where the dependent variable, lnWit, is the log of the real hourly wage (in 1999 colones) of
individual i at time t (1988.1999). The explanatory variables include the log of the real
minimum wage (in 1999 colones) that applies to that worker’s occupation category in each year,
ln MWit. The coefficient α1 is an estimate of the impact on average actual wages of changes in the
legal minimum wage. Other explanatory variables include the vector Xit, of individual specific
human capital variables (years of education, a quadratic in experience, gender, and full
interactions among these variables), and Zit, the value-added in the industry of the individual’s
job in year t.8 We also include dummy variables for each occupation categories, OCCitj (j =
1...350), in order to control for occupation-specific fixed effects and for the endogenous
correlation of wages and minimum wages across occupation categories.9 Finally, to control for
endogenous changes in yearly average minimum wages (as well as other year-specific factors
such as aggregate supply and aggregate demand changes, the timing of minimum wage
changes,10 or design changes in the household surveys) we include a dummy variable for each
year, YRt (t=1989-1999). After including these two sets of dummy variables (for occupation and
years), our resulting estimates of the impact of legal minimum wages on wages are based only on
deviations of each minimum wage from the average minimum wage change within occupation
categories over time. We next argue that these changes can reasonably be thought of as
exogenous.
A major problem plaguing the empirical minimum wage literature is one of endogeneity
bias that arises if minimum wages are set according to changes in demand and supply conditions,
which also affect wages. This occurs, for example, when the level of minimum wages is adjusted
by the amount of inflation in the previous period, a practice followed in Costa Rica and common
in many countries. However, we argue that a special feature of Costa Rica’s minimum wage
policy over this period assures us that we do not have a simultaneity problem in our estimations.
During the period under study, the government of Costa Rica implemented a policy of gradually
reducing the number of minimum wages from over 500 categories (set by the industry and
occupation of the worker) in 1988 to 19 categories (set by skill only) in 1997. The process of
simplification made changes in the relative minimum wage within occupational categories
exogenous over this period. The reason for this is that relative changes in minimum wages within
8 We use the ISIC at the one digit level.
9 These occupation categories correspond, as best as we can make them, to the categories in the 1988 minimum
wage legislation.
10 Minimum wages were set typically in January and July of each year, but sometimes they were set a little earlier or
later. See Appendix Table A.1 for the exact timing minimum wage setting over this period.
9