Decades-old sloppy analysis fuels mistrust of child allowances

Decades-old sloppy analysis fuels mistrust of child allowances
© Getty Images

For the last two years, the city of Stockton, California has randomly selected individuals to receive $500 a month. There are no strings attached to this cash, including  no requirements that recipients have a job, contrasting many federal social programs that have such requirements. People who received this “basic income” were 7 percentage points more likely to keep or obtain full time employment during the program.

The majority of economics studies that examined cash payments also show that they don’t decrease workforce participation. For example, for 40 years the Alaska Permanent Fund has given residents of Alaska a yearly cash dividend from oil revenues — typically around $2,000 per person. Studies have shown that this payment has no negative effect on participation in the workforce and may even slightly increase part-time work and entrepreneurshipMost estimates of cash payment programs show at most slightly negative work effects — earned income decreases by perhaps $1 for every $10 in benefits.

Despite clear evidence that unconditional cash doesn’t meaningfully impact the workforce, many argue against the new “child allowance” programs that Mitt RomneyWillard (Mitt) Mitt RomneyThe Hill's 12:30 Report - Presented by Facebook - US gymnast wins all-around gold as Simone Biles cheers from the stands The Hill's Morning Report - Presented by Facebook - A huge win for Biden, centrist senators The 17 Republicans who voted to advance the Senate infrastructure bill MORE and Joe BidenJoe BidenCDC chief clarifies vaccine comments: 'There will be no nationwide mandate' Overnight Defense: First group of Afghan evacuees arrives in Virginia | Biden signs Capitol security funding bill, reimbursing Guard | Pentagon raises health protection level weeks after lowering it Biden urges local governments to stave off evictions MORE have proposed - and Congress has temporarily enacted as part of the American Rescue Plan. These allowances would give most families a monthly payment up to $300 for each child. For example, Marco RubioMarco Antonio RubioBreak glass in case of emergency — but not for climate change Democrats join GOP in pressuring Biden over China, virus origins Senators introduce bipartisan bill to expand foreign aid partnerships MORE opposes these programs, claiming that they would make families “reliant on those cash benefits” and “[undermine] the inherent value and dignity of work” — as if $3,600 a year is enough to provide for a child at all, let alone make a parent stop working. The research on similar cash programs demonstrates that his fears are unfounded — and in fact those same fears have led to harmful policies. 

ADVERTISEMENT

Recent efforts to make benefits conditional on work have largely failed to increase employment. Many states have instituted “work requirements” for using Medicaid and have not seen any impact in workforce participation. The harm: many people were denied health insurance as a result of the onerous administrative requirements for proving they had worked. 

Unconditional cash payments are also not a new idea: Richard Nixon almost implemented one in the 1970s, and at that time many cities such as Seattle and Denver experimented with programs similar to Stockton’s. However, instead of everyone getting the cash benefit (the “universal” in “universal basic income”), the benefit only went to people with sufficiently low incomes. If recipients’ incomes increased, the amount of money they received from the program decreased. Very naturally some people decided not to report their income to the experimenters in order to continue to receive their payments — something which would not happen in a universal program. Early evaluations of these programs, based on interviews, showed shockingly large decreases in employment rates — and resulted in the cancellation of Nixon’s unconditional cash plan.

In the 1980s, several researchers revisited these experiments using administrative data sources such as Social Security instead of self-reported income. The revised findings show modest reductions in work, similar to the findings of modern experiments. And that conclusion shouldn’t be surprising: A few thousand dollars a year is not enough to push people out of the workforce and may be just enough to help them join it (for example, by enabling them to buy a car to commute.) The belief that people will not work if they are receiving unconditional cash payments is a myth — one that originates in sloppy data analysis many decades ago. 

Unfortunately, public understanding never caught up with these errors. In fact, the myth became the basis for poorly designed policies going forward. For example, Charles Murray’s 1983 book “Losing Ground” referred to the 1970s findings, without noting that the apparent decreases in work were based on faulty data analysis. “Losing Ground” subsequently became a major inspiration for “welfare reform” in the 1990s, resulting  in the creation of the Temporary Assistance for Needy Families program — a program that exacerbated poverty by making benefits harder to access.

Sadly, these same arguments are fueling debate around the child allowance proposals today — and based on the evidence,  we can be confident that these policies could significantly reduce child poverty without meaningfully impacting participation in the workforce.

ADVERTISEMENT

It might be simpler for policymakers if the cause of poverty was the result of laziness or bad incentives. But the truth is more insidious — policymakers have made poverty worse with poorly designed programs based on sloppy analyses and that leaves people who live with low incomes struggling to do their best to survive. 

We can do better by those families if we follow the evidence and the evidence says that cash works.

Matthew Darling is a vice president at ideas42, a non-profit behavioral design lab where he applies behavioral science to labor and workforce development to improve economic justice. He is an expert on behavioral science and labor programs. His work has been published in Governing, Quartz and other publications.