Medigap reform legislation of 1990: a 10-year review.

The 1990 Medigap reform legislation sought to make it easier for consumers to compare policies, provide market stability, promote competition, and avoid adverse selection. Evidence is that the standardization of benefits has simplified consumer choice and is strongly supported by consumers and State regulators. The 1990 reforms also decreased carrier and agent abuses. However, loss ratios (the proportion of premiums paid in benefits versus being retained for administration and profit) have changed little since 1990, bringing into question whether price competition has been enhanced. The prescription drug benefit, which is included in 3 of the 10 standardized plans, provides only limited financial protection yet is expensive, one reason being adverse selection. Access to coverage for Medicare disabled beneficiaries is problematic in most States.

Introduction

Since the inception of the Medicare Program, beneficiaries have purchased individual supplemental insurance—known as Medigap—which reimburses for Medicare cost sharing and selected services that Medicare does not cover. An estimated 27 percent of beneficiaries have Medigap coverage, second only to the 36 percent who have coverage from a former or current employer (Rice and Bernstein, 1999). Approximately 15 percent are covered by Medicare managed care plans. However, enrollment in these plans peaked in November 1999 at 6,356,000, declining to 4,963,000 in August 2002, a 22-percent decline (Centers for Medicare & Medicaid Services, 2003), underscoring the critical role that Medigap plays in supplementing Medicare coverage.

The Medigap market changed dramatically with the passage of the Omnibus Budget Reconciliation Act (OBRA) of 1990. That legislation, most of whose provisions went into effect in 1992, requires that all new Medigap policies sold conform to 1 of 10 standardized benefits packages, labeled Plans A through J. Other provisions of OBRA 1990 include the following:

Higher requirements for loss ratios, which is the percent of the premium dollar that is paid in benefits rather than being used for administration and profit. Minimum loss ratios were set at 65 percent for individual policies and 75 percent for group policies. Failure to meet these standards generates requirements for premium refunds.

Severe penalties on agents or insurers who knowingly sold duplicate policies.

Limitations on agent commissions during the initial year of coverage to no more than twice the commissions for renewal policies. This provision is intended to discourage agents from churning, i.e., inducing beneficiaries to switch policies in order to generate commission income.

A requirement that insurers hold a 6-month open enrollment period when beneficiaries who are 65 or over first enroll in Part B of Medicare. During that period, a person can purchase any policy offered regardless of health status and receive the carrier's most favorable rate.

Limitations on pre-existing condition exclusions to no more than 6-months duration.

We conducted a study from 1992 to 1995 on the impact of this legislation on the performance of the Medigap market (McCormack et al., 1996). That study concluded that most of the objectives of the legislation appeared to have been met. Consumers were better able to make choices because comparing the policies across carriers was facilitated, marketing abuses had declined, and the level of competition was adequate. However, since the most recent data used in that study were from early 1995, less than 3 years after implementation, how the market would behave in the long run was not yet clear.

This article examines the Medigap market a decade after the implementation of the OBRA-1990 reform provisions. Like the previous one, the findings reflect site visits to six States; interviews with insurance carriers, government officials, and policymakers; and analyses of Medigap sales trends. The article addresses the following questions:

What policies are most popular with consumers, including two variants on the standardized plans, i.e., (1) plans with limited networks, known as Medicare SELECT, and (2) plans with high deductibles?

Are Medicare disabled, who do not come within the scope of the Federal open enrollment provisions, able to obtain Medigap policies?

How expensive are Medigap policies, how has this changed over time, and what patterns are there among the 10 plan types?

Is prescription drug coverage available, and is there evidence of adverse selection?

What is the impact of the Medigap reform provisions on consumers?

Should the standardized benefit packages be restructured?

The concluding section raises a number of overarching issues associated with Federal regulation of the Medigap market.

Data Sources

The information compiled for this study was collected from several sources. First, we conducted site visits in fall 2001 to the insurance departments in six States: Florida, Missouri, New York, South Carolina, Texas, and Washington. These States were also site-visited as part of the previous study (McCormack et al., 1996). They were selected to provide variation in geography, population, and regulatory environment. In each of these States, we met with the principal staff to discuss topics such as: trends in the Medigap market, policy benefits, policy filings, premiums, reporting requirements, consumer information, complaints, Medicare SELECT, and impacts of the Medicare+Choice (M+C) program on the Medigap market. Typically, six to eight staff members were interviewed in each State.

Second, trend data were requested from the six States on the number of carriers selling Medigap policies in the State and on consumer complaints. Also, in our previous study (McCormack et al., 1996), data were obtained on premiums and sales for each policy type (A-J) for the five largest carriers in the State for the year 1994. In the current study the same data were obtained for the same carriers for the years 1997 and 2000, thus generating a time series that spans 6 years.

Third, we conducted telephone interviews with 15 insurance carriers that sell Medigap policies. Respondents were queried about such issues as sales trends and other changes in the market, how premiums are set, experience with biased selection, issues regarding prescription drugs, whether the open enrollment provisions for the aged should also apply to Medicare disabled beneficiaries, experiences with Federal and State regulators, as well as their opinions about how public policy and regulation of the market might be changed.

