Prescription Drug Benefits and Consumption of Non-medical Goods: Evidence from the Implementation of Medicare Part D

The precautionary savings theory predicts that individuals reduce consumption and increase savings to insure themselves against unexpected future income shortages or expenditures, such as medical expenses. By lowering the risk of high medical expenses, enrollment in health insurance should reduce the strength of the precautionary saving motive. We examine this effect using the introduction of Medicare Prescription Drug Program (Medicare Part D) in 2006. We use data from the Panel Study of Income Dynamics (PSID) to identify the causal impact of the Medicare Part D programs on household consumption of non-medical goods. About a quarter of seniors lacked coverage for perception drugs prior to the implantation of Medicare Part D. Older adults who regularly used prescription drugs, but lacked coverage, faced substantial out-of-pocket costs. Using a difference-in-difference approach, we show that the program significantly increased consumption of older households and decreased their wealth accumulation.

Several theoretical studies and numerical simulations examine how uncertainty associated with future health status affects consumption and saving behavior. While the theoretical implications are clear, only few empirical studies attempt to measure reduction in consumption specifically attributable to the risk of future medical expenditures. Since individuals self-select into health insurance coverage, the implications of precautionary savings theory remain elusive in the empirical studies, mainly due to the fact that individuals who purchase insurance are also precautionary savers. A limited number of studies simultaneously use quasi experiments in the U.S. and investigate the changes in consumption. Our study aims to fill in this gap.

We use data from the 1999 through 2013 waves of the PSID, which includes detailed information on consumption expenditures, health, and insurance status. To construct comparable “treatment” (age≥65 after 2006) and “control” (age<65 after 2006) groups, we restrict the sample to households with either spouses is between 60 and 70 years old. Our dependent variable is household consumption, which is measured as the sum of household annual expenditures on food at home and away from home, housing, utilities, and transportation related expenses, and child care. For durables such as houses and vehicles, we calculated the service flows. Our regressions estimate the difference in change of consumption between Medicare eligible households (65 and older) and younger cohorts over time. We find that Medicare eligible households increased their consumption by 7% compared to younger cohorts after the introduction of the Medicare Part D. Compared to a pre-implementation average annual consumption of $35,000, this change represents about $2,450 increase. We also find a similar reduction in wealth accumulation. Medicare eligible households decreased their net wealth by 3-4% compared to younger cohorts after the introduction of the Medicare Part D.