Growth Miracles and Growth Debacles Exploring Root Causes

Growth Miracles and Growth Debacles
M2545 - BHATTACHARYYA PRINT.indd i
02/03/2011 10:09
To Anna, Sahana and Swapan
M2545 - BHATTACHARYYA PRINT.indd ii
02/03/2011 10:09
Growth Miracles and

Growth Debacles
Exploring Root Causes
Sambit Bhattacharyya
Research Fellow, Department of Economics, University of
Oxford, UK
Edward Elgar
Cheltenham, UK Northampton, MA, USA
M2545 - BHATTACHARYYA PRINT.indd iii
02/03/2011 10:09
Sambit Bhattacharyya 2011

All rights reserved. No part of this publication may be reproduced, stored in a
retrieval system or transmitted in any form or by any means, electronic,
mechanical or photocopying, recording, or otherwise without the prior
permission of the publisher.
Published by
Edward Elgar Publishing Limited
The Lypiatts
15 Lansdown Road
Cheltenham
Glos GL50 2JA
UK
Edward Elgar Publishing, Inc.
William Pratt House
9 Dewey Court
Northampton
Massachusetts 01060
USA
A catalogue record for this book

is available from the British Library
Library of Congress Control Number: 2010934015
ISBN 978 1 84844 631 1
03
Typeset by Servis Filmsetting Ltd, Stockport, Cheshire

Printed and bound by MPG Books Group, UK
M2545 - BHATTACHARYYA PRINT.indd iv
02/03/2011 10:09
Contents
Preface
Acknowledgements
1
Introduction
PART I
2
3
4
5
7
8
9
HISTORY AND ECONOMIC DEVELOPMENT
The Great Divergence: an account of the growth miracles and

growth debacles since AD 1000
Theories of root causes of economic progress
Empirical evidence
Root causes of economic progress: a unifying framework
PART II
vi
viii
9
15
48
95
PROMOTING GROWTH IN THE CURRENT

ENVIRONMENT: EVIDENCE AND POLICIES
Institutions and trade: competitors or complements in

economic development
Improving institutions with trade policy: myth or a possibility
Which institutions matter most for economic growth?
Making policy work: a road map for future growth
Data appendix
References
Index
119
140
147
166
179
184
199
M2545 - BHATTACHARYYA PRINT.indd v
02/03/2011 10:09
Preface
No new light has been thrown on the reason why poor countries are poor and
rich countries are rich.
Paul Samuelson (1976), Illogic of Neo-Marxian Doctrine, p.107.
Wide variations in living standards are observed across countries. Going

further back in time, we notice that this variation is largely attributable
to economic performance across countries since the start of the sixteenth
century and not just how they performed post-Second World War. This
book presents, in two parts, a detailed account of this process of divergence, the data involved and policies for the future. The first part opens
with theories of institutions, geography, human capital, trade, religion
and culture, and state formation and war explaining this divergence.
Following which, it discusses some empirical results and illustrates the difficulties in quantifying the relationships between root causes and economic
development. The novelty of the book is the unifying framework which
explains the process of development in Western Europe. This framework
is also compared with growth narratives in other countries and continents
namely Africa, China, India, the Americas, Russia and Australia. A narrative style is adopted throughout to create a bridge between the empirical
literature and history. The unifying framework is an attempt to merge
all seemingly disparate theories of economic progress. The main message
is that diseases and geography matter at an early stage of development.
Institutions, however, become much more important as the economy
develops. Geography and, in particular, disease epidemics are a crucial
explanator of lack of development in Africa. In contrast, in China and
India, the Malthusian population growth and disease cycle was broken
fairly early and institutional weaknesses played a crucial role in their
respective declines. In the Americas and Australia, colonial institutions
were a crucial factor. In Russia, it was crippling political institutions of
the nineteenth century and restrictive political and economic institutions
of the Soviet Union that did the damage.
The second part of the book focuses on growth promoting policies.
First,it documents some macro evidence on the role of policy in yielding
growth. It shows that trade can benefit nations in situations when institutions are adequately strong. Property rights and contracting institutions
vi
M2545 - BHATTACHARYYA PRINT.indd vi
02/03/2011 10:09
Preface
vii
are good for growth. Market stabilizing institutions are good for growth
and regulations are important only up to a certain extent. Second, it outlines growth promoting policies namely the first principles of growth.
They are property rights, contracts, regulatory institutions, rule of law,
macroeconomic stabilization, representative politics, human capital
investments, market access and international trade. Third, it discusses the
cases of India and China, two recent success stories. It shows how these
countries have preserved incentives for private investments even without
rigidly following the first principles. But more importantly, they have been
able to create institutions which are well grounded in local traditions and
culture and are also able to create appropriate incentives for investments.
Fourth, the book outlines steps that could be taken to facilitate growth in
situations of state failure, disease trap, poverty trap and scarcity of skilled
workers.
M2545 - BHATTACHARYYA PRINT.indd vii
02/03/2011 10:09
Acknowledgements
I gratefully acknowledge discussions with my colleagues and friends at different stages of development of this project. In particular, I would like to
thank Jeffrey Williamson and Tim Hatton for sparing their time to discuss
some aspects of this project. I would also like to acknowledge help and
encouragement from Prema-chandra Athukorala. I have benefited from
comments by and discussions with Akihito Asano, John Braithwaite,
Steve Dowrick, Gary Magee, Mark Rogers and Jonathan Temple. I also
gratefully acknowledge financial support from Jonathan and Jennifer
Oppenheimer. Last, but not least, I would like to thank my wife and family
for their love, affection and words of encouragement. The majority of the
manuscript was written when I was a research fellow at the Arndt-Corden
Division of Economics of the Australian National University.
viii
M2545 - BHATTACHARYYA PRINT.indd viii
02/03/2011 10:09
1.
Introduction
Hidaya Mohamed is a fifth grade student. She dreams of going to the

Benjamin Mkapa Secondary School, one of the better government schools
in her town, but no one in her family has received secondary education
before. Her older brother does not go to school and her sister is in a government primary school. Her mother, a single parent, earns a living selling
maandazi (traditional buns) which is not enough to finance Hidayas
dream. I dont think I can go there, reflects Hidaya. My mother cant
pay her contribution and I dont know who can help me.1
Hidayas family resides in Tanzania. The average income per capita in
Tanzania today is $1141, which is approximately one-sixty-third of what
it is for Luxembourg, the richest nation in the world.2 Tanzania is poor in
spite of its rich natural resource endowments of diamonds, gold, iron ore,
coal, natural gas and nickel. Tanzanian soil is suitable for coffee, cotton
and clove plantations. In spite of all this, the Tanzanian development
record is disappointing. A colonial history marred with slave trade reveals
a very sad tale of exploitation and underdevelopment. Even after independence in 1961, things did not change much. Almost all of the small and
big development initiatives of the government in independent Tanzania
culminated in disastrous failures. One of the better known accounts of disaster is that of the Morogoro Shoe Factory. The Morogoro Shoe Factory
was established in Tanzania with the help of the World Bank in the 1970s.
It was endowed with labour, machinery and the latest shoe-making technology. But it hardly produced any shoes, only utilizing 4 per cent of its
capacity, largely due to the absence of production incentives for the firm
(Easterly, 2001). The plant was not well designed either. It had aluminium
walls and no ventilation system, which was unsuitable for the Tanzanian
conditions. After two decades of struggle, production finally stopped in
1990.3 Tanzania now stands as one of the poorest countries in the world,
struggling with poverty, high infant mortality, HIV/AIDS and malaria.
By 2001, Tanzania had accumulated external debt worth $6.7 billion
which was cancelled under the Heavily Indebted Poor Countries (HIPC)
Initiative.4 The Tanzanian economy in its current state cannot generate enough wealth to ensure a decent living standard for someone like
Hidaya. Hidayas story is not unique. There are many more like Hidaya
M2545 - BHATTACHARYYA PRINT.indd 1
02/03/2011 10:09
Growth miracles and growth debacles
in the poor tropics and subtropics whose minimum aspirations cannot be

fulfilled because their respective economies fail to generate enough wealth
for them.
The obvious question that follows is the most common yet crucial in
the field of economic development. Why is there a 63-fold difference in
income per capita between Luxembourg and Tanzania? The question is
common because it has been addressed by social scientists on numerous
occasions over the last 50 years. However, it is important to note that this
debate is far from being over. It is perhaps fair to say that we are only
starting to understand the complex process of development. It is crucial
because it has the potential to improve living standards and reduce the
proportion of people who are suffering from starvation, poverty and
deprivation.
In the past, the main focus of the literature was to discover and evaluate
proximate causes (physical capital accumulation, technological progress
and so forth). In recent years, however, there has been a welcome shift in
the focus. Economists have stressed the importance of root causes (institutions, human capital, religion and culture, openness to trade and geography) in their quest for growth. Ever since Francois Quesnay wrote about
the giant economic machine in 1763, the development literature has always
been in search of the factor that propels the machine (see Banerjee, 2007,
p. 125). No wonder that phrases of the likes of Holy Grail of growth and
Elixir of growth have been used so many times in the literature. Perhaps
the idea of an all encompassing law governing society is too irresistible
to the researchers of the dismal science. The recent literature on the root
causes of economic progress is not an exception. A lot of effort has gone
into identifying the root cause. As a result the abovementioned explanations are often posed as competing explanations of economic development. Needless to say, this need not be the case. It will be surprising if the
process of development, as complex as it is, turns out to be a result of a
single factor. Perhaps the most likely outcome is a series of causally interlinked explanations.
I make an attempt to establish some of these linkages. In the first part
of this book, I provide empirical evidence on the role of root causes. I also
offer a unifying framework for Western Europe a development success
story. The framework links the seemingly competing explanations of longrun development. Then the framework is put to the test by comparing and
contrasting it with the historical process of development in the Americas,
Africa, China, India, Russia and Australia. It appears that the framework
does well in explaining the development process in these different continents. In the second part of the book, I focus on policies that could improve
growth. I provide contemporary macroeconomic evidence on policies that
02/03/2011 10:09
Introduction
may have worked better for growth. I then outline the first principles
(property rights, contracts, rule of law, regulation, macroeconomic stability, representative politics, access to markets, human capital investments
and trade) that are essential for growth. I also discuss the case of India
and China, two recent success stories, and how they have applied the
first principles. The book concludes with a discussion of growth policy in
situations of state failure, poverty trap and scarcity of skilled labour. All
along, I take into consideration the following issues. First, the literature is
rich with various explanations and theories of development. However, it is
difficult to find a single source which covers the literature well. I make an
attempt here to fill this vacuum. Hopefully, graduate students working on
this topic will find this useful. However, my attempt to cover the literature
may not be exhaustive. Nevertheless, I do make a conscious effort to cover
all of the major studies in history, economics and politics. Second, I make
an attempt to present the material in a non-technical fashion wherever I
can. However, some technical materials are presented in Chapters 3, 4, 6,
7 and 8. These models are followed by a non-technical summary of the
main message. Therefore, I am assuming that this is not an insurmountable hurdle for readers outside the discipline. Furthermore, readers may
also choose to skip the technical parts and focus on the summary instead.
In my view, this will have very little impact on their understanding of the
main message of the book. To put it simply: readers wont miss much by
skipping the technical details. Nevertheless, only time will tell how successful I am in making it within the reach of the elusive intelligent layperson.
Third, for the sake of brevity, I skipped technical details in some cases.
The analysis proceeds in two parts. The first part focuses on the role
of history and the second part is more concerned about contemporary
growth policy. Chapter 2 in Part I introduces the core empirical fact in the
literature the Great Divergence. Using simple diagrams from previously
published research, I take a brief look at the growth miracles and growth
debacles of the last millennium. I give special attention to the economic
performance of North America, South America, Africa, the United
Kingdom, Western Europe, Southern Europe, Russia, China, India and
Indonesia. I also revisit the reversal of fortune hypothesis proposed by
Daron Acemoglu, Simon Johnson and James Robinson.
Chapter 3 critically evaluates theory and evidence on the root causes of
economic progress. It begins by distinguishing between root and proximate causes. It also shows how the idea of root causes can be augmented
into a standard Solow growth model. Some of the prominent political
economy models and theories of institutions are reviewed in this chapter.
It is followed by a review of theory and evidence on the role of institutions, the role of religion and culture, the role of geography, the role of
02/03/2011 10:09
trade openness, the role of human capital and the role of state formation
and war.
Chapter 4 presents empirical evidence. It highlights some of the problems related to the empirical estimates in the root causes literature. It
shows that malaria is crucial in explaining long-run development in
Africa. It shows that factors such as institutions are statistically insignificant in an African sample.
Chapter 5 presents the unifying framework. The framework describes
the trajectory of capitalist development in Western Europe. It compares and
contrasts the Western European trajectory with the trajectories of Africa,
the Americas, China, India, Russia and Australia and tries to establish
what deep structural factors are behind the development or lack of development of these regions. My conjectures are backed by the existing empirical
evidence. Empirical evidence drawn from published research is presented
with limited technical details. However, the source research papers are cited
so that interested readers can follow it up if they please. In this chapter, I
also try to present a case that the African process of development is different from the rest of the world. I argue that high incidence of malaria has an
adverse impact on household savings which dampens growth.
Part II of the book focuses on contemporary policymaking. Chapter 6
presents evidence on the role of trade in promoting development over the
last two decades. It shows that trade is effective only when a country is
above a certain threshold level of institutional quality. Chapter 7 shows
that a countrys institutions could be improved through trade policy. It
presents theoretical explanation and empirical evidence in support of such
view.
Chapter 8 tackles a related issue of institutional effectiveness. It empirically tests which institutions were more effective in delivering growth
over the last two decades using international panel data. It finds that
property rights, contract and macroeconomic institutions are important.
Regulatory institutions are also important but overregulation is not good
for growth. Political institutions come out to be statistically insignificant.
Finally, Chapter 9 concludes with a list of policies essential for growth. I
call this list the first principles. I argue why they are important for growth.
I also discuss the Chinese and the Indian cases to illustrate how important
it is to apply the first principles so that they are best suited for local conditions. I outline a way forward in situations of state failure, poverty trap
and scarcity of skilled labour.
Without doubt, collectively the literature has made immense progress
towards understanding and identifying the root causes of economic
progress. However, a lot of work remains to be done to establish linkages
between factors. This book is an attempt to explore these linkages.
02/03/2011 10:09
Introduction
NOTES
1. Hidayas story is quoted from an article titled All these children want is a decent education by Sakina Zainul Datoo published in the October 2003 edition of the journal
Africawoman on page 3. The online source of this article is the following:
http://www.africawoman.net/newsdetails.php?NewsID5182&AuthorID553&CountryI
D512&NewsTypeID56&IssueID524
2. These are figures for 2007 expressed in constant 2005 international dollars adjusted
for purchasing power parity (PPP) differences. The source is the World Development
Indicators (WDI) Online, The World Bank.
3. The story of the Morogoro Shoe Factory is quoted from Easterly (2001, p.68).
4. The Heavily Indebted Poor Countries (HIPC) Initiative was established in 1996 as a
joint collaboration between the World Bank and the International Monetary Fund
(IMF). The aim of the initiative is to reduce the excessive debt burdens faced by the
worlds poorest nations. In 1999, the international community endorsed enhancements
to the HIPC initiative allowing more countries to qualify for HIPC assistance, accelerating and deepening the delivery of relief, and strengthening the link between debt relief
and poverty reduction.
02/03/2011 10:09
02/03/2011 10:09
PART I
History and economic development
02/03/2011 10:09
02/03/2011 10:09
2.
The Great Divergence: an account

of the growth miracles and growth
debacles since AD 1000
As I have illustrated in the introduction with the Tanzania and Luxembourg

example, living standards do vary enormously across countries. Lets take
stock of the situation by subjecting this empirical fact to close examination. In Figure 2.1, I plot the trend in per capita income in seven different
countries over the period 1950 to 2004 using data from the Penn World
Table version 6.2. What I notice is, indeed, a significant variation in standard of living across countries. China, India and Brazil seem to be part of a
completely different club when compared with the living standards in the
United States, Australia and the United Kingdom. The other important
point to note is that the difference across countries is persistent and there
is very little change in relative positions since 1950. Therefore, it is quite
11
USA
Australia
UK
Log GDP per capita
10
Brazil
China
India
Ghana
6
1950
1960
1970
1980
1990
2000
2010
Year
Figure 2.1
Evolution of per capita income in USA, Australia, UK, Brazil,

China, India and Ghana over the period 1950 to 2005
9
02/03/2011 10:09
10

USA
10
Japan
Log GDP per capita
China
8
Netherlands
UK
India
Africa
6
0
200
400
600
800
1000
1200
1400
1600
1800
2000
Year
Source: Authors plot using data from Maddison (2004).
Figure 2.2
Evolution of per capita income in USA, UK, the Netherlands,

Japan, China, India and Africa over the period AD 12003
evident that the origin of this divergence is certainly not the post-war
period of 1950 to 2004. To trace out the origin, we need to look further
back.
Looking further back in time, of course, is a challenge. Do we have
the data to do this successfully? The answer is yes, thanks to the excellent research done by economic historians over the last four decades. The
most widely used numbers are from Angus Maddisons (2004) The World
Economy: Historical Statistics. However, there is no consensus among
historians regarding the validity of these numbers. Some historians argue
that the values for China, Japan and other parts of Asia were comparable or even higher than those in Western Europe in the late nineteenth
century. Needless to say, Maddisons figures show the opposite. I shall
come back to some of these debates very briefly later. Nevertheless, this
debate is not central to my story and the majority of historians agree on
the broad patterns of evolution of income across countries since AD 1.
In Figure 2.2 we plot the evolution of per capita income since AD 1
in the United States, United Kingdom, the Netherlands, Japan, China,
India and Africa. What we notice is quite striking. Prior to 1400, living
02/03/2011 10:09
The Great Divergence
11
40
England
Netherlands
Urbanization rate
30
Italy
20
France
10
Russia
Japan
India
China
0
1000
1100
1200
1300
1400
1500
1600
1700
1800
1900
Year
Source:
Authors plot using data from Acemoglu et al. (2002).
Figure 2.3
Urbanization in the world powers over the period AD

10001850
standards across countries were largely comparable. Countries diverged

from each other from the sixteenth century onwards. This difference has
been persistent and widening, particularly over the last two centuries, especially since the Industrial Revolution in the United Kingdom. The only
exception is steady growth in India and China over the last two decades.
But this, without doubt, is not enough to match the standards of living in
the United States and United Kingdom any time soon for these countries.
Indeed, the developed countries are a long way away. Therefore, the gap
that we notice in Figure 2.1 is more of a continuation of the divergence
that started in the sixteenth century.
If we compare urbanization across major world powers over the period
AD 1000 to 1850, we also notice a similar trend (see Figure 2.3). The level
of urbanization across the world was very similar till the twelfth century
with the exception of Italy, where the city states were more urbanized
than the rest of the world. Figure 2.3 shows that the Netherlands experienced rapid urbanization during the thirteenth century and thereafter,
and caught up with Italy, then the most progressive and prosperous
part of Europe, during the early fifteenth century. This rapid growth in
urbanization was unmatched by any other parts of the world till the mideighteenth century when Britain experienced a similar trend. The British
02/03/2011 10:09
12

Western Europe
Urbanization rate
20
15
10
Asia
Eastern Europe
1300
1400
1500
1600
1700
1800
1900
Year
Source: Authors plot using data from Acemoglu et al. (2005b).
Figure 2.4
Urbanization rate in Western Europe, Eastern Europe and

Asia
growth in urbanization was largely propelled by the Industrial Revolution

and an expansion in Atlantic trade. In contrast, the French growth in
urbanization was relatively modest. This is perhaps explained by the slow
progress of industrial development in the European continent following
the Industrial Revolution in Britain. Landes (2003) documents why the
progress of industrial development and technological change in continental Europe was so slow. Urbanization in India, Russia, China and Japan
remained largely unchanged over this millennium again suggesting that
the genesis of divergence between the East and the West (or the North and
the South) is somewhere rooted around the fifteenth century.
This trend in urbanization rate persists even when we look at aggregate
data. Figure 2.4 presents urbanization rates weighted by population at
the continental level from Acemoglu et al. (2005b). The diagram shows
that there is very little change in urbanization rates across continents over
the period 1300 to 1500. It is only after 1500 that Western Europe surged
ahead leaving Asia and Eastern Europe way behind. Eastern Europe experienced some growth after 1700, but the gap with Western Europe in fact
widened. The urbanization rate in Asia declined after 1600 and became on
a par with Eastern Europe in 1850.
So much for urbanization. How about industrial production per capita?
02/03/2011 10:09
Industrial production per capita, UK in 1900 =100
The Great Divergence
13
400
US
300
200
Canada
Australia
New Zealand
100
Brazil
India
0
1750
1800
1850
Year
1900
1950
Source: Authors plot using data from Bairoch (1982).
Figure 2.5
Industrial production per capita relative to the UK in the New

World and emerging economies
Figure 2.5 compares industrial production per capita relative to Britain in

the New World economies and the emerging economies starting from the
mid-eighteenth century. We notice that industrial production in the United
States relative to Britain increased rapidly, starting in 1830. Canada
also experienced rapid growth, starting in 1880. Compared to these, the
growth in industrial production in the other two Anglo-Saxon colonies,
namely Australia and New Zealand, was fairly modest. They did experience growth, but only in the twentieth century. The growth in the United
States was rapid and persistent and, in 1953, the United States industrial
production per capita reached 3.5 times the level of Britain. In contrast,
industrial production in Brazil and India experienced very little growth
during this period.
Finally, to summarize, in this chapter I show that there is a significant
gap in living standards across countries. The genesis of this gap is not the
post-war period. In fact, looking at data on earlier periods reveals that
countries diverged from each other in terms of living standards around
the fifteenth century. Economic historians often describe this as the Great
Divergence. The question, however, is what caused this divergence. We
shall plunge into a pool of answers in the following chapters. But before we
do that, lets take a first pass on the type of explanations that will follow.
In an influential paper published in the Quarterly Journal of Economics,
02/03/2011 10:09
14
Acemoglu et al. (2002) argue that the Great Divergence can be explained
via institutional1 divergence. They show that the fortunes of the nonWestern countries reversed around 1500. This is because of institutional
reversal. The Western colonial model set up extractive institutions in the
resource rich periphery where they never intended to settle. The reverse is
true for settlement colonies where they wanted to settle. So, they set up
institutions that are favourable to capitalism. This led to an institutional
reversal which explains the Great Divergence.
Not everyone, however, agrees with this story. The human capital theorists, the social capital theorists, the free trade group and many others, all
have their own explanation. I shall provide a brief review of all of these
theories in the next section. I shall also explain how this can all be related
to the neoclassical growth model.
NOTE
1. Acemoglu et al. (2002) mainly focus on property rights and contracting institutions. For
a more detailed definition of institutions, see Chapter 3.
02/03/2011 10:09
3.
Theories of root causes of economic

progress
One of the major predictions of the neoclassical growth model is that

given the level of technology and income at a particular point in time,
and a common growth rate of technology, a country accumulating more
physical and human capital grows faster than a country accumulating
less of the same. Empirical evidence suggests that this is only a part of the
story. In a simple regression model with two factors of production, physical and human capital, and log initial income as an additional control, a
large fraction of the cross-national growth variance remains unexplained.
This leads to two types of reaction. First, a search for additional correlates of growth with the majority of the correlates having no or very little
connection with theoretical models of optimizing behaviour. Second,
a move towards more structured growth empirics with the primary
objective of teasing out causality. The second approach is relatively
recent and has gained a fair amount of acceptance of late. However,
the question of how the literature has moved in this direction remains.
To find an answer, we need to trace out the origin and the evolution of
the growth regression. This is exactly what I seek to do in the following
section.
3.1 THE NEOCLASSICAL GROWTH MODEL

Lets kick off with an illustration of the standard neoclassical growth
model and show how the growth regression emerges from it. An earlier
version of this model was independently developed by Robert Solow
(1956) and Trevor Swan (1956) which only focused on physical capital
accumulation. This version was developed by Mankiw et al. (1992) who
augment the original version with human capital to accommodate the
ideas of knowledge-based growth proposed and formalized by Paul
Romer in the 1980s. The model is as follows.
The model considers an economy at a particular point in time t with
the following constant returns to scale production technology producing
income Y:
15
02/03/2011 10:09
16
Yt 5 F (Kt, Ht, Et) 5 Kat H bt E 1t 2a 2b
(3.1)
where Kt and Ht are the stock of physical and human capital, respectively,
Et 5 AtLt is effective labour growing at the aggregate rate of exogenous
population growth n plus technological progress g.
Dividing both sides by Et and suppressing the time subscript t, one can
rewrite the production function in per capita terms as follows
y 5 k ahb
(3.1a)
where k and h are stocks of physical and human capital per efficiency
unit of labour, respectively. It is assumed that a fixed proportion sk and
sh of the per capita output y are invested in physical and human capital,
respectively. The evolution of the economy is determined by the following
equations of motion. Both physical and human capital depreciate at an
exogenous rate d.
#
k 5 sky 2 (n 1 g 1 d) k
(3.2)
#
h 5 sh y 2 (n 1 g 1 d) h
(3.3)
#
#
At steady state k 5 h 5 0 and the long-run steady-state equilibrium
values of per capita physical capital, human capital and output are given
by
k* 5 a
1/1 2a 2b
1/1 2a 2b
s1k 2bsbh
sak s1h 2a
, h* 5 c
, and
b
d
n1g1d
n1g1d
y* 5 c
1/1 2a 2b
sak sbh
d
a
1b
(n 1 g 1 d)
(3.4)
The unique steady-state value is ensured by the Cobb-Douglas technology, which guarantees that the Inada conditions are satisfied. By linearizing around the steady state using the first order Taylor expansion and
integrating, one can derive the following expression for country i:
yiT 5 g1 1 g2 ln yit 1 g3XiT
(3.5)
where yiT ; (1/T) ln [ yiT /yit ] and the vector XiT includes the investment
rates in physical and human capital, rate of depreciation of capital and the
rate of growth of population. It follows from the model that the lower the
level of initial output of a country, the higher will be its growth rate for a
given value of XiT. In other words, all other factors remaining unchanged,
02/03/2011 10:09
17
poor countries grow faster than the rich countries. Therefore, the sign of
g2 is negative. This is known as conditional convergence in the literature.
Therefore, the theory predicts that a part of the variation in crossnational income can be explained by the variation in investment rates in
physical and human capital. In other words, the central problem in economic development is to understand the process of factor accumulation.
This leaves us with an empirical question to answer. Is the prediction made
by the theoretical model supported by the data? This is what motivated
the expansion of the now voluminous empirical growth literature, which
I discuss next.
3.2 GROWTH, FACTOR ACCUMULATION AND

THE MEASURE OF OUR IGNORANCE
One can test the steady-state properties of the neoclassical growth model
by estimating Equation 3.5 using regression techniques. In order to do that
Equation 3.5 is rewritten as follows:
yiT 5 g1 1 g2 ln yit 1 g3XiT 1 eiT
(3.5a)
where eiT is the unobserved residual. Many empirical growth studies estimate this model using a variety of techniques. Mankiw et al. (1992), King
and Levine (1994), Caselli et al. (1996), Hall and Jones (1999) and Easterly
and Levine (2001) are among the most cited ones. The final conclusion,
however, is loud and clear. The residual rather than factor accumulation
accounts for most of the growth differences across nations.
This leads to two types of research strategy in the literature. The first
strategy is to increase the number of regressors to account for the omitted
factors which explains the variance. The inevitable fallout of this line of
research is a never ending list of regressors. This is useful if the regressors
are derived from models of optimizing behaviour indicating the possibility of a strong causal relationship. Unfortunately, this is not the case. The
majority of the regressors are chosen on an ad hoc basis. As a result one
can identify correlates, but not causal relationships. The list of correlates
includes physical capital investment, human capital accumulation, low
income inequality, being located further from the tropics, fewer tropical
diseases, low fertility, less government spending, trade policy openness,
political freedom, British legal origin, rule of law, foreign direct investment, foreign aid, ethno-linguistic fractionalization and so forth. The list
is growing over time and it is far from being exhaustive.
The second strategy is to look at the deep structural features of the
02/03/2011 10:09
18
economy. The idea is that the deep structural features influence proximate
causes (factor accumulation) over the long horizon and shape the longrun economic performance of the economy. To put it more formally, the
proximate causes are functions of deep determinants. Following is the
relationship in mathematical terms:
XiT ; F (ZiT)
(3.6)
where ZiT is a vector of deep determinants. Sachs and Warner (1997a) and
Bhattacharyya (2004) use this structure in their models. The deep structural determinants are trade, institutions, geography, religion and culture
and knowledge. The listing of knowledge and human capital as a deep
determinant, however, is debatable. Rodrik et al. (2004) and Acemoglu
et al. (2001) identify human capital as a proximate determinant. Glaeser
et al. (2004) in the empirical literature and Mokyr (1990) in the economic
history literature identify knowledge as a deep determinant. Given the
weight of evidence in some of the recent research, it is perhaps appropriate
to classify it as a deep determinant. Nevertheless, it is still open to scrutiny.
Therefore, one can rewrite (3.5a) as a reduced form regression:
yiT 5 a ln yit 1 gZiT 1 eiT
(3.7)
A standard way of estimating this equation is to regress Zit on exogenous

historical variables that may have influenced Zit in the past and use the predicted value of Zit from this regression to estimate (3.7). It may appear that
this two-step procedure is a handy way to circumvent endogeneity problems; however, it has its own problems. The two-step procedure applied to
levels regression suffers from identification problems and multicollinearity
issues. Bhattacharyya (2009a) illustrates this problem in a cross-section
model with institutions and human capital. Dollar and Kraay (2003) show
this with institutions and trade. I shall discuss this more in the following
chapters. Nevertheless, this allows researchers to estimate the direct effect
of deep structural determinants on current growth.
The advantage of the structured growth empiric is that it is closer to
theory and it gives us useful insights into the establishment of causality. This is perhaps one of the major reasons behind the majority of the
recently published research taking up this approach. However, further
work needs to be done to identify the chain of causation and to support it
with a general equilibrium model.
02/03/2011 10:09
19
3.3 ROOT AND PROXIMATE CAUSES

OF DEVELOPMENT: WHAT ARE THE
DIFFERENCES?
A major criticism of the root causes literature is that its distinction of root
and proximate causes is somewhat arbitrary. Furthermore, the justification for putting institutions and diseases into the root causes basket is also
unclear. Before we proceed any further with the theories of root causes, it
is perhaps useful to clarify these definitional aspects right now. A quick
review indicates that the literature has used at least three different ways
to justify this distinction. First, root causes are more fundamental than
proximate causes. In other words, they cause long-run economic progress
or decline by influencing the proximate causes (see Acemoglu et al., 2001;
and Rodrik et al., 2004). Second, root causes are durable, whereas proximate causes vary over time (Glaeser et al., 2004). Third, it is difficult to
influence root causes through direct policy intervention, whereas one can
do the same with proximate causes relatively easily (Glaeser et al., 2004).
One can certainly question the merits of making this distinction, the justification provided by the literature and the validity of labelling institutions
and diseases as root causes. Some of the major ones are as follows.
First, is the distinction between proximate and root causes meaningful
and the same in all contexts, especially when there is virtually nothing
except geography that is truly fixed?1 The answer to this is yes, as the distinction between proximate and root causes relies more on causal linkages
between factors than time invariance of a particular factor. The problem
with relying too heavily on time invariance is that no factor is truly fixed
when a big enough time frame is considered. In contrast, a time frame
looking at the chain of causation may provide some useful information
about which factor is more fundamental. For the purpose of this book,
I tend to take the view that institutions and diseases are the root causes,
as they create incentives or disincentives for investments in physical and
human capital, which promote development. Although one should not
write off the possibility of causation running in the opposite direction,
recent evidence shows that the possibility of such occurrence is negligible.2
Second, are the root causes more durable than proximate causes as
it is often the case that they change rapidly due to revolution or technological breakthrough in medicine? Data from the Polity IV project
shows that economic and political institutions are indeed more durable
compared to investments in physical or human or financial capital. Some
empirical research, however, is guilty of using expropriation risk from
the Political Risk Services, a private company which assesses the risk of
expropriation of foreign investment in different countries, as a measure
02/03/2011 10:09
20
of institutions. This measure is not durable as it assesses institutional

outcome rather than institutions themselves, and it is also a measure based
on perception (Glaeser et al., 2004). But, overall, there is evidence both
in pre-independence historical data (Acemoglu et al., 2005b) and postindependence data from Polity IV that institutions persist. The changes
that we observe are due to exogenous shocks consistent with the critical
juncture view. This, however, may not be the case with diseases. But it
is perhaps fair, at least for the purpose of this book, to classify diseases
as a root cause when we are looking at the whole stretch of development
history, including the period of primitive hunter-gatherer societies. After
all, back then, diseases were very much a result of geography, rather than
lack of intervention of medical science.
Third, there is evidence that policy influences institutional quality and
disease environment over time. Therefore, what is the use of labelling
institutions and diseases as root causes when it can also be influenced by
good policy like many other proximate causes? This is perhaps the most
difficult case to argue against. Recent research on policy and institutional
development show that good policy does influence institutions in the short
run (Thelen, 2000). However, the benefits of good policy never get translated into long-run economic development unless they are locked into the
institutional structure. Examples of short-lived policy-driven growth with
poor institutions are not that difficult to find. Argentina in 1870s to 1920s,
Czarist Russia in the decade leading up to World War I, Colombia during
1900 to 1940s, Cte dIvoire in the first two decades after independence
all experienced short-lived growth when the benefits of good policies were
not locked into institutions (Robinson, 2006). Therefore, what is critical
for institutional change is to lock in either the benefits of good policy or
the positives of an exogenous shock at a critical historical juncture. For my
purpose, however, it still makes sense to label institutions as a root cause,
since policies are short lived compared to institutions in the time frame that
I am looking at. I can present the same argument in the case of diseases,
since I start from a period when the Malthusian cycle was operational and
there is no recorded evidence of public health intervention that far back in
time.3 Nevertheless, we shall come back to some of these details in Chapter
4, when I present the unifying framework to explain development history.
3.4 THEORIES OF DEEP STRUCTURAL

DETERMINANTS
In this section, I review the theories and empirics of deep structural
determinants of economic progress. This provides a background for the
02/03/2011 10:09
21
unifying structure to be presented in Chapter 4. Theories of state, institutions, religion and culture, geography, trade and human capital and
knowledge are reviewed. I include human capital and knowledge in this
list in order to review the theories that strongly argue knowledge is a deep
determinant of long-term economic development.
3.5 THE INSTITUTIONS THEORY

Little else is required to carry a state to the highest degree of opulence from the
lowest barbarism, but peace, easy taxes and tolerable administration of justice;
all the rest being brought about by the natural course of things Adam Smith
(cited in Jones, 1981, p. 253)
This is what Adam Smith had to say about the role of institutions in economic development. Even though many scholars have referred to the role
of institutions time and again in their work, institution theory perhaps is
most strongly associated with Nobel laureate Douglass North. According
to North (1994), institutions are humanly devised constraints that structure human interaction. He adds that they are made up of formal and
informal constraints and they depend on the enforcement characteristic of
these constraints. All these taken together form the incentive structure of
the society. North (1989) argues that the long-run economic performance
of a society is shaped by its incentive structure. In this study, he looks at
the effects of the United States Constitution, the British common law and
the centralized enforcement mechanism of Spain on economic performance and argues that institutional differences do have economic consequences. The recent empirical literature on long-run growth endorses this
view (Hall and Jones, 1999; Acemoglu et al., 2001; Rodrik et al., 2004).
This perhaps explains why in the recent past the institutional view has
received so much attention in the growth literature. Empirical studies on
growth use the instrumental variable approach in order to estimate the
relationship. This is essential in order to take account of the possibility of
reverse causality. In other words, as better institutions influence growth
positively, it is also possible that faster growing economies can afford to
develop better institutions.
Knack and Keefer (1995) and Hall and Jones (1999) are among the
first to use institutions as one of the explanators of economic progress.
They argue that Western Europeans are historically associated with high
quality institutions characterized by protected property rights and efficient
enforcement of contracts. Western Europeans also migrated and settled
in large numbers in similar temperate climates. When they migrated, they
02/03/2011 10:09
22
carried these values and institutions with them. Therefore, it can be argued
that the more temperate the climate the better is the institutional quality.
This allows Hall and Jones (1999) to use distance from the equator as an
instrument for institutional quality. They also use the fraction of the population speaking English and the fraction of population speaking other
Western European languages as alternative instruments for institutions.
A high proportion of the population speaking English and other Western
European languages is indicative of a significant presence of Western
Europeans in that country and hence better institutional quality. Their
study reports a strong positive effect of institutions on income per capita.
Acemoglu et al. (2001) put forward a strong argument in favour of the
proposition that institutional quality is the fundamental determinant of
economic development. They use European settler mortality as an instrument for institutional quality. They argue that Europeans resorted to different styles of colonization depending on the feasibility of settlement. In a
tropical environment, the settlers had to deal with the killers malaria and
yellow fever and, hence, a high mortality rate. This prevented colonizers
from settling in a tropical environment and resource extraction became
the most important, if not the only, activity in tropical colonies. In order
to support these activities, the colonizers in the tropics and the subtropics
erected institutions which were extractive in nature. On the other hand,
in temperate conditions European settlers felt more at home and decided
to settle. In these places, they erected institutions characterized by strong
protection of property rights and efficient enforcement of contracts. These
institutions created by the colonizers have persisted in the colonies even
after independence. Therefore, they look at the extent of European settler
mortality in the past in order to use it as an instrument of the current
quality of institutions. In a subsequent paper, Acemoglu et al. (2002)
argue that the settlement decision of European settlers also depended on
the rate of urbanization prior to 1500, the start of the period of colonization, and hence it can be used as an alternative instrument for institutions.
According to this argument, a high density of population in the colonies
ruled out any possibility of settlement for the settlers and they used the
natives to erect extractive institutions without bothering about the overall
protection of private property in the society. On the other hand, a low rate
of urbanization and low population density in the colonies encouraged the
settlers to settle in large numbers and build institutions to protect property rights. They use the urbanization rate in 1500, measured by the then
density of population, as an instrument for institutions in this study. In
both cases they report strong effects of institutions on development.
Rodrik et al. (2004) also report strong effects of institutions on economic development. In a cross-country regression with the level of income
02/03/2011 10:09
23
per capita as the dependent variable, they control for institutions, trade
openness and geography, and institutions comes out to be the only variable having a statistically significant relationship with income. In a similar
exercise, Easterly and Levine (2003) report no direct effect of geography
and policy on economic performance if institutional quality is present as a
control. They find that the only channel through which geography affects
income is institutional quality. They use geographic endowment measures
as their geography variable.
Many more, including the abovementioned empirical studies, identify
institutions as one of the important correlates of growth. This is perhaps
convincing enough to say that institutions matter. However, the question
that still remains is how institutions affect economic performance. I now
discuss three important papers that deal with this issue.
One of the early works on this issue is by Engerman and Sokoloff
(2001). In their study, they trace the origins of the contrasting growth
experiences of the North and South Americas. They argue that the settlement colonies of the North and South Americas differed significantly in
terms of factor endowments creating incongruous initial conditions. The
initial conditions have influenced the growth path of these two continents
that ensued via economic and political institutions. The incongruity in the
initial conditions thus explains the variability in their growth performance.
Engerman and Sokoloff (2001) write,
a hemispheric perspective across the range of European colonies in the New
World indicates that although there were many influences, the factor endowment or initial conditions more generally had profound and enduring impacts
on the evolution of economic institutions, on the structure of the colonial economies, and ultimately on their long-run paths of institutional and economic
development. While all began with an abundance of land and other resources
relative to labour, at least after the initial depopulation, other aspects of their
factor endowments varied contributing to extreme differences in the distributions of land holdings, wealth, and of political power.
They argue that the climatic conditions of the North were favourable
to mixed farming of grains and livestock which exhibited limited economies of scale in production. This encouraged the development of small
family-size farms and a relatively homogeneous population in terms
of the distribution of wealth and political power. The result was better
institutions favouring broader access to economic opportunities, more
extensive domestic markets and better overall growth. The South and
the Caribbean, in contrast, were endowed with vast mining resources
and climates and soil conditions conducive to commercial crops such as
sugar, tobacco and cotton. To exploit these resources, which exhibit large
02/03/2011 10:09
24
economies of scale, it was efficient on the part of the colonizers to set up

large mining firms and plantations run with cheap slave labour. This led
to the establishment of a society where a large section of the population
remained impoverished with no political or economic rights, whereas
a small section, the political and economic elite, took control of all the
wealth. The high level of inequality in the distribution of wealth and
political power created institutions which are extractive, exploitative and
extremely unjust. These institutions persisted over time causing persistent
inequality and lack of growth. The highly concentrated landholdings and
extreme inequality in Mexico, Colombia and Peru during the colonial
period are supportive of their story.
In a recent paper, Acemoglu et al. (2005a) formalize some of the ideas
of Engerman and Sokoloff (2001) using a schematic model. The following
is a discussion of that model. Acemoglu et al. (2005a) argue that a nations
economic and political institutions are broadly endogenous. They are, at
least in part, determined by the society or a segment of it. They distinguish
between two components of political power de jure and de facto. They
define de jure political power as the power that originates from formal
political institutions in the society, for example, Monarchy, Democracy,
Autocracy or Dictatorship and so forth. De facto political power, on the
other hand, originates from informal political institutions which often
challenge the formal ones, for example, economic and social interest
groups and so forth. They assert that the existing political institutions
at the current period determine the distribution of de jure and de facto
political power in the society. The strength of de facto political power also
depends upon the distribution of resources in that period. This shapes
economic institutions of that period and political institutions of the future.
Current economic institutions influence current economic performance
and the future distribution of resources. Future distribution of resources
and future political institutions in turn influence subsequent distribution
of political power, economic institutions and economic performance. The
following schematic diagram summarizes the link.
PIt 1 dj PPt & df PPt

DRt 1 df PPt
} {
1
EIt 1 Yt & DRt11

PIt11
(3.8)
where PI, DR, djPP, dfPP, EI and Y are political institutions, distribution
of resources, de jure political power, de facto political power, economic
institutions and economic performance respectively.
In another related paper Acemoglu et al. (2005b) explore the origin
of capitalist institutions in early modern Europe which can be cited as
02/03/2011 10:09
25
further evidence for the hypothesis that economic and political institutions
influence economic performance and vice versa. They describe the interactive role of trade and institutional development in stimulating economic
growth in certain Western European countries from the sixteenth century
onwards. They assert that Atlantic trade and colonial activity enriched
and strengthened commercial interest groups both economically and
politically in these countries (most notably Britain and the Netherlands),
which then demanded and obtained institutional reforms to protect their
property rights, enabling them to trade and invest more, triggering a
circular and cumulative pattern of economic growth. However, this was
not the case with some of the early starters in the colonization process
(notably Portugal and Spain). Even though they managed to expand their
trading and ransom-seeking activities rapidly by exploiting their early
mover advantage, their growth was not sustainable. A major part of the
explanation is the traders failing to secure their rights from the monarch.
The monarch and the clergy in Portugal and Spain remained firmly in
control of the activities of the trading companies, never allowing them to
accumulate de facto political power to challenge the existing institutions.
As a result, the incentive to undertake investments in physical and human
capital eroded fast and the growth was short lived.
Bhattacharyya et al. (2009) test this hypothesis for the period 1980 to
2004 using international data. They find that there is a threshold level
of institutional quality below which trade is not effective for economic
development. However, countries above this institutional threshold are
able to reap benefits of trade to the fullest extent. They present estimates
of the threshold level of institutional quality (property rights institutions,
in particular) in their paper.
A further extension of the political power theory is presented by
Acemoglu and Robinson (2000). Their argument is akin to the Marxist
theory of the capitalists and the workers class struggle. Marx argued
that capital accumulation and the associated decline of the surplus of
capitalists would intensify exploitation of the workers which would lead
to intense class struggle between the capitalists and the workers eventually
terminating the class society. Acemoglu and Robinson (2000) accept the
basic premise of this theory and argue that social and political reforms in
Western Europe during the nineteenth century were an outcome of deliberate concessions made by the elite, designed to avert political instability,
expropriation and possibly a revolution. As a result of these concessions
inequality reduced in Western society, creating the necessary precondition for the development of property rights institutions protecting private
property, which translated into faster economic growth in the periods that
followed. However, this theory is not free from its critics. Galor and Moav
02/03/2011 10:09
26
(2006) come up with an alternative model to explain the termination of the

class society. They argue that it was not class struggle but human capital
and knowledge that brought about the demise. I discuss the findings of this
study in detail in Section 3.9.
3.5.1
How About Institutions in Africa?
A great deal of research in recent years suggests that institutions are the
key for economic development. Institutions shape the incentive structure
for factor accumulation and technology development which is critically
important for a countrys economic progress. Therefore, success or failure
in development can be traced back to success or failure in institution building. I have reviewed the relevant literature that deals with this view above.
This view, however, is also relevant for Africa (Easterly and Levine, 1997;
Herbst, 2000; Acemoglu et al., 2001; Acemoglu et al., 2002). The institutions school argues that the weak institutions in Africa largely explain
her state of underdevelopment. Easterly and Levine (1997) use several
indicators of public policy which can also be categorized as institutional
measures even though they were not explicit about this in their study.
They argue that the lack of growth in Sub-saharan Africa is associated
with poor public policy, such as low schooling, political instability, underdeveloped financial systems, high government deficits and insufficient
infrastructure. They link the choice of poor public policies on the part of
the African governments to Africas high ethnic fragmentation. In other
words, they argue that high ethnic fragmentation increases polarization
and the degree of social conflict in terms of policy choice in African societies. This increases the cost of conflict resolution. The government often
fails to internalize these costs and selects socially suboptimal policies. This
is perhaps the first systematic attempt to explore African institutional
weaknesses and the reason behind it in our profession. Even though
Easterly and Levine (1997) were the first to apply this idea in Africa, it
is older than this. Alesina and Tabellini (1989) and Alesina and Drazen
(1991) theoretically modelled it almost a decade before their article was
published.
Herbst (2000) also seeks an explanation for institutional weakness in
Africa. Robinson (2002) provides an excellent review of Herbsts book.
Herbsts work predominantly focuses on the following questions: why are
states are so ineffective in Africa? Why are African states often plagued
by chaos and lawlessness? He comes up with an interesting answer.4 He
argues that in Africa, labour was scarce and not land, which is reflected
by the low population density figures across the continent. Therefore, the
pre-colonial states in the continent did not fight over land but over people.
02/03/2011 10:09
27
This explains why property rights over people were so well defined in the
form of slavery, whereas property rights over land were not, as many of
the lands were held communally. This implies that it was not necessary
for the pre-colonial states to defend a well-defined territory, as there was
no fierce competition over land. In other words, there was less need to
invest in defence and maintain bureaucracies. Therefore, tax collection
in these states was poor and often absent. This allowed African states to
survive without having to engage in institution building (tax collection,
investing in defence, maintaining bureaucracies and provide rule of law)
which made them very fragile. The trend of almost no external threat
coupled with low population density persisted during the colonial period.
Therefore the colonizers had little incentive to develop institutions either.
After independence, the situation did not change. The United Nations
decision to enforce the colonial boundaries as national boundaries and
Cold War politics reinforced this trend. Hence, what we observe now
are the weak institutions of contemporary Africa. Herbsts argument is
an extension of the work done by Max Weber on the development of
European states.
Even though Herbsts thesis has received kudos for its intuitive appeal,
it is still not free from criticism. Robinson (2002) criticizes Herbst for not
emphasizing the impact of colonialism on institutions and state formation
in Africa. He also adds that the relationship between population density
and institution building becomes unclear, particularly in the case of Africa
when one takes into account the reversal of fortune effect and the impact
of colonialism (Acemoglu et al., 2001; Acemoglu et al., 2002). The following paragraph deals with this theory in more detail.
Acemoglu et al. (2002) observe that the relatively prosperous and
densely populated areas of the world outside Europe prior to 1500 are
relatively poorer now, whereas the less prosperous and sparsely populated
regions outside Europe before 1500 are relatively prosperous at present.
This empirical fact is termed by Acemoglu et al. (2002) as the reversal
of fortune. They argue that this is because of the nature of the colonial
policy pursued by the European colonizers. In colonies with high population density prior to 1500, the European colonizers faced relatively more
resistance from the natives, which ruled out the possibility of settlement
and, hence, they erected extractive institutions without bothering about
the overall protection of private property. On the other hand, in colonies
with low population density prior to 1500 the European colonizers faced
less resistance from the natives and decided to settle and erect institutions
with better protection of private property. This argument may not be
true in the case of Africa. The African continent has always been a place
with low population density. In spite of the low population density, the
02/03/2011 10:09
28
colonizers decided not to settle in Africa and built institutions of plunder

instead.5 The answer to this apparent paradox perhaps lies in the argument of Acemoglu et al. (2001). Acemoglu et al. (2001) argue that one
of the key factors that influenced the decision of not settling in Africa by
the European colonizers was the continents disease environment, which
is well reflected by the settler mortality figures from the colonial records.
They use settler mortality data mostly constructed from the historian
Philip Curtins work on the disease environment of the European colonies to establish a statistically significant inverse relationship between the
disease environment and institutional quality. As Robinson (2002) puts it,
these factors are not taken into account in Herbsts story.
3.5.2
The Effects of the Institutions of Slavery and Slave Trade in Africa
The theories of the impact of slave trade on Africas current development

can be classified into two broad categories: the depopulation view and the
sociopolitical view.
The strong proponents of the depopulation view are the historians
Joseph Inikori and Patrick Manning. Gemery and Hogendorn (1979) also
add an alternative dimension to the view, basing their argument on trade
theory. Inikori (1992) argues that the slave trade out of Africa during the
period 1450 to 1870 resulted in massive depopulation of the continent.
The population throughout this period remained too low to trigger division of labour, growth in internal trade, specialization and diversification,
transformation of the technology and organization of production in the
manufacturing sector. As a result, manufacturing in pre-colonial Africa
could not develop beyond the handicraft stage. In the agricultural sector,
the effects were more immediate. The low ratio of population to cultivable
land encouraged dispersed settlements throughout the continent, particularly in Sub-saharan Africa. The population moved towards an extensive
rather than intensive form of agriculture, which made subsistence and
local self-sufficiency predominant. This had a dampening effect on technology adoption and production organization in the long run. African
agriculture largely remained primitive and undercommercialized during
this period.
Fage (1978) and Lovejoy (1982) strongly disagree with Inikoris view.
They argue that Inikoris claim of an absolute decline in the African population due to the slave trade is unfounded. Fage (1978) also adds that the
trade had a minimal impact demographically. However, Manning (1981,
1982) and Thornton (1980) show that the demographic effects are significant no matter what the absolute totals are.
Manning (1981, 1982) adds another dimension to the depopulation view.
02/03/2011 10:09
29
He shows that the African continent, faced with an increasing demand for
slaves from the New World, reacted by increasing the supply of slaves.
This increase in availability of slaves also raised the African demand for
slaves. Africans preferred female slaves, whereas young male slaves were
exported across the Atlantic to work in the plantations. This engendered
huge imbalance in African sex ratios, which had a long-term impact on
the continents demographic structure. According to his estimates, the
continents population was held in check during the eighteenth and the
nineteenth century by the slave trade, restricting economic progress.
Finally, Gemery and Hogendorn (1979) add that the mass removal of
the working-age population from the continent caused an implosion of
the African production possibility frontier as the lost labour input affected
virtually all production choices. This resulted in unambiguous reduction
in the welfare of the continent. The secular decline in welfare continued
over more than two centuries, plunging the continent into economic
backwardness.
The major contributors to the sociopolitical view are Inikori (1977,
1992) Manning (1981, 1982) and Miller (1988).
Inikori (1992) argues that a vast majority of slaves that were exported
were free individuals captured by force. The capture took a number of
forms, notably kidnapping, raids organized by the state, warfare, pawning,
via the judicial procedures, tributes and so forth. Firearms were imported
from the Europeans in exchange for slaves, particularly during the period
1750 to 1807, which were used for capturing more slaves (Inikori, 1977).
Rodney (1966) and Meillassoux (1976) show that an increase in the
Atlantic slave trade led to more being captured and expansion of the
African slave trade, often by violence. Lovejoy (1994) shows that warfare,
raiding and kidnapping were the means of enslavement for more than
three-quarters of the slaves captured during 1805 to 1850 from central
Sudan. The judicial process became a tool for enslaving people within the
community (Klein, 2001). Klein (2001) observes that judicial penalties in
the form of compensation, exile or beatings were converted to enslavement. Inikori (2000) argues that the increase in the trade for captives institutionalized banditry and corruption for more than three hundred years in
the continent, which retarded socioeconomic development.
Curtin (1975) partially disapproves of Inikoris view. He argues that it is
improper to view conflicts that arose from political causes in Africa in the
same light as those that originated from purely economic motives, such as
the slave trade. Inikori (1992), however, questions the usefulness of this
dichotomy as these relationships were far more complex. According to
his view, attempts to establish any form of simplistic relationship are
therefore misleading (Inikori, 1992, p. 26).
02/03/2011 10:09
30
Mannings work focuses on Dahomey, which is roughly the area around

the Bight of Benin, during the period 1640 to 1860 (Manning, 1981,
1982). He observes that the immense increase in slave prices with a price
elasticity of supply of 1.5 created the incentive for capturing more slaves.
Institutions were set up during 1640 to 1670 in this area, which were conducive to capturing slaves. The institutions included warfare, raiding, kidnapping, judicial procedures and tributes. Manning (1982) writes, These
structures further reinforced a willingness to tolerate or justify the enslavement of ones enemies or even ones own (p. 9). In the case of Dahomey,
the state became an active participant in the collection and delivery of
slaves, which is evidence favouring Inikoris argument.
Miller (1988) gives an account of the deadly nature of the process
of capture in Angola. He writes that warfare and violence stimulated
the capture of slaves in West Central Africa, often capturing slaves in
exhausted, shaken or physically wounded condition. Populations were
raided consistently by stronger neighbours, and harassed and driven out
from their homes and land. Gemery and Hogendorn (1979) add that the
culture of raiding and warfare created a distinct minority class in African
society who became far more powerful than the rest of the population,
both in economic as well as political terms. The slave traders of Africa
started enjoying European goods, currency and guns in return for slaves.
They made significant gains from trade, at least in the short run, even
though everyone was unambiguously worse off in the long run due to the
implosion of the production possibility frontier (Gemery and Hogendorn,
1979). The extreme inequality of wealth distribution tilted the existing
institutions in favour of the slave traders, which created the foundation
for further inequality in the future. This is consistent with the theory of
institutions, inequality and growth proposed by Engerman and Sokoloff
(1994) and Acemoglu et al. (2005a) with respect to the New World.
These institutions persisted throughout the pre-colonial period, further
intensifying the problem. After colonization, the colonial powers did not
interfere with the existing extractive institutions of the natives. They were
reinforced by them instead (Robinson, 2002). Persistence of weak institutions had a negative impact on the long-run economic development of the
continent.
Nunn (2008) also provides a similar account. He shows that the slave
trade prevented state development, encouraged ethnic fractionalization
and weakened legal institutions. Through these channels, the slave trade
continues to affect current economic development in Africa. However,
Bhattacharyya (2009b) shows that Nunns evidence is not robust to the
inclusion of malaria and other geographic variables.
02/03/2011 10:09
31
3.6 THE RELIGION AND CULTURE THEORY

