What is Wrong With U.S. Manufacturing?
Gregory Tassey Economic Analysis Office National Institute of Standards and Technology
[email protected] http://www.nist.gov/director/planning/
October 2012
Need Institutionalized Process for Managing Science, Technology, Innovation, and Diffusion (STID) Policy:
(1) Demonstrate importance of the policy issue for economic growth
(2) Identify indicators of underperformance at the macroeconomic level Productivity growth Trade balances Corporate profits Employment and earnings
(3) Estimate underinvestment in technology as determinant of growth Specific R&D investment trends Investment by phase of the R&D cycle Technology diffusion rates Scale-up
(4) Identify causes of underinvestment using technology-based models Excessive technical and/or market risk Appropriability, information, expertise, capital market problems
(5) Develop/select policy responses and management mechanisms Policy instruments matched with underinvestment phenomena Metrics for policy management and impact assessment 2 (1) Importance of the Policy Problem – Inadequate Innovative Capacity
Projected slow rate of economic growth is manifestation of inadequate long-term investments in productivity growth
. For the last decade (2000-2010)
Average annual real GDP growth was 1.7 percent
U.S. private nonfarm employment declined 3.3 percent
Real disposable personal income grew 11.9 percent
Median household real income declined 7.0 percent
. 23.5 million people were unable to find full-time jobs, double the 12 million working in U.S. manufacturing (July 2012)
. 47 million Americans on food stamps (June 2012)
. However, the current economic growth policy debate is focused on macrostabilization issues:
Government spending vs. deficit reduction
Monetary base expansion vs. potential inflation effects 3 (1) Importance of the Policy Problem – Wrong Growth Policy Model
Macrostabilization vs. Long-Term Growth Strategies
GDP Asian Long-Term Growth Economies (smoothed pattern)
Business Cycle (actual growth pattern) Western Economies
The Great Recession
Time
Gregory Tassey, “Beyond the Business Cycle: The Need for a Technology-Based Growth Strategy,” Science and Public Policy (forthcoming) 4 (1) Importance of the Policy Problem – Role of Domestic Manufacturing
Economic importance of A National Strategy for Advanced Manufacturing:
. As an independent economy, the value added produced by the U.S. manufacturing sector would rank 9th in the world . Important for diversification in a highly competitive global economy Contributes $1.7 trillion to GDP and employs 11 million workers High-tech service jobs are increasingly ‘‘tradeable’’ and 30 economies have policies in place to promote service exports
. Manufacturing accounts for 70% of US domestic industry-performed R&D 60% of U.S. domestic industry’s scientists/engineers 70% of patents issued to U.S. companies/residents
. Loss of domestic manufacturing would erode U.S. R&D infrastructure . High-tech services must have close ties to its manufacturing base Large high-tech manufacturing firms are integrating forward into services . The majority of trade is in manufactured products—deficits for last 37 years . Yet, neoclassical economists are okay with offshoring all manufacturing 5 (2) Underperformance – The High-Tech Economy
The Bottom Line: The high-income economy must be the high-tech economy 1) Technology is the long-term driver of productivity growth and hence growth in real wages 2) Median wages in BLS “high-tech” occupations exceed median for all industries by 50-100 percent 3) But, the U.S. economy is in long-term decline relative to the global economy due to underinvestment in R&D, capital formation, and workforce training 4) This underinvestment is now being manifested in a range negative economic growth indicators (productivity, real income, trade, GDP) 5) High-tech sector is small part of the U.S. economy, giving it low political influence 6) Existing underinvestment phenomena must be addressed by appropriate policy instruments—requires a STID economic growth model 7) Applying such a framework demands a STID policy analysis infrastructure— hardly exists in the United States 6 (2) Underperformance – Manufacturing
US Manufacturing Employment: 1960–2011
No. of Workers (thousands) 20,000
