#2001-01 January 2001 Restricting Underwriting and Premium Rating Practices in the Medigap Market: The Experience of Three States by: The Lewin Group Steven Lutzky, Lisa Maria B. Alecxih, Veena Pankaj, Stephanie Laud, Genevieve Schaab Gerry Smolka, PPI Project Manager The Public Policy Institute, formed in 1985, is part of Public Affairs at AARP. One of the missions of the Institute is to foster research and analysis on public policy issues of interest to mid-life and older Americans. This paper represents part of that effort. The views expressed herein are for information, debate, and discussion, and do not necessarily represent formal policies of AARP.  2001, AARP. Reprinting with permission only. AARP, 601 E Street, NW, Washington, DC 20049 Table of Contents EXECUTIVE SUMMARY................................................................................................................................. ES-1 BACKGROUND ....................................................................................................................................................ES-1 PURPOSE ............................................................................................................................................................ES-1 METHODOLOGY .................................................................................................................................................ES-1 FINDINGS............................................................................................................................................................ES-2 CONCLUSIONS ....................................................................................................................................................ES-3 I. INTRODUCTION ........................................................................................................................................... 1 OBRA-90 LEGISLATION STANDARDIZED MEDIGAP POLICIES ..................................................................................1 SOME STATES HAVE ADOPTED REGULATIONS MORE STRINGENT THAN THOSE REQUIRED BY OBRA-90 ...................3 II. METHODOLOGY .......................................................................................................................................... 5 SELECTION OF STUDY AND COMPARISON STATES ....................................................................................................5 ANALYSES................................................................................................................................................................6 SOURCES OF DATA ...................................................................................................................................................6 OUTCOME MEASURES ..............................................................................................................................................7 III. FINDINGS...................................................................................................................................................... 10 AVAILABILITY ........................................................................................................................................................11 AFFORDABILITY .....................................................................................................................................................14 IV. CONCLUSIONS............................................................................................................................................ 21 POLICY LESSONS AND CONSIDERATIONS ................................................................................................................22 QUESTIONS FOR FURTHER RESEARCH ....................................................................................................................22 Figures and Tables TABLE 1: Benefits Covered By The Ten Standardized Medigap Plans.....................................................................2 TABLE 2: States With Regulations Exceeding OBRA-90 Requirements – 1997.......................................................4 FIGURE 1: Number Of Companies Actively Marketing Medigap Plans In Study States and Comparison States for 1998........................................................................................................................................................12 FIGURE 2: Number of Companies Offering Guarantee Issue Policies in Study States and Comparison States for 1998........................................................................................................................................................13 FIGURE 3: Average Weighted Premium Increases from 1995 to 1998 in Study States and Comparison States .....16 FIGURE 4: Average Weighted Premium Increased from 1995 to 1998 for Guarantee Issue Policies in Study States and Comparison States ...........................................................................................................................17 FIGURE 5: Average Weighted Medigap Premiums Purchased at Age 65 as Percentage of Median Family Income in Study States and Comparison States for 1995 and 1998 ........................................................................18 FIGURE 6: Average Weighted Medigap Premiums Purchased at Age 80 as a Percentage of Median Family Income in Study States and Comparison States for 1995 and 1998 ....................................................................19 TABLE 3: Premium Growth for AARP Medigap Policies, 1995-1998 ....................................................................20 FIGURE 7: Average Premium Increases Between 1995 and 1998 (Weighted by Number of Policies Sold in Each State) for Policies Offered by AARP......................................................................................................21 Appendix A - Data Included in Graphs APPENDIX TABLE 1: Number Of Companies Offering Plans In 1998 And Changes In Premiums 1995-1998 ......... A-1 APPENDIX TABLE 2: Weighted Average Premium Increases in Plans Offered on a Guarantee Issue Basis, 1995-1998............................................................................................................ A-2 APPENDIX TABLE 3: Weighted Average Premiums, Median Family Income, and Premium as a Percentage of Income in 1995 and 1998 in Study and Comparison States .................................................... A-3 Foreword Because roughly a third of Medicare beneficiaries rely on Medicare Supplemental (Medigap) insurance to help protect them from health care costs not covered by original Medicare, access to affordable Medigap coverage is critical to the financial security of many older persons. Insurance practices can negatively affect access to and the cost of Medigap coverage. Underwriting (using health status information to determine whether or not to cover someone or to charge a higher premium) can deny some applicants the opportunity to buy a policy. Likewise, premiums for Medigap policies can be set to favor young and healthy buyers, and to discourage older and/or sicker buyers. Federal law limits underwriting in the Medigap market in certain situations. When an individual is age 65 or older and first enrolled in Medicare Part A and B, there is a six-month period (commonly referred to as “open enrollment”) when insurers cannot underwrite. The Balanced Budget Act (BBA) of 1997 prohibited underwriting in specific additional situations, such as when a beneficiary’s Medicare+Choice plan terminates its contract with Medicare, when a Medicare beneficiary moves out of the Medicare+Choice plan’s service area, or when a Medicare beneficiary loses employer-sponsored health benefits. Unless state regulation further limits underwriting, Medigap insurers are free to underwrite Medigap coverage outside of the situations covered by federal law. Without a Medigap insurer that will accept them, Medicare beneficiaries may find that they are unable to buy a Medigap policy to help with their health care expenses. During the situations identified above, federal law prohibits insurers from charging individuals higher Medigap