Responsibility for regulating employer-sponsored health benefit plans

Transcription

1 DataWatch Self-Insured Employer Health Plans: Prevalence, Profile, Provisions, And Premiums by Gregory Acs, Stephen H. Long, M. Susan Marquis, and Pamela Farley Short Abstract: Data from three recent surveys indicate that about 40 percent of workers with employmentbased health insurance are enrolled in plans that their employers self-insure. Despite the considerable differences between federal regulation of these self-insured plans and state regulation of employer plans purchased from an insurance company, we find striking similarities in the populations they serve, the benefits they offer, and their premium costs. Implications for health policy are discussed. Responsibility for regulating employer-sponsored health benefit plans is divided between state and federal governments. Traditionally, state governments had the authority to regulate the insurance business. Since its passage in 1974, however, the Employee Retirement Income Security Act (ERISA) preempts state laws affecting self-insured employer plans. Hence, these self-insured plans are subject only to federal oversight by the Department of Labor. State insurance commissioners continue to regulate insurance companies, thereby affecting employer-sponsored plans that are purchased from insurers. Although the division of responsibility is clear in theory, the states, the Department of Labor, the courts, and Congress have struggled to define it in practice. For example, because some avowedly self-insured employers shift much of their risk onto others through reinsurance and capitation arrangements, the definition of self-insurance is an issue. Whether certain state laws have enough of an effect on employee benefits to trigger the ERISA preemption also has been a question. The state and federal regulatory regimes are quite different. State govern ments impose reserve requirements on health insurers to minimize the risk of insolvency; they impose premium taxes to finance state guaranty funds to pay the claims of insolvent plans and, in some states, to establish high-risk pools to cover the uninsurable; and they mandate certain types of benefits in plans marketed in their state. Some states also regulate rating and Gregory Acs is a senior research associate at The Urban Institute in Washington. Steve Long, Susan Marquis, and Pamela Farley Short are senior economists at RAND, also in Washington. Health Affairs, Volume 15, Numher The People-to-People Health Foundation, Inc.

2 DATAWATCH 267 underwriting practices in the small-group market. In contrast, ERISA requirements on health plans are far more limited. They deal with reporting requirements, the disclosure of information to employees, and the regulatory interpretation of the fiduciary responsibility of plan administrators, The ERISA preemption is very controversial. Proponents argue that it has permitted employers to develop innovative solutions to containing health care costs. 1 They assert that employers have led the way in negotiating with provider networks, developing plan performance measures, and developing new risk-based payment arrangements to provide financial incentives to providers to hold down costs. They believe that employers would not have had the flexibility to adopt these strategies without the ERISA preemption. They are particularly concerned that large employers, which often are the source of innovations, would be handicapped by a proliferation of state-specific requirements, because large employers tend to operate in many different states. Furthermore, ERISA proponents contend that permitting states to include self-insured employers in their rating reforms-thereby forcing them to pool their claims costs with otherswould weaken employers financial incentives to reduce costs. On the other hand, opponents of ERISA view it as an obstacle to enacting comprehensive health care reform at the state level. 2 The law limits many of the health care financing and cost containment initiatives that states have considered and encourages legal challenges to these initiatives. 3 Because self-insured plans do not have to comply with statemandated benefits, ERISA prevents states from legislating a minimum benefit package for all of their residents. Opponents also are concerned that ERISA does not provide adequate consumer protection, ranging from methods of dispute resolution to solvency requirements. Despite these strong views about the implications of the ERISA preemption, there is little information about firms that self-insure and the differences between self-insured and fully insured health plans. What shares of the market are now under federal versus state jurisdiction? What distinguishes firms that self-insure from those that purchase insurance? Have the two regulatory regimes fostered two different systems of employer-sponsored insurance, with different benefit provisions and different premiums? Virtually any reform proposal affecting the private employment-based health insurance market must deal with the issue of continuing or modify ing the current regulatory distinction between self-insured and purchased plans. For example, during the recent national health care reform debate there was considerable discussion of whether or not to include multistate, self-insured employers in proposed health insurance purchasing alliances, on either a mandatory or a voluntary basis. Analysts on both sides of the debate were uncertain about the number of employees in self-insured plans,