Respect the altar of every belief. Spanish America, limited to Catholicism to
the exclusion of any other religion, resembles a solitary and silent convent of
nuns . . . To exclude different religions in South America is to exclude the
English, the Germans, the Swiss, the North Americans, which is to say the
very people this continent most needs. To bring them without their religion is
to bring them without the agent that makes them what they are. Juan Bautista
Alberdi (cited in Landes, 2000, p. 5)
Juan Bautista Alberdi, a distinguished Argentinian of the nineteenth

century, believed that religion and culture play the most significant role in
shaping the economic destiny of a nation. This is what he wrote in 1852,
half a century before Max Webers remarkable work The Protestant Ethic
and the Spirit of Capitalism.
In this section, I discuss some of the notable theories that link religion,
culture and economic performance. I classify the religion and culture
theory into five broad categories, as follows: the Protestant work ethic
view, the religious intolerance view, the collective belief view, the middle
class virtues view and the trust view.
The Protestant work ethic view is due to Max Weber and it goes back
to the early twentieth century when he published his thesis entitled The
Protestant Ethic and the Spirit of Capitalism. Weber (1930) argues that
the epoch-making Protestant Reformation changed societal outlook and
sowed the seeds of modern capitalism. Protestantisms emphasis on industry, thrift and frugality, along with its moral approval of risk taking and
financial self-development, created a social environment conducive to
investment and the accumulation of private capital. This became an inspirational force in transforming traditional societies into modern capitalist
ones. Catholicism, on the other hand, considered trade and accumulating riches to be sinful, an attitude which was quite inimical to economic
progress. Samuelsson (1993) points out that Webers theory hardly applies
to Lutheranism, which retained the traditional attitudes of Catholicism
towards trade and commerce at least at the time of its inception. But it
undoubtedly applies to Calvinism, which supports trade and commerce.
One of the central themes of Calvinism is predestination. According
to this belief, some individuals are chosen and saved by God while the
others are not. This process of selection by God is predestined and unalterable. Calvin, in his discourses, did not mention who had been chosen.
Therefore, the only practical way for the followers of Calvin to breed hope
of attaining salvation is by performing intense worldly activities. Weber
(1930, p. 69), also cited in Acemoglu et al. (2005a, p. 15), states that,
however useless good works might be as a means of attaining salvation,
02/03/2011 10:09
32
they are indispensable as a sign of election. They are the technical means,
not of purchasing salvation, but of getting rid of the fear of damnation.
Calvins doctrine encourages economic activity, but condemns enjoying
the fruits of such activity by means of leisure. The followers of Calvin consider morality to be a virtue not because it can bring salvation but because
it helps in acquiring moral credit. All these characteristics of Calvinism are
essential preconditions for economic progress and this is how, according
to Weber (1930), Calvinism influences economic performance. Given the
provocative nature of Webers thesis, it is not hard to understand why it
fails to achieve universal acceptance. In an influential study on religion
and growth, Tawney (1926) rejects Webers view. He argues that the
English economy took off in the sixteenth century when the religious influence in society was replaced by secular attitudes.
The religious intolerance view is due to Landes (1998). This adds another
dimension to the theory that links religion and economic performance. In
his book The Wealth and Poverty of Nations, he argues that the profit
motive of the entrepreneurial class coupled with invention and productivity
gain are core to achieving material progress. Therefore, it is crucial for religious beliefs and practices within the society to encourage hard work and
not inhibit rational and scientific thought and profit making. He uses the
example of eighteenth century Britain to illustrate the effect of Protestant
work ethics and values on economic performance. According to his thesis,
Britains numerous scientific successes after the Industrial Revolution
were made possible due to the supportive and tolerant role played by the
Church. In contrast, the orthodoxy of the Catholic Church in Portugal and
the culture of intolerance diffused by the Catholic Church in Spain and
Italy strangulated scientific thinking and halted progress even though they
were the first to make significant progress in cartography and ocean navigation. The infamous Roman Inquisition in sixteenth century Europe
is a testimony to the religious intolerance and orthodoxy of the Catholic
Church which virtually stalled the European renaissance. To summarize,
Landes (1998) maintains that the culture of intolerance in Catholicism and
Islam limits the potential of a country to grow. Protestantism, on the other
hand, facilitates economic growth by encouraging hard work and showing
a more tolerant attitude towards scientific research.
The collective belief view is due to Avner Grief. Grief (1994) holds that
different cultures and religions generate different sets of beliefs about how
people behave and peoples behaviour has a direct role in shaping economic performance. According to his argument, this can generate different
growth equilibria across the globe even with the same set of institutions.
These beliefs can also influence institutional quality in the long run and
affect economic performance indirectly.
02/03/2011 10:09
33
The middle class virtues view is due to Gregory Clark (2007). Clarks
main proposition is that the lack of growth in the preindustrial world is
due to the lack of middle class virtues of thrift and hard work. He argues
that industrial revolution took place in England because the middle class
virtues spread down the social scale in the sixteenth and seventeenth centuries. The poor experienced relatively high death rates during this time.
Therefore, they did not have enough surviving children. In contrast, the
infant survival rate amongst the rich was higher. As a result, many of
the rich offspring were forced down the social scale carrying middle class
virtues with them. This led to a change in work ethic of the entire society
and rapid economic progress. This, however, did not take place in Asia,
Africa and Latin America. The infant survival rates amongst the rich and
the poor were largely similar in these societies. Therefore, the spread of
middle class virtues was at best slack. This, in Clarks view, explains the
difference between the rich and the poor countries. Clarks view, however,
is criticized by many including Allen (2008) for its lack of evidence. Even
the evidence provided by Clark appears to be unreliable (Allen, 2008).
Finally, the trust view is due to Putnam (1993). He holds that the
Catholic tradition emphasizes vertical bonds with the Church which tend
to undermine horizontal bonds with fellow citizens. This reduces the
level of trust in society and increases the transaction costs of economic
exchange. Therefore, Catholicism has a negative effect on institutions and
hence a negative effect on growth. Trust also has a direct impact on creditworthiness. Lack of trust in society creates a creditor unfriendly environment that limits economic progress. The creditor unfriendly environment
is also due to the anti-usury culture prominent in the Catholic tradition
(Stulz and Williamson, 2001).
In a related thesis, Kuran (1997) argues that a large divergence in individual and social preferences may lead to a decline in social capital. He
formally defines the phenomenon of divergence in individual and social
preferences as preference falsification. To illustrate this further, one could
think of a situation where social preference is moulded by an authoritarian state and is different from individual preferences. In such a situation,
individuals would never reveal their true individual preference in public,
which is different from the social preference put forward by the state, due
to the fear of a backlash. Kuran describes this phenomenon as preference
falsification. In his book, he describes both positive and negative implications of preference falsification. One of the negative implications could
be a decline in social capital and trust which may have negative effects on
long-term development.
02/03/2011 10:09
34
3.7 THE GEOGRAPHY THEORY

The striking differences between the long-term histories of peoples of the
different continents have been due not to innate differences in the peoples
themselves but to differences in their environments . . . if the populations of
Aboriginal Australia and Eurasia could have been interchanged during the
Late Pleistocene, the original Aboriginal Australians would now be the ones
occupying most of the Americas and Australia, as well as Eurasia, while the
original Aboriginal Eurasians would be the ones now reduced to downtrodden
population fragments in Australia. (Jared Diamond (1997), Guns, Germs and
Steel, p. 405.)
In this section, I discuss the theories that link geography with development. I classify the existing theories of geography into five different categories. They are as follows: the climate view, the agriculture view, the market
proximity view, the disease view and the sophisticated geography view.
The climate view holds that the population in the tropics is not industrious enough largely due to the energy-sapping heat. Natural availability of food in abundance also makes tropical people idle. This has a
direct and negative effect on human productivity and hence economic
growth (Montesquieu, 1748). In a recent study, Parker (2000) supports
Montesquieus argument. According to his thesis, an individuals desire
to maximize utility is dependent on motivation, homeostasis and neural,
autonomic and hormonal adjustments. These physiological factors are
governed by the hypothalamus. The activity of the hypothalamus is
heavily dependent on thermodynamics. In hot conditions, the hypothalamus secretes hormones which negatively affect motivation and enterprise
whereas, in cold climates, individuals are naturally hard working. These
tendencies affect the steady-state level of income in these two regions. The
average steady-state income in cold climates is naturally higher than the
average steady state in hot climates. Hence, climate explains two-thirds
of the per capita income differences between the tropics and temperate
regions.
The agriculture view is due to Gallup and Sachs (2000). This view maintains that high relative humidity and high night-time temperature in the
tropics cause high plant respiration and slow down plant growth. They
argue that the deficiency in plant growth in the tropics is also related to the
lack of nutrients in tropical soil. Humid tropical soils (alfisols, oxisols and
ultisols) are typically low in nutrients and organic matter. This limits plant
growth and also causes soil erosion and acidification. In addition, the
lack of frost allows a greater number of pests to survive and breed. These
factors have a debilitating impact on agricultural productivity and inhibit
economic progress. In addition to this, an alternative and sophisticated
02/03/2011 10:09
35
agriculture view exists due to Diamond (1997). According to this view,

prior to the period of colonization, Europeans used technology which was
specifically designed to suit temperate conditions. In the late eighteenth
and early nineteenth centuries, when the Europeans embarked on a colonization drive, they introduced these technologies in the colonies. They
worked well in colonies with temperate conditions, but failed to deliver the
same goods in the tropical environment. This explains the low productivity of tropical agriculture and hence slow economic progress.
The market proximity view is due to Sachs and Warner (1995b), Sachs
and Warner (1997b) and Gallup et al. (1998). There is no disagreement
within the economics community that trade and commerce generate
wealth and prosperity (Smith, 1776 [1976]). One of the important preconditions for trade is easy access to major markets. According to the market
proximity view, unfavourable geographic location characterized by no or
limited access to ports or ocean-navigable waterways, being landlocked,
is a major impediment to trade and commerce. Access to a port or major
markets in this situation often involves crossing international boundaries,
which makes the cost of transportation relatively high and limits international trade. Absence of international trade in these economies confines all
commercial activities to small internal markets. This causes an inefficient
division of labour and underdevelopment. If one looks at inland Africa,
which is also one of the poorest areas in the world, it is quite evident that
most of the countries of this region are landlocked. This prevents these
countries from effectively participating in international trade because
transport costs are too high (Sachs and Warner, 1997b).
The disease view is due to Bloom and Sachs (1998), Gallup et al. (1998)
and Gallup and Sachs (2001). According to this view, infectious malaria
has a debilitating effect on human productivity and directly affects economic performance. Gallup and Sachs (2001) point out that the countries
with intensive malaria grow 1.3 percentage points slower per person per
year than countries without malaria and a 10 percentage point reduction
in malaria might result in a 0.3 percentage point increase in annual per
capita income growth. Bloom and Sachs (1998) also claim that the high
incidence of malaria in sub-Saharan Africa reduces the annual growth rate
by 1.3 percentage points a year. In other words, eradication of malaria
in 1950 would have resulted in a doubling of current per capita income.
Sachs (2003a) and Carstensen and Gundlach (2006) in empirical studies
report strong and negative effects of malaria on economic progress, even
after controlling for institutions and openness.
Finally, the sophisticated geography view is due to Jared Diamond
(1997). Diamond, in his book entitled Guns, Germs and Steel asks the
question: why did history unfold so differently on different continents? He
02/03/2011 10:09
36
argues that geography and biogeography moulded the contrasting fates

of human lives in different continents. In summary, his hypothesis is as
follows.
He argues that geography has endowed mankind with different sources
of food and livestock. The Eurasian climate, especially southwest Asia
which he calls the Fertile Crescent6 was best suited for the growth of
the maximum number of edible wild grains and large mammals. Early
hunter-gatherers living in this region domesticated these wild grains and
adopted a sedentary agriculture-based lifestyle. They domesticated the
large mammals for meat, milk and muscle power which they could use in
farming. The use of large mammals in agriculture immensely improved
farm productivity yielding more and more food surplus. This technology
and knowledge spread all across Eurasia along the same latitude. Close
contact with the large mammals also led to frequent outbreaks of epidemic diseases among the human population in this region. This helped
Eurasians to develop immunity to many of these diseases over the long
run. The gift of food surplus and a sedentary lifestyle allowed them to
invest more time in the development of guns, steel swords, ocean-going
ships and so forth. The societal structure became hierarchical and far more
complex than that of the hunter-gatherers.
This, however, was not the case in other continents. In places like
New Guinea humans were left with very limited choices of food. No wild
grains or large mammals were available for domestication. So humans
in New Guinea remained hunter-gatherers. In the Americas, corn was
the major grain that was domesticated. Other grains were not available.
The continent also lacked large mammals for domestication. The only
available option was the llama, which is weak and yields less meat and
milk than cows, goats or sheep. The indigenous American population
were also not familiar with horses and they lacked immunity to the fatal
Eurasian diseases. As a result the Europeans faced little or no resistance
from the indigenous American population in their colonial conquest. The
Aztecs, the Incas and the Mayas were fighting an unequal battle with the
Europeans powered with guns, steel swords, horses and germs. More than
two-thirds of the population were wiped out by a smallpox epidemic when
they first came into contact with the Europeans. Riding on the power of
guns, germs and steel, the Europeans colonized most of the known world,
getting access to a large pool of resources, which helped them to develop
an advanced industrialized society. The Americas, Australia and New
Zealand gained from European migration and the migration of European
technology along with them. The rest of the world remained largely
impoverished and the gap widened over time with the development of
more and more advanced technology in the West.
02/03/2011 10:09
3.7.1
37
Diseases, War and Urbanization in Europe
A related but somewhat different theory of geography is from Voigtlnder

and Voth (2008). They pose the question of why Europe was prosperous
relative to the rest of the world during the 1700s. They argue that even
though pre-modern Europe was largely Malthusian, the reason behind
the divergence in income between Europe and the rest of the world can
be explained by three factors, namely, diseases, war and urbanization.
War, urbanization and trade-driven disease raised death rates and once
death rates were higher, incomes could remain at an elevated level even
in a Malthusian world. High density of population, rapid urbanization
and the disease-ridden European cities kept the death rates relatively high.
International trade also helped spread diseases across the continent.
High death rates meant labour scarcity and higher wages. This changed
the nature of demand. As a result people started having more children.
However, people still had above subsistence income and they spent their
surplus income mainly on manufactured goods produced in cities. This
reinforced urbanization and trade. Urbanization raised the risk of diseases
and trade raised the risk of war which was financed by taxes from the
cities. Therefore they argue that geography, diseases and war had a positive impact on living standards in pre-modern Europe.
3.7.2
Malaria in Africa
The malaria view can be categorized into two broad categories. The first
deals with the economic burden of malaria in contemporary Africa, and
the second deals with the historical impact of malaria and diseases on the
continents long-run economic development. I call the former the contemporary malaria view, whereas the latter is the historical malaria view.
The contemporary malaria view is due to Sachs (2003b). According to
this view, malaria dramatically lowers labour productivity and the return
on foreign investment and raises transaction costs of international trade,
limiting development, which is typically observed in sub-Saharan Africa. I
have discussed this view in the following sections.
Sachs (2003a) in a cross-country empirical study shows that institutions
do matter, but not exclusively. He highlights the prevalence of malaria
as another important factor which should not be underestimated. In a
series of studies, Bloom and Sachs (1998), Gallup et al. (1998) and Gallup
and Sachs (2001) show that infectious malaria has a debilitating effect on
human productivity and directly affects economic performance. Sachs et
al. (2004) argue that with malaria and subsequent low productivity of agricultural labour, sub-Saharan Africa cannot generate enough marketable
02/03/2011 10:09
38
surpluses which also limits the prospect of market development. Markets,

even if they develop, remain concentrated at a very local level. This is
indicative of a situation of low level equilibrium trap in these economies.
Acemoglu et al. (2003a) strongly disagree with Sachss view. They argue
that the disease environment influences the balance of power between
previously isolated populations when they come into contact. The local
disease environment influences the colonization strategy and settlement
decision, which sets up the path for future institutional development
(Acemoglu et al., 2001, 2002). Therefore, the disease environment affects
the level of development indirectly through institutions. I call this the
historical malaria view. This view is also closely related to the historian
Phillip Curtins work on epidemiology of the New World and Africa
(Curtin, 1968). In this article, Curtin shows that epidemiological factors
have influenced economic decisions and economic patterns of the New
World (particularly tropical America) and Africa.
Another angle within the historical malaria view is a line of argument that is strongly pursued by a section of the historians. Miller (1982)
uses Portuguese travellers records, missionary and church documents to
show that frequent epidemics of malaria and yellow fever caused massive
depopulation in the agriculturally marginal zones of West Central Africa.
Dias (1981), in a similar study of nineteenth and twentieth century Angola
argues the same. They observe that the effects of disease, epidemics and
famines were far more powerful than the slave trade in depopulating the
region. According to their view, the increase in slave trade was an outcome
of local epidemiology and poor agriculture rather than strong Atlantic
demand. Bhattacharyya (2009b) also shows that malaria explains longrun economic development in Africa. All other variables including institutions and the slave trade are statistically insignificant.
Miller (1982) writes, The slave trade appears in some ways less a cause
of depopulation than a consequence of it when viewed in terms of droughts
and demographic changes in West Central Africa. Hence, according to his
view, the disease environment had direct effects on the demography and
economic development of the region. Inikori (1992) and Manning (1981,
1982), however, provide evidence to show that these effects were not strong
enough to have a larger impact on the African population compared to the
strong Atlantic demand for slaves.
3.8 THE TRADE OPENNESS THEORY

The greatest improvement in the productive power of labour, and the greater
part of the skill, dexterity, and judgement with which it is any where directed,
02/03/2011 10:09
39
or applied, seem to have been the effects of the division of labour ... This division of labour, from which so many advantages are derived, is not originally the
effect of any human wisdom, which foresees and intends that general opulence to
which it gives occasion. It is the necessary, though very slow and gradual, consequence of a certain propensity in human nature which has in view no such extensive utility; the propensity to truck, barter, and exchange one thing for another.
(Adam Smith (1776 [1976]), Wealth of Nations, Chapters I & II, pp. 7 and 17)
The openness view of development goes back at least to Adam Smith.

Smith (1776 [1976]) argues that openness to trade increases the size of the
market, which raises the possibility of greater division of labour. Division
of labour in turn improves productivity and productivity improvement
induces faster economic growth.7
The neo-classical theory, however, holds a slightly different view.
According to this theory, reduction in trade barriers opens up the possibility of a more efficient exploitation of comparative advantage through reallocation of factors. Labour and capital move towards their highest valued
uses improving overall productivity and the welfare of the economy.
Growth takes place in the economy due to transitional dynamics. In other
words, growth lasts only for the duration of the transitional period and
stops after the economy reaches its new steady-state levels of capital and
output per worker.
Another significant theory in the openness and growth literature is the
technology transfer view. According to the closed economy neo-classical
growth models, given the technology level at a particular point in time,
a country accumulating more physical and human capital grows faster
than a country accumulating less of the same, but all of them converge to
the same long-run steady-state equilibrium rate of growth (Solow, 1956;
Swan, 1956; Mankiw et al., 1992). In this model, income per capita can
grow in the long run only when there is exogenous technological progress.
In real life, we hardly observe convergence of income. In contrast, what
we notice is that the rich economies are growing faster than the poor ones
and the gap between the rich and the poor is widening over time.8 Coe
and Helpman (1995) point out that this gap is due to the variable rate of
technological progress. They find evidence that technology catch up or
R&D spillover bridges this gap. Dowrick and Rogers (2002), on the other
hand, report that the gap is due to variable rates of capital accumulation
as well as technological progress. Howitt (2000) builds a theoretical model
to show that because of technology transfer, R&D performing countries
converge to parallel growth paths, whereas others stagnate. Several others,
including Coe and Helpman (1995), identify trade openness as a medium
of technology transfer. In other words, what these studies argue is that
the follower economies adopt technology and knowledge developed in
02/03/2011 10:09
40
the advanced economies and the transfer takes place largely through
international trade. Therefore, effective participation in international
trade and opening up by reducing trade restrictions exposes economies to
new knowledge and new technology that helps them to grow in the long
run.
Other influential studies in the openness and growth literature are Sachs
and Warner (1995a) and Frankel and Romer (1999).
Sachs and Warner (1995a) use the growth accounting approach to show
that trade openness is the key determinant of growth. They calculate trade
openness using information on average tariff levels, non-tariff barriers,
black market premiums, state monopoly on major exports and so forth.
Their regression results suggest, on average, open economies grow somewhere between 2 to 3 percentage points faster than closed economies.9
Their study, however, is not free from criticism. Rodrik and Rodriguez
(2000) show that the Sachs and Warner measure of openness suffers from
over-reliance on the black market premium and state monopoly on major
exports criteria.
Frankel and Romer (1999), on the other hand, resort to level accounting
using instrumental variable regressions to conclude that trade openness is
a key factor influencing economic development. They construct an instrument by separating out the influences of income, population size and land
area, which they call the constructed openness. This variable they claim
yields more accurate estimates of the relationship between openness and
per capita income. They report a strong and positive effect of openness on
income per capita.
Even though there has been a fair bit of agreement about the association of trade openness with growth for at least the last couple of decades,
recently these results have come under fire in a series of papers. These
papers argue that trade openness loses importance once institutional
quality is introduced as a control in a regression framework. Rodrik et
al. (2004) reach a similar conclusion, which we discussed previously. One
possible explanation is openness enhancing institutional development.
Therefore, a regression model with openness and institutions as controls
is not picking up any individual effect of openness on income as it is
operating through the institution channel. This view is supported by Wei
(2000) who suggests that more open countries face greater losses from
corruption than less open ones as the loss from corruption is higher in
the case of foreign transactions. Therefore, open countries have a higher
incentive to develop better institutions. Another explanation comes from
Dollar and Kraay (2003). They argue that cross-country regressions of
log-level per capita gross domestic product (GDP) on instruments of
institutions and openness do not reflect the relative importance of these
02/03/2011 10:09
41
variables in the long run. The high correlation between institutions and
openness does not allow regression models to estimate the actual relationship. According to their analysis, regression of changes in decadal
growth rates on instrumented changes in trade and changes in institutional
quality show a significant effect of trade openness on growth. They report
that trade openness and institutions are important for economic growth
in the very long run, but trade has a relatively larger role over institutions in the short run. They do not control for geography, religion and
culture. Also, their study explains changes in growth rates, rather than
growth itself, and growth accelerations are hard to observe in the long run
(Jones, 1995). This also brings about the question of whether the spurt
in growth rate that comes out of changes in institutions and openness
is a mere impulse which tapers off in the long run or something that is
sustainable.
In a recent study, Chang (2008) argues that overtly strong commitment to free trade is not good for growth. As an alternative strategy,
he promotes heterodox capitalism, which includes a combination of
protectionism, government promotion of favoured industries, strong
state-owned enterprise and heavy regulation of foreign direct investments. Changs main piece of evidence in support of this argument is as
follows. He shows that the developing countries grew at a rate two times
faster during the heterodox era of the 1960s and 1970s than during the
1980s onwards free trade era. The average growth rate for developing
countries during the 1960s and 1970s, according to Chang, was 3 per cent.
In contrast, the average growth rate during the period 1980 to 2002 for
the same country group was 1.7 per cent. Easterly (2009), however, finds
several inconsistencies in the figures reported by Chang if the 1980 turning
point is shifted to 1983 or 1982. He also shows that the average growth
rate for developing countries for the period 1983 to 2008 (2.7 per cent)
is in fact higher than the same for the period 1960 to 1982 (2.6 per cent).
Therefore, Changs argument heavily relies on picking 1980 as the turning
point. Easterly (2009) also shows that picking 1980 as the turning point
in itself is misleading as trade liberalization was first pushed by the IMF
and the World Bank in the aftermath of the Third World debt crisis in
1982.
In sum, even though Chang (2008) perhaps appropriately criticizes the
one size fits all free trade policy as a development strategy for all developing countries, he also appears to be guilty of moving to the other extreme
of making strong claims in favour of protectionism as an alternative.
02/03/2011 10:09
42
3.9 THE KNOWLEDGE AND HUMAN CAPITAL

THEORY
Technological progress predated capitalism and credit by many centuries, and
may well outlive capitalism by at least as long. Joel Mokyr (1990), The Lever
of Riches, p.6
The knowledge and human capital theory of development is due to Joseph

A. Schumpeter. In his book entitled The Theory of Economic Development
he describes the notion of creative destruction and how it is central to the
process of economic development. He argues that the entrepreneurs in a
capitalist economy are constantly looking for new ideas which will render
their rivals ideas obsolete. A product developed from the new idea will be
able to capture the entire market and allow the firm to enjoy monopoly
profit. This continuous process of creative destruction raises a firms
output. The possibility of enjoying monopoly profits creates an incentive
for rival firms to invest in R&D and innovate, and the process continues.
The increase in R&D activities in the economy creates positive externalities and reduces the cost of R&D for an individual firm. The economy
grows, powered by the increase in knowledge capital and R&D.
Even though Schumpeter (1934) talks about R&D, he focuses mainly
on the role of innovation rather than invention. Joel Mokyr (1990), on the
other hand, argues that inventions and innovations are complements and
both play an important role in economic development. In his book entitled The Lever of Riches he defines invention as, an increment in the set
of the total technological knowledge of a given society, which is the union
of all sets of individual technological knowledge (Mokyr, 1990, p. 10).
He argues that invention in itself may not be a meaningful concept, but,
without it, it is impossible to develop an application which has immediate
commercial usage. At the other extreme, without innovation, an inventor
will have little economic incentive to pursue new ideas.
In Part III of his book he goes on to argue that the complementarity
between invention and innovation explains why many societies failed to
be technologically creative. Invention requires individual brilliance and an
inventor expects social recognition of her work. Therefore, a society more
receptive to new ideas and willing to reward human inquiry is more likely
to produce more scientists and researchers. Innovation, on the other hand,
requires interaction with other individuals and an in-depth knowledge of
the market. Hence, it has more of a social and economic nature. Therefore,
a society willing to encourage business enterprise is more likely to produce
more innovations. However, neither of them can survive independently in
the long run. A society has to fulfil multiple conditions to generate both
02/03/2011 10:09
43
invention and innovation. He supports his argument using evidence from

the Chinese and the European history of technology.
Even though China entered the early modern era as a technology leader
of the world, soon afterwards it lost its advantage to Europe and suffered
a technological slowdown. Europe, on the other hand, acquired most of
the early technologies from China and the East through direct and indirect
means and went on to develop them further, which led to the Industrial
Revolution. Mokyr asks the question, why does this happen? His explanations are as follows. First, social groups trying to sabotage and block
new technology were far less powerful in Europe than in China. Second,
a market for technology developed fairly rapidly in early modern Europe,
which created an economically as well as socially rewarding environment
for the scientists and innovators. This encouraged private initiatives in the
development of new technology. Therefore, the change in the character of
the state did not affect investments in technology much. Even if it did, it led
to a decline of that state in the European economic hierarchy. This made
every state or monarch in early modern Europe conscious of the costs
of conformism. In contrast, Chinese scientific activities and technology
development were largely state sponsored. This prevented the development of private initiatives and hence a technology market. A change in the
character of the state often led to hostile behaviour towards innovation
and nonconformism. The abrupt ending of the famous Cheng Ho voyages
in 1433 is a classic illustration. A change in the ruler and internal power
structure in the Ming Dynasty in 1433 led to the closing down of shipyards
and production of ocean-going junks. The continuous threat to potential
innovators from rival power groups had perhaps forced them to shy away
from novel ideas.
Even though Mokyr identifies technology and knowledge as the central
factor, he does not rule out the role of culture and institutions in shaping
technology. I will discuss this further in Chapter 4.
Glaeser et al. (2004) go even further to claim that human capital is
more fundamental than institutions. They challenge the Acemoglu et
al. (2001) institutions theory which says that in temperate conditions
European settlers felt more at home and decided to settle. In these colonies, they brought Western values and institutions along with them when
they migrated that were favourable to capitalism. In contrast, Glaeser et
al. (2004) argue that the decision to settle in temperate conditions does
not tell us anything definitive about what the Europeans brought with
them when they migrated. It could well be human capital and ideas and
not institutions. In other words, they argue that institutions have only a
second order effect on economic performance and the first order effect
comes from human and social capital, which shape both institutional and
02/03/2011 10:09
44
productive capacities of a society. They call this the Lipset-PrzeworskiBarro view of the world.
In related research, Allen (2009) argues that the Industrial Revolution
in eighteenth century Britain propelled its economy to new heights which
were never experienced before elsewhere. He argues that industrial revolution took place in Britain and not elsewhere in Europe or Asia because of
two main factors. Eighteenth century Britain was a high wage economy
relative to the rest of the world. Britain was also abundant in cheap coal
energy and the cost of capital was relatively low. This created conditions
for capital- and energy-intensive technological breakthroughs. As a result,
capital- and energy-intensive technologies, such as the steam engine, the
cotton mill and coal-fired metal production were all profitable. The high
wage economy of pre-industrial Britain also meant that schooling and
apprenticeship were affordable to the masses, which, in turn, secured the
steady supply of the highly skilled labour required to operate these technologies. The Industrial Revolution would spread around the world much
later and only when further advances were made by the British engineers
to make these technologies affordable.
Other strong supporters of the human capital view are Galor and Moav
(2006). They provide a counterargument to the political power and institutions theory of Acemoglu et al. (2005a) and Acemoglu and Robinson
(2000).10 Galor and Moav (2006) argue that it was not concessions by the
elite that lead to the demise of the class society. It was the investments in
human capital by the capitalists to sustain their profit rates that caused the
gradual dismantling of the class society.
Their version of the story goes as follows. After the Industrial Revolution
in the late eighteenth and early nineteenth century Europe, production
was heavily dependent on capital and labour. The ownership of capital
was with the capitalists, whereas the ownership of labour was with the
workers. This created the class society characterized by a social division
in line with the ownership of the factors of production. However, in the
latter half of the nineteenth century, the character of capitalist production
experienced a change. Capitalists started realizing the increasing importance of technology and human capital in production. They noticed that
the productivity of a skilled worker is three or four times more than an
average unskilled worker. Therefore, investments in human capital and
technology became crucial for the sustenance of profit rates of the capitalists. These investments in mass schooling by the capitalists paid dividends
in the long run by generating positive externalities for R&D and faster
economic growth.
Galor and Moav (2006) also provide empirical evidence in favour of
their theory using the voting patterns of the legislators on Englands
02/03/2011 10:09
45
education reform, the Balfour Act of 1902. This bill was meant to create
a publicly supported secondary school system. They show that legislators support for this bill was positively correlated with the industrial
skill-intensity level of the counties. Legislators representing counties
with a high industrial skill-intensity level overwhelmingly supported this
bill, whereas the others did not. By assuming that the legislators received
financial backing from the industrialists of their respective counties, they
argue that this result reflects the eagerness of the capitalists owning skillintensive industries to invest in human capital.
Many others debate this view. Both Acemoglu et al. (2001) and Rodrik
et al. (2004) put human capital in their list of proximate determinants
rather than as a fundamental determinant of economic progress. The literature, however, is yet to arrive at a consensus.
3.10 THE STATE FORMATION AND WAR THEORY

All societies must deal with the possibility of violence, and they do so in different ways ... Most societies, which we call natural states, limit violence by political manipulation of the economy to create privileged interests. These privileges
limit the use of violence by powerful individuals, but doing so hinders both
economic and political development. In contrast, modern societies create open
access to economic and political organisations, fostering political and economic
competition. Douglass C. North, John Joseph Wallis and Barry R. Weingast
(2009), Violence and Social Orders, p.1
Max Weber argued that high density of population and scarcity of land
in early modern Europe led to intense territorial competition amongst
monarchs and feudal lords. Therefore, warfare for the control of scarce
resources and land were extremely common and frequent. Taxation was
used as the main instrument by small- and medium-sized states to finance
their war expenditure. In other words, the symbiotic financial relationship between the ruler and his or her subjects became ever more powerful
during this time. This relationship, in Webers view, is fundamental to the
formation of statehood and perhaps is at the heart of the construct of a
modern state.
Recent research by economists and political scientists suggest that state
formation and state capacity play a major role in economic development. Herbst (2000), using Webers thesis, also argues that state weakness
explains the lack of development in Africa. As I discussed earlier, Herbsts
answer depends on the factor intensity of the continent. Historically,
Africa is land abundant and labour scarce. As a result, property rights
over land are weak and institutions to adequately tax land have failed
02/03/2011 10:09
46
to evolve. African states survived for centuries without engaging in the

institution building that facilitates tax collection, investment in defence,
maintaining bureaucracy and providing the rule of law. The lack of these
institutions made African states extremely fragile. The trend changed little
during the colonial period and the Cold War. Hence the weak institutions
in Africa continued to persist.
Besley and Persson (2008) and Besley and Persson (2009) also present
theory and evidence on why the lack of state capacity may lead to poor
development outcomes. They argue that a states institutional capacity
to levy taxes and support markets are typically constrained by its history
of investments in legal and fiscal capacity. States with weak legal systems
and imperfect taxation institutions are unlikely to be capable enough to
adequately levy taxes and support markets. In other words, state capacity
would be fairly limited. This would lead to poor development outcomes.
They also show that investments in common interest public goods, such
as fighting external wars, political stability and inclusive political institutions are all conducive to building state capacity. In contrast, civil war,
internal conflict and restrictive political institutions are all damaging to
state capacity.
In a similar vein, North et al. (2006, 2009) also argue that the nature
of the social order characterized by the nature of the existing state can
explain the process of modern social development. Their framework suggests that in a natural state the actors use political power within the existing political system to limit economic entry and create rents. These rents
are then used to stabilize the political system and limit violence. However,
restrictive economic entry and restrictive political institutions are not
conducive to long-run economic growth as it suffocates enterprise, the
free market and innovation. They argue that this is the natural way most
human societies are organized, even today, and hence the name natural
state. In contrast, a handful of developed countries have developed
open access social orders where open access to economic and political
organizations facilitates competition and supports the market. In other
words, social order in these states is sustained by competition and not rent
creation.
The obvious question, however, is how the transition from natural
state to open access state takes place. According to North et al. (2006),
war and the threat of violence play a crucial role. In the event of an external war, the incumbent state requires more taxation revenue and support
from its subjects in general. To increase revenue, the state resorts to
increasing taxation in return for more widespread political rights. OBrien
(2001, 2005) also shows that in early modern Britain the state was borrowing heavily from its subjects to finance continental war in return for
02/03/2011 10:09
47
political rights. This led to the development of an efficient financial system

conducive to rapid economic growth.
In summary, one can argue that the development of the state has played
a crucial role in prosperity. The factors that influence state development
are taxation, legal institutions, external conflict and internal conflict.
Evidence suggests that external conflict boosts state capacity, whereas the
reverse is observed if the conflict is internal in nature.
NOTES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
Even geography is not truly fixed as we are facing the challenge of climate change.
The idea of reverse causality is from Seymor Martin Lipsets (1959) modernisation
hypothesis which says that institutional quality, education and health improve as
countries develop. However, recent research in political science and economics shows
that institutions diverge not in a systematic manner as outlined by Lipset (1959), but
due to exogenous shocks at critical historical junctures as outlined by the famous work
of Barrington Moore (1966) on institutions (see Thelen, 2000; Acemoglu et al., 2007;
and many others) . For a survey of critical juncture hypothesis, see Thelen (1999).
This argument, however, is not universally applicable. In particular, this does not apply
to the cross-national studies that focus on a shorter time period and also use variation in diseases that are not geography-based and can be influenced by public health
intervention.
Herbsts thesis is not stand-alone. Bates (1983) and Stevenson (1968) also provide some
tentative evidence within Africa which is in agreement with the population density
argument of the thesis.
The exceptions are the settler colonies of South Africa, Zimbabwe, Namibia and
Kenya.
The Fertile Crescent spans a part of modern day Israel, Palestine and the Jordan valley.
For further reference, see map in Diamond (1997) p. 135.
Yang and Borland (1991) use a growth model to formalize Adam Smiths notion of
division of labour.
This is disputed by many if we consider the last 50 years. World income distribution
has increased over the last 50 years, especially with the impressive growth performances
of India and China over the last two to three decades, but the diverging trend certainly
holds over a longer period.
Wacziarg and Welch (2003) improve the Sachs and Warner (1995a) measure and extend
it to the 1990s. Their measure however is not free from the basic criticisms of Rodrik
and Rodriguez (2000).
I have discussed the political power and institutions theory of Acemoglu et al. (2005a)
and Acemoglu and Robinson (2000) in Section 3.5.
02/03/2011 10:09
4.
Empirical evidence
Before introducing the unified framework in Chapter 5, let me start with

a discussion of some of the empirical issues involving the root causes of
economic progress literature. First, I introduce some rough correlations
in the data (both diagrammatically and in tabular form). Then I report
some of the issues relating to the appropriateness of the levels model in
estimation. I follow it up with a discussion on identification which is taken
from my earlier work on this topic (see Bhattacharyya, 2009a). In Sections
4.5 and 4.6, I provide evidence on the role of diseases (malaria in particular) in Africa. The main message is that the causal linkages between deep
determinants of development and economic development are complex and
multidimensional. They also vary across countries and continents. The
cross-section regressions are useful in identifying correlation, but may not
tell us much about the causal links. The unifying framework in Chapter 5
is an attempt to identify such causal links.
4.1 A QUICK LOOK AT THE DATA

The analysis here is based on a dataset which consists of per capita GDP
levels, a measure of institutions, measures of geography, a measure of
openness, measures of religion and a measure of human capital in (up to)
180 countries. Since I am combining data from different sources, I have
to deal with different numbers of observations for different variables. The
data typically also come from different years. In case of institutional measures, I use averages in order to capture the long-run effect. The definitions
and sources of all the variables are summarized in the Data Appendix.
Table 4.1 presents the summary statistics of these measures.
I divide the dataset into six major parts depending upon the variable
that they are measuring. They are as follows: measures of economic development, measure of institutions, measures of geography, measures of
religion, measure of openness and trade and measure of human capital.
Following is a brief outline of each of them. We also plot some of these
variables to look at the correlation.
Economic development is measured by the level of per capita GDP in
48
02/03/2011 10:09
Empirical evidence
Table 4.1
49
Summary statistics
Variables
Number
of obs
Measure of institutions
Rule of law index
Measures of geography
Distance
Malaria risk
Land area within
tropics
Soil suitability
Land area within
100km of
ocean or oceannavigable river
Measures of religion
Catholicism
Islam
Mean
Standard
deviation
171
0.0015
0.9441
178
160
146
19.6293
0.3678
0.4991
17.0873
0.4390
0.4779
0
0
0
63.89
1
1
154
146
13.3965
46.0343
9.8512
37.604
0
0
55.07
100
174
174
31.3805
22.0817
36.1243
34.749
0
0
97.3
99.8
4.0914
0.6384
Measure of openness and trade

Log of trade share
146
Measure of human capital
Enrolment ratio in
88
1900
Measures of economic development
Log initial income
50
(1820)
Log initial income
60
(1870)
Log initial income
39
(1900)
Log initial income
136
(1950)
Log initial income
115
(1960)
Log per capita
147
GDP in 2000
32.34
26.37
Minimum Maximum
2.19
2.58
0.1
2.37
5.76
95
6.58
0.3893
5.98
7.52
6.88
0.5668
5.98
8.09
7.46
0.6096
6.30
8.41
7.29
0.9446
5.67
10.32
6.3153
0.8647
4.72
8.23
8.588
1.1177
6.19
10.799
Note: For a detailed discussion of the definition and source of these variables, see Data
Appendix.
02/03/2011 10:09
50
2000. I also use the level of GDP per capita in 1960 as the initial level of
development.
The average rule of law index is a measure of institutional quality. In
particular, it measures the overall institutional quality and the quality of
governance in a particular country. The range of this index varies from
2.5 to 2.5 with higher values implying better institutional quality. I use
an average value of this index from 1998 through to 2000 for each of 171
countries as the measure. The data show that Somalia has the weakest
institutions in the world and Singapore has the strongest institutions.
In case of geography, I use five different measures: distance, malaria
risk, land area within tropics, soil suitability and land area within 100 km
of the ocean or an ocean-navigable river.
Distance is a measure of the distance from the equator. Following the
argument of Sachs (2003a), I use this as a measure of climate. The greater
the distance from the equator, the further the country is from the tropics
and more temperate or cold it is. I exploit data from 178 countries.
Malaria risk measures the share of population at risk from malaria in
160 countries in the year 1997. I use this as a measure of disease burden. A
higher value indicates greater risk for the population. Typically, tropical
and subtropical countries register higher risk of malaria and the risk of
malaria declines in the temperate climates.
Land area within the geographical tropics is used as a measure of the
effects of geography on agricultural productivity. This is calculated as
a proportion, a higher value implying that a higher proportion of the
countrys land is tropical in nature and, therefore, the countrys agriculture is expected to have tropical characteristics marked by slow and low
plant growth. I use soil suitability as an additional measure of agriculture. Ukraine registers as the country with highest proportion of suitable
soil.
Finally, I use land area within 100 km of the ocean or an ocean-navigable
river as a measure of market proximity. The higher the proportion of
land within 100 km of the ocean or an ocean-navigable river, the greater
is the chance for an economy to participate in maritime trade and have
better access to larger markets. The bulk of the sub-Saharan economies is
observed to be landlocked.
In measuring religion, I use the proportion of population following
Catholicism and Islam in the year 1980 as a measure of the same, respectively. The data shows Spain, Ireland, Portugal and the most of Latin
America to be predominantly Catholic and the Middle East, Indonesia,
Pakistan and some of the ex-Soviet republics to be predominantly
Islamic.
I measure openness to trade by using the log value of the actual trade
02/03/2011 10:09
Empirical evidence
51
share from Frankel and Romer (1999). They calculate the actual trade
share by taking the percentage of imports plus exports to GDP in 1985
from 146 countries. The higher the trade share, the more open is the
economy. Singapore is observed to be the most open economy in the world
and Myanmar is the most closed.
The human capital measure is from Benavot and Riddle (1988). They
record data on the primary school enrolment ratio in 1900. I use this as
a proxy measure of historical human capital attainment. The data shows
that in the United States, 95 per cent of the relevant school going population were enrolled in primary school in 1900 whereas in Equatorial Guinea
the same number was as low as 0.1 per cent.
Pairwise correlations are reported in Table 4.2. It shows that better
institutions, less disease, better agriculture and soil conditions, more trade
and higher initial income are all positively associated with development.
This is also revealed in the following scatter plots (Figure 4.1ai).
The data shows that countries that were wealthier in 1960 were also
likely to be wealthier in 2000. The data also shows that institutions
measured by rule of law, trade and 1900 school enrolment rate are positively correlated to the current level of development. This is in line with
the prediction made by many theorists of root causes of development.
In contrast, malaria risk and land area within the tropics are negatively
correlated to the current levels of development. The scatter plots also
show that countries further from the tropics, countries with proximity to
oceans and ocean-navigable waterways, and countries with Catholicism
as the dominant religion are relatively prosperous. The obvious question is, does this imply causality? I will discuss this issue in detail in what
follows.
4.2 GROWTH OR LEVELS: WHICH ONE IS THE

APPROPRIATE EMPIRICAL FRAMEWORK?
The levels framework adopted by Rodrik et al. (2004) and other recent
papers1 centres on a cross-country regression where the dependent variable is the current level of economic development, typically measured
by the natural logarithm of real GDP per capita, lnyiT, where i indexes
countries and T indicates the year. The explanatory variables consist of a
measure of the quality of contemporary institutions, IiT, and a vector of
other explanatory variables, WiT. The other explanatory variables include
log trade share in the current period and latitude:
ln yiT 5 hIiT 1 FWiT 1 ei
(4.1)
02/03/2011 10:09
52
02/03/2011 10:09
0.6298
0.6265
0.4975
0.3431
0.1629
0.7783
0.2374
0.3310
0.4758
0.2573
0.7073
0.1331
0.3178
0.7387
0.8439
0.6931
0.8299
0.6210
1.0000
0.0160
0.2768
0.7613
0.3310
0.3896
0.6837
0.6551
0.5993
1.0000
0.0857
0.0500
0.6522
0.0907
0.2096
1.0000
0.6647
0.8956
Log initial Average Distance

income rule of law
(1960)
index
0.8182
1.0000
Log per
capita
GDP in
2000
0.1445
0.2190
0.4504
0.1831
0.3874
1.0000
0.6493
Malaria
risk
0.1473
0.0458
0.5681
0.0101
0.0854
1.0000
0.1417
0.2145
0.2025
0.3128
1.0000
Land area Land area

within
within
tropics
100 km
of ocean
or oceannavigable
river
Pairwise correlation of the major variables (obs 5 99)
Log per capita