Finally, telephone interviews were conducted with some 20 Federal officials, representatives of interest groups, and policymakers in order to obtain factual information and to elicit the perspectives of key stakeholders about market trends, successes and failures of Medigap and Medigap regulation, and recommendations for policy changes.

Findings

Policy Offerings and Sales

The period immediately following the implementation of policy standardization in 1992 found fewer insurance carriers offering Medigap plans. Most of the carriers that exited the market had small market shares and had difficulty being price competitive with the large carriers, in part because they faced diseconomies of scale in such activities as marketing, claims processing, and regulatory compliance (McCormack et al., 1996).

Respondents from the six State insurance departments report that both the number of carriers and market share in their respective States have been stable the last several years. Similarly, most insurance carriers interviewed report that their respective market shares have also remained relatively stable, with some experiencing fluctuations based on factors such as the penetration of M+C plans, premium-setting policies, and whether their marketing departments were pushing Medicare supplemental products. The largest carrier, United Healthcare, the agent for AARP, accounts for approximately 21 percent of the market, with a number of (generally, State-based) BlueCross® and BlueShield® plans garnering a large portion of the remainder (National Association of Insurance Commissioners, 2001b). Among carriers that increased their market share in recent years, the main reason is related to M+C plan withdrawals. Some Medigap carriers have advertised their stability in the face of uncertain health maintenance organization (HMO) offerings.

Table 1 displays the benefits covered by each of the standardized plans. It also shows estimates of the distribution of sales in 2000 based on data from two sources: (1) five of the six site-visited States (all but South Carolina, which declined to provide the data) and (2) the Medicare Payment Advisory Commission (MedPac) (Medicare Payment Advisory Commission, 2002). The latter source, which uses data from the NAIC, is more comprehensive; the former is of value principally in that it has been collected longitudinally and thus permits an analysis of trends.

Table 1

Benefits Covered by Standardized Plans and Percent Distribution of Sales: 2000

Benefit	A	B	C	D	E	F	G	H	I	J
Core Benefits	x	x	x	x	x	x	x	x	x	x
Skilled Nursing Facility Coinsurance	–	–	x	x	x	x	x	x	x	x
Medicare Part A Deductible		x	x	x	x	x	x	x	x	x
Medicare Part B Deductible	–	–	x	–	–	x	–	–	–	x
Medicare Part B Excess Charges	–	–	–	–	–	High	Low	–	High	High
Foreign Travel	–	–	x	x	x	x	x	x	x	x
At-Home Recovery	–	–	–	x	–	–	x	–	x	x
Prescription Drugs	–	–	–	–	–	–	–	Low	Low	High
Preventive Medical Care	–	–	–	x	–	–	–	–	–	x
	Percent
Distribution of Sales (Five-State Data)	4.0	12.6	18.6	10.6	2.9	35.6	2.7	2.8	7.4	3.7
Distribution of Sales (MedPac Data)	10.0	10.0	26.0	6.0	2.0	35.0	3.0	2.0	3.0	4.0

NOTES: Core benefits include coverage of all Part A coinsurance for stays longer that 60 days, the 20-percent coinsurance, the Parts A and B blood deductible, and 365 lifetime reserve days of inpatient care. Low Part B excess charge coverage pays 80 percent of the difference between the physician's charge and the Medicare allowable rate; high coverage pays 100 percent of the difference. Low prescription drug coverage has a $250 annual deductible, 50-percent coinsurance, and a maximum annual benefit of $1,250; high coverage is identical except that the maximum annual benefit of $3,000. MedPac is Medicare Payment Advisory Commission.

SOURCES: Site visits to Florida, Missouri, New York, Texas, and Washington (South Carolina declined to provide data), 2001; and (Medicare Payment Advisory Commission, 2002).

Plans C and F remain very popular, accounting for 54.1 percent of sales between them based on the five-State data, an increase from our prior work, which estimated that, in 1994, 47.5 percent of Medigap enrollees were in one of these two plans. MedPAC places the combined enrollment in Plans C and F at 61 percent. Both plans cover all Medicare cost sharing; the difference is that Plan F, but not plan C, covers 100 percent of physician balanced billing. Plan B is also popular, accounting for 12.6 percent of enrollees as estimated by the five-State data and 10 percent in the MedPAC estimates; the difference between Plans A and B is that the latter covers the Part A deductible. The three plans that cover prescription drugs, H, I, and J, together account for only 13.9 percent (9 percent in the MedPAC estimates) compared with our prior estimate for 1994 of 15.5 percent.

Variants of the Standardized Plans

Federal law allows carriers to sell two variants of the standardized plans: Medicare SELECT and high-deductible plans. Medicare SELECT plans have the same benefits design as the standardized plans but with the added restriction that the enrollee must receive services through the carrier's contracted network in order to receive full Medigap benefits. (Medicare benefits are payable regardless.) Medicare SELECT was authorized by OBRA-1990 as a 15-State demonstration and became a national program in 1995. The objective was to allow carriers to contract with a limited network of providers. The intent was that these providers would waive Medicare cost sharing in return for greater patient volume, with the resulting savings being passed along to the consumer in the form of premium reductions. Federal regulations allow waiver of cost sharing for Part A services only, and Medigap carriers that offer a SELECT product have mostly contracted with hospitals to waive the Part A deductible. That deductible had an actuarial value of $210 a year in 2002, equivalent to slightly more than 10 percent of the cost of a typical Medigap policy (Plans C or F).