3 268 HEALTH AFFAIRS Summer 1996 their distribution by size of firm, and the likely demographic and health profile of alliances formed voluntarily. As the impetus for reform has shifted from the national to the state level, and the scope of legislative proposals from universal to incremental, the unanswered questions remain much the same. This paper addresses these questions. Methods We tabulate data from three recent surveys with information about self-insured plans: the 1991 Health Insurance Association of America (HIAA) survey, the 1993 Robert Wood Johnson Foundation (RWJF) Employer Health Insurance Survey, and the 1987 National Medical Expenditure Survey (NMES). The first two sources are surveys of employers, while the third is a survey of households. The HIAA survey contains the responses of 3,323 firms nationwide that responded to a telephone interview; the survey response rate was 71 percent. 4 The RWJF survey, part of the foundation s State Initiatives in Health Care Reform program, permits state-specific estimates for ten states. 5 Telephone interviews with approximately 2,000 establishments-about threefourths with fifty or fewer employees-were completed in each state, for a total of 22,347 interviews. The response rate was 71 percent. 6 Both employer surveys collected data about the number and characteristics of the persons employed by the business and about the insurance plans offered. The survey respondent was asked to report whether each plan was self-insured or fully insured. 7 We use these data to make estimates of the number and characteristics of employees in businesses that offered a selfinsured plan. We refer to these as self-insured employers, although many of them also offered a fully insured plan. We also use the employer surveys to estimate the share of employees enrolled in self-insured and fully insured plans, and to compare the benefits and premiums for workers in self-insured plans with those in fully insured plans. 8 For information about the characteristics of employees who actually enroll in self-insured or fully insured plans, we use NMES, sponsored by the Agency for Health Care Policy and Research in the U.S. Department of Health and Human Services. Approximately 15,000 households were interviewed about every four months to collect information about their health insurance during NMES contacted employers, unions, and insurance companies that provided insurance for household members to verify enrollment, determine premiums and payment sources, and obtain information about the plan, including whether or not it was self-insured. We base our estimates on 6,489 policyholders in the household and follow-up surveys. 9 Among the households completing all rounds of household interviewing,

4 DATAWATCH 269 the response rate for the follow-up survey was 62 percent. All three surveys provide information about the health insurance sponsored by private employers of all sizes. We exclude self-employed persons with no employees. 10 The NMES data include all public employers; the HIAA survey includes only state and local government employers; and public employers are excluded from the RWJF tabulations reported here.. The RWJF and NMES data include union-sponsored plans, but they are not included in the HIAA data. Results Prevalence of self-insurance. The three surveys indicate that percent of workers who enroll for employer-sponsored health benefits are in a self-insured plan (Exhibit 1). 11 This rate varies greatly with firm size. Fewer than one in six insured employees in smaller firms (100 employees or fewer) are enrolled in a self-insured plan, whereas about 60 percent of insured workers in the largest firms are in a self-insured plan. While the majority of insured workers in large firms are in self-insured plans, even in large firms a substantial share of insured workers are in fully insured plans. Moreover, most workers in self-insured plans are in a plan whose risk has been limited by the employer through a reinsurance arrangement. The concentration of self-insured plans in large firms also is evident from the distribution of enrollees in self-insured plans by size of firm (Exhibit 2). The surveys indicate that percent of employees in self-insured plans are in firms with more than 500 employees. In contrast, almost half of Exhibit 1 Percentage Of Insured Employees In Self-Insured Plans, By Firm Size, Estimates From Three Surveys Insured employees in self-insured plans All firms employees employees More than 500 employees Enrollment in self-insured plans with reinsurance All firms employees employees More than 500 employees RWJF HIAA NMES (1993) (1991) (1987) 41.0% 32.3% 42.2% _a _a _a _a Sources: Authors tabulations of the 1987 National Medical Expenditure Survey (NMES), the Health Insurance Association of America s (HIAA s) 1991 Survey of Employer-Sponsored Healrh Benefit Plans, and The Robert Wood Johnson Foundation s (RWJF s) 1993 Employer Health Insurance Survey. a Not available.

