Authors: Sven Wilson, Joseph Price, Hannah Hammond, Katharine Wilson
Mentors: Joe Price
Insitution: Brigham Young University
Research suggests that environmental and behavioral risk factors often shared by spouses will result in similar spousal lifespans. This is seen most powerfully in the well-documented tendency for the hazard of mortality to rise immediately after spousal loss, thereby causing a greater concordance in spousal lifespans. This is sometimes referred to as "broken heart syndrome," and a widowhood effect on mortality has been observed in multiple large population studies in many countries. We employ a new dataset from the United States containing over 2.8 million couples drawn from the 1940 census who were born between 1880 and 1920 who are linked to US Census Schedules through 1940 and to the wiki-like genealogical data source Family Search. To account for the presence of assortative mating, we employ a strategy where the lifespan of the spouse's siblings is used as an instrumental variable for spousal lifespan. We find only a small amount of assortative mating on the basis of lifespan, but considerable sorting based on education and income. Our simple IV model finds a statistically significant causal relationship between the death of both husbands and wives on the lifespan of their respective spouses, and those effects are roughly twice as big as the OLS estimates. An additional year of life for the spouse adds .11 years to the husband’s lifespan; and .157 years to the wife. The larger effect on women’s lifespan is in opposition to most of the literature on the widowhood effect.