Few Medicare SELECT policies have been sold, largely because the savings have not been sufficient to generate significant consumer demand. Most beneficiaries who are willing to accept a limited network can obtain greater savings by enrolling in a M+C plan. Of the 15 carriers that were interviewed as part of this study, only 6 offered SELECT plans, and then typically only in limited geographic areas. Four of the six are BlueCross® BlueShield® plans, which, unlike most commercial carriers, typically have ongoing relationships with most hospitals in their respective service areas. In addition to the limited consumer appeal, carriers expressed concern that offering Medicare SELECT products entailed administrative costs that could not be recouped in premiums. These administrative costs result, for example, from having to develop the provider network, revise marketing materials, educate the sales force, and obtain State regulatory approval.

State representatives confirmed that SELECT plans had limited appeal. In five of the six site-visited States, between one and eight carriers offered SELECT plans, often in only a limited portion of the State. In the sixth State, Florida, 12 carriers marketed SELECT products. Overall, State representatives had little to say about SELECT plans, reflecting it not being a major factor in the market place.

The Balanced Budget Act (BBA) of 1997 also authorized insurers to offer high deductible Medigap policies. These policies correspond to Plans F and J except for an annual deductible of $1,500 in 1998 and 1999, increasing with the Consumer Price Index thereafter. The reductions in premiums from high-deductible plans are substantial. We compared the premiums for the high-deductible Plan F with the standard plan using the buyers guides obtained from the six site-visited States and found substantial reductions, from an average of $1,522 in 2001 to $670, a decrease of 56 percent. In none of the six States did any carrier sell a high-deductible plan for Plan J.

Like SELECT plans, consumers and carriers have shown scant interest in high-deductible plans. Many in the policy field have advocated incorporating higher cost-sharing into Medigap policies in light of the consistent research finding that cost-sharing reduces the use of health services and, hence, Medicare spending (McCormack et al., 1996; Christensen and Shinogle, 1997; McCall et al., 1991; Christensen, Long, and Rodgers, 1987). Thus, to some this lack of consumer interest is disappointing.

Of the 15 carriers interviewed, only 2 offered such plans, although 2 others were considering doing so. The primary reason for the paucity of offerings was the perceived lack of consumer demand; carrier representatives reported that seniors tend to be risk averse and prefer not to face large deductibles. In addition, agents may not actively seek to market the high deductible plans because sales commissions are lower. Also, some carriers expressed the concern that, while benefit payments are lower, high-deductible plans have administrative costs that are comparable to those of the regular plan, making it more difficult to meet the federally-mandated loss ratio requirements. One possible factor that contributes to the low volume of sales is that the high-deductible plans have not been well publicized. For example, many State consumer guides only describe them in footnotes to the corresponding regular plan.

Impact on Medicare Disabled Beneficiaries

In 1998, 5.0 million out of 38.8 million Medicare beneficiaries were under age 65 and qualified for Medicare by virtue of being disabled, a number that has been growing several times faster than the number of aged beneficiaries (Health Care Financing Review, 2001). Although disabled beneficiaries represent 13 percent of all Medicare beneficiaries, they represent only 1 percent of Medigap policyholders (National Association of Insurance Commissioners, 2000a). A major reason is that they are not covered by the Federal 6-month open enrollment period when someone turns age 65, although they are eligible for the 6-month open enrollment period when they turn age 65. For those who are enrolled, claims cost and resulting premiums, are considerably higher, reflecting both the high medical costs of disabled beneficiaries and adverse selection, i.e., sicker disabled disproportionately enrolling in Medigap plans. In 1998, disabled beneficiaries incurred an average Medicare cost-sharing liability of $1,043 per capita compared with $915 for aged beneficiaries, a difference of 14 percent. (Health Care Financing Administration, 2001). In contrast, annual claims cost were an average of 78 percent higher, with the difference between the two percentages being largely attributable to adverse selection (National Association of Insurance Commissioners, 2000a).

Some States have enacted legislation to improve access to Medigap benefits. In 2000 the NAIC reported that 19 States had mandated an open enrollment period for disabled, in some cases limited to Plan A (unpublished compilation prepared by the NAIC). In most of these States, however, carriers can charge more to disabled than to aged beneficiaries. Another approach that States have adopted is allowing beneficiaries to purchase coverage through State high-risk pools, priced at some multiple (e.g., twice) of the typical premium. New York requires continuous open enrollment along with community rating, i.e., carriers cannot make premium distinctions based on age, a provision that applies to disabled as well as aged beneficiaries and results in a single rate for each plan that a carrier sells. State representatives report that between 2 and 3 percent of Medigap enrollees are disabled, compared with only 1 percent nationally.

