13 firms per 1,000 inhabitants; this is 1,000 inhabitants. even lower in the Northern regions, which have 6 firms per 1,000 inhabitants.

INFORMATION FOR POLICYMAKERS

To get scale, policymakers will have to think about solutions that foster agglomeration economies, increase productivity, and allow concentration of people and firms to thrive while also minimizing the negative externalities of concentration (congestion, pollution, etc.).

Creating Scale means supporting the expansion of economies of scale external to firms – for example, by supporting agglomeration economies through better local services or through reducing barriers in the local business environment — and supporting the scale of production within firms or sectors which raises the efficiency of production.

One way in which local governments have often tried to create scale is through Special Economic Zones (SEZs). SEZs can support the creation of scale by allowing agglomeration economies to follow – organized clusters of firms solve a coordination failure and allow efficient provision of infrastructure and services, but only if there is demand for the services they offer, and markets for those industries. In addition, it is essential to accompany them with complementary policies and investments to ensure their success. 10 | EXECUTIVE SUMMARY

TERRITORIAL DEVELOPMENT IN ARGENTINA

Specialization can mean places are able to focus Argentina vs the World in their most productive sectors On average, the top 5 The dominance of the service sector across regions aligns agglomerations in Argentina with weak specialization: have lowe employment in Dominance of services: 64% of GDP; 48% of employment. tradeable sectors compared The KSI analysis reveals weak specialization at the province with the average for both level and city level, with specialization becoming weaker direct (Bangkok, Istanbul, and over time. São Paulo) and best-practice A relatively high share of Argentina’s urban jobs depends on comparator cities (London, Paris, non-tradeable sectors. and Seoul).

Connectivity enables specialization According to the Logistics Performance Index (LPI) in 2016 Patagonia and the North have longer Argentina ranked 66th, after India travel times and the lowest levels of Weak connective infrastructure (35), Kenya (42), Mexico (54), road accessibility. is hampering market integration Brazil (55) and Uganda (58). in Argentina Weaknesses in 'soft' connective infrastructure also dampen integration.

A good business environment can spark investment, Argentina has a product market fostering integration and specialization regulations (PMR) index score of 3.11. This is relatively restrictive Analysis of the Argentina Enterprise Surveys for compared to other OECD 2010 and 2017 suggests that firms both in the countries and other countries informal and formal sectors highlight institutional in the region such as Colombia barriers as the key obstacles for their regular Specialization (1.77), Chile (1.51), Peru (1.66) and operation. Mexico (1.91).

INFORMATION FOR POLICYMAKERS

To achieve specialization, it will be necessary to think about solutions that allow firms to reach larger markets for their inputs and outputs.

Connective infrastructure (soft and hard) is key to integrating regional markets. For example, provinces can connect urban and rural places within their borders and connect themselves to the national economy. Investments in connectivity between places make it cheaper for those places to trade.

When a connective investment is made, relocation responses of people and firmas may be hampered if there are barriers to their mobility, highlighting the need for complementary actions. This includes frictions like land transfer fees, business registration costs, information gaps for job search, and sunk cost investments in buildings. TERRITORIAL DEVELOPMENT IN ARGENTINA: DIAGNOSING KEY BOTTLENECKS | 11

TERRITORIAL DEVELOPMENT IN ARGENTINA

Convergence means convergence in living standards Argentina vs the World

Despire overall poverty reduction, large disparities in Argentina has reached similar levels to living standards persist in Argentina, with the North te OECD average in access to water and particularly lagging. The Northern provinces are home sanitation. to a high share of Argentina's poor population, and they Internationally comparable test score encompass Argentina's poorest provinces as well. data shows that Argentina underperforms While access to water has converged over the relative to its peers both at the primary years, important gaps remain between the and secondary level. The median Northern regions and the rest of the country Argentine performs in mathematics in terms of access to sanitation and basic an equivalent of 2.5 years below the services. average of OECD countries. Argentina is still far behind in terms Primary education has been improving, of infant mortality, under-five mortality shortening the gap between leading and rate, and life expectancy compared to lagging areas. However, disparities remain in secondary education, quality of education, health international standards. Chile has half maternal mortality rate and almost half services, and other indicators. Lagging regions infant mortality rate; in Malaysia mortality in Argentina are less effectively equipping their rates are also lower, even with less populations with skills and productive work. spending per GDP than Argentina.

INFORMATION FOR POLICYMAKERS

For convergence, the right policy mix depends on how the population is distributed across space.

Convergence All localities need access to decent public services and human capital development.

Where population density is low, it will be important to concetrate access to services inregional centers that allow extended access to the region and consider the possibility of providing mobile services. Where population density is high, a more place-based approach may be required, from social protec- tion to – sometimes – more proactive LED policies.

In large cities, where divisions of formality vs. informality can generate deep gaps in the quality of life, spatially targeted policies may be necessary to reduce these divisions.

But to ensure that public services are distributed effectively it will be necessary to strengthen local institutions and build capacities, so that local governments can fulfill their responsibilities in terms of providing basic services and local public goods to their citizens. 12 | EXECUTIVE SUMMARY

REFERENCES

Blankespoor, B.; Bougna Lonla, T.; Garduño Rivera, R.; Selod, Muzzini, E., Eraso Puig, B., Anapolsky, S., Lonnberg, T. and H. (2017). Roads and the Geography of Economic Activities in Mora, V. (2016). Leveraging the Potential of Argentine Cities: Mexico n.º 8226. Washington, D.C.: World Bank. A Framework for Policy Action. Directions in Development. Washington, D.C.: World Bank. FAO (Food and Agriculture Organization of the United Nations) (2011). Land Portal. Densidad vial (cada 100 km2). Naciones NOAA (National Oceanic and Atmospheric Administration) Unidas. (2015). National Geophysical Data Center database. Silver Spring, MD. INDEC (Instituto Nacional de Estadística y Censos) (2010). National Population, Household and Housing Census. OEDE (Observatorio de Empleo y Dinámica Empresarial). Buenos Aires, Argentina. Conjunto de datos de Dinámica del empleo y rotación de empresas. Ministry of Labor of Argentina. Buenos Aires, (2 016). Encuesta Nacional sobre Acceso y Uso de Tecnologías Argentina. de la Información y la Comunicación 2016. Buenos Aires, Argentina. World Bank (2016). Argentina: The Challenge of Reducing Logistics Costs. Washington, DC: World Bank. Ministry of Production of Argentina (2015). Fichas Provinciales 2015. Buenos Aires, Argentina. (2018). Systematic Country Diagnostic Argentina. Washington, DC: World Bank. (2018). Datos de crecimiento. Buenos Aires, Argentina. CHAPTER I

A framework to analyze territorial development: scale, specialization, and convergence

BUENOS AIRES, ARGENTINA Katarzyna Drużbiak © 14 |

CHAPTER I A framework to analyze territorial development: scale, specialization, and convergence

Territorial development is development that is point of view; it is a sign of good economic perfor- aware of where things happen; by using a spatial mance. Spatial concentration of economic activity lens, it brings a multi-sectoral approach to deve- often reflects regional specialization; the concen- lopment. Economic activity, people, jobs, and tration of firms and people in a single location living standards are often not equally distributed gives rise to agglomeration that fosters economic across a country. Hence, when thinking about how activity, allowing products, workers, and ideas to to reach every corner of a country, every citizen flow across markets and firms, thereby building within its boundaries, policy makers must under- economies of scales in the production process.6 stand the landscape of territorial development It benefits workers who can take advantage of along three dimensions: scale, specialization, and increased opportunities provided by the concen- convergence. This chapter explains these three tration of firms in a single area. It also benefits dimensions, which are the pillars of the territo- firms by bringing them close to their suppliers rial development framework used in the report by creating stronger production linkages, to to analyze Argentina, and the provinces of Salta their consumers by reducing transaction costs and Jujuy. Moreover, it summarizes the indicators to buy their products, and to other similar firms, used to measure each of the dimensions. since they can share technologies and information (Duranton and Puga, 2004; Rosenthal and SCALE – People and firms are often concentrated Strange, 2004). Concentration also allows people in a few places. to be closer to services and economic opportuni- International experiences indicate that as coun- ties, increasing their welfare. tries transition from agricultural to indus- SPECIALIZATION – lack of integration between trial to service-oriented production, firms and local markets can lead to weakly specialized workers move to not just larger and denser settle- regions. ments, but also to places in which scale matters. Scale can be fostered if production and distri- When places are well connected, and markets bution in a place become more efficient, which of inputs and outputs are linked, firms and brings benefits to firms, as well as to people. This workers can locate in the areas where they report will refer to these benefits as agglomera- are most productive, enabling the speciali- tion economies. Agglomeration economies are zation of cities and regions. Because places those benefits that come when firms and people are connected, they can trade with each other, are physically close to one another, such as lower each one exploiting its absolute and compara- transport costs or knowledge sharing between tive advantages (Duranton and Venables, 2018). firms (Malmberg et al., 2000). Due to these bene- Building connected, cohesive territories can

fits, people and firms often concentrate in a few 6 The benefits arising from the concentration of people and firms are places. Concentration of economic mass is inev- often referred to as agglomeration economies (Glaeser, E.L. ed., 2010. itable and generally desirable from an efficiency Agglomeration economics. University of Chicago Press.) TERRITORIAL DEVELOPMENT IN ARGENTINA: DIAGNOSING KEY BOTTLENECKS | 15

be a pathway to eliminating extreme poverty affect the quality of life of citizens and have the and promoting shared prosperity. Poverty and potential to negatively impact national unity and inequalit y in rural areas decline faster when such social cohesion. Disparities in living standards areas are well connected to centers of economic across a country can foster political instability growth such as cities. Recent studies in Chile and and increase the possibility of social unrest. But Colombia have shown that stronger urban-rural while spatial imbalances in economic growth are connections are associated with more economic inevitable, disparities in access to opportunities growth and greater reduction of poverty and living standards are not. Understanding the (Berdegue et al., 2015). Instead, when places patterns of spatial imbalances in living standards are disconnected, regions can remain isolated, provides the third piece of the puzzle to charac- leading to regional autarkies, where these areas terizing the challenges for territorial development end up supplying all goods and services for their in a place. local markets, producing them at often higher For each territorial development dimension costs. Broken links between cities and their above, the analysis was steered by key guiding hinterland also impede the spreading of benefits questions, and proxy indicators chosen to throughout broader regions. In line with founda- answer such questions. It is worth noting that tional trade economics, improved opportunities no single indicator can answer the key questions to trade (such as those brought about by connec- around Scale, Specialization, and Convergence, tive investments like roads) encourage greater and hence several indicators are used in the specialization of locations in activities in which analysis. The table below summarizes the main they are more competitive. Connectivity also basic indicators used in this report to assess the increases firms’ access to inputs and consumer status of Argentina along these three dimen- markets, allowing firms to produce at greater sions, including the guiding questions that moti- scale as well as specializing thanks to more vate their use. More details on the indicators are niche inputs and markets. Connective invest- explained in Annex A. ments therefore tend to support productivity growth, by raising scale, specialization, and To tackle the challenges of achieving a balanced agglomeration effects (Redding & Turner 2015). territorial development, the Government of Argentina (GoA) has several options. It can invest CONVERGENCE – Just as the spatial distribution on policies and investments that foster concentra- of firms and people varies across space, so does tion in new areas or support further concentration the distribution of living standards. in existing growth centers; it can strengthen the While scale and specialization are key mecha- links between lagging and leading regions, areas nisms for territorial development, the final of low and high concentration of people and jobs, objective for policymakers is to ensure the by investing in corridors of connectedness; or it welfare of citizens. The dynamics of markets do can prioritize investments in equalizing living not always lead to increasing living standards, and standards with a focus on basic services, health, concentration of economic activity and incomes and education. Box 1.1 provides details of the all too easily leads to widening and persistent most important efforts of the GoA in territorial gaps in living standards. Such disparities can development. 16 | CHAPTER I A FRAMEWORK TO ANALYZE TERRITORIAL DEVELOPMENT: SCALE, SPECIALIZATION, AND CONVERGENCE

Table 1.1: Summary of the guiding questions and indicators used by each dimension

GUIDING QUESTIONS INDICATORS

SCALE

Scale can mean concentration of people – large cities have more scale. The Population and economic following guiding questions should help identify population scale: activity concentration

1. Does the country display a distinct concentration of population? Urban primacy 2. Where is the geographic concentration of people in the country? 3. Does the country experience an urban hierarchy dominated by a particular city?

Scale can also mean concentration of economic activity. GDP per capita and regional The following guiding questions should help identify economic activity scale: GDP. 1. Does the country display large concentration of economic activity in Labor participation rates at particular regions? the national, provincial and 2. Are there significant variations of economic activity across the country? municipal level. 3. Are there distinct patterns of transportation/commerce flows? Flow of flights Flow of passengers Road freight Flow

Finally, scale can mean firm concentration. The following guiding questions Firms by size should help identify firm activity scale: Firm density

1. How are firms distributed across the country and by sector? Firms by Origin 2. Does the distribution of firms vary across regions in the country? Firms by sector 3. Has such distribution changed over time? Fast-growing firms and firm 4. Does the distribution of human capital reflect different graduation geographical/sectorial patterns? Labor force skill level

SPECIALIZATION

Specialization can mean when places can specialize in what they are Distribution of the GDP by sector most productive. The following guiding questions should help identify Distribution of firms by sector economic activity specialization: Distribution of employment by 1. How is the economic activity distributed across the country? sector 2. How is employment distributed by sector and by provinces? TERRITORIAL DEVELOPMENT IN ARGENTINA: DIAGNOSING KEY BOTTLENECKS | 17

GUIDING QUESTIONS INDICATORS

SPECIALIZATION

Location Quotient (LQ) at the 3. What are the employment drivers regionally? national and provincial level. 4. How are exports distributed across the country and regionally? Employment drivers at the national and provincial level. Krugman Specialization Index (KSI) International production openness index

Connectivity enables specialization. The following guiding questions should Road density help identify connectivity topics key for specialization: Road accessibility Logistics Performance Index 1. How is the quantity and quality of infrastructure in the country? Market Access index 2. How is the road accessibility across the country and within Soft connective infrastructure specific regions? 3. What are the logistics challenges for international trade?

A good business environment can spur investment, integration, and specia- Private sector limitations lization. The following guiding questions should help identify competition Stoplight barriers for specialization: Product market regulations 1. What are the main challenges for private investments and World Economic Forum’s Global the business sector? Competitiveness Index (GCI) and 2. What are key competitiveness challenges in the country, if possible, World Bank’s Doing Business by sector? (DB)

CONVERGENCE

Convergence means convergence in living standards. The following Poverty incidence at the prov- guiding questions should help identify convergence: ince and district level Index of access to services for 1. What are the levels of poverty across the country and regionally? water and sanitation 2. How is the access to basic services, nationally and locally? Education coverage at the 3. How is the access to education and health services, nationally national, provincial, and local and locally? levels Maternal mortality rate (MMR) and infant mortality 18 | CHAPTER I A FRAMEWORK TO ANALYZE TERRITORIAL DEVELOPMENT: SCALE, SPECIALIZATION, AND CONVERGENCE

Box 1.1. Territorial Development efforts in Argentina

The GoA has taken important steps towards a more balanced territorial development. To address the variations in planning policies among provinces and municipalities, the federal government launched in 2003 the National Development and Territorial Planning Policy7 which proposed a strategic vision for the deployment of investment in the territory, with two fundamental tools for spatial planning: the strategic territorial plan and a regulatory framework for territorial development, including efforts on land use regulation. First, the Strategic Territorial Plan (Plan Estratégico Territorial, PET)8 is a framework to guide investment in strategic infrastructures and is the main tool to carry out the National Plan. The PET provides a framework to: 1) identify the necessary investments in infrastructure and improvements in capacity that are needed to reduce the gap in development outcomes between regions, and 2) promote economic development in lagging areas that need to increase productivity and strengthen their connection with local and international markets.

The PET divides the national territory in 25 sub-regions grouped into three main categories depending on the level of urbanization, integration, and economic activity. The PET has been updated several times but the latest version (2016) defines three areas of work: (i) areas of intervention (or lagging areas); (ii) connectivity corridors; (iii) polycentric urban system. Within the areas of intervention, the PET identifies 1) the most dynamic areas, with competitive economies and inserted in the international market; these regions need to adapt their infrastructures and equipment to present and future demands, as well as to mitigate the impacts of unplanned growth; 2) areas with a certain dynamism, based on emerging activities that have not yet reached high levels of consolidation; and 3) the country's currently most lagging and least populated areas. However, the PET has advanced relatively little toward the achievement of its objectives, because implementation of the investment plans is at the discretion of the provinces, and no regulatory tools or financial instruments have been provided to the provinces or municipalities for implementing the plans (Muzzini et al., 2016).

Plan Belgrano

To foster coordination across ministries in the lagging Northern regions and execute the priorities of the territorial plan, the GoA launched the Plan Belgrano in 2016. The plan is rooted in the idea that Argentina owes its lagging Northern provinces, particularly Salta and Jujuy, an ambitious social, production, and infrastructure plan to reduce the spatial inequalities that have hampered the economic development of the country. The policies highlighted in the plan have a triple objective of poverty reduction, regional economic

7 Government of Argentina, Ministry of Federal Planning, Public Investment. 2004. Argentina 2016: Política y Estrategia Nacional de Desarrollo y Ordenamiento Territorial. Buenos Aires: Ministry of Federal Planning, Public Investment, and Services, Undersecretary of Territorial Planning of Public Investment.

8 Government of Argentina, Ministry of Federal Planning, Public Investment. 2016a. “Plan Estratégico Territorial (PET), avance III.” Buenos Aires: Ministry of Federal Planning, Public Investment. TERRITORIAL DEVELOPMENT IN ARGENTINA: DIAGNOSING KEY BOTTLENECKS | 19

development, and reduction of drug trafficking. Specifically, it seeks to reduce poverty and achieve equal- ity of opportunities for the population in the northern provinces, encourage the development of several key sectors, and invest in infrastructure to improve the region’s connectivity. A range of national and regional agencies are responsible for planning and implementing different elements of this agenda, with the coordination at the political level being led by the Plan Belgrano Unit, within the Chief of Cabinet’s office (Chief of Cabinet in Argentina, 2016). The Plan proposed over US$16 billion in road, rail, and air transport investments, a historic reparation fund ($50 billion in 4 years), rapid interventions in areas of extreme poverty, subsidies for port costs, prioritization of social services, and investment in education, a plan for 250,000 houses (low income housing, credits, and subsidies) and labor and fiscal incentives to attract investments to the region. Including future projects (Table 1.2), most of the projects are on road infrastructure (26.7 percent), followed by water and sanitation projects (23.8 percent). At the province level, 20.62 percent of the projects in Plan Belgrano are located in the province of Chaco, while Salta and Misiones concentrate 12.89 and 12.63 percent of the investments, respectively (Secretariat of Productive Integration, Ministry of Production, 2018). The renewal of the Belgrano Freight Railway is the largest project in Plan Belgrano, which include the renewal of 1815 km of roads and connects from Rosario Port to Bolivia.

Table 1.2: Projects of Plan Belgrano

STATUS OF PROJECTS IN PLAN BELGRANO TOTAL AMOUNT (IN ARGENTINE PESOS) PERCENTAGE

Projects about to start 2,018,574,918.41 1.2%

Adjudicated projects 1,007,835,000.00 0.6%

Projects in Progress 97,319,872,541.03 56.7%

Projects in Bidding Process 4,532,497,629.47 2.6%

Projects in revision stage 8,260,262,021.23 4-8%

Finished projects 35,982,621,927.63 21.0%

Detained projects 3,215,611,528.26 1.9%

Future Projects 19,380,961,732.07 11.3%

Total 171,718,237,298.11

Source: Secretariat of Productive Integration, Ministry of Production (2018)

The following chapters aim at drawing a diag- chapter provides an overview of the economic nosis of territorial development in Argentina, geography of Argentina by discussing charac- with a zoom-in on the provinces of Salta and teristics of the spatial distribution of people Jujuy, using the three dimensions of Scale, and economic activities and presenting the Specialization, and Convergence. The next challenges the country faces along the three 20 | CHAPTER I A FRAMEWORK TO ANALYZE TERRITORIAL DEVELOPMENT: SCALE, SPECIALIZATION, AND CONVERGENCE

dimensions. This same territorial lens is then Argentina; and (iii) they are in the direct area of applied at the subnational level in Salta and Jujuy influence of the World Bank Northwest Corridor since (i) they are among the poorest, and those Project. Finally, the report presents an over- facing deeper challenges in terms of achieving view that summarizes where Argentina stands in convergence of living standards; (ii) they repre- the world in terms of Scale, Specialization, and sent an economic opportunity for northern Convergence.

