The China Syndrome. Local Labor Market Effects of Import Competition in the United States. David H. Autor, David Dorn, and Gordon H.

The China Syndrome Local Labor Market Effects of Import Competition in the United States David H. Autor, David Dorn, and Gordon H. Hanson AER, 2013 presented by Federico Curci April 9, 2014 Autor, Dorn, Hanson The China Syndrome April 9, 2014 1 / 26

Introduction Research question Study the effect of Chinese import competition between 1990 and 2007 on U.S. local labor markets Treat local labor markets as sub-economies subject to differential trade shocks according to initial patters of industry specialization Expand analysis of effect trade to non-wage variables Employment, unemployment, participation, wages, income, mobility, and transfers Autor, Dorn, Hanson The China Syndrome April 9, 2014 2 / 26

Introduction Research question Study the effect of Chinese import competition between 1990 and 2007 on U.S. local labor markets Treat local labor markets as sub-economies subject to differential trade shocks according to initial patters of industry specialization Expand analysis of effect trade to non-wage variables Employment, unemployment, participation, wages, income, mobility, and transfers Develop robust instrument variable approach Autor, Dorn, Hanson The China Syndrome April 9, 2014 2 / 26

Introduction Stylized facts Pre-1990s limited impact trade on U.S. labor Small imports from low-wage countries (only 9% U.S. manufacturing imports in 1991) Autor, Dorn, Hanson The China Syndrome April 9, 2014 3 / 26

Introduction Stylized facts Pre-1990s limited impact trade on U.S. labor Small imports from low-wage countries (only 9% U.S. manufacturing imports in 1991) Trend decline in U.S. manufacturing employment (from 12.6% in 1991 to 8.4% in 2007) Increase in import competition from China without an offsetting increase in demand for U.S. exports Imports raise to 28% in 2007 (China accounts for 89% growth) U.S. exports to China grew by much less Amplification mechanism because of current account imbalances between U.S. and China Autor, Dorn, Hanson The China Syndrome April 9, 2014 3 / 26

Introduction Stylized facts Pre-1990s limited impact trade on U.S. labor Small imports from low-wage countries (only 9% U.S. manufacturing imports in 1991) Trend decline in U.S. manufacturing employment (from 12.6% in 1991 to 8.4% in 2007) Increase in import competition from China without an offsetting increase in demand for U.S. exports Imports raise to 28% in 2007 (China accounts for 89% growth) U.S. exports to China grew by much less Amplification mechanism because of current account imbalances between U.S. and China Variation in regional manufacturing employment in U.S. and within-manufacturing import Autor, Dorn, Hanson The China Syndrome April 9, 2014 3 / 26

Introduction Stylized facts Figure 1: China s import penetration and U.S. manufacturing employment Autor, Dorn, Hanson The China Syndrome April 9, 2014 4 / 26

Introduction Literature- Trade shocks effects 1980s: modest effect trade on wage premium Autor, Dorn, Hanson The China Syndrome April 9, 2014 5 / 26

Introduction Literature- Trade shocks effects 1980s: modest effect trade on wage premium 1990s: grow literature in order to understand rising U.S. wage inequalities, focus on skill biased technical change 2000s: predominant focus on wages and employment in manufacturing industries or occupations Satisfactory results if labor markets geographically integrated and fully competitive Regional economies as unit of analysis Borjas and Ramey (1995), Chiquiar (2008), Topalova (2005, 2010) Industry or occupation as unit of analysis Autor, Dorn, Hanson The China Syndrome April 9, 2014 5 / 26

Introduction Literature- Trade shocks effects 1980s: modest effect trade on wage premium 1990s: grow literature in order to understand rising U.S. wage inequalities, focus on skill biased technical change 2000s: predominant focus on wages and employment in manufacturing industries or occupations Satisfactory results if labor markets geographically integrated and fully competitive Regional economies as unit of analysis Borjas and Ramey (1995), Chiquiar (2008), Topalova (2005, 2010) Industry or occupation as unit of analysis Bernard, Jensen and Schott (2006), Liu and Trefler (2008), Ebenstein et al. (2010) Autor, Dorn, Hanson The China Syndrome April 9, 2014 5 / 26

Introduction Literature- Trade shocks effects 1980s: modest effect trade on wage premium 1990s: grow literature in order to understand rising U.S. wage inequalities, focus on skill biased technical change 2000s: predominant focus on wages and employment in manufacturing industries or occupations Satisfactory results if labor markets geographically integrated and fully competitive Regional economies as unit of analysis Borjas and Ramey (1995), Chiquiar (2008), Topalova (2005, 2010) Industry or occupation as unit of analysis Bernard, Jensen and Schott (2006), Liu and Trefler (2008), Ebenstein et al. (2010) Feenstra (2010) Autor, Dorn, Hanson The China Syndrome April 9, 2014 5 / 26

