DP17026 What is the optimal minimum wage?
The extensive literature on minimum wages has found evidence for the compression of relative
wages and mixed results for employment. This literature has been plagued by a number
of problems. First, the median-minimum wage ratio is used as the independent variable, where
the median is endogenous. Second, it is difficult to disentangle compression of relative wages
and truncation due to employment effects. Third, all effects are likely to depend on the initial
level of the minimum. Fourth, employment effects are likely to differ between worker types.
We offer solutions for these problems, by using instruments for the median, by using data on
personal characteristics, and by using a flexible specification. We apply our method to US data
starting from 1979, allowing for a wide variation in minimum wages. We find strong compression
and positive employment effects for the lower half of the distribution, persisting for quite
high levels of the minimum.