REFERENCES Berdegué, J. A., Carriazo, F., Jara, B., Modrego, F., & Soloaga, I. Ministry of Federal Planning, Public Investment and Services (2015). Cities, territories, and inclusive growth: Unraveling of Argentina (2004). Argentina 2016: Política y Estrategia urban–rural linkages in Chile, Colombia, and Mexico. World Nacional de Desarrollo y Ordenamiento Territorial. Buenos Development, 73, 56-71. Aires: Ministerio de Planificación Federal, Inversión Pública y Servicios, Subsecretaría de Planificación Territorial de la Duranton, G., & Puga, D. (2004). Micro-foundations of urban Inversión Pública. agglomeration economies. In Handbook of regional and urban economics (Vol. 4, pp. 2063-2117). Elsevier. (2016). Plan Estratégico Territorial (PET), avance III. Buenos Aires: Ministerio de Planificación Federal, Inversión Pública & Venables, T. (2018). Appraising Place-Based Policies: y Servicios. Principles and Argentine Examples. Documento de antecedentes para este informe. Muzzini, E., Eraso Puig, B., Anapolsky, S., Lonnberg, T. & Mora, V. (2016). Leveraging the Potential of Argentine Cities: Glaeser, E. L. (ed.) (2010). Agglomeration Economics. University of A Framework for Policy Action. Directions in Development. Chicago Press. Washington, D.C.: World Bank. Chief of Cabinet of Argentina (Jefatura de Gabinete in Spanish) Redding, S. J. & Turner, M. A. (2015). Transportation Costs and (2016). Presentación: el Desarrollo del Norte Argentino a través the Spatial Organization of Economic Activity. En Handbook of del Plan Belgrano. Buenos Aires: Jefatura de Gabinete, Unidad Regional and Urban Economics (vol. 5, pp. 1339-1398). Elsevier. Plan Belgrano. Rosenthal, S. S., & Strange, W. C. (2004). Evidence on the nature Malmberg, A., Malmberg, B., & Lundequist, P. (2000). and sources of agglomeration economies. In Handbook of Agglomeration and firm performance: economies of scale, regional and urban economics (Vol. 4, pp. 2119-2171). Elsevier. localisation, and urbanisation among Swedish export firms. Environment and Planning a, 32(2), 305-321. Secretaría de Integración Productiva, Ministry of Production. (2018). Informe Regional: Obras Plan Belgrano. Buenos Aires, Argentina. TERRITORIAL DEVELOPMENT IN ARGENTINA: DIAGNOSING KEY BOTTLENECKS | 21 CHAPTER 2

Scale, specialization, and convergence: where does Argentina stand?

TUCUMÁN, ARGENTINA Nicolás Perondi © 22 | CHAPTER 2 SCALE, SPECIALIZATION, AND CONVERGENCE: WHERE DOES ARGENTINA STAND?

CHAPTER 2 Scale, specialization, and convergence: where does Argentina stand?

Where does Argentina stand? With 2.8 million by the concentration of firms in a single area. square kilometers, Argentina is one of the Firms also benefit from being close to their largest agricultural producers in the World. It suppliers, giving them stronger production also has a vast amount of natural resources in linkages and access to larger markets for their energy, with the second highest shale gas and products; they may also benefit from being fourth highest shale oil reserves in the World. closer to other firms as proximity allows Argentina also has significant opportunities in spillover of ideas, information, and techno- some manufacturing subsectors and high-tech logies. Moreover, consumers benefit from innovative services. Nonetheless, compared to having access to larger and more varied product its peers, Argentina’s long-run economic perfor- baskets. Concentration allows people to be mance appears to be below potential. Average closer to services and economic opportunities, long-run economic growth in Argentina has increasing their welfare (Duranton and Puga, been only 2.7 percent—about half the growth of 2004; Rosenthal and Strange, 2004). However, high performing countries in the region and less governments justifiably concerned that concen- than a third of growth of emerging countries in tration of economic activity can leave some Asia. Today, the country faces yet again another areas behind. Also, proximity gives rise to nega- crisis: inflation is rising, and the Argentina peso tive externalities or congestion effects. Without continues to devaluate. Crises are generally additional investments in infrastructure or viewed as dangerous, expensive, and detrac- innovations in urban management and policy, ting from other agendas and priorities, but they cities can become congested, roads and other also pose opportunities to evaluate and redi- public infrastructure more crowded, and crime rect policies and resources to increase their more prevalent. All cities are subject to the effectiveness. This section looks at the territo- opposing forces of agglomeration and conges- rial development challenges of Argentina along tion, but their net outcomes can be influenced the three dimensions of Scale, Specialization, by policy choices, determining their producti- and Convergence as a way to identify where the vity. In what follows, this chapter will look at biggest gaps lie, and how policy and invest- the patterns of scale of economic activity and ments can help breach them. people in Argentina.

LOW SCALE IN ARGENTINA – Argentina has a heavily concentrated low firm density, with little growth of firms economy and population

People and firms often concentrate in a few Argentina does not differ from global trends, places. This benefits workers who can take with population and economic activity being advantage of increased opportunities provided highly concentrated in few places. With a TERRITORIAL DEVELOPMENT IN ARGENTINA: DIAGNOSING KEY BOTTLENECKS | 23

population of 40.12 million (INDEC, 2010), Figure 2.2. Economic concentration in AMBA: nighttime lights Argentina’s population is highly concentrated in its capital city. According to the 2010 Census, 37 percent of the population is concentrated in the Buenos Aires Metropolitan Area (AMBA), and 66 percent is in AMBA and the nearby Pampeana Region. It is also a highly urbanized country, with 91 percent of the country’s population living in urban areas,9 and 70 percent of this urban population is concentrated in just 31 cities (Figure 2.1). Additionally, Argentina’s greater agglomerations – AMBA, Greater Mendoza, and San Juan – host 50 percent of urban population, above Latin American average of 40 percent (Figure 2.2). Mexico and Brazil, for example, host 8 and 19 greater agglomerations, respectively, spreading that Latin American average over more metropolitan areas, making the difference greater with Argentina (Ferreyra Source: NOAA, 2015. & Roberts [eds.] 2018). The Metropolitan Area of Buenos Aires hosts

Figure 2.1. Population density is higher in the center and in close to 40 percent of the population and Buenos Aires contributes to 1/3 of the national GDP. Buenos Aires dominates the economic geography of Argentina. The economy of AMBA is nearly 11 times as large as the second largest city, Cordoba, reflecting the fact that Argentina has one of South America’s highest urban primacy rates with 44.4 percent, just below Paraguay with 58.6 percent and Peru with 48 percent. Brazil, on the contrary, has the lowest primacy rate in South America with 13.1 percent (Ferreyra & Roberts [eds.] 2018). Economic activity is geographically concentrated in the central and coastal areas, with two thirds of the national Gross Domestic Product (GDP) being produced in two regions: Buenos Aires Metropolitan Area and Pampeana.10 High concentration in few places is not necessarily a problem; on the contrary it

Source: WorldPop, 2015. is seen throughout the world. However, AMBA

9 All localities with a population of 2,000 or more are considered urban 10 Data cited is from INDEC (2016). The graph maps the distribution of in Argentina’s official definition. economic activity using night light data (VIIRS) for Argentina. 24 | CHAPTER 2 SCALE, SPECIALIZATION, AND CONVERGENCE: WHERE DOES ARGENTINA STAND?

has not extended the benefits of concentration the negative externalities of such concentration to all regions and has been unable to manage (Box 2.1). Box 2.1. Zooming into Cities: The Case of AMBA

The Metropolitan Area of Buenos Aires (AMBA by its acronym in Spanish) concentrates the largest share of the country’s population and economic activity. AMBA is the third-largest megacity in Latin America after Mexico City and São Paulo, concentrating 14.8 million people, or one third of the country’s population. The metropolitan area generates almost half of the country’s GDP; however, it presents large disparities between the core and periphery.

It is estimated that 25.5 percent of the population in AMBA lives in poverty. But there is a significant difference in poverty rates between the Autonomous City of Buenos Aires (CABA by its acronym in Spanish) and the average in the municipalities that constitute the metropolitan area (9.0 percent and 29.5 percent respectively) (INDEC, 2017b).. This is partly explained by the expansion trends of AMBA. AMBA has registered one of the highest population growth rates in the country, at an annual rate of 1.3 percent between 2001 and 2010. The growing population was mostly accommodated by expansion in the peri-urban areas (Muzzini et al., 2016), resulting in a size- able population living in relatively new low-density neighborhoods. This has made it expensive to extend public infrastructure and public services, while also reinforcing social and economic segregation between the municipalities and central CABA.

The lack of affordable housing in AMBA is acute. AMBA experiences the highest housing deficit when compared to other Argentine cities. There are 53 vulnerable neighborhoods11 in CABA, and 1,062 in the rest of the metropolitan area (Chief of Cabinet, 2016a). There are also significant differences in access to basic services between core and periphery. While access to piped water supply and sewerage is almost 100 percent in CABA, 30 percent lacked piped water and 59 percent lacked sewage access in the metropolitan area in 2010 (INDEC, 2010). Limited public transport options in the metropolitan area has resulted in lower employment opportunities for households, particularly low-income populations without private modes of transportation.

A key reason for territorial disparities in AMBA is the lack of effective coordination and planning at the metropolitan level. While this is a challenge for most large cities around the country, the mag- nitude of the challenge is especially relevant in AMBA given its size and degree of institutional fragmentation. AMBA’s governance involves the National Government, the Government of the Province of Buenos Aires (PBA), the government of CABA, and the surrounding municipalities. His- torically, AMBA made limited progress in the establishment of metropolitan coordination entities.

11 This refers to Barrios Populares, defined by the government as at least 8 families grouped, where more than half of the population does not have the land title or regular access to two or more basic services (water network, electrical network with meter and/or sewerage network). TERRITORIAL DEVELOPMENT IN ARGENTINA: DIAGNOSING KEY BOTTLENECKS | 25

A few specific sectoral institutions were created to foster inter-jurisdictional coordination – such as those for the management of metropolitan transit (Autoridad Metropolitana de Transporte) and the environmental management in the Matanza Riachuelo river basin (Autoridad de Cuenca Man- tanza Riachuelo). More recently however, the governments of CABA and PBA have committed to improving the metropolitan coordination of policies and programs through the establishment of a Metropolitan Cabinet in December 2015. Since its inauguration, it has been holding regular meet- ings to identify and follow up on priority areas of common interest for collaboration, which include, for example, the development of a metropolitan data system (AMBA-Data) and the extension of the CABA’s Emergency Medical Services (SAME) to the rest of the metropolitan area.

Argentina’s concentration of economic acti- Figure 2.3: Argentina’s Regions vities and people across space is explained partly by geography and history. The country has five major regions (Figure 2.3): Cuyo (West), Northeast (NEA), Northwest (NOA), Pampeana (central) and Patagonia region (South). The Northern regions, Cuyo, and Patagonia, with the lowest accessibility in the country, have a more difficult topography than the fertile grass- lands of the central Pampas, where Buenos Aires is located. The mountainous Andean region in Cuyo extends 3,700 km along the western edge of the country from Bolivia to southern Patagonia, forming most of the natural boundary with Chile; while the NEA region is partly composed of dry lowlands (Gran Chaco), an area with a challenging combination of physical conditions which makes it one of the least-inhabited parts of the country. Finally, Patagonia is a cold, parched, windy region that extends 1,900 km south of the Pampas, from the Colorado River to Tierra del Fuego. 12

12 Geographical information was obtained from Encyclopedia Britannica. Source: INDEC. 26 | CHAPTER 2 SCALE, SPECIALIZATION, AND CONVERGENCE: WHERE DOES ARGENTINA STAND?

The North of Argentina (consisting of the Aires overall (43.8) and Pampeana (42.9). The Northeast and Northwest regions;13 see Figure Northeast (NEA) had the lowest employment 2.3 above) contributes significantly below its rate, at 38.9 percent (the only region with values population share to national GDP. Northern below 40 percent). Low labor force participa- Argentina is second in terms of concentration tion in lagging regions confounds unemployment of population, followed by Cuyo and Patagonia, statistics. Thus, although the lowest unemploy- in that order. However, despite experiencing ment rates were in Cuyo (3.2) and NEA (3.9), rapid urbanization in the past two decades, the NEA’s rate reflects a labor force participation North still contributes significantly below its of just 50.8 percent, well-below the national population share to national GDP. The North average of 58 percent, with many discouraged is home to 21 percent of the population but workers that fall into inactivity. Participation contributes only 10 percent to national GDP14 rates are slightly higher in Northwest (NOA) and 7 percent of total exports (INDEC, 2016). (56.6 percent). The national unemployment rate Cuyo and Patagonia, on the other side, host was 7.2 percent. 8 percent and 6 percent, respectively, of the Figure 2.4. Provinces in the NEA and NOA, except national population, and contribute 7 percent Catamarca, have a per capita Gross Geographical Product (GGP) lower than the national average, 2005* and 8 percent of GDP (Ministry of Production, 2015). Figure 2.4 shows that all the provinces in the Northwest and Northeast regions (highlighted in light blue) except for Catamarca15 have a lower per capita gross geographical product (GGP) than the national average. By contrast, the city of Buenos Aires has a higher per capita GGP (as well as more firms per capita) than the average high middle-income country (Muzzini et al. 2016).

Employment and participation rates are relatively homogenous across regions, with the exception of the Northeast, where labor participation rates are low. In 2017 the national employment rate was 43 percent, with most regions between 40-43 percent (INDEC, 2017a). It was the highest by far in the AMBA (50.6), followed by more typical rates in Buenos

Source: SIDEP Ministerio de Hacienda; *Current prices (last data available). 13 Northern Argentina comprises NOA (Northwest) and NEA (Northeast) Provinces ordered according to total population (2005), from least to most regions. The Northwest region comprises the provinces of Catamarca, populated. *Note: The figure shows Gross Geographical Product (‘local GDP’) Jujuy, La Rioja, Salta, Santiago del Estero, and Tucumán; and the per capita by province. Provinces with lower populations are at the top, going Northeast region comprises the provinces of Formosa, Chaco, to the most populated provinces in the bottom. Northeast and Northwestern Corrientes y Misiones provinces are colored in the lightest blue. The national average is shown in 14 GDP in constant prices of 2004; Ministry of Production (2015). grey at the bottom. The high contribution of regions such as Tierra del Fuego is 15 Due to its gold and copper mines a function of the low population in the province. TERRITORIAL DEVELOPMENT IN ARGENTINA: DIAGNOSING KEY BOTTLENECKS | 27

Natural endowments do not fully explain the they are less important in explaining economic concentration of people or economic activi- activity in Argentina than in other countries in ties in Buenos Aires, highlighting the role of the region such as Brazil, Chile, Colombia or public policy and agglomeration economies in Ecuador. This suggests a stronger role for policy the country. Figure 2.5 displays agro-climatic and ‘man-made’ agglomeration economies in the conditions across Argentina. The Pampeana and countr y. For example, analysis by Fajgelbaum and parts of the Cuyo regions display clear natural Redding (2018) shows that the primacy of Buenos advantages,16 and are where most economic Aires was deepened substantially by the opening activity (measured by night lights) is also located. of Buenos Aires to international markets in the While natural endowments may have historically late 1800s, combined with the opening of rail- helped to cement the predominance of Buenos ways that connected Buenos Aires to the more Aires, however, Henderson et al. (2012) find that distant regions by the early 1900s (Box 2.2).

Figure 2.5. Natural endowments are less important in accounting for economic activity in Argentina than in other countries in the region

Average monthly Average monthly Total lights temperature (1960–1990) precipitation (1960–1990)

Length of growing period Distance to natural Land suitability harbor

Source: Henderson et al. 2012. 16 In Cuyo’s case, good soil and climatic conditions 28 | CHAPTER 2 SCALE, SPECIALIZATION, AND CONVERGENCE: WHERE DOES ARGENTINA STAND?

Box 2.2. Connective investments and the primacy of Buenos Aires

The construction of a network of railways connecting Buenos Aires to Argentina’s Northern cities coincided with the opening of Buenos Aires to international trade in the late 1800s and early 1900s. Fajgelbaum and Redding (2018) show that structural transformation, from agriculture to non-agriculture, and across disaggregated goods within the agricultural sector, was central to Argentina’s rapid export-led economic development. The authors provided evidence that the re- ductions in internal transport costs from the construction of the railroad network were important in enabling interior regions to participate in this process of structural transformation and economic development. Results from the study shows that the construction of the railroad network increased the total population of Argentina by 49 percent under free international migration and raised the common real wage across all Argentinian districts by 8 percent under restricted international migration.

However, this deeper internal and external connectivity reinforced the primacy of Buenos Aires and central Argentina (Fajgelbaum and Redding, 2018).Before these investments, it could take 3-4 months to travel from Buenos Aires to Salta, the largest city in the country’s Northwest, and cost around 13 times as much to move a ton of goods from Salta to Buenos Aires as from Liverpool to Buenos Aires (Fajgelbaum and Redding, 2018). However rather than spreading production more evenly across locations, the trade and urbanization that followed reinforced the centrifugal force of Buenos Aires throughout the 20th century, and particularly after the 1930s.

Migration movements within Argentina were also historically directed towards the province of Buenos Aires. These movements started in the early 20th century, as the railroads reduced travel times to the capital city. While towards the end of the 1980s internal migration was also registered to medium-size cities such as Cordoba, Rosario, and Santa Fe, Buenos Aires remained the main destination for internal migration. In recent decades, a larger flow of internal migrants towards Patagonia has been recorded.

In 2010, only 68 percent of the population of AMBA was born in the city, showing that migration into Buenos Aires continues to be strong. This contrasts with more than 90 percent in the Northern provinces; for instance, 91.6, 90.5, and 87.7 percent of people in Chaco, Jujuy, and Salta were born in those provinces, respectively. Reflecting recent trends, Patagonian provinces display a larger share of non-natives, with only 56.2, 71.8, and 72.2 of people originally from the provinces of Santa Cruz, Neuquén, and Chubut, respectively.

Source: Fajgelbaum and Redding, 2018 TERRITORIAL DEVELOPMENT IN ARGENTINA: DIAGNOSING KEY BOTTLENECKS | 29

As well as concentrating Argentina’s popula- (Table 2.1). With overall low density, most firms tion and firms, Buenos Aires acts as Argentina’s are concentrated in Pampeana region (Figure transport hub, with most air transportation of 2.6), hosting 77.5 percent of the country’s firms, passengers and road freight flowing through where 52.4 percent of the firms in the Pampeana the capital. Air transport is becoming increas- region are located in AMBA. Even in Pampeana, ingly important to Argentina, with an 85 percent however, firm density is low by international increase in the number of passengers (for domestic standards, with just 16 firms per 1,000 inhabit- and international flights) from 2006-2016 ants. Firm density is even lower in the Northern (ONDAT, 2018). 71 percent of passengers in 2016 regions, with six firms per 1,000 inhabitants used an airport in the Pampeana region, of which (Table 2.1). 61.4 percent had Buenos Aires city 17 as their origin Figure 2.6. Firms are concentrated in the Pampeana region or return destination. In terms of cargo, 91 percent (number of firms by size in Argentina, 2015) of the total airport freight passes through Ezeiza International Airport in Buenos Aires (ONDAT, 2018). Finally, the greatest flows of cargo and passengers are in the corridor connecting Buenos Aires city with Rosario and Cordoba–which also concentrates much of the population and main economic activity of the country.18 The remaining corridors have significantly lower traffic volumes (Ministry of Transport, 2017).

Lack of scale is reflected in hosting mainly local and micro firms with overall low firm density Firm density in Argentina is two to four times lower than in neighboring Chile and Brazil. Lack of scale is reflected in the country’s overall low firm density, particularly in northern Argentina. Argentina has only 13 firms per 1,000 habitants, which is lower than comparator coun- tries19 such as Chile and Brazil, which have 48 and 25 firms per 1,000 habitants respectively

17 Share of the passengers that used the two airports serving AMBA: Aeroparque Jorge Newbury and Ezeiza International Airport. 18 The population for urban agglomerations of Buenos Aires is 13,588,171 people; Cordoba 1,454,645 and Rosario 1,236,086 (Census 2010).

19 Aspirational comparators (Australia and South Korea) and regional peers (Chile, Uruguay, Mexico and Brazil) were selected for the comparison. Source: Authors' calculation from OEDE dataset. 30 | CHAPTER 2 SCALE, SPECIALIZATION, AND CONVERGENCE: WHERE DOES ARGENTINA STAND?

Table 2.1. Firm Density in Argentina is lower than in Northwest, 64 percent of the micro, small, and international comparators medium-sized enterprises (SMEs) are in agricul-

COUNTRY FIRM EVERY 1,000 INHABITANTS ture and supply the domestic market directly or Australia 88 larger corporations with export potential. While South Korea 66 most of the firms in Argentina are microen-

Chile 58 terprises, 51 percent of the private employ-

Uruguay 48 ment is generated by large firms. This pattern

México 34 is not unique though: International experience

Brazil 25 suggests that in most countries, the majority of

Argentina 13 enterprises (between 70% and 95%) are microen-

NOA 6 terprises but the largest share of employment is

NEA 6 in large firms (Ministry of Production, 2018).

Cuyo 10 Lack of scale is also seen in the dominance of Patagonia 14 microenterprises that do not graduate to higher Pampeana 16 size groups. Firms in Argentina stagnate in their

Source: Author’s calculation using OEDE dataset and Fundación size group over time, with only a few graduating Observatorio PyME (2018) to higher size classes. During the 2008-2015 Further, most firms remain local, with microen- period, most firms remained in their size category terprises dominant. For instance, in AMBA, 92 or moved to a lower one. For example, 57 percent percent of firms are classified as local – similar of microenterprises remained a microenterprise to Pampeana (91 percent), and Argentina as a after 5 years and 40 percent closed. Only 3 percent whole (90 percent). Overall, the commerce (93.3 of the firms managed to exceed the threshold of percent), services (91.3 percent), and agricul- 9 employees and transition from microenterprise ture sectors (91 percent) have the highest share to small companies. Moreover, new microenter- of local firms, while construction (87.6 percent), prises have low employment generation. During industry (85.5 percent), and mining (65.5 percent) 2007-2016 new microenterprises went from just have fewer proportion of local firms. The excep- below 2 to 3 employees. In the case of medium tions in terms of local origin are provinces in firms, only 6 percent grew to become large firms. Patagonia with a high share of extractive indus- The fast-growing firms—those that manage to tries, such as Tierra del Fuego or Santa Cruz, increase employment quickly and sustainably- with 83 and 80 percent of firms classified as local are associated with the development or adoption respectively. Moreover, in 2016, there were 9,648 of new technologies, products, and processes. exporting firms in Argentina (1.6 percent of the Between 2010 and 2016, there were an average total number of firms), a smaller figure than in of 3,700 fast-growing firms per year–but that 2006 (15,075 firms). The number of exporting only represents less than 1 percent of the total firms is a measure of the country's international number of firms (Ministry of Production, 2018). insertion and the capacity of its private sector to Fast-growing companies comprise a very small compete with the world. In 2015, 67 percent of proportion of existing companies in all regions firms were microenterprises, 22 percent small, 6 of the country and they are mostly concentrated percent medium, and 5 percent large firms. In the in AMBA (Table 2.2). TERRITORIAL DEVELOPMENT IN ARGENTINA: DIAGNOSING KEY BOTTLENECKS | 31

Table 2.2. Fast-growing firms represent a small share of existing firms across regions, with the highest share concentrated in AMBA

REGIONAL DISTRIBUTION OF FAST-GROWING COMPANIES

REGION/PROVINCE 2000-2002 2003-2007 2008-2011

AMBA 32.0 27.2 29.6

Pampeana 13.0 13.6 13.3

Cordoba 7.3 7.2 7.3

Santa Fe 8.0 8.0 8.0

NEA 4.0 4.8 4.4

NOA 8.0 8.0 8.0

Cuyo 6.7 6.6 6.6

Patagonia 7.0 7.6 7.3

FAST-GROWING COMPANIES AS A PERCENTAGE OF EXISTING COMPANIES IN THE REGIONS

REGION/PROVINCE 2000-2002 2003-2007 2008-2011

a b a b a b

AMBA 0.6 4.6 1.06 27.2 0.82 4.9

Pampeana 0.3 3.2 0.56 13.6 0.37 3.7

Córdoba 0.4 3.8 0.66 7.2 0.45 3.9

Santa Fe 0.4 3.8 0.8 8.0 0.45 3.4

NEA 0.4 3.3 0.96 4.8 0.72 5

NOA 0.8 4.3 1.24 8.0 0.85 5.1

Cuyo 0.6 3.9 0.96 6.6 0.67 4.5

Patagonia 9,7 4.8 1.35 7.6 0.7 4.6

a : as a percentage of total active Firms; b : as a percentage of firms with at least 10 employees Source: IADB (2014) and OEDE.

