De-industrialization of the Riches and the Rise of China MURAT ÜNGÖR http://www.muratungor.com/ February 15, 2012 Employment Shares by Sector, U.S. 1840-2000 %90 80 70 Agriculture 60 50 40 30 20 10 0 1840 1860 1880 1900 1920 1940 1960 1980 2000 1 / 55 Employment Shares by Sector, U.S. 1840-2000 %90 80 70 Agriculture 60 50 40 Manufacturing 30 20 10 0 1840 1860 1880 1900 1920 1940 1960 1980 2000 2 / 55 Employment Shares by Sector, U.S. 1840-2000 %90 Services 80 70 Agriculture 60 50 40 Manufacturing 30 20 10 0 1840 1860 1880 1900 1920 1940 1960 1980 2000 3 / 55 Sources of Structural Change I Two channels that can account for structural change I Sectoral differences in productivity growth (Baumol 1967; Ngai and Pissarides 2007) I Sectoral differences in income elasticities of demand (Kongsamut, Rebelo, and Xie 2001) I The majority of the existing studies on structural change examine the experience of a country in isolation I Critics of Closed Economy Models I I Buera and Kaboski (2009), Matsuyama (2009) Some Recent Attempts I Yi and Zhang (2011), Teignier (2011), Stefanski (2011) 4 / 55 Contribution I Important questions I How does structural change in one country will slow down or speed up structural change in other countries? I How does an emerging giant alter the global allocation of resources? 5 / 55 Contribution I I Important questions I How does structural change in one country will slow down or speed up structural change in other countries? I How does an emerging giant alter the global allocation of resources? This paper I Investigates impact of the rise of China on structural transformation observed in U.S. between 1978 and 2005. 6 / 55 Contribution I I Important questions I How does structural change in one country will slow down or speed up structural change in other countries? I How does an emerging giant alter the global allocation of resources? This paper I Investigates impact of the rise of China on structural transformation observed in U.S. between 1978 and 2005. I A closed economy model accounts for 32.8 percent of the de-industrialization while an open economy model accounts for 62.6 percent of it. 7 / 55 Contribution I I Important questions I How does structural change in one country will slow down or speed up structural change in other countries? I How does an emerging giant alter the global allocation of resources? This paper I Investigates impact of the rise of China on structural transformation observed in U.S. between 1978 and 2005. I A closed economy model accounts for 32.8 percent of the de-industrialization while an open economy model accounts for 62.6 percent of it. I Open economy model has more explanatory power to explain the de-industrialization in the post-1990 period accounting for more than 80 percent of it. 8 / 55 Structural Change in a Closed Economy 9 / 55 Households and preferences U(Ā, C ) = Ā + log(C ) (1) The instantaneous utility is defined over the agricultural good (Ā) and the composite consumption good (C ): C = (γ 1−η I η + (1 − γ)1−η S η )1/η (2) At each date, and given prices, the household chooses consumption of each good to maximize his lifetime utility subject to the budget constraint: pA Ā + pI I + pS S = ω (3) 10 / 55 Firms and technologies The production function for sector j is given by Yj = θj Lj (4) Firm j problem is given by max pj Yj − ωLj s.t. Yj = θj Lj , Lj > 0 (5) 11 / 55 Equilibrium Given a set of prices {pA , pI , pS , ω}, a competitive equilibrium consists of consumption decisions that are the household’s allocations {Ā, I , S}, and factor allocations for the firms {LA , LI , LS } such that given prices, the firm’s allocations solve its profit maximization problem, the household’s allocations solve the household’s utility maximization problem, and factor and product markets clear: LA + LI + LS = 1 (6) Ā = YA , I = YI , S = YS (7) 12 / 55 Sectoral Employment Shares Employment share in agriculture: LA = Ā/θA (8) 13 / 55 Sectoral Employment Shares Employment share in agriculture: LA = Ā/θA (8) Employment share in industry: LI = ∆(1 − (Ā/θA )) , 1+∆ (9) where ∆ ≡ (γ/(1 − γ))(θI /θS )−η/(η−1) . 