Inequality and Development - A Regression Analysis of the Cross-Country Pattern

The purpose of this paper is to reinvestigate the relaionship between economic development and size distribution of income. In this field, the inverted U-curve, or Kuznets curve, describing the inequality trend in the course of economic development, is often viewed as a "stylized fact" of development literature, established mainly by a number of cross-country studies. Compared to these earlier investigations the present paper uses a different data selection procedure, updates the figures, and is thus testing hypotheses by the means of regression analysis within the framework of a new set containing 64 countries. The U-hypothesis in its GDP per capita version, using a global scale, is rejected. However, the theoretical expectations are supported, if one of the following three modifications is used: application of a subsample of only LDC-countries, logarithmic transformation of GDP, or structuralist interpretation of the development process. In these cases the models have a strikingly high explanatory power, but quite different predictions about the form of the estimated U-curve. It is concluded that the cross-country U-curve is not only heavily dependant on the theoretical assumptions made, but also significantly affected by differences in the data selection procedure and by the appearance of new data.