- Much of the observed cross-state variation in the distribution of income can be explained by a model which includes variables representing eight factors describing economic and demographic characteristics of each state. These eight factors are education, race and sex, employment level, occupational and industrial structure, the level and composition of income, the size of the nonproductive population, and other demographic characteristics such as age and the proportion of unrelated individuals in the state., In the process of developing and testing the above model, the study had two other major objectives: (1) Calculation of measures of income inequality for each state in 1975 from the microdata files of the 1976 Survey of Income and Education (SIE), and comparison of these distributions with those from the 1949, 1959, and 1969 Census of Population. (2) Replication (as nearly as possible) of three prior studies of cross-state inequality differences using data from the SIE, and observation of the effect of changing, successively, the recipient unit, the measure of inequality, and measure of the level of income (adjusting for interstate price level differences)., The model finally tested attempted to explain cross-state inequality differentials in the income received by families and unrelated individuals. In constrast to earlier studies which used the Gini concentration ratio as both the measure of inequality and the dependent variable, our model uses a set of six quantile income shares to represent the income distribution. Each quantile share serves as the dependent variable in one equation of a multi-equation model. In this way, we can observe the effects of the explanatory variables on each part of the distribution. In some cases, this technique provides considerably more information about the relationship between the explanatory variables and inequality than does a single equation, Gini-ratio model., The model shows that cross-state inequality differentials are, in general, positively related to both the level and the standard deviation of education in the states (as predicted by the human capital model and by Thurow's job-competition model). However, these two education variables interact in their effect on income inequality. As a result, there are some unusual relationships, particularly in the middle quantiles., and Other interesting results describe the interaction between the distribution of education and racial balance in state and the importance of a set of occupational variables broadly classified by skill level requirements. In general, the model is an attempt to explain inequality differences in terms of the supply and demand for labor of various types, with other variables to describe the social and institutional structure of the states. An examination of significance levels, both T-values for individual coefficients and F-values for variables across the set of share equations, indicates the usefulness of the model for the explanation of inequality differentials among states.
- Dissertation Note:
- Ph.D. The Pennsylvania State University 1980.
- Source: Dissertation Abstracts International, Volume: 41-09, Section: A, page: 4099.
View MARC record | catkey: 13610998