Globalization, structural change, and economic growth Dani Rodrik - PowerPoint PPT Presentation

Globalization, structural change, and economic growth Dani Rodrik April 2011 Based on a paper with the title “Globalization, Structural Change, and Productivity Growth,” authored jointly with Margaret McMillan (IFPRI). I acknowledge financial support from IFPRI and a joint IL-WTO project on "Making Globalization Socially Sustainable."

Productivity growth through structural change  Importance of mis-allocation  Recent work focuses on within-industry misallocation  Older tradition of dual-economy models focused on broad structural change  This work reminiscent of that older tradition and complements more disaggregated research

Productivity gaps are large in developing countries

Productivity gaps diminish over the course of development -.15 NGA -.2 Component plus residual SEN MWI KEN ETH -.25 ZMB BOL GHA THA VEN CHN MYS BRA COL IDN KOR PER -.3 PHL TWN NLD DNK HKG IND CHL ARG TUR MEX UKM ZAF SGP JPN ESP USA MUS ITA CRI -.35 SWE FRA 7 8 9 10 11 lnlabprod05 Relationship between inter-sectoral productivity gaps and income levels

Growth requires both new activities and ongoing structural change In %, = (agr_lprod_kppp00/non_agr_lprod_kppp00 100 100 2004 2002 2005 1999 2000 2001 80 1998 1997 1960 1996 1961 2003 1962 1995 1964 1963 1994 1967 1970 1965 1968 1969 1992 1993 1971 1966 1975 1973 60 1972 1990 1974 1977 1978 1976 1980 1981 1983 1991 1982 1984 1989 1979 1987 1985 1988 1985 1988 1986 1986 50 1989 1990 1982 1987 1993 1992 1994 1991 1984 1996 1995 1983 1998 1997 1981 40 1999 1979 1980 2001 2000 2002 1978 2003 1973 1974 1975 1976 2004 1972 1977 1971 1970 2005 1968 1991 2002 1967 1990 1989 1999 2000 2001 1969 2003 2004 1965 1966 19921993 2005 1962 1964 1950 1954 1960 1963 1988 1998 1955 1997 1994 1996 1961 1961 1995 1959 1960 1962 1957 1958 1964 1963 1965 1966 1970 1956 1967 1984 1969 1987 20 1968 1985 1986 1971 1983 1982 1972 1979 1981 1973 1978 1974 1977 1976 1975 1980 0 7 8 9 10 11 ln_sumlprod 7 8 9 10 11 ln_sumlprod IND FRA In %, = (agr_lprod_kppp00/non_agr_lprod_kppp00)*100 Fitted values PER Relationship between economy-wide labor productivity (horizontal axis) and the ratio of agricultural productivity to non-agricultural productivity (percent, vertical axis)

Data sources  Start from Groningen Growth and Development Center (GGDC) data base, which provides employment and real valued added statistics for 27 countries disaggregated into 10 sectors (Timmer and de Vries, 2007; 2009) – We converted local currency value added at 2000 prices to dollars using 2000 PPP exchange rates.  Complement with data from national sources for 11 additional countries (China, Turkey, and several African countries)  For the most part, VA comes from national income accounts, while level and structure of employment come from population censuses (and other household surveys) – Since employment data are not based on labor force or industrial surveys (save for extrapolation purposes), coverage of informal sector should be less problematic than otherwise

Sectors

Decomposing labor productivity growth ∑ ∑ ∆ = θ ∆ + ∆ θ Y y y − t i t k i t i t i t , , , , = = i n i n within structural change Y refers to aggregate labor productivity, y is sectoral labor productivity, θ is employment share, Δ is the first-difference operator, i indexes sectors, t -k and t stand for initial and final years.

