What Accounts for Obamacare Numbers?

Yesterday’s CBO report contained more than just a revision of its projections for Obamacare's effect on employment; it also revised the expectations for the controversial “risk corridor” plan.

Yesterday's Congressional Budget Office report contained more than just a revision of its projections for Obamacare's effect on employment; it also revised the expectations for the controversial "risk corridor" plan, which basically taxes health insurers that make a lot more money than expected in the market and gives money to insurers who have bigger losses.

You might think from that description that this means it's budget-neutral: Insurers pay in, insurers take out. But that isn't the case; neither the pay-ins nor the pay-outs are capped. So, in theory, the government could transfer a bunch of money to insurers if the insurance market ends up with fewer young healthy customers than expected -- or, conversely, that the government could end up making money off the program.

And to the surprise of everyone, that's what the CBO is projecting. Over the provision's three-year life, the agency now estimates that insurers will pay in $8 billion more than they take out. The CBO bases this projection on the experience with a similar part of Medicare Part D.

Is Medicare Part D really comparable? my editor asked me this morning. And the answer is, No, I don't think it is. Don't get me wrong -- I understand, and even approve of, what the CBO did. But I still don't expect it to match the reality of the Patient Protection and Affordable Care Act.

To see why, I want to take a little detour into the way accountants think about valuing factories and equipment. No, wait -- don't go. I promise, this will all make sense in a minute.

When accountants value a company's factories or other similar assets, there are a lot of choices they could make. They could look at want ads and see what similar equipment is worth. They could hire an appraiser to give an expert estimate. They could develop some complex formula that tries to account for inflation and wear and tear. However, they tend to use the simplest method: They value the asset at whatever the company paid to acquire it.

You can instantly see the biggest issue with this. If you bought a long-lived asset decades ago, then what you paid is probably less -- maybe even a lot less -- than what you could sell it for. So the book value of your assets will tend to understate the actual value of your company.

Why, then, do accountants use a method that is in many cases almost guaranteed to give an inaccurate value?

Because it's objective. If you let management guesstimate the value of their assets, there's simply too big a risk that the guesstimate will represent what management would most like to think, rather than what is likely to be actually true. (This is still a problem with accounting, obviously. But it could be a much, much bigger one.)

This approach is what accountants call "conservative" -- which, in accounting, is not an ideological market, but rather an attempt to cut down on the freedom of choice that the people who do the books have to inject their own opinion. And it's the best approach, even though the results may differ significantly from the company's "true" value.

The CBO tries to be conservative in the same sense. Any economic forecast has thousands of embedded assumptions and choices of methodology that can significantly affect its results. In order to avoid becoming a partisan football, the CBO frequently chooses to tie its own hands -- to limit its choices in order to avoid, as much as possible, letting personal opinions and beliefs influence its work.

That's why it scores bills on the assumption that all of a bill's components will be enacted as written, even if some of them seem extremely unlikely to be left intact. (Several important components of Obamacare have already been rolled back because, as people noted at the time, they were bad ideas and, even worse, unpopular ones.) And that's why, I think, it chose to let the history of Medicare Part D guide its work: It's the only historically comparable program it has.

So I think that the CBO made the right choice, for institutional reasons. Unfortunately, there's also reason to worry that in this case, the right choice might differ significantly from the actual outcome.

While the Part D program and Obamacare's risk corridors are superficially very similar, they actually have a very different set of problems. In Part D, there are basically two risks a company faces: First, that it will, by chance, get a disproportionate share of the very sick folks who use a lot of prescriptions, and second, that it will get a normal share of very sick folks but that people will use more, or more expensive, drugs than it projected.

In recent years, Americans have consistently spent less on prescription drugs than expected. There are fewer new must-have blockbuster drugs; moreover, companies, doctors and patients have all been aggressively moving toward lower-cost generics whenever possible. So it's not surprising that companies have overestimated what patients were going to cost them -- and, hence, that the government has profited from the risk-adjustment program.

With Obamacare, by contrast, the risk isn't that an individual company is going to misestimate its costs. Or that is one risk -- but the major risk is in the market, not in individual firms. It's the risk that fewer people will buy insurance than expected, and that the people who do sign up will be the older and sicker ones. If that happens, then most of the companies in the market will lose money, and the government will have to pay a considerable portion of all those losses. That isn't a risk with Medicare Part D, because basically every senior who is eligible signs up for it.

Early enrollments in Obamacare, on the other hand, were lower than anticipated. They also showed a marked skew toward older enrollees. That may change, because young people might wait until the last minute to sign up, as opposed to older people who have a lot of medical needs. Humana Inc. has already seen a slight improvement in the demographic mix of the people who are buying its exchange policies, and it says it's confident that it has priced the policies correctly.

But we're certainly not out of the woods yet, and at this point, the CBO is projecting that the exchanges will have 1 million fewer customers this year than originally anticipated. Just as it's reasonable to think that younger people are likely to be the ones who waited to buy insurance, it's also reasonable to think that they are the ones most likely to give up entirely. Not every insurer sounds as bullish as Humana; WellPoint Inc. apparently expects to recover 10 percent of its premiums from the reinsurance funds.

So while I think the CBO used the right method, reporters and others should be cautious about concluding that it therefore got the right answer: that the government is likely to net $8 billion from the program over the next three years. That's certainly within the realm of possibility. But a scenario in which the government loses money seems at least as likely.

(Megan McArdle writes about economics, business and public policy for Bloomberg View. Followher on Twitter at @asymmetricinfo.)

This column does not necessarily reflect the opinion of Bloomberg View's editorial board or Bloomberg LP, its owners and investors.