Essays on Wages in the Danish IDA Database, Selection, and Firm Performance

In this dissertation, I uncover how the wages in the Danish Integrated Database for Labor Market Research (IDA) are estimated. Using old program codes and other nontraditional sources, I document the estimation method and demonstrate that it has significant time series and compositional problems. Due to the gradual decline in working hours in Denmark from 1986 to 1990, the estimated wages of part-time workers increased rapidly just to fall steeply from 1992 to 1993 after the implementation of a labor market pension reform. Moreover, the wages of new hires are too high which makes for negative or zero returns to tenure. I undertake a complete revision and propose a new wage measure that solves both problems. I then use the new wage measure and the longitudinal features of the database to examine the relationship between labor dynamics, firm dynamics, and wages. I find that growth in employment and average wages are negatively related at the establishment level. This is due to a selection effect that operates through both the hiring and separation margins. Relative to wages of continuing workers, low-wage workers make up the bulk of separations and hires, leading to positive selection at shrinking establishments and negative selection at expanding establishments. Further analysis suggests that the wage differences are primarily due to worker heterogeneity. In fact, worker characteristics can account for both the level and the variation of the relative wage of separations. Worker characteristics also account for half of the level and sixty percent of the variation of the relative wage of hires. Job heterogeneity accounts for the remaining part of the levels, and heterogeneity in wage offers to similar workers occupying similar jobs account for most of the remaining variation.