Missouri has mandated a 6-month open enrollment period when someone first becomes eligible as a disabled beneficiary. What is novel is that each carrier is required to set premiums for disabled enrollees for each of the plans that it markets at the average premium for aged enrollees. The effect is a potential cross-subsidy from aged to disabled beneficiaries. The provision became effective July 1998, at which time a 6-month open enrollment period for all disabled beneficiaries was initiated; thereafter, the provision applies only to the first 6 months after Medicare eligibility is first established.

The State estimated at the time the provision was adopted that Medigap costs overall would increase by only 1 percent. It reports no significant change in premium trends as a result of the provision, and no carrier has left the market as a result. The extent to which enrollment among Medicare disabled beneficiaries increased as a result of this provision can only be approximated because the State has tracked these numbers only from 1998 on. As one measure, in 1998 some 758 disabled beneficiaries had Medigap policies compared with 3,283 in 1999 and 3,807 in 2002, indicating that greater access to coverage did result.

Premium Trends

Sales data and premium data (for a 70-year old female) for each of the 10 standardized plans was requested for 1997 and 2000 from the States that were site visited. These data were requested of the same five carriers from our prior study (McCormack et al., 1996). Five of the six States (all but South Carolina) responded.

Table 2 shows average premiums for the selected years for each of the 10 standardized plans. In computing the averages, each of the five States was given equal weight, a procedure used in previous studies of the Medigap market (Rice, Graham, and Fox, 1997; Rice and McCall, 1985; McCall et al., 1991). Within each State, however, the data were weighted by the number of policies sold. For example, if in a particular State one carrier sold 500 policies and another sold 1,000, the latter would receive twice the weight.

Table 2

Average Annual Medigap Premiums and Changes in Premiums in Five States: Selected Years, 1994-2000

Plan	1994	1997	2000	Percent Change
				1994-1997	1997-2000	1994-2000
Plan A	$551	$717	$1,085	30	51	97
Plan B	862	1,036	1,438	20	39	67
Plan C	881	1,131	1,592	28	41	81
Plan D	897	1,091	1,513	22	37	69
Plan E	886	1,005	1,359	13	33	53
Plan F	1,137	1,287	1,716	13	29	51
Plan G	1,076	1,176	1,457	9	24	35
Plan H	1,215	1,617	2,428	33	52	100
Plan I	1,348	1,679	2,382	25	42	77
Plan J	1,851	1,939	2,665	5	37	44

SOURCE: Insurance carrier data provided by Florida, Missouri, New York, Texas, and Washington, 2001.

In 2000, the lowest average premium was $1,085 for Plan A. As would be expected, the most expensive plans were those that cover prescription drugs, with the highest being $2,665 for Plan J. Plans H and I had average premiums of $2,428 and $2,382, respectively. To place these numbers in context, in 1999 the pretax median household income for those age 65 or over was $22,800 (U.S. Bureau of the Census, 2001).

Table 2 also shows the rate of change in premiums between selected years 1994-1997 and 1997-2000. For each of the 10 plan types, premiums rose far faster in the latter period. Averaging the plan types, they rose 20 percent during the first 3 years, and 39 percent during the last 3 years. A major contributor to the increase is the growth in claims costs for hospital outpatient services. A report by the American Academy of Actuaries (2000) states that: “From 1995 to 1998, the increase in outpatient claim costs caused [the] overall trend to be 2.9 percent higher per year than it would have been if the outpatient trend had equaled the average of the other components.” However, carrier representatives report that the trend of high rates of increase has largely been reversed starting in 2001 as a consequence of the initiation that year of the phase-in of the hospital outpatient prospective payment system provisions of the 1997 BBA, which includes limitations in patient balance billing. Plans A, C, and H had particularly high premium rate increases. Various sources said that Plan A had experienced adverse selection because it was promoted by some organizations representing or serving disabled beneficiaries, some of which, including dialysis centers, reportedly subsidize the Plan A premium.

An important regulatory issue is the way in which premium rates are set. There are three different methods. They can be either community rated, meaning that all policyholders in a geographic area are charged the same amount; issue-age rated, whereby premiums are based on the age at initial purchase; or attained-age rated, whereby premiums rise as beneficiaries age. Many State officials and consumer advocates dislike attained-age rating because it results in higher premiums as beneficiaries grow older, when their incomes in real terms are likely to decline. In contrast, most insurance carriers we interviewed held that rating practices should not be restricted in order to give consumers more options, i.e., allow consumers to purchase either an attained-age or an entry-age policy, with the understanding that the former would be the less expensive in earlier years but more expensive in later years. In 2000, a dozen States had either mandated community rating or banned attained-age rating (National Association of Insurance Commissioners, 2000b). Among the six study States, New York and Washington mandate community rating, Florida requires issue-age rating, and starting in 2000, Missouri banned newly issued policies from using attained-age rating.