14 / 55 Sectoral Employment Shares Employment share in agriculture: LA = Ā/θA (8) Employment share in industry: LI = ∆(1 − (Ā/θA )) , 1+∆ (9) where ∆ ≡ (γ/(1 − γ))(θI /θS )−η/(η−1) . Proposition When 1/(1 − η) < 1 (gross complementarity), faster productivity growth in industry leads to Baumol’s prediction: labor goes to the slow-growing service sector. ∂LI > 0, if 1/(1 − η) > 1; ∂θI ∂LI < 0, if 1/(1 − η) < 1. ∂θI 15 / 55 Calibration for the U.S., 1950-2005 I Normalize productivity levels across sectors to 1 in 1950: j θ1950 =1 I Use data on sectoral productivity growth to get time paths: j θt+1 = (1 + gtj )θtj I I Calibrate the preference parameters to match the initial employment shares Consider 1/(1 − η) to be a free parameter in order to study the implication of this parameter value for the findings I Bah (2009), Duarte and Restuccia (2010), Ngai and Pissarides (2004, 2008), Rogerson (2008) argue and cite the empirical literature that 1/(1 − η) ' 0.1 − 0.45 I I study three values of 1/(1 − η) 16 / 55 Sectoral Employment Shares, 1950 - 2005 Agriculture %9 Model Data 7 5 3 1 1950 1960 1970 1980 1990 2000 2010 17 / 55 Sectoral Employment Shares, 1950 - 2005 Industry Services %34 %82 77 28 72 67 22 62 Data 1/(1-η)=0.45 16 1950 1980 2010 57 1950 1980 2010 18 / 55 Sectoral Employment Shares, 1950 - 2005 Industry Services %34 %82 77 28 72 67 22 62 Data 1/(1-η)=0.30 1/(1-η)=0.45 16 1950 1980 2010 57 1950 1980 2010 19 / 55 Sectoral Employment Shares, 1950 - 2005 Industry Services %34 %82 77 28 72 67 22 62 Data 1/(1-η)=0.10 1/(1-η)=0.30 1/(1-η)=0.45 16 1950 1980 2010 57 1950 1980 2010 20 / 55 Problem with the Closed Economy I Differential productivity gains in industry do not necessarily explain de-industrialization I A major problem is that productivity growth in industry relative to services is not high enough to move labor out of the industrial sector I This problem is robust under different values of the elasticity of substitution between industry and services 21 / 55 Does Openness Matter? China has experienced high productivity growth in industry (especially after 1990) If modeled, will improve upon 1. Technology-based explanations (different rates of sectoral productivity growth) 2. Utility-based explanations (different income elasticities for different goods) that can’t account for last 3-4 decades in closed economy framework 22 / 55 U.S. Trade in Goods: Imports from %24 20 Canada 16 12 8 4 0 1978 1988 1998 2008 23 / 55 U.S. Trade in Goods: Imports from %24 20 Canada 16 Japan 12 8 4 0 1978 1988 1998 2008 24 / 55 U.S. Trade in Goods: Imports from %24 20 Canada 16 Dragons Japan 12 8 4 0 1978 1988 1998 2008 25 / 55 U.S. Trade in Goods: Imports from %24 20 Canada 16 Dragons Japan 12 8 Germany 4 0 1978 1988 1998 2008 26 / 55 U.S. Trade in Goods: Imports from %24 20 Canada 16 Dragons Japan 12 Mexico 8 Germany 4 0 1978 1988 1998 2008 27 / 55 U.S. Trade in Goods: Imports from %24 20 Canada 16 Dragons Japan 12 Mexico 8 Germany 4 China 0 1978 1988 1998 2008 28 / 55 Structural Change in an Open Economy 29 / 55 Firm j Problem in Country i max pji Yji − ω i Lij s.t. Yji = θji Lij Lij > 0 where pji is producer price for sector j in country i Yji is output produced in sector j in country i ω i is real wage in country i Lij is labor employed in sector j in country i θji is productivity of sector j in country i 30 / 55 The representative agent’s utility function in country i is U(Āi , C i ) = Āi + log(C i ) 31 / 55 The representative agent’s utility function in country i is U(Āi , C i ) = Āi + log(C i ) 1/η C i = (γ i )1−η (I i )η + (1 − γ i )1−η (S i )η 32 / 55 The representative agent’s utility function in country i is U(Āi , C i ) = Āi + log(C i ) 1/η C i = (γ i )1−η (I i )η + (1 − γ i )1−η (S i )η 1/(1 − η) is substitution elasticity between industry and services γ i is share of industry in non-agricultural consumption in country i 33 / 55 Industrial good consumption in country i 1/ρ 1−ρ I i = (µi )1−ρ (Iii )ρ + (1 − µi ) (Iki )ρ 1/(1 − ρ) is the elasticity of substitution between home and foreign goods µi is home-product consumption bias Consumption bias is a stand-in for trade costs Free borrowing / lending in a complete set of asset markets 34 / 55 Industrial good consumption in country i 1/ρ 1−ρ I i = (µi )1−ρ (Iii )ρ + (1 − µi ) (Iki )ρ 1/(1 − ρ) is the elasticity of substitution between home and foreign goods µi is home-product consumption bias Consumption bias is a stand-in for trade costs Free borrowing / lending in a complete set of asset markets I Solve the World’s social planner problem 35 / 55 Industrial good consumption in country i 1/ρ 1−ρ I i = (µi )1−ρ (Iii )ρ + (1 − µi ) (Iki )ρ 1/(1 − ρ) is the elasticity of substitution between home and foreign goods µi is home-product consumption bias Consumption bias is a stand-in for trade costs Free borrowing / lending in a complete set of asset markets I Solve the World’s social planner problem I A sequence of static resource allocation problems 36 / 55 World’s Social Planner Problem Max αU ĀUS , C US + (1 − α)U ĀCh , C Ch s.t. YAI = θAi LiA = Āi YIi = θIi LiI = Iii + Iik YSi = θSi LiS = S i LiA + LiI + LiS = 1 37 / 55 Quantitative Analysis 1. Numerical experiments to answer I How does openness effect de-industrialization in the United States? I What would have happened if the Chinese economy had productivity in industry same as the U.S.? I How does the elasticity of substitution between home and foreign goods affect the sectoral re-allocation of labor? 