There has been much academic debate about the effects of the Mariel Boatlift—the exodus of 125,000 refugees from Cuba to Miami, Florida in April 1980—on the wages of local workers. The Mariel Boatlift increased the labor force in Miami by about 8 percent, and increased the low-skilled labor force in Miami by 18 percent. Peri and Yasenov (2019), employing a synthetic control method, found no significant effect on wages and employment of non-Cuban high school dropouts.
This paper replicates the main results of Peri and Yasenov (2019) using the synthetic control method with additional robustness checks. The synthetic control method estimates the effect of the Mariel Boatlift by comparing the evolution of wages for local workers in Miami compared to the evolution of wages for a ‘synthetic’ control group. The synthetic control group is a weighted average of cities that did not experience inflows of migrants, constructed in such a way as to approximate the labor market in Miami prior to the inflow of Cuban refugees. Since there is an ongoing discussion about the sample of local workers used for the analysis, the author also replicates the results for a smaller group of local workers, specifically non-Hispanic males between 25 and 59 years old. The analysis is based on the same data used by Peri and Yasenov (2019).
Main findings:
- The results of Peri and Yasenov (2019) are very robust when additional or different control variables are used to construct the control group. Adding age and the share of males as controls does not change the statistical inference of the results.
- There could be negative wage effects for non-Hispanic males ages between 25 and 59 years compared to the synthetic control group, but these effects vanish after some years and are not very pronounced.
The author concludes that the results are robust for the sample chosen by Peri and Yasenov (2019). However, the results suggest that there could be a negative effect on wages for males between 25 and 59, who were most likely to be affected by the supply shock. Finally, the results suggest substantial differences between wage developments of females and males. The author makes the caveat that the dataset for the analysis is quite small.