Globalization, structural change, and economic growthsiteresources.worldbank.org/DEC/Resources/84797-1298663992214/... · Globalization, structural change, and economic ... Margaret

Globalization, structural change, and economic

growth

Dani RodrikApril 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

MWI

ETH

ZMBGHA

KEN

SEN

NGA

BOL

IND

CHN

PHLIDN

BRAPER

THA

COL

CRI

VEN

MEXTURCHLARG

MYSKOR

MUS

ZAF

DNKTWN

ESP

UKM

JPN

SWEITA

NLD

FRA

SGP

HKG

USA

-.35

-.3-.2

5-.2

-.15

Com

pone

nt p

lus

resi

dual

7 8 9 10 11lnlabprod05

Relationship between inter-sectoral productivity gaps and income levels

Growth requires both new activities and ongoing structural change

050

100

7 8 9 10 11ln_sumlprod

In %, = (agr_lprod_kppp00/non_agr_lprod_kppp00)*100 Fitted values

196019611962

19631964

19651966

1967196819691970

1971

19721973

19741975

197619771978

197919801981

198219831984198519861987

1988198919901991199219931994

1995199619971998

199920002001

200220032004

2005

1950 19541955

1956195719581959

19601961

1962196319641965196619671968

1969197019711972

19731974197519761977

1978

197919801981

1982

19831984

19851986

198719881989

1990

1991

19921993

1994

1995

199619971998

199920002001

2002

2003

2004

2005

19601961196219631964

196519661967196819691970197119721973197419751976197719781979

1980198119821983

1984198519861987

1988

19891990199119921993

19941995199619971998

1999200020012002200320042005

2040

6080

100

In %

, = (a

gr_l

prod

_kpp

p00/

non_

agr_

lpro

d_kp

pp00

7 8 9 10 11ln_sumlprod

IND FRAPER

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

∑∑==

− ∆+∆=∆ni

tititini

ktit yyY ,,,, θθ

within structural change

Y refers to aggregate labor productivity, y is sectoral labor productivity, θis 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)

-0.01 -0.005 0 0.005 0.01 0.015 0.02 0.025 0.03 0.035 0.04 0.045

1990 - 2005

1975 - 1990

1950 - 1975

Sectoral productivitygrowthStructural change

Data from Pages, Carmen ed., The Age of Productivity, Inter-American Development Bank, Washington, D.C., 2010.

… and across regions

-0.02 -0.01 0 0.01 0.02 0.03 0.04 0.05

HI

ASIA

AFRICA

LAC

within

structural

Decomposition of productivity growth by country group, 1990-2005

Productivity growth within sectors

Productivity growth due to structural change

The actual numbers

Country details

Patterns of structural change: Asia versus Latin America

agr con

cspsgsfire

man

min

pu

tsc

wrtagr

con

cspsgs

fire

man

min

pu

tsc

wrtagr

con

cspsgs

fire

man

minpu

tsc

wrt

agr

con

cspsgs

fireman

minpu

tsc

wrtagr

concspsgsfireman

min

pu

tsc

wrtagr concspsgs

fire

man

min

pu

tsc

wrt

agrcon

cspsgs fireman

minputsc

wrt

agr

con

cspsgs

fireman

minpu

tsc

wrt

agr

con

cspsgsfire

man

min

pu

tsc

wrt

-10

12

3Lo

g of

Sec

tora

l Pro

duct

ivity

/Tot

al P

rodu

ctiv

ity

-.2 -.1 0 .1 .2Change in Employment Share

(∆Emp. Share)

Fitted values

*Note: β denotes coeff. of independent variable in regression equation: ln(p/P) = α + β∆Emp. ShareSource: Authors' calculations with data from Timmer and de Vries (2009)

