October 25, 2009
In its October 7th analysis of the health care reform proposal that was ultimately approved by the Senate Finance Committee, CBO estimated that roughly 94 percent of the legal nonelderly population would have health insurance in 2019, compared with about 83 percent today. (That estimate did not reflect an analysis of the legislative language that has been drafted to implement that proposal, which was recently released.) The agency has received many questions about that estimate; let me try to answer the two most frequent questions here.
In its preliminary analysis of the Chairmans mark, which was an earlier version of the proposal presented by Senator Baucus, CBO also estimated that about 94 percent of the legal nonelderly population would have health insurance in 2019. Why didnt the committees adoption of an amendment that reduced the penalties for not having insurance reduce the projected percentage of people with insurance?
In general, proposals that have included a mandate to obtain insurance have also included penalties for not having insurance in order to encourage compliance with that mandate. Although the specific features have varied among proposals, the penalties would generally be collected by the Internal Revenue Service (IRS). An amendment adopted by the Senate Finance Committee during its consideration of the Chairmans mark reduced penalties in several ways: It eliminated penalties on children; reduced the penalty on many adults from $950 to $750 per person; phased in that penalty over a period of several years; made it easier for adults to get an exemption from the penalty; and limited the IRSs ability to enforce the intended penalties.
By itself, that amendment probably reduced the number of people who would have health insurance under the proposal by roughly 1 million to 2 millionbut its effects were offset by other changes to the proposal that made insurance more attractive or easier to get. For example, the committee increased the subsidies available through the new insurance exchanges by reducing the share of income that people with income below 400 percent of the federal poverty level (FPL) would have to pay toward insurance premiums. As another example, the committee voted to continue the Childrens Health Insurance Program (CHIP) as a stand-alone insurance program. CBO expects that offering continuity of coverage for several million children in CHIP would keep some children insured who would not have been covered under the Chairmans mark; under the mark as it was originally offered, which would have eliminated CHIP, CBO anticipated that some of those children would be eligible for subsidized coverage in the exchanges but would not be enrolled in an exchange plan (owing at least in part to the higher premiums and higher out-of-pocket costs that they would typically face in such a plan).
During its deliberations, the committee also adopted a number of other amendments that would have small effects on the number of people with insurance coverage. On net, taking all of those changes into account, CBO estimated that the proposal adopted by the committee would lead to roughly the same number of people with insurance coverage as the Chairmans mark, as it was originally proposed, would havealthough substantial uncertainty surrounds both estimates.
Why does CBO estimate that the percentage of people with health insurance would be so high without larger subsidies for buying insurance or larger penalties for not having insurance?
In the December 2008 volume entitled Key Issues in Analyzing Major Health Insurance Proposals, CBO indicated that one option for obtaining near-universal insurance coverage would be to combine an enforceable mandate for individuals to obtain insurance with subsidies for lower-income households to help them pay their required premiums.
However, judgments about the impact of a mandate to buy health insurance and of penalties for not buying insurance are fraught with uncertainty, because relevant evidence is limited. Some insights can be gleaned from the experience with health insurance mandates in Hawaii and particularly in Massachusettswhich has achieved a very high rate of insurance coverageand from the experience with other sorts of mandates that are designed to achieve various policy goals (such as auto insurance coverage, seat belt usage, and income tax compliance). But neither type of experience yields definitive evidence regarding the issues at handbecause of limitations in geographic scope in the one case and limited relevance to health insurance in the other.
The effectiveness of subsidies in encouraging the purchase of insurance is also an important consideration. The proposal approved by the Senate Finance Committee would offer substantial subsidies to many people who would be uninsured under current lawmost of whom have income below 200 percent of the federal poverty level. Those with the lowest income would be offered coverage under Medicaid, for which they would not have to pay premiums; not everyone eligible for that coverage would take advantage of it, but CBO expects that a large share would do so. In addition, CBO analyzed the subsidies available through the insurance exchanges under the proposal approved by the Finance Committee and found that people with income between 133 percent and 200 percent of the federal poverty level would receive subsidies typically covering between 75 percent and 90 percent of their premiums for a relatively low cost insurance plan; those people would also get federal help with their cost-sharing requirements. Reflecting those inducements, CBO estimates, about two-thirds of enrollees in the exchanges would have income below 200 percent of the FPL. Under the proposal, the premium subsidies would be less extensive for higher-income individuals and families, and cost-sharing subsidies would not be availablebut those individuals and families are also more likely to have insurance through their employer.