Prescription Drugs Benefits

Medigap prescription drug benefits have been problematic, with many of the problems being unrelated to standardization. Many carriers do not offer prescription drug coverage. Out of the 15 carriers interviewed, only 8 offer any of the three plans that includes prescription drug coverage (i.e., Plans H, I, or J). Some of the eight carriers offering coverage do so in only a limited number of States that they serve, and only three of the eight offer all three plans. The major policy question, however, is not the proportion of carriers that offer prescription drug coverage but whether consumers have sufficient choices. Indeed, we found that choices were available, although in some States they were few in number, notably in the case of New York, which may reflect that State's regulatory environment. Of particular note is New York State's requirement that carriers have continuous open enrollment, meaning that beneficiaries can purchase any of the 10 plans at any time regardless of health status. To illustrate, a Medicare beneficiary who waited until age 75 before purchasing Medigap coverage could do so from any carrier, although that individual could be subject to a 6-month pre-existing condition exclusion. Table 3 provides the number of plan offerings in each of the six States that were site visited, based on information in State-issued consumer guides.

Table 3

Number of Carriers Selling Medigap Prescription Drug Coverage in Selected States: 2001

State	Plans				Total Number of Carriers
	H	I	J	H, I, J
Florida	7	12	4	12	37
Missouri	8	9	5	13	54
New York	7	3	1	8	16
South Carolina	3	3	3	5	32
Texas	10	10	8	14	59
Washington	8	15	10	17	35

NOTES: Carrier listing in the South Carolina consumer guide is voluntary; thus the data for the State may represent an undercount. Also, although most carriers sell statewide, five of the seven carriers selling Plan H in New York are BlueCross® BlueShield® plans that operate in only a portion of the State. The three carriers that market Plan I and the one that markets Plan J are all statewide.

SOURCE: State-issued consumer guides from Florida, Missouri, New York, South Carolina, Texas, and Washington, 2001.

The bigger problem is its limited scope and high cost of coverage, due in part to adverse selection. The drug benefit in Plans H and I has a $250 annual deductible, 50 percent coinsurance, and an annual maximum on benefit payments of $1,250. Actuarial estimates show that in a typical population of Medicare beneficiaries, the drug benefit would pay on average an estimated 35.5 percent of expenses. The benefit in Plan J differs from Plans H and I only in that the annual maximum is increased to $3,000; it reimburses some 39 percent of drug expenses on average.

In addition to paying only a small proportion of drug expenses, the premiums far exceed the actuarial value of the benefit in an average population of beneficiaries. Plans H, I, and J can each be paired with another plan that is similar, except that it excludes prescription drugs. The close matches make prescription drug benefits nearly equivalent to a freestanding rider, which is likely to generate adverse selection as a result of beneficiaries with chronic conditions disproportionately electing the coverage. The matches and the benefit differences other than drugs are as follows:

Plans H and C differ only in the coverage of the Part B deductible.

Plans I and D differ only in the coverage of excess charges.

Plan J and E differ only in the coverage of the Part B deductible, physician excess charges, and at-home recovery.

For each pair, the impact of benefit differences other than prescription drugs was removed, leaving a dollar amount that can be attributed to drug coverage, as shown in Table 4. The resulting amounts can be compared with the actuarial value of the drug benefit in a cross-section of the Medicare population. That value has been estimated at $403 for Plans H and I and at $435 for Plan J. We increased these numbers by 25 percent to account for a typical administrative load, resulting in estimated values of $504 for Plans H and I and $544 for Plan J. The actuarial value of $504 is significantly below with the actual premium attributable to prescription drug coverage of $974 for Plan H and $954 for Plan I. Similarly, the actuarial value of $544 is less than one-half the actual premium attributable to drug coverage of $1,131 for plan J. Although these numbers represent approximations, they are strongly indicative of the presence of biased selection.

Table 4

Premium Adjustments to Remove Impact of Non-Drug Benefits: 2000

Plan	Adjustment
Plan H to Plan C
Plan H Premium	$2,428
Part B Deductible	-110
Adjusted Premium	2,538
Plan C Premium	1,564
Premium Attributable to Drug Coverage	974
Plan I to Plan D
Plan I Premium	$2,382
Physician Excess Charge—High	-3
Adjusted Premium	2,379
Plan D Premium	1,425
Premium Attributable to Drug Coverage	954
Plan J to Plan E
Plan J Premium	$2,665
Part B Deductible	-110
Physician Excess Charge—High	-3
At-Home Recovery	-2
Adjusted Premium	2,550
Plan E Premium	1,419
Premium Attributable to Drug Coverage	1,131

SOURCE: Actuarial estimates and data provided by Florida, Missouri, New York, Texas, and Washington, 2001.

Table 2 shows that rates of increase in the three plans that include prescription drug benefits for Medigap were only slightly above that of the other seven plans, which would appear to indicate that adverse selection has not worsened over time given the inflationary nature of drug benefits. The American Academy of Actuaries Medicare Supplement Insurance Work Group (2000) conducted a detailed analysis of Medigap carrier claims for the NAIC and found that “…the average annual trend for drug benefits (from 1996 to 1998) for the drug benefits under plans H-J was 15.0 percent. This is thought to be consistent with the generally increasing drug expenditures of the 65+ population, where relatively high increases in drug costs and utilization have been observed.” (American Academy of Actuaries Medicare Supplement Insurance Work Group, 2000.) One limitation of that study is that it covers only a 2-year period.

