Mick Mulvaney isn't pleased with the power wielded by his competitor, the Congressional Budget Office. You can chalk that up in part to a natural rivalry between the CBO and Mulvaney's Office of Management and Budget, which is an arm of the White House. Mulvaney also highlights some reasons everyone concerned with good policy should want reform at CBO.

"At some point, you've got to ask yourself, has the day of the CBO come and gone?" Mulvaney told the Washington Examiner's writers and editors in a recent meeting at the White House complex. Mulvaney specifically tore into a CBO staffer who helped give the Republican healthcare bill a bad score.

"We're hearing now that the person in charge of the American Health Care Act methodology is an alum of the 'Hillarycare' program in the 1990s who was brought in by Democrats to score the ACA."

There's no need to ascribe partisanship or even ideological bias to see a problem with the CBO's score of the AHCA. Health policy experts have noted many of the CBO's curious judgments. Mulvaney, for his part, pointed to the odd assumption that the number of Medicaid recipients — low-income people who get healthcare at no cost, and could still get Medicaid under the AHCA — would drop out of Medicaid because the AHCA would repeal the individual mandate.

Does the CBO think Medicaid recipients would quit the program once the mandate was gone? Or is it assuming fewer new people will sign up? (In which case it's a stretch to call them uncovered, because they would automatically be enrolled in Medicaid once they show up at a doctor's office.) We don't know. So when the CBO says 23 million Americans will lose coverage under the AHCA, we don't know whether this based on an undue faith in mandates.

And that's the problem.

The CBO is mostly opaque. It keeps its models basically secret. It doesn't show its work. And while it is forecasting incredibly complex questions with complex interactions for many years into the future, it always arrives at a very specific answer: 23 million uninsured, or a deficit reduction of $119 billion over 10 years.

Any honest accountant would issue such forecasts as a range. On occasion the CBO has done this, but Congress has demanded CBO produce a single number. And so the CBO gives a single number.

These are both matters of transparency, of candor: CBO should show its work and confess its uncertainty. Would this allow for more picking of nits and cherries? Yes. But as Mulvaney's comments showed, the CBO no longer enjoys immunity from criticism. Maybe it's time to drop the mask of omniscience and open the books.

In the age of the Internet when normal people have high-powered computers, there's a good case to be made for transforming the CBO from a black box that uses an opaque model to spit out a single, authoritative answer into something more open-source. "Rather than produce stark hard-number projections behind a heavy veil," Yuval Levin wrote in National Review this week, "the CBO and [the Joint Committee on Taxation] could act more as developers and stewards of an open public model, available online to anyone with sufficient technical prowess to use it."

Depriving the CBO of its privilege and duty to produce the single authoritative number would disrupt the budget process, where currently the "budget reconciliation" process (which neuters the Senate filibuster) relies on the CBO's judgment. So Mulvaney's idea, fostering multiple budget models and considering an array of competing estimates, would need to be part of a bigger institutional reform of Congress, as Levin argues.

Congress, obviously, is in need of reform as an institution. Maybe the prod it needs is to overhaul its budget office by requiring more transparency.