The Academy committee is not recommending that a reinsurance mechanism be required or that one is even desirable. High risk pools have both advantages and disadvantages. They have the advantage of being able to insulate individual insurers against the risk of adverse selection. They have the disadvantage of tending to reward insurers for their ability to select groups to put into the pool rather than rewarding insurers for controlling claims costs and seeking to better manage care. The committee's only point is that the lack of a risk pool might possibly lead to higher average costs and that the implications of having or not having a risk pool need to be carefully studied.

Applying rating bands will mean that the lowest-cost groups among the currently insured will have to have their premiums raised to subsidize the highest-cost groups among the currently insured. As a result, some low-cost groups will drop coverage altogether. This will cause the average cost of small group insurance to rise.

Rating bands, when combined with guaranteed access, will further encourage adverse selection among groups that become newly insured. The least healthy, highest-cost groups that were not previously insured will be the first to elect coverage under the new standards. This, too, will drive up the average cost for small groups. And here again, the result is likely to be selective disenrollment of the lowest-cost groups who will either move to underwritten individual health insurance products or drop health insurance coverage altogether. In addition, as the average cost rises, more small groups may find self-insurance an attractive alternative despite the 25% surtax imposed on the health care expenditures of self-insured small groups.

Although it seems clear that average costs for insured small groups will rise, there are a number of factors that may mitigate against a substantial across-the-board increase. First, not all currently insured small groups are highly underwritten. This is true for most small groups covered by HMOs and, in some states, for the small groups that participate in Blue Cross/Blue Shield plans. In fact, the adverse impacts of H.R. 3626 can be expected to vary widely among geographic areas. In California, where HMOs and Blue Cross and Blue Shield dominate, only a very small percentage of small groups might initially face substantial rate increases or substantial rate decreases. In some states that have implemented their own health reform measures, the impact of H.R. 3626 might also be relatively small.

Exhibit 3 shows how small the impact of the H.R. 3626 rating bans will be in some settings. The exhibit is based on data for groups of under 30 that are covered by not-for-profit companies in Minnesota (HMOs and the Blue Cross/Blue Shield plan). In the exhibit small groups are divided into cost groups in deciles. The top horizontal bar in the exhibit is for the 10 percent of groups with the lowest current premium rates. Similarly, the bottom bar is for the 10 percent of groups with the highest premium rates. For each of the ten cost groups, the percentage change in cost is shown for rating bands that permit a 30% variation in premium, a 20% variation, and no variation (i.e., community rating). As shown in the exhibit, with a 30% rate band, only the premiums of the 10 percent of groups with the lowest and highest premiums would be substantially affected, and premiums would increase by less than 15% for the lowest cost groups. Exhibit 3 also shows that the number of groups whose rates would change is fairly sensitive to changes in the rating band. When rate variations are restricted to 20%, 40 percent of groups are affected and the premium increases for the lowest cost groups are more than double those required when 30% variation is permitted.

Another factor that could mitigate against substantial across-the- board increases is the ability of highly underwritten small groups to move to managed care settings, such as HMOs and PPOs. In the absence of aggressive underwriting of small groups, these settings may offer lower-cost alternatives than traditional indemnity plans whose morbidity might begin to look much more like that in settings that are currently not heavily underwritten.

Although there may be mitigating factors, nonetheless, some, if not many, of the cross-subsidies envisioned in the bill will not materialize, and average cost will increase for the remaining insureds. Thus, although H.R. 3626 will give some groups access that was previously denied, the induced higher average cost of small group insurance may well increase the proportion of

5

workers in small firms who are uninsured, and this proportion could continue to increase over time if the risk pool of insureds deteriorates because of because of selective disenrollment and high cost new entrants.

The bill addresses the magnitude and predictability of rate increases by introducing rating bands and limiting annual premium increases for any group to the increase in the insurer's base rate for all groups plus 5 percentage points.

Over a period of three or four years, the rating bands and related restrictions should add to the predictability of rate increases. These provisions will eliminate first-year discounts for newly underwritten small groups and will prevent insurers from using tier rating and other durational rating schemes. In the short-run, however, many small groups with very low premiums may experience large rate increases in order to subsidize the groups that currently have the highest

rates.

The hardest hit groups will be the younger, healthy ones that have recently been underwritten or reunderwriten. These groups will pay increased premiums to subsidize the higher cost groups, to pay for changes in the preexisting conditions limitations, and to support the inflationary increase in medical costs and utilization. Finally, the groups paying the lowest premiums may experience substantial premium increases because of the imposition of the new mandated benefit package.

At the same time that some of the lowest-cost groups would be experiencing sharp short-run increases in premiums, imposition of the rating bands would mean that the groups with the highest premiums currently would experience reductions in their rates which in the short-run could be as great as 20-30% or, in extreme cases, even more. Of course, in some cases substantial reductions would not occur because of the mandated enhanced benefits package and the cost of guaranteed issue and other provisions. Moreover, even large reductions that are realized may be short-lived. They could be reversed as the market adjusts to the effects of adverse selection by high-cost groups entering coverage and selective disenrollment of currently insured low-cost groups.