19,000
18,000
17,000
16,000
15,000
14,000
13,000
12,000
11,000
10,000 1960 1964 1968 1972 1976 1980 1984 1988 1992 1996 2000 2004 2008
Source: Bureau of Labor Statistics 7 (2) Underperformance – U. S. Economy
Policy Focus: Multifactor Productivity Trends in Productivity and Income: U.S. Manufacturing Sector, 1987-2010
Index 250
230
210
190
170
150 Multifactor Productivity 130 Labor Productivity
110
90
70 Real Hourly Compensation 50 1987 1990 1993 1996 1999 2002 2005 2008
Source: Bureau of Labor Statistics 8 (2) Underperformance – U.S. Economy
Non-Farm Employment Growth in Post World-War-II Business Recoveries: Percent Change from Recession Trough
12%
10%
8% Average of First Seven Post- World War II Recoveries
6%
4%
1990-91 Recovery 2%
Months 0% from 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 Trough 2009 Recovery 2001 Recovery -2% Source: G. Tassey, The Technology Imperative (updated); BLS for employment data; NBER for recession trough dates 9 (2) Underperformance – Manufacturing
U.S. Trade Balances for High-Tech vs. All Manufactured Products, 1988-2011
$ billions 100 Advanced Technology Products
0 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010
-100 -$99B
-200
-300 All Manufactured
-400 Products
-500
-600 Cumulative drain on GDP over this period is $8.25 -$637B trillion (2011 dollars) -700
Source: Census Bureau, Foreign Trade Division (“Trade in Goods with Advanced Technology Products” and “U.S. Trade in Goods and Services (FT900), Exhibit 15” for the manufacturing trade balance) 10 (2) Underperformance – R&D Intensity and Innovative Output
Rate of Innovation vs. R&D Intensity: Percent of Companies in an Industry Reporting Product and/or Process Innovations, 2003-2007
Index 140 Minimum for R&D- Intensive Industries
120
100
80
60
40
20
0 R&D 0 5 10 15 20 25 Intensity
Source: Gregory Tassey, “Beyond the Business Cycle: The Need for a Technology-Based Growth Strategy,” forthcoming. Index = sum of percent of companies in an industry reporting product innovations and percent reporting process innovations. R&D intensity data from Science and Engineering Indicators 2010 , Appendix Table 4-14 11 (industry and other non-federal funds for R&D); innovation data from Mark Boroush, “NSF Releases New Statistics on Business Innovation,” NSF InfoBrief, October 2010 (2) Underperformance – R&D Intensity and Economic Growth
Relationship Between R&D Intensity & Real Output Growth in Manufacturing
Ave. R&D Intensity, Percent Change in Percent Change in Industry (NAICS Code) 1999-2007 Real Output, 2000-07 Real Output, 2000-09
R&D Intensive: Pharmaceuticals (3254) 10.5 17.9 4.9 Semiconductors (3344) 10.1 17.0 1.1 Medical Equipment (3391) 7.5 34.6 39.5 Computers (3341) 6.1 109.9 147.0 Communications Equip (3342) 13.0 -40.0 -59.7 Group Ave: 9.5 Group Ave: 27.9 Group Ave: 26.6 Non-R&D Intensive: Basic Chemicals (3251) 2.2 25.6 -7.8 Machinery (333) 3.8 2.3 -22.4 Electrical Equipment (335) 2.5 -13.4 -33.4 Plastics & Rubber (326) 2.3 -5.2 -28.0 Fabricated Metals (332) 1.4 2.6 -23.6 Group Ave: 2.5 Group Ave: 2.4 Group Ave: -23.1
Sources: NSF for R&D intensity and BLS for real output. 12 (2) Underperformance – Maintaining Domestic Supply Chains
High-tech offshoring is a multi-step process, driven by (1) increasingly attractive skilled labor, (2) R&D and capital subsidies, and (3) supply-chain synergies
1) Originally, offshoring of manufacturing was for local-market opportunities
2) Increasingly, motivation is to access cheaper but also skilled labor
. Required small amount of supporting R&D . Host country frequently subsidizes plant and equipment
3) Host country gains R&D experience at “entry” tier in high-tech supply chain
4) Host country begins to integrate forward into design or backward into components
. China—backward to components (from assembly) . Taiwan—forward to electronic circuits (from components) . Korea—forward to electronic products (from components)
4) Economic policy point: Co-location synergies are being lost in U.S. and captured in Asia . Between 2011 and 2013, 81 new semiconductor fabs will begin
operations globally—six of them in the United States (SEMI) 13 (2) Underperformance – Manufacturing
Poor Technology Life-Cycle Management:
The United States has been the “first mover” and then lost virtually all market share in a wide range of materials and product technologies, including
• oxide ceramics
• semiconductor memory devices
• semiconductor production equipment such as steppers
• lithium-ion batteries
• flat panel displays
• robotics
• solar cells
• advanced lighting (e.g. OLED)
14 (2) Underperformance – Manufacturing
Trends in Manufacturing R&D Needing Policy Attention
. 37 consecutive years of trade deficits
. Manufacturing sector’s average R&D intensity (3.7 percent) has remained flat since the mid-1980s
has not been helped by offshoring of low R&D-intensive industries
pales compared to truly “R&D-intensive” industries, whose ratios range from 5 to 22 percent
. Need for effective policy response is great
most of the global economy’s $1.4 trillion annual R&D spending targets manufacturing technologies
Technology life cycles are being compressed
U.S. manufacturing firms are increasing offshore R&D at three times the rate of domestic R&D spending
. Government funding of manufacturing R&D increases the sector’s R&D performance intensity from 3.7 to 4.1 percent 15 (3) Underinvestment – Amount of R&D
U.S. R&D Intensity: Constant for 45 Years Funding as a Share of GDP, 1953-2008
3.5%
Total R&D/GDP 3.0%
2.5%
2.0% Federal R&D/GDP
1.5%
1.0% Industry R&D/GDP
0.5%
0.0% 1953 1957 1961 1965 1969 1973 1977 1981 1985 1989 1993 1997 2001 2005
Gregory Tassey, “Rationales and Mechanisms for Revitalizing U.S. Manufacturing R&D Strategies,” Journal of Technology Transfer 35 (2010): 283-333. Data from the National Science Foundation. 16 (3) Underinvestment – Amount of R&D
National R&D Intensities, 2009 Gross R&D Expenditures as a Percentage of GDP
5.0
4.46 4.5
4.0 3.92 3.61 3.56 3.5 3.36 3.10 3.00 2.94 2.90 3.0 2.82 2.72
2.5 2.40 2.27 2.26 1.99 2.0 1.92 1.85 1.70
1.5
1.0
0.5
0.0
Source: OECD, Main Science and Technology Indicators, 2010. 17 (3) Underinvestment – Amount of R&D
Changes in National R&D Intensity, 1995-2009
250%
198.2% 200%
150%
99.1% 100%
73.5% 70.9%
50.2% 50% 28.8% 24.0% 15.5%
0% China Singapore Finland Taiwan South Korea Germany Japan United States
Gregory Tassey, “Beyond the Business Cycle: The Need for a Technology-Based Growth Strategy,” Science and Public Policy (forthcoming). Data from OECD, Main Science and Technology Indicators, 2010/1. 18 (3) Underinvestment – Technology Intensity
Shares of Domestic Manufacturing Value Added from Moderate and High R&D-Intensive Industries 70%
60% 58%
52% 50% 49%
41% 42% 40%
32% 30% 26%
20%
10%
0% Australia Canada United Kingdom United States Japan Korea Germany
Source: OECD STAN Indicators 2009, "Value-added shares relative to manufacturing," Percent of manufacturing sector with 3% or greater R&D intensity, http://stats.oecd.org/index.aspx?r=228903 19 (3) Underinvestment – Composition of R&D
The “Valley of Death” (New Technology Platforms) is Getting Wider Trends in Short-Term vs. Long-Term US Industry R&D, 1993-2011 “Sea Change” Index 70
60
50
40
30
20
10
0 1993 1995 1997 1999 2001 2003 2005 2007 2009 2011 -10
Short-Term -20 (“Current Life-Cycle Applications”) -30 Long-Term (“Technology Platform Research”) -40 Gregory Tassey, “Beyond the Business Cycle: The Need for a Technology-Based Growth Strategy,” Science and Public Policy (forthcoming). Compiled from the Industrial Research Institute’s annual surveys of member companies. 20 (3) Underinvestment – Capital Formation Fixed Private Investment in Hardware & Software (growth by decade in 2005 dollars)
180% 167.7%
160%
140%
120% 108.8% 107.8%
100% 84.9% 80%
60%
40% 26.5%
20% 14.6%
0% 1960-70 1970-80 1980-90 1990-00 2000-10 2009-2011
Source: Gregory Tassey, “Beyond the Business Cycle: The Need for a Technology-Based Growth Strategy,” (Science and Public Policy, forthcoming). Data from Bureau of Economic Analysis, NIPA Table 5.3.5 (includes both equipment and software) and Table 5.3.4 (price indexes for fixed private investment) 21 (4) Causes of Underinvestment – Conventional Wisdom
“Black Box” Model of a Technology-Based Industry
Strategic Market Value Value Production Planning Development Added Added
Entrepreneurial Activity
Proprietary Applied Research Technology
Science Base Basic Research
Private Good Public Good