premiums because of health status or claims experience. Outside of these federally protected situations, state law governs insurers’ premium rating practices. Whereas community rating was the prevailing method used by insurers to set Medigap premiums a number of years ago, issue age rating and attained age rating have largely supplanted community rating in states that do not specify methods for setting Medigap premiums. The growing prevalence of attained age rating has the effect of raising Medigap premiums for older policyholders, making coverage more expensive as they age. Where age rating and community rating coexist in a market, insurers using community rating, such as United HealthCare’s policies for AARP members, may find that policyholders with age-rated policies switch to community- rated products as they age. Since increased age is generally associated with increased costs, this may potentially adversely affect the cost of community-rated products. Some states have gone beyond federal requirements and have further limited insurance underwriting and rating practices in the Medigap market. This study was conducted by The Lewin Group on behalf of AARP. AARP funded this study to learn about the experience under underwriting and rating reforms in three such states and to compare the experience in those states with that in three comparison states where no such reforms had been enacted. Gerry Smolka, Senior Policy Advisor Public Policy Institute i EXECUTIVE SUMMARY Background Although many older Americans rely on Medicare to help pay for medical expenses, individuals still face high out-of-pocket expenses for items that are not covered by Medicare. Medicare Supplemental insurance, also known as Medigap, helps many older Americans shoulder out-of- pocket costs not covered by Medicare. The Omnibus Budget Reconciliation Act of 1990 (OBRA-90) helped simplify consumers’ ability to choose Medigap policies by developing ten standardized Medigap plans. Although OBRA-90 legislation made it easier for consumers to differentiate Medigap plans when selecting coverage, it did not address certain insurance practices that affect the price of and access to Medigap plans. Therefore, Medigap insurers are still permitted to: 1) screen for health conditions (underwriting) when issuing policies; and 2) set premiums using a method which automatically increases rates as policyholders age (attained age rating). These practices can limit policy choices and increase premiums among individuals with health problems. Eleven states have responded to these concerns by adopting requirements that restrict underwriting and/or premium rating practices. Purpose No one has systematically studied the effect of state requirements exceeding those in OBRA-90. Theoretically, these regulations could either help consumers by leveling the playing field among insurers in the market or hurt consumers by inhibiting competition and decreasing affordability of Medigap policies. Insurers have argued that these regulations are harmful to the Medigap market and therefore harmful to consumers. This study explores the effect on the Medigap market of requirements that expand consumers’ ability to purchase policies regardless of their health status (open enrollment) and that restrict certain rating practices. The intent of this study was to assess whether states had viable Medigap markets after the implementation of regulation of underwriting and premium setting practices by examining the availability and affordability of Medigap coverage in those states in comparison to states without such regulation. It is important to note that other factors, such as the growth of Medicare managed care, have likely affected the Medigap market during the timeframe studied. It was beyond the scope of this study to conduct an in-depth analysis of all of the various factors that may have affected the market during the study period. Methodology The Medigap markets in three states that have regulations exceeding the OBRA-90 requirements (New York, Connecticut, and Florida, referred to as “study states”) were compared with markets in three states without additional regulations (Arizona, Ohio, and Virginia, referred to as “comparison states”). We also analyzed the experience of Medigap policies offered to AARP ES-1 members through Prudential and United HealthCare1 (hereafter referred to as AARP Medigap policies) in all states with and without additional regulations. The methodology for contrasting study and comparison states relied on secondary analyses of rate books created by the state insurance departments to inform consumers, and state-specific 1996 Medicare supplement insurance experience reports assembled by the National Association of Insurance Commissioners (NAIC). We supplemented these analyses with interviews with state regulators to determine if companies listed were actively marketing products within the state and to obtain their overall impression of the market. We assessed the viability of the Medigap market by examining the availability and affordability of policies. We assessed the availability of Medigap policies by examining the number of insurers offering policies. We selected several outcome measures to assess the affordability of Medigap policies in study states and comparison states. To assess affordability we examined the rate of increase of Medigap premiums between 1995 and 1998 and the cost of premiums as a percentage of median family income. We conducted separate analyses for all Medigap policies being marketed and for guarantee issue policies, which are available to all consumers regardless of their health conditions. We also analyzed data for AARP Medigap policies nationwide in order to compare the rate of premium increases in states with different levels of regulation. Findings The results of our analyses comparing study states with comparison states suggest that markets in the study states and the comparison states were similar in that: 1. consumers had multiple insurers from which to choose; 2. premiums generally increased at substantially similar rates; and 3. premiums accounted for a similar percentage of median family income. Medigap markets in the study and comparison states differed in that: 1. consumers with health conditions had more choices of insurers in the study states; 2. more companies, on average, offered policies in the comparison states, but the difference in the number of companies offering policies was smaller among insurers with more than one percent of the market. 1 AARP offers Medigap policies to its members. Prior to 1998, AARP offered Medigap policies through Prudential; since 1998 they have been offered through United HealthCare. Policies available to AARP members are particularly relevant because they account for a large share of the Medigap market nationally (up to one- quarter). Premiums for these policies do not differ by age, and are available to members regardless of health (i.e., community-rated and guarantee issue). In many states, policies available through AARP represent the only Medigap options available to all individuals at the same rate regardless of age and health. ES-2 Analyses of premiums for AARP Medigap policies across the country did not show any evidence that premiums were increasing at a faster rate in states that had implemented additional regulations than in other states. Premiums for Plan A were an exception because they generally increased more in states without additional regulation. Conclusions State regulators in the study states concluded that the additional regulations have had their desired effect and have not significantly harmed competition. Restricting underwriting and premium rating practices appears to increase access to Medigap policies for individuals with health conditions, while leaving healthy consumers with multiple affordable choices among insurers. Our findings indicate that while additional regulations decreased the pool of insurers competing in the market, consumers still have choices among insurers. The study also found that affordability of Medigap, measured as premiums as a percent of median family income, appears to have declined substantially