Introduction Literature Growth in low-income country exports, China s transition to market economy and accession to WTO Chen, Jin, and Yue (2010), Hsieh and Klenow (2009), Naughton (2007) Assess gains and losses trade with low-income countries Gains Broda and Weinstein (2006), Goldberg et al. (2010), Bloom, Draca and Van Reenen (2009) Efficiency losses Broda and Weinstein (2006), Goldberg et al. (2010), Bloom, Draca and Van Reenen (2009) Regional adjustment to labor-market shocks Incomplete and slow mobility responses (lowest for non-college workers) Topel (1986), Blanchard and Katz (1992), Glaeser and Gyourko (2005), Artuc, Chaudhuri and McLaren (2010), Dix-Carneiro (2011) Autor, Dorn, Hanson The China Syndrome April 9, 2014 6 / 26

Theoretical predictions China experience productivity growth (transition from central planning to a market economy) and a reduction in its trade costs (accession to WTO), effect to labor market of U.S. region i? Increased competition in markets in which region i sells its output (change in China s export-supply capability) Increased demand for goods in China (change in expenditure in China on each industry) Autor, Dorn, Hanson The China Syndrome April 9, 2014 7 / 26

Theoretical predictions China experience productivity growth (transition from central planning to a market economy) and a reduction in its trade costs (accession to WTO), effect to labor market of U.S. region i? Increased competition in markets in which region i sells its output (change in China s export-supply capability) Increased demand for goods in China (change in expenditure in China on each industry) Model of trade Based on monopolistic competition and variation in industry labor productivity across countries Region i is a SOE, produce both traded and non-traded good Autor, Dorn, Hanson The China Syndrome April 9, 2014 7 / 26

Theoretical predictions Positive shock to China s export supply Decrease region i s wage and employment in traded goods Increase in employment in non-traded Autor, Dorn, Hanson The China Syndrome April 9, 2014 8 / 26

Theoretical predictions Positive shock to China s export supply Decrease region i s wage and employment in traded goods Increase in employment in non-traded Positive shock to China s import demand Increase region i s wage and employment in traded goods Decrease in employment in non-traded Final effect depend on trade imbalance Focus on import competition in U.S. U.S. imports from China vastly exceed U.S. exports to China U.S. market accounts for large majority of demand for most U.S. industries Autor, Dorn, Hanson The China Syndrome April 9, 2014 8 / 26

Empirical strategy Construction of the measure of local labor market exposure to import competition Autor, Dorn, Hanson The China Syndrome April 9, 2014 9 / 26

Empirical strategy Construction of the measure of local labor market exposure to import competition Identification strategy Autor, Dorn, Hanson The China Syndrome April 9, 2014 9 / 26

Empirical strategy Construction of the measure of local labor market exposure to import competition Identification strategy Benchmark results on manufacturing employment Autor, Dorn, Hanson The China Syndrome April 9, 2014 9 / 26

Empirical strategy Construction of the measure of local labor market exposure to import competition Identification strategy Benchmark results on manufacturing employment Other labor market effects Autor, Dorn, Hanson The China Syndrome April 9, 2014 9 / 26

Empirical strategy Construction of the measure of local labor market exposure to import competition Identification strategy Benchmark results on manufacturing employment Other labor market effects Alternative measures of labor market exposure Autor, Dorn, Hanson The China Syndrome April 9, 2014 9 / 26

Empirical strategy Construction of the measure of local labor market exposure to import competition Identification strategy Benchmark results on manufacturing employment Other labor market effects Alternative measures of labor market exposure Quantitative assessment efficiency losses Autor, Dorn, Hanson The China Syndrome April 9, 2014 9 / 26

Local labor market exposure to import competition I PW ui t = j L i j t L u j t M uc j t L i t Autor, Dorn, Hanson The China Syndrome April 9, 2014 10 / 26

Local labor market exposure to import competition I PW ui t = j L i j t L u j t M uc j t L i t Variation IPW caused by Concentration employment in manufacturing or non-manufacturing Specialization in import-intensive industries within local manufacturing Bias Realized U.S. imports from China correlated with unobserved shocks to U.S. product demand and U.S. employment Need instrument for growth in Chinese imports in U.S. ( I PW ) Autor, Dorn, Hanson The China Syndrome April 9, 2014 10 / 26

Identification strategy Instrument to identify supply-driven components of Chinese imports Chinese imports in eight other developed countries Autor, Dorn, Hanson The China Syndrome April 9, 2014 11 / 26

