Some Comments on the CBO Appendix – “Labor Market Effects of the Affordable Care Act: Updated Estimates”

On February 4, the Congressional Budget Office released a report, The Budget and Economic Outlook: 2014 to 2024. It immediately created a furor, not because of the main report or its general projections but because of an eleven page appendix (Appendix C, “Labor Market Effects of the Affordable Care Act: Updated Estimates”). The CBO was probably a bit taken aback by the reaction in political and policy analysis circles and in the news and opinion media. Evidently feeling it necessary to clarify what the CBO had said, the CBO’s Director, Doug Elmendorf, posted on the CBO’s website on February 10 a piece entitled “Frequently Asked Questions About CBO’s Estimates of the Labor Market Effects of the Affordable Care Act”, in order to clarify that they did not mean that 2.5 million people would “lose their jobs in 2024 because of the ACA” (Affordable Care Act).

The sentence critics of the ACA jumped on stated: “The reduction in CBO’s projections of hours worked [due to the ACA] represents a decline in the number of full time equivalent workers of about 2.0 million in 2017, rising to about 2.5 million in 2024.” Republican politicians immediately denounced the ACA as a job killer, and initially the media played along. When the media finally realized that the Republican characterization of the report was misleading at best, or just plain wrong, they subsequently wrote better articles. Nevertheless, some, most notably Chris Cillizza, a political reporter for the Washington Post, wrote a much ridiculed post for his Washington Post blog, “The Fix” (a reference to political junkies). Cillizza argued that it did not really matter what the facts were for a political analyst like himself; what is important is voters’ perceptions. He wrote: “My job is to assess not the rightness of each argument but to deal in the real world of campaign politics in which perception often (if not always) trumps reality. I deal in the world as voters believe it is, not as I (or anyone else) thinks it should be.”

David Weigel of Slate, responded to Cillizza by writing: “Now, if we're talking or reporting on what Republicans are saying at this moment, true: They're talking about the CBO report. What will we be reporting on a few months from now? What'll the attack be? We don't know.” Ezra Klein, Cillizza’s former colleague at the Post, wrote on his public Facebook page:

“I don't quite understand the model of politics underlying the backlash-to-the-backlash over the CBO report. The theory is that though the GOP's initial spin on the report was wrong it's meta-right because the lies will be used to power effective attack ads in the fall -- and in politics, what's true, and what voters can be tricked into believing is true, are two equally valid categories for inquiry…

“Obamacare is an unpopular law that suffered from a disastrous launch. Does anyone think the GOP would've been unable to find a way to write devastating attack copy about it if that CBO report hadn't come out? Of course not.

“The parties often don't have enough money to air the attacks ads they want to air. They often lack candidates with the credibility to make the attack ads stick. They're often lack the economic conditions that predispose the electorate to listen to them. They're typically chasing voters who lack any interest in watching another nasty political commercial.

“There are plenty of real scarcities in American politics that could really change elections if one party or the other solved them. But "things to say in an attack ad" just isn't one of those scarcities.”

In his blog post, Doug Elmendorf wrote:

“Q: Will 2.5 Million People Lose Their Jobs in 2024 Because of the ACA?

“A: No, we would not describe our estimates in that way...

“Because the longer-term reduction in work is expected to come almost entirely from a decline in the amount of labor that workers choose to supply in response to the changes in their incentives, we do not think it is accurate to say that the reduction stems from people ‘losing’ their jobs.

“Here’s a useful way to think about the choice of wording: When firms do not have enough business and decide to lay people off, the people who are laid off are generally worse off and are therefore unhappy about what is happening. As a result, other people express their sympathy to those people for having “lost their jobs” due to forces beyond their control. In contrast, when the labor market is strong and people decide on their own to retire, to leave work to take care of their families, or to cut back on their hours to pursue other interests, those people presumably think they are better off (or they would not be making the voluntary choices they are making). As a result, other people are generally happy for them and do not describe them as having ‘lost their jobs.’

“Thus, there is a critical difference between, on the one hand, people who leave a job for reasons beyond their control and, on the other hand, people who choose not to work or to work less. The wording that people use to describe those differing circumstances reflects the different reactions of the people involved. In our report, we indicated that ‘the estimated reduction [in employment] stems almost entirely from a net decline in the amount of labor that workers choose to supply,’ so we think the language of ‘losing a job’ does not fit…

“There is a broader question as to whether the society and the economy will be better off as a result of those choices being made available. Even though the individuals making decisions to work less presumably feel that they will be happier as a result of those decisions, total employment, investment, output, and tax revenue will be smaller. (Those effects are included in CBO’s budget and economic projections under current law.) To be sure, the health insurance system in place prior to the ACA generated its own distortions to people’s work decisions, but many of the decisions to work less under the ACA will be made possible by government-funded subsidies, the burden of which will be borne largely by other people. Moreover, people’s decisions about work are also affected by taxes and benefit programs apart from those related to health insurance. Hence, whether voluntary reductions in hours worked owing to the ACA are good or bad for the country as a whole is a matter of judgment.

“A tradeoff of this sort—although not necessarily of the same magnitude—is intrinsic in any effort to significantly increase health insurance coverage or to provide other types of benefits that are aimed at low-income people. As we wrote in the report: ‘Subsidies that help lower-income people purchase an expensive product like health insurance must be relatively large to encourage a significant proportion of eligible people to enroll. If those subsidies are phased out with rising income …, the phaseout effectively … discourage[es] work.’ Again, the best way to address that tradeoff is a matter of judgment.”