GDP in 2000
Log initial
income (1960)
Average rule of
law index
Distance
Malaria risk
Land area within
tropics
Land area
within 100 km
of ocean
or oceannavigable river
Log of trade
share
Enrolment ratio
in 1900
Catholicism
Islam
Table 4.2
0.1382
0.0062
0.1415
1.0000
Log of
trade
share
0.0208
0.3452
1.0000
1.0000
0.5104
Primary Catholicism
enrolment
ratio in
1900
1.0000
Islam
Empirical evidence
53
11
Fitted values/per capita GDP in 2000
LUX
10
USA
NOR
SGP
CAN
DNK
CHE
HKG
IRL
AUS
ISL
JPN
NLDSWE
BEL
AUT
FIN
GER
FRA GBR
ITA
NZL
ISR
ESP BRB
TWN
PRT
GRC
KOR
MUS
TTO
ARG
SYC
CHL
GNQ PRI
URY
MEX
MYS
ZAF
BRA
GAB
TUR
VEN
IRNPAN
TUN
BWA THA
DZA
FJI
CRI
COL
DOM
ROM
PRY
PER
NAM
SLV
SYR EGY JOR
GTM
CPVMAR
PHL GUY
IDN
CHN
LKA
ECUJAM
PNG
GIN
BOL
IND
ZWE
COG
HTI
CMR
PAK
HND
CIV
NIC
COM
BGD
SEN
LSONPL MRT
KEN BEN GMB GHA
MOZ
UGA BFA
MLI
RWA
TGO TCD MDG
NER CAF ZMB
NGA
MWI
SLE
ETH GNB
BDI
TZA
6
5
Initial income
Note: See Table 8.2 for country abbreviations.
Figure 4.1a
Log per capita GDP in 2000 and log initial income
11
LUX
USA
NOR
CAN CHE
SG P
HKG IRL
AUS
ISLDNK
JPN BEL
NLD
AUT
SWE
FIN GBR
FRA
ITA
NZL
ISR TWN
ESP
PRT
SVN
GRC
KOR
CZE
KNA
MUS
TTO
SVKARG
SYC
HUN
EST
CHL PRI
GNQ
URY
RUS
MEX
POL
HRV
MYS
BLR
LTU
KAZ
DMA
ZAF
VCT LVA
GAB
BRA
TUR
VEN
BLZIRN
BWA TUN
LCA
THA
PAN
GRDLBN
DZA BGR
CRI
COL
DOM
MKD FJI
GEO
SWZ
PRYUKRROM
PER
NAM
SLV
SYR
GTM
MAR
CPV
JOR
PHL
IDNGUYEGY
CHN
LKA
ECU
ALB
AZE PNG
JAM
KGZ
ARM
GINBOL
IND
COG ZWE
HTI
CMR
MDA
HND
PAK
CIV
COM NICBGD
SEN
NPL
LSO
MRT
TJK
GHA
GMB
STP
KEN
KHM
YEM
BEN
MOZ
UGA
MLI TCD
BFA
CAF
RWA
TGO
NER ZMB
MDG
NGA
MWI
GNB
SLE ETH
BDI
TZA
10
6
2
Rule of law
Figure 4.1b
Log per capita GDP in 2000 and rule of law
02/03/2011 10:09
54

11
LUX
USA
NOR
CAN
SGP
DNK
CHE
HKG
AUS
IRL
ISL
JPN
NLD
BEL
AUT
SWE
FIN
GBR
GER
FRA
ITA
NZL
ISR
ESP
TWN
PRT
SVN
GRC
KOR
CZE
MUS
TTO
SVK
ARG
HUN
EST
PRI
CHL
URY
RUS
POL
HRV
BLR
LTU
KAZ
LVA
ZAF
TUR
VEN BRA
TUN
BGR
DZA
LBN
CRI
DOM
MKD
GEO
UKR
SWZ
ROM
PRY
PER
SLV
EGY
GTM
MAR
JOR
CHN
ALB
AZE
JAM
KGZ
ARMBOL
10
GNQ
MYS MEX
IRN PAN BWA
GAB
BLZ
THA
COL
NAM
IDN
GUY
LKA
PNG
ZWEIND
SYR
ECU
PHL
GIN
COG
HTI
CMR
PAK
HND
CIV
NIC
SEN
GHA
GMB
YEMKHM
BEN
MOZ
UGAA
MLI
BFA
CAF
RW
TCD
TGO
NER
MDG
ZMB
NGA
MWI
GNB
SLE
BDI
TZA
MDA
BGD
NPL
LSO
TJK
MRT
KEN
7
ETH
6
0
0.2
0.4
0.6
0.8
Malaria risk
Figure 4.1c
Log per capita GDP in 2000 and malaria risk
11
LUX
USA
SGP
GER
MAC
10
ATG
SVN
CZE
SVK SYC
EST
GNQ
HRV MYS
BLR
LTU
KAZ
LVA
GAB
HKG
TWN
BRB
TTO
AUS JPN
ISR
KNAMUS
NZL
ESP
PRT
GRC
KOR
ARG
PRI
CHL
URY
MEX
DMA BRA
ZAF
VCT
VEN THA
IRN
TUN
LCA BLZ BWA
PANGRD
DZA
LBN COL
FJI
CRI
DOM
MKD
GEO
UKR
SWZ
PER
PRY
NAM
SLV
EGY
GTM
SYR
CPV
MAR
JOR
PHL
IDN
CHN
LKA
ECU GUY
AZE
JAM
PNG
KGZ
ARM
GIN BOL
IND
ZWE
COG
CMR HND HTI
MDA
PAK
CIV
NIC
COMSEN
BGD
NPL
LSO
MRT
TJK
GHA
GMB
STP
KEN
KHM
YEM
BEN
MOZ
UGA
MLI
BFA
CAF
RWA
TGO TCD
NER
ZMBMWI MDG
NGA
SLE
ETH GNB
BDI
TZA
NOR
CAN CHE
DNK
IS L
NLDIRL
AUTBEL
SWE
FIN
ITA FRAGBR
HUN
POL
RUS
TUR
BGR
ROM
6
0
20
40
60
Distance
Figure 4.1d
Log per capita GDP in 2000 and distance
02/03/2011 10:09
Empirical evidence
55
11
USA
CAN
AUS
10
NOR
CHE
SWE
FIN
URY
RUS
MEX
KAZ
LVA
ZAF
BRA GAB
TUR TUN
VEN
IRN
BWA
THA
DZA
COL
MKD
GEO
ROM
NAM PER PRY EGY
SYR
MARGTM
JOR
CHN
ECU
AZE
KGZ
ARM
GIN
BOL
IND
ZWECOG
PAKCMRCIV
NPL
LSO
TJK
MRT
KEN
BEN
GHA
HUN
MYSPOL BLR
LTU
BGR
UKR
IDN
ALB
PNG
NIC
SEN
UGA
MLI
BFA
CAF
RWA
TCD
NER
ZMB
MWI
ETH
BDI
FRA
ITA
PRT
CZE
SVK
CHL EST
ARG
DN K
HKG
IRLJPN SGP
NLD
GBR
GERBEL
ISR NZLTWN
SVN
KORGRC MUS
TTO
AUT
ESP
HRV
PAN
LBN
CRIM
DO
SLV
PHL
LKA
JAM
HTI A
MD
BGD
HND
GMB
KHM
YEM
MOZ
TGO
NGA
MDG
GNB
SLE
TZA
6
0
20
40
60
80
100
Land area within 100 km of ocean or ocean-navigable river
Figure 4.1e
Log per capita GDP in 2000 and land area within 100 km of
ocean
11
USA
NOR
CAN
DNK
CHE
IRL
JPN
NLD
BEL
AUT
SWE
FIN
GBR
GER
FRA
ITA
NZL
ISR
ESP
PRT
SVN
GRC
KOR
CZE
SVK
ARG
HUN
EST
URY
RUS
POL
HRV
BLR
LTU
KAZZAF
LVA
TUR
IRN
TUN
BGR
LBN
MKD
GEO
UKR
ROM
SYR
MAR
JOR
CHN
ALB
AZE
KGZ
ARM
10
SGP
HKG
AUS
TWN
MUS
TTO
CHL
MEX
MYS
BRA GAB
VEN
THA
PAN
CRI
COL
DOM
PER
SLV
GTM
PHL
IDN
LKA
ECU
JAM
PNG
GIN
BOL
ZWE
COG
HTI
CMR
HND
CIV
NIC
SEN
MRT
GHA
GMB
KEN
KHM
YEM
BEN
MOZ
UGAA
MLI RW
BFA
CAF
TCD
TGO
NER
MDG
ZMB
NGA
MWI
GNB
SLE
ETH
BDI
TZA
BWA
DZA
PRY
EGY
NAM
IND
MDA
PAK
BGD
NPL
LSO
TJK
6
0
0.2
0.4
0.6
0.8
Land area within tropics
Figure 4.1f
Log per capita GDP in 2000 and land area within tropics
02/03/2011 10:09
56

11
LUX
USA
NOR
CAN DNK
CHE
HKG
AUS
IRL
ISL
JPN
NLD
BEL
FIN SWEAUT
FRA GBR
ITA
NZL
ISR
ESP
PRT TWN BRB
GRC KOR
MUS
TTO
ARG
HUN
EST SYC PRI
CHL
URY
RUS
MEX POL
MYS
DMA
ZAF
VCT
GAB
TUR
VEN
IRN BRA
BLZ
TUN
BWA
THA
PAN
GRD LCA
BGR
DZA
CRI
COL
DOM
SWZ
ROM
PER PRY
NAM
SLV
EGY
GTM
SYRIDN
CPV
JOR
PHL MAR
CHN
GUY
ECU LKA
JAM
PNG
GIN
BOL
IND
ZWE
COG
HTI
CMR
PAK
HND
CIV
NIC
COM
BGD
SEN
NPL
LSO
MRT
GHA
GMB
KEN
YEM
BEN
MOZ
UGA RWA
MLI
BFATCD
CAF
TGO
NER
MDG
ZMB
NGA
MWI
GNB
SLE
BDI
10
SGP
TZA
6
2
4
Natural log of trade share
Figure 4.1g
Log per capita GDP in 2000 and log trade share
11
LUX
NOR
DNK
HKG
ISL SGP
JPN
SWE MACGBR
FIN
NZL
ISR
TWNBRBATG
GRC
KORKNA
10
USA
AUS
GER
MUSTTOCZE
EST
RUS
MYS
KAZ ZAFBLRLVA
VCT
TUR
IRN
TUN
BWA
THA
GRD
BGR
DZA
LBN
FJI
MKD
GEO
UKR ROM SWZ
NAM
EGY
SYR
MAR
JOR
IDN
CHN
GUY
ALB LKA
AZE
JAM
PNG
KGZ
ARM
GIN
IND
ZWE
CMR
MDA
PAK
CIV
BGD
NPL SEN
MRT
TJK
GHA KEN
GMB
KHM
YEM
BEN
MOZ
MLI
BFA
TCD
TGOCAF
NER
MDG
ZMB
NGA
MWI
GNB
SLE
ETH
CAN CHE
NLD
FRA
ITA
BEL
AUT
IRL
ESP
PRT
SVN
SVK
ARG
SYC
HUN URY
PRI
CHL
GNQ HRVPOL
MEX
LTU
DMA
GAB
BRA
VEN
BLZ
LCA
PAN
CRI DOM
COL
PER
PRY
SLV
GTM
CPV
PHL
ECU
BOL
HTI
HND
NIC
LSO
STP
UGA RWA
BDI
TZA
6
0
20
40
60
80
100
Catholicism
Figure 4.1h
Log per capita GDP in 2000 and Catholicism
02/03/2011 10:09
Empirical evidence
57
11
HKG
10
FIN
TWN
ITA
PRT
MUS
GRC
ARG
PRI
URY
RUS CHLMEX
ZAF
BRA
VEN
BLZ
TUNPAN
BGR
DZA
CRI
DOM COL
ROM
PER
PRY
NAM
EGY SLV GTM
PHL
IDNSYR
LKA
GUY
ECU
GNQ
BOL
IND
ZWE
CMR
GHA GMB
ESP
BRB
USA
CAN
IRL AUS
FRA
NZL
TTO
SYC
HUN
VCT
LCA
GRD
FJI
JAM
HND
NIC
NOR
DNK
CHE
BEL NLD
AUT
SWEGER
GBR
JPN
LSO
TGOMDG
SLE
MWI
6
0
20
40
60
80
100
Enrolment ratio in 1900
Figure 4.1i
Log per capita GDP in 2000 and enrolment ratio in 1900
It is apparent that this model, which is known as the levels regression

model, is nested within the growth specification reported in (2.7),2 since it
can be rewritten as:
ln yiT 5 (1 1 aT) ln yit 1 hIiT 1 FWiT 1 ei
(4.2)
where g 5 [ hF ] , ZiT 5 sIiT t and 1 1 aT 5 0.

W
iT
The key question is whether (4.1) is the appropriate empirical framework

to test the role of deep structural determinants on long-term economic
progress. Sachs (2003a) argues that economic growth is a dynamic process
rather than a state. Therefore, it is more appropriate for empirical studies
to estimate growth regressions rather than levels regression. Alternatively,
Dollar and Kraay (2003) argue that during the sixteenth century, income
across the world was largely similar. Therefore, the growth model with an
initial income around 1500 essentially boils down to a levels model. Let me
illustrate how. The existing literature implicitly starts from the following
model:
02/03/2011 10:09
58
ln yiT 5 a 1 b ln yiT 2500 1 gIiT 1 dXiT 1 ei
(4.3)
The above is a growth regression with the current level of economic

development, measured by the natural logarithm of real GDP per capita,
ln yiT as the dependent variable, where i indexes countries and T indicates
the year. The explanatory variables consist of the initial level of economic
development (or the level of development 500 years ago), ln yiT 2500, a
measure of the quality of contemporary institutions, IiT, and a vector of
other explanatory variables, XiT, which includes human capital, geography
and trade. Under the assumption that the levels of development in the
distant past were not too different across countries3 the long-run growth
literature generally estimates the following version of the above model
which is similar to (4.1):
ln yiT 5 a 1 gIiT 1 dXiT 1 ei
(4.4)
To estimate the causal effect of institutions on long-run growth, the

literature uses historical and geographical instruments. These instruments
predict the historically and/or geographically determined component of
institutions at the first stage and then they are used in the second stage to
estimate the causal effect of institutions on long-run growth.
Easterly (2009) also supports the levels framework over growth. He
argues growth is inherently unstable and it is difficult to predict in advance
what policy or set of policies would work. Therefore, with a small crossnational sample on growth, it is easy to jump into erroneous conclusions too
quickly. This is what he calls the law of small numbers. He suggests that
the levels framework is more appropriate as it reflects the outcome of entire
previous growth and hence is free from the problems of small numbers.
Nevertheless, to statistically test whether growth or levels is the appropriate framework, we use the following method. The null hypothesis, H0:
(1 1 aT) 5 0, reduces (4.2) to the levels regression, whilst rejection of the
null favours the growth specification.
The economic intuition behind this test is that the levels regression is
implicitly explaining the steady-state distribution of income levels. This
assumption is explicit in the augmented Solow model derived by Mankiw
et al. (1992) who use investment rates as the proximate determinants of
the neo-classical steady state. Mankiw et al. (1992) go on to show that if
economies are not in their steady states, the transitional dynamics of the neoclassical model are captured by the addition of the initial income level in a
growth regression.4 If economies actually are in steady state, as explained by
the right-hand-side variables in Equation 4.1, then the addition of the lagged
dependent variable, as in Equation 4.2, should add no explanatory power.
02/03/2011 10:09
Empirical evidence
Table 4.3
59
Omitted variable test: OLS regressions
Dependent
Variable
Log initial income

(1820)
Log initial income
(1870)
Log initial income
(1900)
Log initial income
(1950)
Rule of law (2001)
Average log trade
Share
Latitude
R2
Log per capita GDP in 2000

OLS
obs 5 44
OLS
obs 5 52
OLS
obs 5 35
OLS
obs 5 108
0.35**
(0.1599)
0.36**
(0.1949)
0.46**
(0.1705)
0.78***
(0.0780)
0.043
(0.1150)
0.003
(0.0046)
0.8303
0.60***
(0.1345)
0.043
(0.1376)
0.001
(0.0053)
0.7522
0.399***
(0.1325)
0.133
(0.1526)
0.006
(0.0070)
0.7992
0.598***
(0.0858)
0.49***
(0.0936)
0.15*
(0.1013)
0.01**
(0.0040)
0.8181
Notes: ***, ** and * indicate significance levels of 1%, 5% and 10%, respectively, against
a one-sided alternative. Figures in parentheses are the respective standard errors. The
standard errors reported in the regressions are heteroskedasticity robust. All the regressions
reported above are carried out with an intercept.
I test the null hypothesis by adding an initial income term into Rodrik
et al.s (2004) levels specification. The results are reported in Table 4.3. In
all the four regressions using 1820, 1870, 1900 and 1950 levels of log per
capita GDP as initial income,5 I observe that the initial income term is
statistically significant. This indicates that the standard levels framework
neglects the role of transitional dynamics and the coefficient estimates
suffer from omitted variable bias.
4.3 IDENTIFICATION ISSUES WITH THE LEVELS

FRAMEWORK
The levels framework also suffers from identification issues when human
capital is added into the mix. I illustrate this point in Bhattacharyya
(2009a).
In Table 4.4, I adopt the same strategy and add schooling into the levels
model to estimate partial effects of human capital and institutions. In
02/03/2011 10:09
60
02/03/2011 10:09
Identification issues in long-run growth empirics: part A
F-test
Sargan test (p)
Latitude
Total years of
schooling (2000)
Avg log trade share
Rule of law (2001)
Explanatory
variables
0.10
0.01
(0.0140)
0.47
(0.2849)
1.55***
(0.2541)
Original model
obs 5 68
(1)
0.08
(0.1219)
0.038**
0.99
1.1
(1.322)
2.47
(4.023)
3.55
(6.224)
Sargan test (p)
Latitude
Total years of
schooling (1995)
Avg protection
against
expropriation
19851995
Model with
Explanatory
schooling obs 5 43 variables
(2)
0.002
(0.0128)
0.96***
(0.2227)
Original model
obs 5 65
(3)
0.98
0.04
(1.428)
1.35
(38.11)
2.72
(104.2)
Model with schooling

obs 5 51
(4)
Acemoglu et al. (2001)

Dependent variable: log per capita GDP in 1995
Panel A: 2SLS Results
Rodrik et al. (2004)

Table 4.4
61
02/03/2011 10:09
0.7311
33.57
1.32***
(0.2243)
2.05**
(0.8444)
1.44**
(0.5645)
0.002
(0.0164)
0.014
(0.0199)
obs 5 56
(2)
1.29***
(0.2949)
0.8061
0.83***
(0.1923)
1.1*
(0.6263)
1.33***
(0.4949)
0.026*
(0.0141)
0.01
(0.0200)
obs 5 56
(3)
Schooling (2000)
R2
F-stat
Avg. prot.
against
expropriation
19851995
Latitude
Log settler
mortality
L. pop. den. in
1500
Dependent
variable
0.2553
10.97
0.02
(0.0155)
0.47***
(0.1448)
Avg prot. against

expropriation
obs 5 67
(4)
0.7195
43.61
0.02
(0.0202)
1.39***
(0.2179)
0.58***
(0.1333)
obs 5 55
(5)
0.7692
0.01
(0.0189)
0.49***
(0.1683)
1.11***
(0.2153)
0.49***
(0.1283)
obs 5 51
(6)
Schooling (1995)
Notes: ***, ** and * indicate significance levels of 1%, 5% and 10%, respectively, against a two-sided alternative. Figures in parentheses
are the respective standard errors. Joint F-test p-value of rule of law and schooling is reported in column 2, Panel A. The standard errors
are heteroskedasticity robust. L. pop. den. in 1500 is log population density in 1500, ENGFRAC is fraction of population speaking English,
EURFRAC is fraction of population speaking European languages. In Panel B the dependent variables are rule of law (2001), schooling (2000),
average protection against expropriation, and schooling (1995).
Rule of law
(2001)
R2
F-stat
Latitude
CONST
0.5472
21.54
0.25***
(0.1039)
0.95***
(0.3311)
0.18
(0.2197)
0.008
(0.0057)
0.02**
(0.0088)
Log settler
mortality
ENGFRAC
EURFRAC
Rule of law
(2001) obs 5 76
(1)
Dependent
variable
Panel B: first stage regressions
62
column 1 (Panel A, Table 4.4) I follow Rodrik et al. (2004)6 and estimate
their preferred model using the same set of instruments as they did in their
study. Rule of law is the only statistically significant variable in this model,
which confirms their basic finding that institutions dominate the influence
of both trade and geography as the fundamental determinant of long-run
development. In column 2, I add schooling into this model and we are
unable to isolate the partial effects of institutions and human capital as
none of the coefficients are statistically significant. However, the F-test
on the joint significance of schooling and rule of law reveals that they are
jointly significant with a p-value of 0.038. To investigate the reason behind
this, I look at the first stage regressions reported in columns 1 and 2 of
Panel B. In column 1, I confirm that there is a strong partial correlation
between the settler mortality instrument and current rule of law a result
well documented in the literature. However, I also find that the settler mortality instrument is correlated with current schooling (see column 2) and
this correlation is independent of the correlation between current schooling and current rule of law (see column 3). I observe that the correlation
between fitted values of rule of law and the fitted values of schooling is as
high as 0.9435. This is perhaps causing a severe multicollinearity problem in
the second stage regression making all coefficients statistically insignificant.
In column 3 (Panel A, Table 4.4), I follow Acemoglu et al. (2001)7 and
estimate their preferred model. I also confirm their finding that institutions
measured by expropriation risk have a causal effect on long-run growth.
In column 4, I add schooling in 1995 into this specification and I confront
the same multicollinearity problem that I encountered in column 2. None
of the coefficient estimates are statistically significant. I take a quick look
at the first stage regressions (Panel B, columns 46). They reveal that the
settler mortality instrument is as good a predictor of schooling as it is for
expropriation risk, and this relationship is independent of the correlation
between schooling and expropriation risk. Also, the correlation between
fitted values of expropriation risk, and schooling is as high as 0.9036. This,
again, is causing a multicollinearity problem at the second stage.
In Table 4.5, I follow Acemoglu and Johnson (2005) and Easterly
and Levine (2003) and estimate their preferred specifications.8 Acemoglu
and Johnson (2005) evaluate the relative importance of property rights
institutions and contracting institutions measured by constraint on the
executive and the legal formalism index, respectively, and find that the
former have a first-order effect on long-run growth and the latter does not
seem to matter.9 Easterly and Levine (2003) show that endowment measured by settler mortality, latitude, landlocked dummy and crops/minerals
dummy affect development through institutions. I confirm their findings
(see Panel A, columns 1 and 3). However, I encounter the same problem
02/03/2011 10:09
63
02/03/2011 10:09
Constraint on
executive
Total years of
schooling (1995)
Sargan test (p)
0.99
0.06
(0.1735)
0.77***
(0.1869)
Original model obs 5 41

(1)
0.12
(0.2755)
0.45
(1.731)
0.22
(0.6945)
0.99

obs 5 37
(2)
Sargan test (p)
Total years of
schooling (1995)
Oil
Institutions index
1.56***
(0.4275)
0.26
2.19***
(0.1735)
Original model
obs 5 72
(3)
0.99***
(0.2909)
0.26***
(0.0671)
0.93**
(0.3638)
0.13

obs 5 55 (4)
Easterly and Levine (2003)

Explanatory
variables
Panel A: 2SLS Results
Acemoglu and Johnson (2005)

Identification issues in long-run growth empirics part B
Legal formalism
Explanatory
variables
Table 4.5
64
02/03/2011 10:09
15.16
36.31
0.6565
44.79
0.7249
1.38***
(0.2048)
0.56
(0.4413)
0.597***
(0.1323)
obs 5 55 (3)
1.32***
(0.2168)
0.38
(0.4387)
0.56***
(0.1353)
0.13
(0.1370)
0.7341
obs 5 53 (4)
Total years of
schooling (1995)
Crops/minerals
(10 variables)
Institutions
index
R2
F-stat
Oil
Landlocked
Log settler
mortality
Latitude
Dependent
variable
Panel B: First Stage Regressions
0.6379
7.17
0.14**
(0.0652)
1.21*
(0.6931)
0.09
(0.1886)
0.25
(0.2028)
2.41**
(0.018)
Institutions
index
obs 5 72 (5)
0.7053
6.84
1.5***
(0.3061)
3.5
(2.997)
0.28
(0.8690)
0.60
(1.02)
1.01
(0.452)
obs 5 55 (6)
1.0***
(0.2398)
1.5
(2.238)
0.26
(0.6415)
1.26*
(0.7609)
1.22
(0.306)
2.4***
(0.4098)
0.8434
obs 5 55 (7)
Total years of schooling (1995)
Notes: ***, ** and * indicate significance levels at 1%, 5% and 10%, respectively, against a two-sided alternative. Figures in parentheses are the
respective standard errors. For crops/minerals (10 variables) the table reports F-test of joint significance of the individual variables with p-value
in parentheses. English legal origin is used as an instrument for legal formalism in Acemoglu and Johnson (2005). L. pop. den. in 1500 is log
population density in 1500. In panel B the dependent variables are constraint on executive, legal formalism, institutions index, and total years of
schooling (1995).
F-stat
0.3285
0.91***
(0.2104)
0.28
(0.4617)
Log settler
mortality
English legal
origin
L. pop. den. in
1500
Constraint on
executive
R2
0.15
(0.1146)
2.02***
(0.2515)
Constraint Legal formalism

obs 5 41 (2)
on executive
obs 5 41 (1)
Dependent
variable
Table 4.5 (continued)
Empirical evidence
65
of multicollinearity when I add schooling into these specifications.10 The

first stage regressions of panel B show that the standard instruments used
in these specifications are as good a predictor of institutions as they are of
schooling. This makes the fitted values of institutions and schooling for the
second stage regressions correlated with each other correlation coefficient
of 0.9226 in the case of the Acemoglu and Johnson (2005) specification and
correlation coefficient of 0.8562 in the case of the Easterly and Levine (2003)
specification. This causes the multicollinearity problem at the second stage.
Furthermore, I also notice similar problems with the model estimated in
Acemoglu et al. (2003b). Acemoglu et al. (2003b) argue that the fundamental cause of post-war instability in many of the least developed countries is
institutional. They show that the volatility in per capita GDP growth and
other macroeconomic indicators have a strong negative relationship with
institutions. They argue, using the settler mortality instrument, that this
negative relationship has its roots in the colonial institutions. They also
show that this finding is robust when they control for religion, latitude,
initial income, log inflation and so forth.
However, this finding also suffers from the identification problem that
I have documented before. In Table 4.6, I outline the problem using a
particular measure of volatility the standard deviation of GDP per
capita growth (197098). In column 1 of Panel A it can be seen that initial
executive constraint, which is a measure of institutions in 1950, 1960
and 1970, negatively impacts volatility in output growth in the following
two decades. This implies that institutional weaknesses in the past lead
to higher volatility in the current growth rate. This result holds when I
control for latitude. But if I replace the initial constraint on the executive
by total years of schooling in 1960, then the same negative relationship
is observed, even though this time it is between schooling and volatility.
This is indicative of the identification problem similar to Acemoglu et al.
(2001), Rodrik et al. (2004) and Acemoglu and Johnson (2005), which
I discussed previously. The first stage regressions in Panel B show that
the log settler mortality instrument is not only correlated with initial
constraint on the executive, but it is also correlated with schooling. The
correlation between schooling and log settler mortality is independent of
the correlation between schooling and initial constraint on the executive.
Therefore, it appears that the cross-sectional data is not ideal for separating out the partial effects of institutions and human capital on long-run
growth. This is because the fitted values of the endogenous variables in the
second stage of a 2SLS regression can be highly correlated with each other
given the set of instruments used. Multicollinearity at the second stage
blows up the standard errors making all coefficient estimates statistically
insignificant.11 However, I may end up with an entirely different result if
02/03/2011 10:09
66
Table 4.6
Identification problems in Acemoglu et al. (2003)
Dependent
variable
Standard deviation of GDP per capita growth (197098)

Model (1)
obs 5 71
Model (2)
obs 5 52
Panel A: second stage of 2SLS

Initial constraint
on executive
Total years of
schooling (1960)
Latitude
0.46**
(0.1941)
0.56***
(0.1603)
0.03
(0.0231)
0.04
(0.0214)
Panel B: first stage for initial constraint on executive and total years of schooling
(1960)
Dependent
variables
Initial constraint
on executive
obs 5 71
Log settler
mortality
Latitude
Initial constraint
on executive
R2
Total years of schooling (1960)

obs 5 52
obs 5 49
0.85***
(0.1957)
1.2***
(0.2208)
0.94***
(0.2685)
0.003
(0.0216)
0.05**
(0.0223)
0.2590
0.5387
0.05**
(0.0224)
0.23*
(0.1339)
0.5718
a two-sided alternative. Figures in parentheses are the respective standard errors. All the
regressions reported above are carried out with an intercept. Log settler mortality is used as
an instrument for initial constraints on executive.
a different instrument set is used. Hence, the appropriate interpretation is

that the regressions are simply uninformative as they provide no information that would alter our theoretical priors one way or the other. In other
words, based on these results, I am unable to comment on the relative
importance of institutions and human capital to long-term development.
Nevertheless, we do learn from the empirical studies conducted that
institutions protecting individual rights and supporting entrepreneurship
matter. We learn that geography plays an important role in economic
development, be it directly or through shaping incentives and economic
institutions. We also learn that geography is critically important for the
continent of Africa (Bhattacharyya, 2009b).
02/03/2011 10:09
Empirical evidence
67
With this as a background, I introduce the unified framework to explain

the process of development in Western Europe. I then compare and contrast that with the history of development in other parts of the world.
4.4 EXISTING RESULTS WITH THE LEVELS

FRAMEWORK
As I have illustrated above, the majority of the existing results in the
root causes of economic progress literature indicate that institutions are
a major driver of economic progress over the long run (Acemoglu et al.,
2001; Rodrik et al., 2004; Acemoglu and Johnson, 2005; Easterly and
Levine, 2003). Sachs (2003a) and Carstensen and Gundlach (2006) show
that it is not just institutions; diseases are also important. In this section, I
show that there is some heterogeneity in the institutions result. Institutions
are no longer important if we divide the sample between high-income
and low-income countries using the World Bank criteria. In a sample of
low-income countries, malaria turns out to be the main explanator of lack
of development. I use the levels framework to demonstrate this effect.
Therefore, I implicitly assume that initial income during the sixteenth
century was similar across countries. The results, however, are robust to
the inclusion of initial income. I report the results as follows.
In order to dig deeper into the malaria result discussed above, I ask a
further question. Is the finding a reflection of the overall variation in the
data or is it largely driven by the significant presence of the low-income
economies (LIEs) in the sample? A closer look at the dataset reveals that 26
out of 39 observations used in the estimation are from LIEs.12 Estimating
the same model using a sample of LIEs shows that malaria is the only
variable with statistically significant coefficient estimates. This is reported
in Table 4.7a. The impact of one sample standard deviation reduction in
malaria (which is 0.44) on per capita GDP is 1.7 fold. This is less than the
pooled sample estimate of a twofold increase. However this is the only
variable that is statistically significant. Estimating the same regression on
a sample of high-income economies (HIEs) fails to yield meaningful estimates because of very few degrees of freedom. The regression relies on as
few as 13 observations. The ordinary least squares (OLS) estimates with
the HIE sample, however, show that diseases are no longer important.
It is only institutions that matter. The partial R2 of the OLS estimates
reported in Table 4.7b show that in a pooled sample 28 per cent of the
variation in per capita GDP is explained by institutions and about 41 per
cent by malaria, whereas in a sample of LIEs, the explanatory power of
institutions is statistically insignificant and 43 per cent of the variation is
02/03/2011 10:09
68
Table 4.7a
Determinants of level of development: the case of low-income

economies (LIEs)
Dependent
variable

OLS
Pooled
sample
obs 5 66
0.45***
(0.0947)
0.04
(0.0938)
1.03***
(0.1839)
Catholicism
2e-04
(0.0012)
Enrolment
3e-04
ratio in 1900 (0.0029)
R2
0.8833
F-test (p-value)
0.0000
Instruments
Average rule
of law index
Log of trade
share
Malaria risk
2SLS
Restricted
sample
(LIEs)
obs 5 32
Restricted
sample
(HIEs)
obs 5 34
Pooled
sample
obs 5 39
Restricted
sample
(LIEs)
obs 5 26
0.23
(0.2050)
0.12
(0.1585)
0.87***
(0.2011)
1.6e-04
(0.0022)
0.002
(0.0083)
0.6742
0.0000
0.41***
(0.1103)
0.02
(0.1073)
0.05
(0.6148)
0.002*
(0.0013)
0.003
(0.0029)
0.7841
0.0000
0.89***
(0.3173)
0.08
(0.1685)
1.61***
(0.3238)
0.004
(0.0032)
0.004
(0.0049)
0.83
(0.7267)
0.45
(0.3345)
1.17***
(0.5040)
0.001
(0.0041)
0.02
(0.0139)
Log settler mortality,

ENGFRAC,
EURFRAC, Constructed
openness, Malaria
ecology
a one-sided alternative. Figures in parentheses are the respective standard errors. LIEs and
HIEs signify low-income economies and high-income economies, respectively. For the 2SLS
regression on the HIEs, malaria no longer remains significant. The 2SLS regression on the
HIEs does not yield meaningful estimates due to the problem of degrees of freedom. There
are only 13 observations for the HIEs and hence are not reported here. The standard errors
reported in the regressions are heteroskedasticity robust. All the regressions reported above
are carried out with an intercept.
explained by malaria. In an HIE sample, the situation reverses and institutions perform 34 per cent of the explaining and the contribution of malaria
is statistically insignificant. Based on the evidence, therefore, it is fair to
conclude that the major effect of malaria on per capita GDP comes from
the LIEs. Institutions become more important in a sample of HIEs.
Therefore, one can put forward the following hypothesis with some
caution to explain this correlation.13 Disease is the most important factor
at a low level of development. Overcoming diseases ensures a basic
02/03/2011 10:09
Empirical evidence
Table 4.7b
69
Partial R2 of OLS estimates (reported in Table 4.7)
Dependent variable

Pooled sample
obs 5 66
Average rule of law

index
Natural log of trade
share
Malaria risk
Catholicism
Enrolment ratio in
1900
Restricted sample
(LIEs) obs 5 32
Restricted sample
(HIEs) obs 5 34
0.283***
0.048
0.344***
0.004
0.021
0.001
0.413***
0.0004
0.0002
0.430***
0.0002
0.002
0.0003
0.088*
0.034
Notes: ***, ** and * indicate significance levels of 1%, 5% and 10%, respectively, against a
one-sided alternative. The partial R2 indicates the proportion of variation in the dependent
variable that is explained by the explanatory variables when other things are held constant.
Multiplying the numbers by 100 will give the percentages.
minimum level of labour productivity and from then on institutions

become important.
One implication could be that the data on economic development has
twin peaks and there are multiple equilibria. The lower peak can be best
explained by geography. Note that this is not just geography per se, but
the poverty traps that emerge out of it.
However, just by looking at the log settler mortality data (which is an
instrument in the 2SLS regression) one may suspect that the strong negative correlation between malaria and current per capita GDP in LIEs is
driven by the large proportion of high malaria and low per capita GDP
African economies in the sample. To find out the true picture, I perform
some quick statistical analysis, which, in fact, confirms the suspicion.
There is a strong association between malaria and low per capita GDP
when the sample size is reduced down to African LIEs. The next section
presents this finding and discusses its possible implications.
4.5 IS THERE ANY CONNECTION BETWEEN LOWINCOME COUNTRIES AND AFRICA?

A quick check of the sample that the 2SLS regression (see Table 4.7a) on
LIEs uses reveals that 10 out of 26 observations are African LIEs, which is
approximately 39 per cent of the sample.
02/03/2011 10:09
70
Table 4.8
Determinants of level of development: the case of LIEs in

Africa
Dependent variable
Average rule of law index

Log of trade share
Malaria risk
Catholicism
Enrolment ratio in 1900
R2
F-test (p-value)
Instruments

OLS obs 5 15
2SLS obs 5 10
0.055
(0.2167)
0.15
(0.1981)
0.9***
(0.2653)
3e-05
(0.0080)
0.01
(0.0172)
0.9282
0.0004
0.07
(0.1725)
0.35
(0.2452)
0.98***
(0.2902)
0.005
(0.0085)
0.04
(0.0221)
Log settler mortality,

ENGFRAC, EURFRAC,
Constructed openness,
Malaria ecology
a one-sided alternative. Figures in parentheses are the respective standard errors. The 2SLS
regression may not yield meaningful estimates due to the problem of degrees of freedom.
There are only ten observations for this regression. The standard errors reported in the
regressions are heteroskedasticity robust. All the regressions reported above are carried out
with an intercept.
The regression estimates reported in Table 4.8 show that malaria risk is
the only explanator that has a statistically significant coefficient estimate
when the sample is exclusively of LIEs from Africa.
Based on these results, is it fair to conclude that the root of Africas
development problem is malaria? The answer is no. The correlation
between malaria and the level of GDP can be due to some omitted factor
specific to Africa which moves in the same direction as malaria. For
instance, the history of the slave trade in Africa may be the true reason for
the continents lack of development. Malaria being an outcome of the lack
of development will also share a high positive correlation with the slave
trade variable. Therefore, running a regression of current per capita GDP
on malaria without controlling for the correlation between slave trade and
current per capita GDP will lead to the error of attributing the movement
in per capita GDP data that is due to slave trade to malaria. A proper
02/03/2011 10:09
Empirical evidence
71
analysis of the true causes should involve estimation of a model which

controls for all the relevant variables.
In summary, this section shows that institutions dont rule. Malaria
is an important correlate of development. The explanatory power of
institutions and malaria are almost equal in the pooled sample. Malaria,
however, is the most importan`t factor for the LIEs. The results also show
that the malaria result in an LIE sample is driven by the large proportion
of African economies in the sample. This certainly makes sense given
the enormity of the malaria problem of that continent. Malaria kills 12
million Africans every year, out of a continent-wide total of roughly 9
million deaths per year. However, this does not convincingly reject all
the other competing theories, which may be correlated with both malaria
and the current level of development. In order to present more convincing
evidence of cause and effect, I construct a model which controls for all the
variables relevant to Africas development or lack of development. I estimate this model to confirm that malaria does matter when a host of other
relevant factors are controlled for. This is to be investigated in Section 4.6.
4.6 COLONIAL INSTITUTIONS, DISEASES AND

FORCED MIGRATION: THE CASE OF AFRICA
It is well known that Africa is falling behind the rest of the world in terms
of economic well-being. Even though global poverty is on the decline due
to rapid economic growth in India, China and other parts of the world,
Africas contribution to this decline is disappointing. Absolute poverty in
many of the African nations is, in fact, rising (Sachs, 2005). What is the
fundamental cause behind this decline? This has been a topic of research
for a few decades now. Even though it is extremely difficult to summarize
this voluminous literature, it is perhaps fair to say that three strands of
thought stand out.14
4.6.1
Towards an Empirical Question
The first is the disease view. According to this view, malaria and other
infectious diseases have fatal as well as debilitating effects on the human
population in Africa. It negatively influences productivity, savings and
investments in physical and human capital, and directly affects the economic performance of the continent (Bloom and Sachs, 1998; Gallup and
Sachs, 2001).15 According to Bloom and Sachs (1998), the high incidence
of malaria in sub-Saharan Africa reduces the annual growth rate of the
continent by 1.3 percentage points a year and eradication of malaria in
02/03/2011 10:09
72
the 1950s would have resulted into a doubling of per capita income. Sachs
(2003a) and Carstensen and Gundlach (2006) using a global cross-national
dataset and Lorentzen et al. (2008) using cross-national and sub-national
datasets also make similar arguments about the role of diseases. Lorentzen
et al. (2008), in particular, argue that higher adult mortality is associated with an increased level of risky behaviour, higher fertility and lower
investment in physical and human capital. Acemoglu and Johnson (2007),
however, question these results. They find that there is no statistically significant effect of improved life expectancy on GDP levels, leading them to
conclude that diseases do not have a direct role in development.
Despite the doubts posed by Acemoglu and Johnson (2007), a significant number of recent studies tend to support the disease view both
at the macro as well as micro level. Weil (2005) and Bloom and Canning
(2005) calibrating the effects of health from a range of micro estimates
into a macro model show that these effects are important at the aggregate
level. Kalemli-Ozcan et al. (2000) and Kalemli-Ozcan (2002) also show
that lower mortality as a result of better health contributes to economic
growth. In related literature, Arndt and Lewis (2000), Bell et al. (2003)
and Kalemli-Ozcan (2006) find that HIV/AIDS is reversing the trends in
demographic transition in Africa and is negatively affecting growth.16 At
the micro level, Knaul (2000), Schultz (2002), Bleakley (2003), Behrman
and Rosenzweig (2004), Miguel and Kremer (2004), and many others find
that improved health leads to better individual economic outcomes.17
The second is the colonial institutions view. According to this view, the
persistent effect of colonial institutions can explain the huge differences in
income across all ex-colonies including Africa (Knack and Keefer, 1995;
Hall and Jones, 1999; Acemoglu et al., 2001; Bhattacharyya, 2004, 2009a;
Rodrik et al., 2004; Nunn, 2007). The story as outlined by Acemoglu et al.
(2001) goes as follows.18 Europeans resorted to different style of colonization depending on the feasibility of settlement. In a tropical environment,
the settlers had to deal with killer malaria and hence a high mortality rate.
This prevented colonizers from settling in a tropical environment and they
erected extractive institutions in these colonies. These colonial institutions
have persisted over time and they continue to influence the economic
performance of the colonies even long after independence. Hence, the
Acemoglu et al. (2001) argument is that diseases affect economic performance only indirectly through institutions. Nunn (2007), using a stylized
model for Africa, shows that colonial extraction, when severe enough,
can cause a society to move from a high to a low production level equilibrium. Due to the stability of a low-level equilibrium, a society can remain
trapped in this equilibrium even after the period of colonial extraction is
over.
02/03/2011 10:09
Empirical evidence
73
Earlier work by Easterly and Levine (1997), Sachs and Warner (1997b)
and Temple (1998) also reports a strong link between quality of institutions
and post-war growth (or the lack of it) in Africa.19 Easterly and Levine
(1997) show that ethnic diversity in Africa has led to social polarization
and the formation of several rival interest groups, which increase the
likelihood of selecting socially suboptimal policies when an ethnic representative in the government fails to internalize the entire social cost of their
rent-seeking policies. Sachs and Warner (1997b), on the other hand, stress
Africas lack of openness to international markets and unfavourable geography as other contributors to poor growth, in addition to poor quality
institutions. Temple (1998) emphasizes the role of social arrangements in
explaining Africas slow growth.
Finally, a third group of explanations relates to the economic impact
of Africas engagement in the slave trade. According to this view, Africas
engagement in the slave trade caused massive depopulation of the continent
over two centuries (see Gemery and Hogendorn, 1979; Manning, 1981;
Inikori, 1992). The result was a significant slowdown in division of labour,
demographic transition,20 human capital accumulation and long-run economic growth (Inikori, 1992). Depopulation also resulted in an implosion
of the continents production possibility frontier21 and an unambiguous
reduction in welfare (Gemery and Hogendorn, 1979). The secular decline
in welfare continued over more than two centuries, plunging the continent
into economic backwardness. In a recent paper, Nunn (2008) also reports a
negative causal relationship between the slave trade and current economic
performance in Africa. He shows that the slave trade prevented state development, encouraged ethnic fractionalization and weakened legal institutions, and through these channels it affected economic development.
These competing theories, even though plausible, do not tell us how
much of the variation in income across countries in Africa they can
explain. One possible way to arrive at an answer is to check the relative
strengths of these theories in explaining the variation when they are pitted
against each other in a regression model. In this section, I investigate
their relative strength by setting up a parsimonious regression model. In
the regression model I use log GDP per capita in 2000 as the dependent
variable and malaria risk, institutions and log total slave exports out of
Africa normalized by land area as explanatory variables. This exercise is
important to demonstrate that diseases and other factors are particularly
important in Africa. In other words, it is not all institutions when it comes
to Africa. In fact, institutions are statistically insignificant in an African
sample and malaria is the only statistically significant variable. This is contrary to the results reported by the proponents of the institutions theory
using cross-national data.
02/03/2011 10:09
74
I deal with the complex causality issues by using appropriate exogenous

instruments for malaria risk, institutions and total slave exports. Malaria
ecology from Kiszewski et al. (2004) is used as an instrument for malaria
risk. Given the controversy regarding exogeneity of malaria ecology, I also
use rain, humidity and frost as alternative instruments. My basic result
survives this test. Institutions and slave exports are instrumented by log
settler mortality22 from Acemoglu et al. (2001) and distance measures from
Nunn (2008), respectively. The results show that malaria matters the most
and all other factors are statistically insignificant. This result survives even
when I use Nunns econometric specification and dataset. I also show that
malaria dampens savings. Increases in mortality and morbidity can be
possible channels through which malaria impacts African development.
Increased mortality induces households to increase current consumption and save less for the future (hence the negative relationship between
savings and malaria). Increased morbidity, on the other hand, adversely
affects productivity, reducing household income and savings. This slows
down capital accumulation and economic development. This discussion
perhaps sheds some light on why malaria is so persistent in Africa.
The result is striking. Malaria is the most powerful explanator (at
least statistically) of long-run economic development (or the lack of it)
in Africa. None of the other factors (including institutions and the slave
trade) are statistically significant. I also provide an explanation for the
persistence of malaria in Africa.
The benefits of looking at an Africa-only sample are twofold. First, it
allows us to statistically test the strengths of competing theories of African
underdevelopment. Second, it allows us to focus on a continent where the
majority of the bottom billions countries are located (Collier, 2007).
4.6.2
Specification and Data
In order to estimate the causal effects of malaria, colonial institutions and

the slave trade on Africas long-run economic development, I use the following model. The implicit assumption here is that initial incomes across
countries in Africa going back several centuries were similar. Therefore, I
no longer need to control for initial income explicitly and the model can be
interpreted as a long-run growth model.
logyi 5 l 1 aMALi 1 bINSi 1 gSLVXi 1 xi F 1 ei
(4.5)
where yi, MALi, INSi and SLVXi are per capita income in country i,
measure of malaria, measure of institutions and measure of slave exports,
respectively. xi is a row vector of additional control variables,23 F is the
02/03/2011 10:09
Empirical evidence
Table 4.9
75
Descriptive statistics
Variable
Log GDP per capita in
2000 (log yi)
Malaria Risk (MALi)
Expropriation Risk in
1985 to 1995 (INSi)
Log total slave exports
normalized by land
area (SLVXi)
Number
of obs
Mean
Standard
deviation
Minimum Maximum
46
7.46
0.815
6.19
9.24
49
35
0.77
5.82
0.386
1.30
0
3
1
8.27
52
3.26
3.89
2.30
8.82
vector of coefficients on other control variables denoted by the vector and

ei is the random error term. We are interested in the size, sign and statistical significance of the three coefficients a, b and g.
The estimation of Equation 4.5 is based on a dataset consisting of per
capita GDP levels, measure of malaria risk, measure of institutions and
measure of slave exports in (up to) 52 countries in Africa. Definitions and
sources of all the variables used in this study are summarized in the Data
Appendix. Table 4.9 presents summary statistics for the key variables of
interest.
GDP per capita in 2000 data is from the Penn World Tables 6.1.
According to these figures, Tanzania was the poorest country in Africa in
2000. I also use per capita income data from Nunn in Table 4.15 when I
check the robustness of the result using Nunns dataset and specification.
Note that Nunn uses income data from Maddison (2004).
Malaria risk is the percentage of population living in areas of high
malaria risk in a country in 1994. It is calculated using GIS software from
a digitized WHO map of the world distribution of malaria and a detailed
database of world population distribution in 1994.24 The variable lies
between 0 and 1 and a higher value indicates greater risk for the population. Most of the countries in the sample register high malaria incidence,
except Algeria, Tunisia and Egypt.
There are at least three measures of institutional quality that have been
used in the literature. Knack and Keefer (1995), Acemoglu et al. (2001)
and many others use expropriation risk averaged over 1985 to 1995 from
the Political Risk Services. Rodrik et al. (2004) use the rule of law index
from the World Bank. Others use the executive constraint from the Polity
dataset. The expropriation risk measure is perhaps the most appropriate
for my purpose as I would like to capture the variation in institutions
02/03/2011 10:09
76
originating from different types of colonial states and state policies (see
Acemoglu et al., 2001). It is also the closest to Douglass Norths (1981)
definition of good institutions25 as it captures the notion of extractive state.
I also check the robustness of the results using rule of law and executive
constraint measures.
Slave export data is from Nunn (2008). Nunn (2008) reports the natural
log of total slaves exported out of each of the African nations normalized
by land area and population in 1400.26 According to Nunn, the maximum
number of slaves exported was from Angola, which accounted for 23.1
per cent of the total slave exports, followed by Nigeria (12.9 per cent) and
Ghana (10.2 per cent). The fewest slaves exported were from Tunisia. I
follow Nunn and use log total slave exports normalized by land area as
my preferred measure.
Identifying good empirical proxies for each of these variables is difficult but perhaps not the most challenging part of the analysis. The major
challenges are to estimate the causal effects. In order for the estimates of
a, b and g to be interpreted as causal effects, they have to overcome some
serious econometric challenges. I list them as follows.
Endogeneity
Economic development is a complex phenomenon. Given the complex
nature of this process, reverse causality is a real possibility. For example,
rather than malaria influencing development, the causality may run the
other way round. The rich economies can afford to invest in the research
and development of drugs that cure or minimize the effect of malaria.
They can also invest in public health programmes to tackle malaria. A
similar argument can be made about institutions. Rich nations have better
institutions, not because they have grown richer due to better institutions,
but because they can afford better institutions. Furthermore, there are
endogeneity concerns with the slave trade. Societies that initially had poor
domestic institutions may have selected into the slave trades. Therefore,
the observed negative relationship between slave exports and development
may not be the causal effect (Nunn, 2008). If this is the case then OLS estimates of a, b and g will be biased away from zero as we will be erroneously
attributing the effects of income or other factors on endogenous variables
to the direct effects of these variables on income.
Measurement error
The slave export data are likely to contain both classical and non-classical
measurement error. One can identify the following sources. First, slave
ethnicities in the dataset may have been misclassified. Slaves with similar
but different ethnicities may have been classified under one ethnicity. But
02/03/2011 10:09
Empirical evidence
77
the possibility of a bias due to errors of this nature is minimal as the data
is aggregated at the country level. Second, measurement error may arise
due to the under-representation of slaves from the interior or due to the
assumption used in the construction of the data that slaves shipped from
a port within a country are either from that country or from countries
directly to the interior. In either case, OLS estimates of a, b and g will be
biased towards zero the classical measurement error (Wooldridge, 2000).
Furthermore, any random measurement error present in the data will also
have the same effect on OLS estimates. Moreover, it is not possible to rule
out non-classical measurement error.
Omitted variable bias
Many of the omitted time-invariant deep factors (culture, ethnic make-up,
colonial or legal origin, religion, climate) influencing long-run economic
development can be correlated with malaria risk, institutions and slave
exports. This has the potential to bias the OLS estimates of a, b and g
away from zero. I control for regional fixed effects, colonizer fixed effects
and legal origin fixed effects to tackle this problem. I also test the robustness of our estimates by controlling for additional covariates. Some of the
obvious ones are trade openness, Catholicism, Islam, historical schooling,
ethnic fractionalization, share of mining, foreign aid and the Gini coefficient. However, as is the case with all empirical modelling, I can never be
entirely sure that I have adequately controlled for all the omitted factors.
To tackle the problems of endogeneity and measurement error, I use the
instrumental variable (IV) estimation. A valid instrument has to satisfy the
twin conditions that it is correlated with the suspected endogenous variables (malaria, institutions and slave exports, in this case) but uncorrelated
with the error term or a measurement error hidden in the error term. It is
obviously a difficult task to find valid instruments. However, the literature
has identified several instruments that could serve my purpose.
Previous studies have used log settler mortality as an instrument for
institutions (Acemoglu et al., 2001; Rodrik et al., 2004; and others). It is
based on the idea that European colonizers erected good institutions only
in the settlement colonies. Elsewhere they erected extractive institutions.
Therefore, the settler mortality instrument is likely to be negatively correlated with the quality of institutions and also orthogonal to the random
error term since it is geography-based. Recently this instrument has come
under intense scrutiny. In Section 4.3, I show that it is almost impossible
to separate out the effects of institutions and human capital on long-run
development in a cross-section model due to multicollinearity. This is true
regardless of the specifications used. The instruments also have issues
with satisfying the exclusion restrictions as they are highly correlated with
02/03/2011 10:09
78
schooling. Albouy (2008) identifies several weaknesses with regard to the