β = -2.6866; t-stat = -1.17

Correlation Between Sectoral Productivity andChange in Employment Shares in Latin America (1990-2005)

agr

con cspsgs

fire

man

minpu

tscwrt

agr

con

cspsgs

fire

man

min

pu

tsc wrt

agr

concspsgs

fire

man

min

pu

tsc wrt

agr

concspsgs

fire

man

min

pu

tsc

wrt

agr

con

cspsgs

fire

man

min

pu

tsc

wrtagr

con

cspsgs

fire

man

min

pu

tsc

wrtagr

concspsgs

fireman

min

pu

tscwrt

agrcon cspsgs

fireman

min

pu

tscwrt

agr

con

cspsgsfire

man

minpu

tsc

wrt

agr con

cspsgsfire

man

min

pu

tsc

wrt

-2-1

01

23

Log

of S

ecto

ral P

rodu

ctiv

ity/T

otal

Pro

duct

ivity

-.2 -.1 0 .1Change in Employment Share

(∆Emp. Share)

Fitted values

*Note: β denotes coeff. of independent variable in regression equation: ln(p/P) = α + β∆Emp. ShareSource: Authors' calculations with data from Timmer and de Vries (2009) and China's National Bureau of Statistics

β = 3.5826; t-stat = 2.20

Correlation Between Sectoral Productivity andChange in Employment Shares in Asia (1990-2005)

Asia Latin America

Argentina

agrcon

cspsgsfire

man

min

pu

tsc

wrt

-.50

.51

1.5

2Lo

g of

Sec

tora

l Pro

duct

ivity

/Tot

al P

rodu

ctiv

ity

-.06 -.04 -.02 0 .02 .04Change 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. ShareSource: Authors' calculations with data from Timmer and de Vries (2009)

β = -7.0981; t-stat = -1.21

Correlation Between Sectoral Productivity andChange in Employment Shares in Argentina (1990-2005)

Brazil

agr

con

cspsgs

fire

man

min

pu

tsc

wrt-10

12

Log

of S

ecto

ral P

rodu

ctiv

ity/T

otal

Pro

duct

ivity

-.1 -.05 0 .05Change 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. ShareSource: Authors' calculations with data from Timmer and de Vries (2009)

β = -2.2102; t-stat = -0.17

Correlation Between Sectoral Productivity andChange in Employment Shares in Brazil (1990-2005)

Nigeria

agr

con

cspsgs

fire

man

min

putscwrt

-4-2

02

46

Log

of S

ecto

ral P

rodu

ctiv

ity/T

otal

Pro

duct

ivity

-.15 -.1 -.05 0 .05 .1Change 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. ShareSource: Authors' calculations with data from Nigeria's National Bureau of Statistics and ILO's LABORSTA

β = -12.2100; t-stat = -1.06

Correlation Between Sectoral Productivity andChange in Employment Shares in Nigeria (1990-2005)

Zambia

agr

con

cspsgs

fire

man

min pu

tscwrt

-2-1

01

23

Log

of S

ecto

ral P

rodu

ctiv

ity/T

otal

Pro

duct

ivity

-.1 0 .1 .2Change 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. ShareSource: Authors' calculations with data from CSO, Bank of Zambia, and ILO's KILM

β = -10.9531; t-stat = -3.25

Correlation Between Sectoral Productivity andChange in Employment Shares in Zambia (1990-2005)

India

agr

concspsgs

fire

man

min

pu

tsc

wrt

-10

12

Log

of S

ecto

ral P

rodu

ctiv

ity/T

otal

Pro

duct

ivity

-.04 -.02 0 .02Change 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. ShareSource: Authors' calculations with data from Timmer and de Vries (2009)

β = 35.2372; t-stat = 2.97

Correlation Between Sectoral Productivity andChange in Employment Shares in India (1990-2005)

Thailand

agr

con

cspsgsfire

man

minpu

tsc

wrt

-10

12

3Lo

g of

Sec

tora

l Pro

duct

ivity

/Tot

al P

rodu

ctiv

ity

-.2 -.1 0 .1Change 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. ShareSource: Authors' calculations with data from Timmer and de Vries (2009)

