The EITC began as a conservative idea and now pays out billions in error

Here is a puzzle for the next time someone tells you that conservatives do not care about poor families. The most important federal program for low-wage workers is a check, delivered through the tax code, that pays you only if you work. Its intellectual grandfather was Milton Friedman. It was expanded by Ronald Reagan's tax reform. It is the program conservatives point to when they say that help should reward effort. And it is also one of the leakiest programs in the federal government.

Both of those things are true, and a serious conservative has to hold them in the same hand. The credit has a sound idea and a flawed machine. Let me define the terms, tell the history, show where the money goes wrong, and argue for the repair.

What the credit is

The Earned Income Tax Credit, or EITC, is a refundable federal tax credit for low- and moderate-income workers. Three words in that sentence matter.

Earned income. The credit is paid only on wages or self-employment income. No work, no credit. A family with no earnings receives nothing.

Credit. A credit lowers your tax bill dollar for dollar. If the credit is larger than your tax, the credit does not simply vanish.

Refundable. The excess is paid to you as a check. A worker who owes no income tax at all can receive thousands of dollars. That is why the credit appears to the Treasury as a payment and, for budget scoring, partly as outlays, as we noted in Who actually pays the federal income tax, read straight from the IRS tables.

The credit has three zones. In the phase-in, the credit grows with earnings: each additional dollar earned raises the credit by a fixed percentage. This is the wage-subsidy part, and it is the strongest incentive to work in the whole safety net. In the plateau, the credit is at its maximum. In the phase-out, the credit shrinks as income rises, until it reaches zero. The size of the credit depends on the number of qualifying children, and in the past it was nearly nonexistent for workers without children. The dollar amounts are adjusted each year, so read the IRS tables for the latest figures.

The conservative origins

The idea begins with Milton Friedman. In Capitalism and Freedom, Friedman proposed a negative income tax: below a certain income, instead of paying tax, a person would receive a payment from the government, shrinking as income rose. Friedman's motive was to replace the thicket of welfare programs, each with its own bureaucracy and its own rules, with a single, simple cash payment that kept a gap between working and not working. His point was that the poor should be helped through a system that respects their choices, not by a pile of in-kind programs run by social workers.

A negative income tax with no work requirement is not an EITC. The credit as passed was a different animal, tied to earnings, and many conservatives preferred it for just that reason. The Earned Income Tax Credit was enacted in 1975, sponsored in the Senate by Russell Long of Louisiana, as a temporary measure and made permanent in 1978. It was small. Its purpose, as its backers said, was to offset the payroll tax burden on low-wage workers and to encourage work.

It then grew. The Tax Reform Act of 1986, signed by President Reagan, expanded it as part of a broader reform package. Reagan's supporters praised the credit as an anti-poverty measure that rewarded work and supported families. Further expansions came in 1990, and then in 1993, in the Omnibus Budget Reconciliation Act signed by President Clinton, which made the credit much larger and added a version for workers without children. Over a generation the credit grew from a modest offset into one of the largest cash-assistance programs in the federal government.

It is worth pausing on the politics of that. The credit grew under presidents of both parties, and it grew because it fit a moral intuition shared across the spectrum: help for people who are working and still poor. Few other programs enjoy that sort of consensus. The conservative case was always that this is how you help the poor without trapping them. A dollar that appears only after you work is a dollar that rewards the habit of work.

The problem: the error rate

Now the other half of the story. The government measures how much of what it pays out is paid wrongly, a figure it calls the improper payment rate. An improper payment is any payment that should not have been made or was made in the wrong amount. It includes overpayments and underpayments. It is not the same as fraud. Many improper payments are honest mistakes, and some are the government's own errors.

By this measure, the EITC has been at or near the top of the federal government's list for many years. The federal government publishes the estimates in its annual reports on payment integrity, and in those reports the EITC's improper payment rate has run, year after year, above 20 percent of the credits paid, a figure far higher than most federal programs. In dollar terms, that has meant well over ten billion dollars a year paid in error by recent estimates. I would rather you read the figures yourself than take them from me, because they change from year to year and the agencies revise their methods. You can find the latest estimate on the federal government's payment accuracy website and in the Treasury's annual financial report.

Think about what that number says. For every four or five dollars the Treasury sends out in this program, roughly one is paid incorrectly. If a private business had a one-in-five error rate in its payroll, it would be out of business. For a program that is supposed to be the pride of conservative anti-poverty policy, it is an embarrassment, and a conservative who cares about the credit should care about the error rate more than anyone.

Why the errors happen

Three features of the design explain most of the problem, and none of them is about the character of the people who claim the credit.

One, complexity of eligibility. To claim the credit with children, a taxpayer must establish that each child is a "qualifying child," a status that depends on the child's relationship to the taxpayer, age, residency in the taxpayer's home for more than half the year, and other tests. Families in the real world are complicated. Children move between households, live with grandparents, and spend time with divorced parents. A rule that turns on where a child slept for more than half the nights of the year is hard for the family to apply and hard for the government to check. Many errors occur when a taxpayer in good faith misjudges who qualifies, and some occur when more than one adult claims the same child.

Two, income reporting. The credit depends on earned income, and so it creates an incentive to report self-employment income at a level that maximizes the credit. The self-employed are harder to verify than wage earners whose pay is reported by an employer on a W-2. The errors here range from honest mistakes to deliberate misreporting.

