Saving for Future Health Expenditures: Evidence from the Health and Retirement Study

Economic theory of inter-temporal choice predicts that rational individuals will save for future needs. One of the motives for saving is as a precaution against expected, but uncertain future health risks. This precautionary motive has received a fair amount of attention in the empirical literature, but has yet to be definitively demonstrated. In this study, we use data from the Health and Retirement Study (HRS) to examine the effect of predicted future health care expenses on current wealth holdings. We find that, controlling for selection effects and endogeneity, predicted out-of-pocket health care expenses have a significant effect on current period wealth holdings. Previous research examining poor health as a motive for savings has addressed both the effect of future spending uncertainty on savings behavior and the effect of expected future spending levels on savings (e.g., Chou, Liu, & Hammitt, 2003; Kennickell & Lusardi, 2006; Kong, Lee, & Lee, 2008; Hsu 2013; Starr-McCluer, 1995; Yilmazer & Scharff, 2014). The evidence has been mixed. Some have found, consistent with theory, that a reduction in uncertainty through insurance (Chou et al., 2003; Kong et al., 2008) leads to a decline in savings, while others have found an unexpected positive effect of insurance on savings (Starr-McCluer 1995). The literature is also unclear about whether health risks and their corresponding expected future expenses lead to increased savings. Kennickell and Lusardi (2006) find that individuals with expected future health expenses have a higher level of desired precautionary wealth, but find no evidence that they act on that preference. Similarly, Yilmazer and Scharff (2014) find that individuals with health risks do not save significantly more. One factor complicating the analysis in each of these studies is the fact that health and wealth are endogenous. For example, while insurance may reduce the need for savings, those who are wealthier are more likely to both have insurance and spend more on health care. More relevant to this study, while higher expected future out-of-pocket spending may theoretically be a good determinant of wealth, operationally, wealth is an equally valid determinant of out-of-pocket spending. Thus, a positive relationship between the two cannot be definitively attributed to precautionary motives. This study utilizes the panel structure of the HRS to predict an individual-level out-of-pocket spending variable, which is independent of wealth. This variable is then used as a determinant of wealth. Our central hypothesis is that future health risks, manifested through predicted future out-of-pocket health care expenses, leads to increased wealth (savings).