Alot has happened since I last wrote about Mitt Romney’s plan to create a child allowance — worth $350 a month for kids under six and $250 a month for older kids up to age 17 — while ending some other welfare spending and special-interest tax breaks to pay for it.
The Democrats have gotten close to enacting their own child allowance, which would be in effect for only a year to start but would not be paid for at all. And a debate has exploded within the Right as to whether a child allowance is a pro-family reform to support parents — or rather would restore the welfare system we had before 1996, when an “underclass” of single mothers got by on welfare without working, or at least working on the books.
Each side got a boost this week. Scott Winship of the American Enterprise Institute (AEI) released his long-awaited report detailing the evidence that payments such as these could push single parents out of the work force or increase single parenthood. And a new “universal basic income” demonstration project released some preliminary results, finding that no-strings-attached money actually increased work among people who got it.
I’m not going to resolve this debate here, but I would like to briefly review these two documents — and suggest a way forward for right-wingers who are at least open to supporting a child allowance.
As I noted in my writeup of Romney’s proposal, no one seriously denies that free money would reduce work at some point. If you paid all parents $150,000 a year, some would choose to live on that rather than keeping their jobs. The question is what happens at lower levels of support.
There are legitimate reasons to worry about the two proposals on offer. As Winship’s AEI colleague Matt Weidinger has pointed out, the Democrats’ child allowance would be worth more than old-school welfare payments in many states, at least for parents with two kids under age six. Others have noted that the allowances could be combined with food stamps — a program that even Romney doesn’t cut much, and which has weak work requirements for parents, especially those with young kids.
The proposals would also dilute pro-work incentives in the tax code. Both would replace the current child tax credit, which today “phases in” as non-workers start to earn their own money, and Romney would kill the reformed cash-welfare system (which has work requirements) and shrink the earned-income tax credit (which subsidizes work for low-income parents).
Then again, unlike the old welfare system, the child allowance wouldn’t phase out until people earned pretty high incomes, so it wouldn’t create a “welfare trap” in which earning an extra dollar actually costs someone a big share of the new money in lost benefits.
What can we conclude about all this from existing research? As Winship shows, it’s complicated, but there are plenty of studies that should concern us — and the research cited by child-allowance advocates is not as compelling as they would have us believe.
For example, experiments decades ago that paid people a “negative income tax” (NIT) often found significant declines in employment for single mothers, an effect that seemed to grow over time:
[Gary] Burtless cites evidence from another NIT study that found that in the fifth year of the experiment at the sites that lasted that long, single mothers worked 32 percent less than did their counterparts who could not receive NIT benefits — over twice as large an effect as the one for the entire experiment period. That would imply something like a 4 percentage point drop in employment among single mothers, or one-third of the three-decade increase since 1989.
Similarly, there are studies of welfare reform and the introduction of food stamps showing substantial impacts on work when benefits were made more or less available. Winship also expresses concern that bigger benefits could encourage single parenthood, given the big increases in nonmarital births that occurred while the old welfare system was in effect.
Importantly, though, even Winship isn’t too strident about all this. He writes:
There is significant ambiguity in the relevant research literatures — much more than child allowance advocates have conveyed. Different people with different priors can point to this or that study if they are intent on asserting a claim about the impacts of a novel program like child allowances, but strong statements are not justifiable based on a comprehensive assessment of the literatures.
Meanwhile, we have a new report on a demonstration project out of Stockton, Calif. As the report summarizes, the project “gave 125 Stocktonians $500 per month for 24 months. The cash was unconditional, with no strings attached and no work requirements.” The folks who got the money were randomly selected from areas with below-average median incomes, and they were compared with similar folks who didn’t get it.
For our purposes, the important result is this:
In February 2019, 28% of recipients had full-time employment. One year later, 40% of recipients were employed full-time. In contrast, the control group saw only a 5% increase in full-time employment over the same one-year period — 32% of those in the control group were employed full-time in February 2019; one year later, 37% of control group participants were employed full-time.
The report argues, based on interviews with the participants, that money often helped people improve their job prospects. They could scale back part-time or gig work to focus on training, internships, or applying for better positions, for instance. Some participants used the money to cover transportation to work or child care.
Unfortunately, however, the Stockton paper is “preliminary.” It simply doesn’t have a lot of detail, and frankly it’s written in a way that often feels more propagandistic than academic. (Mini-profiles of smiling people who got free money are scattered throughout.) I’d like to see some other analyses of the data before drawing strong conclusions. It’s also possible that people would react differently to a temporary project like this as opposed to a new, guaranteed government benefit — or that welfare dependence is something that takes hold in a community gradually over time as mores shift to accommodate the new incentives.
Again: I’m not going to resolve this here. But I’ve long been something of a lukewarm child-allowance supporter — I’m attracted to replacing our bizarre and complicated welfare system with simple, flat, universal, non-stigmatized benefits, but I also think the 1996 reform was an important success story. So I might be well-positioned to offer a way forward for the conservatives who are at least open to the idea.
We should oppose the Democrats’ proposal, though it very well could become law anyway. You shouldn’t just dump a massive child allowance on top of the current welfare state without paying for it, especially at a time when the government has been sending people huge checks for a year already, through both “stimulus” payments and expanded unemployment benefits. Instead, you should start small, and you should pay for the change by cutting other programs — whether it happens now or after the Democrats’ one-year allowance expires.
The current child tax credit is worth $2,000 per kid, but it’s not available to very poor parents who don’t have income- or payroll-tax liability. Start by making the credit fully refundable and sent out monthly ($167 per kid) to those who prefer it that way. In cutting other spending to pay for it, focus on the programs, such as food stamps and housing subsidies, with the weakest work requirements, rather than gutting the most pro-work parts of the safety net as Romney proposed.
Study how that works. Meanwhile, roll out some longer-term demonstration projects with bigger allowances to see how they work too. If the Stockton report is right, and the money doesn’t recreate the conditions we saw under welfare before 1996, then we can start talking about making the nationwide allowance bigger.
This wouldn’t address every last conservative concern. (It’s probably impossible to: There are numerous conservatives who take a keen interest in child credits, and every one of us seems to have a different idea of how such aid is supposed to work.) The expanded allowance might have bad effects but prove hard to repeal, for example, and it wouldn’t preserve the hard distinction between “targeted tax relief for parents” and “welfare” that many insist on. But it would represent a more gradual shift and give policymakers time to correct course, if needed, before families got used to big monthly checks from the government.
The 1996 welfare reform worked so well in part because states had already experimented with similar reforms to see what worked. We should try that approach again.