The high cost of coverage deters its being purchased. Using NAIC data, MedPAC (2002) estimates that in 2000, Plans H, I, and J represented only 9 percent of all standardized plans (Table 2). This figure is lower than the 13.9 percent figure based on our State data. However, because we sampled only large carriers, and United Healthcare/AARP, which is represented in our sample in all States, offers drug coverage nationally, our estimate is likely to be overstated.

What is surprising is the difference between the MedPAC data and that from the Medicare Current Beneficiary Survey (MCBS), a sample survey of beneficiaries. The MCBS found that in 1998, 43 percent of beneficiaries with “individually purchased coverage only” had policies that included drug coverage (Poisal and Murray, 2001). Individually purchased coverage includes beneficiaries with prestandardized policies, but these account for only approximately one-third of all Medigap policies, less than 15 percent of which had prescription drug coverage in the early 1990s (Rice, Graham, and Fox, 1997). We hypothesize that the MCBS substantially overestimates the prevalence of prescription drug coverage. One possible explanation is that many carriers offer drug cards to enrollees without prescription drug coverage. These are essentially discount cards that achieve savings of around 15 percent, largely due to contracted pharmacists' accepting lower dispensing fees. Enrollees with these cards may commonly reply affirmatively to the question of whether they have prescription drug coverage. The difference between the two estimates is important because the MCBS has received widespread attention in policy circles and appears to have resulted in the proportion of beneficiaries with prescription drug coverage being overstated.

The Bush Administration has proposed that carriers be allowed to offer two additional Plans, labeled K and L, that combine drug coverage with reduced coverage of Medicare cost sharing. Plan K would provide the same drug coverage as Plan J; however, it would not cover the Part B deductible and would only cover 50 percent of Medicare cost sharing, with a $4,000 limit on out-of-pocket expenses. Plan L would provide the same coverage as Plans H and I; however, it would not cover the Part B deductible and would pay 75 percent of Medicare cost sharing with a $2,000 limit on out-of-pocket expenses. Such an approach has the potential to be attractive to consumers without generating the extent of adverse selection that has characterized Plans H, I, and J, for which there are close counterparts that do not have drug coverage. A case can be made that, for beneficiaries without any supplemental coverage, the financial exposure is greater for prescription drug expenses than for Medicare cost sharing. However, whether these two plans would be attractive to consumers is not known, particularly in light of the lack of interest in high-deductible plans.

The standardization of prescription drug benefits illustrates the tradeoff between enhancing consumer understanding and encouraging innovation. Left unstandardized, prescription drugs are perhaps the most confusing of all benefits because of the almost infinite variations among health plans with respect to differences in the wording that is used to describe identical benefits, the cost sharing structure, and the nature of the cost-containment measures that are undertaken (Fox et al., 1999). While enhancing consumer understandability, standardization has severely restricted the ability of carriers to implement cost-containment measures. Carriers may engage in educational programs with patients and providers, although these are generally viewed as having limited impact. Furthermore, the high level of cost sharing in Plans H, I, and J creates an incentive for the consumer to be cost conscious.

What current law precludes is a myriad of cost-management techniques that are increasingly being adopted by employee benefit programs in both the private and the public sector. First and foremost is the incorporation of financial incentives to encourage the use of drugs that are on a formulary. Some 79 percent of prescription drug expenses in 1998 were for brand name drugs that did not have generic equivalents (Henry J. Kaiser Family Foundation, 2000). For a significant number of these, multiple products within a given drug class are available that are roughly equivalent in terms of effectiveness. Private-sector purchasers, such as employers, commonly create incentives and take other measures to encourage the use of less expensive drugs. Medigap carriers are also precluded from requiring prior authorization for selected high-cost drugs, something that many private purchasers do if safe, less expensive drugs are available that might be tried first or if the drug can be used for purely cosmetic purposes.

The constraints in OBRA 1990 on Medigap carriers illustrate two dilemmas associated with standardization. First, many of the cost-management activities, which are commonplace today, were rare 10 years ago when the standardized policies were established, yet the legislation has not been modified. One argument for not changing the law is to avoid the inevitable resulting consumer confusion and additional administrative expenses to the carriers. Second, carrier flexibility on such matters as formulary selection would militate against standardization, creating the potential for consumer misunderstanding and for carriers' developing overly restrictive formularies. It would be difficult to standardize the formulary (or other aspects of cost containment), particularly as an actively managed formulary entails making regular changes as new medications enter the market, pharmaceutical manufacturers alter their prices, and new scientific evidence about individual drugs becomes available.

Consumer Impact

Standardization can benefit consumers, first, by facilitating consumer understanding and, second, by enhancing price competition, resulting in premiums being lower than they would have been otherwise. Available evidence is that the first objective has been achieved but not the second. Attitudes towards standardization are overwhelmingly positive. Consumer representatives at both the national and the State level report that Medicare beneficiaries have become knowledgeable about the 10 standardized plans and are comfortable choosing among them.

One measure of consumer impact is the number of Medigap-related complaints received by State insurance departments. Because reliable longitudinal data on consumer complaints are not available nationally, these data were requested from the six study States. All of the States have current tallies on the number of Medigap complaints, but some do not maintain data for previous years. Most States only record a contact as a complaint if it appears justified. Thus, if a consumer calls to complain about high premiums but the company is charging rates that were approved by the State, the contact is generally not recorded as a complaint. Table 5 presents complaint data from 1995 to 2000 in four of the six States— Florida, New York, Texas, and Washington.