Thus, although H.R. 3626 should lead to more predictable rates, during the first two or three years of implementation, erratic changes in rates both up and down may well confuse some members of the small employer community and increase uncertainty among both employers and employees. Moreover, employees and employers who cannot afford to keep their health insurance because of these changes are likely to be highly dissatisfied with the new federal standards.

OPTIONS FOR IMPROVING H.R. 3626

If Congress chooses to adopt the approach set forth in H.R. 3626 several changes could be made that could mitigate the large impact the bill may have on reducing coverage and the short-run disruption the bill would cause for both consumers and insurers.

The most important thing that should be done is to provide for a reasonable transition time from the present system of carefully underwritten small group health insurance to a system with only minimal underwriting. The various provisions of the bill could be phased-in so that those currently insured would not experience dramatic rate changes in a single year. For example, during the first year insurers could be required to implement the new rules for pre-existing conditions. During the second to fourth years insurers might be required to eliminate durational and tier rating practices. During the fifth and subsequent years, insurers might be required to move their rate structure within the rating bands specified in the bill.

There are several other advantages to a phased-in approach in addition to insulating large numbers of consumers from dramatic premium rate changes in a single year. Insurers would be better able to anticipate how their mix of insureds was changing and price this risk appropriately. In addition, spreading the large rate changes that some groups will have to experience over a

6

number of years may help employers adjust to the higher costs and reduce the number of small groups that withdraw from coverage.

Alternatively, some provisions might be dropped from the bill initially until their implementation could be more carefully planned and the cost implications more carefully studied. For example, an initial proposal might only place new restrictions on exclusions for preexisting conditions and eliminate durational rating schemes. These changes would not be highly disruptive to either the consumer or insurer side of the market.

There are many other options for improving H.R. 3626 including allowing more flexibility in the minimum benefit package. The Academy's Committee on Health would be happy to discuss the full range of options with members of the subcommittee or their staff.

KEY PROVISIONS OF H.R. 2121

H.R. 2121, introduced by Representative Stark, would set minimum federal standards for insurance sold to all employer groups for employees who work 30 or more hours per week. (No minimum or maximum group size limits are noted in the bill.) The bill would:

Require that health insurers issue policies to any employer willing to pay the premium, maintain year-round open enrollment for all eligible employees, and guarantee the renewability of the policy.

Require that all policies be issued on a community-rated basis with no recognition of a group's demographic characteristics. Individuals leaving the group could opt to convert to individual coverage at a rate of no more than 150% of the insurer's applicable community rate.

Mandate a minimum standard benefit package that includes all coverages included under
Medicare Parts A and B plus a number of additional coverages.

Limit exclusions for pre-existing conditions to 6 months.

Establish a federal reinsurance trust fund that would reimburse qualified plans for payments of core benefits for an individual after the first $25,000 of covered medical expenses. The reinsurance fund would be financed through a per participant excise tax. The tax rate would be set each year to cover the expected expenditures for that year, and the trust fund arrangement would apply to all qualified plans, including self-insured ones.

IMPACT OF H.R. 2121 ON CURRENT INSURANCE PRACTICES

Mandated Minimum Benefits Package: The package in this bill closely parallels the package included in H.R. 3626. However, the deductibles and out-of-pocket amounts, which may not exceed $500 and $2500, respectively, are only applied to individuals. No deductibles may apply to preventive services for children, nor to pregnancy related services. The minimum covered services provisions also parallel those in H.R. 3626. Since the differences in the bills are minor, the same comments apply that apply to H.R. 3626.

The minimum benefits provisions apply to all health insurance products offered to all employers. Thus, the intent of the core benefit requirements appears to be that a specified minimum level of benefit should be provided if an employer elects to provide health benefits at all. Assuming that is true, extending the core benefit provisions to self-insured plans would seem appropriate and would avoid peculiar cross subsidies between self-insured and insured employer groups that could occur through the reinsurance mechanism.

Guaranteed Access: H.R. 2121 includes a number of provisions to ensure guaranteed access to health insurance. These include year- round open enrollment, guaranteed eligibility of all

7

members within an employer group including dependents, guaranteed issue except for inability to pay premiums, and guaranteed renewal except for non-payment of premiums, fraud, misrepresentation, or non-compliance with plan provisions. Related to these are limitations on the use of preexisting condition limitations.

These changes to ensure access will primarily affect the small group market where they will cause health insurance premium costs to increase. This market is usually defined to include groups of up to 25 or 30 employees. For larger employer groups access to insurance is not a significant problem. Since the guaranteed access provisions of H.R. 2121 are similar to those of H.R. 3626, the same comments apply.

Community Rating: Premium rates for coverage under health insurance contracts will be based on the "average per capita cost" of providing coverage to all individuals covered by applicable contracts issued by the insurer. This means that each person will pay exactly the same premium rate for coverage. This rating approach is commonly referred to as community rating.

