Joanna Short, Augustana College
This paper examines the effect of a sharp rise in mortality, the 1918 influenza epidemic, on life insurance holdings in the U.S. We might expect that the spike in mortality provided a salient reminder of the benefits of life insurance to young families, and as a result, more households sought life insurance. Insurance executives thought 1919 was a banner year for selling life insurance, and largely credit the epidemic for it (Tarbell, 1919). In contrast, though, Kantor and Fishback (1996) have shown no effect from industry accident risk on life or accident insurance holdings in this period. This paper uses two sources of data: The Bureau of Labor Statistics Cost of Living Surveys of 1918-19 and the Spectator Insurance Year Books from 1916-1923. The timing of the Cost of Living Surveys allows some comparison of household consumption before and after the epidemic—some households were surveyed before, and others during or shortly after the worst of the influenza outbreak. To control for regional differences in insurance demand, I exploit city-level differences in influenza severity in a difference-in-differences framework to compare the increase in spending on insurance in cities particularly hard hit by the epidemic, relative to those that were not. Preliminary results indicate that the pandemic increased holdings of industrial insurance, and to a lesser extent, fraternal insurance. In contrast, the holding of ordinary policies was not affected by the disease environment. This is surprising given the especially large increase in sales of ordinary insurance, but perhaps this reflects the short time window of the survey. State-level panel data from the Spectator Life Insurance Year Books over a longer time period confirms that flu severity did not affect the growth in new life insurance written. I consider a few explanations for the smaller-than-expected results.
No extended abstract or paper available
Presented in Session 240. Hazards, Risks and Disasters