Table 5

Consumer Complaints, by Site-Visited States: 1995-2000

State	1995	1996	1997	1998	1999	2000
Florida	406	479	364	362	381	428
New York	162	119	111	115	132	146
Texas	40	91	141	97	123	112
Washington	19	26	18	14	40	45

SOURCE: Primary data provided by Florida, New York, Texas, and Washington, 2001.

Complaints regarding Medigap supplemental insurance filed with State departments of insurance remained fairly steady through 1998 but increased in 1999 and 2000, due largely to the impact of Medicare managed care plans' withdrawing from the market, necessitating that more individuals consider joining Medigap plans. Never-theless, it is important to realize that the numbers are small. For example, the State of New York receives 60,000 insurance-related complaints a year, of which only a small fraction relate to Medigap. McCormack et al. (1996) reported a significant decline in the number of complaints after standardization was implemented. A majority of the complaints referenced delays in claims handling and in applying medical underwriting to prospective applicants.

We find no evidence of greater price competition as a result of standardization. One would expect greater price competition to cause premiums to be lower than they would have been otherwise, resulting in a higher proportion of the premium dollar being devoted to benefit payments rather than administration and profit, i.e., one would expect loss ratios to rise. Table 6 displays loss ratios from 1990 to 2000 and shows no evidence of higher loss ratios since the advent of standardization (National Association of Insurance Commissioners, 2001b). Indeed, the loss ratio in 2000 was 80.4 percent, compared with 83.4 percent in 1991, the year prior to the implementation of standardization.

Table 6

Average Loss Ratios for All Medigap Policies: 1990-2000

Year	Loss Ratio
	Percent
1990	81.2
1991	83.4
1992	79.7
1993	75.9
1994	81.3
1995	85.6
1996	82.5
1997	82.7
1998	79.8
1999	79.3
2000	80.4

SOURCE: (National Association of Insurance Commissioners, 2001.)

State regulatory and consumer information staff endorse standardization because it facilitates beneficiary choice. One unforeseen consequence of standardization, a favorable one, is that it has enabled States to reduce staffing levels devoted to regulating the insurance market because it is neither a significant source of consumer confusion or complaints nor of carrier or agent fraud. Other lines of insurance are viewed as more problematic. Accurately quantifying staffing levels is difficult because staff often work on several lines of insurance. Also, consumer inquiries may be handled either by the State insurance department, a separate department (usually, the Departments on Aging), or by the State Health Insurance Assistance Programs (SHIPs), which are federally funded but may be administered by either a government agency or under contract with private, nonprofit organizations. To illustrate the small levels of staff devoted to regulation in four of the site-visited States: New York has two staff members devoted to rate reviews and two attorneys who review marketing materials and regulations, each of whom spends only 5-10 percent of their time on Medigap matters, and four staff members at the SHIP. In Florida, two individuals spend less than one-half of their time on regulatory matters. Missouri and New York each have 1.5 full-time equivalents devoted to Medigap regulatory issues including rate reviews. South Carolina has only two staff members who spend measurable amounts of time, each between 20 and 25 percent, on Medigap regulatory.

States ameliorate consumer confusion by instituting consumer information efforts. The six States we visited all had information for beneficiaries available online, including consumer guides that explain Medigap plans and offer charts with contract and premium comparison information for individual Medigap carriers. In addition, they mail out consumer guides, conduct outreach and consumer counseling, sponsor toll-free information numbers, and run volunteer programs. The SHIPs provide one-on-one counseling services and group seminars. Two areas of significant consumer confusion are: (1) the Federal guaranteed issue requirements and (2) the difference between issue and attained-age rating in States where attained-age rating is allowed. The 1997 BBA expanded guarantee issue for Medigap policies in three ways:

Beneficiaries who enroll in a Medicare HMO at age 65 and who disenroll within one year can enroll in any Medigap Plan (A-J) within 63 days of disenrollment.

Beneficiaries who switch from a Medigap policy to a Medicare HMO and who disenroll within 1 year can re-enroll in that same Medigap policy type within 63 days of disenrollment.

Beneficiaries can enroll in Medigap Plans A, B, C, or F within 63 days of any of the following qualifying events: (1) their employer terminates its Medicare supplemental plan, (2) their Medicare HMO terminates coverage, (3) they move outside of the HMO's service area, or (4) their Medigap plan becomes insolvent or violates the terms of contract.

Widespread confusion has caused not only beneficiary but also by carrier misconceptions about how and when guarantee issue provisions apply.

Benefit Package Issues

Standardizing benefits inevitably raises controversy as policymakers seek to address the often competing issues of: simplicity and ease of understanding, allowing a reasonable range of consumer choice, achieving public health objectives such as encouraging prevention, and restraining increases in Federal spending resulting from Medigap policies' paying for Medicare cost sharing. The 10 plans include several benefits whose value has been debated, among them: the Part B deductible, preventive care, and at-home recovery. Elimination of these benefits would reduce the number of standardized plans below 10, something that some consumer advocates have favored.

