Structural Transformation Policies
and Economic Growth
Dani Rodrik
November 2012
-.05
0
.05
.1
Latin America has been doing better
(sort of…)
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 trends by region, 1950-2011
Outline
 Structural change: why and how it matters
 Structural transformation and fundamentals as drivers of
economic growth
 Policies for structural transformation
 What can/does go wrong?
Why structural change matters
 Developing and emerging nations are characterized by
structural heterogeneity
 wide dispersion in labor productivity across activities
 Some industries are “escalator industries”
 higher productivity level and trajectory
 typically tradables and manufactures
 though increasingly tradable services exhibit similar features too
 Moving resources (especially labor) to these escalator
industries (i.e., structural change) is a key source of
economy-wide productivity growth
Escalator industries: unconditional
productivity convergence in manufacturing
Source: Rodrik (2012)
Notes: Vertical axis represents relative growth rate of labor productivity, controlling for period fixed effects. Each country enters with most recent
decade for which data are available.
Manufacturing as indicator of
structural change
Manufacturing employment share
0.3
COL
South Korea
THAI
0.25
KOR
0.2
Thailand
0.15
0.1
Colombia
0.05
0
1950
1955
1960
1965
1970
1975
1980
1985
1990
1995
2000
2005
How structural change contributes (or
not) to economic growth
Different patterns of structural change
 Thailand
 Hong Kong
 Colombia
 Argentina
Growth-increasing structural change:
Thailand
3
Correlation Between Sectoral Productivity and
Change in Employment Shares in Thailand (1990-2005)
 = 5.1686; t-stat = 1.27
pu
2
min
1
tsc
man
0
fire
cspsgs
w rt
-1
con
agr
-.2
-.1
0
Change in Employment Share
(Emp. Share)
Fitted values
*Note: Size of circle represents employment share in 1990
**Note:  denotes coeff. of independent variable in regression equation:
ln(p/P) =  + Emp. Share
Source: Authors' calculations with data from Timmer and de Vries (2009)
.1
Growth-increasing structural change:
Hong Kong
 = 2.8359; t-stat = 2.51
pu
1
2
Correlation Between Sectoral Productivity and
Change in Employment Shares in Hong Kong (1990-2005)
fire
0
tsc
min
w rt
cspsgs
man
-1
con
-2
agr
-.2
-.1
0
Change in Employment Share
(Emp. Share)
Fitted values
*Note: Size of circle represents employment share in 1990
**Note:  denotes coeff. of independent variable in regression equation:
ln(p/P) =  + Emp. Share
Source: Author's calculations with data from Timmer and de Vries (2009)
.1
Colombia: Growth-neutral structural
change
 = -2.0921; t-stat = -0.16
pu
2
3
Correlation Between Sectoral Productivity and
Change in Employment Shares in Colombia (1990-2005)
1
min
man
0
cspsgs
con
tsc
fire
agr
-1
w rt
-.1
-.05
0
Change in Employment Share
(Emp. Share)
Fitted values
*Note: Size of circle represents employment share in 1990
**Note:  denotes coeff. of independent variable in regression equation:
ln(p/P) =  + Emp. Share
Source: Author's calculations with data from Timmer and de Vries (2009)
.05
Growth-reducing structural change:
Argentina
2
Correlation Between Sectoral Productivity and
Change in Employment Shares in Argentina (1990-2005)
 = -7.0981; t-stat = -1.21
min
man
tsc
0
.5
1
1.5
pu
con
w rt
-.5
agr
fire
-.06
-.04
-.02
0
Change in Employment Share
(Emp. Share)
Fitted values
*Note: Size of circle represents employment share in 1990
**Note:  denotes coeff. of independent variable in regression equation:
ln(p/P) =  + Emp. Share
Source: Authors' calculations with data from Timmer and de Vries (2009)
.02
cspsgs
.04
The two dynamics of growth
1.
“Fundamentals”: building broad capabilities in the forms of human
capital and strong institutions

2.
