clock menu more-arrow no yes

Medicaid for Kids could pay for itself

A new study finds people who got Medicaid as children were less likely to die young, and more likely to be employed.

Stacey English with her daughter, Addison, 7, in their Houston home in 2017. Addison lost critical services since Texas implemented $350 million in Medicaid cuts to speech, occupational, and physical therapy.
David J. Phillip/AP

One of my favorite types of social science papers is studies that look at the long-term effects of some government policy that had been enacted decades earlier.

Minneapolis Fed economist Andrew Goodman-Bacon recently dropped a stone-cold classic of the genre, examining how the rollout of Medicaid in the late 1960s affected people who were children at the time.

If you got health insurance through the program as a child, he found, you were less likely to die young; you were likelier to be employed and less likely to have a disability as an adult; and all these benefits actually wound up saving the government money.

Today, Medicaid is one of the biggest and most critical welfare programs, providing free health insurance for nearly a quarter of Americans. But at the time, it was kind of a thrown-together program, an attempt by House Ways and Means Chair Wilbur Mills during negotiations on legislation to create Medicare to appease skeptics who thought the federal government should only subsidize health care for low-income people through the states. (Jill Quadagno’s One Nation, Uninsured has a great history of these negotiations.)

Rep. Wilbur D. Mills of Arkansas in 1971.
Bob Daugherty/AP

Mills designed Medicare Part A to appeal to lefties who wanted a traditional social insurance program funded with payroll taxes, and Medicaid as a state-based component, building on an earlier health coverage law he wrote with Sen. Robert Kerr (D-OK) called the Kerr-Mills Act.

But Kerr-Mills had been a miserable failure that covered less than 1 percent of the elderly population it targeted. So this time, the law added a requirement that states cover everyone currently on their welfare programs.

States, then as now, had wildly different rules for who their welfare programs covered. In his paper, Goodman-Bacon exploits this fact to compare life outcomes for children in states with more generous welfare (which experienced a bigger increase in health coverage due to Medicaid) with outcomes for kids in states with stingier welfare (which experienced a smaller increase).

There’s a methodological problem here: What if states that had more generous welfare programs to start with were already better places for kids to grow up — even without Medicaid? Goodman-Bacon helps get around that problem by using a “differences in differences” design, which helps account for differences between the states going into Medicaid’s rollout.

The findings underscore not just the value of Medicaid as a policy but also the larger importance of resisting short-termism when thinking about policy.

Medicaid for Kids pays for itself

Goodman-Bacon found that getting Medicaid as a young child saves lives: It reduces mortality in the affected population. It also reduces rates of disability and receipt of disability benefits, which makes sense if recipients’ health has improved. It also increases employment for recipients once they grow up.

It helps so much, in fact, that it saves the government money in the long term. The fiscal benefits of the policy to the government, Goodman-Bacon concludes, amounted to about $200 billion (in 2017 dollars) if you calculated the net present value in 1965, compared to a coverage cost of roughly $92 billion.

Goodman-Bacon is hardly the only economist to find government social policies that saved more money than they cost.

Dr. Richard Long treats a young Medicaid recipient in 1995.
Steve Liss/The LIFE Images Collection via Getty Images

A few years ago, I wrote about a paper by Harvard’s Nathaniel Hendren and Ben Sprung-Keyser that found a number of policies that saved the government money, and thus had an infinite social return (if there’s no government cost, the denominator is zero, and that gets you to infinity). An early version of Goodman-Bacon’s paper was actually one of the estimates they included.

Overall, they estimated that programs to help young adults enter and pay for college had an infinite payoff. So did programs to equalize funding between majority-nonwhite and majority-white schools. Now we can add Medicaid for kids (MediKids? Kidcaid? I’ll work on it) to the list.

The obvious takeaway here is that national Medicaid/Medicare for all kids might be a good idea.

Why are governments so shortsighted?

But I think the bigger challenge Goodman-Bacon, Hendren, and Sprung-Keyser’s work raises is: If these programs amount, effectively, to multibillion-dollar bills lying on the sidewalk, why aren’t more governments picking them up?

I have a few theories. The most basic one is limited information. Policymakers in 1965 couldn’t be sure Medicaid would save this much money in the future, so they couldn’t use that as a pay-for.

The second is a lack of budget smoothing — governments can’t always use income they’ll have in the future or had in the past to finance current spending. The government is either unable, as a matter of how the Treasury functions, or unwilling to use money saved in 1987 to pay for health coverage in 1966.

A third is short-term political incentives. Wilbur Mills’s career famously ended after he and an exotic dancer named Fanne Foxe were stopped by police as he was driving drunk in Washington, DC, and Foxe tried to escape by jumping into the Tidal Basin. That happened in 1974, and Mills lost his chair position shortly thereafter. His tenure did not last long enough for him to benefit politically from the long-term effects of Medicaid.

That’s an … unusual example, but Mills’s inability to consider the long-term consequences of his actions is sadly typical of a lot of politicians. Most aren’t willing to wait 50-plus years to see if their programs work for people. Usually they’re on reelection schedules that won’t allow that.

A fourth reason is splits within the government. Goodman-Bacon looks at government expenses as a whole, but policymakers focused on Medicaid may not consider that money they spend there could reduce spending on something like the Social Security Disability Insurance program, since that’s a distinct program with its own trust fund.

A more complex version of that problem would arise in situations where different levels of government are involved: If Medicaid saved local governments money, say, by reducing imprisonment, that probably wouldn’t make the federal government or state governments likelier to fund it. They’re not getting the benefit.

All of those reasons make sense. But we might want to find ways to mitigate them, so governments are likelier to eat these kinds of free lunches.

Sign up for the newsletter The Weeds

Understand how policy impacts people. Delivered Fridays.