between 1995 and 1998 in both the study and comparison states. Affordability appears to have declined more quickly for Plan A than for other policies. This is a particularly disturbing trend because this plan may be the only affordable option for individuals with lower incomes. ES-3 Restricting Underwriting and Premium Rating Practices in the Medigap Market: The Experience of Three States I. INTRODUCTION Medicare Supplemental insurance, also known as Medigap, protects individuals with fee-for- service Medicare coverage against out-of-pocket costs not covered by Medicare. Many older persons choose to insure against these costs by purchasing a Medigap policy. Depending on the policy, Medigap can cover costs associated with the Part A deductible ($776 in 2000), Part B deductible ($100 in 2000), coinsurance (generally 20 percent for Part B services,) and other expenses, such as skilled nursing facility care and prescription drugs. OBRA-90 Legislation Standardized Medigap Policies Federal requirements related to the Medigap market, enacted as part of the Omnibus Budget Reconciliation Act of 1990 (OBRA-90), simplified the task of choosing a policy and established important consumer protection features. OBRA-90 established minimum federal standards that apply to all Medigap policies. These standards were then adopted by states2. States can add more stringent regulations as long as they conform to federal standards. OBRA ’90 was designed to encourage competition around price rather than product differentiation. Although subsequent federal laws have added standards on Medicare Select as well as new guarantee issue protections, they are not included in the scope of our discussion. The OBRA-90 legislation authorized the National Association of Insurance Commissioners (NAIC) to develop up to ten standardized Medigap plans. Insurers selling Medigap can offer as many or as few of the standardized plans as they choose, but they must all offer a basic Medigap plan, Plan A. Purchasers may retain policies they held before standardization (pre-standardized policies). Congress chose to require standardized plans to make it easier for consumers to compare plan features and costs. A report examining the impact of this legislation concluded that it has achieved this goal.3 The NAIC developed the ten plans outlined in Table 1. All plans cover the so-called “core benefits.” These benefits include: coverage of all Part A coinsurance for hospital stays longer than 60 days plus coverage for 365 additional lifetime reserve days after Medicare benefits are exhausted; Part B coinsurance which is generally 20 percent of Medicare approved expenses; and the Parts A and B blood deductible. 2 OBRA ’90 included a provision for HCFA to assume responsibility for Medigap regulation if a state failed to adopt the NAIC model regulations. The legislation also grandfathered existing standardized policy structures in Massachusetts, Minnesota, and Wisconsin. 3 McCormack, L.A., Fox, P.D., Rice, T., and Graham, M.L. (1996). The Medigap Reform Legislation of 1990: Have the Objectives Been Met? Health Care Financing Review, 18, 1, 157-174. 1 The standardized plans differ on whether they cover: the deductibles for Medicare Parts A and B; skilled nursing facility coinsurance; medical expenses incurred when traveling to another country; at home recovery; prescription drugs; and preventive care. In addition, Plans F, G, I, and J cover Part B charges in excess of the amount Medicare will allow (balance billing charges.) The need for this benefit has decreased in recent years as the vast majority of doctors have agreed to accept Medicare reimbursement rates, and HCFA has put limitations on the amounts physicians’ can charge above Medicare approved rates. The comprehensiveness of the coverage offered by the different plans generally increases from A through J. For example, Plan A covers only core benefits, Plan F adds coverage of all Medicare-related cost sharing, and Plan J adds all Medicare cost-sharing and additional optional benefits including prescription drugs coverage. TABLE 1 BENEFITS COVERED BY THE TEN STANDARDIZED MEDIGAP PLANS PLAN Benefit A B C D E F G H I J a Core benefits • • • • • • • • • • b SNF coinsurance • • • • • • • • Part A deductible • • • • • • • • • Part B deductible • • • c c c c Part B excess charges High Low High High Foreign travel • • • • • • • • At-home recovery • • • • d d d Prescription drugs Low Low High Preventive medical care • • e % of policies sold 11 14 26 4 1 30 1 6 2 5 a Core benefits include coverage of all Part A (hospital) coinsurance for stays longer than sixty days, the 20 percent Part B coinsurance, and the first three pints of blood. b SNF is skilled nursing facility. c Low 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. d Low prescription drug coverage has a $250 annual deductible, 50 percent coinsurance, and a maximum annual benefit of $1,250; high coverage is similar, but has a $3,000 maximum annual benefit. e From National Bipartisan Commission on the Future of Medicare, “Private Supplement Coverage Summary,” thomas.loc.gov/medical/K-P-1499, HTML, March 19, 1999. Source: NAIC, Medicare Supplement Insurance Minimum Standards Model Act (30 July 1991). OBRA ‘90 also established a six-month open enrollment period when an individual first enrolls in Medicare Parts A and B and is age 65 or older. During this enrollment period, Medigap insurers cannot turn down an applicant or discriminate in setting premiums on the basis of an applicant’s health. Beyond these provisions, OBRA ’90 did not address underwriting or rating practices. 2 Some states have adopted regulations more stringent than those required by OBRA-90 Despite standardization, policies can still differ along two main dimensions: underwriting (screening for health conditions) and rating practices. A 1997 report by The Lewin Group indicated that insurers are increasingly relying on underwriting and rating practices to offer policies at more competitive premiums to younger purchasers.4 The Lewin report found that fewer insurers were selling policies without screening out individuals with health conditions (guarantee issue) and that fewer insurers were charging the same premiums for all policyholders (community rating) or basing premiums on the policyholder’s age at the time of initial purchase (issue age rating). Instead, more insurers appeared to be underwriting policies and basing premiums on the current age of the policyholder (attained age rating). The 1997 Lewin study suggested that these underwriting and rating practices could reduce access to Medigap among older individuals, especially those with health conditions. In addition, because Medigap policies offered to AARP members through Prudential5 (hereafter referred to as AARP Medigap policies) did not vary premium rates by age or health status of the insured, the authors of the report raised the concern that in many markets these policies might become the only community-rated, guarantee issue policies. If this occurred, AARP’s Medigap policies might become the only available or affordable Medigap option in a market for individuals with health conditions. This, in turn, could result in adverse selection for AARP Medigap policies which could cause rates to increase because individuals with health conditions tend to have greater health expenditures. These subsequent increases in premiums for AARP Medigap policies might induce younger and healthier people to switch to less costly alternatives, exacerbating the adverse selection problem and causing rates to spiral even higher. By adopting regulations beyond those required by OBRA ’90, some states have tried to prevent competition based on skimming the healthiest individuals and excluding those with health problems through underwriting and premium-rating practices. According to unpublished data provided by the National Association of Insurance Commissioners, by 1997 nine states restricted either premium rating or underwriting practices or both (see Table 2). This study revealed that two additional states, Arkansas and Idaho, have also implemented regulations that exceed OBRA-90 requirements. 