You can think of the OMB as the Red Sox to the CBO’s Yankees. They’re in the same league — assessing the likely outcomes of proposed legislation — but each team has its own style. (Players, however, may float from one team to another, as my Bloomberg View colleague Peter Orszag did from the CBO to the Obama White House.) The OMB is essentially a creature of the White House, which is to say, political. The CBO, on the other hand, stays so scrupulously out of politics that a running joke in Washington notes that its full name is the “nonpartisan Congressional Budget Office,” as journalists seem to automatically add the word.
Mulvaney’s broadside against the CBO claims that it’s not so nonpartisan at all. “If the same person is doing the score of undoing Obamacare who did the scoring of Obamacare in the first place,” he told the Washington Examiner, “my guess is that there is probably some sort of bias in favor of a government mandate.”
Is that a fair accusation? CBO analysts are people, not calculating machines; no matter how scrupulously they think they’re trying to be fair, it’s undoubtedly true that their own ideas and values are going to affect how they assess various pieces of research, the weights they put on different probabilities, and so forth. To that extent, Mulvaney’s point is inarguable: There is probably some sort of bias. Moreover, as uberwonk Yuval Levin has pointed out to me, the architects of Obamacare spent the better part of a year tossing ideas into the CBO model in order to see how they scored, and then taking out the stuff that didn’t score well, producing a program precision-tailored to the CBO’s scoring methods. So it may even be fair to say that CBO scoring is biased in favor of Obamacare.
To that I’d add an observation: There is something very weird going on with the CBO score of the American Health Care Act. Essentially, it suggests that the Republican health care plan offers almost zero net improvement over the pre-2013 status quo. And that finding is frankly unbelievable.
The AHCA, despite its many flaws, represents billions of dollars of subsidies compared to the pre-Obamacare era. I am as skeptical as the next pundit about the effectiveness of government subsidies. You would be hard-pressed to find very many pundits who have as little faith as I do in the curative power of government programs. But when you tell me that the government is going to spend tens of billions of dollars every year on premium tax credits, and almost no new people will get any insurance out of it? Well, I’m going to want some powerful proof of that proposition.
To be clear, I’m quite sure that the CBO is directionally right: Under the AHCA, we would spend less government money on health care, and the U.S. would have more uninsured people. But the CBO is not in the “direction” business; it provides precise point estimates. And I do not believe these estimates.
Having acknowledged all these things, let me also say that I do not think, as Mulvaney implies, that the CBO is engaged in some sort of conspiracy to make the AHCA look bad. I think that the Republicans did a fine job of that all by themselves, by rushing through the CBO process and choosing a program design that was almost bound to work badly with CBO models.
Do CBO health care analysts like government-sponsored insurance programs more than I do? Perhaps. Probably. (Most people do!) But the CBO is currently headed by Keith Hall, an economist of impeccably conservative credentials. I am skeptical that he is allowing the CBO to run wild with liberally biased estimates. More plausibly, this bill is just unscorable.
Mulvaney’s proposed solution — turn legislative analysis into a Hobbesian war of all-against-all, with competing ideological interest groups producing independent scores of legislation — misses the fundamental point of having the CBO, which is not that its estimates are precisely right, but that they are consistent.
Consistency is tremendously valuable in working with data — so valuable that analysts have to think hard about switching to new estimation methods that will break their data series, even if the new estimation methods are more precise. The CBO’s institutional continuity (quite different from think tanks, which can staff up or slim down programs very quickly), means that you don’t score budget bills one way this year, and then next year, when there’s a new guy in charge, completely switch models so that the estimates aren’t comparable.
Perhaps most important of all, the CBO is the only player in the process that is definitely not starting from an outcome — to support a bill, or to oppose it — and then working backward to a model that will support that outcome. That has enormous value, even if the scores sometimes miss the mark. Without them, people trying to discuss legislation would have no common reference point for the discussion; conservatives would wave the Heritage estimate, liberals one from the Center on Budget and Policy Priorities, and both sides would talk past each other. Partisan think tanks serve a vital function in the policy discussion; they are the warriors who make things happen on the battlefield of ideas, and their contention keeps us moving closer to the truth. But opposing sides also need neutral territory where they can negotiate deals peacefully.
And though conservatives may chafe at the idea of letting center-left analysts have the last budgetary word about their legislative proposals, it’s worth noting that before the CBO, the left had enormous room to propose high-cost programs using rosy projections — programs that turned out to be darned hard to repeal when the real numbers came in. (I’m looking at you, Medicare.) One of the first big impacts the CBO had was providing a reality check on a health care bill proposed by Ted Kennedy, which the office estimated would cost three times what its sponsor claimed. Over the office’s long history, it has helped conservatives to make their case more often than it has hindered them.
There is undoubtedly a serious discussion to be had about ways to make the CBO scoring process better. I hope that we will have that discussion. But until someone comes up with an outside institution that can be reasonably credible to both sides, and which can start from the model rather than the desired output, then the CBO is still useful all around, even when you disagree with it.
— Read Repealing Obamacare Would Cost $353 Billion Over Next Decade on ThinkAdvisor.