Productivity decomposition 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 across regions LAC Productivity growth within sectors AFRICA within structural ASIA Productivity growth due to structural change HI -0.02 -0.01 0 0.01 0.02 0.03 0.04 0.05 Decomposition of productivity growth by country group, 1990-2005

The actual numbers

Country details

Patterns of structural change: Asia versus Latin America Correlation Between Sectoral Productivity and Correlation Between Sectoral Productivity and Change in Employment Shares in Asia (1990-2005) Change in Employment Shares in Latin America (1990-2005) Log of Sectoral Productivity/Total Productivity Log of Sectoral Productivity/Total Productivity β = 3.5826; t-stat = 2.20 β = -2.6866; t-stat = -1.17 3 3 pu min min min pu fire pu pu min min pu 2 pu fire min pu pu pu pu 2 con min minpu min min min pu fire pu min pu tsc min tsc pu min min man pu 1 fire man fire pu man min man wrt man fire con tsc tsc fire 1 tsc cspsgs tsc fire man tsc tsc fire fire tsc man fire man man con fire tsc wrt cspsgs man man con tsc tsc man man tsc man 0 min cspsgs man tsc man pu tsc tsc wrt wrt wrt con con wrt wrt con con man tsc cspsgs cspsgs wrt cspsgs wrt man cspsgs fire cspsgs tsc tsc fire con agr cspsgs con wrt con agr cspsgs 0 wrt agr man cspsgs agr min cspsgs cspsgs fire cspsgs fire wrt fire wrt con con cspsgs wrt con -1 agr cspsgs con cspsgs agr agr agr agr agr cspsgs cspsgs fire wrt agr con con fire agr agr agr con wrt min agr wrt wrt agr agr agr -1 pu wrt con -2 agr -.2 -.1 0 .1 -.2 -.1 0 .1 .2 Change in Employment Share Change in Employment Share ( ∆ Emp. Share) ( ∆ Emp. Share) Fitted values Fitted values *Note: β denotes coeff. of independent variable in regression equation: *Note: β denotes coeff. of independent variable in regression equation: α + β∆ Emp. Share α + β∆ Emp. Share ln(p/P) = ln(p/P) = Source: Authors' calculations with data from Timmer and de Vries (2009) and China's National Bureau of Statistics Source: Authors' calculations with data from Timmer and de Vries (2009) Latin America Asia

Argentina Correlation Between Sectoral Productivity and Change in Employment Shares in Argentina (1990-2005) Log of Sectoral Productivity/Total Productivity β = -7.0981; t-stat = -1.21 min 2 pu 1.5 1 man .5 tsc con 0 agr wrt cspsgs -.5 fire -.06 -.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: α + β∆ Emp. Share ln(p/P) = Source: Authors' calculations with data from Timmer and de Vries (2009)

Brazil Correlation Between Sectoral Productivity and Change in Employment Shares in Brazil (1990-2005) Log of Sectoral Productivity/Total Productivity β = -2.2102; t-stat = -0.17 pu 2 min 1 fire man con tsc 0 cspsgs agr -1 wrt -.1 -.05 0 .05 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: α + β∆ Emp. Share ln(p/P) = Source: Authors' calculations with data from Timmer and de Vries (2009)

Nigeria Correlation Between Sectoral Productivity and Change in Employment Shares in Nigeria (1990-2005) Log of Sectoral Productivity/Total Productivity β = -12.2100; t-stat = -1.06 6 min 4 fire 2 tsc pu wrt man con 0 agr -2 cspsgs -4 -.15 -.1 -.05 0 .05 .1 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: α + β∆ Emp. Share ln(p/P) = Source: Authors' calculations with data from Nigeria's National Bureau of Statistics and ILO's LABORSTA

Zambia Correlation Between Sectoral Productivity and Change in Employment Shares in Zambia (1990-2005) Log of Sectoral Productivity/Total Productivity β = -10.9531; t-stat = -3.25 3 fire 2 min pu con tsc man 1 wrt cspsgs 0 -1 agr -2 -.1 0 .1 .2 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: α + β∆ Emp. Share ln(p/P) = Source: Authors' calculations with data from CSO, Bank of Zambia, and ILO's KILM

India Correlation Between Sectoral Productivity and Change in Employment Shares in India (1990-2005) Log of Sectoral Productivity/Total Productivity β = 35.2372; t-stat = 2.97 2 pu fire tsc min 1 wrt con cspsgs man 0 -1 agr -.04 -.02 0 .02 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: α + β∆ Emp. Share ln(p/P) = Source: Authors' calculations with data from Timmer and de Vries (2009)