38 / 55 Quantitative Analysis 1. Numerical experiments to answer I How does openness effect de-industrialization in the United States? I What would have happened if the Chinese economy had productivity in industry same as the U.S.? I How does the elasticity of substitution between home and foreign goods affect the sectoral re-allocation of labor? 2. Data and Parameters I Feed the path of actual sectoral productivity growth rates I Parameterization of the model 39 / 55 Time Path of Exogenous Variables The exogenous variables are sectoral productivities in each country 40 / 55 Time Path of Exogenous Variables The exogenous variables are sectoral productivities in each country i θj,1978 =1 Normalization 41 / 55 Time Path of Exogenous Variables The exogenous variables are sectoral productivities in each country i θj,1978 =1 Normalization Use data on the growth rate of sectoral value added per worker in country i to obtain the time paths of sectoral productivity i i i , )θj,t θj,t+1 = (1 + gj,t for t = 1978, 1979, ...2004 42 / 55 Parameters Subsistence in agriculture: Āi I to match employment share of agriculture in 1978 Consumption weight on industrial goods: γ i I to match employment share of industry in 1978 Home-product consumption bias: µi I to match the average import ratio η: elasticity of substitution between tradables and nontradables I 0.45 ρ: elasticity of substitution between home and foreign goods I International Economics Literature: 1.5 43 / 55 Sectoral Employment Shares in China Agriculture %72 54 36 Data Model 18 1978 1987 1996 2005 44 / 55 Sectoral Employment Shares in China Industry Services %45 %40 38 33 31 26 24 19 Data Autarky 17 1978 1993 2008 12 1978 1993 2008 45 / 55 Sectoral Employment Shares in China Industry Services %45 %40 38 33 31 26 24 19 Data Autarky Open w/o wedge 17 1978 1993 2008 12 1978 1993 2008 46 / 55 Sectoral Employment Shares in China Industry Services %45 %40 38 33 31 26 24 19 Data Autarky Open w/o wedge Open with wedge 17 1978 1993 2008 12 1978 1993 2008 47 / 55 Sectoral Employment Shares in the U.S. (a) Agriculture %2.9 (b) Industry (c) Services %29 %81 26 78 23 75 20 72 2.2 1.5 Data Autarky Open Data Model 0.8 1978 1993 2008 17 1978 1993 2008 69 1978 1993 2008 48 / 55 What would have happened if the Chinese economy had productivity in agriculture / industry / services / all sectors same as the United States? 49 / 55 Role of Productivity Differentials Industrial Employment Share in the U.S. %29 26 23 20 Data Autarky Open All Sectors 17 1978 1988 1998 2008 50 / 55 Role of Productivity Differentials in Industry and Services (a) Role of Industry (b) Role of Services %29 Industrial Employment Share in the U.S. Industrial Employment Share in the U.S. %29 26 23 20 Data Autarky Open All Sectors Industry Only 17 1978 1988 1998 2008 26 23 20 Data Autarky Open All Sectors Services Only 17 1978 1988 1998 2008 51 / 55 Sensitivity Analysis in the Open Economy Sensitivity for η Sensitivity for ρ %29 %29 26 26 23 23 20 20 Data 1/(1-ρ)=1.2 1/(1-ρ)=1.5 1/(1-ρ)=2.0 Data 1/(1-η)=0.30 1/(1-η)=0.45 1/(1-η)=0.74 17 1978 1993 2008 17 1978 1993 2008 52 / 55 Trade Variables TB to GDP %0.4 TV to GDP RER 2.4 %3.3 2.0 1.6 2.2 log(RER) -0.8 -2.0 1.2 0.8 1.1 0.4 0.0 Data Model -3.2 1978 1993 2008 0 1978 1993 2008 -0.4 1978 1993 2008 53 / 55 Discussion on the Chinese Data Chinese Industrial Productivity 8 Holz Maddison-Wu 7 U.S. Industrial Employment Share %29 26 1978 = 1 6 5 23 4 3 20 Data Autarky Holz Data Maddison-Wu Data 2 1 1978 1992 2006 17 1978 1992 2006 54 / 55 Concluding Remarks I Openness matters for sectoral reallocation of labor! I If the Chinese economy had productivity in industry same as the U.S., then the role of openness is diminished I The higher the elasticity of substitution between home and foreign goods is, the more accelerated structural transformation I Several additional questions can be profitably addressed in future work 55 / 55
© Copyright 2025 Paperzz