Identifying absolute advantages is important for when thinking about trade between countries, Scale when the focus is on sub-national development, To create scale in areas that have not been firms and investors look for locations and sectors picked up by the market, economic princi- with an absolute advantage within the country. ples suggest that absolute, rather than rela- This means the location must be the most effi- tive advantages should be the focus of efforts. cient place to produce the given output. Hence, While comparative advantage is a key concept the first step in thinking about what sectors 32 | CHAPTER 2 SCALE, SPECIALIZATION, AND CONVERGENCE: WHERE DOES ARGENTINA STAND?

could be supported to create scale is to take below provides an example of how a comparative stock of the local assets, and advantages revealed advantage alone will not be enough to attract by the current sectors active in a place. Box 2.3 investment to a lagging region.

Box 2.3: The problem of absolute advantage

‘Comparative advantage’ is the ability of a country to produce an output at a lower opportunity cost than a competitor, even if direct costs may be higher (production is less efficient). A country may produce a given output less efficiently than a competing country, but the competitor may have an advantage in producing another output with even higher returns, to which it instead dedi- cates its scarce resources. This allows the less efficient country to nonetheless specialise in the first output, according to its comparative advantage.

At the sub-national level, the scarcity of resources has different dynamics. Many assets are mobile across space, so can move to whichever place offers higher productive efficiency. Within a country, a particular locality could in theory be disadvantaged across all economic sectors. More com- monly, a location may be advantaged in only very low-value addition sectors, unable to satisfy the aspirations of local governments for job creation and local value addition. This is, for example, the case of small commercial activities or non-tradeable services.

To illustrate the point, take as an extreme example a district that encompasses only an isolated desert, whose only resources are sand and unskilled labor, and is therefore only able to produce sand. With nothing to produce except sand, its opportunity costs are zero. If the district were instead a country, its exchange rate would adjust to make the country’s sand competitive for export, and it would specialize in this sector, importing other products.

As a district within the country, by contrast, sand may still be cheaper in another part of the country, or cheaper to import. This leaves no demand for the district’s produce at all. In this situation, the price of immobile assets – such as land, and people unwilling to out-migrate – decline in the region in response to low demand. However, other assets are mobile and leave the less efficient place (for example, labor that migrates to find work).

Unfortunately, immobile factors found in lagging regions, like land or natural resources, are typically a small share of input costs for productive firms, compared to these mobile factors, and thus the falling cost of land is inadequate to draw back productive firms. This contrasts with the national-level case, in which exchange rates effect a global price response to make the economy’s output (with a comparative advantage) competitive.

In economic terms, a locality within a country must have an absolute advantage, not simply a comparative advantage, in what it produces. It must be the most efficient place to produce an output TERRITORIAL DEVELOPMENT IN ARGENTINA: DIAGNOSING KEY BOTTLENECKS | 33

in which it specializes. Lagging regions thus tend to house sectors in which immobile factors like land are a large share of production costs, while sectors with considerable shares of mobile factor inputs (which are typically more productive) concentrate more heavily, in ‘leading’ areas with higher economic development.

LOW SPECIALIZATION IN ARGENTINA – of specialization at the province and city levels. lack of integration between local markets Then the discussion will focus on how this relates to current connectivity challenges within The concentration of people may not be the country. of concern when connections across areas are good. When links between territories are The dominance of the service sector across regions well established, people and goods can move is in line with weak specialization freely between areas, and hills and valleys of As is common in many developed and emer- economic activity will be linked. When places ging economies, services dominate GDP, while are well connected, people are able to reach manufacturing has been declining substan- more economic opportunities, and firms can tially. In 2017, ser vices (light green bars in Figure reach larger markets for inputs and outputs. 2.7) contributed to 50.4 percent of national GDP, When connections are lacking, capital and labor growing 7.5 percentage points (pp) from 2007. markets are segmented, economies of scale Manufacturing (26.2 percent) and the primary limited, and opportunities truncated. Ideally, the sector (6.1 percent) followed (INDEC, 2018). benefits of concentration can be spatially spread Manufacturing has been losing its importance when places are well connected, and connected in all regions, particularly during the nineties; places can either specialize in their most produc- it represented 25 percent of total private-sector tive industries or diversify and keep innovating. employment in 1996 and represents only 19 The question on the role of diversification or percent in 2016 (Figure 2.8a). Partly, this trend specialization on economic growth and develop- is the result of the loss of competitiveness of the ment has been widely explored in the economic sector (Garcia-Herrero et al. 2014). Employment literature. Several studies show that at low levels patterns reflect those of GDP (Ministry of of income, countries are highly concentrated but Production, 2018). In all regions, employment then they diversify (Imbs and Wacziarg, 2003; has also been mainly concentrated in the services Cadot et al., 2011). Hausmann et al. (2011) shows sector20, about 43 percent nationally in 1996 and that countries in the highest level of income 20 The total employment number at the province level is obtained from and development, tend to have highly diversi- the total number of employees declared working in that province by fied export baskets (see also Hidalgo et al., 2007; each firm. OEDE only registers private-sector formal employment. The latter is divided in seven broad sectors (see Nota Metodologica Hidalgo and Hausmann, 2009; Hausmann and 2016 OEDE), according to the 3-level digits by ISIC: Agriculture, Klinger 2006). This section provides a diag- which includes agriculture, cattle raising, and fisheries; Mining, which nostic of the distribution of economic activity represents mining and oil; Manufacturing, which includes light and heavy manufacturing activities; Commerce, which includes retail and across regions in Argentina, assessing the level wholesale trade; Utilities for electricity, gas and water; all Services 34 | CHAPTER 2 SCALE, SPECIALIZATION, AND CONVERGENCE: WHERE DOES ARGENTINA STAND?

48 percent in 2016 (Figure 2.8 b-d).21 Likewise the most industrialized regions of Argentina; the number of firms by sector22 reflects the thus, manufacturing is also visibly more impor- predominance of the services sector across tant (in employment and the number of firms) regions (Figure 2.9). The second largest number there than in the North or Patagonia. of firms is in the commerce sector, reflecting the Services represent about 39 percent of emplo- smaller size of firms here as well as the growth yment in Northern provinces, and their of the sector since the mid-2000s, particularly prominence has grown over time. During the in the Northern regions. twenty-year period between 1996 and 2016, Figure 2.7. Services dominate economic activity in Argentina, 2017 employment in the service sector increased by 4.5 percentage points, to represent close to 39 percent of total private sector employment in the region (Figure 2.9b). The share of employment in the (non-tradeable)23 commerce sector, which includes retail and wholesale trade, has increased since 2006. However, services in the northern regions also include tradeables, and strong service sector growth was likely driven more by the surge in tourism in the region.

Location Quotients can be used to more systematically compare sectoral employment shares by location (explained in Box 2.4). These show that there is relative specialization in agri-

*Note: Sectoral shares of GDP. The dominant services sector is disaggre- culture in the North, and in extractives in gated into its sub-sectors, colored violeta oscuro. Source: INDEC (2018). Patagonia. Employment in the Northern regions Services contribute greatly to the economy in show a significant concentration in agricul- the central Pampeana and Cuyo. Pampeana’s ture (LQ>2), but its concentration of employ- services sector contributes 78 percent to local ment in services is below the national average. In GDP, well above the national average of 50 Patagonia, the mining sector has a much higher- percent. Services dominate private sector than-average share of employment, driven by the employment as well, and constitute almost half strength of the extractive industries in the local of total employment, followed by manufac- economy of the region (See Figure 2.10). turing and commerce. Pampeana and Cuyo are

(low and high-skill); and Construction. 23 Non-tradeable goods are generally defined as those goods that are not 21 Though these figures distinguish commerce from broader services, traded internationally, due to prohibitive transportation costs. were commerce classified as a service sector, as in the GDP figures, its employment share would be even higher.

22 Firms are classified according to the province where they declare employees, which leads to double-counting if firms declare employees in more than one province, as per OEDE methodology. Local firms are those that declare most of their employees inside the given province (>50%). TERRITORIAL DEVELOPMENT IN ARGENTINA: DIAGNOSING KEY BOTTLENECKS | 35

Figure 2.8. Private-sector employment in Argentina is mainly in Services

b. Distribution of employment by sectors, a. Distribution of employment by sectors Northern Provinces (NOA & NEA)

c. Distribution of employment by sectors, d. Distribution of employment by sectors, Patagonia Pampeana

Note: The figure depicts the evolution of formal private-sector employment, broken down into seven economic sectors, between 1996 and 2016 at the national level, then contrasts it with the evolution in three regions: The Northern Provinces, Pampeana region, and Patagonia.24 Source: Data from OEDE, authors’ computations.

Figure 2.9. Lastly, a high proportion of firms in the country are in the service sector

a. Distribution of Firms by sectors b. Distribution of Firms by sectors & regions, 2016

Source: Data from OEDE, authors’ computations.

24 The used was collected by the OEDE “Observatorio de Empleo y Dinamica Empresarial” of the Ministry of Labor, as it contains the latest publicly available data and historical series on the sectorial distribution of employment and firms across provinces in Argentina. The dataset is constructed from the social security registry. 36 | CHAPTER 2 SCALE, SPECIALIZATION, AND CONVERGENCE: WHERE DOES ARGENTINA STAND?

Figure 2.10. The North of Argentina displays a concentration of employment in agriculture

Location Quotient (2016) By Regions

Source: Data from OEDE, authors’ computations. Box 2.4. Measuring the geographical concentration of economic sectors

This section uses Location Quotients (LQ) to examine the sectorial concentration of employment and employment growth drivers across provinces, using the same OEDE dataset. Here, the LQ simply compares the share of local employment (province-level) with the share of national employment in each major sector (agriculture, services, construction, and so on). A value below one indicates that the share of employment in that sector is lower in the province relative to the entire country, and the contrary is true for a value above one. The LQ is a relatively coarse measure, but its simplicity is helpful in identifying high concentration in major sectors.

The analysis in Figure 2.11 combines location economic sectors.25 The region of Pampeana can quotients of employment with GDP growth in be highlighted since its trends are very close to the same sector and location, to identify impor- the national ones (they are closely aligned with the tant drivers of recent and future employment. axis) and its economic dominance increases their Figure 2.11 shows this analysis for every of the 5 25 This analysis was also made by economic sector, seeing the differences regions, where the different symbols represent between regions. The corresponding graphs are in Annex B. TERRITORIAL DEVELOPMENT IN ARGENTINA: DIAGNOSING KEY BOTTLENECKS | 37

weight in the national economy, making the sizes diversification. In Patagonia, we can see that of the figures bigger. By contrast, from the first construction and utilities are important sectors in figure it is noticeable that both Patagonia and the decline, while services and commerce are growing, Northern regions display growing sectors that are albeit timidly. quite distinct from national trends. The Figure also reveals important differences The figure confirms the importance of mining between the Northeast and Northwest, with as the main employment growth driver in NOA’s growth drivers more specialized and Patagonia, not only led by its growth in the past driven by agriculture. The NEA region does not 10 years but by the size of the sector in the region. show any specific outlying sector; the construction There is, in turn, a smaller share of private sector sector figures as the most prominent emerging employment in services or manufacturing. This sector, but private sector employment in construc- can be expected in this extractive-rich region, tion remains low (close to 5 percent). The agri- where extractives represent more than 10 percent cultural sector has been declining in NEA, and of private employment. This heavy dependence by contrast remains an important employment on minerals can lead to a form of sub-national growth driver in NOA, with both a large share Dutch disease, through which minerals drive up of sectoral GDP relative to national averages and costs of land and business inputs, reducing the significantly higher employment (Figure 2.11). competitiveness of other activities and dampening Figure 2.11: Patagonia and Northwest regions differ from national trends

Employment Drivers – All regions Patagonia

Economic sectors labels

Agriculture Manufacturing

Utilities Construction

Commerce Service

Mining

Difference between regional and national sectoral growth (%) (2006-2016) Northeast Northwest

Difference between regional and national sectoral growth (%) (2006-2016) Difference between regional and national sectoral growth (%) (2006-2016) 38 | CHAPTER 2 SCALE, SPECIALIZATION, AND CONVERGENCE: WHERE DOES ARGENTINA STAND?

Pampeana Cuyo

Difference between regional and national sectoral growth (%) (2006-2016) Difference between regional and national sectoral growth (%) (2006-2016)

Note: The analysis combined Location Quotients with provincial DP statistics to identify sectors that have both a higher employment concentration and are growing faster in the provinces compared to the national average (2006-2016). The size of the figures indicates the size of employment in the sector (total number of employees) in the province, and the shapes represent the sectors. The upper-right quadrant then shows the key employment growth drivers – that is, sectors with a higher level of concentration of employment that are also growing faster than the national average in a given province. The upper-left quadrant contains sectors with low growth but with a high concentration of employment, i.e. declining sectors. Emerging sectors are in the lower-right quadrant. Source: Data from OEDE, authors’ computations.

More lagging locations in Argentina do not Further, a relatively high share of Argentina’s urban simply produce for themselves, behaving as autar- jobs depends on non-tradeable sectors, compared to kies, but focus particularly on goods and services international benchmarks (See Figure 2.12).26 The that are difficult to trade. While weak specializa- share of employment in tradeable sectors is lowest tion at the province level is not unique to Argentina, in the Northern provinces. Tradeables in the North the predominance of non-tradeables suggest a diffi- represent between half to two-thirds of the employ- culty transitioning to high value-added activities. ment they represent in Metropolitan Buenos Aires (35 At larger geographical levels, specialization is often percent) (2012-2014) (World Bank, 2017a).27 lower than at more granular scales, reflecting the This is a concern, as tradeable sectors, such as high- broader diversity of the economy of larger areas. For end services and manufacturing, tend to be more example, evidence from the U.S. shows a decline in productive and able to create more jobs and higher specialization at a larger regional level since the incomes. Tradeables can serve larger (more distant) 1930s (Kim 1995). Instead, for locations like small consumer markets and benefit more from special- to medium-sized cities, the logic of trade tends ization and scale. Rather than self-reliance due to to entail increased specialization with improved strong diversification, the reliance on non-tradeables connectivity. Connectivity improves trade with suggests underlying challenges of weak connectivity other centers, enabling each to specialize in that between locations, raising costs to trade and disincen- in which it is advantaged. Thus, in Mexico, for tivizing the development of more tradeable sectors. example, municipalities have been found to have

specialized over time as connectivity increased 26 The tradeables share of employment is lower in Buenos Aires compared in the period 2004-2014 (Blankespoor et al. 2 0 1 7 ) . to Sao Paulo, Bangkok, Istanbul or Seoul, and while close to 70 percent of urban jobs are in Argentina are in non-tradeables, in most cities this Likewise, in Argentina, at the regional level, Figure would be two thirds.

2.11 shows that services are gaining ground in 27 This analysis is based on the EPH data for the period 2003-2014; the both Northern regions and Patagonia, as services OEDE dataset does not allow for differentiating sectors into tradeables GDP growth has been above the national average. and non-tradeable sectors. TERRITORIAL DEVELOPMENT IN ARGENTINA: DIAGNOSING KEY BOTTLENECKS | 39

Figure 2.12. AMBA has a high share of jobs in non-tradeable sectors relative to international comparators

9% 17% 10% 11% 24% 8% 7% 25% 8% 29% 7% 24% 10% 13% 31% 26% 37% 12% 17% 13% 10%

30%

30% 28% 26% 27% 37% 34%

26% 23% 22% 18% 15% 9% 19% 6% 1% 0% 1% 0% 0% 0%

Financial and Transport and Industry Consumer Public services Agriculture businesses services communications services

Note: Composition of employment in AMBA and comparators. Source: Muzzini et al. 2016.

To better understand patterns of specializa- considered as not specialized (Blankespoor et al. tion, we now look at the more local province 2017). Here, we use this threshold at the city level and city levels and employ a more sophisti- but adjust to smaller thresholds at the provincial cated measure of specialization. The Krugman level given that the distribution is more compact Specialization Index (KSI) can be calculated at at this larger scale (see notes of Tables 2.3 and the more province and city level using the high- 2.4). level sectors available in the OEDE dataset. The analysis reveals weak specialization at The KSI is a widely-used, relatively sophisti- the provincial level between 2006 and 2016, cated specialization measure that captures the with specialization becoming weaker over dissimilarity of the local sectoral composition time. The mean and median KSI are very close of employment compared to that of the refer- to 0.35 for all years analyzed (Table 2.3). Further, ence group. In this case, the provincial and city- in line with international trends for these larger level specialization indices are calculated using geographical levels, and trends at the regional as reference the national structure. The KSI can level discussed above, provinces have become less take values between 0 and 2, with 0 indicating specialized during the period considered. Among that the employment structure of a province is the ten Northern provinces, only one province in identical to the national sectoral composition 2006 and two provinces in 2010 displayed KSI and 2 indicating the largest possible difference levels below 0.30 (lighter pink colors in Figure between the two-employment distributions 2.13a) compared to over five provinces in 2016. being compared.28 In general, highly specialized As supported by further analysis below, this localities have indices higher than 0.75, whereas likely reflects a positive trend of diversification localities with an index below 0.35 can be and upgrading of economic activities towards 28 The formula of the KSI is given by: KSIp=i=1|Sp,i-S|; where Sp,i is the employment share of industry i in province p, and S is the average share of industry i in the employment across all provinces in Argentina. 40 | CHAPTER 2 SCALE, SPECIALIZATION, AND CONVERGENCE: WHERE DOES ARGENTINA STAND?

services over time. Only the Patagonian province in 2010 and 2016 (strong red); Chubut, La Rioja of Santa Cruz, which is rich in natural resources, and Santa Cruz were the ones above 0.50 in 2006. and Formosa in the NEA region are above 0.50

Table 2.3. Trends in Employment Specialization for Provinces

2006 2010 2016

Mean 0.344 0.330 0.311

Median 0.356 0.340 0.306

Minimum 0.094 0.084 0.087

Maximum 0.668 0.644 0.563

KSI < 0.3 0.292 0.375 0.458

KSI (0.3; 0.5] 0.583 0.542 0.458

KSI > 0.50 0.125 0.083 0.083

Notes: Mean, median, max and min values of KSI for all 24 provinces (including CABA) in Argentina. KSI <0.30 'not specialized'; KSI>0.5 'highly specialized' are the threshold used for the provinces.

Table 2.4. Employment Specialization – Nodal cities, 2015

NORTHERN ALL PROVINCES PATAGONIA CUYO PAMPEANA

Mean 0.409 0.485 0.403 0.325 0.388

Median 0.389 0.480 0.407 0.305 0.375

Minimum 0.120 0.179 0.121 0.135 0.125

Maximum 0.984 0.984 0.698 0.519 0.713

KSI < 0.35 0.400 0.286 0.400 0.500 0.437

KSI (0.35: 0.5) 0.318 0.238 0.400 0.333 0.333

KSI (0.5; 0.75) 0.247 0.333 0.200 0.167 0.229

KSI > 0.75 0.035 0.143 0 0 0

Note: Mean, median, max and min values of KSI for all 85 nodal cities in Argentina, as defined by OEDE. KSI <0.35 'not specialized'; KSI>0.75 'highly specialized'. 21 nodal cities in the sample are in the Northern provinces, with average (median) population of 300,421 (207,304). Cities in other regions are 10 in Patagonia, 48 in Pampeana, and 6 in Cuyo. The average (median) city-size in the sample is 1,403,967 (152,891) people.

