ENDOGENOUS ECONOMIC POLICY AND THE STRUCTURE OF PRODUCTION: THEORY AND EVIDENCE
| Published date | 01 May 2007 |
| Date | 01 May 2007 |
| Author | Kwok Tong Soo |
| DOI | http://doi.org/10.1111/j.1467-9485.2007.00412.x |
ENDOGENOUS ECONOMIC POLICY AND
THE STRUCTURE OF PRODUCTION:
THEORY AND EVIDENCE
Kwok Tong Soo
n
Abstract
This paper develops and tests a model that predicts a positive relationship between
absolute levels of capital stock and how favourable are policies toward capital. The
theoretical model we use is a model of campaign contributions and electoral
competition, extended to consider the implications for factor mobility and hence
the structure of production. There are two main predictions. First, countries with
more capital stock tend to implement more pro-capital policies. Second, in a two-
country model, the country that initially has more capital will be able to attract
capital inflows from the other country. Given additional assumptions on the
production side, this yields the prediction that the more different are countries’
policies, the more different will be the set of goods that they produce. These
predictions of the model are confirmed using panel data on cross-state differences
in policies and economic outcomes in India.
I Intro ductio n
Absolute advantage, broadly defined as absolute superiority in technology or
factor endowments, has rarely played a major role in theories of international
trade. Traditional trade theory, from Ricardo to Heckscher and Ohlin, has
stressed the role of comparative rather than absolute advantage as the reason for
trade. New trade theories in the Helpman and Krugman (1985) mould show the
impact of increasing returns in determining the pattern of trade. This paper
seeks to address a simple question: can absolute advantage, through its impact
on policy, play a role in international trade and the structure of production?
We propose a model of trade between two locations
1
that makes the
following two predictions. First, the location with a greater absolute stock of
capital will have policies that are more favourable to capital.
2
Once we allow for
capital to move across locations, the location with greater absolute capital stock
n
Lancaster University.
1
The mechanism underlying the model is more general than the precise two location setup,
and could be extended to multiple locations.
2
A note on terminology: we use ‘locations’ to distinguish between geographical areas. We
prefer this term to alternatives such as ‘countries’ or ‘states’ as the mechanism in the model
Scottish Journal of Political Economy, Vol. 54, No. 2, May 2007
r2007 The Author
Journal compilation r2007 Scottish Economic Society. Published by Blackwell Publishing Ltd,
9600 Garsington Road, Oxford, OX4 2DQ, UK and 350 Main St, Malden, MA, 02148, USA
220
will attract inflows of capital from the other location, changing relative capital–
labour endowments, so that the location with an initial absolute advantage in
capital stock, will end up having a comparative advantage in the capital
intensive good. Therefore, the second theoretical prediction is that locations
with more different policies will also have more different industrial structures.
Both of these predictions are then tested empirically.
The way the model works is as follows. We start with a standard model of
trade with identical technologies but different relative endowments of capital
and labour across locations. Then, comparative advantage implies that each
location will export the good that uses intensively its relatively abundant factor.
However, capital owners are able to lobby the government for policies that
favour them, at the expense of labour; labour owners are assumed to be
unorganised and hence unable to counter-lobby the government. The location
with greater absolute levels of capital will therefore implement policies that are
more favourable to capital, because the cost of policy is spread over more units
of capital.
3
Once we allow for capital mobility, lobbies in each location play a two stage
game in which they first simultaneously choose whether to lobby their respective
governments or not, and then whether to stay in their original location or move
to the other location. In general there are two (pure strategy) Nash equilibria, in
which all the capital in the world locates in either location. However, if initial
absolute capital endowments are different across locations, the two equilibria
are not symmetric, and we can obtain conditions for which we have a unique
Nash equilibrium, that all the capital in the world will locate in the location that
has the larger initial stock of capital. This somewhat extreme outcome is a result
of a simplifying assumption we make on the technology side that makes the
model much simplerto solve. Relaxing this assumption wouldresult in incomplete
capital relocation, which is what is observed in reality. Therefore, while
comparative advantage matters for current patterns of trade, initial absolute
endowments of capital determine policies and hence the future location of
capital, future comparative advantage, and future patterns of production. This
can lead either to the reinforcement or reversal of initial comparative advantage.