construction of the instrument. In spite of the controversy regarding this
instrument, I continue to use it here to facilitate comparison of my results
with previous studies in the literature.
I also follow Nunn (2008) and use sailing distance from the coast to the
closest market of the Atlantic slave trade, sailing distance from the coast
to the closest market of the Indian Ocean slave trade, overland distance
from the centroid to the closest port of export for the trans-Saharan
slave trade, and overland distance from the centroid to the closest port of
export for the Red Sea slave trade as instruments for slave exports. Nunn
(2008) argues that the distance instruments are negatively correlated with
slave exports and also exogenous. Therefore, they are valid instruments.
He also uses the overland distance from the centroid to the coast and log
population density in 1400 as additional instruments. However, he notes
that the additional instruments may not satisfy the exclusion restrictions.
Therefore, I decide not to use these additional instruments.27
Finally, I follow Sachs (2003a) and Carstensen and Gundlach (2006)
and use malaria ecology as an instrument for malaria risk. Malaria ecology
is an ecologically based spatial index and depends on climatic factors and
biological properties of each regionally dominant malaria vector. Hence
it is exogenous to public health interventions and economic conditions,
and thus can serve as an instrumental variable in regressions of economic
performance on malaria risk (Kiszewski et al., 2004).28 Rodrik et al. (2004)
doubt the exogeneity of malaria ecology as they argue that from the
little information provided by Sachs (2003a), it remains unclear whether
malaria ecology can be influenced by human action. Another concern
regarding malaria ecology comes from a previous version of the text
describing the construction of the index as it says the calculation includes
mosquito abundance. Even though both critiques are technically correct,
the doubts about the exogeneity of the instrument may not be justified
for the following reasons. First, the index is vector-based and not affected
by human activity as public health interventions against malaria only
serve to break the transmission cycle, but do not eliminate the presence of
the vector itself. Even today, Anopheles mosquitoes capable of transmitting malaria can be found throughout the US and Europe, places where
malaria has been largely eradicated (see Kiszewski et al., 2004). Second,
observed mosquito abundance enters the index only as a screen for precipitation data, where the independently identified dominant malaria
vector is assumed to be absent from the specific site under consideration
if precipitation falls below a certain level per month (see Carstensen and
Gundlach, 2006). Nevertheless, I also use average rainfall, average humidity and prevalence of frost as alternative instruments for malaria and the
02/03/2011 10:09
Empirical evidence
79
results are robust to these changes. Rainfall, humidity, and lack of frost
are crucial to the life cycle of the parasite and hence could serve as good
instruments. They are also geography-based and hence exogenous to economic conditions.
One concern is that rainfall, humidity and frost may not satisfy the
exclusion restriction because they may affect development through channels other than malaria. Statistically, this will bias my estimates only if the
predicted value of malaria at the second stage is correlated with the error
term. The Hansen J-test for exogeneity of instruments indicates otherwise (see Table 4.11). Nevertheless, I also use them as exogenous control
variables which may directly influence economic performance. My results
survive this test.
In IV estimation, endogenous explanatory variables are replaced by
their predicted values from the first stage equations. The first stage equations are as follows:
MALi 5 m 1 dMEi 1 cLSMi 1 kDCi 1 xiF 1 eMALi
(4.6)
INSi 5 1 hLSMi 1 sMEi 1 nDCi 1 xiF 1 eINSi
(4.7)
SLVXi 5 y 1 wDCi 1 fMEi 1 pLSMi 1 xiF 1 eSLVXi
(4.8)
where MEi, LSMi, and DCi refer to malaria ecology, log settler mortality
and the distance instruments from Nunn (2008). Equations 4.54.8 are
at the core of the empirical results that I report in the next section. I also
report statistical tests (Hausman test, Sargan test and Hansen test) for the
validity of instruments.
An additional concern with IV is the bias due to weak instruments.
Staiger and Stock (1997) and others have shown that the consequence
of weak instruments is a large-sample bias in IV as in effect the model
becomes unidentified. Furthermore, the magnitude of the large-sample
bias increases with the number of instruments. The Staiger and Stock
(1997) results rely on asymptotic properties and asymptotic distribution
theory may not necessarily apply for this small sample. However, the bias
in 2SLS cannot be ruled out. More importantly, the limited information
maximum likelihood (LIML) estimator does not have such bias. It is also
more robust to the weak instruments problem than IV (Stock and Yogo,
2005). My basic results survive when I use the LIML estimator.
4.6.3
Results
Table 4.10 reports the core results. In column 1 of Panel A, I start with
estimating the basic model using OLS. I find that malaria negatively
02/03/2011 10:09
80
Table 4.10
Malaria as a root cause of African underdevelopment: core

results
Panel A: Model log yi 5 l 1 aMALi 1 bINSi 1 gSLVXi 1 ei

Dependent variable
Malaria risk (MALi)

Expropriation risk in
1985 to 1995 (INSi)
normalized by land
area (SLVXi)
Log per capita income in
1960
R2
Hansen J-test (p)
Growth
during
19602000
OLS
estimate
obs 5 33
(1)
2SLS
estimate
obs 5 27
(2)
LIML
Fuller
estimate
obs 5 27
(3)
2SLS
estimate
obs 5 27
(4)
0.86*
(0.4576)
0.18*
(0.0992)
0.08*
(0.0451)
4.19**
(2.105)
0.29
(0.6543)
0.39
(0.3043)
3.3**
(1.758)
0.16
(0.5251)
0.25
(0.2505)
0.04*
(0.0244)
0.004
(0.0049)
0.002
(0.0033)
0.005
(0.0049)
0.59
0.92
Hausman/Sargan test (p)

Cragg-Donald test (p)
Additional controls
Instruments
0.63
0.97
0.71
LPDi, IDCi
ME, LSM, ADC, IODC, SDC, RDC
Panel B: First stage regressions

Dependent variables
MALi
obs 5 27
(1)
INSi
obs 5 27
(2)
SLVXi
obs5 27
(3)
Malaria ecology (MEi)
0.02**
(0.0080)
0.12*
(0.0663)
0.09
(0.0745)
0.00004
(0.00009)
0.07
(0.0588)
0.03
(0.0277)
0.08
(0.2906)
0.26
(0.3741)
0.002***
(0.0007)
0.12
(0.2149)
0.18**
(0.0863)
0.63
(0.6501)
1.64*
(0.8451)
0.001
(0.0022)
0.36
(0.6597)
Log settler mortality

(LSMi)
Log population density in
1500 (LPDi)
Interior distance (IDCi)
Atlantic distance (ADCi)
02/03/2011 10:09
Empirical evidence
Indian distance (IODCi)

Saharan distance (SDCi)
Red Sea distance (RDCi)
R2
F-stat
0.02
(0.0454)
0.12
(0.0969)
0.18*
(0.0912)
0.82
54.75
81
0.13
(0.1680)
0.55*
(0.3154)
0.19
(0.3920)
0.58
6.65
0.24
(0.6589)
2.2*
(1.162)
1.6*
(0.8680)
0.64
3.88
Panel C: Instrument redundancy tests

Instruments tested
ME
LSM
IDC, ADC,
IODC, SDC,
RDC
LM test statistic
p-value
Degrees of freedom
7.7
0.05
3
6.92
0.07
3
19.57
0.08
12
Notes: ***, ** and * indicate significance levels of 1%, 5% and 10%, respectively,
against a two-sided alternative. Figures in parentheses are cluster standard errors and
they are robust to arbitrary heteroskedasticity and arbitrary intra-group correlation.
All the regressions reported are carried out with an intercept. Fullers modified LIML
estimator with a 5 1 (correction parameter proposed by Hausman et al., 2005) is used in
Panel A, column 3. Both Hansen J-test and Hausman/Sargan test p-values are reported.
In both cases, the null hypotheses are that the instruments are jointly exogenous. CraggDonald test p-values for weak instruments are also reported. The null hypothesis, in this
case, is that the instruments are jointly weak. The test statistic follows F-distribution
under the null with degrees of freedom 5 N-L, L1 (N, number of observations; L, total
instruments; L1, excluded instruments). The LM statistic for instrument redundancy tests
are distributed as chi-squared under the null hypothesis that the specified instruments are
redundant with degrees of freedom equal to the number of endogenous regressors times
the number of instruments being tested. The endogenous regressors are MALi, INSi and
SLVXi. The abbreviations used in the table are MEi, malaria ecology; LSMi, log settler
mortality; LPDi, log population density in 1500; IDCi, interior distance; ADCi, Atlantic
distance; IODCi, Indian Ocean distance; SDCi, Saharan distance; and RDCi, Red Sea
distance.
impacts development, institutions are good for development and slave

exports are negatively correlated with development.29 I also plot the OLS
partial effects (see Figure 4.2). The estimates, however, are likely to be
inconsistent as OLS does not account for endogeneity or measurement
error problems. In column 2, I estimate the model using IV. I notice that
the negative effects of malaria survive; however, institutions and slave
exports are statistically insignificant. The magnitude of the malaria effect
is also large. A one standard deviation decrease in malaria risk increases
the income of an average country in Africa fivefold. To put this into
02/03/2011 10:09
82

2
GIN
COG AGO
CMR
SEN
BWA
DZAZAF
GHA CIV
TUN
ZWE MOZ
KEN TGO
MLI BFA EGY
GNB NGA
MDG
UGA
MAR MWI
SLE
ZMB
NER
ZAR
ETH
TZA
SDN
GMB
1.5
1
.5
0
.5
.5
EGY
0
ETH
.5
AGOGINCOG
SDN
GHASEN
TUN
CMR
MLI
GNB BFA
CIV
BWA
NGA MOZZWE
TGO
KEN
MDG
ZAR
GMB
MAR
UGA
SLE
MWI
NER
1
TZA
3
e( expr8595 | X )
coef = .17928689, (robust) se = .09921514, t = 1.81
e( lgdppc2000 | X )
GAB
DZA
ZAF
e( lgdppc2000 | X )
e(lgdppc2000 | X )
GAB
.6
ZMB
.4
.2
0
.2
.4
e( mal94p | X )
coef = .8617306, (robust) se = .45775097, t = 1.88
GAB
COG
SDN GIN
AGO
CMR
BWA
ZAF
SEN
TUNCIV
ZWE
MLI
UGA
BFA
GHA DZA
MDG
EGY
ZAR
NER
KEN
MOZ
GNB
NGA
MAR
GMB
TGO
SLE
ZMB
MWI
TZA
ETH
2
0
2
4
e( ln_export_area | X )
coef = .07661942, (robust) se = .0451374, t = 1.7
Figure 4.2
Partial correlation plot: root causes of African

underdevelopment
perspective, the model explains approximately 92 per cent of the difference in per capita income in Namibia and Nigeria two countries who
also share approximately one standard deviation actual gap in malaria
risk. The Hansen J-test30 and the first stage regressions reported in Panel
B show that the instruments are valid; however, the Cragg-Donald test
for weak instruments suggests that some of the instruments may be weak.
Staiger and Stock (1997) and others have shown that weak instruments
can cause large-sample bias in the IV estimates even when there are multiple instruments. The extent of the bias increases with the number of instruments. They suggest that an F-statistic of less than ten at the first stage is
a cause of concern. They recommend that cutting down on the number
of instruments may help in reducing the large-sample bias. However,
this may not be a useful strategy here as the instruments pass the Hall
and Peixe (2000) instrument redundancy test (see Panel C). Note that the
weak instruments problem is not unique to this study and may well be a
02/03/2011 10:09
Empirical evidence
83
general problem with the empirical comparative development literature.

Stock and Yogo (2005) show that LIML estimators are more robust to
weak instruments than IV. In column 3, I report Fullers modified LIML
estimates with a 5 1 (correction parameter proposed by Hausman et al.,
2005) and I get results similar to IV.31 The magnitude of the coefficient on
malaria risk declines. The model now explains approximately 73 per cent
of the difference in per capita income in Namibia and Nigeria. I choose the
LIML as my preferred estimate since it is the lower bound. The positive
correlation between malaria ecology and the slave trade at the first stage is
certainly noteworthy. This is consistent with the view that the slave trade
was also an outcome of local epidemiology, particularly malaria (see Dias,
1981; Miller, 1982). I also notice that the interior distance is negatively
correlated with colonial institutions. This may be due to the possibility
that proximity to the coast leads to more trade and more trade leads to
better institutions (see Acemoglu et al. 2005b).
Sachs (2003a) predicts 1.6 fold, 1.9 fold and 1.8 fold increases in per
capita GDP due to one standard deviation decline in malaria risk in
AJR, RST and EL samples respectively. Carstensen and Gundlach (2006)
predict a 1.6 fold increase of the same. Both studies are based on a global
sample and they find both institutions and malaria are statistically significant. I find that the malaria effect is even larger (my preferred LIML
estimate predicts a 3.6 fold increase) in an Africa-only sample and all other
factors are statistically insignificant. My results are at odds with the findings of Nunn (2008) who reports that slave exports have a causal effect on
current development in Africa via state development, ethnic fractionalization and weakened legal institutions. I do not find any statistical evidence
of direct and indirect effects of the slave trade on Africas current development. To be completely sure, I also check the robustness of my result
using Nunns specification and dataset (see Table 4.15). The malaria result
survives. I also do not find statistical support for the colonial institutions
view in Africa. This is regardless of the specification and sample.
In column 4, I estimate the causal effect of malaria on growth over the
period 1960 to 2000. The effect is large as one standard deviation reduction
in malaria yields approximately 1.5 per cent growth dividends annually
to an average country in Africa. This suggests that eliminating malaria
alone in 1960 would have resulted in doubling of income in Africa by now.
The relationship between malaria and growth is not surprising as current
income levels and growth in Africa are correlated (approximately 0.7). I
fail to find evidence of causal effects of institutions and the slave trade on
growth.
Tables 4.11, 4.12 and 4.13 report robustness tests with alternative instruments, with fixed effects and with additional covariates. The alternative
02/03/2011 10:09
84
Table 4.11
Dependent
variable
Malaria and African underdevelopment: robustness with

alternative instruments
2SLS
estimate
obs 5 27
(1)
2SLS
estimate
obs 5 27
(2)
2SLS
estimate
obs 5 25
(3)
2SLS
estimate
obs 5 25
(4)
2SLS
estimate
obs 5 27
(5)
2SLS
2SLS
estimate estimate
obs 5 27 obs 5 25
(6)
(7)
Malaria risk
(MALi)
Expropriation
risk in 1985
to 1995
(INSi)
Log total
slave
exports
normalized
by land
area
(SLVXi)
Hansen J-test
(p)
Additional
controls
2.38** 3.6**
3.45** 1.95*
3.35** 6.3
(0.9327) (1.887)
(1.632)
(1.188)
(1.425) (10.70)
0.36
0.33
0.51
0.61
0.29
1.89
(0.5166) (0.5721) (0.7613) (0.6554) (0.5764) (5.629)
Instruments
Replacing Replacing Replacing Replacing

ME by
ME by
ME by
ME by
rain
humidity
frost
rain,
humidity
and frost
0.11
(0.1475)
0.53
0.29
(0.2864)
0.94
0.33
(0.2358)
0.93
0.13
(0.2306)
0.64
LPDi, IDCi
0.24
(0.1989)
0.53
0.85
(2.031)
0.97
0.26
(3.363)
0.52
(0.5102)
0.12
(0.1705)
0.28
LPDi,
IDCi,
rain
LPDi,
LPDi,
IDCi, rain, IDCi,
humidity
rain,
humidity,
frost
ME, LSM, ADC,
IODC, SDC, RDC
a two-sided alternative. Figures in parentheses are cluster standard errors and they are
robust to arbitrary heteroskedasticity and arbitrary intra-group correlation. All the
regressions reported above are carried out with an intercept. P-values of Hansen J-tests are
reported. The null hypothesis is that the instruments are jointly exogenous. The endogenous
regressors are MALi, INSi and SLVXi. The abbreviations used in the table are MEi, malaria
ecology; LSMi, log settler mortality; LPDi, log population density in 1500; IDCi, interior
distance; ADCi Atlantic distance; IODCi, Indian Ocean distance; SDCi, Saharan distance;
and RDCi, Red Sea distance.
instruments strategy is to address the concern that malaria ecology is not

exogenous. The fixed effects and the additional covariates strategies are to
address the omitted variable problem. In Table 4.11 columns 14, I replace
the malaria ecology instrument with geography-based instruments (rain,
02/03/2011 10:09
Empirical evidence
Table 4.12
85
Malaria and African underdevelopment: robustness with fixed

effects
Dependent variable
Malaria risk (MALi)

2SLS estimate
obs 5 27 (1)
2SLS estimate
obs 5 27 (2)
2SLS estimate
obs 5 27 (3)
2.36***
(0.6808)
0.19
(0.4147)
0.09
(0.0915)
1.43
(1.654)
0.23
(0.2615)
0.03
(0.1034)
3.99**
(1.729)
0.07
(0.3751)
0.36
(0.2436)
Expropriation risk in 1985

to 1995 (INSi)
normalised by land area
(SLVXi)
Hansen J-test (p)
0.28
0.09
0.92
Additional controls
LPDi, IDCi
Fixed effects
Colonizer Fixed Region Fixed
Legal Origin
Effects
Effects
Fixed Effects
Instruments
robust to arbitrary heteroskedasticity and arbitrary intra-group correlation. Colonizer fixed
effects, region fixed effects and legal origin fixed effects are dummies representing colonial
origin, region and legal origin, respectively. The endogenous regressors are MALi, INSi
and SLVXi. The abbreviations used in the table are MEi, malaria ecology; LSMi, log settler
mortality; LPDi, log population density in 1500; IDCi, interior distance; ADCi, Atlantic
distance; IODCi, Indian Ocean distance; SDCi, Saharan distance; and RDCi, Red Sea
distance.
humidity and frost) and the malaria result survives.32 One concern is that
rain, humidity and frost may not satisfy the exclusion restriction as they
might influence income through channels other than malaria. To address
this concern, I use these variables as additional controls in columns 57.
The malaria result survives in column 5. The large standard errors and
statistical insignificance of all variables in columns 6 and 7 may be due
to small sample size, degrees of freedom problems and multicollinearity.
The malaria result also survives the inclusion of colonizer fixed effects
and legal origin fixed effects (see columns 1 and 3, Table 4.12). However,
it vanishes when regional fixed effects are added (see column 2). This is
not surprising as I find that the western region indicator dummy and the
eastern region indicator dummy (which are representative of tropical
Africa) predict a negative impact on development. Therefore, it could very
well be the case that these dummies are picking up the negative malaria
effect. Multicollinearity between malaria and the regional dummies can
02/03/2011 10:09
86
02/03/2011 10:09
Mining
3.29**
(1.464)
0.20
(0.4987)
0.25
(0.1943)
0.67
2SLS
estimate
obs 5 27
(1)
1.36*
(0.8251)
0.43
(0.4931)
0.01
(0.1422)
0.08
2SLS
estimate
obs 5 26
(3)
2.39***
(0.8453)
0.02
(0.5876)
0.09
(0.1216)
0.41
LPDi, IDCi
2SLS
estimate
obs 5 26
(4)
2.31***
(0.8267)
0.33
(0.4141)
0.16
(0.1341)
0.45
2SLS
estimate
obs 5 19
(5)
3.22
(2.199)
0.47
(0.6639)
0.32
(0.3083)
0.84
2SLS
estimate
obs 5 27
(6)
Ethnic
Catholicism
Islam
Gini coefficient Foreign aid
fractionalization
3.2**
(1.496)
0.32
(0.5814)
0.29
(0.2359)
0.61
2SLS
estimate
obs 5 27
(2)
Malaria and African underdevelopment: robustness with additional covariates
Schooling
in 1900
1.51***
(0.4838)
0.06
(0.0849)
0.05
(0.0491)
0.95
2SLS
estimate
obs 5 11
(7)
All
instruments
plus CONST
and IDCi
Trade share
2.22***
(0.6315)
0.04
(0.1488)
0.09
(0.0694)
0.42
LPDi
2SLS
estimate
obs 5 26
(8)
Notes: ***, ** and * indicate significance levels of 1%, 5% and 10%, respectively, against a two-sided alternative. Figures in parentheses are
cluster standard errors and they are robust to arbitrary heteroskedasticity and arbitrary intra-group correlation. All the regressions reported above
are carried out with an intercept. The instrument CONST is constructed openness from Frankel and Romer (1999). The endogenous regressors are
MALi, INSi and SLVXi. The abbreviations used in the table are MEi, malaria ecology; LSMi, log settler mortality; LPDi, log population density in
1500; IDCi, interior distance; ADCi, Atlantic distance; IODCi, Indian Ocean distance; SDCi, Saharan distance; and RDCi, Red Sea distance.
Hansen J-test
Control
variables
Additional
covariates
Instruments
SLVXi
INSi
MALi
Dependent
variable
Table 4.13
Empirical evidence
87
also be an issue here as I notice large standard error on the malaria estimate. Alternatively, it may be due to deep cultural or geographic factors
specific to these regions influencing both malaria and income. I am unable
to separate out these effects. The malaria effect also survives the additional covariates test in the majority of cases (seven out of eight) which
are reported in Table 4.13. The additional covariates (mining, ethnic fractionalization, Catholicism, Islam, Gini coefficient, foreign aid, schooling,
trade share)33 are chosen on the basis of previous findings in the literature.
The literature identifies these variables as important correlates of growth
and development. Controlling for all additional covariates together may
not be an option as it weakens the power of statistical tests due to the loss
of degrees of freedom.
Table 4.14 tests the robustness of the malaria result with alternative
measures of institutions and slave exports, and omission of influential
observations. In column 1, I replace the expropriation risk measure of
institutions with Rodrik et al.s (2004) preferred measure, the rule of law
index. I notice that the malaria result survives and the magnitude of the
coefficient is larger than my preferred estimate. In column 2, I replace
it with executive constraints another measure of institutions used by
Acemoglu et al. (2005b) and many others. The malaria result survives in
this case. In column 3, I replace the log slave exports normalized by land
area measure with log slave exports normalized by population. Again, I
notice that the malaria result survives. In column 4, I identify influential
outliers using the DFITS, Cooks distance and Welschs distance formula
(see Belsley et al., 1980) on the OLS regression reported in Panel A,
column 1 of Table 4.10. The DFITS and Cooks distance formula identifies Ethiopia and Gabon as influential observations whereas the Welschs
distance formula identifies Gabon as an influential outlier. I omit these
observations and estimate the model. The malaria coefficient survives the
test. In column 5, I use the DFBETA formula and omit Algeria, Ethiopia,
Gabon and Zambia. The malaria result survives and the coefficient
becomes larger in magnitude.
In Table 4.15, I test the robustness of my malaria result using Nunns
specification and data. In column 1, I estimate Nunns preferred specification34 (see Table 5, column 6 of Nunn, 2008, p. 31). My estimate of
0.20 is marginally different from Nunns 0.188.35 In column 2, I add
malaria into this specification and the statistical significance of the slave
trade variable disappears. Nunn argues that the effect of the slave trade
may be working through institutions. Column 3 checks this possibility by
adding institutions into the mix. The malaria effect survives and neither
institutions nor slave trade are statistically significant. In column 4, I
replace malaria ecology with the geography-based instrument, humidity.
02/03/2011 10:09
88
Table 4.14
Alternative measures and influential observations tests
Dependent
variable
MALi

2SLS
estimate
obs 5 27
(1)
2SLS
estimate
obs 5 25
(2)
2SLS
estimate
obs 5 27
(3)
2SLS
estimate
obs 5 25
(4)
2SLS
estimate
obs 5 23
(5)
4.08**
(1.893)
3.82***
(1.423)
3.97*
(2.262)
0.42
(0.7933)
2.72***
(0.7063)
0.03
(0.2977)
3.92**
(1.793)
0.11
(0.3580)
0.13
(0.0969)
0.25
(0.1990)
INSi
Rule of law
index
Executive
constraint
SLVXi
Log total
slave exports
normalized by
population
Hansen J-test (p)
Controls
Omitted
influential
outliers
Instruments
0.17
(0.7219)
0.37
(0.3051)
0.19
(0.2040)
0.29
(0.1846)
0.48
(0.4206)
0.87
0.47
0.99
LPDi, IDCi
0.32
ETH,
GAB
0.79
DZA, ETH,
GAB, ZMB
robust to arbitrary heteroskedasticity and arbitrary intra-group correlation. Influential
observations are omitted using the following standard rules. In column 4, omit if at least
0 DFITSi 0 . 2
k
4
, 0 Cooksdi 0 . , and 0 Welschdi 0 . 3"k
n
n
holds (see Belsley et al., 1980). In column 5, an additional formula is used which is
uDFBETAiu . 2/!n. Here n is the number of observations and k is the number of
independent variables including the intercept. All the distance formulas are calculated from
the OLS version of the model. The endogenous regressors are MALi, INSi and SLVXi. The
abbreviations used in the table are MEi, malaria ecology; LSMi, log settler mortality; LPDi,
log population density in 1500; IDCi, interior distance; ADCi, Atlantic distance; IODCi,
Indian Ocean distance; SDCi, Saharan distance; and RDCi, Red Sea distance.
02/03/2011 10:09
89
02/03/2011 10:09
Controls
Ethnic
fractionalization
Colonizer fixed
effects
Sargan test
YES
0.20
0.29
0.05
(0.0624)
1.2***
(0.4443)
obs 5 48
(2)
YES
0.20***
(0.0429)
obs 5 52
(1)
0.13
YES
1.67***
(0.5749)
0.04
(0.0864)
0.05
(0.0749)
obs 5 27
(3)
0.19
YES
1.00*
(0.5861)
0.12
(0.1718)
0.05
(0.0721)
0.47
YES
0.04
(0.0567)
0.26
(0.4171)
1.39***
(0.4587)
obs 5 46
(5)
0.57
YES
0.77
(0.7163)
0.11
(0.1403)
2.6*
(1.611)
obs 5 27
(6)
Log income in 2000

obs 5 27
(4)
Robustness with Nunns specification and data
Pre-colonial state
development
Rule of law
SLVXi
INSi
MALi
Dependent
variable
Table 4.15
0.18
0.37
(0.7170)
YES
0.04
(0.0629)
1.42**
(0.6328)
obs 5 48
(7)
Frost
0.70
YES
0.34
(0.2126)
obs 5 43
(8)
Exact
identification
Frost
YES
0.58
(0.7179)
obs 5 25
(9)
90
02/03/2011 10:09
ADC,
IODC,
SDC,
RDC
obs 5 52
(1)
obs 5 25
(9)
LSM
obs 5 43
(8)
ADC,
IODC,
SDC,
RDC
obs 5 48
(7)
ME,
ADC,
IODC,
SDC,
RDC
obs 5 27
(6)
ME,
LSM,
ADC,
IODC,
SDC,
RDC
obs 5 46
(5)
ME,
ADC,
IODC,
SDC,
RDC
obs 5 27
(4)
Humidity,
LSM,
ADC,
IODC,
SDC,
RDC
obs 5 27
(3)
ME,
LSM,
ADC,
IODC,
SDC,
RDC
obs 5 48
(2)
ME,
ADC,
IODC,
SDC,
RDC
Log income in 2000
Notes: ***, ** and * indicate significance levels of 1%, 5% and 10%, respectively, against a two-sided alternative. Figures in parentheses are
cluster standard errors (except for column 1) and they are robust to arbitrary heteroskedasticity and arbitrary intra-group correlation. Colonizer
fixed effects are dummies representing colonial origin. The dependent variable is from Nunn who uses Maddisons figures for per capita GDP in
2000. The endogenous regressors are MALi, INSi, SLVXi and rule of law. The abbreviations used in the table are MEi, malaria ecology; LSMi,
log settler mortality; IDCi, interior distance; ADCi, Atlantic distance; IODCi, Indian Ocean distance; SDCi, Saharan distance; and RDCi, Red Sea
distance.
Instruments
Dependent
variable
Table 4.15 (continued)
Empirical evidence
91
The malaria result survives.36 In columns 57, I check whether the indirect
effects of the slave trade can survive the malaria test. Nunn argues that the
slave trade works through pre-colonial state development, rule of law and
ethnic fractionalization (see Table 8 of Nunn, 2008, p. 37). None of these
variables are statistically significant in the presence of malaria. In columns
8 and 9, I use a more direct approach to test the robustness of the slave
trade result of Nunn (2008) and the institutions result of Acemoglu et al.
(2001). I check what happens to these results when I use frost as an additional control variable. Note that I do not use frost as an instrument to
address the concern that it may not satisfy the exclusion restriction. Also,
note that I choose not to use the controversial malaria ecology variable. It
appears that frost alone is enough to knock off the slave trade and institutions results. This further reinforces the point that the slave trade and
institutions results are extremely weak for the continent of Africa. Malaria
is the only statistically significant variable.
The results are even more unfavourable for the slave trade and institutions if I control for malaria ecology, log population density in 1500, frost,
rainfall and humidity. This result holds if we eliminate log population
density in 1500 from the mix. The only exception is the case when log population density in 1500, frost, rainfall and humidity are used as additional
controls. Slave trade is marginally significant with p-value 0.09. This is not
surprising as it is very close to Nunns original specification. Institutions,
however, are statistically insignificant.
Next, I ask the question: why is malaria so persistent in Africa? The
answer to this question may lie with the mechanism through which
malaria impacts long-term economic performance. In Table 4.16, I report
a strong negative relationship between national savings and malaria in
Africa even after controlling for income. This is perhaps indicative of the
fact that malaria influences long-run development in Africa through the
savings channel. Malaria increases mortality and morbidity. A high mortality rate induces households to save less and consume more. Morbidity
reduces productivity, shrinking the households income and the ability to
save. The result is a low-level equilibrium trap and persistent poverty. This
perhaps helps to explain the persistence of malaria in Africa and also why
malaria is a root cause of African underdevelopment. Chapter 5 explains
this mechanism using an OLG model.
02/03/2011 10:09
92
Table 4.16
Malaria and national savings

S
Model a b 5 1 MALi 1 r log yi + zi
Y i
Dependent
variable
S
Gross savings as percentage of GDP in 2000 a b
Y
OLS Estimate 2SLS Estimate OLS Estimate 2SLS Estimate
obs 5 42
obs 5 42
obs 5 40
obs 5 40
MALi
15.21***
(3.674)
12.29**
(4.997)
log yi
R2
F-Stat
P-value
Instruments
0.30
16.67
0.0002
5.76
0.0211
ME
15.22***
(3.923)
2.58
(2.069)
0.33
7.85
0.0014
12.56**
(5.248)
2.96
(1.919)
3.74
0.033
ME
Notes: ***, ** and * indicate significance levels of 1%, 5% and 10%, respectively, against a
two-sided alternative. Figures in parentheses are cluster standard errors and they are robust
to arbitrary heteroskedasticity and arbitrary intra-group correlation. The endogenous
regressor is MALi. The abbreviation used in the table is MEi, malaria ecology.
NOTES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
Examples of papers using the levels approach are Acemoglu et al. (2001), Glaeser et al.
(2004) and Sachs (2003a).
The growth specification is yiT 5 a ln yit 1 gZiT 1 eiT where yiT ; 1/T [ ln yiT 2 ln yit ] .
Acemoglu et al. (2002) show that around 1500 Western Europe diverged from the rest
of the world in terms of standard of living. Prior to that the living standards were more
or less equal across countries.
A similar point has been made by Caselli et al. (1996) and by Sachs (2003a).
The data on GDP per capita in 1820, 1870, 1900 and 1950 are from Maddison (2004).
For more details see Data Appendix.
The data used is from Rodrik et al. (2004) with data on schooling from Barro and Lee
(2000).
The data used is from Acemoglu et al. (2001) and schooling data is from Barro and Lee
(2000).
The data used are also from Acemoglu and Johnson (2005) and Easterly and Levine
(2003).
See Section IIA, p. 955 of Acemoglu and Johnson (2005) for an explanation on why
constraint on the executive and the legal formalism index are good proxies of property rights institutions and contracting institutions, respectively.
The only exception is the Easterly and Levine (2003) specification where we are able to
isolate the partial effects of institutions and schooling (see Panel A, column 4).
In this sense, the problem is also related to a more general critique of the empirical
growth literature by Levine and Renelt (1992) who show that all growth enhancing
factors are correlated with each other and also with omitted factors, which makes
02/03/2011 10:09
Empirical evidence
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
93
it difficult to estimate the separate effects of these factors on growth using a crosssectional dataset.
The World Bank classifies economies into different income categories using gross
national income per capita. We use this classification to divide our sample into two
groups. The low-income economies (which are the low-and lower-middle-income
economies according to the World Bank list) and the rest are taken to be high-income
economies. The source of this classification: http://www.worldbank.org/data/about
data/errata03/Class.htm.
The observed correlation in the LIE sample could be due to the small sample size or the
observed variation being too limited.
Note that I have summarized some of this literature in Chapter 2. However, it is
perhaps worth refreshing our minds to put this sections empirical exercise into context.
Earlier contributions by historians also suggest that malaria indirectly affected development of the continent by causing massive depopulation in the agriculturally marginal
regions (Dias, 1981; Miller, 1982). They argue that the slave trade was also an outcome
of local epidemiology (particularly malaria) and poor agriculture among other things.
For an alternative view, see Young (2005), who use a calibrated simulation for South
Africa to forecast that survivors of the AIDS epidemic will be economically better off
than they would have been without the epidemic. The intuition in Youngs model is that
women become more cautious about sex due to the fear of infection. As others die out,
female labour becomes more valuable and a consequent reduction in fertility leads to
higher standards of living.
Acemoglu and Johnson (2007) argue that their results are not comparable with the
micro studies as the micro studies do not incorporate general equilibrium effects.
An alternative story of African institutions is from Herbst (2000). He argues that due
to the abundance of land in Africa, there was hardly any competition among precolonial states to defend a well-defined territory. This prevented the development of
state institutions (tax collection, defence, bureaucracy, rule of law and so forth). This
trend of almost no external threat continued during the colonial period. Therefore the
colonizers also had very little incentive to develop good institutions. After independence the situation did not change and what we observe now is the weak institutions of
contemporary Africa.
See Collier and Gunning (1999) for a survey of this literature.
Faced with an increasing demand for slaves from the New World, African demand for
slaves also increased. Africans preferred female slaves whereas young male slaves were
exported across the Atlantic. The result was a huge imbalance in African sex ratio, and
a slow down of demographic transition and economic progress (Manning, 1981).
There wasnt enough labour to support capital and facilitate industrialization in an
already labour scarce continent.
Note that the settler mortality instrument is not free from controversy either. Recently,
Albouy (2008) identified several problems with the construction of the original variable
in Acemoglu et al. (2001) and the revised dataset was published as an MIT mimeo by
the same authors in March 2005 and September 2006. We continue to use the original
variable here to facilitate comparison with all other papers that have used this variable.
I use log population density in 1500 and interior distance as additional control variables
as they might influence current development through other channels.
For more details, see http://www.earth.columbia.edu/articles/view/1932
North (1981) defines good institutions as those that provide checks against expropriation by the government and other politically powerful groups.
These numbers are the aggregate of Atlantic slave trade, Indian Ocean slave trade, Red
Sea and Trans-Saharan slave trade. For more details, see Nunn (2008).
My basic results remain unaffected even if I use these additional instruments.
Detailed information on the construction of the instrument is available online at http://
www.earthinstitute.columbia.edu/articles/view/1932.
Note that including log population density in 1500 and interior distance as additional
02/03/2011 10:09
94
30.
31.
32.
33.
34.
35.
36.

controls does not alter the malaria result in column 1. In fact, institutions and slave
exports become statistically insignificant.
The Hansen J-test is preferred over the Hausman test as it is robust to random or cluster
heteroskedasticity in standard errors.
One could argue that the presence of both the slave trade and institutions in the model
weakens the direct effect of institutions on long-run development. To allay this concern,
I run a direct contest between malaria and institutions leaving out slave trade as a
control. I estimate this model using LIML. Malaria is the clear winner with a coefficient
estimate of 1.37 (se: 0.4510) and institutions are statistically insignificant.
For curiositys sake I also check the robustness of our malaria result using logmort2
from Albouy (2008) as an instrument for institutions instead of Acemoglu et al.s (2001)
settler mortality. The malaria result survives and all other variables are statistically
insignificant.
I also use corruption and the Sachs and Warner openness index as additional covariates. The malaria result survives these tests.
This specification is estimated without interior distance and log population density in
1400 instruments.
Surprisingly, I get very different first stage estimates. I am somewhat puzzled with this
outcome as I am using exactly the same specification, dataset and sample of countries
as Nunn. Not reported here to save space, but available upon request.
Note that the result is qualitatively the same if I use humidity as an additional control
and not as an instrument.
02/03/2011 10:09
5.
Root causes of economic progress: a

unifying framework
The evidence that I have reviewed so far does identify the important
deep factors, namely institutions and diseases that can explain the longrun difference in living standards across nations. This also gives rise to
the apparent conflict between the institutions view and the disease
view. The genesis of the conflict is the statistical significance of the
institutional quality variable in a cross-country regression model estimated at a particular point in time when diseases and other geographic
measures are used as controls. One possible reason behind this empirical
result is perhaps that institutions and diseases are important at different
stages of development and the cross-section regression model is incapable of taking this into account as it solely focuses on a particular point
in time. It could very well be the case that diseases are important at an
early stage of development and institutions become important at an
advanced stage. Bhattacharyya (2009c) presents a case for the existence
of stages of development in the cross-national data. Dividing the crossnational sample into LIEs and HIEs, he finds that diseases explain the
majority of the variation in per capita income in LIEs (which are at an
early stage of development), whereas institutions explain the majority
of the variation in the same in HIEs (which are at an advanced stage of
development).
In this chapter, I make an attempt to build on that finding and explain
the interrelationship between institutions, diseases and economic development by using a method which may not satisfy the purists because of its
somewhat speculative nature. But this method does have some advantages
over the standard cross-country regression modelling approach. It can
throw new light on the complex causality issues by bringing the stages
of development hypothesis, which has been somewhat ignored by the
empirical studies, into the forefront. The strategy is as follows. First, I put
forward a unifying framework, which describes the process of development in Western Europe. This framework lays down the different stages
of development in Western Europe. Second, I compare and contrast the
Western European trajectory with the trajectories in Africa, China, India,
the Americas, Russia and Australia. This allows us to understand where
95
02/03/2011 10:09
96
it went wrong for the other continents. It also allows us to compare and
contrast the stages of development across different continents.
5.1 A UNIFYING FRAMEWORK FOR WESTERN

EUROPE
One could divide the process of economic development in Western Europe
into four different stages. The first stage is the era of the Malthusian cycle
in which geography and epidemic diseases played a crucial part in determining food production. The widely known economic as well as social
impact of the Black Death and other epidemic diseases that hit Western
Europe during the fourteenth century is a testimony of the power of geography and germs. The second stage is characterized by conflict, militarism
and increase in food production and population density. The third stage
is the period of an increased demand for wealth. This epoch is characterized by state investments into daring expeditions to acquire wealth and
resources from foreign lands to finance the soaring costs of war. Finally,
the fourth stage is the stage of change in the nature of the state leading to
rapid technological progress, industrial revolution, mass production and
the rise of the capitalist system.
In order to describe the first stage, one can think of a food production
function which makes use of land, labour, human capital, technology and
climate. Land is fixed and is owned by the small state or group. The entire
population supplies labour except the elite. Human capital is defined
as the knowledge required to use the existing technology successfully.
Technology signifies the development of new tools and it is spasmodic.
Climate is exogenously given. Labour supply is affected by the disease
environment with high incidence of diseases resulting in less labour
supply. In this kind of world, an increase in food production, because of
the positive effects from all of these factors, results in an increase in the
population. Increased population raises the demand for food. However,
there is a limit to what a fixed amount of land can yield given technology,
climate and diminishing returns to labour and human capital. Therefore,
a food crisis ensues and eliminates a large proportion of the population.
This cycle repeats itself in the absence of technological progress. In addition to this natural cycle, food production is also constrained by epidemic
diseases and natural disasters. This is what Thomas Malthus described
as the principle of population in his famous essay in 1750. This process
took place in Western Europe during the thirteenth and the fourteenth
centuries when it was ruled by the small- and medium-sized feudal states.
Robbins (1928) provides evidence of a Europe which is close to what I
02/03/2011 10:09
Root causes of economic progress
97
have described above. In her paper on the impact of the Black Death in
France and England, she records that France was hit by famine on at least
fifteen occasions during the fourteenth century.
In stage two, the production of food may rise due to technological
progress. New technology may evolve due to indigenous effort or due to
technology transfer or can be completely serendipitous. At least in the
case of Western Europe, we know that most of the early technologies were
acquired from the Chinese or the Arabs or from ancient Rome (Mokyr,
1990). There is support for this in Diamond (1997, pp. 40910) as well. He
writes:
Until the proliferation of water mills after about AD 900, Europe west or
north of the Alps contributed nothing of significance to Old World technology or civilization; it was instead a recipient of developments from the eastern
Mediterranean, Fertile Crescent, and China. Even from AD 1000 to 1450
the flow of science and technology was predominantly into Europe from the
Islamic societies stretching from India to North Africa, rather than vice versa.
The arrival of new technology increases agricultural production by

manyfold and creates a situation of food surplus. The food surplus also
increases fertility and reduces mortality, raising the total population.
Rising population puts pressure on land and other resources inducing the
state to get involved in territorial conflicts. This is what we observe during
the age of the Crusades when Europe engaged in repeated conflicts and
wars. The state also gained more in terms of tax revenue in the event of
an increased agricultural yield. A significant proportion of this revenue is
spent in the development of new armoury and the military. The logic is
simple. More lethal weapons and a well-nourished army can win battles
and winning battles is crucial to the very existence of the state. The state
investment in military technology creates positive externalities for civilian
R&D leading to more breakthroughs in technology for agriculture and
crafts. New technology in agriculture and crafts results in steady growth
in output and population, causing more territorial conflict. This pattern
is observed until the fifteenth century in Western Europe when the states
become stronger and stronger. However, the increased frequency of
military conflict did put enormous pressure on the states exchequer and
forced it to look for alternative and richer sources of revenue. Perhaps, this
is what led Western Europe to stage three.
Stage three signifies a state in which investments in maritime expeditions
in order to hunt for alternative sources of wealth and resources become
extremely important. It is also a state which borrows heavily from its subjects to finance expeditions and wars. So, the sometimes heavily indebted
state is more eager to give away political rights to some of its subjects in
02/03/2011 10:09
98
return for loans. This leads to more egalitarian political institutions and
also the development of financial markets. Michael Beaud (2000, p. 14)
describes this process in his book. He writes, Monarchs greedy for greatness and wealth, states battling for supremacy, merchants and bankers
encouraged to enrich themselves: these are forces which inspired trade,
conquests, and wars. These investments lead to the discovery and conquest of new land. The prevailing mercantilist philosophy1 induces explorers to search for bounty in these newly discovered lands and bring it back
to their motherland. This is what Hernan Cortez did when his band of conquistadors came into contact with Montezumas Aztecs in the New World.
Similar was the fate of Atahualpas Inca when Francisco Pizzaros army
of 200 conquistadors defeated them in Cajamarca in 1532. The capture
of Atahualpa by Pizzaros men yielded the largest ransom recorded in
human history. The ceaseless pillage of wealth and precious stones from
the new land triggers inflation in the home country as too much money
chases too few goods. To counter inflation, the state imposes restrictions
on imports, but encourages exports so that it does not run out of wealth.
This policy leads to the expansion of maritime trade and commerce. The
no import ideology also boosts domestic manufacturing, providing it
with a large domestic as well as overseas market. Outward orientation
and trade in manufacturing leads to specialization, division of labour and
increased gains from trade. The nature of the distribution of gains from
trade changes the structure of the political economy and the distribution
of political power. Two distinct patterns emerge. The first is an absolutist
state which takes control of all gains from trade and concentrates political
power. The second is a type of state which allows private accumulation
(such as money lending, trading of precious metals, real estate and so
forth) by the bourgeoisie and hence a relatively equitable distribution of
wealth and political power. In an absolutist state no change occurs in the
institutional structure. However, in the second type of state institutional
changes take place which are favourable to capitalism. The increase in
wealth of the bourgeoisie due to private accumulation provides them with
de facto political power. The bourgeoisie invests in private manufacturing and trade which generates more wealth for the future. This further
strengthens their de facto political power. The bourgeoisie with their new
found political power start demanding institutional change by challenging
the authority of the monarchs. They demand protection of private property and a more equitable distribution of political power. If the de facto
political power of the bourgeoisie is greater than the de jure political power
of the monarchs then the will of the bourgeoisie prevails over the will of
the monarchs. This leads to the establishment of institutions which protect
private property and the political rights of the bourgeoisie. Democratic
02/03/2011 10:09
99
institutions are established to cement the power of the bourgeoisie

and also to make sure that the monarch cannot take over power in the
future.
This pattern of institutional development is observed in Western
Europe from the sixteenth century onwards. The Spanish and Portuguese
monarchs were absolutist in nature and they centralized the process of
manufacturing and trade, discouraging private enterprise (Acemoglu et
al., 2005b). This prevented the development of institutions which provide
incentives to private investment. However, in Britain and the Netherlands,
the state allowed private enterprise, which led to the Civil War in 1642
and the Glorious Revolution in 1688 in Britain and the Dutch War of
Independence which began in the 1570s. Describing the events in Britain
and the Netherlands, Acemoglu et al. (2005b) writes:
The victory of Parliament in the Civil War and after the Glorious Revolution
introduced major checks on royal power and strengthened the rights of merchants. After the Civil War, the fraction of MPs who were merchants increased
dramatically. (p. 564)
...
Dutch merchants always had considerable autonomy and access to profitable
trade opportunities. Nevertheless, prior to the Dutch Revolt, the Netherlands
(in fact, the entire Duchy of Burgundy) was part of the Habsburg Empire, and
the political power of Dutch merchants was limited ... The critical improvement in Dutch political institutions was therefore the establishment of the
independent Dutch Republic, with political dominance and economic security
for merchants, including both the established wealthy regents and the new merchants immigrating from Antwerp and Germany. (p. 566)
Stage four signifies more private as well as state investments in technology, which leads to the development of the factory system and industrial
revolution. The institutional changes of stage three create the ideal incentive structure for private investments in technology development. This
induces rapid technological progress. The rapid improvement in technology increases the cost of moving information relative to the cost of moving
people (Mokyr, 2001). This leads to the rise of the factory system and a
subsequent breakdown of cottage industry. Such a pattern is observed
in Britain and other parts of Western Europe during the period of the
Industrial Revolution (17601830).
Therefore, in sum, the story that I want to get across is as follows.
Western Europe managed to beat the constraints imposed by its geography, in particular diseases, on food production early on and started
a journey on an independent growth trajectory. Availability of food
increased population density which caused territorial conflicts and war.
Ceaseless conflicts induced more investment in military technology. The
02/03/2011 10:09
100
conflicts also put enormous pressure on the finances of the state. The
state commissioned daring naval expeditions to search for bounty so
that it could finance its military expenditure and avarice. These expeditions brought wealth from overseas which also caused inflation. In order
to remedy inflation and also to abide by the principles of mercantilist
philosophy, the state restricted imports of foreign goods and promoted
exports of domestically manufactured goods. This induced specialization
and division of labour in the domestic economy. Institutional changes followed depending upon the initial distribution of the gains from trade. A
non-absolutist state allowed bourgeois accumulation which increased the
power of the bourgeoisie, resulting in major institutional changes favourable to capitalism. In contrast, an absolutist state allowed very little or no
bourgeois accumulation which arrested the prospect of any institutional
change. The states with capitalist institutions attracted private investments in production and technology building. This led to rapid technological progress, the rise of the factory system and industrial revolution.
Therefore, what we learn from the unifying framework is that breaking the disease bottleneck is crucial for future institutional development,
which leads to sustained technological progress and economic growth.
5.2 WHAT WAS DIFFERENT IN AFRICA, CHINA,

INDIA, THE AMERICAS, RUSSIA AND
AUSTRALIA?
5.2.1
Africa
Africa has a long history of diseases. Epidemic diseases such as smallpox,

measles, yellow fever, cholera, tuberculosis, malaria and typhus have
always been a part of African life. Many of these diseases and some new
killers (HIV/AIDS is an example) play a significant role in African life to
date. Africa has also been subject to huge climatic variations. Long dry
seasons were followed by very humid periods with heavy rain (Miller,
1982). These factors have impacted in the past and still continue to impact
Africas growth trajectory. If one seeks an explanation in terms of the
unifying framework that I have outlined in Section 5.1, the obvious question to ask is at what stage did African economics go wrong? My answer is
stage one. Why stage one? The intuitive explanation is as follows.
Geography has always constrained food production in Africa. Long
stretches of drought causing major reductions in cultivation have always
weakened the African population by subjecting them to malnutrition.
Malnutrition made them vulnerable to epidemic diseases. A return of the
02/03/2011 10:09
101
rain also brought diseases along with it, further weakening the labour
force an important input in food production.
Miller (1982) writes, Outbreaks of diseases paralleled the chronology of
drought in an epidemiological sequence familiar from many other regions.
Africans weakened by malnutrition and exhausted by dispersal into the
bush or by flight into lowland became particularly vulnerable to endemic
pathogens (pp. 2223).
A Portuguese observer in eighteenth century Angola commenting on
the increase in disease incidence after the rain writes, Rain brings food
in abundance but leaves no one alive to eat it.2 This situation was further
complicated by the African involvement in the slave trade. Africa had a
long history of slavery as a social institution. However, it was never commercialized on such a large scale prior to the European engagement.3 The
slave trade led to depopulation of the continent, reducing food production
further (Inikori, 1992).4 However, the fact is that without depopulation,
Africa struggled to produce more than a subsistence level of food grains.
This restricted Africa from attaining stages two, three and four and
moving towards the development of a fully home-grown capitalist system.
The engagement with the Europeans during the sixteenth century and
formal colonization during the nineteenth century aborted the independent trajectory of institutional development in Africa. In the colonies with
a high European mortality rate, the colonizers erected extractive institutions. The slave trade encouraged the African elites to go for violent slave
raids inland which institutionalized the culture of violence and lawlessness
in certain parts of the continent. Many of these institutional features have
persisted over time and still exist in the economic and political institutions of modern Africa. These weak institutions continue to influence
the economic performance of the continent. Coupled with diseases and
geographic constraints, poor institutions perhaps explain the bulk of the
African growth tragedy.
Let me introduce an overlapping generation model to explain the story
that I have laid down above. The idea in this model is that the direct
relationship between the rate of time preference and diseases at the representative household level is perhaps the best explanator of the African
situation. In other words, households living in an environment where
malaria incidence and death rate are typically high will choose to save less
for the future and this affects economic progress adversely.
I consider a closed overlapping generation economy consuming and
producing a single homogeneous product. This can be thought of as a
representative economy in the continent of Africa or the whole continent
itself. A typical household in this economy comprises of both young and
old members and each member of the household lives for only two periods.
02/03/2011 10:09
102
The young members of the household work in the first period and retire
in the second period when they are old, and then they die. The members
of this household consume in both periods and the consumption in the
second period is supported by their savings in the first period. Therefore,
at each point in time, members of only two generations are alive. Each
individual within the household maximizes their lifetime utility, which
depends on consumption in the two periods of life. In order to maintain
the simplicity of the structure, I assume away the possibility of bequests or
any altruistic behaviour.
The lifetime utility of a representative individual of generation t can be
expressed as follows:
ut 5
2q
c11t2q 2 1
c12t11
21
1
1 a
ba
b, q . 0, r . 0, [ [ 0, 1 ] (5.1)
12q
11q
12q
where c1tand c2t11 are the consumption of generation t when young and
old, respectively, r is the exogenously determined pure rate of time preference and q is the intertemporal elasticity of substitution.
The survival probability to the old age of the representative individual
depends on the unfavourable geography vector, G.5
5 (G) [ [ 0, 1 ]
(5.2)
G is exogenous to the model and shares an inverse relationship with .

