Comments on Paper by Skinner and Zhou

What are the distributional effects of advances in medical care? This important question has been relatively neglected by health economists. I will discuss some of the reasons for this neglect and speculate about possible answers. But first, I will comment on the Skinner-Zhou (S-Z) paper, looking at the questions they address and the answers they provide. Their first question concerns trends in health care expenditures by income class between 1987-2001. They calculate average Medicare expenditures per beneficiary, ages 65 or over, in each income decile based on zip code of the recipient of care. They find that the rate of growth of expenditures was negatively correlated with income. The increase for the lowest income decile ($2,624) was more than double the increase for recipients in the highest decile ($1,214). In evaluating their conclusion, some caveats seem warranted. First, expenditures are not the same as utilization of care. Medicare reimbursement rates for the same services vary across different areas and the differentials may vary over time. Second, and more important, some changes in expenditures, as noted by S-Z, are the result of changes in coverage. For example, Medicare's changes in reimbursement for home care had a very large distributional effect, but the sharp increase in Medicare home care expenditures for lower income beneficiaries tells us nothing about the distributional effects of medical advances. It is also important to note that Medicare covers only a portion of the total medical expenditures of the elderly, and the proportion covered varies with income. A recent by study by Chen and Escarce