The key contribution of this paper is to combine the microfoundation of
policy with comparative advantage and trade patterns. Following the lead of
Grossman and Helpman (1996), we endogenise the relative weights which the
government places on campaign contributions vs. voter welfare, and show that
the share of tax revenue derived from capital depends on how effective is
lobbying activity, how dispersed are voter preferences for different parties in an
election, and how well informed are voters. Our two main contributions relative
to Grossman and Helpman (1996) are, first, to incorporate a Heckscher–Ohlin
production structure with two sectors that use two factors of production with
would operate irrespective of the geographical area involved, as long as different locations have
different policies and capital is mobile between locations.
3
If the government-provided good is a public good that is non-rival in consumption, e.g.
national defence, then the benefits of government policy do not decline with the units of capital.
ECONOMIC POLICY AND STRUCTURE OF PRODUCTION 221
r2007 The Author
Journal compilation r2007 Scottish Economic Society
different intensities. Second, we allow for capital mobility across locations. Both
of these new dimensions of the analysis substantially change the predictions of
the model.
There have been previous models on the impact of endogenous policy
formation on capital mobility; examples include Persson and Tabellini (1992)
and Haufler (1997). At least two key features distinguish our model from
previous models. First, it is based on a two-good, two-factor model of
international trade; this allows us to derive some results on the impact of factor
mobility on the structure of production. Second, and fairly crucially for our
results, the government in our model maximises the probability that it will win
an election. In the process of doing so, it has to compete with a rival party using
a combination of campaign contributions and policies, but this results in the
capital lobby being able to exploit this rivalry to maximise its own welfare at the
expense of the political parties.
The model of absolute advantage which we present also differs from previous
models in which absolute advantage is an explanation for trade, for example
Copeland and Kotwal (1996), Ricci (1999), or Neary (2003). In Copeland and
Kotwal (1996), absolute (technological) advantage combined with non-homo-
thetic preferences may lead to no gains from trade, so that absolute advantage
reduces trade flows. Ricci (1999) allows for technological differences across
locations in a two-location model of economic geography with increasing
returns at the level of the firm, and finds that agglomeration may occur in the
location with an absolute technological disadvantage, if the difference in
productivities between the two countries is not too large. In Neary (2003), a fall
in production costs in one country in a many-sector oligopolistic general
equilibrium model may lead both countries to specialise less in accordance with
comparative advantage through changes in factor prices. Differently from these
papers, in our model, absolute advantage in terms of absolute capital stock has
an indirect effect on patterns of trade, by influencing the location of capital
through its impact on policy. Trade in goods, taking factor endowments as
given, is still governed by comparative advantage.
In the theoretical model, we model the government’s policy variable as the
tax rate on capital and labour. In our empirical analysis, we take a broader
interpretation of policy. We test for the impact of capital endowments on
economic policy using state-level data on India between 1959 and 1997, using
Besley and Burgess’ (2004) labour regulation indicator as a measure of how
favourable is the policy stance of a state to capital.
As capital stock is endogenous in our model, our empirical strategy uses
instrumental-variables methods to overcome this problem. We instrument
capital stock using electricity generating capacity and bank credit. These are
intuitively appealing instruments as they are highly correlated with capital stock,
but at the same time, are not correlated with the error term, as they are primarily
determined by the central government. We discuss the instruments in greater
detail in Section III.
Both predictions of our theoretical model are confirmed in the data. First,
even after controlling for numerous other factors, absolute capital endowments
K. T. SOO222
r2007 The Author
Journal compilation r2007 Scottish Economic Society
To continue reading
Request your trial