Income inequality and economic progress: an empirical test of the institutionalist approach

American Journal of Economics and Sociology, The,  Jan, 1996  by Kang H. Park

• E-mail
• Print
• Link

In addition to the inequality variable, three additional variables that are logically associated with the magnitude of socio-political instability are considered in model [1]. These are GNP per capita, population, and the separatism index in 1975 available from Taylor and Jodice.(14) GNP per capita is considered to signify the effect of the overall level of well-being of the society compared to the effect of relative well-being (income distribution). Some (e.g. Ogwang, 1995) used the Human Development Index (HDI) developed by the United Nations Development Program to measure the level of well-being or development of the society.(15) The population variable is considered to normalize the frequency of the dependent variable. The number of violent events or deaths is likely to be higher in more populous countries, but not necessarily proportional. The separatism index is developed as a product of two variables, the percentage involved in separatist movements and the intensity of these movements, both available from Taylor and Jodice.(16) This variable is expected to have a detrimental effect on socio-political stability.

No single indicator currently available can properly represent economic progress in model [2] as defined by the institutionalists. However, among the currently available indicators, the rate of growth of GNP per capita would be the most meaningful indicator of the progress in economic well-being in the society. The rate of growth of GNP per capita instead of GNP itself may better represent changes in the standard of living of individual households in the society. The average growth rate of GNP per capita for the period of 1968-1977 was used in this study. In model [2], we are mainly concerned with the effect of income inequality on economic progress. However, we also included the average growth rate of gross domestic investment for the period of 1968-1977 in explaining economic growth for the same period, since a growth model cannot ignore the effect of technology advance or capital accumulation on economic growth.(17)

Results from a series of multiple regressions of model [1] are given in Table 1. Lower [R.sup.2]s are expected because of the nature of cross-sectional regression. The basic model only includes the population variable (in logarithm) and the inequality variable (Gini index in percentage). The magnitude of socio-political instability measured by deaths from political violence (in logarithm) is positively and strongly related to both population and the degree of income inequality as regression 1 reports expected, as well as significant coefficients for both variables. GNP per capita and the separatism index are added in regression 2 and regression 3 respectively. GNP per capita has a significant negative effect on socio-political instability operationalized by deaths from political violence while the separatism index is positively related to this instability. The addition of either variable increases the explanatory power by about .15 (61% improvement). When all four variables are included in regression 4, all the variables have theoretically expected signs and significant coefficients at the 1% or 5% levels and there is an improvement of [R.sup.2] to .477 (94% improvement over regression 1). Replacing the Gini index by the income share of the top 20 percent of population or replacing GNP per capita by HDI do not change the regression results much in terms of the significance of the coefficients and the explanatory power of the model. The Pearson correlation coefficient between Gini index and the top 20 percent income is .946, and the HDI is correlated with GNP per capita at .886 and with the Gini index at -.431 for the matching sample.(18)