6 4 Alecxih, L., Lutzky, S., Sevak, P., and Claxton, G., “Keys Issues Affecting Accessibility to Medigap Insurance,” prepared for The Commonwealth Fund, August 1997. 5 Prior to 1998, AARP offered Medigap policies through Prudential; since 1998 they have been offered through United HealthCare. 6 Missouri adopted additional Medigap regulations that took effect on January 1, 2000. Missouri’s regulations ban attained age rating and mandate an annual open enrollment period for individuals who have held a Medigap policy for one year. 3 TABLE 2 STATES WITH REGULATIONS EXCEEDING OBRA-90 REQUIREMENTS - 1997 State Mandatory Ban Continuous Annual Open Year Community Attained Open Enrollment Regulations Rating Age Rating Enrollment Went into Effect Arkansas X 1993 California X 1997 Connecticut X X 1993 Florida X 1993 Georgia X 1991 Idaho X 1995 Maine X X 1993 Massachusetts X X 1994 Minnesota X 1993 New York X X 1993 Washington X 1995 All States 7 3 2 3 Notes: Mandatory community rating requires insurers to charge the same premium for all eligible purchasers. Attained age rating involves automatic increases in premiums as the policyholder ages. Continuous open enrollment requires that individuals 65 and older may purchase a policy at any time, regardless of their health conditions. Annual open enrollment allows eligible Medicare beneficiaries to purchase a policy during a certain time period during the year regardless of their health condition. Source: Unpublished National Association of Insurance Commissioners research conducted in May 1997 and authors’ research. Theoretically, these regulations could either help consumers by leveling the playing field among insurers in the market or hurt consumers by inhibiting competition and decreasing affordability of Medigap policies. Insurers have argued that these regulations are harmful to the Medigap market and therefore harmful to consumers. Legislation that expands open enrollment requirements, such as requiring an annual open- enrollment period or mandating that all policies be sold continuously on a guarantee issue basis, provides greater access to individuals with health conditions that would preclude them from passing an underwriting test. If all insurers were required to hold an open enrollment period during which they would have to accept all applicants regardless of their health status, the few Medigap insurers that currently do not underwrite coverage would no longer be the only choice for many “bad risks.” As a result, any source of coverage that is not underwritten, such as United HealthCare’s plans for AARP members, would be less likely to experience a cycle of deteriorating experience and escalating premiums. However, requiring an open-enrollment period 4 could exacerbate adverse selection problems for the entire Medigap market. If individuals can purchase Medigap insurance when they are ill, they may wait until a condition arises before they buy a policy. It could also encourage individuals to rely on a Medicare+Choice plan when they are healthy and only purchase a Medigap policy when they require care and want greater choice of providers and services. However, additional rating and enrollment requirements would improve the number of choices for some individuals while limiting the choices for others. For example, banning attained age policies and converting these policies to some other form of rating would increase premiums for younger individuals, possibly reducing access to coverage by making it unaffordable for some. Legislation that standardizes rating structures, such as requiring issue-age premiums for all policies sold, could facilitate the ability of consumers to base decisions on a consistent pricing approach. However, such legislation could further restrict consumer choice, since the initially lower priced attained-age policies would no longer be available. In addition, insurance companies might choose not to sell a product in a state with additional regulation because of the administrative costs associated with complying with a state’s unique regulations. This study assesses the effect of rating restrictions and open enrollment regulations on the market in three states that have adopted them. The intent of this study was to assess whether states had viable Medigap markets after the implementation of additional regulations. It is important to note that other factors, such as the growth of Medicare managed care, have likely affected the Medigap market during this timeframe. It was beyond the scope of this study to conduct an in-depth analysis of all the various factors that may have affected the market during the study period. II. METHODOLOGY Our methodology involved comparing the Medigap markets in three states that have regulations exceeding the OBRA-90 requirements (study states) with markets in three states without additional regulations (comparison states). We also analyzed the experience of AARP Medigap policies available to AARP members in all states with and without additional regulations. AARP Medigap policies are particularly relevant because they account for a large share of the Medigap market nationally (up to one-quarter), and in many states they represent the only community- rated, guarantee issue policy available. Selection of Study and Comparison States We selected our study states from among those states that have implemented regulations for underwriting or rating practices that exceed the requirements of the OBRA-90 legislation that standardized Medigap insurance (see Table 2). We selected three study states (New York, Connecticut and Florida) based on policy relevance and data availability. New York and Connecticut were deemed to be policy relevant because they represented states with the most restrictive regulations. We selected Florida because it took a more incremental approach than New York and Connecticut, and it is a state with a large older population. 5 Resources available for the project constrained the research design by limiting the number of states we could study to three. Hence, we built on work done in earlier research for the Commonwealth Fund to provide the comparisons. We obtained data from three states without requirements exceeding the OBRA-90 standards (i.e., comparison states) that we had studied previously (Arizona, Ohio, and Virginia). Since the comparison states were originally selected as part of the earlier study, they had been selected to demonstrate the diversity of experience in terms of size, geographic coverage, and Medicare managed care penetration, in addition to data availability rather than to be comparable to the study states used here. We updated information on the comparison states for the current study. Analyses We explored the experience of three states implementing regulations related to open enrollment and rating practices and compared it to that in three states not implementing additional regulations. This methodology relied on secondary analyses of rate books created by the state insurance departments to inform consumers and on state-specific 1996 Medicare supplement insurance experience reports assembled by the National Association of Insurance Commissioners (NAIC). Since some, but not all, states produce Medigap consumer information booklets to help consumers compare rates and rating methods for policies that are currently available, availability of booklets was a secondary factor in selecting states. We supplemented these secondary analyses with interviews with state regulators to determine if companies listed were actively marketing products within the state and to obtain their overall impression of the market. To simplify the analyses, we restricted our review to data for four of the ten standardized plans: A, C, F, and I. We chose Plans C and F because they are the policies most often bought by consumers.7 Plan A was chosen because it is the most basic and least expensive policy, and the OBRA-90 legislation requires all companies offering standardized Medigap policies to offer it. We chose Plan I because we wanted to include a plan that covered prescription drug benefits, and our preliminary review of information from Virginia revealed that more insurers offered Plan I than the other plans with prescription drug coverage. However, we note that national data,8 which became available after the data for this study were collected, indicate that the other two policies that include drug coverage, H and J, attract more purchasers than Plan I. Sources of Data We relied on five sources of data: 7 National Bipartisan Commission on the Future of Medicare, “Private Supplement Coverage Summary,” thomas.loc.gov/medical/K-P-1499, HTML, March 19, 1999. 