If G is too high then can be too low.
Using Equation 5.2, one can rewrite the lifetime utility of the representative individual as follows:
ut 5
2q
c11t2q 2 1
c12t11
21
1 s (G) a
b
12q
12q
(5.3)
where s (G) 5 (G) /1 1 r is the effective rate of time preference. If G is

too high then s is too low.
The representative individual supplies one unit of labour inelastically when young and receives a wage income wt. Therefore the budget
constraint faced by this individual in period t is given by the following
expression:
c1t 1 st 5 wt
(5.4)
where st is the amount of grain saved in period t. The saved grain also
grows at a rate rt11 when planted. Therefore in period t11 the individual
02/03/2011 10:09
103
consumes the amount saved in period t plus the growth in the grains. So
the consumption in the second period is as follows:
c2t11 5 (1 1 rt11) st
(5.5)
Using Equation 5.5 one can rewrite the budget constraint as

c2t11
5 wt
1 1 rt11
c1t 1
(5.6)
Each individual treats wt and rt11 as given and maximizes their lifetime
utility subject to the budget constraint. This yields the following Euler
equation:
1
c2t11
5 [ s (1 1 rt11) ] q
c1t
(5.7)
Using Equations 5.6 and 5.7 and solving for c1tand c2t11 yields
1
c1t 5 wt [ 1 2
1
2q
s
c2t11 5 wt [
] and
(1 1 rt11) 1 2 q 1 1
1 1 rt11
1
2q
]
1
(5.8)
(1 1 rt11) 1 2 q 1 1
I assume that the representative household uses a Cobb-Douglas

technology to produce the grains and the factors of production are paid
according to their marginal product. The production function is represented as follows:
yt 5 Akat
(5.9)
where yt and kt are per capita output and per capita capital stock, respectively. In this economy, net investment has to be equal to total income less
consumption. Therefore, capital stock in this economy evolves as follows:
Kt11 2 Kt 5 AKat L 1t 2a 2 dKt 2 c1tLt 2 c2tLt21
(5.10)
I assume that the economy starts off with an initial capital stock K1 that
is owned by the L0 elderly in period 1. Also each individual wants to end
02/03/2011 10:09
104
up with no assets when they die. This yields the result that the savings of
the young in this period are equal to the next periods capital stock. Hence
we have the following relationship:
Kt11 5 stLt4t $ 2
(5.11)
In this economy, the population grows at a rate n. Then using the above
relationship and logarithmic preferences6 I get the steady-state value of per
capita capital stock as
k* 5 c
(1 2 a) As
1 2a
d
(1 1 n) (1 1 s)
(5.12)
It is evident from the above equation that a high value of s increases the
steady-state level of per capita capital stock and hence enhances growth.
In Africa, due to the high incidence of killer malaria and other diseases,
one would expect Gto be very high and s to be fairly low. The economy
in this case is in danger of getting trapped into a low income, low capital
stock equilibrium which I have described as stage one of the unified framework. This result is also in line with the empirical findings of Bloom et
al. (2003). They show that economies experience lack of growth not due
to geographical bottlenecks per se but due to the poverty trap situation
emerging out of the bottleneck. This perhaps explains the situation in
Africa which is supported by the empirics in Bhattacharyya (2009b).
Another observation is that if contemporary Africa is stuck at stage one
due to diseases and other geographic constraints then the data is going to
show a strong correlation between the current level of development and
these factors. The correlation between institutions and other factors will
not be visible if it is a poverty trap situation similar to stage one. This is
precisely what the data shows in Chapter 4.
5.2.2
China
The case of China is somewhat surprising. The Chinese were at the forefront of the Old World technology and knowledge till the mid-fifteenth
century. Cast iron, the compass, gunpowder, paper, printing and many
others were first invented in China. The Chinese also invented sophisticated irrigation canals which increased rice production by many-fold
(Diamond, 1997). In light of this long list of technological breakthroughs,
why have the Chinese failed to achieve the same heights as the Western
Europeans? Why did they waste their early technological advantage?
Why was it Britain and not China that progressed towards building an
02/03/2011 10:09
105
industrial society? The answer lies with Chinese institutions. The following paragraph attempts to provide an intuitive explanation in terms of the
broad structure.
Food production developed in China as early as 7500 BC (Diamond,
1997, p. 100). By the start of the millennium, Chinese agriculture was able
to support a large population and the hierarchical structure of Chinese
society was comparable to the social institutions of stage two and three of
the proposed broad structure. One can claim that by the fourteenth and
the fifteenth centuries, China has taken significant steps towards reaching
stage four. The treasure fleets of the early fifteenth century, the discovery
of gunpowder and the compass suggest that the Chinese were incredibly
close to making it to stage four. However, the question remains, what went
wrong? The fate of the treasure fleet after it returned in 1433 gives us a clue
to the answer. After the return of the fleet in 1433, the composition of the
Chinese state changed significantly. The previously powerful eunuchs were
overthrown by their opponents within the Chinese court. This was partly
triggered by Li Zichengs rebellion and the collapse of the Ming Dynasty
into the hands of the Manchu-led Qings. The eunuchs were in favour of
technology, scientific discovery and daring expeditions. Ming rule, under
which the eunuchs were influential, saw a rapid growth in private maritime trade, especially with Portugal and Spain,7 the size of the navy, and
construction projects related to infrastructure (see Ebrey, 1999; Ebrey et
al., 2006).8 The commander of the treasure fleet, Cheng Ho, was himself
a eunuch. When their opponents assumed power, they aborted all the
activities that the eunuchs were involved in, either directly or indirectly.
Gradually, they dismantled the entire infrastructure that was put in place
to encourage these activities. The absolutist nature of their regime also did
not allow private initiatives in these activities. In this way the absolutist
regime destroyed all the institutional incentives for technological research
(Landes, 1998) and China went backwards in the next 500 years. This is a
good example of the theoretical claim that bad institutions can destroy all
the incentives for economic progress even when the region is endowed with
the right geography.
In summary, the Chinese experience shows that escaping the poverty
trap is a necessary, but not sufficient, condition for development. In
other words, a transition from stage two to stages three and four is not
automatic. It also demonstrates that institutions are a deeper cause of
development than technology as technological progress is not sustainable
without strong institutions.9 The contemporary Chinese success story is
perhaps another strong example of institutional reversal, where local level
institutions are more favourable towards business and economic progress
(Rodrik, 2000). How this change happened, however, is a different matter.
02/03/2011 10:09
106
5.2.3
India
Pre-modern India
India escaped the strong grips of the Malthusian cycle (stage one) long
before the British arrived. India was an exporter of industrial goods and an
importer of primary and intermediate goods when Sir Thomas Roe visited
the court of the Mughal emperor Jahangir in 1615.10 After the European
conquest of the Americas, international trade increased manyfold. India
also became an important part of this triangular trade network (Prakash,
2004). Europeans got access to precious metals in South America through
conquest. They traded precious metals (especially silver) in return for
manufactured goods from India. Initially it was the Portuguese who
established a monopoly over this trade but later on they were successfully
challenged by the British and the Dutch East India Companies.
In spite of an increase in trade this period delivered very little in terms
of overall economic growth in India. This could be explained by institutional weaknesses. Monopoly rents of the landed elite were protected
by the state through violence, higher taxes and other forms of coercive
instruments. The state created entry barriers for the outsiders (especially
farmers, merchants and artisans) through higher taxes and violence.
Revenue collected from farmers and artisans in the form of taxes was
largely spent on conspicuous consumption. There was very little investment in public goods such as roads and transportation networks. This
seriously constrained private investments and discouraged technological
innovation. As a result, growth dividends were fairly limited in spite of
an increase in trade. In summary, property rights institutions were weak
because political power was concentrated in the hands of the elite and the
merchants were unable to constrain the actions of the elite. Trade produced very few growth dividends in an institutional environment which
was weak. Bhattacharyya (2011) presents a more detailed account of this
period.
In spite of sluggish growth in pre-modern India some progress was
made during the Mughal period. The structure of the Mughal Empire
was already very hierarchical with power concentrated in the hands of
the minority elites. It also generated enormous amounts of wealth for the
nobility. In support of this fact Landes (1998) writes, India also had a
large and skilled industrial workforce, whose products circulated throughout the region. As a result, the Indian yielded a substantial surplus that
supported rulers and courts of legendary opulence (p. 156).
Therefore, it is perhaps fair to say that the Mughal Indian society
achieved living standards and institutional structure comparable to stage
three of the proposed broad structure. However, this pattern reversed as
02/03/2011 10:09
107
the British started gaining more political control during the late eighteenth
century. The obvious question that one would like to ask is why?
The answer is not as complicated as it may seem. Acemoglu et al. (2002)
talk about an institutional reversal that brought about this change. Their
definition of institutional reversal, however, is very broad. They argue that
the British colonizers never considered India and other tropical colonies
as possible settlements and therefore they erected extractive institutions in
these colonies. These extractive institutions reversed the trend of economic
performance. In the case of India, however, it wasnt only the lack of settlement opportunities that persuaded the British colonizers to erect extractive
institutions. It was also a direct result of the then prevailing political economy
in both countries. Dutt (1992) argues that strong parliamentary lobbying
by the British cotton manufacturers against the import of Indian textile
forced the East India Company to resort to policies which led to a systematic
destruction of the Indian textile industry. He writes, Even in 1813, witness
after witness in the Select Committee of the House of Lords testified that free
Indian textile imports (of both finer and coarser varieties) would damage
British industry (pp. 148149). The British East India Company resorted
to policies of imposing internal tariffs and transit duties on Indian goods,
dislocation and direct exploitation of the artisans, and forceful reduction of
market demand to destroy the industry.11 Indian textiles also lost their overseas market due to the imposition of high import tariffs in Britain.
The company had an influence on the land tenure system and property
rights during that time. In many areas the existing landlords received proprietary rights in land. The company extracted rents from them without
caring much about investment. The landlords passed on this burden of
rent to the farmers and the poor farmers struggled to make investments in
capital and technology. This system of rent-seeking significantly reduced
agricultural productivity and trapped farmers in a vicious cycle of poverty.
One such institution is the Permanent Settlement concluded by the
Cornwallis administration in 1793. It was a grand contract between the
company, the government and the Bengal landlords. Under the contract,
the landlords were admitted into the colonial state system as the absolute
proprietors of landed property and the government was barred from
enhancing its revenue demands from the landlords. This arrangement
institutionalized the alliance between the landlords and the colonial rulers.
It also legitimized rent-seeking. In a recent study, Banerjee and Iyer (2005)
show that these institutional arrangements had and continue to have a
significant impact on economic outcome within India. Areas where proprietary rights were given to the landlords have significantly lower agricultural investments and productivity than areas where rights were given to
the cultivators.
02/03/2011 10:09
108
Therefore, colonization by the British led to institutional reversal which

prevented India from reaching stage four and developing a home-grown
efficient capitalist system. The progressive forces within Indian society
which had the capacity (at least theoretically) to push the economy
towards large-scale industrialization were systematically destroyed by
the existing polity. The domestic extractive institutions were allowed to
continue and it strengthened the feudal landlords both economically and
politically. These institutional changes systematically destroyed the incentives for private investments in land, capital, and technology. As the incentives changed, so did the comparative advantages. India soon became a
net exporter of raw materials and primary products and a net importer of
industrial goods. What ensued is two centuries of deindustrialization and
economic slowdown.
Deindustrialization during the eighteenth and nineteenth centuries
In recent research using the long-run relative price series, Clingingsmith
and Williamson (2008) document that India suffered from deindustrialization during the eighteenth and nineteenth centuries.12 The novelty of
Clingingsmith and Williamsons approach is that they use relative price
data as opposed to employment and output share numbers used in previous studies (see Clark, 1950; Thorner, 1962; Bagchi, 1976a,b; and Prakash,
2005). The key question, however, is why Indian manufacturing suffered
deindustrialization. Bhattacharyya (2011) provides a detailed account
of the series of events that led to deindustrialization. Without going into
detail, I will summarize the economic theory here that may help explain
deindustrialization in India.
Indian manufacturing (especially cotton textiles) faced both supply
side and demand side shocks. On the supply side, frequent climate shocks
(droughts and floods) had a profound impact on productivity. On the
demand side, globalization forces in the form of declining transport
cost and productivity improvements in British manufacturing made it
very difficult for Indian manufacturing to compete in the global market.
Globalization and productivity growth in British manufacturing significantly reduced the market share of Indian manufacturing in the international market. Domestic demand for Indian manufacturing also took a hit
with the disintegration of the Mughal Empire and the subsequent decline
of royal courts who were major buyers of Indian manufacturing.
Indian manufacturing succumbed to these shocks because of weak
institutions. In the chaos that ensued after the disintegration of the
Mughal Empire, local powers increasingly resorted to tax farming and
expropriation of private property to finance a seemingly never ending
conflict (Bhattacharyya, 2011). This led to very little investment from the
02/03/2011 10:09
109
local merchants and artisans in machinery and technology to keep up

with the productivity improvements in British manufacturing. As a result
deindustrialization ensued in the eighteenth and nineteenth centuries. In
summary, Indian manufacturing was unable to withstand shocks because
of weak institutions during the late Mughal period. These rent-seeking
institutions persisted throughout the period when the British East India
Company took control in 1757 after the Battle of Plassey and when the
British Crown took over in 1858. The Permanent Settlement of Bengal
between the East India Company (headed by Lord Cornwallis) and the
Bengali landlords, concluded in 1793, made tax farming legal (Banerjee
and Iyer, 2005). The agreement between the two parties was that the landlords would transfer a fixed magnitude of the revenue raised from land to
the company.
Post-independence India
After independence in 1947 India resorted to a development strategy
heavily reliant on public sector investments. Private property and private
investments were not illegal but they were discouraged. This model of
development was inspired by the success of the Soviet Union at that
time. Bhattacharyya (2011) provides a detailed account of the economic
history of that time. This strategy of development was heavily reliant on
command and control and it brought in layers of inefficient bureaucracy
and regulatory institutions. The culture of rent-seeking became wellentrenched in the system and India experienced very slow growth during
this time. Rodrik and Subramanian (2004) document that Indias growth
rate in per capita income over the period 1950 to 1980 was as low as 1.7 per
cent. This was mainly an outcome of poor regulatory institutions which
encouraged inefficiency and corruption.
1980s policy shift and turnaround
There was good news for the Indian economy after a policy shift in
the 1980s. The Indian economy experienced rapid growth in per capita
income in the 1980s (Rodrik and Subramanian, 2004; De Long, 2003; and
Williamson and Zagha, 2002). The cause of the 1980s turnaround was
mainly attitudinal (Rodrik and Subramanian, 2004). In a gesture demonstrating clear disconnect from the past, the government of the day became
pro-business and encouraged private investments without any significant
change in policy. This accidentally came about when Indira Gandhi was
seeking support from the business community to ward off the electoral
threat from the Janata Party (Kohli, 1989). This in itself was enough to
kick-start growth in the 1980s.
The 1990s reforms were a more serious policy shift (Bajpai and Sachs,
02/03/2011 10:09
110
1999; Ahluwalia, 2002; Srinivasan and Tendulkar, 2003). Trade openness,

economic liberalization and privatization led to improvements in institutional quality. Better institutions and pro-business policies were able to
deliver rapid growth (Bhattacharyya, 2011).
5.2.4
The Americas
When the Europeans first arrived in the Americas in the late fifteenth
century, the indigenous American civilization of the Incas and the Aztecs
was quite developed both economically and politically. The Incas and
the Aztecs developed agriculture, which was capable of supporting large
populations. Their political structures were also very advanced and somewhat similar to the Europeans. The majority of the political power was
concentrated in the hands of the minority elites and the ruling nobility. If
one wants to make a comparison between the then states of the Europeans
and the indigenous Americans, one would be able to point out that there
were certain things that the indigenous Americans were able to achieve
and there were certain things that they failed to achieve. Whatever they
may be, they are secondary to my focus. The important issue is that the
European arrival stalled the independent process of development in the
Americas. The indiscriminate massacre of the indigenous population
and epidemic diseases, such as smallpox, contracted from the Europeans
rapidly reduced the indigenous population to an inconsequential level.
This allowed the Europeans to grab more indigenous land and erect
institutions, which were along the lines of institutions in Western Europe.
However, in the case of the Spanish colonies in South America, the
Spanish colonial rulers continued with the Inca tribute system and other
rent-seeking institutions for their own benefit. Engerman and Sokoloff
(2001) argue that the institutional differences between North and South
America after the European conquest stem from the factor endowment of
the two continents. The following is their theory.
They argue that the factor endowment in South America supported
resource extraction and rent-seeking. Huge reserves of precious metals
supported mining. The climate in many of the southern colonies was
suited to growing sugar, which can be efficiently produced in large plantations. To enjoy economies of scale and extract maximum value, the owners
of mines and plantations employed a large population of slave labour.
These labourers had no rights and no assets. This contributed to the
extreme differences in the distributions of landholding, wealth and political power, which shaped future institutions in the south. In contrast, the
factor endowment in the northern colonies supported small family-sized
farms and farming of grains and livestock. This led to the development
02/03/2011 10:09
111
of a society with relatively equitable distribution of wealth and political

power, and institutions that honoured private property rights. The better
institutions of the North contributed to development as an advanced
capitalist society, whereas for the South it was always a struggle thereafter.
Robinson (2006), in an excellent survey of North American development history, also finds support for the Engerman and Sokoloff theory.
He writes,
British American colonies were founded by entities such as the Virginia
Company and the Providence Island Company whose aim was to make profits.
The model that they had in mind was not so different from that adopted by
the Spanish or Portuguese (a system that other British colonizing entities,
such as the East India Company, used to great effect). Yet these colonies did
not make money and indeed both the Virginia Company and the Providence
Island Company went bankrupt. A colonial model involving the exploitation
of indigenous labor and tribute systems was simply infeasible in these places,
because of lack of large indigenous population and the absence of complex
societies. (p. 28)
A manorial system, as envisaged by Charles I, failed to materialize in

places such as Maryland due to the acute shortage of labour. As a result,
institutions in these settlements ended up giving far more economic
and political rights (access to land, property rights and suffrage) to the
migrants than they originally wished. Therefore, institutions encouraged
private enterprise and investments, which resulted into economic growth
in the long run.
In contrast, the Spanish and the Portuguese colonies of the south were
abundant in indigenous labour and natural resources, which the colonizers used to good effect to set up extractive colonies. These colonies ran
on exploitation of indigenous labour and the native tribute system. After
Pizzaros conquest of Peru, he set up several institutions to extract rent
from the indigenous population (Robinson, 2006). Among these institutions were: (1) encomienda (forced labour), (2) mita (forced labour used
in the mines), and (3) repartimiento (forced selling of goods at a higher
price to the native population). Many of these institutions persisted till
independence and they discouraged private enterprise and investment
throughout. This is perhaps a major reason for the lack of growth and
economic stagnation in Latin America.
In summary, the above discussion shows that two different styles of
colonization policies and, hence, institutions created two different types
of capitalist societies in the Americas after the European conquest. Many
of the old indigenous institutions were replaced after the conquest. The
economic performance of the two continents thereafter depended on the
02/03/2011 10:09
112
new institutions. The Engerman and Sokoloff (2001) theory shows that
the difference in living standards of the two continents can be explained
by institutional differences, which have their roots in the respective factor
endowments.
5.2.5
Russia
It is perhaps appropriate to say that by the start of the nineteenth century

Russia had managed to successfully overcome the Malthusian disease bottleneck. Therefore, it is perhaps best described as an economy operating
in stage three of our unifying framework. However, Russia experienced
a decline relative to Western Europe over the next two centuries. This
is somewhat puzzling given that Russia was not very far away from the
technological breakthroughs and the social changes that were taking place
in Western Europe during that time. Therefore, the key question is what
explains this decline.
To understand, we may have to go back to the nature of resource endowments that Russia had at the start of the nineteenth century. Russia, at that
time, was a predominantly land-abundant and labour-scarce country with
most of its population living in the European parts of the country. It was
also not as mechanized as the Western European economies of that time.
Therefore, the elite held most of their wealth in the form of land and serfs.
The country was virtually run by an alliance between the monarch and the
landed elites. In August 1812, Napoleon Bonaparte invaded Russia from
the west. It is well documented, including in Tolstoys classic novel War and
Peace, that the elite and the peasants fought side by side for the patriotic
cause. This and several other factors led to the belief in some sections of the
liberal nobility that the future of Russia rested on the shoulders of the serfs
and the peasants. Inspired by this idea, a group of liberal noblemen (including Prince Sergey Volkonsky) led a popular, but unsuccessful, uprising on
14 December 1825. Legendary Russian poet Alexander Pushkin was also
sympathetic to this idea and to part of this group, who were known as the
Decemberists. The main idea behind the Decemberist uprising was to grant
more democratic rights to the peasants and the serfs. The landed elite and
the monarch vigorously resisted any such democratic pressure. This is not
surprising in a country where the majority of wealth of the nobility came
from the land. Land being an asset easy to expropriate or impose tax upon,
it is not difficult to understand why the elite felt so threatened by the prospect of relinquishing political power. Therefore, the outcome was a high
level of inequality and weak institutions, which persisted over a century.
This led to very little investment relative to Western Europe in the creation
of market-supporting public goods, such as road or rail networks.
02/03/2011 10:09
113
The lack of growth and high level of inequality persisted over a century
and led to the Bolshevik Revolution in 1917 and the creation of the Soviet
Union. What ensued was an institution reversal of a different kind, where
the state takes control of all private property and investments. These
institutions were not conducive to private investments and markets. The
lack of economic incentives for private individuals was crippling to growth
over the long term and ultimately led to the collapse of the socialist system
in 1991.
In summary, it was not disease or geography that hindered economic
progress in Russia. It was the crippling political institutions during the
nineteenth century Czarist Russia and restrictive political and economic
institutions of the Soviet Union that limited growth over two centuries.
5.2.6
Australia
The initial resource endowments in Australia during the early part of colonization in the 1820s and 1830s were similar to Argentina and other Latin
American countries. There was abundant land with a very small native
population and a minuscule European population consisting of convicts,
ex-military servicemen and the military. Therefore, the key question is why
Australia took such a different course from Argentina over the next two
centuries.
The answer may lie in Australian institutions. By the 1840s, the British
colonizers moved towards granting self-government to Australia. This
step by the colonial rulers was perhaps motivated by the earlier loss of
American colonies and also the problems in Upper and Lower Canada.
The decision to not grant self-rule and more representation in the North
American colonies cost the British colonizers dearly. So they may have
learned a lesson or two from that experience. Therefore, they did not risk
it with Australia. Furthermore, the political trends in Britain may have
also influenced this decision. Nevertheless, Australia established fairly
representative political institutions to support its markets from a very
early stage.
In contrast, the landed interests in Argentina fiercely resisted democratic
pressures. This may have been due to their fear of ownership change and
taxation in the event of democratization (Acemoglu and Robinson, 2006).
In a land-abundant economy, where the elite hold most of their wealth in
the form of land, the elite may strongly resist democratic pressures. This is
because of the relative ease with which land can be taxed or expropriated
by the new regime in the event of a regime change. The resistance is likely
to be fiercer if the agriculture sector is closely aligned with international
trade as it reinforces the political power of the landed elite. This perhaps
02/03/2011 10:09
114
explains why Argentina did not undergo democratization at that time,

which led to a relative decline in living standards, especially during the
twentieth century.
Therefore, to sum up, the key to Australias success is institutions; the
geography bottleneck was not a factor. Better institutions may have also
helped Australia to diversify and industrialize during the twentieth century
in the event of frequent commodity price shocks, which were larger in
magnitude in some cases than those that an average African commodity
exporting country experienced (Bhattacharyya and Williamson, 2009).
5.3 SUMMARY
In summary, the framework presented above shows that diseases are
important at an early stage of development. But as technology coupled
with population growth and some good luck allows a society to escape this
early stage then institutions become important. The interactions of institutions, technology and trade drive the economy to a sustained growth path
thereafter. This is perhaps an appropriate way of describing the process of
economic development in Western Europe.
In China and India, the Malthusian cycle was broken fairly early on and
institutional weaknesses played a crucial role in their respective decline
or stagnation. In the Americas and Australia, colonial institutions were
a crucial factor. In Russia, it was the crippling political institutions of the
nineteenth century and the restrictive political and economic institutions
of the Soviet Union that did the damage.
The African case was somewhat different from the others as the continent struggled to escape from the strong grip of the Malthusian cycle. The
long history of the slave trade and colonial institutions complicated the
story even more later on.
Therefore, the unifying framework does show that there is a case for
dealing with diseases and institutions in the same framework rather than
in isolation. This is a major value added to the literature given that the
existing studies tend to view these two factors as being mutually exclusive,
rather than interlinked.
Even though it is necessary, understanding history is not sufficient for
better policy outcomes in contemporary times. In Part II of this book, I
specifically look at policy outcomes over the last two to three decades. I
ask the questions: whether it is possible to influence institutional quality
through trade policy? Whether trade can improve growth on its own?
Which institutions are important for economic growth? I provide empirical evidence and propose a roadmap for future growth.
02/03/2011 10:09
115
NOTES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
The major argument of the mercantilist philosophy is that a nations wealth depends on
the amount of precious metal it has.
Cited in Miller (1982), p. 23.
The Islamic slave trade started in AD 700, but it never reached the epic proportion of
the Atlantic slave trade.
The historians are yet to reach any agreement on this. For alternative views, see
Lovejoy (1982).
One argument made by Chakraborty and Das (2005) in a recent paper is that the
households can influence by investing in health. However, in this case we assume that
increase in requires huge investment, which is often beyond the scope of a private
investor or a household. This is because malaria is predominantly geographic in nature
and its reduction would require considerable public health intervention.
Logarithmic preference implies q 5 1.
Initially, the Ming court wanted to control trade by using some formal rules. However,
these rules became impossible to sustain with the advent of international trade with the
Europeans. In support of this, Ebrey et al. (2006, p. 277) write, In the sixteenth century,
this formal system [of containing trade] proved unable to contain the emergence of
an international East Asian maritime trading community composed of Japanese,
Portuguese, Spanish, Dutch, and Chinese merchants and adventurers. Because the
profits to be had from maritime trade were high, both open and clandestine trade took
place all along the coast.
Irrigation projects, restoration of the Grand Canal and the Great Wall are some of the
construction projects that the Ming Dynasty undertook.
This is somewhat similar to the Portuguese and the Spanish experience: they failed to
capitalize on their initial technological advantage in maritime trade and shipbuilding
largely due to absolutist institutions (Acemoglu et al., 2005b).
Sir Thomas Roe was the emissary of King James I and he gained for the British the right
to establish a factory at Surat, a port city where the British East India Companys ships
first arrived in India.
According to Dutt (1992), many artisans were subjected to flogging, imprisonment,
and worse. Cutting off the thumbs of winders of raw silk has been documented. The
domestic demand for textiles also reduced significantly due to the decline of the Indian
royal courts, as they were the major buyers of the quality products.
Note that whether deindustrialization took place in India is a long-standing debate in
economic history. We deliberately avoid getting dragged into that debate here. We start
with the premise that deindustrialization took place. For a more detailed account of the
deindustrialization debate, see Bhattacharyya (2011).
02/03/2011 10:09
02/03/2011 10:09
PART II
Promoting growth in the current

environment: evidence and policies
02/03/2011 10:09
02/03/2011 10:09
6.
Institutions and trade: competitors

or complements in economic
development
6.1 THE BACKGROUND

The empirical relationship between trade and development at the crossnational level has been a topic of research for several decades now. Many
policymakers argue that trade openness is beneficial for growth. Outward
orientation of a country promotes efficiency amongst the local firms and
also facilitates technology transfer. Both of these are good for growth. The
root causes literature has reinvigorated the debate over the effectiveness of
policy and trade policy, in particular, in promoting growth.
Until recently, it appeared that a growing academic as well as policy
consensus was emerging on the positive effects of trade on development.
Dollar (1992), using an index of real exchange rate distortion and an
index of real exchange rate variability, shows that outward orientation is
good for economic growth. Sachs and Warner (1995a) construct an index
that combines all aspects of trade policy and show that countries with an
open trade regime, on average, perform better than countries with closed
trade regimes. Ben-David (1993), on the other hand, shows that trade
liberalization leads to less dispersion in income across countries and hence
convergence.
In another influential study, Frankel and Romer (1999) show that there
is a positive relationship between trade volume and national income to
the extent that an increase in trade volume is the result of a reduction in
natural or geographical barriers to trade rather than trade policy. They
use the geographical component of trade volume as an instrument to identify the effects of trade on income.
In the policy arena, the World Bank (2002) emphasizes the advantages
of trade openness, especially for developing economies. In their report
entitled Globalization, Growth and Poverty, they write:
Some 24 developing countries with 3 billion people have doubled their ratio
of trade to income over the past two decades. The rest of the developing world
trades less today than it did 20 years ago. The more globalized developing
119
02/03/2011 10:09
120
countries have increased their per capita growth rate from 1 per cent in the
1960s to 3 per cent in the 1970s, 4 per cent in the 1980s, and 5 per cent in the
1990s ... much of the rest of the developing world with about 2 billion people
is becoming marginalized. Their aggregate growth rate was actually negative
in the 1990s. (pp. 45)
Nevertheless, this growing consensus was shattered by Rodrik and

Rodriguezs (2000) critical survey of the literature. They showed that the
findings of the empirical literature are not robust due to the difficulties in
measuring openness, statistically sensitive specifications, the collinearity
of protectionist policies with other bad policies, and other econometric
problems.1 In an empirical study using data since 1870, Vamvakidis (2002)
finds no support for a positive growth-openness connection before 1970.
In a recent paper, Rodrik et al. (2004) also challenge Frankel and
Romers (1999) result. Using an instrumental variable estimation technique and a cross-country study, they show that institutions dominate the
influence of both trade and geography as the fundamental determinant of
long-run economic development. Rigobon and Rodrik (2005) analyse the
interrelationship between rule of law, democracy, openness and income.
They find that openness negatively impacts on per capita income levels.
The conclusion drawn by Rodrik et al. (2004), however, is different
from Dollar and Kraay (2003). Dollar and Kraay (2003) argue that crosscountry regressions of the log-level of per capita GDP on instrumented
measures of trade and institutional quality are uninformative about the
relative importance of trade and institutions in the long run because of the
very high correlation between the latter two variables. Using an empirical
growth model and panel data, they show that improvements in trade and
institutions have positive effects on growth.
Given the doubts that these studies have created about the empirical
relationship between trade and economic development, further research
on this topic is certainly called for. In this chapter we take a fresh look at
the empirical relationship between trade and development in a model that
also accounts for the effects of institutions on development. Our major
contributions are as follows.
Unlike previous studies which look at the partial effects of trade and
institutions in a linear regression model and hence treat them as competitors in economic development, I join forces with Steve Dowrick
and Jane Golley and consider here the potential for complementarities
between these two variables (see Bhattacharyya et al., 2009). We do this
by introducing an interactive variable in the model which is the product
of our measures of institutional quality and trade openness. We observe
that the coefficient on the interactive variable is positive and statistically
significant, which is indicative of the complementary effects of trade and
02/03/2011 10:09
Institutions and trade
121
institutions on development. We also observe that in order for a country

to benefit from trade, its institutional quality needs to be above a certain
threshold level. This is indicative of the fact that to reap the full rewards
of trade liberalization, countries should also focus on improving their
institutional quality. Using a similar framework, we also test the interrelationship between trade policy openness, institutions and economic
development.
6.2 EMPIRICAL STRATEGY

To examine the potential for complementarities between institutions and
trade share in economic development, we estimate an equation of the
form:
log
y srt 5 ar 1 bt 1 g1TRsrt 1 g2INSsrt 1 g3TRsrt*INSsrt 1 XrsrtL 1 esrt
logy
(6.1)
where log ysrt is the natural logarithm of real GDP per capita purchasing
power parity (PPP) in country s in region r averaged over years t24 to t, ar
represents regional dummy variables controlling for region-specific timeinvariant unobserved heterogeneity,2 bt represents year dummy variables
controlling for time-varying global shocks, TRsrt is the trade share of GDP
in country s in region r averaged over years t24 to t, INSsrt is the quality
of institutions in country s in region r averaged over years t24 to t, Xsrt is a
vector of other control variables, L is a vector of coefficients on additional
control variable Xrsrt and e is a random error term.
One of the major challenges in obtaining unbiased estimates of the
above model is the potential for endogeneity or two-way causality. That
is, while it is a possibility that better institutions and an open trade regime
cause economic development, it is also possible that economic development triggers improvements in institutional quality and trade openness
(see Rodrik et al., 2004). If the latter is true, then a regression based on
Equation 6.1 would spuriously attribute a direct effect of trade openness
and institutions on income that is really due to income affecting them
instead. To tackle this problem, we use the instrumental variable method
of estimation. An instrumental variable has to satisfy the twin conditions
that it is correlated with the suspected endogenous variables (trade share
and institutions) but uncorrelated with the error term in the levels regression.We use log settler mortality, log population density in 1500, fraction
of population speaking English (ENGFRAC), fraction of population
speaking other European languages (EURFRAC), Frankel and Romer
02/03/2011 10:09
122
(1999) constructed openness (CONST), landlocked dummy and land area

as instruments. We confirm below through Hausman and overidentification tests that these are appropriate instruments for the task at hand.
The point estimate of the direct impact of trade openness on development is (g1 1 g3INSsrt) . We expect the coefficient g3 to be positive if there
are complementarities between institutions and trade openness. If g3 is
positive then there are two possible scenarios. First, for a positive estimate
of g1 the net impact of trade openness on development is always positive
given positive values of INSsrt. Second, for a negative estimate of g1 the
net impact of trade openness on development will only be positive when
the value of INSsrt is above a certain threshold level. In other words, only
countries with institutional quality above the threshold level will benefit
from trade.
We estimate another model to uncover potential complementarities
between trade policy and institutions in economic development. The equation that we estimate is as follows:
log
ysrt 5 ar 1 bt 1 y1posrt 1 y2POsrt 1 y3INSsrt 1 y4POsrt*INSsrt
logy
1 XrsrtL 1 xsrt
(6.2)
where posrt is the trade policy openness in country s in region r averaged

over years t24 to t, POsrt is the longer term trade policy measured by the
fraction of open trade policy years in country s in region r since 1950 until
year t and xsrt is a random error term. We also estimate this equation using
the instrumental variable method. The interpretation of the partial effects
in this case is similar to Equation 6.1 discussed above.
6.3 DATA
We use panel data. The master dataset consists of over 209 countries and
covers the period 1950 to 2004. However, our preferred models (columns 4
and 12, Table 6.2) cover 58 countries and the time period 1980 to 1995. A
list of 58 countries is provided in Table 6.5 and there is a maximum of four
data points (1980, 1985, 1990 and 1995) for each country.
The major variables that we use in this study are: log GDP per capita,
trade share, short-run trade policy openness, long-run trade policy openness and expropriation risk. Table 6.1 presents summary statistics of the
major variables and the Data Appendix reports definitions of all variables
used here.
The natural logarithm of GDP per capita PPP (in current international
02/03/2011 10:09
Table 6.1
123
Summary statistics
Variable
Log GDP per

capita (log ysrt)
Trade share
(TRsrt)
Trade policy
openness since
t24 (posrt)
Trade policy
openness since
1950 (POsrt)
Expropriation
risk (INSsrt)
Number Mean Standard

of obs
deviation
(overall)
Standard Standard Minimum Maximum

deviation deviation
(between) (within)
1684
7.67
1.36
1.06
0.87
4.08
10.87
1425
74.20
43.73
39.58
19.25
1.53
362.53
1406
0.46
0.48
0.33
0.36
1384
0.31
0.38
0.34
0.18
445
6.81
2.30
1.84
1.45
10
Notes: For a detailed discussion of the definition and source of these variables, see Data
Appendix.
dollars) is used as a proxy measure of economic development. According to

our sample, the country with the highest per capita income is Luxembourg
in 2004 and the country with the lowest per capita income is China in 1950.
Measuring trade openness is always difficult. We use two measures of
openness in our study. First, the trade share of GDP is used to reflect the
degree of a countrys engagement in trade. This has the obvious advantage of being clearly defined and well measured. However, it does not tell
us anything about why some countries trade more. For example, large
countries have larger internal markets and, therefore, they are more likely
to trade less externally. Therefore, the trade share itself does not tell us
whether this is the case. In our sample, Hong Kong has the highest trade
share and Myanmar has the lowest trade share.
Second, a trade policy measure is used to try and capture a range of
policies that explain why some countries trade more than others. Sachs
and Warners (1995a) trade policy index, which runs from 1950 to 1990, is
the most well-known attempt to quantify trade openness along these lines.
We use this index through to 1990. Wacziarg and Welch (2003) update the
Sachs and Warner (1995a) index and extend it to 2000, which we use for
the 1990s. The Sachs and Warner index is a dummy variable which classifies a country closed (and hence takes the value 0) if any of the following
conditions apply: (1) its average tariff rate on imports of capital or intermediate goods is above 40 per cent; (2) its non tariff barriers cover 40 per
cent or more of its imports of capital and intermediate goods; (3) its black
02/03/2011 10:09
124
market premium is 20 per cent or more; (4) it has a socialist economic

system; or (5) it has a state monopoly on major exports.
Rodrik and Rodriguez (2000) provide the most well-known critique of
Sachs and Warners openness index. They show that the index suffers from
measurement problems and is also correlated with other non-trade-related
bad policies, which make any econometric estimation of its effect on economic development unreliable. However, several recent studies including
one by Rodrik use this index (see Hausmann et al., 2005; Giavazzi and
Tabellini, 2005 and many others).3 We use the Sachs and Warner index
because of its wide coverage (both cross-section and time series) and easy
availability. We also use average tariff, black market premium and real
exchange rate distortions as proxy measures of trade policy. Our results
are not robust to these measures. This is hardly surprising as these measures are only imperfect proxies of trade policy openness (see Wacziarg,
2001; Dollar and Kraay, 2003). They also have a very limited country and
time coverage. In comparison, the Sachs and Warner index appears to be
superior in capturing trade policy openness (Wacziarg, 2001). Also note
that Wacziarg and Welch (2003) accounted for some of the old weaknesses
in the index and significantly improved it in the updated version the
version that we use.
Using the Sachs and Warner index we calculate two indicators of trade
policy openness indicator. First, is an indicator of short-run policy openness which is constructed by dividing the number of years of trade policy
openness between t24 and t by 5. Second, is an indicator of long-run
policy openness which is constructed by dividing the number of years of
trade policy openness between 1950 and t by t21950. The first measure is
expected to be endogenous so we use instrumental variables in our estimation. In contrast, we treat the second measure as exogenous, as one would
expect long-term trade policy going back to the 1950s not to be influenced
by the current level of GDP per capita.
Finally, we use the Political Risk Services index of expropriation risk
as our measure of institutional quality. The measure ranges from zero
to ten where higher values indicate a lower probability of expropriation
of private property by the state. There are other measures of institutions
(such as rule of law, repudiation of contracts, executive constraints, corruption and democracy) used in the literature. However, none of these
measures is close to Douglass Norths notion of good institutions. North
(1981) defines good institutions as those that provide checks against
expropriation by the government and other politically powerful groups
(see North, 1981, pp. 2027). Expropriation risk is perhaps the closest to
Norths definition as it captures the notion of an extractive state. Hence
we use expropriation risk as our measure of institutions. Nevertheless, I
02/03/2011 10:09
125
will also present the empirical estimates of the role of different types of
institutions (rule of law, repudiation of contracts, executive constraints,
corruption and democracy) on development.
6.4 EVIDENCE
This section systematically tests whether institutions and trade are complements in economic development. We first provide the basic results and
then conduct some robustness tests.
6.4.1
Basic Results
Table 6.2 presents the basic results. In column 1 we look solely at the
partial relationship between contemporaneous trade shares and log GDP
per capita. We observe that a one sample standard deviation (43.7 per
cent) increase in the trade share of an average country results in a 1.2
fold increase in per capita GDP.4 This is undoubtedly a large effect and
is statistically significant. However, the coefficient estimate of this model
is a suspect of omitted variable bias. Trade may be correlated with other
factors and with institutions, in particular, which also influence income,
in which case our estimate will show an inflated effect of trade. In order
to tackle this issue, we add institutions in column 2. We observe that both
trade and institutions have positive effects on income and are statistically
significant. The effect of trade on income diminishes from the unconditional estimate of column 1. However, there are still issues of reverse
causality that this model does not address. In column 3 we estimate the
model in column 2 using the IV method to tackle this problem. Our use of
the IV method in this case is statistically valid on the grounds that institutions and trade share variables fail the Hausman test of exogeneity and
the instruments pass the overidentification (OID) test of exogeneity. Note
that institutional quality is the only variable that is statistically significant,
while the coefficient of trade share is insignificantly different from zero.
This confirms previous findings of Rodrik et al. (2004) that institutional
quality is the dominant explanator of long-run variations in economic
development.
The problem with this specification and the specifications adopted by
previous studies is that it implicitly assumes institutions and trade are
competitors in economic development. In column 4 we attempt to address
this issue by introducing an interactive term which is the scalar product of
each countrys trade share and institutional quality to capture any possible
complementarities between the two. This is our preferred model with trade
02/03/2011 10:09
126
02/03/2011 10:09
YES
YES
181
Controls
Region dummies
Year dummies
Countries
POsrt*INSsrt
posrt*INSsrt
TRsrt*INSsrt
(2)
(3)
(4)
117
YES
YES
0.22***
(0.0222)
58
YES
YES
0.78***
(0.1198)
58
YES
YES
0.43**
(0.2171)
0.008**
(0.0037)
0.005*** 0.003***
0.002 0.057**
(0.0005) (0.0008) (0.0026) (0.0285)
(1)
138
YES
YES
1.27***
(0.1062)
(5)
(7)
106
YES
YES
0.21***
(0.0214)
106
YES
YES
0.13***
(0.0374)
0.15***
(0.0266)
0.02 0.88***
(0.0908) (0.2602)
(6)
(8)
138
YES
YES
1.79***
(0.1299)
Log GDP per capita (log ysrt)

(9)
(10)
(12)
0.04
0.68
(0.1039) (0.9398)
(11)
106
YES
YES
0.21***
(0.0202)
106
YES
YES
0.17***
(0.0501)
0.17***
(0.0228)
106
YES
YES
58
YES
YES
0.18*** 1.11**
(0.0507) (0.5064)
0.17***
0.35*
(0.0245) (0.2054)
0.03 1.22*** 1.27*** 8.2**

