Author's personal copy Marriage, gender and obesity in later life

A large body of literature argues that marriage promotes health and increases longevity. But do these benefits extend to maintaining a healthy body weight, as the economic theory of health investment suggests they should? They do not. Using the Health and Retirement Study (HRS), I find that entry into marriage among both men and women aged 51–70 is associated with weight gain and exit from marriage with weight loss. I evaluate three additional theories with respect to the cross-sectional and longitudinal variation in the data. First, it may be that a broader set of shared risk factors (such as social obligations regarding meals) raises body mass for married couples. However, the shared risk factor model predicts that the intra-couple correlation should increase with respect to marital duration. Instead, it declines. Second, scholars have recently promoted a ‘‘crisis’’ model of marriage in which marital transitions, not marital status, determine differences in body mass. The crisis model is consistent with short-term effects seen for divorce, but not for the persistent weight gains associated with marriage or the persistent weight loss following widowhood. And transition models, in general, cannot explain significant cross-sectional differences across marital states in a population that is no longer experiencing many transitions, nor can it account for the prominent gender differences (in late middle-age, the heaviest group is unmarried women and the lightest are unmarried men). Third, I argue that pressures of the marriage market, in combination with gendered preferences regarding partner BMI, can account for all the longitudinal and cross-sectional patterns found in the data. 2012 Elsevier B.V. All rights reserved. * Tel.: +1 8014229018. E-mail address: [email protected]. 1 This literature is vast. See Wood et al. (2007) for a review of some of the more recent evidence. With respect to marriage and mortality, Rendell et al. (2011) find a consistent survival advantage for married over unmarried men and women, though the effect for women is smaller. A recent meta-analysis of published studies shows a relative risk of mortality of 0.88 (95% CI: [0.85–.91]) for the married compared to the unmarried (Manzoli et al., 2007). 2 Not all scholars view obesity as a central public health problem. Campos et al. (2006), for instance, argue that the obesity epidemic is a ‘‘moral panic’’ more than a public health crisis and that high body mass is a very weak predictor of mortality.