8 Ibid. 6 1. Consumer rate books - Most states publish consumer rate books that contain information on premiums and rating and underwriting practices for consumers who interested in purchasing a Medigap policy to use as a reference. We obtained books from 1995 to 1998, the most recent years available at the start of the analyses from the state insurance departments for the study and comparison states. We conducted two analyses for New York state because their rate books provide separate rates for different areas of the state. We focused on the New York City and Rochester metropolitan areas. In Arizona, rate information is compiled by a source other than the state insurance department, Arizona Senior World Newspaper. 2. National Association of Insurance Commissioners (NAIC) Medicare Supplemental Insurance Experience reports9 for 1996 for each of the individual states in this study. At the time we conducted the analyses, 1996 was the most recent year for which data were available. 3. Interviews with state regulators – We conducted brief telephone interviews with state regulators responsible for overseeing Medigap insurers. We obtained information about their impression of the effects of changes in the market and additional information on policies not included in the rate books. We also verified whether certain companies had left or entered the market and resolved discrepancies relating to changes in an insurer’s status in the market (e.g., an insurer’s Medigap business being sold to another insurer). 4. The policy premiums for Medigap coverage offered to AARP members and number of policies sold were supplied by AARP. 5. The 1995 and 1998 March Supplement of the Current Population Survey (CPS) for state- specific information about median income. Outcome Measures We selected several outcome measures to assess how viable the market is and how accessible Medigap policies are in states with and without additional regulations. We defined viability and accessibility to include the availability and affordability of policies. We conducted separate analyses of availability for all Medigap policies being marketed and for guarantee issue policies that are available to all consumers regardless of health conditions. The following are the outcome measures used in our analyses: 1) Availability. We included two measures of availability. 9 These reports contain information about premiums paid in, claims paid out, covered lives, and loss ratios (claims paid divided by premiums paid). 7 • The number of companies marketing policies. We conducted two analyses of the number of companies offering a policy in the market. We considered a company to be in the market if it had filed a policy with the insurance department and covered at least 50 lives in the state. We obtained information about whether a policy was filed by referring to the state’s rate book. We supplemented this with information directly from state regulators for insurers listed on the 1996 NAIC experience report. We also obtained the information about the number of covered lives from the NAIC report. We used the 50-lives cutoff to eliminate companies, because our interviews with state regulators suggested that companies not meeting this cutoff were not actively marketing in that state and had likely only filed a policy for individuals who had moved from another state. Our examination of the data for insurers meeting the 50-lives cutoff suggested that many of these insurers also appear to be only peripheral players in any particular market. Combined premiums for all insurers covering fewer than 50 lives accounted for less than one percent of the market in all study and comparison states. While these insurers have filed a policy for sale with the state insurance department, they do not appear to have invested enough resources in the market to attract consumers. Thus, while these policies are technically available, they do not appear to be readily available to consumers. For example, our interviews with state regulators suggested that some of these insurers were only offering policies to members of an association. We addressed this issue by conducting separate analyses for all insurers with 50 or more lives and for insurers with one percent or more of premiums sold in the particular market. • The total number of companies marketing guarantee issue policies. We also examined the number of insurers offering guarantee issue10 policies. We used the same two criteria for “actively marketing” (described above). 2) Affordability. We included two measures of affordability. • Weighted average premium increases for all policies and increases for guarantee issue policies in the study and comparison states. We assessed changes in premiums using information from 1995 and 1998 rate books supplemented with information from the state for companies that were not included in the books. We weighted premiums to reflect market share (in terms of premiums sold) in 1996 using data from the state-specific NAIC Experience reports. 10 Companies that only offered guarantee issue policies during the open enrollment periods required by OBRA-90 or the Balanced Budget Act of 1997 were not classified as guarantee issue. For Plan I, if an insurer’s underwriting only addressed prescription drug use (as the AARP policy does), the plan was classified as being offered on a guarantee issue basis. 8 • Cost of premiums as a percent of median family income. We calculated the 1995 and 1998 weighted average Medigap premium as a percentage of the median family income for older individuals in each of the study and comparison states11. In calculating the average premium, we weighted each company’s premiums by its market share, measured in terms of premiums paid. We derived state-specific estimates of median12 family income using CPS data.13 3) State-specific changes in premiums for AARP Medigap policies in all states. We analyzed changes in premiums for these policies from 1995 to 1998 in all states that had OBRA-90 standardized Medigap plans.14 We weighted premiums in each state by number of policies sold so that averages reflected the distribution of the market. Hypotheses: Indications that additional regulations have adversely affected the market would include: ⇒ large decreases in the number of companies offering Medigap; ⇒ premiums in study states that are rising faster than in the comparison states (both in the analyses of the study and comparison states and the national analyses of AARP Medigap policies); or ⇒ premium prices in study states that are substantially less affordable (i.e., are a significantly higher percent of median income). Limitations Data availability presented several limitations for this study. 11 We only analyzed premiums at age 65 and 80 because of data availability issues. These were the ages that were most commonly presented in the rate books. 12 We used median rather than average family income because income distributions tend to be skewed causing the mean to be higher than the median. Median income may be a better measure of “typical” experience because it is less affected by the small number of individuals with very high incomes. 