(0.1385) (0.3740) (0.3965) (3.608)
Institutions and trade: competitors or complements in economic development
Trade share
(TRsrt)
Trade policy
openness since
t24 ( posrt)
Trade policy
openness since
1950 (POsrt)
Expropriation
risk (INSsrt)
Dependent
variable
Table 6.2
127
02/03/2011 10:09
1334
0.9928
393
0.9950
216
0.000
0.92
216
0.000
0.21
1258
0.9652
374
0.9955
374
0.9956
1246
0.9666
374
0.9955
374
0.9956
374
0.9956
0.000
0.31
219
Notes: ***, **, and * indicate significance levels at 1%, 5%, and 10%, respectively, against a two-sided alternative. Figures in parentheses are the
respective standard errors. Columns 3, 4, and 12 report instrumental variable (IV) estimates. The instruments used are log settler mortality, log
population density in 1500, fraction of population speaking English (ENGFRAC), fraction of population speaking other European languages
(EURFRAC), Frankel and Romer (1999) constructed openness (CONST), landlocked dummy and land area. The sample years are every fifth year
from 1950 to 2004. All variables are five-year averages except POsrt, which shows policy openness since 1950. Regressions involving INSsrt cover the
period 1980 to 1995 because of data limitations.
Observations
Adjusted R2
Hausman test
OID test
128
share as the measure of trade.5 We notice that the coefficient estimate of

the trade share is negative and the coefficient estimate of the interactive
term is positive. Both coefficients are statistically significant. The positive
coefficient on the interactive term indicates that institutions and trade are
complements in economic development. The negative and positive signs
on the coefficient estimates of the trade share and the interactive term,
respectively, are indicative of the fact that the relationship between trade
and development is non-linear. There is a threshold level of institutional
quality for an average country beyond which the partial effect of trade
on development is positive. For countries with institutions below this
threshold, the partial effect of trade on development is in fact negative.
Our model predicts that the threshold level of institutional quality is 7.1
which sits well within the sample range of 1 and 10.6 The point estimate
suggests that a one sample standard deviation increase in trade share in a
country with an average institutional quality of 7.2 over the period 1980
to 1995 will lead to a 1.03 fold increase in per capita GDP. To put this
into perspective, the model explains 2.2 fold of the tenfold per capita GDP
difference between India (with a trade share of 21.78 and an institutional
quality of 9.9) and the UK (with a trade share of 55.9 and an institutional
quality of 10) in 1995.
The model predicts a positive impact of trade on development in the
United States, Canada, New Zealand, Australia, Hong Kong, Gambia,
Malaysia, India, Brazil, Papua New Guinea, Chile, South Africa, Ivory
Coast and Mexico in a sample of 58 former colonies. Gabon and
Indonesia are predicted to have small negative effects and Haiti, Mali,
Sudan and Zaire register large negative effects. Table 6.5 and Figure 6.1
show predicted values for all 58 countries (see column TRhat).
In columns 512 we explore the relationship between trade policy openness and economic development. Point estimates reported in columns 57
are unreliable due to endogeneity. To tackle the endogeneity problems
associated with the trade policy measure, we adopt two strategies. First,
we use a long-run trade policy measure (fraction of openness years since
1950) instead of a contemporaneous trade policy measure (fraction of
openness years since t 2 4). We assume the long-run trade policy measure
to be exogenous. Second, we estimate our model using the instrumental
variable method. In column 8 we look at the unconditional correlation
between long-run trade policy openness and economic development. The
point estimate is positive and statistically significant. In column 9 we add
institutions into the specification and we observe that the long-run trade
policy openness variable is no longer statistically significant. In column
10 we note that the coefficient estimate of the interactive term is positive
and statistically significant. In column 11 we add the contemporaneous
02/03/2011 10:09
129
02/03/2011 10:09
.02
ZAR
HTI
MLI
SDN
NGA
BOL
HND
BGD
COG
NIC
SLV
GTM
BFA
MDG
UGA
INS
IND
BRA
PNG
CHL
7.7
ZAF
CIV
MEX
GAB
IDN
TTO
COL
VEN
ECU
PRY
TZA
CRI
JAM
GIN
URY
DZA
CMR
EGY
LKA
KEN
DOM
TUN
PER
TGO
MAR
SLE
GHA
PAN
SEN
ARG
AGO
PAK
NER
Partial effect of trade on development
See Table 8.2 for country abbreviations.
.04
Figure 6.1
Note:
Partial effect of TR on log y
GMB
MYS
HKG
AUS
10
CAN
NZL
US A
130
trade policy openness variable. The model suggests that the long-run trade
policy openness matters and the contemporaneous trade policy openness
variable is no longer significant. This model may suffer from endogeneity.
This is also revealed by the Hausman test of exogeneity of explanatory
variables reported in column 12. The test rejects the null of exogeneity
of the explanatory variables. Therefore, we estimate the model using the
instrumental variable method in column 12. The estimates predict that the
threshold level of institutional quality required for a positive partial effect
of long-run trade policy openness on per capita GDP is 7.4.7 A one standard deviation increase in long-run trade policy openness in a country with
institutional quality 7.5 will result in a 1.1 fold increase in its per capita
income. In this case the model predicts a 2.8 fold difference in per capita
income between India (open trade regime for 2 out of 45 years) and the
UK (always open) in 1995, which is also less than the actual difference of
tenfold. South Africa, Ivory Coast and Mexico are omissions from the list
of countries with a positive partial effect when we use the long-run trade
policy measure of openness. Table 6.5 and Figure 6.2 report predicted
values for all 58 countries (see column POhat).
In Table 6.3 we report the first stage regressions of the IV estimates
reported in columns 4 and 12 of Table 6.2. The instruments that we use
for the IV estimation are valid as they are correlated with the suspected
endogenous variables and also exogenous to the model. We perform an
overidentification test to check the exogeneity of the instruments. The
test p-values are reported in columns 4 and 12 of Table 6.2. The high
p-values indicate that we fail to reject the null of exogeneity of the instruments. Countries with higher constructed openness (CONST) are likely
to trade more (see column 1, Table 6.3 and also Frankel and Romer,
1999, for detailed explanation). Furthermore, countries with a higher
density of population generally trade less externally as there is a larger
market for internal trade (note negative coefficient on LPOPDEN). A
similar explanation may apply for area. The negative partial correlation
between EURFRAC and trade share may be driven by low trade share in
Argentina, the United States and Uruguay. In column 2 we report the first
stage of the institutional quality variable. Countries with high European
settler mortality in the past inherited extractive institutions from their
colonizers and hence the negative relationship between log settler mortality and institutional quality (see Acemoglu et al., 2001). Furthermore,
countries with a high proportion of population speaking a Western
European language (high EURFRAC) are observed to have strong institutions (see Hall and Jones, 1999). The major partial correlations reported
in columns 3, 4 and 5 can also be explained using the abovementioned
arguments.
02/03/2011 10:09
131
02/03/2011 10:09
4.6
ZAR
HTI
MLI
SDN
NGA
BOL
HND
BGD
COG
NIC
SLV
GTM
BFA
MDG
UGA
INS
7.3
ZAF
CIV
MEX
GAB
IDN
TTO
COL
VEN
ECU
PRY
TZA
CRI
JAM
GIN
URY
DZA
CMR
EGY
LKA
KEN
DOM
TUN
PER
TGO
MAR
SLE
GHA
PAN
SEN
ARG
AGO
PAK
NER
Partial effect of long-run trade policy on development
See Table 8.2 for country abbreviations.
6.2
Figure 6.2
Note:
Partial effect of PO on log y
IND
BRA
PNG
CHL
GMB
MYS
HKG
AUS
10
CAN
NZL
USA
132
Table 6.3
Determinants of institutions and trade: first stage for the core

specifications
Dependent
variables:
Constructed
openness
(CONST)
Landlocked
dummy
Log settler
mortality
(LSM)
Log population
density
in 1500
(LPOPDEN)
ENGFRAC
EURFRAC
AREA
Trade Expropriation TRsrt*INSsrt

share
risk (INSsrt)
(TRsrt)
Trade
POsrt*INSsrt
policy
openness
since t24
(posrt)
(1)
(2)
(3)
(4)
(5)
2.05***
(0.2382)
0.015
(0.0115)
16.1***
(1.973)
0.0002
(0.0012)
0.01**
(0.0044)
7.67
(6.989)
1.04
(2.555)
0.53
(0.3379)
0.21*
(0.1235)
74.6
(57.88)
29.9
(21.16)
0.05
(0.0326)
0.008
(0.0155)
0.25*
(0.1393)
0.03
(0.0539)
6.66***
(2.052)
0.13
(0.0992)
66.4***
(16.99)
0.02*
(0.0106)
0.15***
(0.0414)
4.48
(9.948)
23.2**
(11.07)
0.004***
(0.0016)
0.37
(0.4809)
0.98*
(0.5354)
0.0001
(0.0001)
11.5
(82.38)
163.7*
(91.70)
0.03**
(0.0131)
0.01
(0.0611)
0.10
(0.0632)
0.000005
(0.000009)
0.88***
(0.0389)
0.12
(0.1966)
0.02
(0.2204)
0.00002
(0.00003)
6.92***
(0.1710)
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
59
219
0.5365
59
219
0.6100
59
219
0.5820
66
705
0.6742
60
226
0.9535
Trade policy
openness
since 1950
(POsrt)
Controls
Region
dummies
Year dummies
Countries
Observations
Adjusted R2
Notes: ***, **, and * indicate significance levels at 1%, 5%, and 10%, respectively, against
a two-sided alternative. Figures in parentheses are the respective standard errors. The
variables used as instruments are correlated with the suspected endogenous variables which
make them valid instruments.
02/03/2011 10:09
Table 6.4
133
How much does within-country variation matter?
Dependent
variable:
Trade share
(TRsrt)
Trade policy
openness since
t24 (posrt)
Trade policy
openness since
1950 (POsrt)
Expropriation risk
(INSsrt)
TRsrt*INSsrt

(1)
(2)
0.002
(0.0013)
0.0004
(0.0129)
(3)
0.42***
(0.1116)
0.009
(0.0122)
0.00003
(0.0001)
0.42***
(0.0971)
0.0003
(0.0018)
0.01
(0.0136)
0.16*** 0.41
(0.0481) (0.6185)
0.016
(0.0129)
0.34***
(0.1187)
0.05**
(0.0230)
1.13**
(0.4760)
2.12***
(0.6060)
0.05***
(0.0159)
POsrt*INSsrt
Latitude
Countries
Observations
Adjusted R2
(5)
0.83*** 7.6**
(0.2843) (3.423)
posrt*INSsrt
Controls
Country dummies
Region dummies
Year dummies
(4)
2.31***
(0.4664)
YES
NO
YES
NO
YES
YES
YES
NO
YES
YES
NO
YES
NO
YES
YES
117
393
0.9997
28
108
106
374
0.9995
106
374
0.9995
29
112
Notes: ***, **, and * indicate significance levels at 1%, 5%, and 10%, respectively,
against a two-sided alternative. Figures in parentheses are the respective standard errors.
Columns 2 and 5 report instrumental variable (IV) estimates. The instruments used are log
settler mortality, log population density in 1500, fraction of population speaking English
(ENGFRAC), fraction of population speaking other European languages (EURFRAC),
Frankel and Romer (1999) constructed openness (CONST), landlocked dummy and land
area.
In Table 6.4 we explore in greater detail where the complementarities

between trade and institutions are coming from by looking at the source
of identification of the complementarities. If most of the identification is
due to cross-sectional differences between countries that are permanent in
nature then we will not find anything in fixed effect regressions. However,
02/03/2011 10:09
134
Table 6.5
Partial effects of trade on development given institutional

quality
ISO code
Country
INS
TRhat
POhat
USA
CAN
NZL
AUS
HKG
GMB
MYS
IND
BRA
PNG
CHL
ZAF
CIV
MEX
GAB
IDN
COL
TTO
VEN
ECU
PRY
TZA
CRI
JAM
GIN
URY
DZA
CMR
EGY
LKA
KEN
DOM
TUN
PER
TGO
MAR
SLE
GHA
PAN
SEN
ARG
United States
Canada
New Zealand
Australia
Hong Kong
Gambia, The
Malaysia
India
Brazil
Papua New Guinea
Chile
South Africa
Cte dIvoire
Mexico
Gabon
Indonesia
Colombia
Trinidad and Tobago
Venezuela
Ecuador
Paraguay
Tanzania
Costa Rica
Jamaica
Guinea
Uruguay
Algeria
Cameroon
Egypt
Sri Lanka
Kenya
Dominican Republic
Tunisia
Peru
Togo
Morocco
Sierra Leone
Ghana
Panama
Senegal
Argentina
10
9.7
9.7
9.4
8.6
8.3
8.3
7.9
7.8
7.7
7.6
7.3
7.2
7.2
7.1
7.1
7
7
6.9
6.8
6.8
6.8
6.7
6.7
6.6
6.6
6.5
6.3
6.3
6.3
6.2
6.1
6.1
6
6
5.9
5.9
5.8
5.8
5.8
5.7
0.023
0.0206
0.0206
0.0182
0.0118
0.0094
0.0094
0.0062
0.0054
0.0046
0.0038
0.0014
0.0006
0.0006
0.0002
0.0002
0.001
0.001
0.0018
0.0026
0.0026
0.0026
0.0034
0.0034
0.0042
0.0042
0.005
0.0066
0.0066
0.0066
0.0074
0.0082
0.0082
0.009
0.009
0.0098
0.0098
0.0106
0.0106
0.0106
0.0114
2.9
2.567
2.567
2.234
1.346
1.013
1.013
0.569
0.458
0.347
0.236
0.097
0.208
0.208
0.319
0.319
0.43
0.43
0.541
0.652
0.652
0.652
0.763
0.763
0.874
0.874
0.985
1.207
1.207
1.207
1.318
1.429
1.429
1.54
1.54
1.651
1.651
1.762
1.762
1.762
1.873
02/03/2011 10:09
ISO code
Country
PAK
NER
NGA
BOL
HND
BGD
COG
NIC
SLV
GTM
BFA
MDG
UGA
HTI
MLI
SDN
ZAR
Pakistan
Niger
Nigeria
Bolivia
Honduras
Bangladesh
Congo, Republic of
Nicaragua
El Salvador
Guatemala
Burkina Faso
Madagascar
Uganda
Haiti
Mali
Sudan
Congo, Democratic Republic
135
INS
TRhat
POhat
5.6
5.5
5.2
5.1
5.1
5
5
5
4.9
4.8
4.7
4.6
4.6
4.1
4
3.9
3.6
0.0122
0.013
0.0154
0.0162
0.0162
0.017
0.017
0.017
0.0178
0.0186
0.0194
0.0202
0.0202
0.0242
0.025
0.0258
0.0282
1.984
2.095
2.428
2.539
2.539
2.65
2.65
2.65
2.761
2.872
2.983
3.094
3.094
3.649
3.76
3.871
4.204
Notes: TRhats is the partial effect (20.057 1 0.008 3 INSs) of a percentage point increase
in trade share in country s where INSs is the average expropriation risk over the period 1980
to 1995. Similarly, POhati is the partial effect (28.2 1 1.11 3 INSs) of a percentage point
increase in PO in country s.
it could also be because of some common omitted factors such as culture

or geography driving both the complementarities effect and per capita
income. We try to explore this by introducing country fixed effects into our
preferred models. Column 1 of Table 6.4 reports a regression involving the
trade share when country fixed effects are introduced. We observe that the
interactive term TRsrt*INSsrt is no longer statistically significant which may
indicate that permanent cross-country differences are driving the complementarities effect and within-country differences over time do not seem to
matter that much. Geography can be a possible source of identification. In
column 2 we test this by replacing country fixed effects with latitude. We
notice that the coefficient on latitude is positive and statistically significant
and the interactive term is no longer significant. This is perhaps indicative of the fact that geography is driving both the differences in living
standards and complementarities between trade and institutions over the
very long run. This is consistent with previous findings in the literature
that geography and disease environment shape the long-run evolution
of institutions and trade (see Gallup et al., 1998; Acemoglu et al., 2001;
02/03/2011 10:09
136
Rodrik et al., 2004). Column 3 checks the impact of country fixed effects
on the complementarities between contemporaneous trade policy openness and institutions. We find that this effect survives even in the presence
of country fixed effects since the coefficient on posrt*INSsrt is positive and
statistically significant. This holds even when we add long-run trade policy
openness. In column 4 we see that the coefficient on POsrt*INSsrt is positive
and statistically significant. In column 5 we check how much of this effect
is due to permanent cross-country differences such as geography or culture
and how much is due to within-country differences. We do this by replacing country fixed effects with latitude. We observe that the complementarities effect survives, which indicates that some part of this effect is coming
from the within-country variation.
The fundamental difference in the results between the trade share openness model and the trade policy openness models with the inclusion of
country fixed effects and latitude could be because of the difference in the
construction of these two measures. Trade shares perhaps reflect a part of
openness which is deeper and largely time-invariant8 whereas trade policy
is more, influenced by short-term changes in the policy environment.
However, we concede that in the absence of truly exogenous variation, our
analysis does not fully resolve all the identification issues.
6.4.2
Robustness
We check the robustness of our basic results by using additional control

variables (schooling, investment, foreign aid, ethnic fractionalization,
black market premium and mining share of GDP) that are reported by
previous studies to be correlated with development. We also check the
robustness of our basic results across different continental sub-samples.
Our results are reasonably robust with both the trade share and the longrun trade policy as measures of openness. Results are reported in Tables
6.6 and 6.7.
6.5 SUMMARY
The results suggest that there is a reasonably robust correlation between
trade openness and economic development when the complementarities
between institutions and trade are taken into account. This is done by
introducing an interactive variable which is the scalar product of institutional quality and trade into the model. The coefficient on the interactive variable is positive and statistically significant which is indicative
of the complementary effects of trade and institutions on development.
02/03/2011 10:09
Table 6.6
Robustness check: trade share
Dependent
variable
Trade share
(TRsrt)
Expropriation
risk (INSsrt)
TRsrt*INSsrt

(1)
(2)
(3)
(4)
0.04
(0.0301)
0.48*
(0.2852)
0.01**
(0.0039)
0.08*
(0.0499)
0.90**
(0.4149)
0.01*
(0.0066)
0.071**
(0.0309)
0.25
(0.2994)
0.01**
(0.0040)
0.074*
(0.0452)
0.78**
(0.3314)
0.01*
(0.0059)
YES
YES
YES
YES
YES
YES
YES
YES
Controls
Region dummies
Year dummies
Additional
controls
137
(5)
(6)
0.09* 0.08*
(0.0583) (0.0478)
0.82**
0.69**
(0.4262) (0.3612)
0.01*
0.01*
(0.0074) (0.0063)
YES
YES
YES
YES
Schooling Investment Foreign

Ethnic
Black
Mining
aid
fractional- market
ization
premium
Countries
Observations
50
189
58
216
55
204
59
219
55
184
59
219
against a two-sided alternative. Figures in parentheses are the respective standard
errors. All regressions are estimated using the instrumental variable estimation method.
The instruments used are log settler mortality, log population density in 1500, fraction
of population speaking English (ENGFRAC), fraction of population speaking other
European languages (EURFRAC), Frankel and Romer (1999) constructed openness
(CONST), landlocked dummy and land area.
However, questions regarding identification remain due to the absence

of truly exogenous variation in macro trade data. The co-movement in
trade institutions and development may be due to other factors (such
as culture, geography or both) which are driving all three of them. Still,
the fact that the results are robust to the inclusion of schooling, investment, foreign aid, ethnic fractionalization, black market premium and
share of mining to GDP is encouraging enough for an interpretation of
this effect as the combined impact of trade and institutions on economic
development.
It is also observed that in order for the partial effect of trade on development to be positive, a countrys institutional quality has to be above
a threshold level. This is indicative of the fact that trade alone may not
promote development. Institutions play a crucial role.
02/03/2011 10:09
138
Table 6.7
Robustness check: trade policy
Dependent
variable
Trade policy
openness since
(1)
(2)
(3)
(4)
(5)
(6)
0.37
(0.8357)
1.73
(2.402)
0.21
(1.326)
0.72
(1.434)
1.04
(1.964)
1.07
(1.521)
9.1***
(3.458)
23.3*
(12.72)
16.4***
(4.974)
13.1*
(7.788)
17.8** 13.2**
(6.951) (5.802)
0.34
(0.2676)
1.36***
(0.4814)
0.51
(0.4754)
3.3*
(1.773)
0.10
(0.3078)
2.43***
(0.7132)
0.57
(0.3812)
1.8*
(1.076)
0.29
0.53
(0.4641) (0.3240)
2.51**
1.8**
(1.004) (0.8010)
YES
YES
YES
YES
YES
YES
YES
YES
t24
( posrt)
Trade policy
openness since
1950
(POsrt)
Expropriation
risk (INSsrt)
POsrt*INSsrt
Controls
Region dummies
Year dummies
Additional
controls
Schooling Investment
Countries
Observations
50
190
58
217
Foreign
aid
55
205
YES
YES
YES
YES
Ethnic
Black
Mining
fractional- market
ization
premium
59
220
55
182
59
220
against a two-sided alternative. Figures in parentheses are the respective standard
errors. All regressions are estimated using the instrumental variable estimation method.
The instruments used are log settler mortality, log population density in 1500, fraction
of population speaking English (ENGFRAC), fraction of population speaking other
European languages (EURFRAC), Frankel and Romer (1999) constructed openness
(CONST), landlocked dummy and land area.
NOTES
1. Lee et al. (2004) answer Rodrik and Rodriguez (2000). They show that trade has a small
but positive effect on growth when the reverse causality issues are handled appropriately.
2. The region dummies cover Europe and Central Asia, East Asia and the Pacific, Latin
America, the Middle East and North Africa, South Asia and sub-Saharan Africa. Timeinvariant factors such as geography and culture are often region specific.
3. Also see Warner (2003) for a reply to Rodrik and Rodriguez (2000).
4. The partial effect of a one sample standard deviation increase in trade share is
Dlog
y 55log
y11 2
y00 5
Dlogy
logy
2log
logy
5 log
loga
y1
y0
b 5 0.005 3 43.7 < 0.22,
02/03/2011 10:09
5.
6.
7.
8.
139
which can also be expressed as y1 5 e0.22y0 < 1.2y0 where y1 and y0 are final and initial
incomes, respectively.
We also estimate our preferred model using log trade share and the results are qualitatively similar.
The partial effect of trade on development is given by 20.057 1 0.008INST, where INST
is the threshold level of institutional quality. The threshold is calculated by equating the
partial effect to zero and solving for INST. We perform an F-test to check the statistical
significance of the threshold estimate and we fail to reject the null (p-value 5 0.55) of
H0: 2g1 /g3 5 7.1 against a two-sided alternative. We also reject the H0: g1 1 g3 5 0 (pvalue 5 0.048). The coefficient estimate is 20.05 (se: 0.0248) suggesting that an average
country in the sample is not benefiting from trade.
We perform an F-test to check the statistical significance of the threshold estimate and
we fail to reject the null (p-value 5 0.97) of H0: 2y2 /y4 5 7.4 against a two-sided alternative. We also reject the H0: y2 1 y4 5 0 (p-value 5 0.023). The coefficient estimate is
27.1 (se: 3.114) suggesting that an average country in the sample is not benefiting from
trade policy openness.
See Frankel and Romer (1999) for a discussion on this; they argue that an increase in
trade volume reflects a reduction in natural or geographical barriers to trade.
02/03/2011 10:09
7.
Improving institutions with trade

policy: myth or a possibility
7.1 INSTITUTIONS AND TRADE POLICY

As I described in Part I of this book, a growing number of economists
argue that institutional divergence across countries can be best explained
by colonial history. In brief, the story is as follows. The Europeans
resorted to different styles of colonization in different parts of the world
depending on the feasibility of settlement. In tropical climates, the mortality rates among European colonizers were extremely high which prevented them from settling there and they erected extractive institutions
characterized by weakly defined property rights and weak enforcement
of contracts. However, in temperate climates, the mortality rates among
colonizers were low, which made them ideal for settlement and they
erected strong institutions characterized by well-defined property rights
and enforcement of private contracts. These institutions persisted over
time and they continue to influence the current institutional and economic
performance of these countries. Therefore, colonial history shaped institutions in most countries around the globe and perhaps policy has very
little or no role to play.
But the fact that we occasionally do notice improvements in institutional
quality due to good policy suggests that good institutions are not entirely
determined by history. This is an important policy question for growth
given that institutions are a crucial determinant. A good illustration of
this is perhaps post-independence India.1 India inherited relatively good
institutions from the British in 1947. A democratic polity, an independent
judiciary and secure property rights were among many other positives that
were enshrined into the constitution of independent India. High scores in
executive constraint (consistently around 7) and democracy (consistently
around 9) from Polity data during the 1950s are indicative of the fact that
the institutions were strong, at least on paper. India also embarked on
an import substituting industrialization policy during this time, relying
on high tariffs and quantitative restrictions to prevent imports. This led
to the well-known problem of a rent-seeking society (see Bhagwati and
Desai, 1970; Krueger, 1974). As a consequence of the widespread culture
140
02/03/2011 10:09
Improving institutions with trade policy
141
of rent-seeking and bad policy, good institutions, on paper, often yielded

poor institutional outcomes over the next three decades. A quick comparison of the Polity score (which measures institutions on paper) and the
ICRG score (which measures institutional outcome) during 1980 to 2000
supports this view.2 The Polity executive constraint index remained consistently at 7 throughout the period. In contrast the ICRG expropriation
risk index was as low as 6 in 1982 when the import licensing system was
fully operational and it became as high as 10 in 1993 and thereafter when
India liberalized its economy.3
In this chapter, I explore this topic in more detail. I start with a discussion of some of the existing theories linking trade and institutional development. This is followed by evidence from both cross-section and panel
data. The chapter concludes with a discussion of what this may mean for
the policymakers.
7.2 THEORIES OF TRADE AND INSTITUTIONAL

DEVELOPMENT
The theoretical literature on institutional change over time perhaps
originates from North (1981). North (1981) emphasizes that a change
in per capita capital stock due to population growth and technological
progress brings about institutional change over time. What North does
not mention is the impact of international trade on population and technological progress. International trade increases the size of the market
which is equivalent to an increase in the size of domestic population (see
Smith, 1776 [1976]). It is also a widely accepted view that trade induces
technological progress via technology transfer (see Romer, 1990; Coe and
Helpman, 1995). Therefore, potentially, engagement in international trade
can bring about institutional change in a country. In related research,
Rogowski (1989) also shows that trade affects domestic political alignments through changes in factor prices. He, however, does not focus on
the impact of trade on institutions.
Acemoglu et al. (2005b) document historical evidence in favour of the
trade induced institutional change view. Their hypothesis, however, is different from Norths (1981) capital stock theory as they focus on trades
impact on the distribution of political power and subsequent institutional
change. They show that Western Europes engagement in Atlantic trade
induced institutional change by strengthening commercial interests which
resulted in rapid economic growth in countries where the initial political
institutions were non-absolutist.
In a related study, Acemoglu and Robinson (2006) show that trade
02/03/2011 10:09
142
induced transfer of skill-biased technology increases the income share of

the middle class. This increases their political power relative to the rest of
society and they impose checks and balances on the existing institutions to
protect their property rights and contracts. With a larger share of income,
the powerful middle class also favours taxation, which is less redistributive. This makes the elite more willing to accept the checks and balances
on institutions imposed by the middle class.
Without doubt, there is enough theory floating around to believe that
trade liberalization may have an impact on institutions. The key issue,
however, is how much of it is supported by the data. I report some empirical results supporting the link in Section 7.3.
7.3 EVIDENCE
A simple plot of the data in Figure 7.1 shows that indeed there is a positive relationship between long-run trade policy openness and institutional
quality. Long-run trade policy openness is measured in the same manner as
in Chapter 6. It is the fraction of years a country has been open since 1950.
2
CHE
FIN
DNK
NOR
SWE
AUT
AUS
NLD
GBR
CAN
DEU
IRL SGP
USA
NZL
BEL
JPN
FRA
HKG
PRT
ESP
TWN
MUS
ITA
GRC
Fitted values/Rule of law
CHL
SVN
ISR
HUN
EST
CZE
CRI
POL
URY
LTU
LVA
NAM
SVK
MNG
TTO
TUN
LKA
ZAF
HRV
MAR
EGY
IND
BGR TUR
LSO
ROM
GHA
MDG
SEN
CHN
MEX
GAB
BRA
MRT
MWI
VNM
MKD
SYR
BEN PER
DOM
ARM
ETH
SLV
MDA
TZA
NPL
PHL
GMB
ERI
ZMB
DZA
MLI
BFA
IRN
NIC
MOZ
TGO
PAK
ARG
COL
GIN
BGD
NER
RUS
AZE
HND
UKR
PNG
KGZ
GTM
UGA
CAF
KAZ
ALB
TCD
CUB
YUG
GNB
RWA
LAO VENKEN
BLR PRY
TKM
UZB
GEO
CIV
COG
SLE
TJK
CMR
ZWE
NGA
SDN
LBR
BDI
AGO
BUR
HTI
ZAR
KOR
BWA
MYS
JOR
THA
JAM
ECU
BOL
IDN
SOM
0.2
0.4
0.6
Trade openness
0.8
Figure 7.1
Long-run trade policy openness and institutions
02/03/2011 10:09
Table 7.1
143
Summary statistics
Variable
Rule of law
Trade policy
openness
Total years of
schooling
Gini decadal average
Log GDP per capita
in 1960
Number
of obs
Mean
Standard
deviation
175
137
0.002
0.255
1.002
0.399
90
3.489
2.531
111
115
43.24
6.32
10.65
0.865
Minimum Maximum
2.05
0
0.074
21.46
4.72
2.03
1
9.56
64.29
8.23
Hence, by construction, the value of this variable varies between 0 and 1.

Institutional quality, on the other hand, is measured by the rule of law
index. Table 7.1 reports summary statistics of all the important variables.
To explore further, I also estimate a model regressing long-run trade
policy openness on institutional quality measured by rule of law. But there
are, of course, technical challenges to appropriately estimate the effect
of trade policy on institutional quality. It is possible that institutional
quality influences trade policy rather than the other way round. This is
commonly known as the problem of reverse causation. This would lead
to inaccurate estimates from the model. In order to tackle reverse causation, I also estimate the model using the instrumental variable method.
Instrumental variables are exogenous to the model but correlated to trade
policy. Therefore, this allows us estimate the effect reasonably accurately.
Frankel and Romer constructed openness is one such instrumental variable. A detailed discussion on this variable is given in Chapter 4. The results
are reported in Tables 7.2 and 7.3. We notice that one standard deviation
increase in trade policy openness would lead to one standard deviation
increase in institutional quality. This implies that if Tanzania were to
become open then her institutional quality would improve by one-quarter
of the institutional quality of Botswana. This result also holds when I consider a sample of former colonies.
In a related study using panel data covering the periods 18651940 and
19802000, and 31 and 103 countries, respectively, I also find that the
variation in economic institutions (namely property rights and contracts)
can be explained by trade liberalization (Bhattacharyya, 2010). To address
the reverse causality concern I use export partner growth and rainfall as
instruments for trade liberalization and log GDP per capita, respectively,
and estimate our model using the LIML Fuller estimation method. I find
02/03/2011 10:09
144
Table 7.2
Trade policy and institutions

Dependent variable: rule of law
Full sample
(obs 5 72)
OLS
estimate
Trade policy
openness
Other controls
Total years
of schooling
Gini decadal
average
Log GDP per
capita in 1960
R2
F-test (p-value)
Hausman
test for
endogeneity
(p-value)
1.05***
(0.1784)
0.18***
(0.0449)
0.003
(0.0072)
0.12
(0.1214)
0.7974
0.0000***
Full sample
(obs 5 72)
2SLS
estimate
AJR sample
AJR sample
without Neo(obs 5 42)
Europe (obs 5 38)
2SLS
2SLS estimate
estimate
2.05***
(0.5753)
2.93**
(1.302)
2.87**
(1.219)
0.091
(0.0716)
0.013
(0.0122)
0.11
(0.1474)
0.039
(0.1798)
0.012
(0.0179)
0.44
(0.3857)
0.076
(0.2049)
0.001
(0.0170)
0.52
(0.3901)
0.0000***
0.022**
0.0001***
0.013***
a two-sided alternative. Standard errors are reported in parentheses. The p value of the
Hausman test for endogeneity is also reported. It is a t-test in this case. Rejection of the
null in this case indicates that trade policy openness is endogenous. Trade policy openness
is instrumented by Frankel and Romer (1999) constructed openness (CONST) and
landlocked dummy. The AJR sample without neo-Europe excludes Australia, Canada,
New Zealand and United States from the AJR sample.
that one sample standard deviation increase in trade liberalization would

lead to a 2.1 point increase in the property rights institutions index. To
put this into perspective, let us focus on Myanmar. Myanmar is a virtually closed economy with very poor institutions during our sample period.
The average trade liberalization (poit) score and the average property
rights (PRINSit) score over the sample period in Myanmar are 0 and 5.7,
respectively. If the trade liberalization index in Myanmar is to increase by
0.5, then the corresponding increase in the institutional score would be 37
percentage points. I also find similar results for contracting and regulatory institutions. The basic result holds after controlling for country fixed
effects, time-varying common shocks and various additional covariates. It
is also robust to various alternative measures of liberalization and institutions, as well as across different samples.
02/03/2011 10:09
Table 7.3
145
Trade policy and institutions: first stage regressions

Dependent variable: trade policy openness
Full sample
(obs 5 72)
OLS estimate
AJR sample
(obs 5 42)
2SLS estimate
AJR sample without

Neo-Europe
(obs 5 38)
2SLS estimate
0.01**
(0.0031)
0.07
(0.1043)
0.098***
(0.0273)
0.02***
(0.0042)
0.04
(0.0811)
0.5650
0.0000
0.011**
(0.0052)
0.035
(0.1775)
0.13**
(0.0389)
0.01*
(0.0057)
0.16
(0.1249)
0.3904
0.0024
0.01**
(0.0052)
0.05
(0.1753)
0.14***
(0.0490)
0.01
(0.0061)
0.19
(0.1244)
0.2763
0.0552
Constructed openness
(CONST)
Landlocked dummy
Total years of
schooling
Gini decadal average
Log GDP per capita
in 1960
R2
F-test (p-value)
a two-sided alternative. Standard errors are reported in parentheses. The p-value of the
Hausman test for endogeneity is also reported. It is a t-test in this case. Rejection of the
null in this case indicates that trade policy openness is endogenous. Trade policy openness
is instrumented by Frankel and Romer (1999) constructed openness (CONST). The AJR
sample without Neo-Europe excludes Australia, Canada, New Zealand and United States
from the AJR sample.
7.4 SUMMARY
Many studies argue that institutional variation across countries originates
from colonial history. Without disagreeing with that view, I show here
that trade policy may also influence institutional quality within a country.
One could, of course, point towards anecdotal evidence on how trade
policy reforms led to institutional development. This may work through
realignment of political power within the country. This may also work
through transfer of institutions. In this chapter, I present evidence using
cross-section as well as panel data showing that trade liberalization indeed
impacts positively on institutions. The challenge, however, is to take this
beyond the broad framework and work out a detailed understanding of
the channels through which trade liberalization impacts institutions. The
next step would be to identify several channels through which trade liberalization leads to importation of institutions from abroad. Nevertheless,
02/03/2011 10:09
146
we do learn that institutional quality could be improved through good

policy. Better institutional quality, of course, could significantly improve
the growth prospects of a country.
Institutions, however, are by definition a broad construct and it is often
not obvious in policy terms what is meant by improving institutional
quality. In the next chapter, I try to deconstruct institutions into four different types market creating, market regulating, market stabilizing and
market legitimizing institutions. I also identify proxy measures of each of
these institutions and estimate their impact on growth since 1980.
NOTES
1. Other good examples are the Philippines and Mexico. In post-independence Philippines,
restrictive trade policy led to institutional decline till the late 1980s when it was gradually reversed. See Shepherd and Alburo (1991) for a summary of trade policy in postindependence Philippines. Haber (2006) documents that import substitution policies
were adopted in Mexico in the early twentieth century to protect the businesses of the
rich elite aligned with the government of the dictator Porfirio Diaz. This damaged subsequent institutions. Haber (2006) also reports similar consequences of import substitution
in Brazil.
2. In both the Polity and the ICRG indices, higher scores signify better institutional quality.
3. India liberalized its economy in 1991 and the positive effect on expropriation risk started
to show in 1992 when the index score jumped from 6.2 to 8.2.
02/03/2011 10:09
8.
Which institutions matter most for

economic growth?
8.1 UNBUNDLING INSTITUTIONS

Institutions are often defined as a collection of many attributes. The
most commonly used measure of institutions in the institutions and
development literature is the rule of law. The implicit assumption made
in these studies is that better implementation of the rule of law is an indication of the overall quality of institutions. However, from the point of
view of policy, it remains crucial to find out which type of institutions
are important for growth and development. In this section, I make an
attempt to unbundle institutions and estimate their impact on growth. I
adopt Rodriks (2005) four-way classification (market creating, market
regulating, market stabilizing and market legitimizing) of institutions and
include them as explanatory variables of growth in my model. This is in
contrast to other studies in the contemporary literature that focus on the
importance of property rights and contracts in creating the right incentive
structure for investment and entrepreneurship. But long-run economic
growth requires more than just a boost in investment or entrepreneurship. To sustain growth momentum requires institutions that can facilitate
exchange in a world of imperfect information, handle random shocks to
the economy and facilitate socially acceptable burden sharing in the event
of a shock. The four-way classification attempts to cover all aspects of
institutions and also distinguish them on the basis of their functions.
Market-creating institutions are those that protect property rights and
ensure contract enforcement. They are called market creating since, in
their absence, either the markets do not exist or they perform very poorly.
Ideally one would expect expropriation risk or executive constraint to be
good measures of market-creating institutions, but they are not entirely
suitable for our purpose, as we will see below. As an alternative, I choose
the ICRG law and order index. The underlying assumption is that a
country with strong law and order is expected to have better property
rights and contract enforcement. The law subcomponent of the measure
assesses the strength and impartiality of the legal system and the order
subcomponent is an assessment of popular observance of the law. The
147
02/03/2011 10:09
148
assessment is made on a six point scale with a high score implying better
law and order. I observe that 79 per cent of the variation in law and order
is between countries and the remaining 21 per cent is within countries.
Sri Lanka ranks lowest on law and order with a score of 0.3 in 1990 and
Luxembourg ranks highest with a score of 6 throughout the period. One
would expect better law and order would lead to less risk of expropriation
and better contract enforcement.
There are at least four other measures that have been used as a proxy
for market-creating institutions in the literature the rule of law index
of Rodrik et al. (2004), the expropriation risk of Acemoglu et al. (2001),
the executive constraint and the legal formalism index of Acemoglu and
Johnson (2005).1 Acemoglu and Johnson (2005) use executive constraint
as a proxy for property rights institutions and legal formalism as a proxy
for contracting institutions to separately estimate their effects on long-run
growth. Therefore, the obvious question is why I chose the ICRG law and
order index as the preferred measure. Even though far from perfect, there
are at least two clear-cut advantages of using the law and order index.
First, it has a long time dimension running from 1984 to 2004, which is
useful for the dynamic panel data estimation. In contrast, the rule of law
index of Rodrik et al. (2004) and the legal formalism index of Acemoglu
and Johnson (2005) are only available in cross-section and the expropriation risk measure of Acemoglu et al. (2001) only covers the period 1982
to 1997 which is too short a time dimension for estimating our dynamic
panel model. Second, it can be counted upon as a measure that represents
both property rights and contracting institutions. I do, however, check the
robustness of the result using alternative measures of institutions.
The other question is why I chose not to separate out the effects of
property rights and contracts. This is largely due to the unavailability of
a reasonable proxy for contracting institutions. Acemoglu and Johnson
(2005) argue that an ideal measure of contracting institutions is the cost of
enforcing private contracts. They use the legal formalism index as a proxy.
This measure, however, is not available in a panel. As an alternative, I use
repudiation of government contracts, which measures the cost of enforcing
government contracts and is far from ideal.2 It primarily focuses on institutions that define the relationship between the state and its subjects and
not on institutions that provide the legal framework that enables private
contracts to facilitate economic transactions. This measure also drastically
reduces the sample size. Nevertheless, I report these findings in Section 8.3.
Market regulating institutions are those that prevent market failure
and help to sustain growth momentum over the long run. Rodrik
and Subramanian (2003) mention regulatory agencies in telecommunication, transport and financial services as examples of market-regulating
02/03/2011 10:09
149
institutions. I proxy this by using Gwartney and Lawsons (2005) composite index of regulation (MR) in the credit market, labour market and
business in general since this is the closest to Rodrik and Subramanians
examples. The index operates on an 11 point scale ranging from 0 to 10
with a high score implying fewer regulations. I notice that 67 per cent of
the variation in the measure is between countries. Romania records the
lowest level of regulation and New Zealand records the highest level of
regulation in the dataset. This is the only proxy of regulatory institutions
that I could locate which has the desirable time dimension.
Market-stabilizing institutions are those that build resilience towards
shocks, reduce inflationary pressure, minimize macroeconomic volatility and avert financial crises. Examples of market stabilizing institutions
include central banks, exchange rate regimes and budgetary and fiscal
rules. Finding a suitable proxy for this is a challenge. I think that it is not
unreasonable to assume that good market-stabilizing institutions (independent central banks) do look to minimize inflationary pressure and volatility in the long run.3 Hence, I use Gwartney and Lawsons (2005) sound
money index (SM) as a proxy which takes into account: (1) average annual
growth of the money supply in the last five years minus average annual
growth of real GDP in the last ten years, (2) standard inflation variability
in the last five years, and (3) recent inflation rate. This index also ranges
from 0 to 10 with a high score implying better market-stabilizing institutions. I observe that 56 per cent of the variation in market-stabilizing institutions is between countries. Canada, France, Denmark and Switzerland
are among countries with very strong market-stabilizing institutions,
whereas some of the Latin American countries, including Argentina and
Brazil, fare among the worst especially during the 1980s.
Market-legitimizing institutions are those that handle redistribution,
manage social conflict and provide social protection and insurance in the
event of a shock. In other words, they help to minimize the idiosyncratic
risk to economic growth and employment in a modern market economy.
Rodrik (2005) suggests democracy as a proxy for market-legitimizing
institutions as there is a positive relationship between the effectiveness
of democratic institutions and the quality of social insurance. Therefore,
the Polity IV democracy index suits such a purpose as it measures the
effectiveness of democratic institutions by capturing different shades of
democracy and it ranges from 0 to 10 with a high score implying a more
democratic system. Among the most democratic nations in the dataset
are the United States, Japan and the United Kingdom, and among the
least democratic are Morocco, Nigeria and Equatorial Guinea. A total
of 78 per cent of the variation in this series is between countries and 22
per cent of the variation is within countries. Several existing studies (see
02/03/2011 10:09
150
Barro, 1996; Rodrik, 1999; Tavares and Wacziarg, 2001; Acemoglu et al.,
2005c; Acemoglu et al., 2008; Bhattacharyya and Hodler, 2010) use this as
a measure of democracy. I also use the freedom house political rights index
as an alternative measure of market-legitimizing institutions.4
There is also a strong view in this literature that a simple dichotomy
between democracy and non-democracy is the most appropriate empirical
definition (see Przeworski et al., 2000). However, this definition is not suitable for our purpose as it does not provide any information on the effectiveness of democratic institutions. Recently, Persson and Tabellini (2006)
used a similar empirical definition and looked at the impact of regime
change and different styles of democracy on within-country growth. They
found that countries liberalizing their economy before extending political
rights do better in terms of growth than countries following the opposite
sequence. Their measure also focuses on democratic transitions and hence
is not suitable for our purpose.
Finally, there is a view in the literature that corruption should be
treated as an institution. Hall and Jones (1999) and Knack and Keefer
(1995) include corruption in their overall measure of institutions. I am,
however, slightly sceptical about classifying corruption as an institution.
In my view, corruption is the outcome of poor market-creating institutions
rather than an institution itself. This is perhaps reflected by the measure of
corruption that Mauro (1995) uses. Mauro (1995) uses an indicator that
reflects experts assessments of the degree to which business transactions
involve corruption or questionable payments averaged over the period
198083. This is an index of perception and is reflective of institutional
outcome rather than institutions themselves. Nevertheless, I do include
ICRG index of corruption as an additional control variable (see Table 8.6,
column 3) and my basic results are robust to this inclusion.
8.2 DATA
The dataset includes measures of schooling, institutions, per capita GDP
growth, per capita GDP levels and trade share. There are six time points
in the data, spanning the period 19802004, with each time point being
approximately five years on average. There are missing observations in the
data and hence I have an unbalanced panel.
I follow the literature and proxy human capital using schooling data
from Barro and Lee (2000). It measures the average schooling years in
the total population. A total of 96 per cent of the variation in schooling is
between countries.
As indicated in the previous section, I use four different measures of
02/03/2011 10:09
Table 8.1
151
Summary statistics
Variable
Number
of obs
Economic growth and development

Growth (yiT)
629
Initial income (yiT25)
629
Market creating institutions
Law and order (LO)
507
Market regulating institutions
Regulation of credit,
616
labour, business (MR)
Market stabilizing institutions
Sound money Index (SM)
634
Market legitimizing institutions
Democracy index
659
(DEMOC)
Schooling
Total years of schooling
482
(TYS)
Trade openness
Log trade share of GDP
754
(LTRS)
Mean
Standard Minimum Maximum

deviation
0.014
8.4
0.031
1.13
0.10
6.01
0.35
10.88
3.7
1.5
0.3
5.6
1.1
2.5
8.8
7.5
2.5
9.8
5.0
4.1
10
5.4
2.9
0.4
12.3
4.19
0.59
2.43
6.01
institutions. There are a number of conceptual and empirical challenges

that I have to face in the analysis. It is hard to rule out potential overlaps
between these measures of institutions. The pairwise correlations reported
in Table 8.3 show that the measures of institutions are correlated. The correlation between law and order and democracy is 0.5321 and the correlation between democracy and the regulation of credit, labour and business
is 0.5066. However, none of the correlations is large enough to cause any
serious problems of multicollinearity.
The GDP per capita PPP measured at a constant 2000 international
dollars is obtained from the World Development Indicator (WDI). The
WDI data provides larger time coverage and allows me to extend the study
to 2004. Annualized growth rates of GDP per capita are calculated using
the formula yiT ; 1/5 ( ln yiT 2 ln yiT 25) .
The trade share measure is also from WDI and is widely used (see
Dollar and Kraay, 2003; Rodrik et al., 2004). It is perhaps the simplest
measures of trade openness where trade is expressed as a share of GDP.
Table 8.1 presents the descriptive statistics for the important variables
used in the study and Table 8.2 presents a list of countries.
02/03/2011 10:09
152
Table 8.2
List of countries
Code
Country
Code
Country
Code
Country
AGO
ARG
ATG
AUS
AUT
BDI
BEL
BEN
BFA
Angola
Argentina
Antigua
Australia
Austria
Burundi
Belgium
Benin
Burkina
Faso
Bangladesh
Belize
Bolivia
Brazil
Barbados
Botswana
Central
African
Republic
Canada
Switzerland
Chile
China
Cte
dIvoire
Cameroon
Congo, Republic
of
Colombia
Comoros
Cape Verde
Costa Rica
Cyprus
Dominica
Denmark
Dominican
Republic
Algeria
Ecuador
Egypt
Spain
Ethiopia
Finland
Fiji
France
Gabon
United
Kingdom
Germany
Ghana
GIN
GMB
Guinea
Gambia,
The
GuineaBissau
Equatorial
Guinea
Greece
Grenada
Guatemala
Guyana
Hong
Kong
Honduras
Haiti
Hungary
Indonesia
India
Ireland
Iran
Iceland
Israel
Italy
Jamaica
Jordan
Japan
Kenya
St. Kitts
& Nevis
Korea,
Republic
of
St. Lucia
Sri Lanka
Lesotho
Luxembourg
Morocco
Madagascar
Mexico
Mali
Mozambique
Mauritania
Mauritius
Malawi
Malaysia
Namibia
Niger
Nigeria
Nicaragua
Netherlands
NOR
NPL
NZL
Norway
Nepal
New
Zealand
Pakistan
Panama
Peru
Philippines
Papua New
Guinea
Poland
Puerto Rico
Portugal
Paraguay
Romania
Rwanda
Senegal
Singapore
Sierra Leone
El Salvador
Sao Tome
and Principe
Slovak
Republic
Sweden
Swaziland
Seychelles
Syria
Chad
Togo
Thailand
Trinidad
&Tobago
Tunisia
Turkey
Taiwan
Tanzania
Uganda
Uruguay
USA
St.Vincent
&
Grenadines
Venezuela
South
Africa
Congo,
Dem. Rep.
Zambia
Zimbabwe
BGD
BLZ
BOL
BRA
BRB
BWA
CAF
CAN
CHE
CHL
CHN
CIV
CMR
COG
COL
COM
CPV
CRI
CYP
DMA
DNK
DOM
DZA
ECU
EGY
ESP
ETH
FIN
FJI
FRA
GAB
GBR
GER
GHA
GNB
GNQ
GRC
GRD
GTM
GUY
HKG
HND
HTI
HUN
IDN
IND
IRL
IRN
ISL
ISR
ITA
JAM
JOR
JPN
KEN
KNA
KOR
LCA
LKA
LSO
LUX
MAR
MDG
MEX
MLI
MOZ
MRT
MUS
MWI
MYS
NAM
NER
NGA
NIC
NLD
PAK
PAN
PER
PHL
PNG
POL
PRI
PRT
PRY
ROM
RWA
SEN
SGP
SLE
SLV
STP
SVK
SWE
SWZ
SYC
SYR
TCD
TGO
THA
TTO
TUN
TUR
TWN
TZA
UGA
URY
USA
VCT
VEN
ZAF
ZAR
ZMB
ZWE
02/03/2011 10:09
153
8.3 EVIDENCE
Table 8.4 reports the dynamic regressions. They are estimated using the
Generalized Method of Moments (GMM) Blundell and Bond estimator.
In column 1, I start with a simple specification regressing law and order
and schooling on growth. The other control variables are initial income
and trade. In this specification, law and order is statistically significant,
but schooling is not. This may be due to endogeneity of the instruments
as they fail the Hansen J-test (p-value 0.006). In column 2, I replace law
and order by MR and I observe that schooling is statistically significant,
but MR is statistically insignificant. In column 3, I replace MR with SM.
In this specification, both schooling and SM are statistically significant.
In column 4, I replace SM with democracy. Schooling is still statistically
significant, but democracy is statistically insignificant. In column 5, I estimate my preferred model with a full set of control variables. I add schooling, law and order, MR, SM, democracy, trade and initial income into the
specification. I find that schooling, law and order and SM are statistically
significant and all institution measures are jointly significant. One sample
standard deviation increase in schooling increases the annual growth rate
in an average country by 1.7 percentage points. MR and democracy do not
seem to matter. Comparable increases in law and order and SM both have
growth effects of 0.75 percentage points. In column 6, I make an attempt
to unbundle market creating institutions and separate out the effects of
property rights and contracting institutions. I use executive constraint
from the Polity IV dataset as a measure of property rights institutions. In
the absence of time series data on the legal formalism index, I use repudiation of government contracts as an alternative measure of contracting
institutions, which, of course, is far from ideal. I notice that both property
rights and contracts have positive and statistically significant effects on
growth. This is in contrast to Acemoglu and Johnson (2005), who report
that property rights matter more than contracts for growth in the long
run. The magnitude of the coefficient on schooling remains unchanged
and is statistically significant. However, a major drawback of this specification is that it drastically reduces the sample size. In this case, T 5 2 and
the estimates are equivalent to 2SLS and they can be poorly identified or
suffer from weak instrument problems (Wooldridge, 2002). In order to
explore the possibility of non-linearity in MR, I introduce MR2 in column
7. The public interest theory of Pigou (1938) suggests that unregulated
markets are comparatively more prone to failure and should be associated with socially inferior outcomes. In contrast, the public choice theory
of Shleifer and Vishny (1998) suggests that regulation leads to corruption
and hence is harmful for development.5 Therefore, it is possible that there
02/03/2011 10:09
154
02/03/2011 10:09
1.0000
0.7243
0.6824
0.2753
0.4437
0.8511
0.1250
0.3075
0.1812
0.3071
0.1729
Initial
income
(yiT25)
1.0000
0.1887
0.3631
0.1639
Growth (yiT)
0.1969
0.7217
0.4287
0.3770
1.0000
0.5321
0.0626
0.6237
0.5066
0.1796
1.0000
Law and Democracy

order (LO) (DEMOC)
0.2707
0.2746
0.4089
1.0000
Sound
money
index (SM)
Unbundled institutions, human capital and growth: pairwise correlation
Growth (yiT)
Initial income (yiT25)
Law and order (LO)
Democracy
(DEMOC)
Sound money index
(SM)
labour, business
(MR)
Total years of
schooling (TYS)
Log trade share of
GDP (LTRS)
Table 8.3
0.2740
0.4899
1.0000
0.1497
1.0000
Regulation Total years

of credit, of schooling
(TYS)
labour,
business
(MR)
1.0000
Log trade
share
of GDP
(LTRS)
155
02/03/2011 10:09
Other controls
Initial
income (yiT25)
DEMOC2
Executive constraint
(EXCONST)
Repudiation of govt
contracts (REPU)
labour, business
(MR)
Sound money index
(SM)
Democracy index
(DEMOC)
Total years of
schooling (TYS)
MR2
YES
0.004
(0.0029)
0.004**
(0.0020)
(1)
YES
0.007**
(0.0034)
0.001
(0.0029)
(2)
YES
0.005*
(0.0028)
0.003***
(0.0009)
(3)
YES
0.0004
(0.0011)
0.01**
(0.0038)
(4)
YES
0.003***
(0.0009)
0.0003
(0.0009)
0.006**
(0.0026)
0.003
(0.0024)
0.005**
(0.0020)
(5)
0.005**
(0.0020)
(7)
YES
0.002
(0.0013)
0.007***
(0.0026)
0.006*
(0.0034)
0.003
(0.0025)
0.005***
(0.0018)
(8)
0.023
(0.0148)
0.005***
(0.0018)
(9)
YES
YES
YES
0.002**
0.003*** 0.002**
(0.0009)
(0.0009)
(0.0009)
0.001
0.001
0.0003
(0.0010)
(0.0021)
(0.0021)
0.007*** 0.005**
0.006***
(0.0024)
(0.0025)
(0.0024)
0.002*
0.002*
(0.0013)
(0.0013)
0.0001
0.0001
(0.0002)
(0.0002)
0.012***
(0.0042)
0.007***
(0.0017)
0.006*
0.023
(0.0036)
(0.0153)
(6)
Dependent variable: annualized growth (yiT)
Unbundled institutions, human capital and growth: dynamic regressions
Law and order (LO)
Table 8.4
156
02/03/2011 10:09
(continued)
363/93
0.325
0.291
348/87
0.109
0.001
YES
YES
0.060
0.001
(2)
(1)
369/93
0.032
0.009
0.000
YES
(3)
370/95
0.251
0.157
0.001
YES
(4)
323/83
0.174
0.351
0.001
0.0027
YES
(5)
242/83
0.480
0.034
0.0000
YES
(6)
(7)
323/83
0.166
0.670
0.002
0.0079
YES
323/83
0.175
0.776
0.001
0.0030
YES
(8)
323/83
0.174
0.866
0.002
0.0044
YES
(9)
Notes: ***, ** and * indicate significance levels of 1%, 5% and 10%, respectively, against a two-sided alternative. Standard errors are reported
in parentheses. Hansen test is the test of the H0: the instruments as a group are exogenous. Hansen test statistic from two-step Arellano and Bond
estimations is reported, which is robust to heteroskedasticity or autocorrelation. Arellano and Bond AR(1) and AR(2) tests in residuals are also
reported. Note that to pass these tests, one has to reject the null of no AR(1) and fail to reject the null of no AR(2).
Observations/countries
Specification tests
(P-values)
Joint F-test
(Institutions)
Hansen test
Test for AR(1) in
residuals
Test for AR(2) in
residuals
Log trade share

(LTRS)
Table 8.4
157
is an optimum level of regulation. In other words, the relationship between

regulation and growth is non-linear and hence it justifies our inclusion of
MR2 in column 7. I observe that the coefficient on MR2 is negative and statistically significant, but the coefficient on MR is positive and statistically
insignificant. This is perhaps because of the non-linear effect of regulation.
In other words, perhaps there is a particular level of regulation that maximizes growth. This level turns out to be 4.9 when I equate the partial derivative of growth with respect to MR to zero.6 One possible interpretation of
this result is that too much or too little regulation is not good for growth.
Too little regulation encourages anti-competitive behaviour among firms
and can lead to market failure. Too much regulation, on the other hand,
can lead to red tape which has tangible costs to the economy. However,
I treat this result with caution as the coefficient on MR is statistically
insignificant. In column 8, I make an attempt to check for non-linearity in
democracy by adding democracy2 into the preferred model. I find that the
statistical significance of schooling, law and order and SM survives, but
both democracy and democracy2 are statistically insignificant. In column
9, I add both MR2 and democracy2 into the preferred model. The effects
of schooling, law and order and SM survive. The non-linear effect of MR
also survives.
To summarize, I find that both human capital proxied by schooling and
institutions have positive and statistically significant effects on growth.
The effect of human capital is relatively larger in size compared to the
effect of institutions. Among institutions, market-creating institutions
proxied by law and order and market-stabilizing institutions proxied by
SM have positive and equal effects on growth. Regulatory institutions
proxied by MR seem to matter only to a certain extent and marketlegitimizing institutions proxied by democracy do not seem to matter.
This holds even when I use the Freedom House Political Rights index as
a measure of democracy. The democracy result is in conformity with the
existing evidence in the literature. Previous studies have documented that
the evidence in favour of democracy yielding subsequent growth is at best
weak (see Barro, 1996; Przeworski et al., 2000; Acemoglu et al., 2008).
In Table 8.5, I deal with an important technical issue. Is the use of
Blundell and Bond estimation technique appropriate? Nickell (1981)
shows that when fixed effects are correlated with explanatory variables
then OLS overestimates the effect of the lagged dependent variable, fixed
effect underestimates it and system GMM should be in between. In a
recent Monte Carlo study, Hauk and Wacziarg (2004) showed that fixed
effect and Arellano and Bond GMM can in fact overestimate the effect of
the lagged dependent variable and bias towards zero the effect of other
variables in the presence of measurement error. They showed that OLS
02/03/2011 10:09
158
Table 8.5
Is the Blundell and Bond method appropriate?