Source: Data from OEDE, authors’ computations. TERRITORIAL DEVELOPMENT IN ARGENTINA: DIAGNOSING KEY BOTTLENECKS | 41

At a smaller scale again, small to medium-sized than in medium-size ones. This suggests that for cities show high levels of employment specia- services firms, the positive forces of agglomeration lization, particularly in the Northern provin- (such as knowledge spillovers, a skilled and diverse ces.29 In Figure 2.13b, light green dots stand for labor force, and proximity to consumers) supersede cities with low specialization index and the orange negative forces of congestion in Argentina’s larger and red for higher specialization. The blue shapes cities. The preference of manufacturing firms for around the city point displays its size. 14 percent medium-sized cities is in line with international of the cities in the North are highly specialized trends and suggests the larger impacts of conges- against only 3.5 percent on average across the tion (such as high land costs), and reduced benefits country, and all highly specialized Northern cities of a highly diverse local economy and labor force, are small to medium-sized. These cities tend to for the sector. See Box 2.5 for international context. be specialized in non-tradeable and agriculture Figure 2.13. Specialization of Employment is weak across sectors, however. Thus, their high specialization provinces in Argentina reflects weak economic diversity and sophistica- a. Provinces (2016) tion, rather than strong market integration. Larger cities, notably Metropolitan Buenos Aires, show low levels of specialization, with KSI at 0.181; these are weakly specialized thanks to their more diver- sified economic base and human capital pool. In Mexico, similar analysis suggests higher specialization, which has been linked to early investments in connectivity infrastructure; over ten percent of cities show high specialization, with a KSI above 0.75 (Blankespoor et al. 2017).

In Argentinean nodal cities30 in 2015 (OEDE),31 employment densities in the services sector are

32 higher in large cities , while manufacturing b. Cities (2015) employment densities are higher in medium- size cities. While this is true across regions of Argentina, the difference is more pronounced in the Northern provinces, where large cities display twice the employment density in the services sector

29 Cities are defined here by OEDE as Local Economic Areas, where smaller cities or “localities” are grouped according to the proximity and employment flows to the closest provincial capital and main city (see OEDE Metodologia AEL for details). 30 See Note of Table 2.6 31 The data only exists as a cross-section and unfortunately it is not possible to measure the evolution of clusters.

32 Large cities have been defined as those with population above 500,000;

while medium size cities are those with population between [100,000; Source: Data from OEDE, 500,000[. The total size of the dataset is 85 cities. authors’ computations. 42 | CHAPTER 2 SCALE, SPECIALIZATION, AND CONVERGENCE: WHERE DOES ARGENTINA STAND?

Box 2.5. Do Service Firms Prefer Small Cities?

In the U.S., Europe, and China, manufacturing and services behave very differently across space, with manufacturing becoming more equally spread across space over time, and the service sector becoming increasingly concentrated in medium-sized locations. Medium-sized cities seem to be a good location for services firms, partly because agglomeration economies dominate congestion forces there (Desmet & Rossi-Hansberg 2009). The service sector is more heavily reliant on information exchange and knowledge spillovers, typically facilitated by dense and diverse local economies and populations. However, an explanation for these trends relates to the impact of Information and Communication Technologies (ICT), which deepens the knowledge spillovers and exchange enabled by spatial concentration and the emergence of clusters of economic activities (Desmet & Rossi-Hansberg 2009).

However, not all fast-growing developing economies display these patterns. India has shown con- trasting trends, with services and (though to a lesser extent) manufacturing increasingly concentrating in larger, high-density clusters (Desmet et al. 2013). Casual observation suggests that while the costs of congestion in India’s mega cities are huge, implying that there should be decreasing returns to further expansion, these mega cities may also benefit from relatively large agglomeration economies compared to medium-sized cities. In turn, medium-sized cities might suffer from market access problems, lack of intermediate goods and infrastructure, and other impediments to fast growth (Desmet et al. 2013).

Source: Desmet et al. 2013 and OEDE Connectivity challenges

Connectivity can support or reduce the growth led to further concentration in the capital city of productive activities and scale in connected (Fajgelbaum and Van Patten, 2018, discussed in locations. In Mexico, for instance, transportation Box 2.2). improvements over the last three decades have Weak physical and soft connective infra- increased access to markets, and these improve- structure is hampering market integration in ments have been found to have had a positive Argentina. Significant spatial disparities in terms impact on employment and the economic special- of access to connective infrastructure persist ization of Mexican localities (Blankespoor et al. across Argentina, and largely align with economic 2017). Evidence from the modern Argentinian under-performance. economy suggests that projects to better connect Argentina’s cities would tend to raise national Hard connective infrastructure is weak, parti- welfare. However, in the late 19th and early 20th cularly in the North and Patagonia. The North centuries, investments that improved connec- suffers lower road density33 than the national

tivity between Buenos Aires and the North only 33 Measured as the ratio of the length of total roads (meters) over the province total area (square meters). TERRITORIAL DEVELOPMENT IN ARGENTINA: DIAGNOSING KEY BOTTLENECKS | 43

average, Pampeana, or Cuyo. Except for Jujuy, road Logistics costs are likewise unequal across density in the NOA region was 8 km and 5.9 km regions, again particularly hampering economic every 100 km2 for national and provincial roads integration of the North. Logistics costs for the respectively in 2015 (Consejo Vial Federal, 2018). northern provinces have been estimated to be The Patagonia region also ranked low in this indi- up to 50 percent higher than in the neighboring cator, with Chubut being the province with the Center and Cuyo regions (World Bank, 2016a). lowest total road density in the country (3.7 km The distances between the provincial capi- every 100 km2). When looking at the quality of tals of the NOA and the port of Buenos Aires, roads,34 91 percent of the total roads in Argentina the main container port of the country, exceed were paved in 2015, with less spatial variation 1,000 km. Due to these distances and the lower between regions in terms of this metric. economic activity in the North, only 25 percent of the freight traveling from North to the South Figure 2.14 shows that areas in Patagonia, returns directly to the Northwest, either carrying together with areas East and West of the Jujuy- inputs or finished goods (Minister of Transport Salta-Cordoba corridor, have the lowest levels of Argentina, 2017). Instead, after unloading in of road accessibility and relatively higher travel Buenos Aires or Rosario, trucks attempt to mini- times. In contrast, provinces in central Argentina mize unproductive routes by diverting to other are well connected, reflecting a dense transporta- cities, leading to significant delays on the South‐ tion network between cities. North route (World Bank, 2016a). Further, during Figure 2.14. Travel times are consistently longer throughout the North and Patagonia the last twenty years, the cargo transported by the Belgrano railway network35 has declined dramatically. Towards the end of the 1980s, it moved around 4 million tons, which represented a quarter of the total rail cargo in the country. After years of disinvestment and constant changes in its operational management, in 2008 it only transported 0.9 million tons - less than 4 percent of the country's total rail freight (World Bank, 2017a). Moreover, according to the Logistics Performance Index (LPI)36 in the 2016 Argentina ranked 66th, after India (35), Kenya (42), Mexico (54), Brazil (55), and Uganda (58). Within the LIP, Argentina had the lowest performance in customs and international shipments. Argentina had also a low performance in infrastructure, which is not surprising given their low investment in the transport sector. In 2015, public and private investment in the

35 The Belgrano Railway Network is the largest of the Argentine railway Notes: The map measures the travel time, using land or water-based travel, to the nearest city of more than 50,000 inhabitants.. Source: OSM network and connects the capital city with the Northern regions.

36 A comprehensive measure of the efficiency of international supply 34 Percentage of national routes paved out of the total number of routes chains. (World Bank, 2016b) 44 | CHAPTER 2 SCALE, SPECIALIZATION, AND CONVERGENCE: WHERE DOES ARGENTINA STAND?

transport sector represented less than 2 percent There is also wide variation in provinces’ integra- of the GDP in Argentina, while in the same tion with international markets, with Pampeana year it represented between 10 and 12 percent in best integrated, and the North East particularly Colombia and Peru.37 poorly connected. National exports are dominated by Pampeana, contributing 74 percent Across provinces, accessibility to local markets to total exports. Together, the NOA and NEA is positively correlated with economic activity. regions account for just 7 percent of the country’s Figure 2.15 shows that there is a positive correla- total exports. The index of international production between access to markets38 and two different tion openness compares a locality’s trade share measures of economic activity across provinces of GGP with the national trade share in GDP, of Argentina, even when controlling for associ- with one indicating an equal trade share.39 For ated factors such as natural endowments, and 2015, the Northeast had an index of just 0.348, the education of the workforce. This underlines indicating a much smaller share of exports and the importance of improving connectivity within imports than national averages. The Northwest, lagging regions and connecting such areas with by contrast, had an index closer to the national more successful locations. average, at 0.925.

Figure 2.15 Market access is positively correlated with local value-added across provinces

Notes: The figure plots the relationship between measures of economic activity at the provincial level against market access by road at the same spatial level (N=23). The measures of economic activity are (left) (log) calibrated Night-Lights (Viirs) 2010 aggregated at the provincial levels; and (right) (log) of local GDP 2010 (PPP 2005). Market access is a (log) market access index (time-distance measured travelling by car). Values are residuals; elevation, ruggedness of the terrain, and total years of education of the population, have been partialled out. Coefficients in both graphs are statistically significant at 1% level. Source: NOAA (2015), INDEC (2010), 2010 Census, authors’ computations

37 Infrolatam database 39 The index is developed by INDEC based on data from CEPAL. It 38 The markets considered include fairly small towns of 50,000 people, measures the trade insertion of the province and is computed as the making this a measure of local-level connectivity rather than connec- share of provincial value-added aimed at exports over the same ratio tivity to major national hubs or international markets. for the country level. TERRITORIAL DEVELOPMENT IN ARGENTINA: DIAGNOSING KEY BOTTLENECKS | 45

In the case of ‘soft’ connective infrastructure, Neuquén and Chubut displaying close to 2 mobile the NOA and NEA regions are the most disad- lines per person and Santa Cruz less than 1 (all vantaged in terms of internet access, hindering in Patagonia). In AMBA there were 4.5 lines per integration. In 2016, 21 percent of people had person.41 residential broadband connections, well below Weaknesses in institutional barriers also dampen the national average of 41 percent. In compar- integration ison, 53 percent of people had residential broad- Further, institutional barriers are found to be band connections in Patagonia, and 50 percent in a key constraint for further investments and AMBA. Since 2016 the Government of Argentina private sector expansion. Analysis using infor- (GoA) has implemented the Federal Internet Plan mation from the Argentina Enterprise Surveys with the goal of reducing service gaps throughout for 2010 and 2017 suggests that firms both in the country by installing a 33,000 Km federal the informal and formal sector highlight insti- network of fiber optics; 22 percent of the connec- tutional barriers as the key obstacles for their tion points are in the northern provinces.40 The regular operation. Among the key issues high- same differences are observed in other telecom- lighted are factors such as tax rates, tax admin- munications: there were on average 146 mobile istration, political instability, and corruption. lines for every 100 people in Argentina in 2017, Box 2.6 below provides additional details on this but only 119 in the Northern provinces compared analysis. for instance to 152 in Patagonia. But these numbers mask inequalities within regions, with

Box 2.6. Quality of institutions is the main limitation for private sector development in Argentina.

Tables 2.5 and 2.6 display the main obstacles to business reported by formal and informal firms in

Argentina (respectively). 42 Obstacles are divided according to four categories: infrastructure, regulations, factor markets, and institutions. The color scale classifies the severity of the obstacle from green to red, with obstacles that are deemed ‘major’ or ‘severe’ by less than 30 percent of firms in green, those thus classified by 30-59 percent of firms in yellow, and those seen as major or severe by more than two thirds of firms in red.

The main obstacle relates to institutional factors such as tax rates, tax administration, political instability, and corruption. In the case of informal firms (Table 2.6), access to electricity is the main obstacle across both cities. This obstacle is seconded by institutional factors, in this case crime and corruption.

41 By the end of 2016, 78.9 percent of people in cities in Argentina used 40 http://datos.arsat.com.ar/dashboards/19767/plan-federal-de-internet/ a cellphone. (National Survey on Access and Use of Information and Communication Technologies. INDEC [2016]).

42 World Bank Enterprise Survey, 2017d. Here, we only present manufacturing and services sectors, given that most firms in the sample are in these two categories. 46 | CHAPTER 2 SCALE, SPECIALIZATION, AND CONVERGENCE: WHERE DOES ARGENTINA STAND?

Firms across the country reported similar bottlenecks for operations in the formal and informal43 sectors, with some small variations by firms’ size and sectors. Relevant for next chapter: the low quality of institutions is also the main constraint for growth in the NOA, where Salta and Jujuy are located.

Table 2.5 Argentina: Main obstacles for formal private firms’ performance

SECTOR MANUFACTURING SERVICES

SIZE SMALL MEDIUM LARGE SMALL MEDIUM LARGE Infrastructure Electricity 49 52 31 56 33 16

Telecommunications 40 39 43 52 37 31

Transport 17 35 23 26 14 14 Regulations Customs and Trade Regulations 15 27 22 18 12 8

Practices of competitors in informal sector 36 41 22 22 28 12

Labor Regulations 48 60 58 46 39 47 Factor Markets Access to Land 9 13 21 16 10 25

Inadequately Educated Workforce 27 45 34 48 21 29

Access to finance 26 31 21 15 24 18

Institutions

Crime 14 13 5 14 7 24

Tax Rates 75 87 74 73 67 81 Tax Administrations 49 71 61 62 52 77

Business Licensing and Permits 13 14 12 36 21 53

Political Instability 54 50 53 66 44 50

Corruption 39 45 38 63 42 66

Courts 30 41 30 44 35 78

N 222 240 150 119 83 59

Scale

Note: percentage of firms answering that the variable is a major or very severe obstacle; sampling weights used. Small firms are those with less than 19 employees, medium sized firms are firms with between 20 and 99 employees; large firms have at least 100 employees. The colors are scaled according to all values in the table from smallest percentage to largest (green to red).

43 See Appendix 2 for descriptive statistics of informal firms. TERRITORIAL DEVELOPMENT IN ARGENTINA: DIAGNOSING KEY BOTTLENECKS | 47

Cuadro 2.6. Principales obstáculos en negocios informales

SECTOR TOTAL BUENOS AIRES CHACO*

Infrastructure

Electricity 50 48 54

Water supply 24 21 29 Market factors

Access to land 23 23 24

Access to finance or loans 31 32 29 Institutions

Crime 44 38 62

Corruption 34 27 53

N 384 284 100

Note: percentage of firms answering that the variable is a severe obstacle. *Chaco is the province, while the city is Resistencia.

Factors of the institutional business environment were also regressed against employment growth outcomes for Argentine firms. Access to finance, labor regulations, infrastructure and security are positively correlated with employment growth for Argentine firms. While statistical significance of point estimates is not robust across specifications, some trends can still be inferred. For instance, increasing trade credit (access to finance) is positively associated with firm-level employment growth, while security costs, power outages, and obstacles in labor regulations are negatively associated with employment growth, suggesting these factors might be important for firms’ performance across cities in the sample. Results for labor productivity are less clear, surely due to higher endogeneity in these regressions, and conclusions are harder to draw. Findings in Reyes et al. (2017) across cities also find basic protections, access to finance, and infrastructure to be critically important factors for firm growth. While we could not do a more granular analysis by age, size, and sector affiliation due to sample size limitations, the original paper finds the effect of business environment to be heterog- enous according to these firms’ characteristics. Their results point at the fact that small firms tend to need a stronger Basic and Refined BE than larger and medium size firms; and that firms in the service sector are more sensitive to most elements of BE relative to those in manufacturing.

Source: Data from OEDE, authors’ computations 48 | CHAPTER 2 SCALE, SPECIALIZATION, AND CONVERGENCE: WHERE DOES ARGENTINA STAND?

Argentina also faces significant obstacles to (1.66) and Mexico (1.91). In terms of regulatory competition. The Product Market Regulation barriers, those related to state control, e.g., public (PMR)44 indicator goes from 0 to 6 (least to most ownership and involvement in business operation restrictive), for which Argentina had an index score (price controls and control regulations), repre- of 3.11. This is relatively restrictive compared to sented the largest share in Argentina.45 (See Box other OECD countries and other countries in the 2.7 for more measurements of competitiveness region such as Colombia (1.77), Chile (1.51), Peru and business environment in Argentina). Box 2.7. Competitiveness in Argentina

Argentina ranks 92nd out of 137 countries included in the World Economic Forum’s Global Competi- tiveness Index (GCI) and 116th out of 190 countries included in the World Bank’s Doing Business (DB) report. Both indices are useful tools to identify and monitor challenges to competitiveness and identify key bottlenecks for increasing competitiveness.

Despite its overall low performance, Argentina’s competitive conditions have improved in the GCI 2017-2018, climbing 12 positions in the rankings. The largest improvements were in the basic requirements sub-index (+7) and the Innovation and sophistication sub-index (+12). However, Argentina is still at the bottom of the world ranking in many crucial factors of competitiveness. Moreover, Argentina ranks towards the bottom of the distribution in all pillars related to efficiency of markets, both for final goods and services and for factors of production (Figure 2.16). The World Bank’s DB report provides areas of policy prioritization to improve the business environment and boost productivity and growth. The DB indicators’ scores reveal the main challenges for Argentina to be in resolving insolvency (41.24), paying taxes (49.34), dealing with construction permits (49.27).

Figure 2.16: Argentina in the Global Competitiveness Report and Doing Business Report a. Argentina is below the median in LAC for the pillars of infrastructure, macroeconomic environment, and those linked to the efficiency of markets

44 The PMR indicators database is a comprehensive and internationally- Source: authors with data from the Global Competitiveness Index comparable set of indicators that measure the degree to which policies 2017-2018 from the World Economic Forum. promote or inhibit competition in areas of the product market where competition is viable. 45 Koske, I., et al. (2015). TERRITORIAL DEVELOPMENT IN ARGENTINA: DIAGNOSING KEY BOTTLENECKS | 49

b. The main challenges in Argentina are Resolving Insolvency, Paying Taxes, and Dealing with Construction Permits.

Source: authors with data from the World Bank Doing Business 2018 report

Argentina’s performance in both indices reveals early results from efforts and reforms to improve competitiveness and the business environment, but also reveals challenges and a long road ahead that will require the strengthening of institutions, coordination mechanisms within government and between government and the private sector, and policy prioritization for adequate sequencing of reforms to achieve the greatest impact and build on early wins.

Additionally, to make progress, Argentina requires subnational measures of competitiveness. Many of the factors that determine competitiveness at the national level, but more importantly, one of the drivers of economic geography and the high concentration of economic activity, is differences in conditions between provinces and regions. The Global Competitiveness Report Framework lends itself to subnational implementation. Countries like Colombia have successfully implemented the framework at the province and city level. The Doing Business framework for assessing the challenges in local business environments can also be implemented at the city level and expanding it beyond Buenos Aires. Measuring competitiveness at the subnational level would allow for more targeted policies as well as drive learning and reform by example and imitation.

Source: Adapted from the Global Competitiveness Report 2017-2018, World Economic Forum, and Doing Business Report (2018)

LACK OF CONVERGENCE IN concentration to other regions, many others have ARGENTINA – some progress is visible, not. In Colombia, the multidimensional poverty but gaps still exist index suggests large disparities across regions, with areas like Bogota having poverty rates lower Spatial concentration of people and firms can than 12 percent while departments in the Pacific be harnessed to improve living conditions have rates of over 40 percent (DANE, 2016). In everywhere. While some countries have managed Source: authors with data from the Global Competitiveness Index some countries the disparities are even larger: to extend the benefits brought by such economic 2017-2018 from the World Economic Forum. poverty rates in states in the south of Mexico such 50 | CHAPTER 2 SCALE, SPECIALIZATION, AND CONVERGENCE: WHERE DOES ARGENTINA STAND?

as Chiapas, Oaxaca, and Guerrero range between than 80 percent of the population do. A composite 76.4 and 66.4 percent, while the poverty rate in index of access to ser v ices for water and sanitation 46 the northern state of Nuevo Leon is 14.5 percent confirms that the Northern provinces are below (CONEVAL, 2018). national averages, with the provinces of Santiago

Gaps remain in key indicators del Estero, Chaco, and Formosa displaying values up to 2 standard deviations below the mean, and There are large spatial disparities in terms of AMBA showing values almost 1.5 standard devia- access to basic services. While access to water has tions above the mean. converged over the years, important gaps remain between the Northern regions and the rest of the Inequalities in living standards in the northern country in terms of access to sanitation and basic regions of Argentina services (Figure 2.17). The Northeastern prov- Despite overall poverty reduction, large dispari- inces (Chaco, Formosa, Santiago del Estero) have ties in living standards persist in Argentina, with the lowest proportions of households with access the North particularly lagging. Even today, the to water: 75 percent. This is not that far from the provinces in northern Argentina struggle with proportion of households with water access in extreme poverty and considerable gaps in living the city of Buenos Aires, where the proportion is Figure 2.17. Convergence in water access, but lower convergence in sewage access, by province over time a. Water access rates by province over time b. Sewage access rates by province over time

Source: authors with data from 2010 Census of Argentina

99 percent (World Bank, 2018b). However, in terms standards when compared to the rest of the 47 of sanitation, only 18 percent of the population in country. While overall poverty rates have

the Northeastern province of Misiones have access 46 The index is computed as the mean of z-scores of three variables to sewage, while in Chaco and Formosa, fewer measured by INDEC at the provincial level. These 3 variables are (i) the percentage of households with access to drinking water; (ii) the than 35 percent do. While in the Northwest, prov- percentage of households living in a dwelling connected to sewages; inces like Salta and Jujuy have a better coverage and (iii) the percentage of households with a toilet inside their (64.71 and 63.35 percent, respectively), in the city dwelling. 47 Poverty level here refers to the one calculated by INDEC (National of Buenos Aires 98.73 percent of the population Institute of Statistics and Censuses) through the regular household has access to sewage. In the provinces of Tierra del survey “Encuesta de Hogares permante (EPH). Here we use data Fuego, Santa Cruz, and Chubut in Patagonia, more from the 1st semester 2017. The caveat with this dataset is that it only TERRITORIAL DEVELOPMENT IN ARGENTINA: DIAGNOSING KEY BOTTLENECKS | 51

been falling since 2002, poverty incidence in Northern regions still display the highest shares the Northwest and Northeast regions are at and absolute numbers of households with unmet 25.4 and 33.2 percent respectively and are the basic needs. Only localities in the Northern prov- highest in Argentina (See Figure 2.18). At the inces have more than 35 percent of households provincial level, Santiago de Estero (in the NOA with unmet basic needs; in one outlier, a locality region) has the highest poverty incidence in the in Formosa, 68 percent of households have unmet country, followed by Corrientes (NEA). Further, basic needs, followed by one in Salta, where 49 the North is home to the highest share of house- percent of households do. holds with unmet basic needs (Necesidades Figure 2.19. At the district level, pockets of poverty are found Básicas Insatisfechas, NBI)48 – 15.9 percent of across the territory a. Poor Places: Percentage of district house- households, compared to 9.1 percent nationally. holds that have unmet basic needs The Northern provinces of Formosa (19.7) and Salta (19.4) have the highest share of unmet basic needs nationally. At a smaller level of aggregation (department or district levels) (Figure 2.19), Figure 2.18. Households with unmet basic needs are predominantly in the North

b. Poor people: Percentage of the total number of poor people in Argentina living in each district

Notes: Percentage of households with unmet basic needs at the province level, as measured by INDEC’s NBI (2010). Source: INDEC (2010).

measures poverty across urban areas. The index of unmet basic needs presented later is the only nationally representative measure, calculated during the Census 2010.

