Information on local unemployment rates is used to measure youth labor market conditions that might have exerted an exogenous influence on youth labor market experiences, as explained more fully in the next section. These unemployment rates come from the NLSY Geocode data file, and are based on state and area labor force data from the May Employment and Earnings covering March of each year; for most states and areas, these rates are based on more than just the household data from the CPS. The Geocode file includes the rates for metropolitan areas for which unemployment rates are reported in Employment and Earnings, for individuals residing in those areas, and rates for the rest of the state for other individuals.
Descriptive statistics for the samples of men and women are reported in Table 1. Because interest centers primarily on the regression estimates, these statistics are unweighted. The proportions non-white are high because of the NLSY oversample of blacks and Hispanics. The figures reveal many typical features. Men’s wages are higher by 26 percent on average, and the average experience and tenure of men are higher, while women have higher average schooling. The average year of labor market entry, as defined here, is in the middle of 1980. The average annual unemployment rates experienced by individuals in the sample range from 7.85 to 9.86 percent over the five-year post-schooling period. Finally, both the longest tenure attained in the five-year post-schooling period and the number of jobs held are slightly higher for men than for women, suggesting that men go through more jobs initially, but also settle into a somewhat longer job in this post-schooling period.
The equation that is estimated is of the form
where w is the adult wage, X is a vector of standard contemporaneous labor market characteristics included as controls, and S is a variable measuring early job stability. To control for simple individual heterogeneity that may be correlated with both wages and early job stability, the AFQT test score (standardized for age) and parents’ education are also included in X. Although the estimated regression is cross-sectional, time subscripts are included to clarify the time at which the variables are measured. Those with a t subscript are measured as an adult, and those with a t’ subscript are measured during the initial five-year post-schooling period. payday loan debt assistance