Identification strategy Instrument to identify supply-driven components of Chinese imports Chinese imports in eight other developed countries Identification assumption Common within-industry component of rising Chinese imports to U.S. and other high-income countries stems from China s rising comparative advantage and/or falling trade sectors Possible threats Product demand shocks may be correlated across high-income countries U.S. productivity shocks may be driving growth in imports from China Growth in imports from China may reflect technology shocks that adversely affect labor-intensive industries in high-income countries Autor, Dorn, Hanson The China Syndrome April 9, 2014 11 / 26

Identification strategy Instrument to identify supply-driven components of Chinese imports Chinese imports in eight other developed countries Identification assumption Common within-industry component of rising Chinese imports to U.S. and other high-income countries stems from China s rising comparative advantage and/or falling trade sectors Possible threats Product demand shocks may be correlated across high-income countries U.S. productivity shocks may be driving growth in imports from China Growth in imports from China may reflect technology shocks that adversely affect labor-intensive industries in high-income countries However, evidence that China s export growth strongly related to factor specific to China Autor, Dorn, Hanson The China Syndrome April 9, 2014 11 / 26

Identification strategy Instrument validity Autor, Dorn, Hanson The China Syndrome April 9, 2014 12 / 26

Benchmark estimation Data sources (from 1991 to 2007): UN Comtrade, U.S. Census, American Community Survey, Bureau of Economic Analysis, Social Security Administration Autor, Dorn, Hanson The China Syndrome April 9, 2014 13 / 26

Benchmark estimation Data sources (from 1991 to 2007): UN Comtrade, U.S. Census, American Community Survey, Bureau of Economic Analysis, Social Security Administration 2SLS model in first difference Second stage: L m i t = γ t + β 1 I PW ˆ ui t + X i t β 2 + e ct First stage: I PW ˆ ui t = ˆβ 3 I PW oi t + ɛ ct Assumptions: E [ I PW oi t I PW ui t ] 0; E [ I PW oi t e ct ] = E [ X i t e ] ct = 0 Demographic and labor force controls Share of manufacturing in a CZ s start-of-period employment Region dummy Start-of-period share of population with college education, foreign born and working age women Susceptibility of a CZ s occupations to substitution by technology or task offshoring Autor, Dorn, Hanson The China Syndrome April 9, 2014 13 / 26

Benchmark estimation Results with no controls Figure 3: Benchmark estimation - no controls Autor, Dorn, Hanson The China Syndrome April 9, 2014 14 / 26

Benchmark estimation Augmented regression Figure 4: Benchmark estimation - augmented model Autor, Dorn, Hanson The China Syndrome April 9, 2014 15 / 26

Benchmark estimation Results Chinese import competition explains 21 percent of decline manufacturing employment in U.S. from 1991 to 2007 (982 thousand workers) Results robust to drop of housing and construction sectors, computer industry and consumer good industries No change estimation by adding other low-income countries Autor, Dorn, Hanson The China Syndrome April 9, 2014 16 / 26

Labor Market effects estimation Reallocation of worker Figure 5: Reallocation of workers estimation Autor, Dorn, Hanson The China Syndrome April 9, 2014 17 / 26

Labor Market effects estimation Reallocation of worker If mobility response is large, unlikely to find indirect effects of trade on local labor markets (initial local impact will rapidly diffuse across regions) No robust evidence that shocks to local manufacturing lead to substantial change in population Several explanations Shocks to manufacturing from China too small to affect outcomes in broader CZ Good markets very well integrated nationally, local labor markets adjust without a mobility response Autor, Dorn, Hanson The China Syndrome April 9, 2014 18 / 26

Labor Market effects estimation Reallocation of worker If mobility response is large, unlikely to find indirect effects of trade on local labor markets (initial local impact will rapidly diffuse across regions) No robust evidence that shocks to local manufacturing lead to substantial change in population Several explanations Shocks to manufacturing from China too small to affect outcomes in broader CZ Good markets very well integrated nationally, local labor markets adjust without a mobility response Cost of moving geographically and between sectors and transfers and house price bear part of the incidence of labor demand shocks Autor, Dorn, Hanson The China Syndrome April 9, 2014 18 / 26

Labor Market effects estimation Labor force Figure 6: Labor force estimation Autor, Dorn, Hanson The China Syndrome April 9, 2014 19 / 26

Labor Market effects estimation Labor force Reduction in manufacturing employment no offset by a rise non-manufacturing employment Autor, Dorn, Hanson The China Syndrome April 9, 2014 20 / 26