Three, refundability and speed. The refund goes out within weeks of filing, and the IRS cannot check every claim before it pays. Once the money is out the door, recovering it is slow and costly, and often it is never recovered. Congress responded partly in 2015, in the PATH Act, which required the IRS to hold refunds for claimants of the EITC and a related credit until mid-February so that wage reports from employers could be matched first. That was a sound idea and it helped. The error rate is still high.

I hear the objection from the left in advance: the error rate is overstated, because many improper payments are underpayments and honest mistakes, and the cure of stricter verification would keep eligible families from getting money they deserve. There is something to that. Studies of the credit have found that some eligible families fail to claim it, and that audits impose real burdens on families who are in fact eligible. Let me take the objection at full strength and answer it.

The steelman and the answer

The EITC is one of the most effective anti-poverty programs we have. It lifts millions of people, many of them children, above the poverty line. It rewards work. It encourages single mothers to enter the workforce, and research has found increases in employment among the group the credit targets. A high improper payment rate is a measure of how complicated it is to deliver help through the tax code, not a reason to reduce the help. Cutting the credit to fix the errors would hurt working families.

I agree with the first half, and I disagree with the conclusion. The numbered answer:

1. The credit's virtues are real, and they are the reason to fix it, not abandon it. A program that rewards work and has bipartisan roots deserves to be defended against both neglect and waste. Conservatives who attack the credit wholesale give up one of the best arguments for work-based assistance.

1. A high error rate is not a victimless inefficiency. Money paid in error is taxpayer money, and it is also a danger to the families who receive it. A family that receives a credit it is not entitled to may face a bill, penalties and a ban on claiming the credit for years. The people harmed most by a complex rule are the ones who made an honest mistake.

1. Accuracy and generosity are not enemies. A simpler rule, in which eligibility is easy to determine and verify, is a better deal for honest claimants and a harder target for abusers. A family that cannot understand whether it qualifies is not served by the program.

1. A program that rewards work should be one the public trusts. The credit survives because it is popular, and its popularity depends on the belief that it goes to workers who deserve it. A reputation for waste is a threat to the program's survival. Defending the credit includes insisting that it be clean.

What a repair would look like

Here are reforms that keep the work incentive and cut the waste. Some are old, and each has costs.

• Simplify the qualifying-child tests. Use a single definition of the child, shared with the child credit and the dependent rules, so that families answer one question once. A uniform definition would remove a large source of error.
• Verify before paying, where it is cheap. Match employer wage reports against claims before sending refunds, as the 2015 change began to do. Third-party reporting is the strongest control in the tax system, because it does not rely on the taxpayer.
• Move toward third-party data for the self-employed, through payment-platform reporting and better business income checks, without creating a mountain of new paperwork.
• Smooth the phase-out. A steep phase-out creates the high marginal rates we examined in benefit cliffs can make a raise cost a single mother real money, and a rate that discourages the next hour of work defeats the credit's purpose. See also the marginal rate is the one that decides whether you work the extra shift.
• Consider paying the credit in smaller amounts through the year, as some reform proposals do, to give families help when they need it rather than in a single large refund, though that option brings its own trade-offs for compliance and for families who value the lump sum as savings.
• Strengthen the credit for workers without children, a proposal that has drawn support from some conservatives who see it as a better way to reward work than a minimum wage increase. We weigh that wage law in the minimum wage's 1938 birth and the workers it was meant to shut out.

None of these reforms is a cut in the name of austerity. They are repairs to a machine that conservatives built and should want to run well.

The Founding and the work ethic

The values behind the credit are older than the tax code. The Declaration of Independence lists "the pursuit of Happiness" among the unalienable rights, and the tradition of earned success, the idea that a person rises by his own labor, runs through the Founders' writing and through the Judeo-Christian view of work as a calling. A conservative approach to helping the poor asks how to support people in a way that strengthens that tradition instead of replacing it. The credit, in its design, tries to do that. In its execution, it falls short, and that is a failure of administration, not of principle.

There is a constitutional dimension as well. Congress has the power to tax and spend for the general welfare, as Article I, Section 8 provides, and a wage subsidy delivered through the tax code is within that power. But a program that sends out money with a one-in-five error rate is not exercising that power with the care that the public's money deserves. The Constitution gives Congress the power of the purse, and with it the duty of stewardship.

The case in brief

1. The EITC is a refundable credit paid only on earned income, an idea descended from Milton Friedman's negative income tax and enacted in 1975, expanded in 1986 and again in 1990 and 1993.
2. It rewards work and has had support from presidents and members of both parties, which is why conservatives should defend its purpose.
3. Federal payment integrity reports have put the credit's improper payment rate above 20 percent for many years, among the highest of any federal program, though improper payments are not all fraud.
4. Complex eligibility rules for qualifying children, hard-to-verify self-employment income and fast refunds explain most of the errors.
5. The remedy is a simpler definition of the child, more verification before payment, a smoother phase-out and a stronger credit for workers without children, all of which keep the work incentive while cutting waste.

What you can do: look up the latest improper payment figure for the EITC on the federal payment accuracy website, then write your representative to ask what the House or Senate committee plans to do to simplify the credit.