Export-led Growth in South Asia: A Panel Cointegration Analysis

Abstract

In this paper, we attempt to examine the export-led and manufacturing export-led growth hypothesis for four South Asian Countries; namely, India, Pakistan, Bangladesh and Sri Lanka, using Pedroni’s panel cointegration technique for the period 1980–2002. In this context we estimate growth accounting equations to investigate the impact of exports, manufacturing exports and other important physical and human capital variables on both total GDP and non-export GDP. The study finds long-run equilibrium relationship between GDP (and non-export GDP) and exports along with other variables supporting export-led growth hypothesis. The results also substantiate the existence of manufacturing export-led growth hypothesis. Further, we find that export, fixed capital formation, public expenditure on health and education have statistically significant coefficients re-emphasizing the importance of these variables for higher economic growth.

Keywords:

• Exports
• economic growth
• South-Asia
• panel cointegration

Acknowledgements

The authors thank Dr Arup Mitra, Dr Geethanjali Nataraj and Dr Sabyasachi Kar for their comments on the first draft of the paper. We are grateful to two anonymous referees for their useful comments and suggestions. However, the usual disclaimer applies.

Notes

¹See Kruger (1975) for the rent seeking activities in India.

²During 1981–1991, East Asian countries Korea, Malaysia and Thailand experience 9.1, 5.7 and 8.2% of GDP growth and 11.3, 11.8 and 15.5% of exports of goods and services respectively. Growth of GDP and exports of these countries during the 1990s fell slightly below the average growth of 1980s due to economic crisis of these countries in 1997–98. During 1991–2001, these three countries registered 5.4, 6.1 and 3.2% of GDP and 15.9, 10.5 and 8.5% of exports respectively.

³Sri Lanka started liberalization its economy in late 1970s, Pakistan in early 1980s and India and Bangladesh in early 1990s.

⁴See Baltagi & Kao (2000) for detail discussion on advantage of panel cointegration.

⁵South Asian countries include India, Pakistan, Srilanka and Bangladesh. Here we have excluded Nepal due to data unavailability.

⁶The crisis year 1991 has been deducted while calculating the growth rate for 1990–1991 to 2000–2001.

⁷The growth rate during 1991–2001 would have been much higher than the existing rate in the latter period had not it faced negative growth rate by−1.4% in 2001.

⁸See Medina-Smith (2001) for an extensive review of the literature.

⁹It is quite surprising that two studies such as Ekanayake (1999) and Doganlar (2004) find different results even after using the same techniques, same variables and equal time period. When the former study finds bidirection causality between real export and real GDP for Thailand, Pakistan and Sri Lanka, the latter on the other hand shows export to growth for the first country and growth to export for the last two countries.

¹⁰Alternative measures of growth include per capita GDP and labour productivity (Marin, 1992).

¹¹It is important to note that this strategy has been widely used by researchers engaged in testing the ELGH for both cross-section and country case studies of developing countries and even for industrialized nations.

¹²Physical capital needs two sets of information, namely the initial base year for the capital stock and the rate of depreciation, which are difficult to obtain.

¹³In these two equations and alternative equation (given in Appendix B), we have dropped two variables such as life expectancy and labor after checking their unit roots.

¹⁴The missing figure of this variable is calculated using the method similar to the life expectancy variable.

¹⁵See Baltagi & Kao (2000) for a detailed discussion on the advantage of panel cointegration.

¹⁶For detailed estimation procedures see Levin & Lin (1992).

¹⁷Generally, the panel cointegration tests developed by Kao & Chiang (1999), Pedroni (1995, 1997) are also widely used.