Initial
income (yiT25)
Law and order (LO)
labour, business
(MR)
Sound money index
(SM)
Democracy index
(DEMOC)
Total years of
schooling (TYS)
Log trade share
(LTRS)
Specification tests
(P-values)
Hansen test
Test for AR(1) in
residuals
Test for AR(2) in
residuals
F-test that all i50
F-test for overall
significance
R2
Observations/
countries
(2)
GMM
Arellano
and Bond
(3)
GMM
Blundell
and Bond
(4)
0.01***
(0.0025)
0.005***
(0.0012)
0.001
(0.0015)
0.05***
(0.0059)
0.004***
(0.0013)
0.001
(0.0018)
0.08***
(0.0117)
0.004**
(0.0015)
0.004**
(0.0021)
0.01*
(0.0063)
0.005**
(0.0020)
0.003
(0.0024)
0.002***
(0.0005)
3e-04
(0.0005)
0.003***
(0.0009)
0.003
(0.0024)
8.5e-04
(0.0006)
4e-04
(0.0005)
0.003*
(0.0017)
0.024***
(0.0057)
Pooled
OLS
Fixed
effect
(1)
1e-05
(0.0007)
8e-05
(0.0007)
0.002
(0.0029)
0.022***
(0.0070)
0.003***
(0.0009)
3e-04
(0.0009)
0.006**
(0.0026)
0.008
(0.0058)
0.0004
0.0191
0.351
0.001
0.344
0.174
240/83
323/83
0.0000
0.0000
0.2211
323/83
323/83
a two-sided alternative. Standard errors are reported in parentheses. Hansen test is the
test of the H0: the instruments as a group are exogenous. Hansen test statistic from twostep Arellano and Bond estimations is reported which is robust to heteroskedasticity or
autocorrelation. Arellano and Bond AR(1) and AR(2) tests in residuals are also reported.
Note that to pass these tests, one has to reject the null of no AR(1) and fail to reject the null
of no AR(2).
02/03/2011 10:09
159
and Blundell and Bond perform the best in this situation. The advantage
of Blundell and Bond over OLS is that it tackles the endogeneity problem
better. My results match with Hauk and Wacziarg (2004), as I notice in
columns 2 and 3 of Table 8.5 that fixed effect and Arellano and Bond
overestimate the effect of the lagged dependent variable (20.05 and 20.08,
respectively) and underestimate the contribution of schooling (0.003 and
0.002, respectively). In the case of Arellano and Bond, the effect of schooling is statistically insignificant. Furthermore, the Hansen test statistic in
this case also documents the weak instrument problem with Arellano and
Bond when there is a short panel with persistence in the data. Therefore,
the evidence suggests that GMM Blundell and Bond is the appropriate
way to go.
Table 8.6 reports robustness results of the preferred model with additional covariates. I add additional covariates (investment and population
growth, corruption, foreign direct investments (FDI), foreign aid, real
exchange rate distortions and credit to the private sector)7 and I find that
my schooling, law and order and SM result survives. Column 1 reports the
preferred specification. In column 2, I add Solow style variables (investment share and population growth) as additional controls. The law and
order and schooling result survives as these variables register positive and
statistically significant effects on growth. The magnitude of the effects
on growth, however, reduces by 0.1 percentage points in the case of law
and order and by 0.58 percentage points in the case of schooling. This
implies that there is an upward bias in my estimates in the absence of these
variables. However, the extent of the bias is not severe enough to alter the
direction and statistical significance of the effects. In column 3, I control
for corruption. Coefficients on law and order, schooling and SM survive
this test. This indicates that including corruption in our set of institutions
will not alter the major result. In columns 4 to 7, I examine the effect of
adding FDI, foreign aid, real exchange rate distortion and credit to the
private sector, respectively. Credit to the private sector is used as a proxy
for financial liberalization in the literature (Levine et al. 2000). In all the
cases, my main result survives. In column 8, I control for all covariates
taken together. Only the effect of schooling survives. However, controlling for all factors drastically reduces the sample size. Also, it suffers from
the weak instrument problem and perhaps multicollinearity, and hence is
unreliable. In column 9, I replace law and order by expropriation risk an
alternative measure of market-creating institutions. This measure has been
used by Acemoglu et al. (2001). I find that expropriation risk, SM and
schooling are statistically significant with positive effects on growth. The
impact of schooling on growth increases by 0.6 of a percentage point as a
result of this change. The impact of market-creating institutions proxied
02/03/2011 10:09
160
02/03/2011 10:09
0.0004
(0.0008)
0.001
(0.0007)
0.004**
(0.0021)
0.003***
(0.0009)
0.0003
(0.0009)
0.006**
(0.0026)
0.351
Specification
tests (P-values)
Hansen test
0.891
0.002
(0.0025)
0.004**
(0.0016)
0.003
(0.0024)
0.005**
(0.0020)
0.715
0.002***
(0.0009)
0.0004
(0.0009)
0.005**
(0.0025)
0.003
(0.0022)
0.006***
(0.0020)
Preferred Investment Corruption

specification share and
population
growth
(2)
(3)
(1)
0.868
0.002***
(0.0009)
0.0002
(0.0008)
0.006**
(0.0024)
0.003
(0.0024)
0.998
0.002**
(0.0007)
0.0005
(0.0007)
0.006**
(0.0026)
0.002
(0.0027)
0.007***
(0.0019)
(5)
(4)
0.004**
(0.0017)
Foreign
aid
FDI
0.851
0.003***
(0.0008)
0.0004
(0.0008)
0.004*
(0.0023)
0.0002
(0.0021)
0.003*
(0.0017)
Real
exchange
rate
distortions
(6)
0.877
0.002***
(0.0009)
0.0001
(0.0009)
0.005**
(0.0024)
0.004*
(0.0023)
0.004**
(0.0020)
Credit
to the
private
sector
(7)
0.468
0.001
(0.0011)
0.001
(0.0009)
0.005*
(0.0030)
0.001
(0.0032)
0.002
(0.0022)
(8)
0.002
(0.0022)
0.007**
(0.0019)
(10)
Alcala
Ciccone
openness
0.230
0.328
0.003** 0.002***
(0.0013) (0.0008)
0.001 0.0002
(0.0012) (0.0008)
0.008** 0.007***
(0.0041) (0.0025)
0.003**
(0.0014)
0.005
(0.0032)
(9)
All
With
covariates EXPR
Unbundled institutions, human capital and growth: robustness with additional covariates
Law and Order

(LO)
Expropriation
Risk (EXPR)
Regulation of
Credit, Labour,
Business (MR)
Sound Money
Index (SM)
Democracy
Index (DEMOC)
Total Years of
Schooling (TYS)
Additional
Covariates
Table 8.6
161
02/03/2011 10:09
LTRS,
yiT25
LTRS,
yiT25
319/82
0.462
0.174
323/83
0.008
0.001
yiT25
LTRS,
322/83
0.179
0.002
yiT25
LTRS,
319/82
0.116
0.001
yiT25
LTRS,
231/61
0.036
0.008
yiT25
LTRS,
300/77
0.098
0.002
yiT25
LTRS,
319/82
0.146
0.001
yiT25
LTRS,
209/55
0.044
0.058
323/83
0.156
0.002
yiT25
yiT25
LTRS, LROPEN,
242/83
0.013
in parentheses. Hansen test is the test of the H0: the instruments as a group are exogenous. Hansen test statistic from two-step Arellano and Bond
estimations are reported, which are robust to heteroskedasticity or autocorrelation. Arellano and Bond AR(1) and AR(2) tests in residuals are
also reported. Note that to pass these tests, one has to reject the null of no AR(1) and fail to reject the null of no AR(2). In each regression, the
standard controls are log trade share (LTRS) and Initial income (yiT25) except column 10 where Alcal Ciccone real openness (LROPEN) is used.
Other controls
Observations/
Countries
Test for AR(1)

in residuals
Test for AR(2)
in residuals
162
by expropriation risk is 0.11 percentage points less than the impact of

market-creating institutions proxied by law and order. The impact of SM,
however, remains unchanged. A major drawback of using expropriation
risk is that it drastically reduces the sample size. Finally, in column 10, I
use Alcal and Ciccones (2004) log real openness index instead of the log
trade share measure. This index is a ratio of trade to PPP GDP.8 My major
findings survive even when I use this measure of trade.
In Table 8.7, I check the robustness of my basic finding in alternative
samples. I find that the schooling, law and order and SM result survives
in the base sample without British legal origin countries, the base sample
without French legal origin countries, the base sample without Africa,
the base sample without Neo-Europe9 and the base sample without oil
exporters. The magnitude of the coefficient on law and order varies from
0.004 to 0.007, which is 0.2 per cent of its sample standard deviation. The
coefficient on schooling varies from 0.005 to 0.008, which is 0.1 per cent
of its sample standard deviation. All other variables except law and order
become statistically insignificant in the Africa sample. This is perhaps
emphasizing the importance of market-creating institutions in Africa.
8.4 SUMMARY
It is well documented that institutions are important for growth. In this
chapter, I make an attempt to go beyond the broad institution continuum.
I use a four-way classification of institutions (market-creating, marketregulating, market-stabilizing and market-legitimizing institutions) and
identify a proxy for each of them. This allows me to unbundle institutions
and take the analysis beyond property rights and contracts. I estimate the
contributions of market-creating, market-regulating, market-stabilizing
and market-legitimizing institutions to growth. I find that strong marketcreating institutions and market-stabilizing institutions are good for
growth. There exists a growth-maximizing level of market regulation
beyond which red tape creates disincentives for investment. Marketlegitimizing institutions do not seem to matter for growth. Our basic result
survives across different samples and the additional covariate test.
This exercise allows us to learn about the statistical importance of these
institutions on growth. However, these models, in spite of their sophistication, are not free from limitations. One needs to bear in mind while
interpreting these coefficients that they are average relationships across
countries. In order to draw policy lessons, one needs to look into more
detailed analysis of the role of these factors. I embark on such an exercise
in Chapter 9.
02/03/2011 10:09
163
02/03/2011 10:09
Specification tests
(P-values)
Hansen test
Test for AR(1) in
residuals
Test for AR(2) in
residuals
0.987
0.002
0.610
0.351
0.001
0.174
0.006**
(0.0024)
0.002
(0.0022)
0.002*
(0.0009)
0.001
(0.0008)
0.008***
(0.0022)
(2)
(1)
0.005**
(0.0020)
0.003
labour, business (MR) (0.0024)
Sound money index
0.003***
(SM)
(0.0009)
Democracy index
0.0003
(DEMOC)
(0.0009)
Total years of schooling
0.006**
(TYS)
(0.0026)
Base sample
without British
legal origin
countries
Base sample
0.152
1.000
0.087
0.002
(0.0022)
0.01***
(0.0032)
0.006***
(0.0014)
0.0004
(0.0008)
0.005*
(0.0027)
Base sample
without
French legal
origin
countries
(3)
0.929
1.000
0.021
0.327
0.935
0.002
0.007***
(0.0022)
0.002
(0.0023)
0.002**
(0.0010)
0.0004
(0.0009)
0.007***
(0.0025)
(5)
(4)
0.007***
(0.0025)
0.005
(0.0036)
0.0009
(0.0007)
0.0005
(0.0008)
0.0007
(0.0032)
Base sample
without
Africa
African
countries
0.165
0.500
0.002
0.005**
(0.0020)
0.003
(0.0026)
0.003**
(0.0010)
0.0003
(0.0009)
0.007**
(0.0027)
(6)
Base sample
without
Neo-Europe
Unbundled institutions, human capital and growth: robustness with alternative samples
Law and order (LO)
Table 8.7
0.102
0.554
0.002
0.004**
(0.0021)
0.003
(0.0024)
0.002**
(0.0010)
0.0001
(0.0010)
0.006**
(0.0027)
Base
sample
without
oil
exporters
(7)
164
02/03/2011 10:09
LTRS,
yiT25
LTRS,
yiT25
157/40
Base sample
without
French legal
origin
countries
(3)
LTRS,
yiT25
LTRS,
yiT25
247/63
(5)
(4)
76/20
Base sample
without
Africa
African
countries
LTRS,
yiT25
311/80
(6)
Base sample
without
Neo-Europe
LTRS,
yiT25
304/78
Base
sample
without
oil
exporters
(7)
in parentheses. Hansen test is the test of the H0: the instruments as a group are exogenous. Hansen test statistic from two-step Arellano and
Bond estimations is reported which is robust to heteroskedasticity or autocorrelation. Arellano and Bond AR(1) and AR(2) tests in residuals are
also reported. Note that to pass these tests, one has to reject the null of no AR(1) and fail to reject the null of no AR(2). In each regression, the
standard controls are log trade share (LTRS) and initial income (yiT25) . Neo-Europe includes Australia, Canada, New Zealand and United States.
LTRS,
yiT25
Other controls
(2)
(1)
210/54
Base sample
without British
legal origin
countries
Base sample
323/83
(continued)
Observations/countries
Table 8.7
165
NOTES
1. Note that Hall and Jones (1999) use the social infrastructure index as an overall measure
of institutions. They combine law and order, bureaucratic quality, corruption, risk of
expropriation and government repudiation of contracts from ICRG and the Sachs and
Warner (1995a) index of openness to construct the index. This, however, is not suitable
for the purpose of unbundling institutions. Also, the Sachs and Warner openness index
used by Hall and Jones is not a measure of institutions.
2. Knack and Keefer (1995) use this as a measure of contracting institutions.
3. Barro (1995) reports a negative relationship between inflation and central bank independence in a cross-section of countries.
4. This measure is suitable since it ranges from 1 to 7 and distinguishes between different
shades of democracy.
5. See Djankov et al. (2002) for a survey of this literature.
6. The value 4.9 lies well within the sample range of MR (which is 0 to 10).
7. Previous studies have reported strong correlation between these variables and growth.
8. Alcal and Ciccone (2004) argue that this measure performs better than the standard
measure in the presence of trade-driven productivity change. However, this measure is not
free from controversy. Rodrik et al. (2004) show that the real openness index (Ropen)
and trade volume (Open) are linked by the identity logRopen 5 logOpen 1 logP,
where P is a countrys price level. Also, from the Balassa-Samuelson argument P is well
known to be closely associated with a countrys income/productivity level. Rodrik et
al. (2004) plot logRopen 2 logOpen on logGDP and find a positive relationship. They
also find very little correlation between logP and logOpen. Based on this, they argue
that augmenting P (which has very little or no correlation to Open) into the standard
trade volume measure is likely to spuriously attribute the correlation between logP and
logGDP on the correlation between openness and logGDP.
9. Neo-Europe includes Australia, Canada, New Zealand and United States.
02/03/2011 10:09
9.
Making policy work: a road map for

future growth
9.1 WHAT HAVE WE LEARNED SO FAR?

Part I of this book focuses on the historical determinants of long-run economic development. In Chapter 2, I document wide variations in growth
rates and living standards across countries over the post-war period. The
key questions, however, are what explains this variation and what explains
its persistence. Looking further back, I document that this variation is not
just attributable to how countries performed over the post-war period but
also from the start of the sixteenth century. Some countries diverged from
the rest in terms of living standards starting from that period. Chapter
2 presents a detailed account of this process of divergence and the data
involved.
In Chapter 3, I present some of the existing theories explaining this divergence. I present a case for exploring the root causes. I also review theories
of institutions, geography, human capital, trade, religion and culture, and
state formation and war. Chapter 4 presents evidence and illustrates the
difficulties in quantifying the relationships between root causes and economic development. It also shows that the empirical results are strikingly
different in Africa compared to the rest of the world. Diseases (malaria, in
particular) are the only significant explanator of economic development in
Africa. All other variables including institutions are statistically insignificant. This is, however, not the case in a more general sample where both
malaria and institutions are important for development.
Chapter 5 makes an attempt to relate these empirical results to economic
history. Empirical results are in reduced form and, therefore, are not able
to explain the dynamics of economic development across continents over
centuries. In order to make sense out of them, one needs to study them in
conjunction with historical narratives. Keeping this in mind, Chapter 5
presents a unifying framework to explain the process of development in
Western Europe. This framework is then compared with the growth narrative in other countries and continents namely Africa, China, India, the
Americas, Russia and Australia. A narrative style is adopted throughout
this section to create a bridge between the empirical literature and history.
166
02/03/2011 10:09
Making policy work
167
The framework is also an attempt to merge all seemingly disparate theories

of economic progress over the long term. This, without doubt, is the main
value added of this book. The main lesson of Chapter 5 is that diseases and
geography matter at an early stage of development. Institutions, however,
become much more important as the economy develops. Geography, and
in particular disease epidemics, are a crucial explanator of the lack of
development in Africa. Diseases impact African development through discouraging economic agents to save for the future. A high discount rate on
future consumption due to diseases and the resulting lack of saving translates into a poverty trap. In contrast, in China and India, the Malthusian
cycle was broken fairly early and institutional weaknesses played a crucial
role in their respective declines. In the Americas and Australia, colonial
institutions were a crucial factor. In Russia, it was the crippling political
institutions of the nineteenth century and restrictive political and economic institutions of the Soviet Union that did the damage.
Part II of the book focuses on empirical evidence using contemporary
data. The aim is to identify policies that have been successful in delivering
growth in the recent past. Chapter 6 focuses on the effectiveness of trade
policy and trade openness in general. The results suggest that trade on its
own is not capable of delivering growth dividends. The effect of trade is
conditional on the quality of institutions. In other words, I identify that
there is a threshold level of institutional quality below which trade is not
beneficial for growth. I also present a list of countries that are most likely
to benefit from trade at the margin, given their institutional quality.
In Chapter 7, I deal with a related question. What is the likelihood
of improving institutional quality through trade? I empirically show
that indeed institutional quality improves with trade. The estimates
suggest that if the trade liberalization index in Myanmar (almost a closed
economy) was to increase by 0.5, then the corresponding increase in the
property rights institutions score would be 37 percentage points. I also find
similar results for contracting and regulatory institutions. The effect could
work through realignment of local political forces. Alternatively, it could
also work through institutional transfer.
Chapter 8 explores the empirical question, which institutions matter for
growth. I use a four-way classification of institutions (market-creating,
market-regulating, market-stabilizing, and market-legitimizing institutions) and identifying a proxy for each of them. Market-creating institutions are characterized by the adequate protection of private property and
contract enforcement. The main role of market-stabilizing institutions is
to ensure macroeconomic stability and avoid undertaking policies that
lead to market distortions. Market-regulating institutions tackle market
failure and restrict anti-competitive behaviour by firms. Market-stabilizing
02/03/2011 10:09
168
institutions, on the other hand, take care of redistributive policies to make

sure that the benefits of growth are distributed fairly across society. The
four-way classification allows me to unbundle institutions and take the
analysis beyond property rights and contracts. I estimate the contributions
of market-creating, market-regulating, market-stabilizing and marketlegitimizing institutions to growth. I find that strong market-creating
institutions and market-stabilizing institutions are good for growth. There
exists a growth-maximizing level of market regulation beyond which red
tape creates disincentives for investment. Market-legitimizing institutions
do not seem to matter for growth.
In summary, we have learned that institutions and geography are
important factors influencing economic development over the very long
run. Evidence suggests that geography is particularly important in an early
stage of development. In fact, in Africa, malaria is the only significant
explanator of lack of development. We also learn that policies such as
trade openness could work only in an environment where institutions are
strong. However, trade has the potential to influence institutional quality.
Even though informative, the empirical results reported and discussed,
so far, are average trends. Therefore, they may not be accurate in individual country situations. In order to inform policymakers, one needs to
go beyond the average results of cross-national regressions and focus on
details. This is what I aim to do in Section 9.2. I start with a discussion
of the first principles of institutions and policies for development. Then I
discuss aberrations and what could be done in such situations.
9.2 POLICIES FOR THE FUTURE

9.2.1
First Principles: Setting up Conditions for Growth
Property rights
Protection of private property is essential for the development of any entrepreneurial activity. In situations where private property rights are weak or
non-existent, it is almost impossible to attract investments from private
investors. Predatory behaviour either by the state or other non-state actors
on private investment creates uncertainty for the investors. The risk of
predation makes the cost of investment prohibitively high. This leads to
very low levels of investment and poor growth outcomes. Therefore, it is
absolutely essential that private property rights are guaranteed by law and
also properly enforced by the government in case of disputes. In many
instances, the law guaranteeing property rights may exist, but predation
by private individuals or non-state actors, or expropriation by the state
02/03/2011 10:09
Making policy work
169
or bureaucrats may be rife. In such instances, long-term growth outcomes

may be poor because of investment risks. As a result, the true economic
potential of that country may not be realized.
Rule of law
Like property rights, the rule of law is also a key ingredient for growth.
The rule of law ensures that private individuals are protected by law and
makes it difficult for a state to actively engage in expropriation or predatory behaviour. In other words, it makes it difficult for powerful individuals within the state to unlawfully use the state apparatus against private
individuals. If such a situation arises, then there are provisions to hold
that individual to account in a court of law. Therefore, it provides an
important layer of protection for investors. In the absence of rule of law,
investment risks multiply and investors only invest in low-risk low-return
projects (for example, in the subsistence sector). Furthermore, violent
conflict may arise making investments unworthy. Therefore, the upshot is
a poor growth outcome in the absence of the rule of law.
Contract enforcement
A strong contracting environment is essential not only for investments
but also for the development of financial markets. Business and enterprise
flourish in an environment where it is easier for two completely unrelated
private individuals to draft contracts and take action in situations where
these contracts are reneged. The law of the land should facilitate this
process. The easier and faster the process, the better it is for new investments and innovations. In situations where this is not the case, growth
dividends are relatively smaller. However, one could argue that formal
contracting institutions are redundant in an environment of high social
capital and trust. In reality, trust and social capital could operate within
relatively small groups. Therefore, the cost of seeking a contract or getting
access to credit becomes difficult for individuals who are not part of this
network. In such an environment, potentially high pay-off projects may
fail to get funded if they come from an individual who is not part of the
network. Hence, by construction such institutions limit growth as they fail
to ensure free entry.
Regulatory institutions
Strong regulatory institutions are important for growth. Private economic
agents in the market are driven by greed and animal spirit. Therefore,
if left unregulated, economic agents may engage in predatory and anticompetitive behaviour, which may lead to market failure and net welfare
loss to society. Recent events in financial markets across the world, which
02/03/2011 10:09
170
led to the global financial crisis, underscore the importance of regulation.

Regulators in the previous decade were increasingly of the view that selfregulation is Pareto superior to regulation by the state or independent
bodies. In retrospect, one could now clearly identify that error. Therefore,
it is important to put in place independent regulatory bodies overseeing the behaviour of market actors to limit anti-competitive behaviour,
monopoly practices and collusion. It is important for the regulators to
be independent of the government so that they are relatively free from
political pressures and also independent of market actors so that they are
relatively free from industry pressures. A strong regulatory environment
may foster investments and economic growth. However, there is a risk of
over-regulation. Excessive regulation, on the other hand, may lead to red
tape, increased cost of investment and low growth. Therefore, in an ideal
world, regulatory institutions should optimize the level of regulation.
Macroeconomic stabilization
Macroeconomic stabilization is crucial for growth over the long term.
Investment, by definition, is a risky activity. In most cases, there is a
gap between investment in a project and the point in time when it starts
delivering returns. This is commonly known as the gestation gap. In an
environment of macroeconomic volatility, it becomes difficult for investors to foresee costs and future returns. Price volatility could impact on
both input costs as well as future returns. Interest rate volatility could
also impact on costs through the credit channel. Therefore, it is crucial for
macroeconomic institutions such as the central bank and the treasury to
deliver a stable macroeconomic environment for growth. Central banks
are best able to do that when they are independent. In most successful
economies, central banks are mandated to tackle inflation by using interest rates as an instrument. However, government borrowing also has an
impact on interest rate. If a government is borrowing from the market for
government spending, it is likely to make credit dearer and increase the
interest rate. This would have an impact on the balance sheet of private
businesses seeking credit from the market. The cost of credit for private
businesses would go up, crowding out investments. Therefore, the treasurys actions would also have an impact on investments and the economy.
To sum it up, the use of a central banks monetary policy to keep inflation
low and the treasurys efforts to keep debt low are important for a stable
macroeconomic environment and growth.
Representative politics
The process of economic growth can be harsh at times. Growth in a
capitalist economy, by its nature, picks winners and losers. Rapid growth
02/03/2011 10:09
Making policy work
171
driven by technological progress or expansion of a particular sector in

an economy might make other sectors unprofitable. This would lead to
redundancies, unemployment and factory closures. Therefore, redistribution would be an important policy tool to help the unemployed receive
training and re-engage with the workforce. Examples of such support
mechanisms are social safety nets, unemployment benefits and social
insurance. These policies are capable of minimizing social unrest and help
to maintain a climate conducive to further investments. Representative
political institutions are best to deliver on redistribution. Some have
argued that democratic institutions are the most efficient form of representative political institution. However, there are examples of other types
of institutions that are also capable of delivering redistributive outcomes
as well as a democracy. Of course, China comes to mind. Even though
inequality in China has risen over the last three decades, significant steps
have been taken by the Chinese authorities to reduce the gap through
retraining and re-employing the unemployed workforce, allowing migration and providing unemployment benefit. Nevertheless, it is perhaps fair
to conclude that the more representative the political institutions are, the
better it is for redistribution.
Human capital investments
Having an educated population is also crucial for growth. Without an
educated population it would be difficult to properly run institutions.
Running institutions would require properly trained bureaucrats. With
a low level of schooling in a country, there would be a limited supply of
properly trained bureaucrats to run the institutions important for growth.
It would also be difficult for businesses to recruit an adequately skilled
workforce. This may lead to low levels of investment and also investments
in low-return projects. On the other hand, in a country with high levels
of schooling, the workforce would be skilled enough to handle multiple
tasks and hence would be much more employable. Therefore, they would
be extremely attractive to investors in more than one sector. Hence, it
is crucial for countries to improve the level of schooling in order to run
national institutions, attract investments and deliver economic growth.
Access to markets
Infrastructure bottlenecks in developing countries often scuttle economic
growth. Lack of roads, railway links, seaports and ocean-navigable waterways limit access to markets and thereby restrict growth. Landlocked
economies in Africa and Central Asia suffer from the lack of access to
international trade routes. Lack of telecommunication networks also limit
business growth and reduce productivity as sellers struggle to contact
02/03/2011 10:09
172
potential buyers of their products. This could cost economies dearly in

terms of future economic growth. In contrast, an economy well connected
to the global markets both physically as well as through telecommunication networks can engage in internal trade and international trade much
more easily. The positive externalities from a functioning road and railway
network could be large. They facilitate institutional delivery at the local
level, in addition to helping business and trade. It makes it much easier
for the central government to keep in contact with local level institutions
and deliver public goods such as health services and materials for schools
effectively. Therefore, infrastructure investments to overcome geographic
bottlenecks are essential for growth.
International trade
International trade increases the size of the market. It can also induce technology transfer and transfer of institutions. All of these factors are potentially growth enhancing. Therefore, trade in general is believed to be good
for growth even though hard evidence from macro data is difficult to find.
However, there is a caveat for developing economies. Trade and thereby
specialization in one or two products may make a developing economy
vulnerable to volatility in international prices. If the specialization is in
primary products and natural resources, the risk of Dutch Disease also
multiplies. In such a situation, a developing economy may not benefit
from trade over the long term. It may also have an adverse impact on its
institutions as natural resource exports are likely to increase inequality
and increase social friction. Diversification, expanding export product mix
and curbing over-reliance on natural resource exports are perhaps the best
way forward in such situations. Therefore, trade openness is perhaps best
for developing countries when the aim is to diversify their exports.
9.2.2
Recent Success Stories and First Principles
Recent growth successes in China and India have allowed these countries
to significantly reduce absolute poverty and improve the quality of life
for millions. How much of their success relies upon the first principles
described above? This could be useful for countries that have not been
able to enjoy such success so far. In this section, I make an attempt to
find out. First I discuss the case of China. This is followed by the case of
India.
China
The Chinese economy has grown rapidly over the last three decades. In
particular, China has been the fastest growing economy since the 1990s.
02/03/2011 10:09
Making policy work
173
What is the reason behind Chinas phenomenal growth? Many economists argue that trade orientation and FDI are central to Chinas success.
However Chinas economic reforms cover something more than that.
China embarked on economic reforms in 1978, led by Deng Xiaoping,
the then leader of the Communist Party. The main idea was to generate enough economic surplus to eradicate imbalances across regions in
the mainland Chinese economy. To this effect, the Chinese government
embarked on a large-scale modernization project for the economy. The
first step was to institute household responsibility systems in the farms.
The idea was to transfer responsibility for cultivation and, in effect,
property rights from the Soviet-style collective farms to the households.
Under this system, the farmers were also allowed to sell their surplus crop
on the open market. This immediately delivered significant productivity
dividends to Chinese agriculture. This reform was followed by the establishment of Township Village Enterprises (TVEs). TVEs were mainly
small- to medium-sized industries owned by townships and villages. TVEs
were also a success story. Throughout this period, the Chinese government
invested in modernizing the infrastructure. New roads, highways, airports,
sea ports and telecommunication systems were being built. During the late
1980s, the Chinese government introduced an open door policy towards
international trade and FDIs. Export processing zones were set up along
the coast and foreign manufacturing firms were allowed to set up factories
using inexpensive Chinese labour. This turned out to be huge success and
helped lift millions out of poverty. As a result, the Chinese economy, at
present, is dominated by the private sector. However, state controlled
enterprises dominate utilities and the resources sector.
Without doubt, China did not follow the first principles of growth to
the letter. There were no private property rights in China. Property rights
in TVEs and household run farms were shared. Nevertheless, it managed
to create enough incentives for long-term investments. The emphasis on
FDI in manufacturing to exploit low labour costs was also successful and
generated employment opportunities for millions. This was made possible
because of an unwritten guarantee of non-predatory behaviour by the
state. Even though there is little rule of law in the Western sense of
the word, the Chinese state has developed a symbiotic relationship with
the multinationals investing in China through mutual trust. Economic
progress is crucial for the legitimacy of the one-party Chinese state and
the state is well aware that the multinational firms are important contributors to that process. Furthermore, Chinese political institutions remained
far from democratic throughout this process. However, they are likely to
remain reasonably representative as long as the Communist Party is in
touch with what is happening on the ground. In spite of rising inequality,
02/03/2011 10:09
174
it is perhaps fair to say that the Chinese state has managed it effectively
through redistribution programmes and minimized social unrest.
Therefore, what we notice in retrospect is that the Chinese Communist
Party perhaps has used the first principles as a guideline and not as gospel.
In the process, they have developed institutions which are rooted in local
customs and culture and best represent local conditions to deliver economic growth.
India
Indias economic reforms started in the 1980s with privatization of some
state-owned enterprises and unilateral reductions in tariffs and quantitative
restrictions. However, the reform process gained momentum in 1991 following a debt crisis. Growth in per capita income experienced a significant
boost following the 1991 reforms. However, the institutional change, on
paper, was not significant. Post-independence India always had property
rights and laws to enforce private contracts. She also had the rule of law
and constitutional rights for individuals to buy and sell goods and services
on the open market. However the sociopolitical environment was not pro
business till the late 1980s. A commonly held view amongst the public and
politicians pre-reform was that a socialist system is best suited for India
to deliver shared prosperity. As a result, the economy was held hostage to
numerous regulations and licensing procedures commonly known as the
licence raj. With reforms, the entire discourse shifted and the government
and its institutions became more pro business. This delivered improvements in institutional quality and rapid growth. The economy experienced
expansion in both manufacturing and services at the expense of agriculture.
However, problems remained at the front of public goods delivery and
infrastructure. With a rapidly expanding economy, the demand for infrastructure and other public goods quadrupled. Laying out infrastructure
efficiently within the current institutional framework is always a challenge
in India. The rapidly growing demand for land to lay out infrastructure and
other projects in a growing economy, such as India, is often faced with challenges of land disputes. The sheer number of these disputes poses challenges
for the court system and other institutions. Nevertheless, in spite of all these
hurdles, Indias progress over the last two decades has been impressive.
In summary, the paths of development chosen by India and China are
distinctly different. Both countries have successfully utilized local conditions and developed institutions that are well aligned with their initial
conditions. However, there is little debate on the fact that they are built on
the appropriate foundation of incentives. In other words, even though the
first principles are not applied to the letter in both of these countries, they
are applied to create the right incentives for growth.
02/03/2011 10:09
Making policy work
9.2.3
175
Growth Amidst Chaos: The Case for Multilateral Interventions
A prerequisite for economic growth is a functioning state which is able

to tax its citizens and provide public goods in return. Therefore, it is
extremely difficult to start growth in an environment of state failure and
chaos. In recent times, conflict-ridden parts of Africa, Central Asia and
the Middle East have experienced such situations. The principal challenge
in such situations is to form a functioning state. In this section, I analyse
whether multilateral interventions can help in a situation of state failure.
Internal conflict and lack of effective government are a serious developmental challenge in many parts of Africa, Central Asia and the Middle East.
What is the best possible way to react in such situations so that an effective
government can be formed? History tells us that conflict and often violent
conflict is integral to institution formation. Conflict shapes the distribution
of economic and political power, which determines institutional outcomes
over the long term. One could argue that the contemporary economic and
political institutions in Europe are a product of two centuries of internal as well as external conflict. The North American economic as well as
political institutions are an outcome of the American War of Independence
and the American Civil War. Many historians argue that contemporary
Australian political institutions also have their roots in the American War of
Independence (see Section 5.2). British colonizers gave in to democratic pressures in Australia because of their fear of a repeat of the North American war
and losing the Australian colonies. However, violent conflict is not an option
in the contemporary world as it leads to loss of life and property damage. In
most cases, what we observe are interventions by the international community to contain violence and build peace. Forming a functioning government
and economic reconstruction are integral parts of that process. The key question, therefore, is how to make interventions work.
Intervention, by definition, is risky. It distorts the distribution of political and economic power in a country across different power groups by
bringing in an external player. Often the external players are not impartial
and are driven by their own security and economic interests. This has an
influence on local distribution of power and makes institution building
even more challenging. In order to form an effective government and
other ancillary institutions, all these interests need to converge. This is,
of course, easier said than done. In situations when they do not converge,
interventions may end up making the situation even more unstable and
prolonging the conflict. Without going into detail, in the following paragraphs I construct a list of strategies that may be useful in such situations.
However, note that these strategies are far from comprehensive and are
perhaps a topic of research for another book.
02/03/2011 10:09
176
Ideally, institution building in a conflict-ridden country should be

based on representative politics and also raising the cost of conflict for the
parties involved. This could be done by making the state as representative
as possible and also providing military support to the state. Multilateral
military presence may have an advantage over unilateral military presence. Both the locals and the international community may perceive
multilateral military presence as more legitimate than unilateral military
presence. While constructing a state, legitimacy is a key issue. At the local
level, efforts should be made so that all sections of society are well represented in any form of government. This, on one hand, may involve creating incentives for warring groups to work together and, on the other hand,
may seek representation of all ethnic groups and tribes in the government.
The more broad-based the government is, perhaps the better it is for
legitimacy. Efforts should be made to create a fair tax system and provide
services such as policing. This may also help the quest for legitimacy of any
young government.
Multilateral interventions may have advantages over unilateral interventions for the following three reasons. First, the perceived legitimacy of
multilateral interventions is, in general, higher than country A- or country
B-led interventions. Second, multilateral interventions are more likely to
be relatively impartial and, as a result, any distortions in the local distribution of power would be less damaging. At least, distortions would not
be driven by security or economic interests of any individual intervening
country. This would also help achieve legitimacy. Third, multilateral interventions guided by institutions such as the United Nations are more likely
to be willing to make long-term commitments, if required. This is unlikely
to be the case with unilateral interventions as the government involved is
more likely to be driven by their own countrys short-term political cycle.
One could, of course, point towards British success in Sierra Leone as
an example of successful unilateral intervention. But these examples are
extremely rare and multilateral interventions are perhaps the best way
forward. Nevertheless, there is no denying the fact that these situations
are extremely challenging and very difficult to handle. It is also important
to note that these efforts are extremely costly in terms of human lives and
require long-term commitment.
9.2.4
Diseases and Foreign Aid
Debilitating and killer diseases often limit an individuals ability to work.

Therefore, they have a direct effect on productivity and economic growth.
In a poor country, often the economic agent may not have the resources
to fight such conditions on their own. Fiscally strapped poor country
02/03/2011 10:09
Making policy work
177
governments also may not have the resources to engage in large-scale

investment in public goods to tackle such conditions. Therefore, what
ensues is a poverty trap. Foreign aid is vital in such situations.
The dominance of disease, especially malaria, in Africa described in
Chapters 4 and 5 does make a strong case for aid. However, aid itself in
these circumstances is not an engine of growth. Aid should be used in a targeted fashion to tackle diseases and create a healthy workforce capable of
undertaking work in the agriculture or manufacturing sectors. Injection of
foreign aid to improve health should be combined with creating the right
incentives for investments in sectors that have potential for growth. Some
have argued in favour of a big-push-style policy of aid to pull poor countries out of the poverty trap. Any big-push-style policy relies on planning
and any planning exercise disregards the often unsolvable information and
incentive problems attached to it (Easterly, 2006). In particular, the incentive of the aid worker and the information problems associated with aid
infrastructure are often overlooked by macro-level planning. Therefore, a
step by step approach to create the right incentives for investments seems
to be the appropriate way forward, rather than a big-push-style approach
funded by foreign aid.
9.2.5
The Role of Technical Assistance
Technical assistance is crucial for the development of a well-functioning

bureaucracy in a developing country. Countries in their early stage of
development may not have the talent pool to choose from for recruitment of government officials, university teachers, school teachers and
police officers. These services are essential for economic growth and also
to supply the next generation of smart individuals in the country capable
of tackling governance and creating a favourable environment for business. Technical assistance from developed countries may go a long way in
helping young countries and governments to address these issues. Creating
university places for developing countries citizens and also creating training opportunities for their government officials are desirable. This would
boost the countries ability to provide public goods in an efficient manner,
which is vital for economic growth.
9.3 EPILOGUE
This book is an attempt to understand the process of economic progress
and why it varied across countries and continents over the last two centuries. The principal message is that history and geography are important
02/03/2011 10:09
178
in explaining this difference. The limited data that we have indicate that
economic progress is an outcome of a combination of multiple factors
over a very long period of time. Geography shapes history and also
shapes incentives that affect economic development. Geography also
affects development directly. It determines factor endowments and natural
resource availability. Incentives remain crucial for economic development,
and geography and resource endowment often set the tone of what type of
incentives are going to evolve over the long term. The historical narratives
chapters (Part I) show that these incentives evolved in a varied way across
continents and countries, which gave rise to different types of institutions.
Hence, the development outcome varies across continents.
The second part of the book focuses on methods that could be used to
influence incentives and institutions. First, it documents some macro evidence on the role of policy in shaping these incentives. It shows that trade
can benefit nations in situations where institutions are adequate. Property
rights and contracting institutions are good for growth. Market-stabilizing
institutions are good for growth and regulation is important only to a
certain extent. Second, it also outlines policies that may be beneficial
in promoting growth. It outlines the first principles of growth namely
property rights, contracts, regulatory institutions, the rule of law, macroeconomic stabilization, representative politics, human capital investments,
market access and international trade. Third, it describes the cases of India
and China, two recent success stories. It shows how these countries have
preserved incentives for private investments even without following the
first principles. But more importantly they have been able to create institutions which are well grounded in local traditions and culture and also able
to create appropriate incentives for investments. Fourth, the book also
outlines steps that could be taken to facilitate growth in situations of state
failure, disease trap, poverty trap and scarcity of skilled workers.
The main contribution of the book is to show that economic growth in
the past has been a product of multiple factors. Institutions and geography
are perhaps the most important ones that create incentives for investment
and growth. Theories are presented and empirical results are discussed to
argue in favour of such a viewpoint. It also provides a list of policies that
countries could follow to bring about growth and reduce poverty.
Of course, the book does not provide us with a magic formula to bridge
the gap between the rich and the poor. However, it does flag the importance of history, geography, politics and culture in economic development.
It also presents a strong case for studying these issues in combination with
each other and not as disparate theories. A better understanding of the
interconnection between these factors may increase our knowledge of how
countries grow in the future.
02/03/2011 10:09
Data appendix
CHAPTER 4
Measures of Economic Development
Log per capita GDP in 2000: Natural log of real GDP per capita in
2000. Real GDP figures are measured in US$ in current prices and the
figures are PPP converted. For Botswana, Cambodia, Fiji, Guyana,
Mauritania, Namibia and Papua New Guinea, I use 1999 values as
an approximation. For Central African Republic, Haiti, Puerto Rico
and Taiwan, I use 1998 values. Source: Penn World Table (PWT) 6.1,
Heston et al. (2002).
Log initial income (1820): Natural logs of per capita GDP (1820) in 1990
international Geary-Khamis dollars. Source: Maddison (2004).
Log initial income (1960): Initial level of per capita GDP (1960) in natural
logs and PPP figures. Source: Penn World Table (PWT) 6.1, Heston et
al. (2002).
Measure of Institution
Average rule of law index: This variable measures the quality of public
service provision, the quality of the bureaucracy, the competence of civil
servants, the independence of the civil service from political pressures,
and the credibility of the governments commitment to policies. The
main focus of this index is on inputs required for the government to be
able to produce and implement good policies and deliver public goods.
This variable ranges from 22.5 to 2.5, where higher values equal higher
government effectiveness. This variable is measured as the average from
1998 through 2000. Source: Kaufman et al. (2002).
179
02/03/2011 10:09
180
Measures of Religion
Catholicism: Identifies the percentage of population of each country that
is Catholic in 1980. Source: LaPorta et al. (1999).
Islam: Identifies the percentage of population of each country that was
Muslim in 1980. Source: LaPorta et al. (1999).
Measure of Openness and Trade
Log of trade share: Natural log of trade share calculated by taking log
values of figure obtained by dividing volume of trade with GDP.
Source: Frankel and Romer (1999).
Measure of Human Capital
Enrolment ratio in 1900: It is the ratio of the number of students enrolled
at the primary level and the relevant school age population. Source:
Benavot and Riddle (1988).
Measures of Geography
Distance: Absolute distance from the equator measured in latitude.
Source: Hall and Jones (1999).
Malaria risk: Percentage of the population at risk of malaria transmission
in 1994. Source: Glaeser et al. (2004).
Land area within tropics: The proportion of a countrys land area within
the geographical tropics. Source: Gallup et al. (1998).
Land area within 100 km of ocean or ocean-navigable river: Proportion of
countrys total land area within 100 km of the ocean or ocean-navigable
river, excluding coastline above the winter extent of sea ice and the
rivers that flow to this coastline. Source: Gallup et al. (1998).
SECTION 4.6
Log per capita GDP in 2000 (log yi): Penn World Table (PWT) 6.1.
Log income in 2000: Nunn (2008), originally from Maddison (2004).
Expropriation risk (INSi): Risk of outright confiscation and forced
nationalization of property. Source: ICRG.
Executive constraint in 2000: A seven category scale, 1 to 7, with a higher
score indicating more constraint. Source: Polity IV.
Rule of law index: See Rodrik et al. (2004) for details.
02/03/2011 10:09
Data appendix
181
Pre-colonial state development: Nunn (2008).

