Jared Bernstein, a former chief economist to Vice President Biden, is a senior fellow at the Center on Budget and Policy Priorities and author of the new book 'The Reconnection Agenda: Reuniting Growth and Prosperity.'
Sen. Bernie Sanders (I-Vt.) got in a heated discussion about the estate tax, health care and President Trump's proposed budget with White House budget director Mick Mulvaney on May 25 at the Capitol. (Senate Budget Committee)

It was thoroughly unsurprising to hear President Trump’s budget director inveigh against the Congressional Budget Office for not delivering the score he wanted on the benighted health-care bill he supports (the American Health Care Act — you know, the one the CBO says would strip insurance coverage from 23 million people). Office of Management and Budget Director Mick Mulvaney mused whether the CBO’s time had “come and gone,” for which he was appropriately pilloried from all sides. His response was to double down, claiming that the “CBO has exhibited a history of flawed analysis and inaccurate findings.”

First off, let’s think about the messenger here. For Mulvaney to attack the CBO for the quality of its analysis is like a kid who can’t add challenging Einstein to a math duel. It’s not just that the Trump budget over which he presided had gimmicks; that’s not uncommon. It was the extent of the gimmicks, most notably, the double counting of tax cuts that not only paid for themselves but magically raised another $2 trillion in revenues. Enter “Mick Mulvaney double counting” in Google, and you get about 43,000 results.

Second, you can’t take anything this guy (or his boss) says seriously. Remember how Trump inveighed against the Bureau of Labor Statistics jobs numbers until he got into office and the BLS released a good jobs report? To great laughs from the press corps (arrgh!), White House press secretary Sean Spicer — yet another guy you can’t trust —announced that, at least for that month, the president believed the numbers. Let’s see the distance Mulvaney puts between himself and the CBO when the agency releases a score he likes.

Mulvaney is also the guy who, when confronted with the pain caused in the lives of the poor by cutting spending on programs that help them by $2.5 trillion, responded: “[W]e are no longer going to measure compassion by the number of programs or the number of people on those programs. We’re going to measure compassion and success by the number of people we help get off of those programs and get back in charge of their own lives.” That’s some deeply cynical rhetoric.

It’s not surprising that those who spout such nonsense and so blatantly cook the books would go after the refs, which obviously makes the refs that much more important. Let’s be crystal clear about this: Facts are not the friends of the Trump administration, especially regarding budgets, deficits, and implausible economic assumptions, all the stuff the nonpartisan CBO tracks.

But just because Mulvaney is in way over his head, that doesn’t mean the CBO is perfect. However, his criticism evinces a fundamental lack of understanding of the budget agency’s inherent assignment and their inevitable limitations. The most essential, most valuable aspect of the CBO — the reason it’s not just important, but increasingly important — is its nonpartisanship. If the people in power got to choose their own scorekeeper, Mulvaney-style bookkeeping would be the last word. Instead, once the CBO scores the budget, we’ll be able to see its impact on our fiscal accounts that does not double-count and that uses widely accepted methods and growth assumptions.

Of course, “widely accepted” does not mean correct. CBO faces the same limits as any other forecasting operation, and if there’s one thing economists do badly, it’s forecasting. However, what matters most in CBOs case is not that its guesstimates are any more or less accurate than anyone else’s; it’s that the agency is not leaning one way of the other to get a preferred result. CBO gives us the mainstream results, the ones you’d get if you weren’t trying to stack the deck one way of the other, and, given the proliferation of dealers from the bottom of the deck, such information is essential.

A legitimate pushback to the above is, “Sure, that makes sense regarding budgets, but Mulvaney was whining about the health-care score, and that’s less mainstream because widely accepted assumptions don’t exist for aspects of that policy.” How, for example, does one model which states will apply for waivers from various requirements under the law when there’s no historical record for such actions? (In fact, the CBO handled this smartly, I thought, by looking carefully at states’ health coverage practices before the ACA was upon the land.)

That’s a fair critique, one taken up in a thoughtful piece by Yuval Levin in the National Review. While praising CBO’s professionalism and integrity, Levin argues for at least two reforms; One, the agency should be more transparent, releasing, for example, their models so outside experts can better understand how they get their results and the sensitivities of the embedded parameters. Two, they should produce their results with confidence intervals, as opposed to just point estimates, the way pollsters often do (“Candidate X is ahead 2 points within a 3-point margin of error”).

I like both of these ideas, even though they come with some risks and limits. Levin notes that policy makers who understand the CBO’s models often game the system by crafting proposals they know will score well, based on known modelling assumptions. There’s a risk that broader exposure to the models would lead to more such gaming, but I think that’s a risk worth taking. The benefits of shining a light into black boxes outweigh their costs.

I, too, have long advocated for error bands around estimates, a practice which is fairly common at similar agencies in other countries. But coming up with such ranges is often impossible. Take the state waivers example above. Since there’s no historical record, there’s no probabilistic model I know of that would enable reliable error margins. On the other hand, budget forecasts can be so modeled. In fact, the OMB makes precisely such estimates, but don’t tell Mulvaney!

The pushback on this point is that Congress needs a point estimate, not a range. But again, I believe the benefits of being transparent about our uncertainty outweigh the false certainty of presenting the best guess by itself.

These are all nuances that may be fun to talk about for wonks like me but are beside the larger point, which is this: Those in power don’t care about the facts, the models, the margins of error. Their agenda is to abuse their power on behalf of their wealthiest supporters at great expense to the rest of us.

CBO is one bulwark against such abuse, something Trump and Mulvaney instinctively recognize. That’s the source of their attacks, and it is why those attacks must be aggressively rebuffed.