13 We estimated median family income figures at the state level for all states except New York. In New York, we developed estimates for the Rochester-Buffalo combined metropolitan areas, and also for the New York City vicinity (including northern New Jersey and Long Island). These figures were calculated for those aged 65 to 74 because inadequate sample sizes did not allow for reliable state-level estimates for 65 year-olds only. As a result, national estimates for median personal and family income were produced for those aged 65 to 74 and well as for those age 65. From these figures, an adjustment factor that is directly related to the national difference in income for these two populations was derived. This allowed the initial state-by-state and MSA-level median income estimates for those age 65 to 74 to be increased using the adjustment factor in order to estimate median income for those aged 65. We used a similar procedure for those age 75 and older population to derive state-specific estimates at age 80. It should be noted that we did not divide couples income in two. 14 Massachusetts, Minnesota, and Wisconsin are exempt from OBRA-90 standardization because they had standardized plans in place before the legislation was passed. 9 We were not able to obtain data about the Medigap markets in the states prior to implementation of the regulations. The absence of these data prevented us from conducting a pre-post analysis of the effect of additional regulations. We only had premium information for an insurer’s total block of Medigap business and did not have information for individual plan types. Using premium volume information to weight some of the analyses creates the implicit assumption that the distribution of plan types sold was the same across all insurers. Insurers directly submit to states premium information in the consumer rate books and the total premiums paid in the NAIC experience reports. The extent to which these data are audited differs by state and, therefore, the accuracy may differ by state. The NAIC does not audit the information. This study is not intended as an in-depth analysis of the Medigap market in the selected states. It is only intended to provide an assessment of whether states adopting additional regulations still have a viable Medigap market and whether premiums for the policies offered have remained affordable. III. FINDINGS The results of our analyses comparing study states to comparison states suggests that markets in the study states were similar to those in the comparison states in that: 1. consumers had multiple choices among insurers; 2. premiums generally did not increase at a substantially greater pace; and 3. premiums accounted for a similar percentage of median family income. The markets in the study states and the comparison states differed in that: 1. healthy consumers in the study states had far fewer choices of insurers than in the comparison states; 2. consumers with health conditions had more choices in the study states than in the comparison states. Analyses of premiums for AARP Medigap policies nationwide did not show any clear trend across all policies which supports the hypothesis that the additional regulations had an impact on changes in premiums. We discuss each of the outcome measures below. The data supporting the graphics are presented in the Appendix. 10 Availability While additional regulations reduce the number of insurers offering Medigap coverage, a choice of four or more insurers remain for consumers. Although more companies were actively marketing in states that permit underwriting, residents in New York and Connecticut, which ban underwriting, still had at least five insurers from which to choose Medigap coverage (Figure 1)15. However, in both the study and comparison states the number of available insurers was more limited for Plan I, which covers prescription drugs. The number of insurers offering Plan I was particularly limited in New York and Connecticut. Florida’s regulation, which prohibits attained age rating but allows underwriting, did not appear to have discouraged companies from offering policies. Similar numbers of insurers offered policies in Florida and in the comparison states. On average, there were more companies offering policies in states that do not restrict underwriting practices. However, among insurers with more than one percent of the market, the difference between the study states and the comparison states was more modest. This suggests that a major effect of the additional regulations was to keep peripheral insurers from offering policies. Legislation mandating open enrollment periods appears to have achieved its aim of increasing access for individuals who would not pass an underwriting screen due to their health status. Figure 2 illustrates that in both regions of New York and in Connecticut, with the exception of Plan I, more guarantee issue plans were available than in the comparison states. 15 Data supporting this and other figures appear in the Appendix. The number of companies actively marketing Medigap in 1998 was determined with the help of information in the 1996 NAIC Medigap Experience Reports on the number of covered lives. The 1995 and 1998 premium analysis was weighted using the 1996 NAIC Medigap Experience Reports. 11 Figure 1 Number of Companies Actively Marketing Medigap Plans in Study States and Comparison States for 1998 Premium Underwriting and Premium Rating Rating Regulations Regulation Comparison States PLAN A 50 NY- NYC Connecticut Florida Arizona Ohio Virginia Rochester 40 30 20 10 0 PLAN C 50 40 30 Number of Companies 20 10 0 PLAN F 50 40 30 20 10 0 PLAN I 50 40 30 20 10 0 Any market presence More than one percent of the market Source: The Lewin Group analyses of state consumer rate book data and NAIC Medigap Experience Reports for 1996. (See footnote 15) 12 Figure 2 Number of Companies Offering Guarantee Issue Policies in Study States and Comparison States for 1998 Premium Underwriting and Premium Rating Rating Regulations Regulation Comparison States PLAN A NY- NYC Connecticut Florida Arizona Ohio Virginia 14 Rochester 12 10 8 6 4 2 0 PLAN C 14 12 10 8 Number of Companies 6 4 2 0 PLAN F 14 12 10 8 6 4 2 0 PLAN I 14 12 10 8 6 4 2 0 Any market presence More than one percent of the market Source: The Lewin Group analyses of state consumer rate book data and NAIC Medigap Experience Reports for 1996. (See footnote 15) 13 The contrast is even more striking when looking at companies with more than one percent market share offering guarantee issue policies. Consumers with health conditions that would preclude coverage if underwriting were permitted had at least five insurers from which to choose one or more Medigap plans in Connecticut and New York. In contrast, in both Ohio and Virginia, an AARP Medigap policy would have been these consumers’ only source of Medigap. Affordability Additional regulations do not appear to have caused premiums to rise faster in study states than in comparison states While there is considerable variation in the size of premium increases across states and Plan types, there is no indication that average weighted premiums increased at a faster rate between 1995 and 1998 in study states than in the comparison states. In fact, as Figure 3 indicates, the comparison states experienced a slightly larger average weighted premium increase for Plan A policies than did the study states. During the three-year period, the average weighted premium increase ranged from 57 to 62 percent in Arizona, Ohio, and Virginia, while the premium increase ranged from 40 to 45 percent in the study states. The premium increases for Plans C, F, and I showed no pattern of differences among the study and comparison states. For Plan C premiums increased at the fastest and slowest pace in the two states that had the most stringent regulations (54 percent increase in New York City and 23 percent increase in Connecticut). Plan F premium increases showed similar variation; however, a comparison state, Ohio, showed the largest increase. Premium increases for Plan I were also varied, except Virginia, a comparison state, showed the lowest increases. For three of the four plans studied, premiums for guarantee issue policies do not appear to have increased at a different pace in states that permit underwriting compared to those in the states that prohibit it. However, rates increased faster for Plan A in the states that permit underwriting (see Figure 4 and Appendix Table 2). In the states that permit underwriting, these larger increases in Plan A suggest that Plan A policies that are available on a guarantee issue basis may be purchased more than other plans by individuals with health conditions who