Introduction of a community rating requirement will have an enormous effect on the insurance market. Nearly all groups will be affected by the introduction of community rating. Groups who currently pay the lowest premium rates will have large rate increases. Groups with the highest current premium rates will have rate decreases. Some of the groups who have significant rate increases will elect to drop coverage rather than pay the increased cost. As a result, the "subsidy" from these groups (i.e., the difference between the community rate and the group's expected costs) will be unavailable, and the average community rate will increase to cover the shortfall.

With an immediate shift to community rating and a guaranteed issue environment, employers, through their insurance brokers, would be able to compare readily the cost of purchasing coverage from many companies. A situation where this leads to financial losses for all competitors is easy to imagine. For example, assume two insurance companies. Company A anticipates that a premium of $120 is needed for its community, and Company B develops a rate of $100. Both companies have based the rates on an expected cost for their enrollment. Because comparison shopping could be readily done, Company A is likely to lose many groups to Company B. Company B, on the other hand, is likely to suffer large short-run losses, because its $100 premium is inadequate to support the costs associated with Company A's groups. These types of problems are especially likely during the transition period when insurers will be dealing with great uncertainty about their expected costs.

The premium rate calculation described in H.R. 2121 is silent on issues of geographic variation. This is problematic, since some health insurers operate nationally, while others are regional or statewide. HMOs, which provide coverage to a growing membership, generally have very concentrated service areas. Geographic differences in costs are substantial, and area rating factors may differ by more than 100 percent. The large variations in costs among geographic areas and the variation in geographic market areas of insurers underscore the need to explicitly address treatment of geography.

Requiring a single community rate for all employer group sizes is also a concern because of variations in insurers operations by group size. These variations affect the cost of operations in the different markets. Examples of the variations include:

The result of requiring a community rating approach for all insured groups may be that larger groups subsidize the costs of smaller groups. To avoid this situation, increasingly smaller employers are likely to elect to self insure, which will raise the average rate needed to cover the expenses for the remaining groups.

8

Health Reinsurance Trust Fund: The Health Reinsurance Trust Fund would effectively provide stoploss insurance to health insurance plans, self-insured employer plans and HMOs for health expenses in excess of $25,000 for members of employer groups. Claims in excess of $25,000 per person represent 10-15% of total health claims. Since the program is applied to all plans, this proposal would have the federal government provide catastrophic insurance protection for employer-based health benefit programs. If this is the intent of the trust fund, is it equitable to provide these benefits only to persons who are covered by employer plans? Perhaps the benefit should be provided to any individual, whether insured in some way or not.

The proposed $25,000 stop-loss limit raises a number of mechanical issues. A common method of determining when the $25,000 threshold is met would be needed. This is a problem because insurers and HMOs pay providers under a number of different arrangements. These include, but are not limited to, billed charges, negotiated discounts, per diem hospital rates, per admission hospital rates, fee schedules and capitations. Using a predetermined rate schedule as a basis for determining whether the limit was met would be a possible solution, but would be an additional administrative expense for insurers.

Another issue is the accumulation of claims over time. Is the $25,000 intended to be an annual threshold or is it cumulative? If the amount is cumulative, must all claims be incurred while an individual is insured by a particular health insurer or are the accumulated claims portable? The $25,000 reinsurance threshold raises policy as well as technical issues. Treatment of many medical conditions that are relatively common today will produce expenses in excess of the $25,000 threshold. Will the federal government become involved in addressing ways to control these costs? An example might be a program dedicated to contracting and/or controlling costs associated with transplants. This, in turn, leads to other questions on the possible role of the government in rationing access to high cost services. The threshold amount needs to be carefully considered both as to its impact on the insurance market and its implications for public policy. Health Insurance Stop-Loss Excise Tax: The expense incurred by the Health Reinsurance Trust Fund will be supported by a tax on health insurance premiums. This tax will also be applied to self-funded plans and HMOs.

Using a tax on employer-sponsored health benefits as a financing mechanism for the Health Reinsurance Trust Fund should be considered along with the question of whether benefits under the Trust Fund should be extended to persons who are not insured by employer programs. If the reinsurance trust fund is to apply to all persons, then some other funding source, such as an employer or individual income tax, a change in the tax deductible status of employer provided health benefits, or a sin tax could be considered.

IMPACTS OF H.R. 2121 ON AVAILABILITY

AND COST OF GROUP INSURANCE

H.R. 2121 includes many provisions that closely parallel H.R. 3626, but takes a somewhat different approach to solving the problems of access and coverage. Although the bill applies to all employer sizes, many of its provisions will primarily affect the small group market, and many of the impacts on that market will be similar to those of H.R. 3626.

Like H.R. 3626, H.R. 2121 will not successfully address the high cost of small group insurance coverage. Average premium costs will rise because of an enhanced benefit plan for many small groups, limitations on preexisting conditions exclusions, and guaranteed access. However, under H.R. 2121 the cost increase is potentially greater because of year-round open enrollment. Despite the protections against preexisting conditions for late enrollees, the potential for selecting coverage only when it is needed is much greater.

The average cost increases imposed by the above provisions of H.R. 2121 will also be exacerbated by the requirement of full community rating. The effect of the immediate implementation of community rating, without regard to employer size and without any transition

9