Robust estimation of the cross-sections of United States wages and stock returns

The returns to worker characteristics are studied using the conditional mean function and conditional quantile function and decompositions of the Lorenz curve and Gini coefficient. The change in the returns to worker characteristics is also examined, as well as the change in the 90th and 10th percentiles of the distribution of wages, the Lorenz curve, and the Gini coefficient. The Lorenz curve and Gini coefficient are expressed using the conditional quantile function yielding measures of wage inequality that are consistent with the parametric model of the structure of wages. The changes in these measures of wage inequality over time are decomposed into the portions associated with the changing returns to the covariates and the changing distributions of the covariates, yielding a deeper understanding of the decline in wage inequality and how it is associated with the difference in the returns to and distributions of worker characteristics. The majority of the decline in wage inequality found in this study is attributable to the changing returns to worker covariates, but the changing distribution of the covariates also caused wage inequality to decline.; The purpose of the study of the cross-section of U.S. stock returns is to examine how the relationship between returns and a conventional measure of risk differs across the conditional distribution of returns. The conditional quartile functions show that the relationship is not constant across the conditional distribution implying that the independent variables have differential effects on the stock return given the quartile examined. The relationship between stock returns and size (market equity) and book-to-market value is not constant across the conditional distribution of returns. The size effect has its usual sign, negative, for the largest returns, but is positive for the median and smallest returns. Book-to-market value has an inconsistent relationship with the return on stocks depending on the other included independent variables. There is some evidence, however, that the large returns are positively related to book-to-market value while the median and small returns have no relationship.