Gregory Tassey, The Technology Imperative, 2007; and, “The Disaggregated Technology Production Function: A New 22 Model of Corporate and University Research”, Research Policy, 2005. (4) Causes of Underinvestment – Market Failure
Neoclassical economics ignores market failure rationales for government support of STID policies:
. Excessive risk (increase costs)
. Knowledge spillovers (reduce benefits)
. Long development time (reduce benefits)
. Economies of scope (reduce benefits)
. Prices spillovers (reduce benefits)
. Information asymmetries (increase costs)
. Coordination inefficiencies (increase costs)
23
(4) Causes of Underinvestment – Composition of R&D
Technology Platforms: Overcoming the Early-Phase R&D Risk Spike
Risk
Risk Spike Science Base R0
“Valley of Death” Commercial Products Measurement & Standards Infrastructure
Basic Technology Platform Applied Research Research Research and Development R&D Cycle 10 years 6 years Commercialization
Source: Gregory Tassey, “Underinvestment in Public Good Technologies”, Journal of Technology Transfer 30: 1/2 (January, 2005); and, “Modeling and Measuring the Economic Roles of Technology Infrastructure,” Economics of Innovation and New Technology 17 (October, 2008) 24 (4) Causes of Underinvestment – Composition of R&D
Infratechnologies for Efficient Transactions in High-Tech Markets
Quality Reliability
Safety Privacy
Seller Market Interface Buyer
Interoperability Security
Performance Environment
25 G. Tassey, The Technology Imperative, Edward Elgar, 2007 (4) Causes of Underinvestment – Composition of R&D
Federal R&D Portfolio is not Optimized for Economic Growth
. Historical focus has and continues to be on “mission” R&D programs (national objectives such as defense, health, energy, space, environmental)—90 percent of federal R&D
National defense and health account for 81 percent of the federal R&D budget
Using NAICS codes to track federally funded R&D performed by industry,
75 percent of federal R&D allocated to the manufacturing sector goes to two NAICS 4-digit industries: aerospace and instruments
These two industries account for 15 percent of company-funded R&D and about 10 percent of high-tech value added
Policy Implication: While economic activity is stimulated by this skewed funding strategy, the federal portfolio is not close to being optimized for economic growth
Example: “technology platform” (proof-of-concept) research
Defense (DARPA): $3.1 billion
Energy (ARPA-e): $400 million
General economic growth (NIST’s ATP/TIP): $60 million
Sources: National Science Foundation: Federal R&D Funding by Budget Function, FY 2008-10, Table 2; Science and Engineering Indicators 2010, Appendix Table 4-13; Bureau of Economic Analysis R&D Satellite Account 26 (5) Policy Response – Technology-Element Growth Model
Technology-Element Model
Strategic System Market Value Production gPlanning Integration Development Added
Entrepreneurial Risk Reduction Activity
ProprietaryProprietary TechnologiesTechnologies
Technology Private Good Platforms
Public Good Science Base Quasi-Public Good
Gregory Tassey, The Technology Imperative, 2007; and, “The Disaggregated Technology Production Function: A New Model of Corporate and University Research,” Research Policy, 2005. 27 (5) Policy Response – Technology-Element Growth Model
Application of the Technology-Element Model: Nanotechnology Platforms
Technology Platforms Commercial Science Base Infratechnologies Product Process Products . carbon-based . biological detection . carbon nanotubes . epitaxy . hardened nanomaterials and analysis tools . dendrimers . nanoimprint nanomaterials for . cellulosic . in silico modeling & . hybrid nanoelectronic lithography machining/drilling nanomaterials simulation tools devices . nanoparticle . flame-retardant . magnetic . in-line measurement . ultra-low-power manufacture nanocoatings nanostructures techniques to enable devices . rapid curing . sporting goods . molecular closed-loop process . self-powered techniques . solar cells nanoelectronic control nanowire devices . self-assembling & . sunscreen/cosmetics materials . sub-nanometer . nanoparticle self-organizing . targeted delivery of . quantum dots microscopy fluorescent labels for processes anticancer therapies . optical . high-resolution cell cultures and . scalable deposition . biodegradable and metamaterials nanoparticle diagnostics method for polymer- lipid-based drug . solid-state detection . metal nanoparticles & fullerene photovoltaics delivery systems quantum-effect . thermally stable conductive polymers . inkjet processes for . self-repairing & long- nanostructures nanocatalysts for for soldering/bonding printable electronics life wood composites . functionalized high-temperature . nanoparticle sensors . purification of fluids . anti-microbial fluorescent reactions with nanomaterials coatings for medical nanocrystals . roll-to-roll processing devices . quantum-confined . nanoscale motion structures microscopes
