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

Researchers and politicians are constantly reassessing the tax preferences given to health expenditures. While the recent focus is on the exclusion of employer sponsored premiums from income and payroll tax, the role and possible interaction of the other tax preferences for medical expenses, namely the itemized medical expense deduction, looms large. This deduction has existed in one form or another for over 30 years yet empirical analysis is scarce. Understanding of the utilization of the deduction will characterize current law and will shed some light on the consequences of several prominent proposals (e.g. Hubbard-Cogan) to relax the requirements for out-of-pocket expenditures to be tax preferred.

Many taxpayers with sizeable medical expenses will not take the medical expense deduction because it requires the taxpayer to itemize and, under current law, is allowable only to the extent expenses exceed 7.5% of taxpayer's adjusted gross income (10% for those on the Alternative Minimum Tax). For example, a taxpayer with qualified medical expenses of $5000 who has an adjusted gross income of $40,000 could itemize $2000 (5000 - .075*40000). With a standard deduction in the range of $5000 this taxpayer may decide against the deduction. Who then takes the deduction and is it sizeable? Does it function as a second source of subsidized insurance? Is it the same people taking the deduction year after year or is the population constantly churning? In this paper, we employ a 1999-2003 random sample panel of 20,000 tax returns to characterize the population taking the itemized medical expense deduction and the persistence in the deduction.

The share of tax payers taking the itemized medical expense deduction grew steadily from 1999 to 2003. Itemizing became more common and, with the rapid growth in health expenditures, the share taking the medical deduction also increased. In 1999, 35% of federal income tax returns were itemized and 15% of those took the medical expense deduction. By 2003 40% of returns were itemized and 20% of those took the medical expense deduction.

The share taking this deduction grew yet the age distribution remained relatively constant. In 1999 and in 2003, those 65 or older accounted for 12% of itemizers but 38% of those taking the deduction. Even after controlling for other factors, age, filing status, and adjusted gross income significantly influence the likelihood of taking the medical expense deduction conditional on itemizing. According to our Probit model estimates, returns with the primary filer age 65 or over are 37% more likely to take this deduction than are 18-24 years old and 28% more likely than 55-64 year olds, all else equal. Moreover, according to our Tobit model results, aged 65 or over returns deduct an average of $4530 more than the 18-24 year olds.

We find substantial persistence in the population of returns taking the medical expense deduction. Moreover, panel regression results suggest that conditional on taking the medical deduction one year, the older tax payers are more likely than younger groups to take the deduction in the following year.