An Application of Panel Regression to Pseudo Panel Data

Abstract

This article illustrates how, in the absence of true panel data, multivariate regression analysis can be used in conjunction with a pseudo panel data set to identify variables that were related to the increase in the proportion of two-income spouses in the United States between 1940 and 2000. We present the procedures used to form the pseudo panel data set, construct and estimate the various models used to analyze the pseudo panel data, and interpret the results produced by those models. Our analysis revealed an inverse relationship between the proportion of two-income spouses and the presence of young children as well as an increasing trend across generations in the proportion of two-income spouses.