As a consequence of this earlier work, the minimum unemployment rate faced by the individual since beginning his or her current (adult) job is also constructed. This minimum rate is included in the wage equation, and then the earlier unemployment rates (those from the first five years in the labor market) are used as instruments. Replicating the Beaudry and DiNardo result, the evidence generally shows that the minimum unemployment rate on the current job is strongly negatively related to contemporaneous wages, and that once this variable is included, neither the unemployment rate at the beginning of the job nor the current unemployment rate is significantly negatively related to contemporaneous wages.  More specifically related to the analysis of this paper, the evidence typically indicates that the early unemployment rates are valid instruments (in the sense of passing overidentification tests) only when the minimum unemployment rate on the current job is included, as the Beaudry and DiNardo results would lead us to expect; these results are reported in the tables below. payday loans lenders

Of course, the unemployment rate that an individual faces in a geographic area is not necessarily exogenous, since there is always the possibility of mobility. Because of this, the sensitivity of the results to using the unemployment rate in different forms is explored. First, the actual unemployment rate for each individual for each year, Uijt, is used, where i, j, and t index individuals, entry year, and year in the labor market (i.e., 1-5). Next, the cohort/year variation, which is more plausibly exogenous, is isolated by constructing
where Njt is the number of individuals in entry cohort j in year t, and i e jt indicates that the summation is taken over all members of a particular entry cohort j in year t in the labor market. U jt measures the average unemployment rate faced by a member of cohort j in their t-th year in the labor market. Although this average rate is still potentially endogenous because labor market conditions could affect the timing of labor market entry, this measure removes the cross-sectional variation within a year that is most prone to endogeneity bias from migration and residence decisions.