The Essays in Health and Economic Status of Elderly Married Couples

This study investigates similarity in health status between elderly married couples and a structural model is developed to analyze how characteristics of marriage, behavioral choices, and out-of-pocket medical expenditure allocations affect household health capital formations. The first question to answer is whether it is the case that people in good health have healthy spouses and people in poor health have unhealthy ones. I find strong evidence supporting positive spousal health similarity using the Health and Retirement Study (HRS), a data set which contains individuals born in 1931-1941 and their spouses; two waves from 1994 to 1996 are used. In the HRS, there are 5 categorized self-reported health status (excellent, very good, good, fair, and poor) which I use as a main health measure. A man who is in excellent health status in 1994 has a 34% chance to have a wife in excellent health status and has only 0.63% probability to have a wife in poor health. On the other hand, 33% of men in poor health are married to a woman in fair health and only 8% have a chance to have a wife in very good health. For women, 7% among women in poor health has a husband in excellent health and 22% of them are married to men in the same health group. A woman in excellent health has a 33% chance of being married to a man in excellent health and has a 1% chance to have a husband in poor health. When I construct a numerical ordering (1 to 5), respectively from poor to excellent in corresponding self-reported health status, the correlation coefficient is 0.27 in 1994. The concept of correlation in health status comes from Wilson (2002), which documents correlations in health status among married couples in later life and provides a simple theoretical model. While Wilson (2002) uses reduced form regression models, this study constructs a structural model and out-of-pocket medical expenditures are introduced. The second question of this study is: how can spousal resemblance in health status be explained? I explore the extent to which alikeness in health status is a consequence of resemblance in characteristics of spouses such as education and ages when couples got married, and behavioral risk factors like smoking and exercise or the result of living together. Individual characteristics at the time of marriage account for 47% of the correlation after controlling ages. Furthermore health behaviors of couples including smoking and exercise during marriage explain about 9% of the correlation. Household allocations of out-of-pocket medical expenditures and health insurances provided from their spouse affect extra 32% of the resemblance in health. The importance of direct health effects between couples during marriage are captured in the unexplained portion of the correlation. Direct health effects exist, but there is only limited evidence.

Health Behavior

21600