48 These are households meeting at least one of the following criteria: (1) precarious housing; (2) no access to private toilet in their dwelling; (3) living in conditions of overcrowding; (4) at least one child between ages 6 to 12 does not go to school; and/or (5) the household head did not finish primary school and there is a ratio of 4-1 per employed member. Censo Nacional de Población y Vivienda (INDEC), 2010. Notes: Households with unmet basic needs as measured by INDEC’s NBI (2010). Source: INDEC (2010) 52 | CHAPTER 2 SCALE, SPECIALIZATION, AND CONVERGENCE: WHERE DOES ARGENTINA STAND?

Within the North, rural areas lag significantly areas can require distinct policy tools, compared behind their urban counterparts. For example, to locations with more widespread poverty, like in Formosa, 31.3 percent of households in rural the North.

a rea s had u nmet ba sic needs in 2 0 10 , compa red to Challenges for lagging regions to effectively equip 17 percent in urban areas; for Salta the numbers their populations with skills for productive work were 36.6 percent and 16.9 percent respectively. Primary education has been improving, shor- Similar patterns are observed across Argentina, tening the gap between leading and lagging with consistently higher living standards in areas. In 1970, AMBA already had a primary urban versus rural locations in every province. education completion rate of 90.2 percent, while Although poverty is more common in the the provinces of Chaco, Misiones, Formosa, North, there are pockets of distress across Santiago del Estero, Neuquén, and Corrientes Argentina, including within some of the had completion rates lower than 55 percent; most successful locations. For instance, in the province of Chaco had the lowest, at 46.8 Patagonia, but also in the central regions and percent. However, there has been a convergence Buenos Aires (Pampeana Region), there are in primary education rates nationally. According districts with large numbers of households with to the last census (2010) the national average is unmet basic needs, as visible in Figure 2.19b. at 92.3 percent, and the range at provincial level Despite its higher average living standards, from 88.5 to 97.4 percent (Figure 2.20). AMBA houses a large number of the country’s However, the Northern region still lags in human poor. Three districts in the AMBA have the capital development, with provinces in the North largest total number of households with unmet Figure 2.20. Convergence in primary education, less convergence in secondary education a. Primary completion rates by province over time b. Tasa de finalización de escuela secundaria por provincia a lo largo del tiempo

Source: authors using data from the 2010 Census in Argentina systematically ranking among the bottom prov- basic needs of any district (i.e. Distrito Federal, inces in education indicators. Buenos Aires enjoys La Matanza, and General Sarmiento49). Such near-full primary completion, at 97.4 percent. Two pockets of distress in otherwise more successful Northern provinces, Misiones and Santiago del Estero, have the lowest primary completion rates 49 The district of General Sarmiento was split into three districts in 1995; however, it was still used in the 2010 Census. TERRITORIAL DEVELOPMENT IN ARGENTINA: DIAGNOSING KEY BOTTLENECKS | 53

nationally, at 86 and 86.9 percent respectively These differences in human development (data from the 2010 Census of Argentina). All the outcomes affect livability and productivity provinces in the North have secondary comple- (World Bank, 2018a). Evidence indicates that tion rates below or at 50 percent. Of all provinces human capital is a foundational source of produc- below the median (45.8 percent), seven are in the tivity across cities throughout Latin America. North, with the remaining three in Patagonia, one On average, an additional year of education in in Pampeana, and one in Cuyo. Similarly, the illit- countries of the region raises nominal wages by eracy rate in the region (at 3.7 percent) is twice as 9.2 percent, and an additional percentage point high as the national average. in the share of adults with college education raises wages by 2 percent (Ferreyra and Roberts Also, relevant for human capital, the North [eds.], 2018). Further, holding density and access lags in health indicators and health services constant, a 1-percent increase in skill (measured provision. Nationally, there are approximately as average years of schooling) is associated with 0.14 health centers and 0.19 schools per 1000 a 0.57 percent increase in city productivity across people, yet the Northwest region has just 0.04 LAC cities. While the contribution of skill to health centers and 0.09 schools per 1000 people productivity varies across countries, it is signifi- (World Bank, 2017b). The NOA region has one of cant and positive for all LAC countries (Ferreyra the lowest percentages of population with medical & Roberts [eds.] 2018). insurance: only 52.5 percent and 59.9 percent of the population in Santiago del Estero and Jujuy However, even given human capital, institu- have medical insurance, compared to 93.4 and tional constraints such as poor land markets, 86.1 percent in Santa Cruz and Tierra del Fuego information or communication can prevent (both in Patagonia) and 85.3 percent in AMBA workers from pursuing economic opportuni- (INDEC, 2017). Regarding health outcomes, the ties across space. This is suggested by recent infant mortality rate (at 16.2 per 1000 live births) work by Quintero & Roberts (2017), which found is well above the national average of 13.3 per that in Argentina, many workers who would gain 1000. Salta and Jujuy have the highest maternal higher nominal wages by relocating (given their mortality rate in the country (8.1 and 7.5 per skills and expected returns to those skills in more 10000 live births respectively), followed by Chaco productive places) do not do so (Box 2.8). Misiones and San Juan, all in the North.

Box 2.8. Evidence of Soft Constraints to Worker Mobility

Quintero & Roberts (2017) estimate the impact of an urban versus rural location on the productivity of workers in 15 Latin American and Caribbean (LAC) countries (including Argentina). They assess the degree to which productivity differences (measured through nominal wages) between urban and rural areas can be attributable to the different characteristics of urban workers (“worker premium”), or other factors associated with cities (“urban premium”). They 54 | CHAPTER 2 SCALE, SPECIALIZATION, AND CONVERGENCE: WHERE DOES ARGENTINA STAND?

find that across the countries, the average rural nominal wage would be 38 percent higher if the characteristics of rural workers were changed to be the same as those of urban workers.

Their results in Argentina suggest that much of the difference in productivity across cities is due to the sorting of workers with higher skills into larger cities, rather than other advantages offered by those high-productivity cities. This follows a similar pattern as other LAC countries. Figures 2.21 a and b show the results of this work for Argentina, with the left panel showing overall differences in productivity, and the right panel showing the “urban premium” after accounting for differences in workers characteristics. On both panels, the location of the dots indicates the cities included in the sample. On the left panel, darker shades of orange indicate a higher “urban premium” or higher productivity. As expected, and as shown in panel a, productivity is typically higher in and around larger and more densely populated cities, such as Buenos Aires. Further, productivity is 3 to 4 times lower in cities in the Northern provinces. However, once differences in worker characteristics are accounted for, the differences in productivity—as measured by the true “urban premium”—are much lower (see panel b).

However, differences in productivity do not disappear after controlling for sorting of workers between areas (Figure 2.21), suggesting that factors other than the chance to gain higher nominal wages affect worker location. This could result from workers sacrificing wages in return for better amenities (such as services or remaining near family and friends) (Rosen- Roback model) (Glaeser, 2007), which, all else being equal, could persuade them to accept a lower-wage in a higher-amenity Northern city. Given lower amenities in most Northern cities, however, it is likely that more external barriers also prevent workers from moving freely across space to work, preventing the optimal allocation of resources, and exacerbating the large inequalities across areas. Thus, weaknesses in ‘soft’ connectivity (such as institutions, land markets, information, and communication) may also be encouraging a suboptimal allocation of labor and capital across locations, dampening incomes and productivity. TERRITORIAL DEVELOPMENT IN ARGENTINA: DIAGNOSING KEY BOTTLENECKS | 55

Figure 2.21. Differences in productivity persist after controlling for concentration of workers between different areas (subnational variations in nominal wages)

a. Without controlling for differences in b. Controlling for differences in worker worker characteristics characteristics

Source: Fereyra y Roberts (2018)50 Given the considerable differences in by thriving areas could benefit from infrastruc- economic profiles and outcomes between ture improvements that link its product and labor regions, the natural question is how to markets to the thriving area. Alternatively, lack achieve a more balanced territorial deve- of connective infrastructure, geographical busi- lopment. Different types of outcomes, and ness environment, or social divisions may prevent interventions, are appropriate for different the benefits of economic activities in one district places; it depends on their characteristics and from spilling over its borders and benefiting the their neighbor’s characteristics. For example, broader regional or national economy. an underperforming district surrounded Moreover, some densely populated areas may

50 Points correspond to major cities/urban agglomerations. remain disconnected from economic opportu- Location premium in the maps is calculated as exp and nities even when connective infrastructure is expressed in 2005 PPP exchange rates, where is the estimated coefficient from a series of country-specific regressions on provided, due to factors such as social divisions, a location dummy, Li,li,t, which takes the value one when a institutional friction, and lack of appropriate worker i lives in a location l in the year t and zero otherwise. human capital. This is often the case for pockets These regressions also include survey-year fixed effects. In (a), the “location premium (before sorting)” is the location of deprivation within large cities. Here, improve- premium without controlling for observable worker char- ments to connective infrastructure can help acteristics. In (b), the “location premium (worker sorted)” is connect, for example, slum dwellers to the main the location premium after controlling for observable worker characteristics (age, age^2, years of schooling, gender, marital center of employment, and better service delivery status) and survey-year fixed effects. may raise health and human capital. In addition, 56 | CHAPTER 2 SCALE, SPECIALIZATION, AND CONVERGENCE: WHERE DOES ARGENTINA STAND?

however, targeted actions may be required. For García-Herrero, A., Dos Santos, E., Urbiola, P., Dal Bianco, M., example, better functioning land markets may be Soto, F., Hernández, M., Rodríguez, A. and Sánchez, R., 2014. Competitividad del sector manufacturero en América Latina: needed to enable people to migrate, or to improve tendencias y determinantes. Documento de Trabajo, (14/10).

their own neighborhoods, while more place-based GeoData Institute (2015). Conjuntos de datos de WorldPop. interventions such as social and urban integration Universidad de Southampton, Southampton, Reino Unido. of vulnerable neighborhoods programs or even job Glaeser, E. L. (2007). Entrepreneurship and the City (No. w13551). creation may be necessary to improve the overall National Bureau of Economic Research. condition of the poor in such areas. Thus, inter- Hausmann, R., & Klinger, B. (2006). The evolution of comparative advantage: the impact of the structure of the product ventions for scale, specialization, and conver- space. Center for International Development and Kennedy gence may be warranted where their respective School of Government, Harvard University. challenges collide. , Hidalgo, C., Bustos, S., Coscia, M., Chung, S., Jimenez, J., Simoes, A. & Yildirim, M. A. (2011). The Atlas of Economic Complexity: Mapping Paths to Prosperity. Cambridge, MA. REFERENCES Henderson et al. (2012). Measuring Economic Growth from Outer Space. American Economic Review 102(2), pp. 994-1028. Blankespoor, B.; Bougna Lonla, T.; Garduño Rivera, R.; Selod, H. (2017). Roads and the Geography of Economic Activities in Hidalgo, C. A. & Hausmann, R. (2009). The Building Blocks of Mexico (inglés). Documento de trabajo sobre investigaciones Economic Complexity. Proceedings of the National Academy of de políticas n.º 8226. Washington, D.C.: World Bank Group. Sciences of the United States of America, 106(26): 10570-10575.

Cadot, O., Carrère, C., & Strauss-Kahn, V. (2011). Export diver- , Klinger, B., Barabási, A. & Hausmann, R. (2007). The sification: what's behind the hump?. Review of Economics and Product Space Conditions the Development of Nations. Statistics, 93(2), 590-605. Science, 317(5837): 482-487.

Consejo Vial Federal (2018). Información técnica: Red vial IADB (Interamerican Development Bank) (2014). Dinámica nacional. Buenos Aires, Argentina. http://www.cvf.gov.ar/ empresarial y emprendimientos dinámicos: ¿contribuyen al red_vial_provincial.php empleo y la productividad? El caso argentino. Washington, D.C. https://publications.iadb.org/handle/11319/6581#sthash. DANE (Departamento Administrativo Nacional de Estadística) acWN5LHu.dpuf (2016). Boletín Técnico: pobreza monetaria y multidimensional en Colombia (2016). Bogotá, Colombia. Imbs, J. & Wacziarg, R. (2003). Stages of Diversification. American Economic Review, 93(6), pp. 63-86. Desmet et al. (2013). Urban Accounting and Welfare. American Economic Review, 103(6), pp. 2296-2327. INDEC (Instituto Nacional de Estadística y Censos) (2010). Censo Nacional de Población, Hogares y Viviendas. Buenos & Rossi-Hansberg, E. (2009). Spatial Growth and Industry Aires, Argentina. Age. Journal of Economic Theory, 144(6), pp. 2477-2502.} (2016). Encuesta Nacional sobre Acceso y Uso de Duranton, G., & Puga, D. (2004). Micro-foundations of urban Tecnologías de la Información y la Comunicación 2016. agglomeration economies. In Handbook of regional and urban Buenos Aires, Argentina. economics (Vol. 4, pp. 2063-2117). Elsevier. (2017a). Encuesta de Hogares Permanente (4.º trimestre). Fajgelbaum, P. & Redding, S. J. (2018). External Integration, Buenos Aires, Argentina. Structural Transformation and Economic Development: Evidence from Argentina 1870-1914 (n.º w20217). National (2017b). Incidencia de la pobreza e indigencia en 31 Bureau of Economic Research. aglomerados urbanos. Segundo Semestre 2017. Buenos Aires, Argentina. https://www.indec.gob.ar/uploads/informesde- Ferreyra M. M. & Roberts, M. (eds.) (2018). Subamos el estándar prensa/eph_pobreza_02_17.pdf para ciudades productivas en América Latina y el Caribe. Washington, D.C.: World Bank Group. (2018). Informe de avance del nivel de actividad. Estimaciones provisorias del PIB para el cuarto trimestre de Fundación Observatorio PyME (2018). Informe especial: 2017. Buenos Aires, Argentina. https://www.indec.gob.ar/ natalidad empresarial y desarrollo productivo. Buenos uploads/informesdeprensa/pib_03_18.pdf Aires, Argentina. http://www.observatoriopyme.org.ar/project/ natalidad-empresarial-y-desarrollo-productivo/ Chief of Cabinet of Argentina (2016a). Relevamiento Nacional de Barrios Populares. Diciembre de 2016. Buenos Aires, Argentina. TERRITORIAL DEVELOPMENT IN ARGENTINA: DIAGNOSING KEY BOTTLENECKS | 57

(2016b). Presentación: el Desarrollo del Norte Argentino a Reyes, Jose Daniel; Roberts, Mark; Xu, Lixin Colin. 2017. The través del Plan Belgrano. Buenos Aires: Jefatura de Gabinete, Heterogeneous Growth Effects of the Business Environment Unidad Plan Belgrano. : Firm-Level Evidence for a Global Sample of Cities. Policy Research Working Paper;No. 8114. World Bank, Washington, Kim, S. (1995). Expansion of markets and the geographic distri- DC. bution of economic activities: the trends in US regional manufacturing structure, 1860–1987. The Quarterly Journal of Rosenthal, S. S., & Strange, W. C. (2004). Evidence on the Economics, 110(4), 881-908. nature and sources of agglomeration economies. In Handbook of regional and urban economics (Vol. 4, pp. 2119- Koske, I. et al. (2015). The 2013 Update of the OECD’s Database 2171). Elsevier. On Product Market Regulation: Policy Insights for OECD and Non-OECD Countries. Documento de trabajo del WorldPop (2015). High-resolution gridded population data- Departamento de Economía de la OCDE, n.º 1200, OECD sets for Latin America and the Caribbean, Scientific Data. Publishing, París. http://dx.doi.org/10.1787/5js3f5d3n2vl-en www.worldpop.org

Ministry of Finance of Argentina (2018). Sistema de Indicadores World Bank (2016a). Argentina - The challenge of reducing logis- de Desarrollo Provincial (SIDEP). Dirección Nacional de tics costs. Washington, D.C.: World Bank Group. Asuntos Provinciales. Buenos Aires, Argentina. (2016b). Connecting to Compete: trade logistics in the Ministry of Production of Argentina (2015). Fichas Provinciales global economy (English). Washington, D.C.: World Bank 2015. Buenos Aires, Argentina. Group.

(2016). Informe Productivos Provinciales: CABA. Buenos (2017a). Diagnóstico de Logística para el NOA y priori- Aires, Argentina. zación de centros logísticos a lo largo del Corredor Vial del NOA. Washington, D.C.: World Bank Group. (2018). Datos crecimiento. Buenos Aires, Argentina. (2017b). Documento de evaluación de proyecto: Corredor Ministry of Transport of Argentina (2015). Plan Estratégico Norte (P163115). Washington, D.C.: World Bank Group. Federal de Transporte, Movilidad y Logística. Buenos Aires, Argentina. https://issuu.com/institutoargentinodeltransporte/ (2017c). Strengthening Argentina’s Integration in the Global docs/pfetra2015_569cdd20df840b Economy: The Role of Trade, Investment and Competition Policies. Washington, D.C.: World Bank Group. (2017). Estudio nacional del flujo de carga. Buenos Aires, Argentina. (2017d). Encuesta de Empresas en Argentina (Enterprise Survey). Washington, D.C.: World Bank Group. Muzzini et al. (2016). Leveraging the Potential of Argentine Cities: A Framework for Policy Action (inglés). Direction in (2018a). World Development Report 2017 – Learning to Development. -- Countries and Regions. Washington, D.C.: Realize Education’s Promise. Washington, D.C.: World Bank World Bank Group. Group.

NOAA (National Oceanic and Atmospheric Administration) (2018b). Sytematic Country Diagnostic Argentina. (2015). Base de datos National Geophysical Data Center. Washington, D.C.: World Bank Group. Silver Spring, MD. (2018c). Informe Doing Business 2018. Washington, D.C.: OEDE (Observatorio de Empleo y Dinámica Empresarial). World Bank Group. Conjunto de datos de Dinámica del empleo y rotación de World Economic Forum (2018). The Global Competitiveness empresas. Ministerio de Trabajo de Argentina. Buenos Report 2017–2018. Washington, D.C. Aires, Argentina.

ONDAT (Observatorio Nacional de Datos de Transporte) (2018). Base de datos sobre transporte urbano. Centro Tecnológico de Transporte, Tránsito y Seguridad Vial - Universidad Tecnológica Nacional. Buenos Aires, Argentina. http://ondat.fra.utn.edu.ar/

Quintero, L. & Roberts, M. (2017). Explaining Spatial Variations in Productivity: Evidence from 16 LAC Countries. Documento de antecedentes. Washington D.C.: World Bank.

Redding, S. J. & Turner, M. A. (2015). Transportation Costs and the Spatial Organization of Economic Activity. En Handbook of Regional and Urban Economics (vol. 5, pp. 1339-1398). Elsevier. 58 | SECTION NAME CHAPTER 3

Scale, Specialization, and Convergence: a closer look at the Provinces of Salta and Jujuy

PURMAMARCA IN JUJUY, ARGENTINA Telmo Filho © TERRITORIAL DEVELOPMENT IN ARGENTINA: DIAGNOSING KEY BOTTLENECKS | 59

CHAPTER 3 Scale, Specialization, and Convergence: a closer look at the Provinces of Salta and Jujuy

The North of Argentina lags relative to other Road Development Corridor Project which has regions in terms of welfare and economic been recently launched by the government with the activity. The analysis of the spatial disparities support of the World Bank. in the previous chapter identified the North as a Figure 3.1: Salta and Jujuy are two provinces in the Northwest lagging region in Argentina. The Northern region51 (NOA) of Argentina. only contributes to 11 percent to national GDP52 and 7 percent of total exports (INDEC, 2016), while it is home to 20 percent of the population. In terms of welfare, the provinces in the North continue to struggle with extreme poverty and considerable gaps in living standards when compared to the rest of the country. Other indicators such as access to services, education, and health, also show that the northern regions still lag in human development outcomes.