Public Mixed Technology Goods Private Technology Goods Technology Goods
Gregory Tassey, The Technology Imperative, Edward Elgar, 2007 28 (5) Policy Response – R&D Intensity and Innovative Output
Policy Imperative: Increase Investment in Science, Technology, Innovation, and Diffusion (STID)
. New roles for governments
Compete as much as private companies
Future roles will be more complex rather than larger
Must manage the entire technology life cycle
. Major real asset categories determining competitiveness:
Technology (intellectual capital)
Skilled labor (human capital)
Hardware and software (non-human capital)
Technical infrastructure (public capital)
Industry structure & behavior (organizational capital)
29 (5) Policy Response – Technology-Element Growth Model
Needed: Total-Technology-Life-Cycle Growth Strategy
. Germany has a trade surplus in manufacturing, even though compared to the United States, it has a
Approximately same R&D intensity (2.82 percent vs. 2.90 percent for U.S.)
26 percent higher average hourly manufacturing labor compensation
. Reason: Germany has more comprehensive/intensively managed STID policy
Coordinated government R&D programs
Strongly integrated R&D and manufacturing
Highly skilled labor force across all technology occupations
Optimized industry structure (support for both large firms and SMEs)
Highest % of manufacturing value added from R&D-intensive industries
. Other nations using similar growth strategies
U.S. projected to fall behind EU27/AP regions in nanotech products (Lux Research)
30 (4) Causes of Underinvestment – Life-Cycle Mismanagement
Compression of Technology Life Cycles Raises Technical and Market Risk
Performance/ Price Ratio
New Global Life Cycles Old Domestic Life Cycles
Time A C B G. Tassey, The Technology Imperative, Edward Elgar, 2007 31 (5) Policy Response – Technology Platforms
Life-Cycle Acceleration: Technology Platform Development
Performance/Price Ratio D•
New Technology
A • C′ • • C
B Current • Technology
Time
Gregory Tassey, “Rationales and Mechanisms for Revitalizing U.S. Manufacturing R&D Strategies,” Journal of Technology Transfer 35 (2010): 283-333. 32 (5) Policy Response – Infratechnologies
Life Cycle Market Failure: Infratechnology
Performance/Price Ratio
C′′ New • Technology
A • C′ • • C
B• Current Technology e.g., semiconductor industry has ~1600 standards
Time
Source: Gregory Tassey, The Technology Imperative, Edward Elgar, 2007 33 (5) Policy Response – Technology-Element Growth Model
Managing the Entire Technology Life Cycle: Science, Technology, Innovation, Diffusion (STID) Policy Roles
Market Targeting Joint Industry- Assistance and Government Scale-Up Procurement Planning Incentives Incentives Strategic System Market Value Production Planning Integration Development Added Interface Standards (consortia, standards groups) Risk Acceptance Technology Entrepreneurial Reduction Transfer/Diffusion (MEP) Activity Test Standards Intellectual Property and National Rights (DoC) Test Facilities (NIST) Tax Incentives Proprietary Technologies Incubators (states)
National Labs National Labs Technology (NIST), Direct Funding of Firms Platforms Consortia & Universities (DARPA, ARPA-E, NRI, AMTech) Science Base 34 Gregory Tassey, “Rationales and Mechanisms for Revitalizing U.S. Manufacturing R&D Strategies,” Journal of Technology Transfer 35 (2010): 283-333. (5) Policy Response – Technology-Element Growth Model
Summary: Complexity of Modern Manufacturing Technologies Leads to Three Targets of R&D Policy
. Amount of R&D Create financial incentives for private companies to increase investments in R&D and increase the R&D intensity of the manufacturing sector Increase Federal investment in research aimed at broad sector growth objectives in addition to those related to agency missions
