Means-testing and the real marriage penalties (wonkish)

For decades (at least as long as I’ve been alive), people have used the expressions “marriage bonus” and “marriage penalty” to describe what happens to the income tax liability of a couple before and after they formally wed. But focusing attention on changes in income tax liability has resulted in a totally inadequate amount of attention being paid to the actual marriage penalties faced by low-income people considering marriage.

The classic marriage “bonus”

Consider two people, one of whom has an IRS form 1040 line 22 total income of $100,000 and one of whom has a line 22 total income of $0. With no other adjustments, in 2018, their combined income tax liability would be $15,410: the high-earner would pay income taxes on an adjusted gross income of $88,000 (after her $12,000 standard deduction), and the low-earner would pay $0 on his $0 income.

Now consider if the same two people wed before the end of the calendar year, and file their taxes jointly. The same $100,000 in total income would be reduced by the new $24,000 standard deduction, and they would owe $8,739 on the remaining $76,000. The difference between the $15,410 owed when unmarried and $8,739 owed after marriage is a marriage “bonus” of $6,671.

Now consider two people, each of whom has a form 1040 line 22 total income of $50,000 in 2018. Each of them would pay $4,370 on $38,000 in taxable income. If they wed before the end of the year, they’ll end up paying $8,739 on their combined $76,000 in taxable income.

In this stylized example, when two people with taxable income wed, their total income tax bill remains unchanged, while when a person with taxable income weds a person without taxable income, the government pays them $6,671 in reduced income tax liability.

Means-testing creates enormous marriage penalties

I am not, in general, concerned with the effects of marriage on income tax liability. I think the “marriage bonus” described above is pretty absurd, and we’d be much better off eliminating the standard deduction entirely and replacing it with a universal basic income, but that’s beside the point of today’s post.

The real marriage penalty doesn’t appear in the income tax tables, it appears in the multiplicity of ways marriage can affect people’s eligibility for means-tested benefits. Let’s look at a few stylized examples.

• Health Insurance. In a Medicaid-expansion state, a person earning $16,753 is entitled to premium-free Medicaid coverage and a person earning $48,560 is entitled to premium tax credits that limit the cost of coverage on the Affordable Care Act exchanges to 9.6% of their income ($4,662). If the two people wed, their income will still be slightly under 400% of the federal poverty line, but 9.6% of that amount will now be $6,270, a marriage penalty of $1,608. Meanwhile, they receive a marriage “bonus” of just $95, leaving them $1,513 worse off, assuming they require no medical treatment whatsoever. If they do require medical treatment, the marriage penalty balloons further since, unlike Medicaid, Affordable Cart Act plans also include substantial deductibles and co-payments for treatment.
• Earned Income Credit. Often described as one of the most effective programs embedded in the tax code, beloved by Democrats (anti-poverty) and Republicans (pro-work) alike, the Earned Income Credit also features an enormous marriage penalty: a single filer with $6,700 in earned income is eligible for the maximum refundable earned income credit of $510. Two such workers, each with $6,700 in earned income, when filing jointly are eligible for a refundable earned income credit of $510, resulting in a marriage penalty of $510.
• Retirement Savings Contribution Credit. The RSCC is based on adjusted gross income, so it can get extremely complicated depending on your ability to game your AGI, but let’s again use a stylized example to illustrate how marriage penalties work in practice. In 2018, one single filer with an AGI of $19,000 ($31,000 in total income) and one single filer with an AGI of $19,001 ($31,001 in total income) would be able to claim a RSCC of $1,000 and $400, respectively, for a total credit of $1,400. If wed, their credit would be just $800, resulting in a $600 marriage penalty.

Means-testing is an assault on the family

All forms of means-testing involve a certain incoherence, even if that incoherence is at times fashionable. The earned income credit, for example, is widely praised by economists for phasing in quickly and phasing out slowly, since economists believe that marginal incentives drive behavior. But if marginal incentives drive behavior, why should the earned income credit be phased out at all? And why should it be halved on the occasion of your wedding day (let alone retroactively to the first day of the year you wed)?

An incoherent view towards the family and towards poverty is at the very heart of the conservative movement in the United States:

• if marriage is the best way out of poverty, then you might want to eliminate the financial penalties for marriage I described above, in order to encourage more people to wed;
• but if anti-poverty programs are keeping people from getting married, then you might want to eliminate benefits for the poor in order to cast them deeper into poverty and encourage them to marry solely to escape it.

Thankfully, I’m not a conservative, so I don’t have to try to resolve the contradictions inherent in this worldview, but merely point them out.

Conclusion

I do not believe that the American family is “in crisis,” but the more seriously you take the institution of the family in American life, the more seriously you should take the very real attacks on family formation built into our welfare state and tax code. A $510 penalty here, $610 there, and even $1,513 over there won’t on their own destroy the institution of the family. But if you take that institution seriously, you ought to fight to eliminate those penalties wherever they arise.