Congressional debate over health care hinges on numbers projected a decade forward to make sure the plan can be paid for. But it's hard enough pinning down today's numbers.
The Census Bureau estimates that the number of uninsured amounts to 45.7 million people. But the agency might be overcounting by millions due to faulty assumptions. Another problem: That 45.7 million figure includes undocumented immigrants, even though they aren't likely to be covered under new laws.
But that hasn't stopped both parties in Congress from using the flawed numbers liberally, as they debate health-care overhaul this summer. That's a reprise of what happened 15 years ago, when the Clinton health plan foundered under differing cost estimates wielded by opponents. But such projections are even more uncertain than today's fuzzy count of the uninsured, depending on tricky assumptions about people's economic choices.
"There is a range of uncertainty in health legislation that probably exceeds that of most other issues before Congress," says Robert D. Reischauer, who headed the Congressional Budget Office when it was analyzing the Clinton health plan.
These sorts of numbers made headlines last week when the CBO dealt a blow to a bill introduced by Sen. Edward Kennedy. The congressional budget watchdog, which relied in part on the 45.7 million uninsured number, said the bill would cost $1 trillion over 10 years, and reduce the number of uninsured Americans by just 17 million, leaving tens of millions of people without coverage.
But the CBO was evaluating just one piece of a larger proposal. For one thing, it omitted a proposed expansion of Medicaid, which would have reduced the number of uninsured further. "They didn't model his whole plan," says Jonathan Gruber, an economist at the Massachusetts Institute of Technology.
The committee's Democrats sought the analysis to respond to Republicans' protests that they were being asked to debate and mark up a bill without cost estimates. "The best guess is there was some sort of crossed wire" between the CBO and the committee, according to Paul Van de Water, senior fellow at the Center on Budget and Policy Priorities. He thinks the committee expected an analysis that would make certain assumptions about the complete bill. These numbers have weighty political consequences -- both the $1 trillion cost projection of the Kennedy plan and a larger estimate for a different Democratic plan sent lawmakers back to the drafting board last week in search of cost cuts.
Clouding future projections of uninsured are tricky methods of counting them today. Even though legislation won't cover many of them, illegal immigrants are especially difficult to enumerate: Few raise their hands to be counted. Prof. Gruber estimates they make up about 13% of the uninsured today, or nearly six million people of that 45 million number.
Of the rest, some people are eligible for health insurance but don't know it and many can afford it but don't want it. About 43% of uninsured nonelderly adults have incomes greater than 2.5 times the poverty level, according to a report released Tuesday by the business-backed Employment Policies Institute.
Meanwhile, Census's state-by-state counts of the uninsured tend to be much higher than state surveys, which have their own flaws. For instance, some don't reach people without landline phones.
The national agency assumes that people who don't answer its health-care questions are much more likely to be uninsured. But that overstates the number of those without coverage, according to Michael Davern, an assistant professor at the University of Minnesota's public health school, who has been studying the discrepancy under a contract with the Census Bureau. To adjust for that overestimate, he recommends that Census adjust its national count of uninsured people downward by 2.5 million. The agency is still considering whether to implement that change.
These misaligned numbers had very real consequences for Massachusetts as it implemented its health-care overhaul three years ago. State surveys indicated there were about 400,000 uninsured people; Census put the figure closer to 650,000.
While Census tends to shoot too high, state surveys can undershoot. "In hindsight it's clear that the higher number was closer to the right number," says Dick Powers, spokesman for the state's Connector Authority, which manages the new health-care choices. The greater-than-expected demand drove costs higher than projections this fiscal year and last.
Given these uncertainties, it's no surprise that projections involve some guesswork. Mr. Reischauer, the former CBO chief who is now president of think tank Urban Institute, says of his former agency's staffers, "They're paid to be straight shooters, and they are." However, he adds that projections "depend on 30 other assumptions. [Some of] those 30 assumptions are going to turn out to be wrong."
In 1993, the agency estimated that health-care costs' share of GDP would rise to 18.9% by 2000. Instead the rate held steady at around 13% for the rest of the decade. "That turned out to be horrendously wrong," Mr. Reischauer says.
The CBO spent five years preparing for its outsize role as a scorekeeper in the health-care debate this time around, working on what it calls a micro-simulation of the health-insurance market. This program models the current state of coverage and the options available to consumers and employers. Analysts then input changes in these conditions, such as government subsidies or penalties for those without insurance, to project the effect of proposed legislation.
The models are based on economic analyses of prior changes in health-care policy and related research. Some of it is well-founded, say researchers who have created their own models and advised the CBO. Other areas, such as the impact of taxing employee health benefits, are more like guesswork, says Elizabeth McGlynn, associate director of RAND Health.
One way costs may be overestimated by the CBO is in judging provisions seeking to end unnecessary and expensive medical procedures. It may be low-balling cost savings, Prof. Gruber says of the CBO.
But others think caution about savings from efficiency gains is warranted. Eugene Steuerle, vice president of the Peter G. Peterson Foundation, an anti-deficit group, points out that every improvement in medical care, such as a cure that could extend life, could lead to more costs.
The biggest uncertainty may be the reaction of employers to a new government plan. The Massachusetts experience isn't explicitly incorporated in the CBO model, but provides insight. Government subsidies pushed some employers to drop their own coverage, but the state's penalties for failing to obtain coverage had the opposite effect. Prof. Gruber notes that, contrary to expectations, employer-provided health insurance rose in Massachusetts.
Such a surprising result "is a somewhat humbling experience in modeling work," Prof. Gruber says. Though one can embrace the flaws: "The nice thing about these models being living entities is, you can update them as you learn more."
Article