Takes time, requires broad-based complementary investments, and
produces steady but moderate growth
“Structural transformation”: emergence and expansion of modern
industries

Requires narrower range of reforms (that are often sectoral) to
remove/compensate for costs modern industries face, and produces
rapid growth until dualism eliminated
The set of policies required to foster these two dynamics overlap, but are
not same
In particular, role of unconventional policies to stimulate new industries
(as in East Asia)
A typology of growth outcomes
Structural transformation
slow
rapid
low
(1) no growth
(2) episodic growth
high
(3) slow growth
(4) rapid, sustained growth
Investment in
fundamentals
A typology of growth outcomes
Structural transformation
slow
rapid
low
(1) Sub-Saharan Africa
(2) Many countries under
ISI, China?
high
(3) post-1990 Latin
America
(4) South Korea, Taiwan
Investment in
fundamentals
Structural transformation is impeded by both
market failures and government failures
 Government failures
 generic
 poor labor laws, inadequate property rights, lack of contract
enforcement, red tape, corruption, macro instability, high taxes, …
 sectoral/micro
 specific regulations and taxes; lack of specific public inputs
 Market failures
 Demonstration effects and learning spillovers from introduction of
new products or new technologies (“discovery”)
 Coordination and agglomeration externalities (e.g., clusters…)
“Orthodox” development policy
 Government failures
 generic
 poor labor laws, inadequate property rights, lack of contract
enforcement, red tape, corruption, macro instability, high taxes, …
 sectoral/micro
 specific regulations and taxes; lack of specific public inputs
 Market failures
 Demonstration effects and learning spillovers from introduction of
new products or new technologies (“discovery”)
 Coordination and agglomeration externalities (e.g., clusters…)
… unsatisfactory outcomes in Latin America since 1990
The new industrial policy
 Government failures
 generic
 poor labor laws, inadequate property rights, lack of contract
enforcement, red tape, corruption, macro instability, high taxes, …
 sectoral/micro
 specific regulations and taxes; lack of specific public inputs
 Market failures
 Demonstration effects and learning spillovers from introduction of
new products or new technologies (“discovery”)
 Coordination and agglomeration externalities (e.g., clusters…)
… is about removing bottlenecks to new economic activities,
whether due to market or government failures
The wrong questions
 Should we have IP?
 Case for IP no different than case for government policies in other
areas such as education, health, infrastructure, macroeconomic
stabilization
 Failures occur in all these areas too, but are not an argument for non-
intervention
 How to do it better, rather than whether to do it
 Most countries are engaged in IP, whether they say so or not
 Sector-specific incentives pervasive in Colombia
 Better to do it explicitly and self-consciously than surreptitiously
 But can governments pick winners?
 Of course not
 Good IP requires much less from government:
 Ability to learn, revise policies, and to “let losers go”
 Mistakes are a natural, expected feature of good IP
 Too few mistakes indicate too timid IP
The right question: how to do IP?
 Traditional IP:
 A list of sectoral priorities + sectoral incentives
 Top-down, relying on quality of bureaucracy (honesty,
competence, implementation)
 Presumes solutions are known
 Modern IP:
 A process of institutionalized collaboration and dialog
 Focused on identification of constraints and opportunities
 And the generation of pragmatic private-public solutions
 Continuous monitoring and evaluation
 Presumes only that solutions can be discovered
Institutional design for industrial policy
Must be built on three ideas, each of which leads to a different
“design principle”:
1. The requisite knowledge about the existence and location of
the spillovers, market failures, and constraints that block
structural change are diffused widely within society
=> “embeddedness”
2.
Businesses have strong incentives to “game” the government
=> carrots and sticks, discipline
3.
The intended beneficiary of IP is neither bureaucrats nor
business, but society at large
=> accountability
The many ways in which IP can fail
 No institutionalized dialog
 No clear targets
 Market failures? Protecting rents? Social policy?
 No clear instruments
 No clear commitments
 No monitoring
 No review/revisions
 Not enough coordination across government agencies
 No political leadership and ownership
 No transparency
Industrial policy as work in progress
 No country can get everything right
 East Asia, for example, typically falls short on accountability and
transparency
 But this is not an argument for not doing it
 Need to look at IP as a normal government function, that can be
performed better or worse
 IP is a “craft”, which improves with experience
What industrial policy cannot do
 Substitute for the lack of “fundamentals”
 human capital
 proper institutions
 Compensate for macro imbalances
 fiscal unsustainability
 large external deficits
 across-the-board uncompetitiveness due to high
prices/overvalued currency