On June 27, 2012 I testified before the House Subcommittees on Human Resources and Select Revenue Measures on the interaction of various welfare and tax credit programs and work incentives. My full testimony, “Marginal Tax Rates, Work, and the Nation’s Real Tax System” can be found here, and the entire hearing can be watched online here. Below is a shorter summary of my remarks.
The nation’s real tax system includes not just the direct statutory rates explicit in such taxes as the income tax and the Social Security tax, but the implicit taxes that derive from phasing out of various benefits in both expenditure and tax programs. These “expenditure taxes,” as I call them, derive largely from a liberal-conservative compromise that emphasizes means testing as a way of both increasing progressivity and saving on direct taxes needed to support various programs. Although low- and moderate-income households are especially affected, you can’t turn around today without spotting these hidden taxes in Pell grants, the AMT, and in dozens, if not hundreds, of programs.
At the Urban Institute we have done quite a bit of work on calculating these rates. One case study we examined was a single parent with two children in Colorado in 2011, to find the maximum benefits for which such a person may be eligible and how they phase out as income increases. (Our net income change calculator (NICC) is now available on the WEB site and can demonstrate tax rates for families in a variety of circumstances in every state.) Rates are low or negative up to about $10,000 to $15,000 of income. Thereafter, they rise quickly. We also looked at the effective tax rate for a household whose income rises from $10,000 to $40,000, and found that income and payroll taxes take away about 30 percent of earnings. The phase out of universally available benefits such as EITC and SNAP or food stamps raises the rate to about 55 percent, and the household getting maximum subsidies, such as welfare and housing, sees a rate of about 82 percent. What used to be called a poverty trap has now moved out to what Linda Giannarelli and I have labeled the “twice poverty trap.” That is, the high rates especially hit households that earn more than poverty level incomes.
Many studies have attempted to show that the effect of these rates on work, and the results are mixed. Work subsidies such as the EITC generally encourage labor force participation but may tend to discourage work at higher income levels, particularly for second jobs in a family or moving to full time work. Design matters greatly. For instance, Medicaid will discourage work among the disabled more than a subsidy system such as the exchange subsidy adopted in health reform; on the other hand, health reform will probably encourage more people to retire early. For the same amount of cost, a program that requires work will indeed lead to more work than one that does not. EITC and welfare reform have done better on the work front than did AFDC.
Other consequences must be examined. Means testing and joint filing have resulted in hundreds of billions of dollars of marriage penalties for low and middle income households. Not marrying is the tax shelter for the poor. Many programs do help those with special needs, although they vary widely in their efficiency and effectiveness. There is some evidence that well-developed programs can improve behaviors such as school attendance and maternal health. At the same time, long-run consequences are often hard to estimate.
Just as a classic liberal-conservative compromise got us to this situation, so might it require a liberal-conservative consensus get us out of it. Among the many approaches to reform are (a) seeking broad-based social welfare reform rather than adopting programs one-by-one with multiple phase-outs, (b) starting to emphasize opportunity and education over adequacy and consumption; (c) putting tax rates directly in the tax code to replace implicit tax rates, (d) making work an even stronger requirement for receipt of various benefits, (e) adopting a maximum marginal tax rate for programs combined, and (f) letting child benefits go with the child and wage subsidies go with low-income workers rather than combining the two.