have low to moderate incomes in these states. These individuals may be responding to the higher premiums associated with guarantee issue policies by choosing Plan A, which has less comprehensive coverage and lower premiums. However, more research on this phenomenon is needed before drawing any conclusions. While the examination of premium increases did not show large differences that would be characteristic of a rate spiral in states with more stringent regulations, these analyses do not rule out the possibility that the regulations had an effect on premium levels. Since regulations in the study states took effect in 1993, some of the change in premiums could have occurred before 1995, the first year of our analyses. To assess differences in affordability in study and comparison 14 states, we examined average premiums as a percent of median family income in both 1995 and 1998. Average weighted premiums were not a significantly higher percentage of median income in study states than in comparison states In general, the affordability of Medigap premiums measured as a percent of median income declined in both study and comparison states. As can be seen from Figures 5 and 6, there is no clear indication of whether Medigap premiums were substantially less or more affordable in the study states than in the comparison states. Within the study states, policies appear to have required a smaller percent of median income at age 8016 than at age 65. We attribute this to mandatory community rating in New York and Connecticut 17. We conducted additional analyses on the experience of AARP Medigap policies, because they can have a strong direct effect on the market experience where they account for a substantial share of the market. We also looked at them in all the states where the ten standardized plans are sold18. For example, policies purchased through AARP Medigap policies accounted for 26 to 40 percent of Medigap premiums in the study states and 21 to 38 percent of premiums in comparison states. They also have an indirect effect on the market because previous research found that other companies appear to shadow price the AARP price19. 16 We were unable to include affordability data at age 80 for Florida, because the state does not publish data on rates at that age in its consumer rate books. 17 Our measure assesses affordability of a newly purchased policy. The percentage of income among all policyholders would differ because individuals who previously purchased issue age policies would have lower premiums. Unfortunately, data were not available for these estimations. 18 Three states that had standardized their Medigap markets prior to OBRA-90 have different standardized plans. These states were not included in the analysis. 19 Alecxih et al, 1997. 15 Figure 3 Average Weighted Premium Increases from 1995 to 1998 in Study States and Comparison States Premium Underwriting and Premium Rating Rating Regulations Regulation Comparison States PLAN A NY- NYC Connecticut Florida Arizona Ohio Virginia 70% Rochester 60% 50% 40% 30% Average Weighted Premium Increases (1995-1998) 20% 10% 0% PLAN C 70% 60% 50% 40% 30% 20% 10% 0% PLAN F 70% 60% 50% 40% 30% 20% 10% 0% PLAN I 70% 60% 50% 40% 30% 20% 10% 0% Note: Based on all companies offering plans. For Plans other than I, four or more companies are represented. For Plan I, in some states only two companies are represented. Source: The Lewin Group analyses of state consumer rate book data and NAIC Medigap Experience Reports for 1996. (See footnote 15) 16 Figure 4 Average Weighted Premium Increases from 1995 to 1998 for Guarantee Issue Policies in Study States and Comparison States Underwriting Regulations No Underwriting Regulation PLAN A NY- NYC Connecticut Florida Arizona Ohio Virginia 80% Rochester Weighted Premium Increases for GI Policies (1995-1998) 60% 40% 20% 0% PLAN C 80% 60% 40% 20% 0% PLAN F 80% 60% 40% 20% 0% PLAN I 80% 60% 40% 20% 0% Note: Based on the subset of companies offering guaranteed issue plans. As a result, there is often only one company in the comparison states. Source: The Lewin Group analyses of state consumer rate book data for age 65 and NAICMedigap Experience Reports for 1996. (See footnote 15) 17 Figure 5 Average Weighted Medigap Premiums Purchased at Age 65 as Percentage of Median Family Income in Study States and Comparison States for 1995 and 1998 Premium Underwriting and Premium Rating Rating Regulations Regulation Comparison States PLAN A NY- NYC Connecticut Florida Arizona Ohio Virginia 8.0% Rochester 6.0% 4.0% 2.0% 0.0% PLAN C Premium as a Percent of Family Income 8.0% 6.0% 4.0% 2.0% 0.0% PLAN F 8.0% 6.0% 4.0% 2.0% 0.0% PLAN I 8.0% 6.0% 4.0% 2.0% 0.0% 1995 1998 Source: The Lewin Group analyses of state consumer rate book data and NAIC Medigap Experience Reports for 1996. (See Footnote 15) 18 Figure 6 Average Weighted Medigap Premiums Purchase at Age 80 as a Percentage of Median Family Income in Study States and Comparison States for 1995 and 1998 Underwriting and Premium Rating Regulations Comparison States PLAN A NY- NYC Connecticut Arizona Ohio Virginia 12.0% Rochester 10.0% 8.0% 6.0% 4.0% 2.0% 0.0% Premium as a Percent of Family Income PLAN C 12.0% 10.0% 8.0% 6.0% 4.0% 2.0% 0.0% PLAN F 12.0% 10.0% 8.0% 6.0% 4.0% 2.0% 0.0% PLAN I 12.0% 10.0% 8.0% 6.0% 4.0% 2.0% 0.0% 1995 1998 1/ Florida was not include due to lack of data. Source: The Lewin Group analyses of state consumer rate book data and NAIC Medigap Experience Reports for 1996. (See footnote 15) 19 Table 3 and Figure 7 demonstrate that premium growth for policies purchased through AARP did not differ substantially in states with additional regulations and in other states. Notable findings from these analyses of AARP Medigap policies are: ⇒ Premiums for all Medigap policies available through AARP increased substantially over the study period. This is consistent with the increases in weighted average premiums in the Medigap market noted for the study and comparison states. ⇒ Premium increases were especially large for the most basic plan, Plan A. ⇒ On average, premium increases for AARP’s Plan A were lower in states with the most restrictive regulations than in other states with less restrictive regulations. This is consistent with the finding regarding premium increases for guarantee issue policies using data from rate books. The higher increase may be caused by the fact that consumers with health conditions and few resources are likely to enroll in the least expensive plan which is available to them on a guarantee issue basis. This could indicate that adverse selection is a problem that primarily affects Plan A in states that permit underwriting. TABLE 3 PREMIUM GROWTH FOR AARP MEDIGAP POLICIES, 1995-1998 # Weighted average premium growth by plan State regulations States A C F I All states with additional regulations 9 78% 58% 40% 50% States without additional regulations 39 108% 64% 42% 41% States mandating open enrollment & community rating 3 67% 54% 40% 55% Connecticut 111% 51% 36% 56% Maine 154% 82% 61% 70% New York 55% 52% 39% 54% Mandated open enrollment only (California) 1 104% 62% 32% 26% States only limiting premium rating practices 5 75% 59% 43% 45% Arkansas 133% 75% 53% 45% Florida 60% 56% 44% 43% Georgia 137% 65% 44% 51% Idaho 48% 36% 28% 40% Washington 58% 54% 40% 41% Source: The Lewin Group analyses of AARP Monthly Plan Rates. 