Employment, Structural Change, and Economic Development Dani Rodrik March 15, 2012
A remarkable reversal in fortunes since 1990s -.04 -.02 0.02.04.06 1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010 year Developed Developed smoothed Developing Developing smoothed Growth trends in developed and developing countries, 1950-2011
Growth was widespread, for once -.05 0.05.1 1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010 year Africa Asia (excl. Japan) Latin America smoothed Latin America Africa smoothed Asia smoothed Developing country growth trend by region, 1950-2011
Growth has been good for poverty reduction Source: http://siteresources.worldbank.org/intpovcalnet/resources/global_poverty_update_2012_02-29-12.pdf
Two interpretations Orthodox view: economic growth is conditional on good fundamentals defined in practice in terms of policies such as openness, deregulation, privatization, protection of property rights, contract enforcement, low inflation and budget deficits, Structuralist view: economic growth is conditional on rapid structural change which requires policies that stimulate employment in manufacturing industries (and modern services), including industrial policies and undervalued currencies While there is overlap between the two sets of policies, there are also tensions and differing priorities
Disappointments of orthodox policy reform 0.04 0.035 Brazil Mexico 0.03 0.025 0.02 0.015 0.01 0.005 0 1950-1980 1980-2000 2000-2008 Growth rates of Brazil and Mexico by period
Why structure of employment matters: Large productivity gaps
which tend to diminish over the course of development dispersion in labor Component productivity plus across residual broad sectors -.35 -.3 -.25 -.2 -.15 MWI ETH KEN ZMB GHA SEN NGA BOL IND CHN THA IDN BRA COL PER PHL VEN MYS KOR TWN NLD DNK HKG MEX TURCHL ARG ZAF UKM SGP ESP JPN USA CRI MUS ITA SWE FRA 7 8 9 10 11 Economy-wide lnlabprod05 labor productivity Relationship between inter-sectoral productivity gaps and income levels
Some sectors are special: industrial employment and growth growth (part orthogonal to other regressors) -.05 0.05.1 0.1.2.3.4.5 industrial employment share Each observation is a country over a 5-year period. Initial income and country and time fixed-effects are included.
Some sectors are special: unconditional convergence in manufacturing Pooled decadal industry convergence regressions, 1990-2007 unconditional (left panel) and conditional (right panel) growth of labor productivity (orthogonal part) -.4 -.2 0.2.4 growth of labor productivity (orthogonal part) -.4 -.2 0.2.4-4 -2 0 2 4 log initial labor productivity (orthogonal part) coef = -.03045731, (robust) se =.00399677, t = -7.62-4 -2 0 2 4 log initial labor productivity (orthogonal part) coef = -.0629993, (robust) se =.00519013, t = -12.14 Each observation corresponds to a 4-digit manufacturing industry in a specific country. Right panel includes country dummies.
Labor productivity convergence in specific industries Productivity convergence in individual industries, 2000-2005 -.6 -.5 -.4 -.3 -.2 -.1 GEO LTU SVK NLD LVA QAT CZE HUN ECU AUS AUT BGR POLSVN ERI MYS IDNMAR EST ZAF PRTESP CYP COL FRA KOR ITA FIN SWE DEU IND TUR MLTISR USA ETH BRAOMN GBR JPN JOR IRN URY SGP -.5 -.4 -.3 -.2 -.1 GEO BGR ZAF SVK QAT LTU ESTCZE HUN JORPOL SVN NOR IDN LVA MLT COL CYP ESPDNK PRT IRL FRA DEU AUT IND MAR ECU ITA GBR KOR SWE NLD FIN BRA ISRAUS OMN ETH MYS URY SGP ERI TUR IRN USA JPN LUX 4 6 8 10 12 footwear (isic 1920) 7 8 9 10 11 plastic products (isic 2520) -1 -.8 -.6 -.4 -.2 ERI GEO BGR MAC ECU ZAF LVA ALB ETH LTU QAT SVK AUS HUN POL CZE EST CYP FIN IND JORPSEMLT OMN SWE DNK MAR SVN BRA COL ESP FRA DEU ITA NOR AUT GBR NLD IRL USA KOR JPN SGP MYS TUR URY IRN MKD -.4 -.3 -.2 -.1 0 GEO SVK BGR ALB CZE HUN LTU EST MAC ECU ZAF POL COL QAT OMN LUX ERI LVA MLTCYP ESPSWE GBR IDN IND PRT SVN FRA AUT FIN DEU ITA NLD DNK NOR KOR MAR PSE USA BRA AUS ETH YEMJOR ISR JPN MYS SGP MKD IRN TUR URY 6 8 10 12 glass and glass products (isic 2610) 6 7 8 9 10 11 furniture (isic 3610)
Distribution of convergence coefficients, by 4- digit industry (normalized by standard error)
How employment structure determines growth: an accounting decomposition beta = convergence coefficient alpha = employment share theta = relative productivity within country g = exogenous growth rates in non-c sectors y = labor productivity y* = frontier labor productivity i industries; j countries; C : set of escalator industries
How structural factors explain growth slowdown in Latin America Productivity decomposition in Latin America across different periods (annual growth rates) 1950-1975 1975-1990 1990-2005 Sectoral productivity growth Structural change -0.01-0.005 0 0.005 0.01 0.015 0.02 0.025 0.03 0.035 0.04 0.045 Data from Pages, Carmen ed., The Age of Productivity, Inter-American Development Bank, Washington, D.C., 2010.