The effectiveness of a mandate, penalties, and subsidies in encouraging the purchase of insurance would depend, in part, on the health status of the individuals who enrolledwhich in turn would affect the premiums that insurers charged. With other factors held equal, relatively healthy individuals would be somewhat less likely to purchase health insurance than relatively unhealthy individuals, and the more limited penalties that were approved by the Senate Finance Committee would do less to offset that tendency for adverse selection than would the originally specified penalties. The greater the extent of that adverse selection, the higher the average premiums would beand such premiums, in turn, could further discourage some people from obtaining coverage. However, the Finance Committees proposal also contains two features that would mitigate those effects: First, it would provide up to $20 billion in funding for reinsurance payments, which would cover some of the costs of insurers who attracted particularly sick enrollees and thus would help limit the impact of adverse selection on the premiums that those insurers charged. Second, the income-based caps on what enrollees would have to pay would lessen the risk of an adverse selection spiral because many enrollees would not have to pay more if premiums were higher (though federal costs would tend to be higher as a result).
For all of the proposals involving mandates that CBO has evaluated this year, the agency has considered both the monetary and nonmonetary factors that would affect peoples decisions about whether to purchase health insurance. The monetary factors include the cost of buying insurance (which depends on insurance premiums and any subsidies provided by the government) and the cost of not buying insurance (which depends on statutory penalties and the likelihood of the governments collecting those penalties). There would also be a nonmonetary cost of not buying insurance, which would reflect a mix of factors, including personal values about following rules, social norms or pressures, awareness of the mandate and the penalties associated with it, perceptions of fairness in the penalties, and the hassle of complying with the mandate relative to that of not complying (including the perceived costs of dealing with the IRS). In CBOs judgment, those nonmonetary factors would play an important role in determining whether people obtained coverage, although their impact would depend on the nature of the mandate and the enforcement procedures. As a result, CBO generally expects that a clear mandate to buy health insurance, in conjunction with substantial government subsidies, would significantly increase peoples purchasing of insurance compared with that under current law, even if the monetary penalties for not doing so were relatively small.
To get a quantitative sense of the importance of a mandate, penalties, and subsidies in CBOs estimates, consider the agencys analysis of three recent proposals:
- During the Senate Finance Committees markup, CBO analyzed an amendment that would impose no mandate to purchase insurance and concluded that it would reduce the number of uninsured in 2019 by an estimated 10 million to 15 million. It contained a combination of the subsidies provided under the Chairmans mark, a late-enrollment penalty, an option for insurers to exclude coverage of preexisting conditions in some cases, and an expansion of Medicaid that would have been optional for states (instead of the mandatory expansion proposed in the Chairmans mark).
- In contrast, CBOs preliminary analysis of H.R. 3200 (the health care reform proposal released by the House tri-committee majority group on July 14) concluded that it would reduce the number of uninsured in 2019 by about 37 million; that proposal contained not only an individual mandate, penalties for noncompliance, and a mandatory expansion of Medicaid, but also subsidies for the purchase of insurance through new exchanges that were more extensive (and more expensive) than those adopted by the Senate Finance Committee.
- CBOs preliminary analysis of the proposal adopted by the Finance Committee projected a result in between the two described abovea reduction of about 29 million in the number of uninsured in 2019.
Those figures for the reduction in the number of uninsured under each proposal can be compared with CBOs projection that roughly 54 million nonelderly people would be uninsured in 2019 under current law. Thus, the no-mandate amendment considered by the Finance Committee would lead to about one-fourth of the uninsured obtaining coverage; the proposal adopted by the Finance Committee, to about half; and the July 14 version of H.R. 3200, to about two-thirds. Taking into account the limited available evidence, CBO believes that those differences are reasonable and appropriate. However, there is clearly a substantial range of uncertainty surrounding those estimates, because of the complexity of the changes involved and the range of possible responses by individuals, insurers, employers, and other key participants in the health care system.