However, not all the provinces in the North, and particularly in the NOA, exhibit the same trends. This chapter focuses on the provinces of Salta and Jujuy (highlighted in Figure 3.1), and their capital cities, Salta and San Salvador de Jujuy. These provinces and cities are important within the region, and display several interesting spatial characteristics, such as their international borders and internal diversity. These provinces were selected since (i) they are among the poorest, and face deeper challenges in terms of achieving convergence of living standards; (ii) they represent an economic oppor- Source: INDEC tunity for the Northern region and (iii) they are in the direct area of influence of the Northwestern LOW SCALE IN SALTA AND JUJUY – firms’ low growth and productivity 51 The Northern Provinces refer to the provinces in the NEA (Chaco, Corrientes, Misiones y Formosa) and NOA (Jujuy, Salta, Catamarca, The provinces of Salta and Jujuy host a high Tucuman y Santiago del Estero) provinces of Argentina, that is the Northeast and Northwest, respectively. concentration of people within the Northern

52 GDP in constant prices of 2004; Ministry of Production (2015). 60 | CHAPTER 3 SCALE, SPECIALIZATION, AND CONVERGENCE: A CLOSER LOOK AT THE PROVINCES OF SALTA AND JUJUY

regions of Argentina, but their firms’ low growth Figure 3.3. The per capita Gross Geographical Product in and productivity imply lack of scale Salta and Jujuy is lower than the national average, 2005

Despite contributing less than 2.5 percent to national GDP, the provinces of Salta and Jujuy, and their capital cities, are hubs of economic activity and population within NOA. With 4.8 percent of the national population, the provinces of Salta and Jujuy contribute less than 2.5 percent of national GDP (1.5 and 0.8 percent respectively). Their per capita GGP is less than half the national average (Figure 3.3) (Ministry of Finance of Argentina, 2018). However, Salta and Jujuy are important provinces within Argentina’s lagging Note: Current prices (last data available). Northwest (NOA). They occupy Argentina’s Source: Muñoz and Trombetta (2015), Ministry of Finance (2018) Northwestern corner, with three international The cities of Salta and San Salvador de Jujuy borders (Chile to the West, Bolivia to the North, are the capitals of Salta and Jujuy provinces and Paraguay to the Northeast), and a combined respectively, and also the largest cities in their area more than half the size of Paraguay. 53 Within provinces. The provinces of Salta and Jujuy have the Northwestern region (NOA), they represent 87.1 and 87.5 percent of their population living 41.2 percent of the population,54 and 42 percent in urban areas (INDEC, 2010).The city of Salta of the economic activity (26.9 and 15.3 percent of is the largest capital of the Northwest region, Geographical Gross Product- GGP-, respectively and the seventh largest city in the country (with – see figure 3.2). Only Tucuman has a higher 521,483 people, according to the 2010 census). It GGP within the region (Figure 3.2), while smaller lies within the Metropolitan Area of Salta (MAS), 55 Catamarca has the highest GGP per capita which compromises more than half of the popula- (Figure 3.3). tion of the province (around 616,367 people). San Figure 3.2. Salta and Jujuy represent 42 percent of the regional Geographical Gross Product Salvador de Jujuy is smaller than the city of Salta, ranking 15th in terms of population size nationally, with 237,751 inhabitants in the 2010 census 7.9% 15.3% (INDEC, 2010). It is located at the junction of the Xibi River and the Río Grande de Jujuy, and lies on 33.6% National Route 9 that connects La Quiaca (180 mi) on the Bolivian border with the province of Salta 26.9% (75 mi). It is 1,525 km (948 mi) from Buenos Aires. 16.3% The cities of Salta and San Salvador de Jujuy are important economic hubs for the region (Figure

Note: Share of regional Geographical Gross Product (GGP)

53 Salta’s area is 155,499km2 54 1,888,479 combined population 55 Which includes seven other municipalities

TERRITORIAL DEVELOPMENT IN ARGENTINA: DIAGNOSING KEY BOTTLENECKS | 61

3.4).56 The Local Economic Areas (AELs) of Salta57 Figure 3.4. Economic activity in the provinces of Jujuy and and San Salvador de Jujuy58 represent most of the Salta is concentrated in their respective capital cities: San Salvador de Jujuy and the city of Salta. private economic activity within their respective provinces, hosting 78 percent and 82 percent of the total firms in the formal sector respectively. Within the region, Salta stands out as a more important economic hub, ranked 2nd in the region in terms of the number of firms, compared to 4th for Jujuy. Likewise, in 2015, Saltahosted 20 percent San Salvador of the jobs in NOA, compared to 10 percent in San de Jujuy Salvador de Jujuy; Salta’s job share was topped only by San Miguel de Tucuman. The concentration of City of Salta firms follows the same patterns, with 20 and 10 percent of total employment located in Salta and Jujuy, respectively.

However, on the national scale, the city of Salta’s Source: NOAA, 2015. GDP per capita is low. The city of Salta59 has a Figure 3.5. Putting things into perspective: San Salvador de GDP per capita of US$ 5,341, while the GDP per Jujuy, city of Salta, and Buenos Aires. capita of top six largest cities in Argentina ranges from US$16,351 (city of Buenos Aires) to US$6,724 Buenos Aires (San Miguel de Tucuman). Despite there being no specific figures for San Salvador de Jujuy, differ- City of Salta ences in economic activity using nightlights are shown in Figure 3.5, where Buenos Aires’s spike is considerably higher than those in San Salvador de Jujuy and the city of Salta. Box 3.1 benchmarks San Salvador de Jujuy economic performance in the city of Salta.

56 All following economic statistics at the city level refer to the AELs of Salta and San Salvador de Jujuy, as the OEDE data is organised at this level. Source: NOAA, 2015. 57 Salta’s AEL has the City of Salta as a node and 9 additional municipalities from 5 departments: from the department Rosario de Lerma, Lack of scale is evidenced by the low growth and Campo Quijano and Rosario de Lerma; from Cerrillos, Cerrillos; from productivity of firms, high informality and unem- Chicoana, Chicoana and El Carril, from General Guemes, El Bordo and General Guemes, from the Capital, Villa San Lorenzo, and Gobernador ployment, and low skill levels Manuel Sola. 58 San Salvador de Jujuy’s AEL has San Salvador de Jujuy as a node and 8 Small, local firms dominate in the provinces of additional municipalities from 6 departments: from the department of Salta and Jujuy, though this reflects trends across Tilcara, Maimara and Tilcara; from Palpala, Rio Blanco and Palpala; Argentina. The previous chapter highlighted that from El Carmen, Perico; from San Antonio, Perico de San Antonio; from Tumbaya, Purmamarca; from Doctor Manuel Belgrano, Yala. across the country, in both leading and lagging 59 Source: Oxford Economics Database, 2017. There is no data available provinces, the vast majority (around 90 percent) of for San Salvador de Jujuy. 62 | CHAPTER 3 SCALE, SPECIALIZATION, AND CONVERGENCE: A CLOSER LOOK AT THE PROVINCES OF SALTA AND JUJUY

firms are of local origin, and most are also SMEs. municipality of Salta is 45.1 percent, the highest in In NOA, as in the rest of the country, about 94 the country (IADB, 2015). Informality is typically percent of registered companies are SMEs, and associated with low-scale firms that are poorly informality in the provinces of Salta and Jujuy is specialized, entailing low productivity, invest- high. In fact, the share of informal firms in the ment, and survival (UCA, 2014). Box 3.1. Benchmarking Economic Performance in the City of Salta

To put Salta’s economic performance in context, we select a set of similar, or similar but aspirational, cities from the region and world to act as comparators.

Selection of Comparators Direct comparators for Salta at the regional and international levels were selected based on the following criteria: GDP per capita, Industry and Public Services share of GGP, and economic characteristics such as Tourism relevance and Industrial Parks development. First, all cities with a GDP per capita between 1 and 1.5 times that of Salta in 2015 were selected. From that list, a subset was selected, to include all cities with a Public Services and Industry share of Gross Value Added (GVA) between 25-60 percent, excluding capital and largest cities. After this process, four cities were selected as direct comparators (Arequipa in Peru, Ibagué in Colombia, Bloemfontein in South Africa and Kuantan in Malaysia), and the rest were excluded on the following grounds: Size/relative economic importance in the country, no focus on tourism and different structural dynamics compared to Salta (e.g. growing share of transport, financial services, no agriculture sector, too industrialized, low public services employment). Aspirational comparators were identified as examples of a potential economic path for the city, such as an important focus in Tourism with a strong industrial sector. Only secondary cities in upper middle-income countries were considered.

Table 3.1 Descriptive data on Salta and its comparators

SHARE OF AVG. GGP AVG. EMP POPULATION GDP PER CAPITA NATIONAL GDP GROWTH GROWTH 2015 (MILLONES) 2015 (USD) 2015 (PERCENT) (2005-15) (PERCENT) (2005-15) (PERCENT)

Salta 0.6 5,300 0.5 3.2 2.3

Arequipa 0.9 8,400 3.6 7.4 2.1 Direct comparators LAC Ibagué 0.5 5,900 0.7 4.9 2.7

Bloemfontain 0.9 7,800 1.7 3.3 1.0 Direct comparators Global Kuantan 0.5 9,200 1.2 5.4 3.5

Aguascalientes 0.8 12,200 0.8 4.1 2.4 Aspirational cities Merida 1.0 10,000 0.8 2.6 2.2

Source: Oxford Economics database, 2017. TERRITORIAL DEVELOPMENT IN ARGENTINA: DIAGNOSING KEY BOTTLENECKS | 63

Regional direct comparators (Latin America)

Arequipa (Peru): Arequipa is an important commercial and industrial hub for the southern Andes. It serves as a link between the coast and the mountains and hosts three Parque Industriales. Over the past century, many factories related to the primary sector have emerged in the city, especially in textiles and agriculture.

Ibagué (Colombia): The economy of Ibagué is based primarily on the industrial, tourism, and agricultural sectors, with its textile industry being the third largest in Colombia.

Bloemfontein (South Africa): Bloemfontein is the judicial capital of South Africa. The city’s economy is driven by government sector with some of South Africa's largest retail companies’ headquarters located in the city.

Kuantan (Malasia): Kuantan is a transportation link between Indochina, India, and China. It is now one of the largest special economic zones (SEZs) in Asia, covering 390,000 hectares and aiming to generate around $23 billion of investments and provide over 90,000 new jobs by 2020. One of its major economic activities is tourism.

Global direct comparators

Aguascalientes (Mexico): Capital of its state, Aguascalientes has one of the most business- friendly environments in Mexico, receiving more than 200 multinationals in the last three decades. Moreover, Aguascalientes is a cultural town, with national and international tourism especially around its most important yearly festival, the Feria de San Marcos.

Merida (Mexico): Capital of Yucatan, the city of Merida hosts the best historical and natural attractions in Mexico. Its main native industry is agricultural, coming from the henequen harvest.

Performance Vis-à-vis Comparators

The regional comparator countries selected have similar sectoral compositions and populations, yet Salta´s GDP per capita is by far the lowest among them (Figure 3.6a). Additionally, Salta’s disposable income growth during the last 5 years has been far lower than any of the comparators, staying below 1 percent (Figure 3.6b), showing low dynamism in the economy. This is not surprising given the weak performance of the city’s GDP over that same period (growing only 0.7 percent), and the stagnation of sugar production in the province over those same years (Ministry of Finance, 2017). 64 | CHAPTER 3 SCALE, SPECIALIZATION, AND CONVERGENCE: A CLOSER LOOK AT THE PROVINCES OF SALTA AND JUJUY

Figure 3.6: Salta’s GDP per capita and disposable income has been growing slower than comparators

a. GDP per capita, 2015 and GDP growth 2011-2015: Salta and comparators GDP growth 2005-2015 (5-year moving average)

PIB per cápita 2015 ($)

775 world cities Direct comparators LAC Aspirational cities

Direct comparators Global Salta

b. Disposable income per capita 2015 & growth trend 2011-2015: Salta and comparators Disposable income per capita (2015, USD) Disposable income per capita growth (annualized, 2011-2015) growth capita per (annualized, income Disposable

Source: Oxford Economics database, 2017. TERRITORIAL DEVELOPMENT IN ARGENTINA: DIAGNOSING KEY BOTTLENECKS | 65

Salta’s overall productivity (measured as value added per worker) is low, and has been declining since 2008 (Figure 3.7). Labor productivity in Salta is the lowest within the group of comparators (except for Ibagué) (Table 3.2). Despite their importance in employment shares, labor productivity in agriculture and industry are far below national averages in Salta.

Table 3.2. Productivity in Salta is the lowest among comparators, except for Ibague. (value added per worker in USD, 2015)

TRANSPORT & FINANCIAL & CONSUMER PUBLIC CITIES TOTAL AGRICULTURE INDUSTRY COMMUNICATIONS BUSINESS SVCS SERVICES SERVICES

Salta 12,921 5,346 16,000 12,708 16,889 13,973 12,404

Arequipa 15,203 8,000 28,673 12,741 26,821 8,169 12,385

Ibagué 11,976 10,400 20,558 9,435 21,120 5,660 12,213

Bloemfontain 23,515 12,000 23,193 54,438 43,250 17,460 20,402

Kuantan 21,874 25,000 35,341 19,455 17,632 21,257 13,612

Aguascalientes 28,296 20,333 44,980 39,500 50,909 14,319 17,836

Merida 20,310 6,429 23,030 29,760 47,389 13,676 15,835

Source: Oxford Economics database, 2017. *No data for San Salvador de Jujuy.

Figure 3.7. Productivity has decreased since 2018, productivity change 2005-2015

Aguascalientes Merida Arequipa Bloemfontein Ibague Kuantan Salta

Source: Oxford Economics database, 2017 66 | CHAPTER 3 SCALE, SPECIALIZATION, AND CONVERGENCE: A CLOSER LOOK AT THE PROVINCES OF SALTA AND JUJUY

Lack of scale is evidenced by the low growth and The provinces of Salta and Jujuy have even productivity of firms, high informality and unem- lower firm density than the national average. ployment, and low skill levels The provinces of Salta and Jujuy have 6 and 5 firms per 1,000 people, which are among the lowest Small, local firms dominate in the provinces firm densities in the country. In the Northwest of Salta and Jujuy, though this reflects trends region Catamarca and Tucuman have a higher across Argentina. The previous chapter high- firm density (7 firms per 1,000 people) and Buenos lighted that across the country, in both leading Aires has the highest firm density at the provincial and lagging provinces, the vast majority (around level (Figure 3.8). 90 percent) of firms are of local origin, and most are also SMEs. In NOA, as in the rest of the Firm under-performance is partially reflected in country, about 94 percent of registered compa- elevated unemployment rates and reduced labor nies are SMEs, and informality in the provinces of participation in the city of Salta. Unemployment Salta and Jujuy is high. In fact, the share of is 7.9 percent in Salta municipality, and 5.6 percent

Figure 3.8. Low firm density in Salta and Jujuy, compared with other provinces in Argentina

Source: IADB (2015) and OEDE. informal firms in the municipality of Salta is 45.1 in San Salvador de Jujuy. These rates are compa- percent, the highest in the country (IADB, 2015). rable to (or better than) the national unemploy- Informality is typically associated with low-scale ment rate of 7.2 percent, and similar to other firms that are poorly specialized, entailing low municipalities in Argentina such as Cordoba and productivity, investment, and survival (UCA, Rosario (around 7 percent). However, labor force 2014). participation is somewhat lower than average TERRITORIAL DEVELOPMENT IN ARGENTINA: DIAGNOSING KEY BOTTLENECKS | 67

in the Northwest, at 56.6 percent compared to LIMITED SPECIALIZATION IN SALTA 58 percent nationally, reflecting a slightly higher AND JUJUY – public sector services share of discouraged workers who become inac- dominance and weak connectivity tive. The rates are also considerably higher than The public sector dominates the service sector, shel- some of the comparator cities in Latin America. tering Salta and Jujuy’s economies and limiting For example, Arequipa (Peru) and Merida (Mexico) opportunities for scale and specialization have respective unemployment rates of 4.9 percent and 3.3 percent, despite similar economic profiles. As on the national and provincial levels, services sector is the largest component of the Skill levels of the labor force in the cities of Salta Geographical Gross Product in the provinces and San Salvador de Jujuy remain low. Over 40 of Salta and Jujuy. Services sectors represent 65 percent of the labor force in both cities has only and 61 percent of the total GGP in the provinces incomplete secondary education (see Figure 3.9 of Salta and Jujuy respectively (versus 50.4 percent panel b). This is quite important, given that lack nationally), and a majority of Salta and Jujuy’s of human capital is one of the main concerns of private employment was also in services in 2016 the productive value chains of the North (CEPAL, – with 57 percent and 52 percent, respectively. As 2017). discussed in the previous chapter, services GGP is Figure 3.9: The city of Salta has high levels of labor force growing faster in the North than nationally. The participation, but low levels of skills high services share reflects the national trend a. Working-age population in Salta and Comparators: 2015. highlighted earlier of higher services shares in large60 cities, particularly in the North. The major private service sectors (representing 8-11 percent of local GGP) in Salta include transportation and ICT (11 percent), commerce (9 percent), financial services (9 percent), and real estate and professional services (8 percent). In Salta, the next largest sector, agriculture, has just 12 percent of GGP, followed by manufacturing with 9 percent.61 Also in Salta, tourism appears as an b. Level of skills in Salta. important growth sector. Tourism accounted for 15 percent of employment in 2015, and empirical analyses suggest that the implementation of the Tourism Development Policy between 2003 and 2013, generated an average increase in employment in the sector of 11 percent per year (Castillo et al. 2015). Tourism is particularly important as

60 Large cities were defined as those with populations above 500,000

61 GGP disaggregated: Public services (24%), Agriculture (12%), Transportation (11%), Manufacturing (9%), Financial Services (9%), Commerce (9%), Real estate services (8%), Utilities (6%), Construction Source: Oxford Economics database (2017) and INDEC (2010). (5%), Mining (3%), Hotels and Restaurant (2%), and Other services (6%). 68 | CHAPTER 3 SCALE, SPECIALIZATION, AND CONVERGENCE: A CLOSER LOOK AT THE PROVINCES OF SALTA AND JUJUY

a tradeable service sector; as highlighted earlier, 2016 and 2017). Fruits, soy, oil, and sunflowers are the North is particularly reliant on non-trade- other important agricultural sectors in Salta and ables, offering limited opportunities for scale and Jujuy (CEPAL, 2017). specialization. A large proportion of Salta and Jujuy’s services Although typically seen as a low value-added sectors involve public services. Public services sector, the agricultural industries of the are almost a quarter (24 percent) of total GGP provinces of Salta and Jujuy include produc- in Salta province. Moreover, 36.5 percent of the tive firms which process agricultural outputs, city of Salta’s residents were employed by the operate at scale, and are integrated into larger public sector in 2015, whereas the median public- vertical supply chains. In the sugar cane industry, sector employment share (at the provincial level) for instance, the five sugar mills of Jujuy and Salta in Argentina was 34.6 percent in 2015. This is produce about 95 percent of total sugar output reflective of the wider NOA region (highlighted in the NOA region, responsible for 11.6 percent in green in Figure 3.10); however, the city of Salta of Gross Provincial Product. Jujuy is responsible is relying less on government employment than for 24.6 percent of national sugar production (the other NOA provinces (except for Tucuman, which second biggest producer in the country), while is below the Argentinian median with 32 percent Salta produces 11.6 percent. Salta and Jujuy also employed in the public sector). This share distin- have important tobacco and bean industries. Salta guishes the city of Salta from its comparator cities, is responsible for 28.3 percent of the national even considering cities like Bloemfontein in South tobacco production and Jujuy for 41 percent. Salta Africa, which is the judicial capital of the country substantially dominates in beans, producing 75.3 (Bloemfontein has around 9.6 percentage points percent of the national bean value added, whereas lower employment in public services than Salta) Jujuy produces only 7 percent (Ministry of Finance, (Figure 3.11d). The share of public employment Figure 3.10. Public employment in Salta and Jujuy is higher than the Argentine median, with more than half of Jujuy’s employment relying on the public sector.

Note: Salta and Jujuy are highlighted in red, the other three provinces in the Northwest are in green, the Argentinian average and median are in black, and the rest of provinces are in blue. Source: MECON, 2017. TERRITORIAL DEVELOPMENT IN ARGENTINA: DIAGNOSING KEY BOTTLENECKS | 69

is even higher in Jujuy, where over half the labor figures), created just eight thousand jobs in the force are publicly employed (51.7 percent).62 Only period. Public employment does not appear to be four provinces in Argentina have higher shares of creating positive multiplier effects in the local public employment than Jujuy. economy. International experience suggests that continued reliance on the government for produc- The city of Salta’s rate of job creation is compa- tive activities indicates low levels of productivity rable to that in the selected comparator cities, in a territory, and an unsustainable burden on but most of them are in the public sector (Figure public accounts (Rodríguez-Pose, 2018). Typically, 3.11), hindering productivity and impacting the power of a city or a province to create jobs and public accounts. More than 60 percent of jobs spur growth depends on its ability to attract and (30,000 jobs) created in Salta from 2005-2015 expand the private sector (Gaubert, 2018). were in the public sector. The next largest job creator, tourism (‘consumer services’ in the below

Figure 3.11. Public sector drives GVA and employment in Salta. a. Evolution of GVA by sector: Salta. b. Gross Value Added in 2015: Salta and comparators

100% 100% 90% 90% 80% 80% 70% 70% 60% 60% 50% 50% 40% 40% Share of total GVA 30% 30% 20% Share of gross value added, 2015 20% 10% 10% 0% 0% 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 Bloemfontein

Public services Finance & business services Transport, storage, ICT Consumer services Industry Agriculture

c. Growth rate and number of jobs created between 2005 d. Share of total employment 2015. and 2015: Salta and comparators 100%

90% 80% 70% 60% 50% 40% 30% Number of jobs ('000) 20% Share of total employment, 2015 of Share 10% 0% Bloemfontein

Number of jobs (‘000) Growth rate of jobs Source: Oxford Economics database, 2017.