. Composition of R&D Create incentives for private-sector investment in early phases of R&D cycle Create public-private partnerships to meet industry’s long-term research needs and IP management requirements through support for technology clusters Fund research aimed at manufacturability to overcome scaling issues Target the “other” 90% of manufacturing value added (outside NAICS 3345 and 3364) Eliminate barriers to private investment in new firms (high technical risk, appropriability, skilled labor acquisition, and process-capability barriers)
. Efficiency of R&D Improve timing and content of R&D through road mapping and portfolio management techniques Increase rates of return and shorten the R&D cycle through technology clusters Build in technology transfer through cluster design and co-located supply chain 35 (5) Policy Response – Technology-Element Growth Model
Summary: Major Targets of Technology-Based Growth Policy
1) Increase industry’s aggregate investment in R&D and thereby leverage the Nation’s R&D intensity 2) Increase technology platform and infratechnology R&D support in early phases of R&D cycle 3) Deliver this support through more efficient policy mechanisms 4) Foster technology infrastructures that enable market entry by SMEs and adjust over technology life cycle 5) Update and expand the educational infrastructure 6) Increase the speed and breadth of the diffusion of new technologies 7) Enable rapid scale-up to commercially efficient volumes of production 8) Improve policy management techniques through planning & evaluation 36 (5) Policy Response – Technology-Element Growth Model
Advanced Manufacturing Policy Targets and Implementation Mechanisms Policy Target Policy Action Required Specific Targets Economic Impact Increase aggregate • Switch to a single flat 10 • 5 percent average R&D intensity for • Recognizes complementary investment in R&D percent credit manufacturing sector roles for industry and • Increase government funding • Reverse 50-year decline in Federal R&D intensity government
Increase amount and • Promote joint industry- • Expand funding through government-wide • Removes suboptimal efficiency of proof-of- government funding of proof- portfolio management mechanism portfolio structure resulting concept research that of-concept research targeted • Establish regional technology-based clusters from existing set of results in new at broad sector growth involving universities, government, and industry individual agency portfolios technology platforms • Promote partnering for proof- • Government laboratories • Facilitates strategic planning, Increase efficiency of of-concept technology complementary asset entire R&D cycle research sourcing, risk pooling, and • Fund infratechnology research rapid diffusion of technical knowledge
Foster effective • Ensure research consortia • Provide interface standards early in R&D cycle • Promotes market entry by industry structures within clusters include SMEs • Focus financial and technical infrastructure firms of all sizes, thereby and supporting • Improve infrastructure that support on young high-tech startups enabling maximum product technical supports start-ups • Provide university entrepreneur training and diversity and efficient system infrastructures incubators/accelerators integration • Provide infrastructure support for distribution Modernize and • Adjust college curricula • Increased STEM support through scholarships • Broadens and deepens expand educational • Change roles of high schools and career promotion skilled labor pool infrastructure and community colleges • Vocational training, apprenticeships through clusters Enable rapid scale-up • Expand manufacturing process • Support process demonstration and actual • Accelerates to commercially technology infrastructure shared production facilities commercialization and efficient production • Accelerate capital formation & • Provide procurement specifications through market share growth, volumes and diversity market penetration clearing house counteracting incentives • Investment tax credit from Asian nations 37 “Sooner or later, we sit down to a banquet of consequences”
– Robert Louis Stevenson
38