Duke 2008 | ASHE | Equity and Efficiency in Health and Healthcare

Most previous studies use current income as a proxy for economic purchasing power. However, it is generally accepted that consumption decisions are driven by lifetime economic resources rather than solely by current income. If the assumption that current income is an adequate proxy for lifetime resources is correct than asset holdings would not matter for health insurance take up decision. In this paper, we have two goals: (1) to present data on the asset holdings and net wealth of individuals by health insurance status. We also present differences in assets controlling for income levels and age as both income and age would affect asset accumulation; and (2) to estimate the effect of assets on insurance take up decisions controlling for income, sociodemographics, and health risks as well as preferences for health insurance.

We use data from the Medical Expenditure Panel Survey for 2002 and 2003. MEPS collects data on financial assets, nonfinancial assets and debt. Data are collected for eleven asset types including residential home, second homes, other real estate, farm or business, transportation vehicles, recreational vehicles, checking accounts, and other savings such as jewelry and debt. We construct three measures of wealth: liquid assets, financial assets, and net worth. 'Liquid assets' include checking accounts, savings accounts and money market funds. 'Financial assets' include liquid assets plus stocks, government and corporate bonds, mutual funds, certificates of deposits, IRA, Keogh and 401K accounts. ?Net worth? includes financial assets plus the net value of residential property, other real estate, business equity, transportation vehicles, and other savings minus debts. We use direct measures of nonhuman wealth (assets) and we proxy for human wealth with current disposable labor income, age, education and job type (proxied by industry and occupation) following previous literature.

We find that asset holdings and net wealth are significantly lower among the uninsured even after controlling for current income. Privately insured adults have much higher levels of assets than the uninsured. Median liquid assets are $2,900 among privately insured adults, while the median among the uninsured is $10. Median financial assets among the privately insured is $11,600 versus $50 among the uninsured. The median net worth of privately insured adults is $60,175 versus $1,000 for the uninsured. We also find that the median assets of privately insured adults always exceed those of uninsured adults in the same age and income group. Next we study the effect of assets and wealth on health insurance take-up decisions. Our results suggest that income is an insufficient proxy for economic resources. People with higher net worth, people who own homes, and those with financial assets are significantly more likely to take up insurance controlling for current income. Furthermore, we find that assets are more important in health insurance purchasing decisions for those without access to the employer-sponsored insurance.