β = 5.1686; t-stat = 1.27

Correlation Between Sectoral Productivity andChange in Employment Shares in Thailand (1990-2005)

What’s going on? Some possibilities:

Some countries have more “surplus labor” in agriculture than others

Role of comparative advantage: primary products versus manufactures

Labor market rigidity: large hiring/firing costs may prevent employment growth in dynamic sectors

Trade/industrial/currency policies have:– Not encouraged new tradable activities sufficiently– Exposed tradables to import competition too early

and excessively

Large reservoir of “excess labor” helps, but only conditionally

SGP

HKG

USAUKMSWENLD

DNKFRAITAJPNESP

ARG

TWN

VEN

ZAFMUS

KORCHL

MEXBRA

MYSCRI

COLPER

BOL

PHL

TUR

NGA

IDN

ZMB

CHN

THA

GHA

SEN

IND

KEN

MWI

ETH

-.08

-.06

-.04

-.02

0.0

2C

ompo

nent

plu

s re

sidu

al

0 .2 .4 .6 .8 1Labor share of agriculture, 1990

SGP

HKG

USAUKM

SWE

NLD

DNK

FRA

ITA

JPN

ESPARG

VEN

ZAFMUS

KOR

CHL

MEX

BRAMYS

CRI

COL

PER

BOLPHL

TURNGA

IDN

ZMB

CHN

THA

GHA

SEN

IND

KEN

MWIETH

-.04

-.02

0.0

2.0

4C

ompo

nent

plu

s re

sidu

al0 .2 .4 .6 .8 1

Labor share of agriculture, 1990

Association between the initial labor share in agriculture and the contribution of structural change to growth

unconditional conditional

Comparative advantage in primary products is bad news

JPN

HKG

KOR

ITASWETWN

UKMUSASGP

PHL

TUR

CHNESPFRA

IND

MUSTHA

MEX

MYSDNKNLDZAF

BRA

SEN

KENCRI

IDNETH

COLARG

GHA

PERCHL

BOL

ZMB

MWI

VEN

NGA

-.1-.0

8-.0

6-.0

4-.0

20

Com

pone

nt p

lus

resi

dual

0 .2 .4 .6 .8 1index_exp_rawmat

Partial association between the share of primary products in exports and the contribution of structural change to growth

But policy can clearly help: currency undervaluation

JPN

DNK

SWE

NLDFRA

UKMITA

NGA

ESPUSA

ZMB

MEX

TWNSGP

KOR

HKGTURPERARG

SEN

VEN

BRA

CRIZAF

BOL

KEN

GHA

CHL

MWI

MYSCOL

THA

ETH

PHL

CHN

IDN

IND

MUS

-.06

-.04

-.02

0.0

2C

ompo

nent

plu

s re

sidu

al

-1 -.5 0 .5 1underval

Partial association between an index of currency “undervaluation” and the contribution of structural change to growth

But policy can clearly help: labor market rigidity

Partial association between an index of labor market rigidity and the contribution of structural change to growth

HKG

USA

SGP DNK

NGA

UKM

MYS

COL

THA

JPN

KENMUSCHL

MWI

ARG

ZMB

GHA

ETH

PHLINDCHN

ZAFTUR

SWE

KOR

ITA

PER

CRI

IDN

MEX

NLD

BRAESP

FRA

SEN

VEN

BOL

-.06

-.04

-.02

0.0

2C

ompo

nent

plu

s re

sidu

al

0 .2 .4 .6 .8Employment rigidity index (0=less rigid, 1=more rigid)

Conclusions

The presence of a large convergence gap ensures significant potential for rapid economic growth in developing world, regardless of what happens in the rich countries

Fulfilling this potential requires ongoing process of diversification and structural change

This process is not automatic, especially in countries with an initial comparative advantage in primary products

It necessitates pragmatic, experimental policies that support new industries– along with an external environment that supports, rather than

misdirects, such efforts – implications for WTO rules, and for WB/IMF policy advice