As it turns out, the hue and cry over what the CBO said about the ACA and jobs has died down. The media appears to have some remorse in the way they initially reported the story and Republican politicians who have mischaracterized the CBO report have been pilloried by liberal political websites. The story for the most part has disappeared, at least for now.

Still, there has been some more intelligent commentary from some critics of the ACA. The argument they make is that the ACA’s phase out of health care subsidies as lower income people earn more income acts like an implicit marginal tax and discourages work. (For example, see these blog posts from Charles Blahous and Keith Hennessey.) This is in fact the main reason the CBO cites for its estimate of the reduction in labor supply.

If one looks at the structure of the subsidies for health insurance bought on the exchanges, the phaseouts are poorly constructed. In some cases, an extra dollar of income would mean a loss of a subsidy of more than a dollar if one’s income is just below the cutoff level for the subsidy. However, in general, the amount of subsidy that an individual gives up with extra income appears to be in the neighborhood of 10 to 16 percent of the additional income. (One can play around with examples with this handy subsidy calculator at the Kaiser Family Foundation website.) The subsidies end at 400 percent of the Federal Poverty Level (“FPL”). Income is defined to be modified adjusted gross income. (For 2013, the FPL was $11,490 for a single adult and $23,550 for a family of four.)

It is not clear how the CBO estimated how much this would reduce the amount of labor supplied either by workers reducing the number of hours worked or dropping out of the labor market entirely. While significant, these implicit marginal tax rates by themselves are neither huge nor confiscatory. Also, it is worth remembering that one cannot live on health insurance subsidies alone. People who drop out of the labor force because of the reduction in subsidies as their income increases need to have other sources of income or wealth. 

If workers take a full time job that provides health insurance, they lose the subsidy. Whether or not that loss acts as an implicit marginal tax from the point of view of the individual worker depends on the particular facts. One would have to look at the change in the net cost of health insurance premiums and expected out of pocket costs to do a full analysis. To do this in the aggregate would seem to be a daunting task for economic forecasters.

Also, it should be noted that for states that have chosen to expand Medicaid, subsidies for health insurance begins at 138% of the FPL. Below that amount, people get coverage through Medicaid. While losing Medicaid can be viewed as a cost, it is hard to assign an implicit marginal tax rate for earning income that crosses the 138% line, because the private health insurance plans offered on the exchanges are likely to be much better than Medicaid. The number of doctors who accept Medicaid is quite limited and there can be severe frustrations, which differ among the states, in dealing with the Medicaid bureaucracy. (There can be frustrations in dealing with private insurance companies, but my impression is that these are probably not quite as bad as dealing with some state bureaucracies and their legal processes.)

In states that have chosen not to expand Medicaid, there will be people who will neither be eligible for exchange subsidies nor Medicaid. For these people, the law will provide an incentive to work at least enough, if they can, to get to the 138% of FPL in order to benefit from the subsidy. 

Finally, people who are covered by Medicare are not eligible for the subsidies provided for health insurance on the exchanges.

As many have pointed out in the discussion of the CB0 report, means testing a government benefit leads to an increase in implicit marginal taxes until the benefit is totally phased out. This is not new to the ACA. For example, welfare benefits, the earned income tax credit, and food stamps are all subject to means testing. What would be useful for policy analysis would be a comprehensive analysis of means-tested government programs, along with federal, state, and local income taxes, and what this might mean to particular classes of workers. Conservatives are of course right when they say that high marginal tax rates discourage additional work. Devising solutions to this issue is not simple.

However, focusing on the ACA in isolation and criticizing it for this does not do much to advance solutions. Some critics of the ACA would like to go back to the status quo ante, but that is not going to happen. Other critics would like to reduce health care coverage substantially for poor people by giving everybody the same fixed amount (a “refundable” tax credit) to buy private health insurance. That also is not going to happen.

As far as what the ACA does to economic growth, this is unclear. In the first instance, as the CBO admits, there is “substantial uncertainty” concerning its estimates of the effects of the ACA on labor market participation. Even taking the estimates as given, former Federal Reserve Vice Chairman Alan Blinder writes in the Wall Street Journal:

“…The CBO now estimates that the ACA will reduce labor supply by about 1.5% to 2% over a decade. That's small and, arguably, a good thing. Let's count it as antigrowth, though certainly not as a ‘job killer.’ But what about the potential savings in health-care costs from the ACA—which are clearly pro-growth?

“Remember, one of the principal objectives of health-care reform is to slow down cost increases, thereby reducing the burdens on both businesses and the federal budget. And health-care costs have indeed slowed dramatically. The following little tidbit is buried deep in the CBO report, and has gotten little attention: CBO ‘lowered their estimate of average premiums for insurance coverage through exchanges in 2014 by about 15 percent.’ Let me repeat that: 15% lower, just since last May. That's huge.”

As a final point, as long as we have substantial unemployment, that the ACA reduces labor market participation should not be a source of concern. Finally, I would note that many of those arguing that the ACA is fatally flawed because of its effects on labor market participation and economic growth also have been in favor of our current contractionary fiscal policy, which has served to slow the recovery and which monetary policy cannot totally counteract.