Instead of targeting welfare, charge after-the-fact

The IRS can sort it out for you on tax day

The U.S. welfare system has ballooned into dozens of departments, each with their own bureaucracy. For every dollar of welfare spending, approximately 10% is spent on administration¹.

The problem is, all of these departments are doing essentially the same job. The goal is to:

A. Identify a group of people who need help and determine what kind of aid to provide, given a budget.

B. Deliver that aid as effectively as possible to the group.

‘A’ is a decision that needs to be made by legislators, social workers, and the people they are trying to help. This is the true role of social programs, and it’s a hard problem.

But these departments also spend time and money on ‘B’. While ‘A’ is unique to each department, ‘B’ is essentially the same problem for all programs. There’s no need for these agencies to reinvent the wheel; we already have a single, competent agency that can handle welfare targeting. It’s the Internal Revenue Service.

To avoid making subjective decisions about who deserves aid, departments often use verifiable data about an individual such as household income, number of dependents, marital status, and age. The IRS already has this data, and handles an enormous number of individual payments.

So here’s my suggestion:

1. Welfare departments publish a list of criteria that qualify an individual for aid and report how much it costs per unit of aid.

2. Anyone can walk into the department, provide some identification, and receive aid, no questions asked. The department reports a list of recipients to the IRS.

3. Come tax season, the IRS determines if recipients qualified for aid that year, and charges them the cost of any unqualified aid².

Here’s a few examples of how it might work:

A. Dan isn’t making much income, and so he signs up for the EITC to make ends meet. The IRS finds that he qualified for the aid and there is no change to his tax bill.

B. Bella, a bartender, loses her job and signs up for SNAP benefits. At the end of the year the IRS determines that she qualified for more aid than she used. It informs her of which departments she should contact in the future³.

C. John is a single father who’s fallen on hard times. He claims an unemployment check, signs up for the EITC, gets SNAP benefits, and enrolls in Medicaid. On tax day, the IRS determines that he took slightly too much aid, and raises his tax bill that year (more on this later).

D. Tiffany, a well-paid programmer, gets laid off from her tech job. To get some cash while she’s applying for jobs, she walks into the unemployment department and claims her monthly check. A few months later, she’s found a new job and doesn’t need the money. On tax day, she pays back every cent she received⁴.

This approach has the benefit of being proactive. With a streamlined process for getting support, nobody has to wait to start receiving unemployment, they can get financial assistance the day they lose their job. This fast response has a stabilizing effect and can help people get on their feet faster, which can potentially increase tax revenue.

It can also save the tens of billions of dollars spent on administration costs. The IRS already handles everyone’s taxes with a budget of only $14 billion.

Punting these decisions to the IRS increases the legibility of the welfare system. It gives us a record of who was helped by which department, tells us how much money was spent, and the payback system helps us learn how aid is being misapplied.

Addressing Issues

So this one weird trick might lower the cost of welfare while increasing it’s effectiveness. What are the risks?

My main worry is that people might sign up for more aid than they qualify for and take on a large tax liability. One way to avoid this is to grant departments the ability to certify that an individual is eligible for aid, protecting them from future tax liability. We could also cap the total amount of aid a person can receive; if they want more, they have to go through a special check to ensure they aren’t going to end up with a huge tax bill.

Another objection I can see is that sending this information to the IRS creates privacy concerns, since there is now a centralized database of welfare recipients. However, it’s unclear if the existing welfare system does a better job of protecting peoples privacy than the IRS. This doesn’t feel like a huge issue, but some improvements to the IRS’s security and separation of welfare and tax data should be sufficient to improve the status quo.

Finally, the requirement to supply identification in order to receive aid might exclude people who receive aid under the current regime. Many U.S. citizens lack proper identification, and welfare departments sometimes get around this by using less stringent requirements such as an address or state ID. But the IRS would likely require a social security number in order to link a person’s records. Additionally, some state welfare programs extend their benefits to undocumented immigrants, who would be cut out of any system administered by the IRS. Because of this, I think a few departments will have to opt out of using the IRS to handle administration. That being said, a large majority of welfare departments that already have strict ID requirements, so this will not be a huge issue.

Conclusion

I don’t think any of the potential problems are significant enough that we shouldn’t experiment with this system. We can start by giving the IRS funding to start a new department assessing welfare claims, and allow departments to opt-in to the new system, outsourcing their paperwork while having more money to spend helping people.

The retroactive approach to welfare eligibility has the opportunity to make social programs more streamlined and effective, targeting aid precisely when it’s needed. This will not only save the billions spent on paperwork, but get people on their feet faster. Social workers can turn to the reason why they joined in the first place, helping the people and communities that need them most.

1. The Heritage piece here lists the administrative costs of some major social programs here in note 23.

2. What about in-kind transfers such as food stamps? From the government’s perspective, everything has a dollar value; if a department gives out food and is later paid back the money they spent on food, then the net effect is the same.

3. Should the IRS just give her the money she was eligible for? No, I would argue that involuntarily providing aid infringes on a person’s free speech right. Some people simply don’t want aid out of principle. Giving aid at the end of the year also defeats the purpose of timely, targeted assistance.

4. In cases like these, it may make sense to charge interest on the funds received.