62 Due to data availability, total employment used to compute the share on public services is excluding informal activity, employees in private households, and independent workers. 70 | CHAPTER 3 SCALE, SPECIALIZATION, AND CONVERGENCE: A CLOSER LOOK AT THE PROVINCES OF SALTA AND JUJUY

There are signs the government has attempted In Salta province, agriculture provides 18 to diversify the economies of the provinces of percent of private jobs, while in Jujuy, agri- Salta and Jujuy during this economic slump. culture employs 16 percent. At the city level, Salta inaugurated an industrial park in 2011, agriculture is less important in the city of Salta a year in which Salta saw 554 new firms – the compared to the wider Salta province (11 percent greatest increase in the number of private firms of employment); this is expected, with urbaniza- in the province over 2009-2015 and reflecting 7.3 tion accompanied by the shift to off-farm sectors. percent firm growth. However, between 2013 and However, in Jujuy, agriculture is concentrated in 2015, Salta lost 33 firms, which needs to be looked San Salvador de Jujuy´s AEL—that includes eight at further to avoid additional challenges. San other municipalities—which has 85 percent of Salvador de Jujuy has also announced a Special the province’s agricultural jobs. Its agricultural Economic Zone in the area, aiming to boost and employment shares are likewise slightly higher diversify employment and productivity. than those in the wider province, at 19 percent

Agriculture, however, is an important economic versus 16 percent in the province (Table 3.3).

driver for the provinces of Salta and Jujuy Jujuy province has an important manufacturing Services has the largest share of private emplo- base, providing 20.8 percent of its private sector yment in both Salta and Jujuy, representing employment. Salta’s manufacturing employ- 57.3 and 52.4 percent respectively. Within the ment share is considerably lower, at 13.5 service sector, commerce (generally a non-trade- percent; by number of firms, it has a large share able service) had the largest share of employment, of the Northwest’s manufacturing firms, though with 18.1 and 17.1 percent, respectively. A similar this is influenced by the small size of those firms. pattern is seen at the city level, with 19 percent Within Jujuy, manufacturing is in fact concen- of jobs in commerce in the city of Salta63 and 20 trated not in San Salvador de Jujuy, but the percent in San Salvador de Jujuy64 (Table 3.3). smaller towns of Libertador General San Martin and San Pedro de Jujuy; these concentrate 74 Table 3.3. Private sector employment in service sectors is dominant in both provinces percent of Jujuy’s manufacturing employment, compared to 26 percent in San Salvador de Jujuy. SALTA JUJUY Agriculture 18.2% 15.9% Thus, manufacturing represents just 8 percent of Mining 1.2% 3.3% employment in San Salvador de Jujuy, compared Manufacturing 13.5% 20.8% Utilities 0.8% 0.8% to 21 percent at province level. The strength of Construction 9.0% 6.8% Commerce 18.1% 17.1% the manufacturing base of Jujuy is important as Hotels and restaurants 4.4% 2.7% manufactured goods are relatively tradeable and Transportation & ICT 6.8% 8.9% Financial services 2.3% 1.8% their production congenial to scale and special- Real estate and professional services 9.4% 6.6% Education 7.4% 6.6% ization (Table 3.3). Health and social services 3.7% 4.1% Other services 5.2% 4.7% As well as considering static sectoral shares, Note: services sector in blue. Figure 3.12 puts the provinces of Salta and Jujuy’s Source: Data from OEDE, authors’ computations sectoral performance in a national context and offers a more dynamic analysis of employment 63 29 percent of firms, or 8,552 firms 64 28 percent of firms, or 4,375 firms growth drivers. This uses Location Quotients TERRITORIAL DEVELOPMENT IN ARGENTINA: DIAGNOSING KEY BOTTLENECKS | 71

(LQ), which were applied earlier in the previous With this analysis, it is clear that agriculture chapter at the provincial level, to compare the is considerably more important in the prov- share of provincial employment in a sector with inces of Salta and Jujuy’s employment profiles national employment share in the same sector; a than nationally (LQ>2). This mirrors the finding higher score shows that a sector is more dominant of the previous chapter for both LQs and KSIs, in the province than is typical nationally.65 The which showed agriculture to be an important figure also compares how local employment in the employment driver in the Northwest and may sector is growing over time, compared to national include a higher impact of agro-processing and employment growth in the sector. agricultural diversification.

Figure 3.12. As in the rest of the NOA region, agriculture is the main employment driver in the provinces of Salta and Jujuy.

Location Quotient vs. Employment Growth Analysis 2006-2016 in Salta and Jujuy AELs

Agriculture Commerce Manufacturing Services Utilities Mining Construction

65 A value below one indicates that employment in that sector is less Source: Data from OEDE, authors’ computations. concentrated in the province relative to the entire country, the contrary is true for a value above one. 72 | CHAPTER 3 SCALE, SPECIALIZATION, AND CONVERGENCE: A CLOSER LOOK AT THE PROVINCES OF SALTA AND JUJUY

Weak connectivity to global, national, and regional Although the province of Salta is doing well markets is holding back productivity gains in the within the region, given the position of both provinces of Salta and Jujuy, limiting specialization provinces with three international borders between them, the export shares for both Salta Within NOA, the province of Salta is relati- and Jujuy are disappointing. Connecting to vely well externally-connected, while nearby foreign markets is key to boosting the scale and Jujuy lags. The previous chapter highlighted specialization of production and can increase a that Northern provinces contribute just 7 percent region’s productivity and employment. Chile is an of national exports, and that the Northeast has a important trading partner for Argentina, being particularly low share of exports in its local GPD, its fifth largest export market, while exports to whereas the Northwest has a trade share closer to Bolivia and Paraguay are just 1 percent and 1.7 the national average. The Northwestern trend is percent of Argentina’s total exports respectively. reflected in Salta, whose export share in GGP is In the case of each export market, trade may be almost identical to the national average (0.98 on dampened by the relatively poor economic perfor- the Index of International Production Openness mance of the regions across Salta/Jujuy’s borders, – higher than average for the region). Within the but it is also possible Salta and Jujuy are not North, Salta is topped only by the province of taking full advantage of the trade opportunities Santiago del Estero in the North. On the other created by their borders. Barriers to trade should hand, Jujuy has an index of only 0.75, well below be diagnosed and addressed. The limited connec- the NOA average of 0.925 (Figure 3.13). tivity and related high cost of transport in the

Figure 3.13. Salta and Jujuy have a larger share of exports relative to other provinces in the region

3.5

2.5

1.5

Índice de apertura de la producción 0.5

Source: Data from INDEC, authors’ computations. TERRITORIAL DEVELOPMENT IN ARGENTINA: DIAGNOSING KEY BOTTLENECKS | 73

Northern province appear to be bottlenecks for with 64 percent of roads unpaved (Federal Road local economic growth. Council o Consejo Vial Federal in Spanish, 2014).

Hard and soft connective infrastructure are weak Salta and Jujuy’s provincial-level roads are partic- within and between Salta and Jujuy ularly low quality, with only 12.5 and 14.8 percent paved, respectively (Figure 3.14). These statistics The provinces of Salta and Jujuy suffer poor suggest that Salta and Jujuy are relatively poorly road infrastructure. Nationally, Argentina connected internally, which hinders local-level has 2.98 km of paved roads66 per 100 km2 (2016) trade, scale, and specialization. (INDEC, 2016). This figure is already lower than international norms by almost an order of magni- The provinces of Salta and Jujuy are also gene- tude, influenced though not fully explained by rally behind in soft connective infrastructure, Argentina’s concentrated population. The equiva- particularly internet and computer access. lent figure in Salta was just 1.48km per 100km2, Nationally, there are 40.8 broadband connections and in Jujuy 2.38. Salta province in fact has the per 100 inhabitants, whereas Jujuy and Salta lowest road density in the NOA region. Likewise, have almost half as many (24 and 22.2 respec- Salta and Jujuy have a much higher than average tively). Nationally, there are 145.7 cellphone lines share of unpaved roads (25 percent in Salta and 37 per 100 people, whereas in Salta there are 130.6 percent in Jujuy, compared to 9 percent nation- (Salta is ranked 17th out of the 23 provinces in ally). Their high share of paved roads makes them the country). Jujuy’s cellphone connectivity, by second and third worst in the country on this indi- contrast, is higher than the national average, at cator, with only Tierra del Fuego below them 151 cellphones per 100 people; given Jujuy’s Figure 3.14.: 52.4 and 75.2 percent of the provincial roads in Salta and Jujuy are not paved

Paved Improved Dirt

66 Primary and secondary networks are considered, excluding local and Source: Consejo Vial Federal (2014). municipal roads. 74 | CHAPTER 3 SCALE, SPECIALIZATION, AND CONVERGENCE: A CLOSER LOOK AT THE PROVINCES OF SALTA AND JUJUY

generally poorer socio-economic indicators, expanding local sections of National Route (NR) this may reflect residents’ reliance on cellphone 34, NR 66 and NR 1V66, to a four-lane highway connectivity given poor transport and internet (from two-lanes) with stronger all-weather connectivity. surfacing, divides, and improved maintenance. The project aims to better link Jujuy’s cities At the city level, cellphone access is slightly, with the province of Salta, and with national and computer access far, weaker than national transport route NR 34 (which runs North to the averages. 90 percent of households nationally border with Bolivia, and South to San Miguel de have cellphone access, compared to 44.3 and 44.9 Tucuman, and eventually Cordoba, Rosario, and percent in the cities of Salta and San Salvador de Buenos Aires). The Corridor Project’s projected Jujuy. 67 percent of households nationally have area of influence is concentrated in Jujuy prov- computer access, compared to just 44 percent ince, specifically the city of San Salvador de Jujuy in the cities of Salta and Jujuy. Lack of such soft (and the smaller San Pedro de Jujuy), and south to infrastructure may impede residents’ and firms’ the border with Salta Province. However, the city abilities to integrate with and benefit from more of Salta is within the extended area of influence, successful regions of the national economy, such which will ensure the two cities become better as to find jobs or inputs, buy or sell products, or linked (Figure 3.15 and 3.16). share and gain knowledge and skills from others. Figure 3.15. Jujuy and Salta cities are located in the direct However, Salta’s far greater export perfor- area of influence of the NW corridor mance relative to Jujuy also points to the potential gains of better integrating the provinces with each other. As discussed above, Jujuy has higher concentration in certain important tradeables such as manufacturing. Connecting Jujuy to nearby Salta (discussed below under the Northwestern Road Development Corridor Project) may help its firms enjoy similar connectivity to regional, national, and international markets.

Current efforts to connect Jujuy to other economic hubs in the NOA, the Bolivian border, and the national economy are underway and include investment under the Northwestern Road Development Corridor Project. The province of Salta and Jujuy are in the direct area of influence of the Nor thwestern Road Development (NW) Corridor Project. The 96.2 km NW Corridor Project with World Bank financing will make improvements to roads from Jujuy, specifically Source: authors with data from Global Human Settlement Layer, GHSL, 2014 TERRITORIAL DEVELOPMENT IN ARGENTINA: DIAGNOSING KEY BOTTLENECKS | 75

Figure 3.16: The area of influence of the NW corridor project Chapter 2 highlighted the potential for secon- covers 27 percent of the built-up area in Salta and Jujuy dary cities to serve as anchors for regional Provinces development, and how connective investments (like a local development corridor) can become an important channel for growing the local economy and reducing spatial disparities. The upgraded Corridor has the potential to facilitate trade between Jujuy’s main cities, and with the city of Salta. This may facilitate development of regional value chains in key sectors such as mining, agriculture, and tourism. For residents, investments in connectivity can also contribute to improve access to goods, services, and jobs in the neighboring locations.

Note: the shaded oval is the area of influence. Source: Global Human Settlement Layer, GHSL, 2014.

Box 3.2. Modelling the impacts of the NW Corridor

Fajgelbaum and Van Patten used a spatial equilibrium model to simulate the potential aggregate and regional effects of investment in transportation infrastructure and public goods in different cities of Argentina. The model combined administrative data of formal employment from OEDE and GIS data on the shape of the road network. The study considers multiple scenarios in terms of the cities being affected by the investment and the type of the policies (increase of public goods other than roads, decrease of trading cost, and both) to assess the impact of such improvements on national welfare, and on employment and GDP in the directly affected cities as well as other specific cities across Argentina.

Returns to Investment across scenarios

First, the study finds that the Northwestern Road Development Corridor Project has the highest returns per 100 miles of road improved (road length being a proxy for cost). For each 100 miles of road over which transport costs fall by 10 percent, it predicts a 0.16 percent national welfare gain when the investment takes place in the Jujuy-Salta region alone, compared to gains of 0.07 percent or less for all larger project scenarios involving Buenos Aires. 76 | CHAPTER 3 SCALE, SPECIALIZATION, AND CONVERGENCE: A CLOSER LOOK AT THE PROVINCES OF SALTA AND JUJUY

The above gain is of course small, as it spreads the local benefit across the whole country. Zooming into the major cities touched by the improved road (San Salvador de Jujuy and San Pedro de Jujuy), however, a 2 percent employment increase is projected, while economically linked small towns like General San Martin also benefit. The study also anticipates economic development to be spread more evenly across the territory: the employment impacts of roads are projected to be greater for smaller towns, with San Pedro (a city of 6 thousand formal workers) enjoying a much higher percentage increase in employment than in San Salvador (40 thousand formal workers).

The study shows that overall effects are larger when multiple types of investment take place. The results suggest complementariness between investment in public services and transport infrastructure are larger in cases involving public services valued by workers only. Comparing Table 3.4a (only one type of government spending change) with Table 3.4b (multiple types of investment take place) shows there are larger benefits when there are complementary investments. Welfare gains in column (c) in Table 3.4b are larger than those in Table 3.4a. For example, when there is complementariness, there is a 0.25 percent increase in welfare (Table 3.4b), compared with 0.18 percent in Table 3.4a in Jujuy-Salta, Moreover, these complementarities are particularly strong in the face of investment linking Buenos Aires to the NOA region. Also, the study provides some evidence suggesting complementarities between transport network and public services for the cities of San Pedro, San Salvador, and San Martin. Similar patterns are also found when the city of Salta also receives public goods investments.

Table 3.4: Welfare changes are higher when multiple types of investment take place a. Welfare changes (only one type of government spending is changing)

c x x CASE (a) G n (b) τηι (c) G n , τηι (d) RATIO (a)+(b)

Jujuy 0.05 0.09 0.12 0.85

Jujuy-Salta 0.09 0.09 0.18 0.99

Buenos Aires – Northwest 0.38 0.79 1.19 1.02

Buenos Aires – Mesopotamia 0.43 0.49 0.93 1.00

Buenos Aires – Patagonia 0.36 0.05 0.40 0.98

Northern cities 0.19 0.14 0.31 0.9 TERRITORIAL DEVELOPMENT IN ARGENTINA: DIAGNOSING KEY BOTTLENECKS | 77

b. Welfare changes (multiple types of government spending are changing)

c u x u x CASE (a) G n,G n (b) τηι (c) G n G n , τηι (d) RATIO (a)+(b)

Jujuy 0.09 0.09 0.15 0.86

Jujuy-Salta 0.17 0.09 0.25 0.94

Buenos Aires – Northwest 0.93 0.79 1.72 1.00

Buenos Aires – Mesopotamia 1.09 0.49 1.53 0.97

Buenos Aires – Patagonia 0.74 0.05 0.84 1.06

Northern cities 0.32 0.14 0.43 0.93

Source: Fajgelbaum and Van Patten, Forthcoming

The impacts on Jujuy could be transforma- From 1975-2014, the share of built-up land in this tive if accompanied by efforts to comple- area of influence increased by 20.7 percentage ment them with local public goods. As in the points more than the increase in built-up land box above, Fajgelbaum and Van Patten predict a in Salta and Jujuy provinces as a whole. In addi- 2 percent employment boost for San Salvador de tion, the affected percentage of the built-up area Jujuy and San Pedro de Jujuy from the new road. increases if we include municipalities around the The road is well-located to affect large shares of city of Salta, which are indirect beneficiaries of the current and future population: 26.7 percent of the NW corridor project (Figure 3.17). In terms of the most rapidly expanding built-up areas in the population, the extended area of influence covers provinces of Salta and Jujuy lies mostly within 1,047,264 people (78.6 percent of the total popu- the area of influence of the NW corridor project. lation of both provinces) and 269,105 households. Figure 3.17: Expansion of the built-up area and population density are tightly concentrated around the capital cities of Salta and Jujuy b. Built-up area of San Salvador de Jujuy in 1975 (red) and in 2014 a. Built-up area of the city of Salta in 1975 (red) and in 2014 (mustard color) (mustard color)

d. Densidad poblacional en San Salvador de Jujuy 78 | CHAPTER 3 SCALE, SPECIALIZATION, AND CONVERGENCE: A CLOSER LOOK AT THE PROVINCES OF SALTA AND JUJUY

c. Population density in the city of Salta d. Population density in San Salvador de Jujuy

Source: GHSL 2014 and INDEC 2010. LITTLE CONVERGENCE IN SALTA AND JUJUY – low-quality and inaccessible basic basic needs, Salta has the highest proportion services, and low human capital (19.4) and Jujuy (15.5) the third highest in the NOA Important gaps in living standards persist within region (INDEC, 2010). these provinces There is considerable variability in welfare and Urban agglomerations in the provinces of Salta service access within each province. Figure 3.18 and Jujuy continue to struggle with poverty. The maps basics needs unmet at the district level in previous chapter highlighted that the North lags Salta and Jujuy. This shows Salta containing loca- behind on most measures of welfare and poverty. tions with high shares of households in distress However, even within the North, Salta and Jujuy (particularly in the East/North East), as well as are poor performers. Poverty rates in Salta and areas (particularly closer to the city of Salta and Jujuy rank highest among all urban agglomera- the West) with lower shares of unmet basic needs. tions nationally. Salta and Jujuy have a poverty In Jujuy, there are noticeably fewer districts where incidence of 24.8 and 24.2 percent respectively, more than 41 percent of households have unmet 67 topped only by Santiago del Estero (38.3 percent). basic needs. Around 20 percent of households in 67 Poverty level here refers to the one calculated by INDEC (National each province lack access to clean drinking water. Institute of Statistics and Censuses) through the regular household survey “Encuesta de Hogares permante (EPH). Here we use data Sewage coverage is below the national average from the 1st semester 2017. The caveat with this dataset is that it only in the provinces of Salta and Jujuy (76.4 and 82.7 measures poverty across urban areas. The index of unmet basic needs percent), but both provinces have higher-than- presented later is the only nationally representative measure, calculated during the Census 2010. average coverage compared to northern Argentina. Increasing access to water to 100 percent in the TERRITORIAL DEVELOPMENT IN ARGENTINA: DIAGNOSING KEY BOTTLENECKS | 79

area of influence of the NW corridor project will However, quality of the water service is a challenge, bring water to 6,086 households, but more signifi- at least in the city of Salta; the water service is not cantly, achieving 100 percent sewage coverage will continuous (19 hours a day) and around 36 percent bring sewage access to 72,363 households. of households report having suffered problems in the provision of running water (UCA, 2014). Figure 3.17: Expansion of the built-up area and population density are tightly concentrated around the capital cities of Salta and Jujuy

Source: Census 2010 INDEC

As expected, the cities of Salta and San Salvador Additionally, substandard housing and public de Jujuy have better coverage of basic services than infrastructure such as lighting and roads remain the rest of the cities in their provinces but quality major challenges for these cities issues remain Over 16 percent of houses in the city of Salta and The municipalities of Salta and San Salvador de San Salvador Jujuy are below habitability stan- Jujuy have a high access to basic services relative dards (INDEC, 2010). The percentage of houses to the rest of the cities in their provinces. This whose conditions are below the habitability stan- is expected, due to the lower costs of provision dards is 16.8 percent for the city of Salta and in more densely populated locations, and higher 16.6 percent for San Salvador Jujuy. This qualita- incomes in the cities, reflecting the national- tive housing deficit refers to houses that require level trends highlighted in the previous chapter. repairs, that are located in risky areas, or do The areas with the best sewage coverage in each not have access to basic services such as water province are within their respective capital cities and sewage. However, once again these cities (Figure 3.19). Specifically, 86.1 and 73.5 percent have a lower percentage of houses below habit- of the households in Salta city and San Salvador ability standards than do the provinces at large; de Jujuy have access to sewage – higher than the provinces of Salta and Jujuy have qualitative cities such as Cordoba, Rosario, and Santiago del housing deficits of 25.7 and 24.0 percent respec- Estero. Access to water is also higher in each city tively (INDEC, 2010). Moreover, while there is not than their provinces at large, though levels are much data on informal settlements, a recent study similar to those in other major cities in Argentina. on the city of Salta reported that 13.7 percent of houses are on informal settlements (IADB, 2015). 80 | CHAPTER 3 SCALE, SPECIALIZATION, AND CONVERGENCE: A CLOSER LOOK AT THE PROVINCES OF SALTA AND JUJUY

Figure 3.19: Even within city boundaries, variations in living standards persist

a. City of Salta sewage access b. San Salvador de Jujuy sewage access

c. City of Salta unmet basic needs d. San Salvador de Jujuy unmet basic needs

Source: INDEC, 2010 TERRITORIAL DEVELOPMENT IN ARGENTINA: DIAGNOSING KEY BOTTLENECKS | 81

There is also a quantitative housing deficit68 for Almost all countries have to confront the problem the provinces of Salta and Jujuy, estimated at 16.9 of unequal or ‘lumpy’ economic activity across percent and 15.3 percent, respectively (INDEC, their territory; their challenge is to work to make 2010). Regionally, both cities have a quantita- people’s living standards equal regardless of tive housing deficit lower than the NOA region where they live. These statistics demonstrate (17.1 percent), but relative to the main cities in that Argentina’s government has struggled to Argentina, housing deficits are high. For example, ensure that its people enjoy the same basic decent the city of Córdoba and Rosario have a qualitative quality of life everywhere. Education in the prov- deficit of 8.3 and 7.8 percent substandard housing, inces of Salta and Jujuy also has not provided citi- while even cities in the NOA region such as San zens an ‘escape route’ from poor living standards.

Miguel de Tucuman and Santiago del Estero have Low human capital and schooling quality are better housing indicators (INDEC, 2010). persistent challenges in the two provinces

Cities also lack infrastructure such as pave- Education coverage and quality are core concerns ment and public streetlights around houses. The for both provinces. Illiteracy rates in the provinces percentage of houses with at least a paved block of Salta and Jujuy are low (3 percent), but much ranges between 4 and 42 percent in the largest higher than the national average (1.9 percent). municipalities in Argentina. Specifically, San Chapter 2 showed a near-complete convergence Salvador de Jujuy has the largest percentage (41.3 across provinces in access to primary education, percent) among the cities in the North. Only 3.6 but wide disparity remaining for secondary educa- and 4.8 percent of the houses in the cities of Salta tion completion. Salta and Jujuy are among the and San Salvador de Jujuy have streetlights outside provinces falling behind in the latter. Secondary their houses, yet that is not far behind munici- education coverage69 is less than half (47.9 and 45.1 palities with large populations, where streetlight percent in Salta and Jujuy respectively), compared coverage is less than 7 percent (Table 3.5).