Malaria risk: Percentage of the population at risk of malaria transmission
in 1994. Source: CID datasets, Harvard University.
Malaria ecology (ME): Kiszewski et al. (2004).
Log total slave exports normalized by land area (SLVXi): See Nunn (2008).
Log total slave exports normalized by population: See Nunn (2008).
Log settler mortality (LSM): Acemoglu et al. (2001).
Log population density in 1500 (LPD): Acemoglu et al. (2001).
Interior distance (IDCi), Atlantic distance (ADCi), Indian distance
(IODCi), Saharan distance (SDCi) and Red Sea distance (RDCi):
Nunn (2008).
Frost: Masters and McMillan (2001), see Carstensen and Gundlach (2006)
for details.
Rain: Minimum of monthly average rainfall, Nunn (2008).
Humidity: Maximum of monthly afternoon average humidity (%), Nunn
(2008).
Legal origin: LaPorta et al. (1999).
Schooling in 1900: Benavot and Riddle (1988).
Log trade share in 2000: WDI online.
CONST: Constructed openness, Frankel and Romer (1999).
Ethnic fractionalization: Alesina et al. (2003).
Mining: Share of mining in GDP, Hall and Jones (1999).
Catholicism and Islam: LaPorta et al. (1999).
Gini coefficient: World Bank.
Foreign aid and national savings: WDI online.
Corruption: ICRG, PRS dataset.
Sachs and Warner openness: Sachs and Warner (1997).
CHAPTER 6
Dependent and Explanatory Variables
Log GDP per capita (log ysrt): Natural log of GDP per capita PPP (current
international dollars), CGDP. Source: Penn World Table, PWT 6.1.
Trade share (TRsrt): (exports 1 imports)/GDP. Source: WDI Online, The
World Bank Group.
Trade policy openness since t24 (posrt): Fraction of years open between t
and t24. Source: Sachs and Warner (1997).
Trade policy openness since 1950 (POsrt): Fraction of years open between
t and t21950. Source: Sachs and Warner (1997).
Expropriation risk (INSsrt): Risk of outright confiscation and forced
02/03/2011 10:09
182
nationalization of property. It covers the period 1982 to 1997. Average

of 1982 to 1983 is used as a proxy for 1980, average of 1984 to 1987 as a
proxy for 1985, average of 1988 to 1992 as a proxy for 1990, and average
of 1993 to 1997 as a proxy for 1995. Source: ICRG.
Instruments
Constructed openness (CONST): Frankel and Romer (1999).
Landlocked dummy: Sachs and Warner (1995).
Log settler mortality (LSM): Acemoglu et al. (2001).
Log population density in 1500 (LPOPDEN): Acemoglu et al. (2002).
ENGFRAC: Fraction of the population speaking English. Source: Hall
and Jones (1999).
EURFRAC: Fraction of the population speaking Western European languages (English, French, German, Portuguese, Spanish). Source: Hall
and Jones (1999).
AREA: CID Harvard Geography datasets.
CHAPTER 8
Growth (yiT) : Calculated using the formula yiT ; 1/5 (yiT 2 yiT 25) for 127
countries. Source: WDI Online, The World Bank Group.
Initial income (yiT 25) : Log GDP per capita PPP (constant 2000 international $). Source: WDI Online, The World Bank Group.
Law and Order (LO): Source: ICRG, The PRS Group.
Regulation of credit, labour and business (MR): Source: Gwartney and
Lawson (2005).
Sound money index (SM): Source: Gwartney and Lawson (2005).
Democracy (DEMOC): Source: Polity IV.
Rule of law (RULE): Source: Rodrik et al. (2004).
Expropriation risk (EXPR): Source: ICRG, The PRS Group.
Executive constraint (EXCONST): Source: Polity IV.
Repudiation of government contracts: Source: ICRG, The PRS Group.
Total years of schooling (TYS): Source: Barro and Lee (2000).
Log trade share (LTRS): Source: WDI Online, The World Bank Group.
Log real openness (LROPEN): Source: Alcal and Ciccone (2004).
Investment share: Source: PWT 6.2.
Population growth: Source: PWT 6.2.
Corruption: Source: ICRG, The PRS Group.
FDI: Foreign direct investment as a share of GDP. Source: WDI Online,
The World Bank Group.
02/03/2011 10:09
Data appendix
183
Foreign aid: Foreign aid as a share of GDP. Source: WDI Online, The
World Bank Group.
Real exchange rate distortions: Real overvaluation. Source: WDI Online,
The World Bank Group.
Credit to private sector: Domestic credit to private sector as a share of
GDP. Source: WDI Online, The World Bank Group.
02/03/2011 10:09
References
Acemoglu, D. and S. Johnson (2005), Unbundling Institutions, Journal
of Political Economy, 113(5), 949995.
Acemoglu, D. and S. Johnson (2007), Diseases and Development: The
Effect of Life Expectancy on Economic Growth, Journal of Political
Economy, 115(6), 925985.
Acemoglu, D. and J. Robinson (2000), Why Did the West Extend
the Franchise? Democracy, Inequality and Growth in Historical
Perspective, Quarterly Journal of Economics, 115, 11671199.
Acemoglu, D. and J. Robinson (2006), Economic Origins of Dictatorship
and Democracy, New York: Cambridge University Press.
Acemoglu, D., S. Johnson and J. Robinson (2001), The colonial origins
of comparative development: an empirical investigation, American
Economic Review, 91(5), 13691401.
Acemoglu, D., S. Johnson and J. Robinson (2002), Reversal of Fortune:
Geography and Institutions in the Making of the Modern World
Income Distribution, Quarterly Journal of Economics, 117, 12311294.
Acemoglu, D., S. Johnson and J. Robinson (2003a), Disease and
Development in Historical Perspective, Journal of the European
Economic Association, 1, 397405.
Acemoglu, D., S. Johnson, J. Robinson and Y. Thaicharoen (2003b),
Institutional Causes, Macroeconomic Symptoms: Volatility, Crises
and Growth, Journal of Monetary Economics, 50, 49123.
Acemoglu, D., S. Johnson and J. Robinson (2005a), Institutions as the
Fundamental Cause of Long-Run Growth, Handbook of Economic
Growth, 1, 385472.
Acemoglu, D., S. Johnson and J. Robinson (2005b), The Rise of Europe:
Atlantic Trade, Institutional Change and Economic Growth, American
Economic Review, 95, 546579.
Acemoglu, D., S. Johnson, J. Robinson and P. Yared (2005c), From
Education to Democracy?, AEA Papers and Proceedings, 95(2), 4449.
Acemoglu, D., S. Johnson, J. Robinson and P. Yared (2006), Income and
Democracy, American Economic Review, forthcoming.
Acemoglu, D., S. Johnson, J. Robinson and P. Yared (2007), Reevaluating
the Modernization Hypothesis, NBER Working Paper No. 13334,
August.
184
02/03/2011 10:09
References
185
Ahluwalia, M. (2002), Economic Reforms in India Since 1991: Has

Gradualism Worked, Journal of Economic Perspectives, 16(7), 6788.
Albouy, D. (2008), The Colonial Origins of Comparative Development:
An Investigation of the Settler Mortality Data, NBER Working Paper
No. 14130, June.
Alcal, F. and A. Ciccone (2004), Trade and Productivity, Quarterly
Journal of Economics, 119(2), 612645.
Alesina, A. and A. Drazen (1991), Why Are Stabilizations Delayed?,
American Economic Review, LXXXI, 11701188.
Alesina, A. and G. Tabellini (1989), External Debt, Capital Flight and
Political Risk, Journal of International Economics, 27, 199220.
Alesina, A., A. Devleeschauwer, W. Easterly, S. Kurlat, and R. Wacziarg.
(2003), Fractionalization, Journal of Economic Growth, 8, 155194.
Allen, R. (2008), A Review of Gregory Clarks A Farewell to Alms: A
Brief Economic History of the World, Journal of Economic Literature,
46(4), 946973.
Allen, R. (2009), The British Industrial Revolution in Global Perspective,
Cambridge: Cambridge University Press.
Arndt, C. and J. Lewis (2000), The Macro Implications of HIV/AIDS
in South Africa: A Preliminary Assessment, South African Journal of
Economics, LXVIII, 856887.
Bagchi, A. (1976a), Deindustrialisation in India in the Nineteenth
Century: Some Theoretical Implications, Journal of Development
Studies, 12, 135164.
Bagchi, A. (1976b), Deindustrialisation in Gangetic Bihar 18091901,
in B. De et al. (eds), Essays in Honour of Prof. S.C. Sarkar, New Delhi:
Peoples Publishing House.
Bairoch, P. (1982), International Industrialization Levels from 1750 to
1980, Journal of European Economic History, 11, 269333.
Bajpai, N. and J. Sachs (1999), Fiscal Policy in Indias Economic
Reforms, in J. Sachs, A. Varshney, and N. Bajpai (eds), India in the Era
of Economic Reforms, New Delhi: Oxford University Press.
Banerjee, A. (2007), Making Aid Work, Cambridge, MA: Boston
Review, MIT Press.
Banerjee, A. and L. Iyer (2005), History, Institutions, and Economic
Performance: The Legacy of Colonial Land Tenure Systems in India,
American Economic Review, 95, 11901213.
Barro, R. (1995), Inflation and Economic Growth, NBER Working
Paper No. 5326, October.
Barro, R. (1996), Democracy and Growth, Journal of Economic Growth,
1, 127.
Barro, R. and J.W. Lee (2000), International Data on Educational
02/03/2011 10:09
186
Attainment: Updates and Implications, Manuscript, Harvard

University.
Bates, R. (1983), The Centralisation of African Societies, Essays on the
Political Economy of Rural Africa, Berkeley: University of California
Press, Chapter 2.
Beaud, M. (2000), A History of Capitalism: 15002000, New York:
Monthly Review Press.
Behrman, J. and M. Rosenzweig (2004), The Returns to Birthweight,
Review of Economics and Statistics, 86(2), 586601.
Bell, C., S. Devarajan and H. Gersbach (2003), The Long-Run Economic
Costs of AIDS: Theory and an Application to South Africa, Working
Paper, June.
Belsley, D., E. Kuh and R. Welsch (1980), Regressions Diagnostics:
Identifying Influential Data and Sources of Collinearity, New York:
John Wiley & Sons.
Benavot, A. and P. Riddle (1988), The Expansion of Primary Education,
18701940: Trends and Issues, Sociology of Education, 61(3), 191
210.
Ben-David, D. (1993), Equalizing Exchange: Trade Liberalization and
Income Convergence, Quarterly Journal of Economics, 108(3), 653679.
Besley, T. and T. Persson (2008), War and State Capacity, Journal of the
European Economic Association, 6(23), 522530.
Besley, T. and T. Persson (2009), The Origins of State Capacity: Property
Rights, Taxation, and Politics, American Economic Review, 99(4),
12181244.
Bhagwati, J. and P. Desai (1970), Planning for Industrialization: A
Study of Indias Trade and Industrial Policies Since 1950, Cambridge:
Cambridge University Press.
Bhattacharyya, S. (2004), Deep Determinants of Economic Growth,
Applied Economics Letters, 11(9), 587590.
Bhattacharyya, S. (2009a), Unbundled Institutions, Human Capital, and
Growth, Journal of Comparative Economics, 37, 106120.
Bhattacharyya, S. (2009b), Root Causes of African Underdevelopment,
Journal of African Economies, 18(5), 745780.
Bhattacharyya, S. (2009c), Institutions, Diseases and Economic Progress:
A Unified Framework, Journal of Institutional Economics, 5(1), 6587.
Bhattacharyya, S. (2010), Trade Liberalization and Institutional
Development, Unpublished Manuscript.
Bhattacharyya, S. (2011), Five Centuries of Economic Growth in India:
The Institutions Perspective, in R. Jha (ed.), Routledge Handbook of
South Asian Economics, New York: Routledge, forthcoming.
Bhattacharyya, S., S. Dowrick and J. Golley (2009), Institutions and
02/03/2011 10:09
References
187
Trade: Competitors or Complements in Economic Development?,

Economic Record, 85(270), 318330.
Bhattacharyya, S. and R. Hodler (2010), Natural Resources, Democracy
and Corruption, European Economic Review, 54, 608621.
Bhattacharyya, S. and J. Williamson (2009), Commodity Price Shocks
and the Australian Economy since Federation, Australian Economic
History Review, forthcoming.
Bleakley, H. (2003) Disease and Development: Evidence from the American
South, Journal of the European Economic Association, 1(23), 376386.
Bloom, D. and D. Canning (2005), Health and Economic Growth:
Reconciling the Micro and Macro Evidence, Working Paper, February.
Bloom, D. and J. Sachs (1998), Geography, Demography, and Economic
Growth in Africa, Brookings Papers on Economic Activity, 2, 207295.
Bloom, D., D. Canning, and J. Sevilla (2003), Geography and Poverty
Traps, Journal of Economic Growth, 8, 355378.
Carstensen, K. and E. Gundlach (2006), The Primacy of Institutions
Reconsidered: Direct Income Effects of Malaria Prevalence, World
Bank Economic Review, 20, 309339.
Caselli, F., G. Esquivel, and F. Lefort (1996), Reopening the Convergence
Debate: A New Look at Cross-Country Growth Empirics, Journal of
Economic Growth, 1(3), 363389.
Chakraborty, S. and M. Das (2005), Mortality, Human Capital and
Persistent Inequality, Journal of Economic Growth, 10, 159192.
Chang, H.-J. (2008), Bad Samaritans: The Myth of Free Trade and the
Secret History of Capitalism, New York: Bloomsbury Press.
Clark, C. (1950), Conditions of Economic Progress, London: Macmillan.
Clark, G. (2007), A Farewell to Alms: A Brief Economic History of the
World, Princeton and Oxford: Princeton University Press.
Clingingsmith, D., and J. Williamson (2008), Deindustrialization in 18th
and 19th century India: Mughal decline, climate shocks and British
industrial ascent, Explorations in Economic History, 45, 209234.
Coe, D., and E. Helpman (1995), International R&D Spillovers, European
Collier, P. (2007), The Bottom Billion: Why the Poorest Countries Are
Failing and What Can Be Done About It, New York: Oxford University
Press.
Collier, P. and J.W. Gunning (1999), Explaining African Economic
Performance, Journal of Economic Literature, 37(1), 64111.
Curtin, P. (1968), Epidemiology and the Slave Trade, Political Science
Quarterly, 83(2), 181216.
Curtin, P. (1975), Economic Change in Precolonial Africa: Senegambia in
the Era of the Slave Trade, Madison: University of Wisconsin Press.
02/03/2011 10:09
188
De Long, B. (2003), India Since Independence: An Analytical Growth

Narrative, in D. Rodrik (ed.), In Search of Prosperity: Analytical
Narratives on Economic Growth, Princeton, NJ: Princeton University
Press, Princeton.
Diamond, J. (1997), Guns, Germs and Steel: The Fate of Human Societies,
New York: W.W. Norton & Co.
Dias, J. (1981), Famine and Disease in the History of Angola, 18301930,
Journal of African History, 21, 349378.
Djankov, S., R. LaPorta, F. Lopez-de-Silanes, and A. Shleifer (2002), The
Regulation of Entry, Quarterly Journal of Economics, 117(1), 137.
Dollar, D. (1992), Outward Oriented Developing Economies Really Do
Grow More Rapidly: Evidence from 95 LDCs, 19761985, Economic
Development and Cultural Change, 40, 523544.
Dollar, D., and A. Kraay (2003), Institutions, Trade, and Growth,
Journal of Monetary Economics, 50, 133162.
Dowrick, S. and M. Rogers (2002), Classical and Technological
Convergence: Beyond the Solow-Swan Growth Model, Oxford
Economic Papers, 54, 369385.
Dutt, A. (1992), The Origins of Uneven Development: The Indian
Subcontinent, American Economic Review, 82, 146150.
Easterly, W. (2001), The Elusive Quest for Growth: Economists Adventures
and Misadventures in the Tropics, Cambridge, MA: The MIT
Press.
Easterly, W. (2006), The Big Push Dj Vu: A Review of Jeffrey Sachss
The End of Poverty: Economic Possibilities for Our Time, Journal of
Economic Literature, XLIV, 96105.
Easterly, W. (2009), The Anarchy of Success, New York Review of
Books, LVI(15), 2830.
Easterly, W. and R. Levine (1997), Africas Growth Tragedy: Policies and
Ethnic Divisions, Quarterly Journal of Economics, 112(4), 12031250.
Easterly, W. and R. Levine (2001), Its Not Factor Accumulation:
Stylized Facts and Growth Models, World Bank Economic Review,
15(2), 177219.
Easterly, W. and R. Levine (2003), Tropics, Germs, and Crops: How
Endowments Influence Economic Development, Journal of Monetary
Economics, 50(1), 339.
Ebrey, P. (1999), The Cambridge Illustrated History of China, Cambridge:
Cambridge University Press.
Ebrey, P., A. Walthall and J. Palais (2006), East Asia: A Cultural, Social,
and Political History, Boston: Houghton Mifflin.
Engerman, S. and K. Sokoloff (1994), Factor Endowments: Institutions,
and Differential Paths of Growth Among New World Economies: A
02/03/2011 10:09
References
189
View from Economic Historians of the United States, NBER Working

Paper No.w066.
Engerman, S. and K. Sokoloff (2001), Inequality, Institutions, and
Differential Paths of Growth Among New World Economies, Working
Paper, University of California, Los Angeles.
Fage, J. (1978), A History of Africa, Hutchinson & Co: London.
Frankel, J. and D. Romer (1999), Does trade cause growth?, American
Gallup, J. and J. Sachs (2000), Agriculture, Climate, and Technology:
Why are the Tropics Falling Behind?, American Journal of Agricultural
Economics, 82, 731737.
Gallup, J. and J. Sachs (2001), The Economic Burden of Malaria,
American Journal of Tropical Medicine and Hygiene, 64(12), 8596.
Gallup, J., J. Sachs and A. Mellinger (1998), Geography and Economic
Development, NBER Working Paper No.w6849, December.
Galor, O. and O. Moav (2006), Das Human-Kapital: A Theory of the
Demise of the Class Structure, Review of Economic Studies, 73, 85
117.
Gemery, H. and J. Hogendorn (1979), The Economic Costs of West
African Participation in the Atlantic Slave Trade: A Preliminary
Sampling for the Eighteenth Century, in H. Gemery, and J. Hogendorn
(eds), The Uncommon Market: Essays in the Economic History of the
Atlantic Slave Trade, New York: Academic Press.
Giavazzi, F. and G. Tabellini (2005), Economic and political liberalizations, Journal of Monetary Economics, 52, 12971330.
Glaeser, E., R. LaPorta, F. Lopez-de-Silanes and A. Shleifer (2004), Do
Institutions Cause Growth?, Journal of Economic Growth, 9, 271303.
Grief, A. (1994), Cultural Beliefs and the Organization of Society: A
Historical and Theoretical Reflection on Collectivist and Individualist
Societies, Journal of Political Economy, 102(5), 912950.
Gwartney, J., and R. Lawson, (2005), Economic Freedom of the World:
2005 Annual Report, Vancouver: The Fraser Institute. Data retrieved
from www.freetheworld.com.
Haber, S. (2006), It Wasnt All Prebischs Fault: The Political Economy
of Latin American Industrialization, in Victor Bulmer-Thomas, John
Coatsworth, and Roberto Cortes Conde (eds), The Cambridge Economic
History of Latin America, Cambridge: Cambridge University Press.
Hall, A., and F. Peixe (2000), A consistent method for the selection of
relevant instruments, Econometric Society World Congress.
Hall, R., and C. Jones (1999), Why do some countries produce so much
more output per worker than others?, Quarterly Journal of Economics,
114(1), 83116.
02/03/2011 10:09
190
Hauk, W., Jr., and R. Wacziarg (2004), A Monte Carlo Study of Growth
Regressions, NBER Working Paper No. T0296, January.
Hausman, J., J. Stock, and M. Yogo (2005), Asymptotic Properties of the
Hahn-Hausman Test for Weak Instruments, Economics Letters, 89(3),
333342.
Herbst, J. (2000), States and Power in Africa: Comparative Lessons in
Authority and Control, Princeton: Princeton University Press.
Heston, A., R. Summers and B. Aten (2002), Penn World Table Version
6.1, Center for International Comparisons (CICUP), University of
Pennsylvania, October.
Howitt, P. (2000), Endogenous Growth and Cross-Country Income
Differences, American Economic Review, 90(4), 829846.
Inikori, J. (1977), The Import of Firearms into West Africa, 17501807: A
Quantitative Analysis, Journal of African History, 18, 339368.
Inikori, J. (1992), The Chaining of a Continent: Export Demand for
Captives and the History of Africa South of the Sahara, 14501870,
Mona, Jamaica: Institute of Social and Economic Research, University
of the West Indies.
Inikori, J. (2000), Africa and the Trans-Atlantic Slave Trade, In T.
Falola (eds), Africa Volume 1: African History before 1885, Durham,
North Carolina: Carolina Academic Press.
Jones, C. (1995), Time Series Tests of Endogenous Growth Models,
Quarterly Journal of Economics, 110(2), 495525.
Jones, E. (1981), The European Miracle: Environments, Economies and
Geopolitics in the History of Europe and Asia, Cambridge: Cambridge
University Press.
Kalemli-Ozcan, S. (2002), Does the Mortality Decline Promote Economic
Growth?, Journal of Economic Growth, 7(4), 311349.
Kalemli-Ozcan, S. (2006), AIDS, Reversal of the Demographic Transition
and Economic Development: Evidence from Africa, University of
Houston, mimeo.
Kalemli-Ozcan, S., H. Ryder, and D. Weil (2000), Mortality Decline,
Human Capital Investment, and Economic Growth, Journal of
Development Economics, 62, 123.
Kaufmann, D., A. Kraay and P. Zoido-Lobaton (2002), Governance
matters II updated indicators for 2000/01, World Bank Policy
Research Department Working Paper No. 2772, Washington D.C.
King, R., and R. Levine (1994), Capital Fundamentalism, Economic
Development, and Economic Growth, Carnegie-Rochester Conference
Series on Public Policy, 40, 259292.
Kiszewski, A., A. Mellinger, A. Spielman, P. Malaney, S. Sachs and J.
Sachs (2004), A Global Index Representing the Stability of Malaria
02/03/2011 10:09
References
191
Transmission, American Journal of Tropical Medicine and Hygiene,

70(5), 486498.
Klein, M. (2001), The Slave Trade and Decentralized Societies, Journal
of African History, 42, 4965.
Knack, S., and P. Keefer (1995), Institutions and Economic Performance:
Cross-country Tests using Alternative Institutional Measures,
Economics and Politics, 7(3), 207227.
Knaul, F. (2000), Health, Nutrition, and Wages: Age at Menarche and
Earnings in Mexico, in W. Savedoff and T. Schultz (eds), Wealth
from Health: Linking Social Investments to Earnings in Latin America,
Washington, DC: Inter-American Development Bank.
Kohli, A. (1989), Politics of Economic Liberalization in India, World
Development, 17(3), 305328.
Krueger, A. (1974), The Political Economy of the Rent-Seeking Society,
American Economic Review, 64(3), 291303.
Kuran, T. (1997), Private Truths, Public Lies: The Social Consequences
of Preference Falsification, Cambridge: MA: Harvard University
Press.
Landes, D. (1998), The Wealth and Poverty of Nations: Why Some Are So
Rich and Some So Poor, London: Abacus.
Landes, D. (2000), Culture Makes Almost all the Difference, in L.
Harrison and S. Huntington (eds), Culture Matters: How Values Shape
Human Progress, New York: Basic Books.
Landes, D. (2003), The Unbound Prometheus: Technological Change and
Industrial Development in Western Europe from 1750 to the Present,
Cambridge: Cambridge University Press.
LaPorta, R., F. Lopez-de-Silanes, A. Shleifer and R. Vishny (1999), The
Quality of Government, Journal of Law, Economics and Organization,
15(1), N1, 222279.
Lee, H. Y., L. Ricci and R. Rigobon. (2004), Once again, is openness
good for growth? Journal of Development Economics, 75, 451472.
Levine, R., and D. Renelt (1992), A Sensitivity Analysis of Cross-Country
Growth Regressions, American Economic Review, 82(4), 942963.
Levine, R., N. Loayza and T. Beck. (2000), Financial intermediation and
growth: Causality and causes, Journal of Monetary Economics, 46, 3177.
Lipset, S. (1959), Some Social Requisites of Democracy: Economic
Development and Political Legitimacy, American Political Science
Review, 53, 69105.
Lorentzen, P., J. McMillan and R. Wacziarg (2008), Death and
Development, Journal of Economic Growth, 13, 81124.
Lovejoy, P. (1982), The Volume of the Atlantic Slave Trade: A Synthesis,
02/03/2011 10:09
192
Lovejoy, P. (1994), Background to Rebellion: The Origins of Muslim

Slaves in Bahia, Slavery and Abolition, 15(2), 151180.
Maddison, A. (2004), The World Economy: Historical Statistics, Paris:
OECD.
Mankiw, G., D. Romer, D. Weil (1992), A Contribution to the Empirics
of Economic Growth, Quarterly Journal of Economics, 107(2), 407
437.
Manning, P. (1981), The Enslavement of Africans: A Demographic
Model, Canadian Journal of African Studies, 15(3), 499526.
Manning, P. (1982), Slavery, Colonialism and economic Growth in
Dahomey, 16401960, Cambridge: Cambridge University Press.
Masters, W., and M. McMillan (2001), Climate and Scale in Economic
Growth, Journal of Economic Growth, 6(3), 167186.
Mauro, P. (1995), Corruption and Growth, Quarterly Journal of Economics,
110, 681712.
Meillassoux, C. (1976), The Role of Slavery in the Economic and Social
History of Sahelo-Sudanic Africa, in J. Inikori (ed.), Forced Migration:
The Impact of the Export Slave Trade on African Societies, London:
Hutchinson & Co., pp. 7499.
Miguel, E., and M. Kremer (2004), Worms: Identifying Impacts on
Education and Health in the Presence of Treatment Externalities,
Econometrica, 72(1), 159217.
Miller, J. (1982), The Significance of Drought, Disease and Famine in
the Agriculturally Marginal Zones of West Central Africa, Journal of
African History, 23, 1761.
Miller, J. (1988), Way of Death: Merchant Capitalism and the Angolan
Slave Trade, 17301830, Madison: University of Wisconsin Press.
Mokyr, J. (1990), The Lever of Riches: Technological Creativity and
Economic Change, New York: Oxford University Press.
Mokyr, J. (2001), The Rise and Fall of the Factory System: Technology,
Firms, and Households since the Industrial Revolution, CarnegieRochester Conference Series on Public Policy, 55, 145.
Montesquieu, C. (1748 [1989]), The Spirit of Laws, New York: Cambridge
University Press.
Moore, B. (1966), Injustice: The Social Bases of Obedience and Revolt,
New York: White Plains.
Nickell, S. (1981), Biases in Dynamic Models with Fixed Effects,
Econometrica, 49(6), 14171426.
North, D. (1981), Structure and Change in Economic History, New York:
Norton.
North, D. (1989), Institutions and Economic Growth: An Historical
Introduction, World Development, 17(9), 13191332.
02/03/2011 10:09
References
193
North, D. (1994), Economic Performance Through Time, American

North, D., J. Wallis and B. Weingast (2006), A Conceptual Framework
for Interpreting Recorded Human History, NBER Working Paper No.
12795, December.
North, D., J. Wallis and B. Weingast (2009), Violence and Social Order,
New York: Cambridge University Press.
Nunn, N. (2007), Historical Legacies: A Model Linking Africas Past to
its Current Underdevelopment, Journal of Development Economics,
83(1), 157175.
Nunn, N. (2008), The Long-Term Effects of Africas Slave Trades,
OBrien, P. (2001), Fiscal Exceptionalism: Great Britain and its European
Rivals, Department of Economic History Working Paper No. 65/01,
London School of Economics, October.
OBrien, P. (2005), Fiscal and Financial Preconditions for the Rise of
British Naval Hegemony 14851815, Department of Economic History
Working Paper No. 91/05, London School of Economics, November.
Parker, P. (2000), Physioeconomics: The Basis for Long-Run Economic
Growth, Cambridge, MA: The MIT Press.
Persson, T. and G. Tabellini (2006), Democracy and Development: The
Devil in the Details, NBER Working Paper No. 11993.
Pigou, A. (1938), The Economics of Welfare, London: Macmillan and
Co.
Prakash, O. (2004), Bullion for Goods: European and Indian Merchants in
the Indian Ocean Trade, 15001800, New Delhi: Manohar.
Prakash, O. (2005), The Great Divergence: Evidence from Eighteenth
Century India, Paper presented at the GEHN Conference on Imperialism,
Istanbul, September 1112.
Przeworski, A., M. Alvarez, J. Cheibub, and F. Limongi (2000), Democracy
and development: Political institutions and material well-being in the
world, 19501990, New York: Cambridge University Press.
Putnam, R. (1993), Making Democracy Work: civic traditions in modern
Italy, Princeton: Princeton University Press.
Rigobon, R. and D. Rodrik (2005), Rule of Law, Democracy, Openness,
and Income: Estimating the Interrelationships, Economics of Transition,
13(3), 533564.
Robbins, H. (1928), A Comparison of the Effects of the Black Death on
the Economic Organization of France and England, Journal of Political
Economy, 36(4), 447479.
Robinson, J. (2002), States and Power in Africa by Jeffrey I. Herbst: A
Review Essay, Journal of Economic Literature, 60, 510519.
02/03/2011 10:09
194
Robinson, J. (2006), Equity, Institutions and the Development Process,

Nordic Journal of Political Economy, 32, 1750.
Rodney, W. (1966), African Slavery and Other Forms of Social Oppression
on the Upper Guinea Coast in the Context of the Atlantic Slave Trade,
Rodrik, D. (1999), Democracies Pay Higher Wages, Quarterly Journal of
Economics, 114, 707738.
Rodrik, D. (2000), Institutions for High-Quality Growth: What They
Are and How to Acquire Them, Studies in Comparative International
Development, 35(3), 331.
Rodrik, D. (2005), Growth Strategies, Handbook of Economic Growth,
1(1), 9671014.
Rodrik, D. and F. Rodriguez (2000), Trade Policy and Economic
Growth: A Skeptics Guide to Cross-National Evidence, in B. Bernanke
and K. Rogoff (eds), NBER Macroeconomics Annual 2000, Cambridge,
MA: MIT Press.
Rodrik, D. and A. Subramanian (2003), The Primacy of Institutions
and what this does and does not mean, Finance & Development, 40(2),
3134.
Rodrik, D. and A. Subramanian (2004), From Hindu Growth to
Productivity Surge: The Mystery of the Indian Growth Transition,
NBER Working Paper No. 10376, March.
Rodrik, D., A. Subramanian and F. Trebbi (2004), Institutions Rule: the
Primacy of Institutions over Geography and Integration in Economic
Development, Journal of Economic Growth, 9, 131165.
Rogowski, R. (1989), Commerce and coalitions: How trade affects Domestic
Political Alignments, Princeton: Princeton University Press.
Romer, P. (1990), Endogenous Technological Change, Journal of
Political Economy, 98(5), S71S102.
Sachs, J. (2003a), Institutions Dont Rule: Direct Effects of Geography on
Per Capita Income, NBER Working Paper No.w9490, February.
Sachs, J. (2003b), Institutions Matter, but Not for Everything, Finance
and Development, 40(2), 3841.
Sachs, J. (2005), The End of Poverty: How We Can Make it Happen in our
Lifetime, London: Penguin Books.
Sachs, J. and A. Warner (1995a), Economic reform and the process
of global integration, Brookings Papers on Economic Activity, 1,
1118.
Sachs, J. and A. Warner (1995b), Natural Resource Abundance and
Economic Growth, NBER Working Paper No.w5398, December.
Sachs, J. and A. Warner (1997a), Fundamental sources of long-run
growth, American Economic Review, 87(2), 184188.
02/03/2011 10:09
References
195
Sachs, J. and A. Warner (1997b), Sources of Slow Growth in African

Economies, Journal of African Economies, 6(3), 335376.
Sachs, J., J. McArthur, G. Schmidt-Traub, M. Kruk, C. Bahadur,
M. Faye, and G. McCord (2004), Ending Africas Poverty Trap,
Brookings Papers on Economic Activity, 1, 117240.
Samuelson, P. (1976), Illogic of Neo-Marxian Doctrine of Unequal
Exchange, in David A. Belsley et al. (eds), Inflation, Trade and Taxes:
Essays in Honor of Alice Bourneuf, Columbus: Ohio State University
Press, pp. 96107.
Samuelsson, K. (1993), Religion and Economic Action: the Protestant
Ethic, the Rise of Capitalism, and the Abuses of Scholarship, Toronto:
University of Toronto Press.
Schultz, P. (2002), Wage Gains Associated with height as a Form of
Health Human Capital, American Economic Review, 92(2), 349353.
Schumpeter, J. (1934), The Theory of Economic Development, Cambridge,
MA: Harvard University Press.
Shepherd, G. and F. Alburo (1991), The Philippines, in D. Papageorgiou,
M. Michaely and A. Choksi (eds), Liberalizing Foreign Trade,
Cambridge, MA: Basil Blackwell.
Shleifer, A. and R. Vishny (1998), The Grabbing Hand: Government
Pathologies and their Cures, Cambridge, MA: Harvard University Press.
Smith, A. (1776), An Inquiry into the Nature and Causes of the Wealth of
Nations, reprinted in E. Cannon (ed.) (1976), The Wealth of Nations,
Chicago: University of Chicago Press.
Solow, R. (1956), A Contribution to the Theory of Economic Growth,
Srinivasan, T. and S. Tendulkar (2003), Reintegrating India with the World
Economy, Washington D.C.: Institute for International Economics.
Staiger, D. and J. Stock (1997), Instrumental Variables Regression with
Weak Instruments, Econometrica, 65, 557586.
Stevenson, R. (1968), Population and Political Systems in Tropical Africa,
New York: Columbia University Press.
Stock, J. and M. Yogo (2005), Testing for Weak Instruments in Linear
IV Regression, in D. Andrews and J. Stock (eds), Identification and
Inference for Econometric Models: Essays in Honor of Thomas Rothenberg,
Cambridge: Cambridge University Press, pp. 80108.
Stulz, R. and R. Williamson (2001), Culture, openness and finance,
NBER Working Paper No.w8222, April.
Swan, T. (1956), Economic Growth and Capital Accumulation, Economic
Record, 32, 334361.
Tavares, J. and R. Wacziarg (2001), How Democracy Affects Growth,
European Economic Review, 45(8), 13411379.
02/03/2011 10:09
196
Tawney, R. (1926), Religion and the Rise of Capitalism, London: John

Murray.
Temple, J. (1998), Initial Conditions, Social Capital and Growth in
Africa, Journal of African Economies, 7(3), 309347.
Thelen, K. (1999), Historical Institutionalism in Comparative Politics,
American Review of Political Science, 2, 369404.
Thelen, K. (2000), Timing and Temporality in the Analysis of Institutional
Evolution and Change, Studies in American Political Development, 14,
101108.
Thorner, D. (1962), Deindustrialisation in India 18811931, in D.Thorner
and A.Thorner (eds), Land and Labour in India, Bombay: Asia Publishing
House.
Thornton, J. (1980), The Slave Trade in the Eighteenth-Century Angola:
Effects on Demographic Structures, Canadian Journal of African
Studies, 14(3), 417427.
Vamvakidis, A. (2002), How Robust is the Growth-Openness Connection?
Historical Evidence, Journal of Economic Growth, 7, 5780.
Voigtlnder, N. and H.-J. Voth (2008), The Three Horsemen of Growth:
Plague, War and Urbanization in Early Modern Europe, Working
Paper, Department of Economics, Universitat Pompeu Fabra.
Wacziarg, R. (2001), Measuring the Dynamic Gains from Trade, World
Bank Economic Review, 15(3), 393429.
Wacziarg, R. and K. Welch (2003), Trade Liberalization and Growth:
New Evidence, NBER Working Paper No. 10152, December.
Warner, A. (2003), Once more into the breach: economic growth and integration, Center for Global Development Working Paper No. 34, December.
Weber, M. (1930), The Protestant Ethic and the Spirit of Capitalism,
London: Allen and Unwin.
Wei, S.-J. (2000), Natural Openness and Good Government, NBER
Working Paper No.w7765, June.
Weil, D. (2005), Accounting for the Effects of Health on Growth, NBER
Working Paper No. 11455, July.
Williamson, J., and R. Zagha (2002), From Slow Growth to Slow Reform,
Washington, DC: Institute for International Economics.
Wooldridge, J. (2000), Introductory Econometrics: A Modern Approach,
Cincinnati: South-Western Publishing.
Wooldridge, J. (2002), Econometric Analysis of Cross Section and Panel
Data, Cambridge, MA: MIT Press.
World Bank (2002), Globalization, Growth and Poverty, New York:
Oxford University Press.
Yang, X. and J. Borland. (1991), A Microeconomic Mechanism for
Economic Growth, Journal of Political Economy, 99, 460482.
02/03/2011 10:09
References
197
Young, A. (2005), The Gift of the Dying: The Tragedy of AIDS and
the Welfare of Future African Generations, Quarterly Journal of
Economics, CXX(2), 423466.
02/03/2011 10:09
02/03/2011 10:09
Index
access to markets 1712
Acemoglu, D. 12, 14, 18, 22, 24, 25,
27, 28, 38, 44, 45, 62, 65, 72, 74,
75, 76, 87, 99, 107, 141, 148,
159
agriculture 50, 97
China 105
economic growth and development
and 345
aid programmes 1767
Alberdi, Juan Bautista 31
Albouy, D. 78
Alcal, F. 162
Alesina, A. 26
Allen, R. 33, 44
Angola 101
impact of slavery on 30, 76
Argentina
20
institutions 11314, 149
international trade 130
Arndt, C. 72
Australia 9
industrial production 13
Banerjee, A. 107
Barro, R. 150
Beaud, Michael 98
Behrman, J. 72
Bell, C. 72
Benavot, A. 51
Ben-David, D. 119
Besley, T. 46
Bhattacharyya, S. 18, 25, 30, 59, 95,
106, 108, 109
Bleakley, H. 72
Bloom, D. 35, 37, 71, 72
Botswana 143
Brazil 9
institutions 149
Calvinism 312
Canada
institutions 149
Canning, D. 72
Carstensen, K. 35, 72, 78, 83
Caselli, F. 17
causes of economic growth see root
causes of economic growth
Chang, H.-J. 41
Chile 128
China
agriculture 105
9, 10, 11, 71, 1724
technology in 43, 104
urbanization 12
Ciccone, A. 162
Clark, Gregory 33
climate 34, 50
Clingingsmith, D. 108
Coe, D. 39
Colombia 20, 24
colonialism 14, 22, 25, 27, 71
colonial institutions view 723, 778,
1079, 11012
decolonization 27
conditional convergence 17
contract enforcement 169
Cortez, Hernan 98
Cte dIvoire (Ivory Coast) 20,
128
culture and economic growth 313
Curtin, Philip 28, 29
199
02/03/2011 10:09
200
Dahomey 30
decolonization 27
Deng Xiaoping 173
Denmark 149
depopulation, slavery and 289
development aid 1767
Diamond, Jared 34, 35, 97
Dias, J. 38
diseases 96
and 35, 378, 114
Africa 100104
empirical evidence 70, 712, 73,
74, 789, 81, 83, 87, 91
foreign aid and 1767
division of labour 39
Dollar, D. 18, 40, 57, 119
Dowrick, Steve 39, 120
Drazen, A. 26
Dutt, A. 107
Easterly, W. 17, 23, 26, 41, 58, 62, 65,
73
economic growth see individual topics
education 62
empirical evidence on economic
growth 48
appropriate empirical framework
519
colonial institutions, diseases and
forced migrations 7191
connection between low-income
countries and Africa 6971
data 4851
existing results with levels
framework 679
identification issues with levels
framework 5967
Engerman, S. 23, 24, 30, 110
English language 22
Equatorial Guinea 51, 149
exchange rates 119
factor accumulation 1718
Fage, J. 28
foreign aid 1767
France
institutions 149
urbanization 12
Frankel, J. 40, 50, 119, 120, 143
future of economic growth 1668

case for multilateral intervention
1756
diseases and foreign aid 1767
recent success stories and first
principles 1724
role of technical assistance 177
setting up conditions for growth
16872
Gabon 128
Gallup, J. 34, 35, 37
Galor, O. 256, 44
Gambia 128
Gemery, H. 28, 29, 30
geography theory of economic growth
348
geography theory of economic growth
and development 348
Ghana 76
Glaeser, E. 18, 43
Golley, Jane 120
Grief, Avner 32
Gundlach, E. 35, 72, 78, 83
Gwartney, J. 149
Haiti 128
Hall, A. 82
Hall, R. 17, 21, 22, 150
Hauk, W. 157, 159
Heavily Indebted Poor Country
(HIPC) Program 1
Helpman, E. 39
Herbst, J. 26, 27, 45
heterodox capitalism 41
history of economic growth since AD
1000 914
Hogendorn, J. 28, 29, 30
Hong Kong 128
Howitt, P. 39
human capital 51, 62, 96, 171
and 425
India
colonialism in 1079
9, 10, 11, 71, 174
institutions 10610, 114, 140
02/03/2011 10:09
Index
urbanization 12
Indonesia 50, 128
infrastructure bottlenecks 1712
Inikori, Joseph 28, 29
institutions 14, 20
building 176
improving institutions with trade
policy 14041, 1456
evidence 1425
theories of trade and institutional
development 1412
institutions and trade as competitors
development 11921, 1368
data 1225
empirical strategy 1212
evidence 12536
institutions theory of economic
growth and development 216,
114
Africa 2830, 73, 75, 100104
Americas 11012
Australia 11314
China 1045
colonial institutions view 723,
778, 1079, 11012
empirical evidence 58, 62, 656, 67
India 10610
institutional weakness in Africa
268
Russia 11213
Western Europe 98100
measure of institutional quality 50,
756
unbundling 14750
which institutions matter most for
economic growth 14750, 162
data 15052
evidence 15362
improving institutions with trade
policy 14041, 1456
evidence 1425
theories of trade and institutional
development 1412
institutions and trade as competitors
development 11921, 1368
201
data 1225
empirical strategy 1212
evidence 12536
measurement of openness to 5051
slave trade 2830, 73, 75
trade openness theory of economic
growth and development 3841
interventions 1756
Ireland 50
Italy 11
Ivory Coast 20, 128
Iyer, L. 107
Japan 10
institutions 149
urbanization 12
Johnson, S. 62, 65, 72, 148
Jones, C. 17, 21, 22, 150
Kalemli-Ozcan, S. 72
Keefer, P. 21, 75, 150
King, R. 17
Klein, M. 29
Knack, S. 21, 75, 150
Knaul, F. 72
knowledge, economic growth and
425
Kraay, A. 18, 40, 57, 120
Kremer, M. 72
Kuran, T. 33
labour supply 96
Landes, D. 12, 32, 106
languages 22
Lawson, R. 149
Lee, J. W. 150
Levine, R. 17, 23, 26, 62, 65, 73
Lewis, J. 72
living standards 13
Lorentzen, P. 72
Lovejoy, P. 28, 29
Luxembourg 148
1, 2, 9
macroeconomic stabilization 170
Maddison, Angus 10, 75
malaria 35, 378, 50, 177
and 35, 378
02/03/2011 10:09
202
empirical evidence 70, 712, 73,

74, 789, 81, 83, 87, 91
Malaysia 128
Mali 128
Malthus, Thomas 96
Mankiw, G. 15, 17, 58
Manning, Patrick 28, 29, 30
markets
access to 1712
and market proximity 35
market-creating institutions 1478
market-legitimizing institutions
14950
market-regulating institutions 1489
market-stabilizing institutions 149
Mauro, P. 150
Meillassoux, C. 29
Mexico 24, 98, 110, 128
Miguel, E. 72
Miller, J. 29, 30, 38, 101
Moav, O. 256, 44
Mokyr, Joel 18, 42, 43
Montesquieu, Baron de 34
Morocco 149
Morogoro Shoe Factory 1
multilateral interventions 1756
Myanmar 51, 144
Namibia 82, 83
neoclassical growth model 1517, 39
factor accumulation and measure of
our ignorance 1718
root and proximate causes of
development 1920
Netherlands 10, 99
colonialism and 25
urbanization 11
New Zealand
Nickell, S. 157
Nigeria 82, 83
impact of slavery on 76
institutions 149
North, Douglass 21, 45, 46, 76, 124, 141
Nunn, N. 30, 72, 74, 75, 76, 78, 83, 87,
91
OBrien, P. 46
Pakistan 50
Papua New Guinea 128
Parker, P. 34
Peixe, F. 82
Persson, T. 46, 150
Peru 24, 98, 110
Pigou, A. 153
Pizarro, Francisco 98
political power theory 25
Polity IV Project 1920
population 96
slavery and depopulation 289
Portugal 50, 99
colonialism and 25, 111
predestination 312
property rights 1689, 173
public choice theory 153
public interest 153
Pushkin, Alexander 112
Putnam, R. 33
Quesnay, Franois 2
regulatory institutions 1489, 16970
religion 50
and 313
representative politics 17071
research and development (R&D) 42
Riddle, P. 51
Rigobon, R. 120
Robbins, H. 96
Robinson, J. 25, 26, 27, 28, 111, 141
Rodney, W. 29
Rodriguez, F. 40, 120, 124
Rodrik, D. 18, 22, 40, 45, 51, 58, 59,
75, 78, 87, 109, 120, 124, 125, 147,
148, 149
Roe, Thomas 106
Rogers, M. 39
Romer, D. 40, 50, 119, 120, 143
Romer, Paul 15
root causes of economic growth 956
Africa 100104
Americas 11012
Australia 11314
China 1045
India 10610
Russia 11213
Western Europe 96100
02/03/2011 10:09
Index
Rosenzweig, M. 72
rule of law 62, 169
Russia
20
urbanization 12
Sachs, J. 18, 34, 35, 37, 40, 50, 57, 71,
72, 73, 78, 119, 123, 124
Samuelsson, K. 31
Schultz, P. 72
Schumpeter, Joseph A. 42
Shleifer, A. 153
Sierra Leone 176
Singapore 50, 51
slavery 27, 38, 70, 78, 101
effects of institutions of slavery and
slave trade in Africa 2830
empirical evidence 73, 75
measurement problems 767
Smith, Adam 21, 39
Sokoloff, K. 23, 24, 30, 110
Solow, Robert 15
Somalia 50
South Africa 128
Spain 50, 99
colonialism and 25, 111
Sri Lanka 148
Staiger, D. 79, 82
state failure 175
state formation theory of economic
growth and development 457
Stock, J. 79, 82, 83
structured growth 18
theories of deep structural
determinants 2021
Subramanian, A. 109, 148
Sudan 128
Swan, Trevor 15
Switzerland 149
Tabellini, G. 26, 150
Tanzania 12, 9, 75, 143
Tawney, R. 32
taxation 456, 97
technology 43, 96, 97, 104
technical assistance 177
transfer 3940
203
Temple, J. 73
Thornton, J. 28
trade see international trade
Tunisia 76
Ukraine 50
unbundling institutions 14750
United Kingdom 99
colonialism and 25, 1079, 113
9, 10, 11, 445
institutions 149
urbanization 1112
United Nations 27
United States of America 51
9, 10, 11
institutions 149
international trade 128, 130
urbanization 11
colonialism and 22
Uruguay 130
Vamvakidis, A. 120
Vishny, R. 153
Voigtlnder, N. 37
Voth, H.-J. 37
Wacziarg, R. 123, 124, 157, 159
Wallis, John Joseph 45
war 175
and 457, 99100
Warner, A. 18, 35, 40, 73, 119, 123,
124
Weber, Max 27, 312, 45
Wei, S.-J. 40
Weil, D. 72
Weingast, Barry R. 45
Welch, K. 123, 124
Williamson, J. 108
World Bank 11920
Yogo, M. 83
Zaire 128
02/03/2011 10:09

Growth Miracles and Growth Debacles Exploring Root Causes

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Growth Miracles and Growth Debacles Exploring Root Causes

Uploaded by

Copyright:

Available Formats

Growth Miracles and Growth Debacles

M2545 - BHATTACHARYYA PRINT.indd i

To Anna, Sahana and Swapan

M2545 - BHATTACHARYYA PRINT.indd ii

Growth Miracles and

M2545 - BHATTACHARYYA PRINT.indd iii

Sambit Bhattacharyya 2011

A catalogue record for this book

ISBN 978 1 84844 631 1

Typeset by Servis Filmsetting Ltd, Stockport, Cheshire

M2545 - BHATTACHARYYA PRINT.indd iv

HISTORY AND ECONOMIC DEVELOPMENT

The Great Divergence: an account of the growth miracles and

PROMOTING GROWTH IN THE CURRENT

Institutions and trade: competitors or complements in

M2545 - BHATTACHARYYA PRINT.indd v

Wide variations in living standards are observed across countries. Going

M2545 - BHATTACHARYYA PRINT.indd vi

M2545 - BHATTACHARYYA PRINT.indd vii

M2545 - BHATTACHARYYA PRINT.indd viii

Hidaya Mohamed is a fifth grade student. She dreams of going to the

M2545 - BHATTACHARYYA PRINT.indd 1

Growth miracles and growth debacles

in the poor tropics and subtropics whose minimum aspirations cannot be

M2545 - BHATTACHARYYA PRINT.indd 2

M2545 - BHATTACHARYYA PRINT.indd 3

Growth miracles and growth debacles

M2545 - BHATTACHARYYA PRINT.indd 4

M2545 - BHATTACHARYYA PRINT.indd 5

M2545 - BHATTACHARYYA PRINT.indd 6

History and economic development

M2545 - BHATTACHARYYA PRINT.indd 7

M2545 - BHATTACHARYYA PRINT.indd 8

The Great Divergence: an account

As I have illustrated in the introduction with the Tanzania and Luxembourg

Log GDP per capita

Evolution of per capita income in USA, Australia, UK, Brazil,

M2545 - BHATTACHARYYA PRINT.indd 9

Growth miracles and growth debacles

Log GDP per capita

Source: Authors plot using data from Maddison (2004).

Evolution of per capita income in USA, UK, the Netherlands,

M2545 - BHATTACHARYYA PRINT.indd 10

The Great Divergence

Authors plot using data from Acemoglu et al. (2002).

Urbanization in the world powers over the period AD

standards across countries were largely comparable. Countries diverged

M2545 - BHATTACHARYYA PRINT.indd 11

Growth miracles and growth debacles

Source: Authors plot using data from Acemoglu et al. (2005b).

Urbanization rate in Western Europe, Eastern Europe and

growth in urbanization was largely propelled by the Industrial Revolution

M2545 - BHATTACHARYYA PRINT.indd 12

Industrial production per capita, UK in 1900 =100

The Great Divergence

Source: Authors plot using data from Bairoch (1982).

Industrial production per capita relative to the UK in the New

Figure 2.5 compares industrial production per capita relative to Britain in

M2545 - BHATTACHARYYA PRINT.indd 13

Growth miracles and growth debacles

M2545 - BHATTACHARYYA PRINT.indd 14

Theories of root causes of economic

One of the major predictions of the neoclassical growth model is that

3.1 THE NEOCLASSICAL GROWTH MODEL