William Davidson Institute Working Paper 479
2.2 Employment Effects
To estimate the employment affect of the minimum wage we estimate separate
regressions for the covered and uncovered sector workers using two different measures as a
dependent variable: a) the log of the number employed (by industry and year) and b) the total
number of hours worked in an average week (by industry and year). These measures allow us to
learn about the reallocation of workers and hours worked across industries over time. We also
examine whether there is a reallocation of labour within an industry between covered and
uncovered employment as the “toughness” of minimum wages rises. For this we use the percent
working in the covered sector in each industry (and year) as a dependent variable. All three of
these measures are regressed on various specifications of the “toughness” measure with industry
and time fixed effects.
2.3 Data
The analysis is based on aggregated data from the Costa Rican annual Household Survey
of Employment and Unemployment (HSEU). The minimum wage data were taken from the
Gazetta published by the Ministry of Labor. Since it was not possible to match the occupational
codes in the HSEU with those in the Gazetta, we have selected the lowest minimum wage in
each one-digit industrial sector as our benchmark. The toughness measure used in our analysis is
plotted in Figure 1. In the regression analysis we were not able to use the data from the 1976-
1979 household surveys because no questions were asked to distinguish self-employment from
other forms of employment in 1977-79.