Congressional Budget Office's sunny forecast carries big uncertainties

By Neil Irwin
Washington Post Staff Writer
Friday, March 19, 2010; A01

The latest estimate of what health-care reform would mean for the government's finances was such a hot document Thursday that at times the Congressional Budget Office's Web site couldn't handle the traffic.

But as much as the 25-page "score" of the legislation was treated as holy writ in Washington -- Democrats eagerly flagged its conclusion that the package they aim to pass this weekend would cut the deficit by $138 billion over the coming decade -- the reality is considerably messier.

Budget experts generally have high praise for the work of CBO analysts, the non-ideological technocrats who crunch the numbers to estimate the fiscal impact of legislation. But their work is often more art than science, and although the forecasts that accompany legislation are always filled with uncertainty, this one contains more than most.

One major reason is the sheer complexity of the legislation. If Congress were considering, say, a 20-cent increase in the gasoline tax, the CBO could easily analyze how that would affect gas consumption and do some simple math to calculate how much money it would raise. The same goes for figuring out the cost of legislation that offers a new benefit, such as an expansion of food stamps.

But the proposal on the table contains sweeping changes that would touch almost all corners of the health-care system, and the changes interrelate in hard-to-predict ways. For example, the legislation contains subsidies for those who would not be able to afford health coverage on their own -- but the cost of those subsidies could vary a lot depending on how much other elements of the legislation change the price of health insurance, such as through provisions requiring minimum coverage levels.

Although some data can help budget analysts estimate the fiscal impact of those policies, such as when similar policies were enacted in Massachusetts, the range of possible outcomes is especially wide because of the complexity involved.

"The health-care sector is incredibly complicated, with patients and doctors and insurers and hospitals and so on," said William Gale, a senior fellow at the Brookings Institution. "There are a lot of layers and interacting agents, so it's very difficult to predict the outcome of policy changes that affect everyone's incentives."

If the CBO did its job well, the errors in estimating the costs and revenue from various elements of the legislation will cancel each other out, and the fiscal impact will be roughly according to the forecast. But budget experts identify two scenarios that could cause the results to vary dramatically.

The bill contains numerous provisions meant to reduce the long-run growth in the cost of health care, such as by funding "comparative effectiveness research" to figure out what treatment strategies offer the most bang for the buck. Insurance companies would push their customers to pursue those treatments and thus keep costs down.

But the CBO was cautious, predicting zero savings from that program and others meant to develop ways to make the health system more cost-effective over time.

Some health experts have argued that the agency was too conservative in its approach and that those programs could lead to vast savings in the cost of health care and make the legislation a boon for the federal budget.

But budget experts are more wary, concerned that the programs could just as easily produce few savings, or even cause higher costs.

"I think the CBO is quite right not putting an estimate on these things," said Rudolph G. Penner, a fellow at the Urban Institute and former CBO director. "If you look at things like the spread of medical information, it seems to me you can't even be certain that the research won't tell you that people need more treatments than they get now. If they turn out to offer a lot more savings than I expect, that will be gravy."

But perhaps the biggest risk that could cause the budget impact to diverge from the CBO estimates comes from Congress. The estimates assume that the legislation plays out as written over the coming decade, which would mean reining in the growth of payments to doctors and hospitals and implementing a tax on high-cost health insurance plans.

Those two policies are responsible for bringing in the revenue and cost savings that allow the plan to expand coverage to 32 million more Americans yet, according to the projections, bring down the deficit.

But that falls apart if a future Congress finds the cuts or taxes too painful to handle and overturns them.

There is precedent for that. The alternative minimum tax, for example, is a policy that under law would increasingly affect more middle- and upper-middle-income people and bring the government tens of billions of dollars. But Congress invariably adjusts the tax every year, preventing it from ensnaring those additional American families.

"The risk is that you've put in these things to create higher savings over time, but then when you get there, people say, 'We can't do that,' " said Maya MacGuineas, president of the Committee for a Responsible Federal Budget. "The huge risk is that the constituencies that didn't win their lobbying effort to stop the bill will be there every single day between now and when these things hit, trying to reduce their impact."

Nonetheless, she and other budget experts say that gradual phase-ins, particularly on the tax on expensive insurance plans, might be the best way to minimize the pain and maximize the chance that future Congresses allow them to take effect.

CBO analysts attempt to make forecasts that are symmetrical -- that have a roughly equal chance of being better or worse than expected. And one recent example offers reason for optimism: The Medicare prescription drug benefit was passed in 2003 -- and its costs have come in lower than the CBO forecast.

© 2010 The Washington Post Company