20 Figure 7 Average Premium Increases Between 1995 and 1998 (Weighted by Number of Policies Sold in Each State) for Policies Offered by AARP Weighted Premium Increases 120% 100% 80% 60% 40% 20% 0% States with Mandating open Mandating open Only banning States without additional enrollment and enrollment only* attained age rating additional regulations community rating regulations Plan A Plan C Plan F Plan I *California only Source: The Lewin Group analyses of experience data provided by AARP GURE 7: AVERAGE PREMIUM INCREASES BETWEEN 1995 AND 1998 (WEIGHTED BY NUMBER OF POLICIES SOLD IN EACH STATES) FOR POLICIES OFFERED BY AARP IV. CONCLUSIONS Each of the study states appears to have maintained a viable Medigap market despite more restrictive regulations. The additional regulations were intended to level the playing field across companies by preventing companies from skimming the healthiest beneficiaries. These regulations have prohibited companies from lowering prices by excluding individuals with health conditions and pricing policies so that premiums are lower at younger ages. State regulators in the study states all concluded that the additional regulations have had their desired effect and have not significantly harmed competition. The clearest findings indicate that while additional regulations resulted in a smaller pool of insurers competing in a market, consumers still had choices among insurers. By regulating underwriting and rating practices, the study states have reduced healthy consumers’ choice of insurers from 10 - 15 to five, but increased less healthy consumers’ choice of insurers from one to five insurers. However, these data also strongly suggest that affordability for Medigap, measured as average premiums as a percent of median family income, appears to have declined substantially between 1995 and 1998 in both study and comparison states. Much of this decline in affordability appears 21 to reflect increases in premiums. This evidence supports the conclusion of previous research20 that found rising Medigap premiums. Affordability appears to have declined more quickly for Plan A than for other policies. This is a particularly disturbing trend because this Plan may be the only affordable option for individuals with lower incomes. Policy Lessons and Considerations The experience of the study states has important lessons for other states. Overall, the most stringent regulations adopted have had the desired effect of leveling the playing field without unduly harming competition. We should note that by achieving the desired objective of improving access for vulnerable individuals, these additional regulations somewhat reduced the number of available insurers for healthier individuals. However, the effect on the number of insurers competing in a given market could be further mitigated if more states adopt additional regulations, because insurers may be more willing to adapt policies that apply to a large number of states in which they do business. Questions for Further Research Restricting underwriting and premium rating practices appears to increase access to Medigap policies for individuals with health conditions, while leaving healthy consumers with multiple affordable choices. Future research may wish to try to differentiate the effects of different combinations of regulations. Analyzing the experience of states with additional regulations not studied in this research may assist in this effort. In addition, this research did not examine the effect of regulations that increase access to Medicare beneficiaries with disabilities under age 65 on the Medigap market. Future research could conduct analyses similar to the ones in this study to explore the effect of these regulations. 20 Alecxih et al, 1997. 22 Appendix A Data Included in Graphs Appendix Table 1 Number of Companies Offering Plans in 1998 and Changes in Premiums 1995-1998 Plan A Plan C Plan F Plan I Rochester, New York Number of Companies Offering Plan in 1998 9 7 8 3 Number with more than 1% of the market 5 4 4 2 Total Average Premium Increase 1995-1998 28.3% 29.7% 24.2% 25.8% Total Weighted Average Increase 1995-1998 41.5% 40.6% 28.5% 43.6% Maximum Increase 1995-1998 44.0% 41.0% 31.5% 43.6% New York City Metropolitan Area Number of Companies Offering Plan in 1998 9 7 7 3 Number with more than 1% of the market 6 4 4 2 Total Average Premium Increase 1995-1998 34.2% 32.1% 30.7% 33.5% Total Weighted Average Increase 1995-1998 45.3% 54.0% 42.1% 58.9% Maximum Increase 1995-1998 59.5% 56.1% 54.7% 59.0% Connecticut Number of Companies Offering Plan in 1998 13 10 12 2 Number with more than 1% of the market 5 4 5 2 Total Average Premium Increase 1995-1998 36.7% 36.3% 21.3% 30.3% Total Weighted Average Increase 1995-1998 41.4% 23.4% 18.1% 27.0% Maximum Increase 1995-1998 114.5% 82.2% 60.6% 35.6% Florida Number of Companies Offering Plan in 1998 36 31 32 14 Number with more than 1% of the market 15 14 13 6 Total Average Premium Increase 1995-1998 21.3% 24.3% 16.8% 11.7% Total Weighted Average Increase 1995-1998 39.9% 41.5% 21.9% 35.9% Maximum Increase 1995-1998 60.0% 55.9% 46.0% 43.6% Arizona Number of Companies Offering Plan in 1998 44 36 39 14 Number with more than 1% of the market 8 8 8 3 Total Average Premium Increase 1995-1998 46.7% 42.8% 29.6% 46.6% Total Weighted Average Increase 1995-1998 61.9% 45.6% 34.4% 46.6% Maximum Increase 1995-1998 93.5% 82.3% 54.8% 46.6% Ohio Number of Companies Offering Plan in 1998 40 35 34 10 Number with more than 1% of the market 13 11 10 2 Total Average Premium Increase 1995-1998 36.5% 34.0% 27.1% 23.7% Total Weighted Average Increase 1995-1998 57.1% 41.6% 44.9% 45.4% Maximum Increase 1995-1998 87.1% 81.6% 82.5% 55.5% Virginia Number of Companies Offering Plan in 1998 22 17 19 4 Number with more than 1% of the market 6 6 6 4 Total Average Premium Increase 1995-1998 29.8% 26.0% 11.1% 22.4% Total Weighted Average Increase 1995-1998 61.4% 39.5% 23.0% 16.3% Maximum Increase 1995-1998 140.0% 79.4% 51.8% 28.5% Source: The Lewin Group analyses of state consumer rate book data for age 65 and NAIC Medigap Experience Reports for 1996. A-1 Appendix Table 2 Weighted Average Premium Increases in Plans Offered on a Guarantee Issue Basis, 1995-1998 A C F I Study States With Additional Underwriting Restrictions NY-Rochester 41% 41% 28% 44% NYC 45% 54% 42% 59% Connecticut 41% 23% 18% 27% Study State Without Additional Underwriting Restrictions Florida 49% 50% 26% 43% Comparison States Without Additional Underwriting Restrictions Arizona 68% 44% 33% 47% Ohio 80% 53% 57% 42% Virginia 65% 40% 24% 16% Source: The Lewin Group analyses of state consumer rate book data for age 65 and NAIC Medigap Experience Reports for 1996. A-2 Appendix Table 3 Weighted Average Premiums, Median Family Income, and Premium as a Percentage of Income in 1995 and 1998 in Study and Comparison States Study States Comparison States NYC NY-Roch. CT FL AZ OH VA Weighted average premiums at age 65 Plan A 1995 $651 $528 $616 $648 $534 $474 $393 1998 $928 $745 $866 $897 $843 $723 $629 Plan C 1995 $973 $964 $1,027 $1,028 $892 $807 $740 1998 $1,491 $1,354 $1,266 $1,484 $1,295 $1,131 $1,031 Plan F 1995 $1,124 $1,110 $1,107 $1,392 $1,020 $868 $853 1998 $1,591 $1,426 $1,302 $1,663 $1,365 $1,255 $1,048 Plan I 1995 $1,480 $1,480 $1,787 $1,460 $1,236 $1,107 $1,104 1998 $2,124 $2,124 $2,304 $1,991 $1,812 $1,608 $1,284 Family income at age 65 1995 $23,292 $23,311 $29,670 $25,502 $24,814 $24,725 $29,440 1998 $32,418 $26,498 $37,657 $27,354 $27,306 $27,893 $30,083 Weighted average premiums at age 65 as percentage of family income Plan A 1995 2.8% 2.3% 2.1% 2.5% 2.2% 1.9% 1.3% 1998 2.9% 2.8% 2.3% 3.3% 3.1% 2.6% 2.1% Plan C 1995 4.2% 4.1% 3.5% 4.0% 3.6% 3.3% 2.5% 1998 4.6% 5.1% 3.4% 5.4% 4.7% 4.1% 3.4% Plan F 1995 4.8% 4.8% 3.7% 5.5% 4.1% 3.5% 2.9% 1998 4.9% 5.4% 3.5% 6.1% 5.0% 4.5% 3.5% Plan I 1995 6.4% 6.3% 6.0% 5.7% 5.0% 4.5% 3.8% 1998 6.6% 8.0% 6.1% 7.3% 6.6% 5.8% 4.3% Weighted average premiums at age 80 Plan A 1995 $651 $528 $616 NA $725 $620 $450 1998 $928 $745 $866 NA $1,140 $941 $724 Plan C 1995 $973 $964 $1,027 NA $1,159 $1,074 $899 1998 $1,491 $1,354 $1,266 NA $1,676 $1,503 $1,248 Plan F 1995 $1,124 $1,110 $1,107 NA $1,311 $1,079 $1,033 1998 $1,591 $1,426 $1,302 NA $1,771 $1,561 $1,268 Plan I 1995 $1,480 $1,480 $1,787 NA $1,236 $1,424 $1,284 1998 $2,124 $2,124 $2,304 NA $1,812 $1,908 $1,500 Family income at age 80 1995 $18,042 $21,693 $22,643 NA $18,923 $18,333 $19,701 1998 $19,608 $23,577 $25,305 NA $20,202 $27,515 $19,884 Weighted average premiums at age 80 as percentage of family income Plan A 1995 3.6% 2.4% 2.7% NA 3.8% 3.4% 2.3% 1998 4.7% 3.2% 3.4% NA 5.6% 3.4% 3.6% Plan C 1995 5.4% 4.4% 4.5% NA 6.1% 5.9% 4.6% 1998 7.6% 5.7% 5.0% NA 8.3% 5.5% 6.3% Plan F 1995 6.2% 5.1% 4.9% NA 6.9% 5.9% 5.2% 1998 8.1% 6.0% 5.1% NA 8.8% 5.7% 6.4% Plan I 1995 8.2% 6.8% 7.9% NA 6.5% 7.8% 6.5% 1998 10.8% 9.0% 9.1% NA 9.0% 6.9% 7.5% Source: The Lewin Group analyses of state rate books, NAIC Medigap Experience Reports, and 1995 and 1998 CPS data. A-3