Labor Market effects estimation Labor force Reduction in manufacturing employment no offset by a rise non-manufacturing employment 1000$ per worker increase in CZ s import exposure reduces employment to population ratio by 0.77 percent 3/4 because of loss in manufacturing and 1/4 because of non-manufacturing 1/4 reduction in employment because of rise in unemployment and 3/4 because of higher non-participation Autor, Dorn, Hanson The China Syndrome April 9, 2014 20 / 26

Labor Market effects estimation Labor force Reduction in manufacturing employment no offset by a rise non-manufacturing employment 1000$ per worker increase in CZ s import exposure reduces employment to population ratio by 0.77 percent 3/4 because of loss in manufacturing and 1/4 because of non-manufacturing 1/4 reduction in employment because of rise in unemployment and 3/4 because of higher non-participation More pronounced effects for no college adults Autor, Dorn, Hanson The China Syndrome April 9, 2014 20 / 26

Labor Market effects estimation Wages Figure 7: Wage estimation General negative effect import exposure on average weekly earnings Autor, Dorn, Hanson The China Syndrome April 9, 2014 21 / 26

Labor Market effects estimation Wages Partial and incomplete labor market adjustments Labor and product markets are not sufficiently integrated (shock to local manufacturing decrease total CZ employment) Manufacturing wages are downwardly rigid Non-manufacturing subject to negative demand shocks and positive labor supply shocks GE effect within but not across local labor markets labor markets Autor, Dorn, Hanson The China Syndrome April 9, 2014 22 / 26

Labor Market effects estimation Transfers and income Figure 8: Transfer estimation Negative total effect on total income considering change in wage, business invest and transfers Autor, Dorn, Hanson The China Syndrome April 9, 2014 23 / 26

Alternative measures of trade exposure Results are robust to different measures of trade exposure Include competition in foreign markets Consider that imports from China include both final goods purchased by U.S. consumers and intermediate inputs purchased by U.S. firms Consider net import from China Apply gravity residual: replace growth in U.S. imports from China with inferred change in China s comparative advantage Use factor content of U.S. net imports from China (change in net import of effective labor services) Autor, Dorn, Hanson The China Syndrome April 9, 2014 24 / 26

Assess efficiency losses Not assess gains trade with low-income countries Consumers: lower product prices or increased varieties Firms: lower input prices, greater varieties and higher invest in innovation Deadweight loss associated with government transfers Annual per capita transfers increase by 58 $ for every 1000 $ of additional import exposure per worker Gruber (2010): estimation of marginal excess burden of taxation required to fund transfers is 40 cents on the Dollar Range of deadweight losses range from 22 $ to 44 $ Autor, Dorn, Hanson The China Syndrome April 9, 2014 25 / 26

Assess efficiency losses Not assess gains trade with low-income countries Consumers: lower product prices or increased varieties Firms: lower input prices, greater varieties and higher invest in innovation Deadweight loss associated with government transfers Annual per capita transfers increase by 58 $ for every 1000 $ of additional import exposure per worker Gruber (2010): estimation of marginal excess burden of taxation required to fund transfers is 40 cents on the Dollar Range of deadweight losses range from 22 $ to 44 $ Involuntary reductions in labor force participation Deadweight losses if the market wage of involuntary displaced workers exceeds their value of leisure Estimated deadweight loss due to involuntary employment is between 87 $ and 137 $ Autor, Dorn, Hanson The China Syndrome April 9, 2014 25 / 26

Conclusions Extend effects of import from low-income countries to local labor markets Autor, Dorn, Hanson The China Syndrome April 9, 2014 26 / 26

Conclusions Extend effects of import from low-income countries to local labor markets Effect Chinese import competition are through reducing manufacturing employment, increasing unemployment, decline in wages, increase in transfer payments and reduction in average earnings Chinese import competition explains 21 percent of decline manufacturing employment in U.S. from 1991 to 2007 (982 thousand workers) Autor, Dorn, Hanson The China Syndrome April 9, 2014 26 / 26

Conclusions Extend effects of import from low-income countries to local labor markets Effect Chinese import competition are through reducing manufacturing employment, increasing unemployment, decline in wages, increase in transfer payments and reduction in average earnings Chinese import competition explains 21 percent of decline manufacturing employment in U.S. from 1991 to 2007 (982 thousand workers) Partial and incomplete labor market adjustments Autor, Dorn, Hanson The China Syndrome April 9, 2014 26 / 26

Conclusions Extend effects of import from low-income countries to local labor markets Effect Chinese import competition are through reducing manufacturing employment, increasing unemployment, decline in wages, increase in transfer payments and reduction in average earnings Chinese import competition explains 21 percent of decline manufacturing employment in U.S. from 1991 to 2007 (982 thousand workers) Partial and incomplete labor market adjustments Robust results Autor, Dorn, Hanson The China Syndrome April 9, 2014 26 / 26