Distributional Consequences of Globalisation: Empirical Evidence from Panel Data

Abstract

This paper investigates the impact of globalisation on cross-country inequality using a large panel dataset. The findings reveal that location and capital are the main determinants of inequality; trade intensity and foreign direct investment make only a small contribution (approximately 4%). The relative contributions of trade and foreign direct investment to inequality have changed little over time and have certainly not increased at the same rate as the rise in global trade and investment activity. Hence, globalisation does not emerge as a significant factor in driving cross-country inequality. Differences emerge when countries are grouped by relative income, but the main findings persist.

Notes

1. The term globalisation encompasses a wide array of phenomena, including falling trade barriers and increased labour mobility, capital flows, technology transfer, exchange of information and culture, and so forth. In this paper, however, we follow the norm in economic literature and define globalisation as increased international movement of goods, services, and capital flows only.

2. At least at the time of writing in 2006. Earlier studies using this approach have focused on decomposing inequality within countries, for example, Wan et al. (2007).

3. Two notable exceptions to these results are Dikhanov and Ward (2001) and Milanovic (2002). Both studies use PPP-adjusted income measures with countries weighted by their population shares and report that world inequality, measured by the Gini coefficient, increased markedly from 1988–1993. This is because these studies use actual household income and expenditure surveys to calculate within country inequality and find that the rise in income disparities between the rural and urban areas, especially in China and India, makes world income distribution highly unequal.

4. See Harrison (1996) for an overview on different measures of trade openness.

5. Some other widely used measures of institutions include the rule of law index constructed by Kaufman et al. (2002) and the index of protection against expropriation used by Acemoglu et al. (2001). We, however, prefer to use the freedom index because of its wide coverage.

6. The basic specification of the gravity model of bilateral trade flows assumes that trade is determined by a country's economic size, measured by its income, population and geographical land area, and the proximity of its trading partners, measured by distance or the transportation costs.

7. The economic freedom index takes into account factors like trade openness, foreign investment inflows, conduct of monetary policy, government intervention, protection of property rights. Each country is assigned an integer between one and five (with one representing freedom and five representing economic repression). Since data for this variable is available from 1995–2004 only, the sample size is reduced from 991 to 529 observations when we include it in Equation (1).

8. Some earlier studies find a significantly positive relationship between economic freedom and per capita income in cross-country regressions (for example, Islam, 1996; Miles et al., 2006). However, our results indicate that the effect of economic freedom on per capita income is insignificant (see Table A3). This could be attributable to high collinearity between this index and the trade and FDI variables included in our equation.

9. The decomposition is performed using a Java programme developed by the United Nations University-World Institute for Development Economics Research (UNU-WIDER). This software allows decomposition of inequality under any functional form and any inequality index.

10. The relative contributions are calculated using the total explained portion of the inequality as the denominator and therefore sum to 100 per cent.

11. The detailed decomposition results are presented in the Appendix.