Table 3.5: Public lighting and paved roads are a challenge for Salta and San Salvador de Jujuy

PERCENTAGE OF HOUSES WITH STREETLIGHTS PERCENTAGE OF HOUSES WITH AT LEAST CITY OUTSIDE THEIR HOMES A PAVED BLOCK

Córdoba 6.3 9.9

Rosario 6.6 4.4

San Miguel de Tucumán 2.9 17.3

Salta 3.6 21.1

San Salvador de Jujuy 4.8 41.3

Santiago del Estero 3.8 21.7

Source: authors, using data from INDEC (2010)

68 Quantitative housing deficit refers to an insufficiency of the existing 69 This refers to the population 13 to 17 year-olds who attend an educa- housing stock, either due to a shortage of supply, or due to the material tional establishment conditions of the occupied dwellings. 82 | CHAPTER 3 SCALE, SPECIALIZATION, AND CONVERGENCE: A CLOSER LOOK AT THE PROVINCES OF SALTA AND JUJUY

to AMBA’s coverage of 71.2 percent; this signif- With 90 percent of Salta and Jujuy’s populations icantly limits the forms of employment open to born in the province, these schooling dispari- Salta and Jujuy’s workers (Caruso and Lugo, 2017). ties already trickle into the labor force. Very low Within schools, the quality of education is also shares of the labor force have higher education in lacking. Around three quarters of high schoolers San Salvador de Jujuy and the city of Salta – only in Salta and Jujuy have below-satisfactory knowl- 5.3 and 7.1 percent respectively. Additionally, one edge in math (Figure 3.20). Primary school perfor- in six young people in the city of Salta (between 18 mance is better, but 44.5 and 47.2 percent of the and 24 years – the typical age for tertiary educa- students in 6th grade of primary school have math tion) are disengaged from productive activities, skills considered below satisfactory. Thus, the neither studying, working, or seeking work; this provinces are struggling to equip the next genera- share is also high in Jujuy, and is concentrated in tion of workers with the skills needed to catch up the lowest socioeconomic strata in both provinces with national incomes and productivity. (Caruso and Lugo, 2017).

Figure 3.20. Provinces of Salta and Jujuy have a high proportion of students in high school with basic and below-basic knowledge in math

Percentage of students with below-basic or basic performance

Percentage of students with satisfactory or advanced performance Percentage

Jujuy Chaco Salta AMBA La Rioja San Juan Chubut Misiones Tucumán Entre Ríos NeuquénFormosaSan Luis MendozaSanta FeCórdoba Catamarc a Santa Cruz Corrientes Río Negro La Pampa Buenos Aires Tierra del Fuego Santiago del Estero

Source: Aprender 2016, Argentine Ministry of Education and Sports

Evidence highlighted in the previous chapter cities like Buenos Aires points not only at the indicates that poor human capital outcomes importance of improving local workers’ skillsets and local amenities may be responsible for the but also improving the provision of local ameni- economy’s underperformance in Salta and Jujuy. ties as a way to attract high-skill workers into the The analysis of Quintero and Roberts (2017) found region. that productivity in Northern provinces was 3 Table 3.6 provides a summary of the diffe- to 4 times lower than in Buenos Aires; however, rent challenges that the provinces of Salta and once differences in worker characteristics in Jujuy face, identified along the three dimensions each location were accounted for, the location of territorial development highlighted in this premium became smaller but persisted. The fact report: scale, specialization, and convergence. that workers sort into larger and most successful TERRITORIAL DEVELOPMENT IN ARGENTINA: DIAGNOSING KEY BOTTLENECKS | 83

Table 3.6: Summary of the Challenges in the provinces of Salta and Jujuy

SCALE SPECIALIZATION CONVERGENCE

Low contribution to Public sector services High poverty rates and the national GDP. dominate employment concentration of poor Low firm growth and in the services sector Lack of access to productivity. Weak connectivity to sewage and other basic Province High labor informality global, national, and services, including secon- & unemployment regional markets dary education Limited skills Weak hard and soft Low human capital and High proportion of connective infrastructure schooling quality small and local firms

Low GDP per capita Large public sector in Qualitative and quantita- and economic terms of employment tive housing deficit Local growth and GVA Lack of urban infrastruc- High labor Low skill levels in the ture (i.e pavement and informality labor force streetlights)

REFERENCES

Caruso, G. & Lugo, M. A. (2017). Convergence of Human INDEC (Instituto Nacional de Estadística y Censos) (2010). Censo Opportunities: Lessons from Argentina. World Bank, Nacional de Población, Hogares y Viviendas. Buenos Aires, Background document. Argentina.

Castillo, V. et al. (2015). Tourism Policy, a Big Push to Employment: (2016). Encuesta Nacional sobre Acceso y Uso de Tecnologías Evidence from a Multiple Synthetic Control Approach. Inter- de la Información y la Comunicación 2016. American Development Bank. MECON (2017). Sector público provincial: ocupación y gastos sala- CEPAL (Comisión Económica para América Latina y el Caribe) riales provinciales 2015. Información fiscal: provincial y muni- (2017). Territorio, infraestructura y economía en la Argentina: cipal. Ministerio de Hacienda y Finanzas Públicas. http://www2. restricciones al crecimiento de distintos complejos productivos. mecon.gov.ar/hacienda/dncfp/provincial/gasto_medio.php https://repositorio.cepal.org/handle/11362/42060 Ministry of Education and Sports of Argentina (2016). Sistema Consejo Vial Federal (2014). Información técnica. Buenos Aires, Abierto de Consulta - Aprender 2016. Buenos Aires, Argentina. Argentina. http://www.cvf.gov.ar/red_vial_provincial.php http://aprenderdatos.educacion.gob.ar/binarg/RpWebEngine. exe/Portal?lang=esp Fajgelbaum, P. & Van Patten, D. (de pronta publicación). Spatial and Aggregate Effects of Transport Network and Public Service Ministry of Finance of Argentina (2016). Informe productivo de la Investments in Argentina. provincia de Jujuy. Buenos Aires, Argentina.

Gaubert, C. (2018). Firm Sorting and Agglomeration (n.º w24478). (2017). Informes productivos de la provincia de Salta. National Bureau of Economic Research. Subsecretaría de Planificación Económica.

Global Human Settlement Layer (GHSL). (2014). European (2018). Sistema de Indicadores de Desarrollo Provincial Commission. (SIDEP). Dirección Nacional de Asuntos Provinciales. Buenos Aires, Argentina. http://www2.mecon.gov.ar/hacienda/dinrep/ IADB (InterAmerican Development Bank) (2015). Salta Sostenible. sidep/index.phpç Washington, D.C.: Inter-American Development Bank. https://issuu.com/ciudadesemergentesysostenibles/docs/ Ministry of Production of Argentina (2017). Informe Productivo de salta_plan_de_acci__n la provincia de Salta. Buenos Aires, Argentina. 84 | CHAPTER 3 SCALE, SPECIALIZATION, AND CONVERGENCE: A CLOSER LOOK AT THE PROVINCES OF SALTA AND JUJUY

Muñoz, F. & Trombetta, M. (2015). Indicador Sintético de Actividad Quintero, L. & Roberts, M. (2017). Explaining Spatial Variations in Provincial (ISAP): un aporte al análisis de las economías regio- Productivity: Evidence from 16 LAC Countries. Documento de nales argentinas. Investigaciones Regionales, (33), 71. http:// antecedentes. Washington D.C.: World Bank. www.aecr.org/images/ImatgesArticles/2016/1/4_munoz_trom- Rodríguez-Pose, A. (2018). The Revenge of the Places that Don’t betta.pdf Matter (and What to Do About It). Cambridge Journal of Regions, OEDE (Observatorio de Empleo y Dinámica Empresarial). Conjunto Economy and Society, 11(1), 189-209. de datos de Dinámica del empleo y rotación de empresas. UCA (Universidad Católica Argentina) (2014). Estado del Desarrollo Ministerio de Trabajo de Argentina. Buenos Aires, Argentina. Humano y Social en el Gran Salta. Región del Noroeste Argentino. Oxford Economics (2017). Oxford Economics Dataset for Cities Informe n.º 1. Observatorio de la Deuda Social de la Infancia. 2017. Oxford, England. Centro de Estudios para el Desarrollo Humano y la Prosperidad. TERRITORIAL DEVELOPMENT IN ARGENTINA: DIAGNOSING KEY BOTTLENECKS | 85 CHAPTER 4

Summary: scale, specialization, and convergence in a nutshell

CERRO OTTO IN BARILOCHE, ARGENTINA Geronimo Giqueaux © 86 | CHAPTER 4 SUMMARY: SCALE, SPECIALIZATION, AND CONVERGENCE IN A NUTSHELL

CHAPTER 4 Summary: scale, specialization, and convergence in a nutshell

This report provided a methodological framework concentration and specialization do not necessa- and diagnostics along three dimensions of rily lead to disparities in living standards across territorial development in Argentina: Scale, regions. Instead, they can be used to achieve Specialization, and Convergence. Scale refers to the convergence in a territory. By looking at the extent of concentration of people, economic acti- economic geography of a country or region along vity, and firms in space. People and firms often these three dimensions, you can inform the iden- concentrate in a few places, because proximity tification of challenges and help define priorities is valuable. However, economic concentration in a territorial development strategy. The tables might lead to the appearance of lagging areas, below summarize in a nutshell the key messages when the benefits of such concentration are of the report, pointing at the key findings and not spread equally across all regions. Moreover, providing some benchmarking to assess how places that are well connected and take advan- Argentina compares to other countries around tage of the agglomeration benefits that arise the world in these dimensions. Key messages for from concentration might be able to specialize policy makers on where to go from here are also in sectors where they are most productive. But included for each dimension. TERRITORIAL DEVELOPMENT IN ARGENTINA: DIAGNOSING KEY BOTTLENECKS | 87

TERRITORIAL DEVELOPMENT IN ARGENTINA

Scale can mean concentration of people Argentina vs the World

INFORMATION FOR POLICYMAKERS

TERRITORIAL DEVELOPMENT IN ARGENTINA

INFORMATION FOR POLICYMAKERS

To achieve specialization, it will be necessary to think about solutions that allow firms to reach larger markets for their inputs and outputs.

TERRITORIAL DEVELOPMENT IN ARGENTINA

Convergence means convergence in living standards Argentina vs the World

INFORMATION FOR POLICYMAKERS

For convergence, the right policy mix depends on how the population is distributed across space.

Convergence All localities need access to decent public services and human capital development.

In large cities, where divisions of formality vs. informality can generate deep gaps in the quality of life, spatially targeted policies may be necessary to reduce these divisions.

ANNEXES

ANNEX A. PROXY INDICATORS Tables below show details on the proxy indicators used to measure Scale, Specialization, and Convergence.

Table A.1: Proxy indicators used for Scale POPULATION – SCALE

GUIIDING INDICATORS DEFINITIONS QUESTIONS

Scale can mean concentration of Population and economic activity Maps display the concentration people – large cities have more concentration of people and economic activity scale. The following guiding ques- for different areas of the country. tions should help identify population At the regional level, indicators scale: should show population concen- 1. Does the country display tration in the largest cities in the a distinct concentration of provinces. population? 2. Where is the geographic concen- Urban primacy tration of people in the country? Percentage of a country's urban 3. Does the country experience an population living in that country's urban hierarchy dominated by a largest metropolitan area. particular city?

ECONOMIC ACTIVITY – SCALE

GUIDING INDICATORS DEFINITIONS QUESTIONS

Scale can also mean concentration GDP per capita and regional GDP Per capita GDP is a measure of of economic activity. Concentration the total output of a country that of economic mass is a sign of takes the gross domestic product good economic performance, but it (GDP) and divides it by the can sometimes leave some areas number of people in that country. behind. The following guiding ques- The per capita GDP is espe- tions should help identify economic cially useful when comparing one activity scale: country to another and between provinces, because it shows 1.Does the country display large the relative performance of the concentration of economic activity country or province. For comple- in particular regions? mentary indicators, if data is avail- 2. Are there significant variations able, the analysis could include of economic activity across the indicators such as disposable country? income per capita and productivity per worker. 92 | ANNEXES

ECONOMIC ACTIVITY – SCALE

GUIDING QUESTIONS INDICATORS DEFINITIONS

3. Are there distinct patterns of Labor participation rates at the Percentage of people who are transportation/commerce flows? national, provincial, and municipal either employed or are actively level. looking for work.

Flow of flights Number of flights between origin/ destination

Flow of passengers Number of passengers that travel between two points. This indicator is displayed as a map.

Road Freight Flow Number of passengers that travel between two points. This indicator is displayed as a map.

FIRM ACTIVITY– SCALE

GUIDING INDICATORS DEFINITIONS QUESTIONS

Finally, scale can mean firm concen- Firms by size Number of firms by size in tration. Firms often concentrate in a each region, province, or Local few places and benefit from agglo- Economic Area. meration economies. The following guiding questions should help iden- Firm density Number of firms per 1,000 inhabit- tify firm activity scale: ants. If data is available, bench- mark with comparators. This 1. How are firms distributed across indicator can also be calculated the country and by sector? by the number of firms per 1,000 2. Does the distribution of firms workers in the labor force. vary across regions in the country? 3. Has this distribution changed Proportion of local firms out of over time? Firms by origin the total firms. If data is avail- 4. Does the distribution of human able, provide share of firms in the capital reflect different geogra- informal and formal sector. phical/sectorial patterns? TERRITORIAL DEVELOPMENT IN ARGENTINA: DIAGNOSING KEY BOTTLENECKS | 93

FIRM ACTIVITY – SCALE

GUIDING INDICATORS DEFINITIONS QUESTIONS

Firms by sector Distribution of the firms by sector. Percentage of firms in services, industry, construction, agriculture and other sectors, if available.

Fasr-growing firms and firm Numbers of firms that manage to graduation increase employment and move to an upper category in each period of time.

Labor force skill level Share of the labor force with incomplete secondary education and share of the labor force with a university degree. These indicators will be helpful in assessing human capital skill levels.

Table A.2: Proxy indicators used for Specialization ECONOMIC ACTIVITY - SPECIALIZATION

GUIDING INDICATORS DEFINITIONS QUESTIONS

Specialization can mean when Distribution of the GDP by sector Classification of the national or places can specialize in what provincial GDP by sector. they are best at producing. The following guiding questions should Distribution of firms by sector Share of firms by sector, at the help identify economic activity national and provincial level. specialization:

1.How is the economic activity Share of employment by sector at Distribution of employment by distributed across the country? the national and provincial level. If sector 2. How is employment distributed data is available classify GDP by by sector and by provinces? tradeable and non-tradeable. 3. What are the employment drivers regionally? 94 | ANNEXES

ECONOMIC ACTIVITY - SPECIALIZATION

GUIDING INDICATORS DEFINITIONS QUESTIONS

4. How are the exports distributed Share of local employment Location Quotient (LQ) at the across the country and regionally? (province-level) over the share national and provincial level. of national employment in each major sector (agriculture, services, construction, and so on). A value below one indicates that the share of employment in that sector is lower in the province relative to the entire country, and the contrary is true for a value above one.

Location Quotients vs GDP growth Employment drivers at the national in the same sector and location. and provincial level. This indicator allows the identification of sectors that have both a higher employment concentration and are growing faster in the provinces compared to the national average over a period of time.

KSI =Σ|S –S | ; where S is Krugman’s Specialization Index p p,i p,i the employment share of industry (KSI) i=1 i in province p, and S is the average share of industry i in the employment of all provinces in Argentina. The KSI can take values between 0 and 2, with 0 indicating that the employment structure of a province is identical to the national sectoral composition and 2 indicating the largest possible difference between the two employment distributions being compared.

International production openness Measures the trade insertion of index the province and is computed as the share of provincial value- added aimed at exports over the same ratio for the country level. TERRITORIAL DEVELOPMENT IN ARGENTINA: DIAGNOSING KEY BOTTLENECKS | 95

CONNECTIVITY - SPECIALIZATION

GUIDING INDICATORS DEFINITIONS QUESTIONS

Connectivity enables specializa- Road density at the national, Ratio of the length of total roads tion. Well-connected places foster regional, and provincial level (meters) over the province total trade, specialization of cities, and area (square meters). If data is regions. The following guiding ques- available, the analysis can extend tions should help identify connectivity to quality of roads: Percentage of topics key for specialization: national routes paved out of the total number of routes. This basic 1. How is the quantity and quality indicator will allow the assessment of infrastructure in the country? of connectivity differences across 2. How is the road accessibility the country. across the country and within specific regions? 3. What are the logistics challenges for international trade? Road accessibility Measures the travel time, using land or water-based travel, to the nearest city of more than 50,000 inhabitants. This indicator reveals which regions are less connected to markets and how dense the transportation network is within the country.

Logistics Performance Index Benchmarking tool to help countries identify the challenges and opportunities they face in their performance on trade logistics and what they can do to improve their performance.

Time-distance measured travelling Market Access index by car.

Soft connective infrastructure Percentage of people or families with residential broadband connections and mobile line. 96 | ANNEXES

COMPETITION - SPECIALIZATION

GUIDING INDICATORS DEFINITIONS QUESTIONS

A good business environment can Private sector limitations Stoplight Qualitative summary of the main spark investment, fostering inte- obstacles to business reported gration and specialization. Barriers by formal and informal firms in to competition and obstacles to the a country. This analysis will help private sector can dampen specia- identify key constraint for further lization. The following guiding ques- investments and private sector tions should help identify competition expansion. barriers for specialization:

1. What are the main challenges Product market regulations The PMR indicators database is for private investments and the a comprehensive and internation- business sector? ally-comparable set of indica- 2. What are the key competitive- tors that measure the degree ness challenges in the country, if to which policies promote or possible by sector? inhibit competition in areas of the product market where competition is viable.

Indices are useful tools to identify World Economic Forum’s Global Competitiveness Index (GCI) and and monitor the challenges in World Bank’s Doing Business (DB). terms of competitiveness and iden-

tify key bottlenecks to increasing

competitiveness.

Table A.3 Proxy indicators used for Convergence

CONVERGENCE INDICATORS

GUIDING INDICATORS DEFINITIONS QUESTIONS

Convergence means convergence Poverty incidence at the province Percentage of households with in living standards. Concentration of and district level unmet basic needs. This indicator economic activity and incomes often will reveal pockets of poverty leads to gaps in living standards. across the territory. This indicator These spatial disparities can affect can be used for rural areas as the quality of life of citizens. The well, if data is available. following guiding questions should help identify convergence: TERRITORIAL DEVELOPMENT IN ARGENTINA: DIAGNOSING KEY BOTTLENECKS | 97

CONVERGENCE INDICATORS

GUIDING INDICATORS DEFINITIONS QUESTIONS

1. What are the levels of poverty Index of access to services for water The index is computed as the across the country and regionally? and sanitation mean of z-scores of three vari- 2. How is access to basic services, ables: (i) the percentage of house- nationally and locally? holds with access to drinking 3. How is access to education and water; (ii) the percentage of health services, nationally and households living in a dwelling locally? connected to sewage; and (iii) the percentage of households with a toilet inside their dwelling. This indicator can show the disparities in access to basic services. If the data is available for a period, the indicator will show whether or not there is convergence over time.

Education coverage at the national, Primary and secondary comple- provincial, and local level tion rates. Also, if data is available for a period, the indicator will show whether there is convergence over time. Complementary indicators can be indicators related to the quality of education and schools per 1,000 of inhabitants.

Maternal mortality rate (MMR) and MMR is the annual number of infant mortality female deaths per 100,000 live births from any cause related to or aggravated by pregnancy, and Infant mortality is the number of infant deaths for every 1,000 live births. Complementary indicators can be indicators related to the number of health centers per 1,000 inhabitants and medical insurance coverage 98 | ANNEXES

ANNEX B. IMPORTANT DRIVERS OF RECENT AND FUTURE EMPLOYMENT (BY ECONOMIC SECTOR) Tables below show location quotients of employment with GDP growth in the same sector and location Figure B.1. Important drivers of recent and future employment (by economic sector) Agriculture Commerce Location Quotient 2016 Location Quotient 2016

Difference between regional and national sectoral growth (%) (2006-2016) Difference between regional and national sectoral growth (%) (2006-2016)

Construction Manufacturing Location Quotient 2016 Location Quotient 2016

Difference between regional and national sectoral growth (%) (2006-2016) Difference between regional and national sectoral growth (%) (2006-2016)

Mining Services Location Quotient 2016 Location Quotient 2016

Difference between regional and national sectoral growth (%) (2006-2016) Difference between regional and national sectoral growth (%) (2006-2016) TERRITORIAL DEVELOPMENT IN ARGENTINA: DIAGNOSING KEY BOTTLENECKS | 99

Utilities Note: The analysis combined Location Quotients with provincial DP statistics to identify sectors that have both a higher employment concentration and are growing faster in the provinces compared to the national average (2006-2016). The size of the figures indicates the size of employment in the sector (total number of employees) in the province. The upper-right quadrant then shows those provinces for which the key employment growth driver is the sector analyzed – that is, the sector analyzed has a higher level of concentration of employment that is also growing faster than national average in a given province. The upper-left quadrant contains provinces where the sector analyzed has low growth but a high concentration of employment, i.e. declining sectors. Provinces where the analyzed sector is an Location Quotient 2016 emerging sector are in the lower-right quadrant.

Source: Data from OEDE’s database, authors’ computation.

Difference between regional and national sectoral growth (%) (2006-2016)

TERRITORIAL DEVELOPMENT IN ARGENTINA: DIAGNOSING KEY BOTTLENECKS | 101 102 | EXECUTIVE SUMMARY