It turns out that, despite the study’s title, the authors’ argument is not that female-named hurricanes are stronger, but that people don’t take hurricanes with female names seriously and consequently don’t prepare properly. However, when I looked at the study. I found several compelling reasons for skepticism.
First, the study includes pre-1979 data, when all hurricanes were given female names. These data are problematic because the average number of deaths per hurricane was 29 during the all-female era and 16 afterward. Perhaps there were more fatalities in earlier years because hurricanes tended to be stronger (the average hurricane category was 2.26 during the all-female era and 1.96 afterward), the infrastructure was weaker, or there was less advance warning. It is more scientifically valid to analyze storms since 1979, when male and female names were assigned randomly.
In addition, the reported conclusion is based on U.S. fatalities from Atlantic hurricanes that made landfall in the United States. If the implicit-sexism theory is correct, it should also be true of fatalities from Pacific hurricanes, from tropical storms, from storms that do not make landfall, and for non-U.S. fatalities.
The study’s analysis can also be questioned for using a noisy index of the femininity of hurricane names, the estimation of a large number of models with various combinations of variables and functional forms, and the use of endogenous monetary damages as an exogenous explanatory variable.
When a strong, surprising conclusion is drawn from restricted data, it can be instructive to see whether the conclusion is robust with respect to the myriad decisions used to restrict the data. I tested the robustness of the reported results using a more inclusive set of data: tropical storms as well as hurricanes, Pacific storms, storms that did not make landfall or made landfall in other countries, and non-U.S. fatalities.
A direct comparison of the frequency with which male-named and female-named storms cause fatalities, cause 1 to 99 fatalities, or cause more than 99 fatalities does not show a consistent pattern, let alone statistically significant differences. A comparison of the average number of fatalities from male-named and female-named storms for all storms, storms with fatalities, storms with 1 to 99 fatalities, and storms with more than 99 fatalities found that male-named storms generally have a higher average number of fatalities, though again none of the differences are statistically significant at the 5% level.
The study’s assertion that female-named storms are deadlier than male-named storms is not robust, evidently because it relied on the questionable statistical analysis of a narrowly defined set of data, including data from years when all hurricanes had female names.