5 270 HEALTH AFFAIRS Summer 1996 Exhibit 2 Percentage Distribution Of Employees In Self-Insured And Fully Insured Plans, Bv Firm Size Employees in self-insured plans All firms employees employees More than 500 employees Employees in fully insured plans All firms employees employees More than 500 employees RWJF HIAA NMES (1993) (1991) (1987) 100.0% 100.0% 100.0% Sources: Authors tabulations of the 1987 National Medical Expenditure Survey (NMES), the Health Insurance Association of America s (HIAA s) 1991 Survey of Employer-Sponsored Health Benefit Plans, and The Robert Wood Johnson Foundation s (RWJF s) 1993 Employer Health Insurance Survey. a Not available. employees in fully insured plans work for firms with 100 or fewer employees; fewer than one-third work for the largest firms. All three data sources suggest that there are geographic differences in the extent of self-insurance. The two national surveys show regional variation. There are more workers in self-insured plans in the Midwest and South than in the Northeast and West (Exhibit 3). However, looking at the state-specific results from the RWJF survey, we find that differences between states within a region are as large as or larger than differences across regions. Others have shown that interstate differences in insurance regulation were a factor leading firms to self-insure in the early and mid-1980s. 12 Choice among plans. Many workers in firms with a self-insured plan ( self-insured businesses in Exhibit 4) can choose among multiple plans offered. Between 50 and 65 percent work for employers with more than one health insurance option. Because self-insurance and choice are both common among large employers, those who work for self-insured employers have more health plan choices than those who work for employers with only fully insured plans. However, when we look at large employers only, we see similar proportions of employees in firms with self-insured plans and with only fully insured plans who are offered multiple options. Forty percent of workers in self-insured firms also are offered a fully insured plan. The most common option is a health maintenance organization (HMO). About 30 percent of workers in firms that offer a self-insured plan have a choice between it and an HMO, and about 9 percent are offered both an HMO and a fully insured indemnity or preferred provider organization (PPO) plan in addition to the self-insured plan.

6 DATAWATCH 271 Exhibit 3 Percentage Of Insured Employees In Self-Insured Plans, By Region And State Northeast New York Vermont Midwest Minnesota North Dakota South Florida Oklahoma RWJF HIAA NMES (1993) (1991) (1987) 28.1% 35.6% 33.7% West Colorado New Mexico Oregon Washineton Sources: Authors tabulations of the 1987 National Medical Expenditure Survey (NMES), the Health Insurance Association of America s (HIAA s) 1991 Survey of Employer-Sponsored Health Benefit Plans, and The Robert Wood Johnson Foundation s (RWJF s) 1993 Employer Health Insurance Survey. Profile of employees. Next, we consider whether there are differences between workers in self-insured businesses or plans and those in fully insured businesses or plans. From the two employer surveys (RWJF and HIAA), we compare characteristics of all workers in businesses offering a self-insured plan with those of workers in fully insured businesses, whether or not the workers enroll in a plan. This comparison provides information about differences in the workforces of employers that make different decisions about self-insuring. From NMES, we compare the workers who enroll in a self-insured plan with workers who enroll in a fully insured plan. This comparison provides information about differences in the populations served by self-insured and fully insured plans. Exhibit 5 shows few differences between employees in businesses that do and do not offer a self-insured plan and between enrollees in self-insured plans and fully insured plans. All three surveys indicate that employees in self-insured businesses or in self-insured plans have higher earnings, but these differences are small. Differences in other demographic characteristics are also slight and do not show a consistent pattern across the surveys. 13 Employees who work for businesses that offer a self-insured plan are more likely than those in other businesses are to be union members; hence, employees in self-insured plans are more likely to be union members. This occurs because large firms are more likely to offer a self-insured plan and more likely to have union workers.

7 272 HEALTH AFFAIRS Summer 1996 Exhibit 4 Percentage Distribution Of Employees In Self-Insured And Fully Insured Businesses, Bv Number And Type Of Plans Offered Number of plans offered All self-insured businesses One Two Three or more All fully insured businesses One Two Three or more Self-insured businesses with more than 500 employees One Two Three or more Fully insured businesses with more than 500 employees One Two Three or more Plan types offered, self-insured businesses Self-insured plan(s) only Self-insured and fully insured HMO only Self-insured and other fully insured plan only Self-insured, fully insured HMO, other fully insured plan RWJF HIAA (1993) (1991) 35.9% 50.4% _a 28.4 _a 4.6 _a 9.2 _a Sources: Authors tabulations of the Health Insurance Association of America s (HIAA s) 1991 Survey of Employer-Sponsored Health Benefit Plans and The Robert Wood Johnson Foundation s (RWJF s) 1993 Employer Health Insurance Survey. Note: HMO is health maintenance organization. a Not available. Provisions and premiums. The benefits available to workers in selfinsured plans are quite similar to those available to workers in fully insured indemnity or PPO plans (Exhibit 6). The employer surveys indicate that the benefits held by those in self-insured plans are somewhat broader and deeper than those held by employees in fully insured indemnity and PPO plans. 14 However, the NMES estimates suggest that these differences are small when translated into actuarial value, expressed as the average share of health care spending that a plan covers. Workers in fully insured HMO plans have more generous coverage than workers in self-insured plans or other fully insured plans have, when judged by this actuarial value measure. All of our sources also show that the average monthly premiums for workers in self-insured and fully insured indemnity or PPO plans are similar (Exhibit 7). The difference between self-insured and fully insured premiums for single coverage ranges between 3 percent and 13 percent ($5-$20

