A Comprehensive Analysis of Inequality in the Size Distribution of Income for the United States, 1952-1981

Abstract

Numerous studies have attempted to estimate inequality in the United States size distribution of income. One standard procedure has been to conjecture that a particular hypothetical statistical distribution function approximates the actual empirical income distribution and then to equate the statistics of the actual empirical income distribution to the parameters of the hypothetical distribution. Kakwani [7] offers a survey of this work. All of these studies have focused on the narrow total income aspect of analyzing comprehensive economic inequality. Some have questioned whether so much emphasis should be placed on income inequality, per se, since income is only one aspect of welfare. Franklin Fisher [5] long ago made this point, but only recently did researchers begin to utilize other available information in making welfare valuations. Since income received and expenditures made by consumers are not statistically independent, a more efficient method of measuring inequality in the marginal distribution of income is to derive the parameters of the marginal distribution of income from the estimated parameters of the joint distribution of income and expenditures. Maasoumi [9] and Jorgenson and Slesnick [6] recognized the importance of inequality in attributes such as schooling, expenditures, and the quality of life in constructing a comprehensive measure of inequality. They attempt to find a single measure that describes the inequality inherent in all the attributes. They did not, however, utilize information about the interdependence of these attributes to derive more efficient measures of the inequality in any particular attribute. This paper utilizes a joint distribution of components of income and commodity group expenditures to analyze inequality in the marginal distributions of various components of income. Estimation efficiency' is gained by recognition of the interdependence of the income