Coverage of the annual $100 Part B deductible is commonly described as dollar trading, rather than insurance, since in any given year roughly 90 percent of beneficiaries spend the deductible amount, and the cost of the coverage is typically priced above $100 after factoring in administrative costs. However, consumer choice of plans indicates that some beneficiaries want their Medigap policies to cover all of the Medicare cost sharing and, therefore, like this benefit. Another reason for wanting coverage is to minimize the burden of filing claims.

The prevention benefit reimburses up to $120 in charges for virtually any preventive service. It is included in Plans E and J, which are rarely purchased (Table 1). This was done at the behest of some consumer representatives, not as a form of insurance, but to further public health objectives. The experience of two very large Medigap carriers (which wish to remain anonymous) is that the value of this benefit, including the administrative load, for those who have the benefit is around $1 per year, indicating that it is hardly ever used. It is possible that doctors are able to justify most medical services as illness related. Also, beneficiaries may not understand, or particularly want, this benefit. It was controversial when it was adopted in the early 1990s, and its value may be less today because Medicare now covers more prevention services than it did then, mostly as a result of expansions in the 1997 BBA.

The at-home recovery benefit also adds little to premiums—between $2 and $3 a year for the two benefits combined—indicating that beneficiaries who have this benefit rarely use it, although it may provide an element of financial security. The athome recovery benefit is poorly understood and is difficult to administer. It covers “…short-term, at-home assistance with activities of daily living for those recovering from an illness, injury, or surgery.” There is a series of limitations that are confusing and that are not fully described in any of the consumer guides of the States that we site visited. They are also not articulated in the consumer-oriented information that the Medicare Program provides beneficiaries, e.g., the requirement that the total number of at-home recovery visits not exceed the number of Medicare-approved home health visits.

The underlying benefit structure of the Medigap policies can also be questioned. The major focus is to cover most or all (depending on the specific plan) Medicare cost sharing. From the plan choices of beneficiaries, one can infer that beneficiaries want all cost sharing covered. However, the effect is to increase the utilization of health services and, hence, Medicare expenditures (McCormack et al., 1996).

Conclusion

Overall, the Medigap reform legislation has had a favorable impact. Consumer confusion has lessened, and consumers like the Medigap reforms that OBRA 1990 introduced. State regulators are supportive, both because the changes it introduced are popular with consumers and because it has lessened both the number of beneficiary complaints and carrier or agent abuses, thereby decreasing the enforcement burden. However, we find no evidence that it has enhanced price competition.

As one would expect, issues remain. One question is whether to change the 10 standardized plans in light of both the experience gained with the OBRA 1990 provisions over the last decade and the changes in Medicare, e.g., the expanded coverage of preventive services. Questions arise as to whether the continued inclusion of certain benefits is desirable, whether first-dollar coverage should be avoided in order to reduce the cost-increasing impact on the Medicare budget of the 10 current plans, and whether some level of prescription drug coverage should be mandated in all of the plans. Any restructuring of the standardized benefits will generate dislocations and costs. Medicare beneficiaries would have to become re-educated to a new set of benefits; whether and how to convert existing plan designs to the new set of benefits would have to be addressed; and carriers would face administrative expenses as a result of having, for example, to rewrite their policies, educate existing enrollees, and revise marketing materials. Some argue that any changes in the 10 standardized plans should await broader Medicare reform, although the timing of such reform is at best is conjectural.

Notwithstanding these reservations, there are certain changes that would involve minimal beneficiary confusion or dislocations. First, carriers selling the plans with drug coverage (H, I, and J) could be allowed to engage in broader cost-management activities, such as encouraging use of contracted pharmacies so that the beneficiary does not pay full retail price and allowing carriers to require prior authorization for expensive drugs. Such measures have the potential for lowering premiums. Federal standards would be desirable, e.g., regarding network size or the types of drugs that could be subject to prior authorization. Second, some of the benefits that are hardly ever used—e.g., inhome services, which is also poorly understood—could be eliminated, which could lead to a reduction in the number of offerings. Whatever their theoretical merit, the fact that enrollees hardly ever use these benefits indicates a lack of consumer appeal, something that was not clear at the time the 10 standardized packages were designed. Finally, greater effort might be made to publicize the availability of high deductible plans.

Some analysts believe that the standardization provisions enacted in 1990 were too limited and should have included standardizing the policy forms themselves, which now vary from State to State, increasing regulatory compliance cost for multi-State carriers. The main aspect of Medigap that has not been standardized is the manner in which age is reflected in premiums. Most carrier representatives with whom we talked felt that beneficiaries should be able to choose between attained or entry-age rating. However, few beneficiaries understand the difference. Some States have handled the issue by precluding attained-age rating, and consideration might be given to the Federal Government standardizing age-rating practices.

The absence of any open enrollment provision in Federal law for Medicare disabled beneficiaries remains problematic. Many of the representatives of the carriers with whom we spoke oppose such a provision, fearing adverse selection. However, the program in Missouri, which not only requires open enrollment in all 10 plans, but also sets a limit on what may be charged based on the premiums charged the aged population, was not opposed by the industry and has resulted in only minor increases in premiums.