Executive Summary

Today, more than 4-in-10 families in America receive some kind of means-tested government assistance, from Medicaid to food stamps. The expanding reach of the welfare state means that a substantial share of lower-middle-class couples with children receive such aid—and many of these couples receive more generous support if they are unmarried. That’s especially the case if their total income as a cohabiting couple is not considered in determining their eligibility for assistance, as often appears to be the case. To use the term of public policy analysts, these couples face a “marriage penalty,” where it makes more financial sense for them to cohabit rather than marry. Our analysis of American couples whose oldest child is two years or younger indicates that 82 percent of those in the second and third quintiles of family income ($24,000 to $79,000) face this kind of marriage penalty when it comes to Medicaid, cash welfare, or food stamps. By contrast, only 66 percent of their counterparts in the bottom quintile (less than $24,000) face such a penalty.

But do the marriage penalties associated with means-tested policies in the U.S. matter when it comes to family formation?

The evidence from a new IFS/AEI report, Marriage, Penalized, is mixed. Three findings are particularly noteworthy:

1. Our data show that the presence of marriage penalties does not affect marriage patterns among unmarried couples in urban America who have just had a baby, or among couples with children two and under whose income falls close to the lower threshold of the marriage penalty facing such couples, that is, for couples whose joint income is close to a level where they would still qualify for means-tested benefits were they to marry. Most in this latter group are in the lowest two quintiles of family income for families with children two and under (less than $48,000). We also find no evidence that marriage penalties associated with Temporary Assistance to Needy Families (TANF) reduce the odds of marriage among lower-income couples.
    
2. However, we do find that couples whose oldest child is two or younger whose income falls closer to the upper threshold of the marriage penalty—couples where each partner’s individual income is near the cut-off for means-tested benefits—are about two to four percentage points less likely to be married if they face a marriage penalty in Medicaid eligibility or food stamps. Most of these couples are in the second and third quintiles of family income for families with children two and under ($24,000 to $79,000).
    
3. Almost one-third of Americans aged 18 to 60 report that they personally know someone who has not married for fear of losing means-tested benefits.

In sum, our findings suggest that marriage penalties related to means-tested benefits do not discourage marriage among the poorest families in the U.S. But marriage penalties may play a role in discouraging marriage among lower-middle-class families.

Given the possibility that at least two major means-tested programs, food stamps and Medicaid, reduce the odds that American couples with children marry, we think that the federal government should consider the following four strategies to address the marriage penalties embedded within social-welfare policies:

1. In determining eligibility for Medicaid and food stamps, increase the income threshold for married couples with children under five to twice what it is for a single parent with children under five. Such a move would ensure that couples just starting a family do not feel pressured to forgo marriage just to access medical care and food for their families. The cost of this policy change would be limited, since it would only affect families with young children.
    
2. Offer an annual, refundable tax credit to married couples with children under five that would compensate them for any loss in means-tested benefits associated with marrying, up to $1000. This would send a clear signal that the government does not wish to devalue marriage and, for couples, it would help to offset any penalties associated with tying the knot.
    
3. Work with states to run local experiments designed to eliminate the marriage penalty associated with means-tested policies. States could receive waivers to test a range of strategies to eliminate penalties in certain communities, and to communicate to the public that the penalties are no longer in force there. Successful experiments could then be scaled up to the national level in future efforts to reform means-tested policies.
    
4. Encourage states and caseworkers working with lower-income families to treat two-parent families in much the same way as they do single-parent families. For instance, states could ease the distinctive work requirements that many have in place for two-parent families receiving cash welfare. Reforms such as this one would put two-parent and single-parent families on a more equal footing when it comes to public assistance. More generally, policymakers and caseworkers should try to eliminate policies and practices that effectively discriminate in favor of single-parent families.

Steps like these are worth exploring, since this report and a number of other studies conclude that marriage penalties, or other aspects of means-tested assistance, may discourage lower- and moderate-income couples from marrying. In the 1960s and 1970s, social-welfare programs may have been most consequential for family formation among the poorest couples, largely because they were the ones most likely to participate in such programs. Today, however, the design of social-welfare policy may have the most influence on couples in the lower middle class. In light of the important role that stable marriage plays in preserving the American Dream for men, women, and children, federal and state policymakers should seek to reduce marriage penalties that increasingly confront lower-middle-class families. No one wants to see government efforts to support the lower middle class undermine the stability of the very families they are intended to help.