This project continues research into the operation of labor markets. Under the previous award the investigator published a number of papers focusing on a variety of related topics including: (a) the determinants of earning and the distribution of income, (b) an analysis of the impact of manpower training programs on earnings, the value of longitudinal data, and the more general issues of the econometrics of selection-bias models, (c) the development and testing duration analysis models. This project will extent the previous work in terms of three broad topics. The first is on the development and assessment of non-experimental estimators for selection-bias problems. The methodology is applied to the further evaluation of the impact of manpower training programs on earnings, employment, and unemployment. The benefits of various types of econometric estimators will be assessed. The benefits of longitudinal and experimental data well also be explored, and data sets analyzed in the previous work will be re-analyzed and a variety of new estimators will be developed. The second topic is on econometric duration analysis. An estimator for duration models using data with unobservable components will be developed and evaluated. The third topic focuses on the analysis of black wage-structures and earnings in the South Carolina labor market during the 1910 to 1980 period. Data from this period will be used to assess the relative importance of government pressure, the growth in schooling, and secular and cyclical market factors in accounting for the breakthrough in black employment and wages observed in that state during the mid 1960's.