Income Distribution in the Third World: Its Estimation Via Proxy Data

Abstract. A two‐Stage least squares estimate of the distribution of income in the Third World is derived in this paper using the per capita ownership of cars, infant mortality rates, and the average daily caloric requirement along with the per capita Gross Domestic Product. Previous work by Kitznets 1955 had established a relationship between the distribution of income and GDP/CAP, but with the inclusion of the three additional “proxy” variables, the distribution of income is estimated with a great deal more precision. For example, the R‐squared for the estimate of the share of income earned by the poorest 20% of households increases from 0.30 to 0.68 by incorporating the proxy variables. Using the parameters estimated via two‐stage least squares on a set of 23 countries for which the distribution of income Is known, the paper then estimates the distribution of income for a set of 43 countries for which this data is unknown. The results indicate that countries like Singapore and Sri Lanka have relatively even distributions of income for their stage of development, and countries like Brazil, Kenya, Bolivia, and Gautemala have highly skewed distributions of income for their level of GDP/CAP.