Determinants of economic growth in OPEC: an extreme bounds analysis

• Date: 01 December 2014
• Author: Fatimah Hosseinpour,Mansour Zarra‐Nezhad
• Published date: 01 December 2014
• DOI: http://doi.org/10.1111/opec.12034

Determinants of economic growth in OPEC:

an extreme bounds analysis

Mansour Zarra-Nezhad* and Fatimah Hosseinpour**

*Professor of economics, Department of Economics, Faculty of Economics and Social Sciences, Shahid

Chamran University, Ahvaz, Iran

**Assistant Professor, Department of Marine Economics, Faculty of Marine Economics and Management,

Khorramshahr University of Marine Science and Technology,Iran. Email: fatemeh.hosseinpour@yahoo.com

Abstract

The aim of this study is to identify robust determinants of economic growth in Organization of Petro-

leum Exporting Countries (OPEC). Toexamine the robustness of 101 potential determinants of eco-

nomic growth in OPEC about 11 million regressions are estimated. The results of extreme bounds

analysis with the unbalanced panel data of 21 years shows that varietyof trade policy measures are

robust and support export-led growth hypothesis. Proxies of export passed very strict test of Levine

and Renelt, while oil export is not robust. The results also support conditional-convergence

hypothesis.

1. Introduction

Since Adam Smith’s time, the economic growthhas been one of the interesting subjects in

economic theories. In last decades, economic growth and its determinants have received

great attention in both theoretical and empirical aspects. One of the main reasons for this

attention is that empirical evidence has showedthat quality and standard of life differ sig-

nificantly in different countries. Forexample, evidence shows per capita income in Taiwan

increased to over 10 times during 1960–2000, while in the same period per capita income

of Congo reduce to 20 per cent of that in 1960.1Research on reasons of this disparity led to

appearance of growth theories in the late 1930s and early of 1940s. Since that time until

now growth theories attempt to find a solution for the problem of the vast differences in

standard of living over time and across countries.

Parallel to theoretical studies, empirical studies tried to identify determinants of

growth and explain differences in standards of living across countries. Evolution of eco-

nomic growth model creates hope to identify determinants of economic growth to obtain

steady state growth, but despite this hope, yet there is no consensus on true model of

growth. Hence, identifying determinants of growth needs to embark on empirical work.

This problem has more importance in developing countries than developed countries.

Sturn (1991) argue that growth theory has contributed to our understanding of how growth

424

© 2014 Organization of the Petroleum Exporting Countries. Published by John Wiley & Sons Ltd, 9600 Garsington

Road, Oxford OX4 2DQ, UK and 350 Main Street, Malden, MA 02148, USA.

is determined and influenced, but it has in manyways missed some of the crucial issues for

developing countries. Economic growth of developingcountries is not directly concerned

with the long-run rate of growth in the sense of the steady-states in the models but is con-

cerned with a medium term of some considerable duration. Therefore the focus on the long

run in theories may have been, at least in part, diversionary.

For specification of empirical growth regressions, if weaccept the variables that intro-

duced by theories and confirmed by empirical researches as determinants of growth,

multiplicity of theories and empirical studies cause to introduce large number of growth

determinants. For example, Durlauf et al. (2005) in their outstanding review introduced

about 150 variables, each of these variables at least have been statistically significant in

one empirical study and its sign has been compatible with a growth theory.It is worthy to

mention that economic growth theories, as Brock and Durlauf (2001) mentioned,are open

ended, in the sense that they are compatible with each other. Hence, if there is a set of K

theories that all of them are logically compatible, there exist2k− 1 possible specifications

of growth regression that each regression is based on a special combination of theories.

Therefore selecting explanatory variables is often ad hoc, and the results are likely to be

sensitive to the selected variables.

These issues along with measurement considerations persuaded economists to

examine variables between set of variables that identified until now as determinants of

growth, instead of following solely theories. Many empirical studies tried to determine

variables that influence economic growth onlyusing one or few regressions. Although the

results of these studies might be logical and compatible with the theories, but the results

could differ when changing the specification. Thus rely on these results might be diver-

sionary. This weakness has been pointed out, among others, by Leamer (1983) where he

emphasised that under uncertainty of model selection one must show how much the result

depends on which variable are included in the regression. Therefore one should subject

regressions to change in specification. This sensitivity analysisprovides a convincing jus-

tification for removal or inclusion of individual variables in the probably true regression.

One of the best approaches for selecting main determinant among vast potential determi-

nants is extreme bounds analysis (EBA). This approach is attributed to Leamer (1983,

1985). Levine and Renelt (1992) applied Leamer’s extreme bounds test for the first time to

identify robust empirical relations discussed in the economic growth literature. Levineand

Zervos (1993) pointed out the EBA helps clarify the degree of confidence that can be

placed to the partial correlations between growth and individual variables. If an indicator

is roboustly correlated with long-run economic growth, then one should feel more confi-

dent about its association with growth than an indicator that has a fragile link.

As Wong(2009) has pointed out, there are some countries that have different resources

of government revenues. These countries have a great quantity of natural resources that

made remarkable revenue for their governments. Special composition of government

Determinants of economic growth 425

OPEC Energy Review December 2014© 2014 Organization of the Petroleum Exporting Countries