and the difference between slow and rapidly growing regions LAC Productivity growth within sectors AFRICA within ASIA HI structural Productivity growth due to structural change -0.02-0.01 0 0.01 0.02 0.03 0.04 0.05 Decomposition of productivity growth by country group, 1990-2005
Growth-reducing structural change: Argentina Log of Sectoral Productivity/Total Productivity -.5 0.5 1 1.5 2 Correlation Between Sectoral Productivity and Change in Employment Shares in Argentina (1990-2005) β = -7.0981; t-stat = -1.21 man agr -.06 -.04 -.02 0.02.04 Change in Employment Share ( Emp. Share) min pu Fitted values con wrt tsc fire cspsgs *Note: Size of circle represents employment share in 1990 **Note: β denotes coeff. of independent variable in regression equation: ln(p/p) = α + β Emp. Share Source: Authors' calculations with data from Timmer and de Vries (2009)
Growth-increasing structural change: Thailand Log of Sectoral Productivity/Total Productivity -1 0 1 2 3 Correlation Between Sectoral Productivity and Change in Employment Shares in Thailand (1990-2005) β = 5.1686; t-stat = 1.27 agr pu min tsc fire con man cspsgs -.2 -.1 0.1 Change in Employment Share ( Emp. Share) wrt Fitted values *Note: Size of circle represents employment share in 1990 **Note: β denotes coeff. of independent variable in regression equation: ln(p/p) = α + β Emp. Share Source: Authors' calculations with data from Timmer and de Vries (2009)
Structural change in Africa Log of Sectoral Productivity/Total Productivity -1 0 1 2 3 agr Correlation Between Sectoral Productivity and Change in Employment Shares in Africa (1990-2005) β = 6.9711; t-stat = 0.23 min wrt tsc pu fire con man cspsgs -.04 -.02 0.02.04 Change in Employment Share ( Emp. Share) Fitted values *Note: Size of circle represents employment share in 1990 **Note: β denotes coeff. of independent variable in regression equation: ln(p/p) = α + β Emp. Share Source: Authors' calculations Weighted average of nine African countries: ETH, GHA, KEN, MUS, MWI, NGA, SEN, ZAF, ZMB.
and with unemployment included Square Root of Sectoral Productivity/Total Productivity 0 1 2 3 4 5 Correlation Between Sectoral Productivity and Change in Employment Shares in Africa (1990-2005) β = -6.4365; t-stat = -0.30 agr wrt min tsc pu fire man con cspsgs -.06 -.04 -.02 0 Change in Employment Share.02.04 ( Emp. Share) unemp Fitted values *Note: Size of circle represents employment share in 1990 **Note: β denotes coeff. of independent variable in regression equation: sqrt(p/p) = α + β Emp. Share Source: Authors' calculations Weighted average of nine African countries: ETH, GHA, KEN, MUS, MWI, NGA, SEN, ZAF, ZMB.
What has worked: productivist policies Sound fundamentals Market-friendly policies Macro stability But also: Industrial policies in support of new economic activities A certain degree of repression of finance, to enable: Development banking Subsidized credit Undervaluation Undervalued currencies to promote tradables The more a country relies on industrial policies, the less the need for currency undervaluation, and vice versa
Will this time be different? Prognosis Policies are much better in conventional sense Macro stability and low inflation Openness to world economy Improved governance But these tend to increase resilience without necessarily igniting or sustaining growth Growth requires diversification and structural change, which is not automatic It necessitates pragmatic, experimental policies that support new industries These are often unconventional policies, difficult to employ effectively And they require a supportive external environment Will WTO/IMF, advanced countries condone such policies?