Results: Licensing Has Little Impact on Health and Safety
Put simply, this study finds no substantive evidence that licensing or more stringent licensing is necessary to protect public health and safety. Specifically, it finds no substantive difference in health inspection outcomes for businesses in licensed or more stringently licensed states and businesses in unlicensed or less stringently licensed states. This holds true whether looking at the raw numbers or the results of my more sophisticated analysis focused on businesses likely to be most similar based on their proximity to the border. These results suggest that businesses with unlicensed or less onerously licensed workers do not present a greater risk to public health and safety than do their counterparts with licensed or more onerously licensed workers.
By the raw numbers, inspection outcomes were favorable across the board, regardless of licensing conditions. Looking at all businesses, most nail salon health inspections in both Connecticut and New York resulted in zero violations. Similarly, in both Alabama and Mississippi, virtually all barbershop health inspections were passed. This was also the case when looking at businesses within the relatively short distance on either side of the state borders for which businesses and business environments are likely to be most similar. Within that bandwidth, not only were inspection outcomes again favorable across the board, but outcomes were actually better, if only slightly, where licensing was absent or less burdensome. Figure 8 shows the average rate of violations per possible violation for nail salons within the bandwidth around the Connecticut/New York border. In Connecticut, salons near the border passed 98% of standards per inspection on average, while in New York, they passed 95%. Figure 9 shows that 98% of inspections for barbershops within the bandwidth were passed on the Alabama side, compared to 95% on the Mississippi side.
Figure 8. Average Nail Salon Violation Rates
On average, nail salons near the border had low violation rates in both Connecticut and New York, though Connecticut’s rate was slightly lower
Figure 9: Barbershop Inspection Pass Rates
Barbershops near the border had high inspection pass rates in both Alabama and Mississippi, though Alabama’s rate was slightly higher
These results were borne out by those from my more sophisticated regression discontinuity analyses. In fact, the results suggest that the presence of licensing was associated with more violations among manicurists and that greater licensing burdens were associated with more failed inspections among barbers. The differences were statistically significant, though their magnitudes were small. In other words, the differences are likely real—that is, more than just statistical noise—but they are not practically meaningful. Thus, the more sophisticated analyses confirm what the raw numbers show: There was very little difference in inspection outcomes on either side of the borders.
For the comparison of nail salons in Connecticut and New York, results for both the model using the violation z-score and the one using violation rate indicated that licensing was associated with increased violations. This was also true when I reran the analysis adjusting for income, education, and population, as well as when I reran the analysis with businesses closest to the border excluded. These results offer support for the conclusion that licensing increases health and safety violations among manicurists. However, as noted, the effect was very small. Compared to a comparable nail salon in Connecticut, a nail salon in New York might exhibit one more violation over the course of nearly two dozen inspections—and those inspections would typically result in zero violations. Considering the results holistically, a conservative conclusion is that there is no substantive difference in health and safety violations for nail salons with licensed workers compared to ones with unlicensed workers.
For the comparison of barbershops in Alabama and Mississippi, results again indicated that licensing was associated with worse inspection outcomes. Inspections at barbershops in Mississippi were about 8% less likely to result in a pass compared to inspections at barbershops in Alabama. My results were about the same when adjusting for income, education, and population and when excluding businesses closest to the border. However, with the base rates for passing an inspection hovering close to 100%, one could predict the outcome of any given inspection without any other information, such as licensing conditions. Thus, as with the comparison of nail salons in Connecticut and New York, a reasonable and conservative conclusion is that licensing has no meaningful impact on health inspection outcomes. See Appendix B for the full details of my results.