8 DATAWATCH 273 Exhibit 5 Characteristics Of Employees In Self-Insured Businesses And Plans Characteristic Average earnings All employees (payroll per worker) RWJF Insured employees (annual wage rate) NMES a Percent earning less than $14,000 per year All employees RWJF HIAA (less than $10,000) Insured employees NMES a Self-insured Fully insured All insured business business businesses or plan or plan or plans $26,117 $25,020 $25,522 28,200 26,000 26, % 28.4% 26.7% Percent under age 25 All employees RWJF HIAA (under age 30) Insured employees NMES Percent over age 55 All employees RWJF HIAA (over age 50) Insured employees NMES Percent male Ail employees RWJF Insured employees NMES Percent union members All employees RWJF Insured employees NMES Sources: Authors tabulations of the 1987 National Medical Expenditure Survey (NMES), the Health Insurance Association of America s (HIAA s) 1991 Survey of Employer-Sponsored Health Benefit Plans, and The Robert Wood Johnson Foundation s (RWJF s) 1993 Employer Health Insurance Survey. a Earnings from the 1987 NMES have been inflated to 1993 by the Consumer Price Index. per month) across the surveys; the range for family coverage is 1 percent to 11 percent. However, self-insured premiums are not consistently higher or consistently lower than fully insured premiums when we compare across data sets, nor when we compare differences between single and family premiums within a data set. This similarity in premiums might be expected in light of the similar benefits and enrollee characteristics in self-insured

9 274 HEALTH AFFAIRS Summer 1996 Exhibit 6 Benefits In Self-Insured And Fully Insured Plans Average annual deductible Percent with out-of-pocket maximum Amount of out-of-pocket maximum Percent with coverage for Prenatal care (RWJF) Maternity care (RWJF) Outpatient drug Outpatient mental Alcohol abuse treatment Dental (RWJF) Percent with limits on preexisting conditions Self-insured Fully insured Fully indemnity/ PPO insured HMO $205 $253 $ _a 89.9% 85.2% 30.6% _a $1,808 $1,873 $1,502 1,363 1,506 _a 97.5% 92.7% 98.8% _a Percent with medical underwriting (RWJF) Actuarial value of plan (NMES) Averaee percent of expenses covered Sources: Authors tabulations of the 1987 National Medical Expenditure Survey (NMES), the Health Insurance Association of America s (HIAA s) 1991 Survey of Employer-Sponsored Health Benefit Plans, and The Robert Wood Johnson Foundation s (RWJF s) 1993 Employer Health Insurance Survey. Notes: PPO is preferred provider organization; HMO is health maintenance organization. a Not available and fully insured plans. Nevertheless, we might expect premiums for workers in self-insured plans to differ from those in fully insured plans because the self-insured primarily work for large employers, and administrative costs decrease in proportion to benefits as the size of the group rises. 15 However, we found similar premiums for workers in self-insured and fully insured plans within size groups, as well as overall. Although total premiums are similar for all types of plans, employees in self-insured plans pay a somewhat smaller share of the total, especially for family coverage, than employees in other plans pay. This finding also holds

10 DATAWATCH 275 Exhibit 7 Premiums In Self-Insured And Fully Insured Plans Average monthly premiums Single coverage a NMES (1987) a Family coverage a NMES (1987) a Average employee contribution rates Single coverage NMES (1987) Family coverage NMES (1987) Self-insured Fully insured Fully indemnity/ PPO insured HMO $156 $151 $ % 17% 18% Sources: Authors tabulations of the 1987 National Medical Expenditure Survey (NMES), the Health Insurance Association of America s (HIAA s) 1991 Survey of Employer-Sponsored Health Benefit Plans, and The Robert Wood Johnson Foundation s (RWJF s) 1993 Employer Health Insurance Survey. Note: PPO is preferred provider organization; HMO is health maintenance organization. a Premiums inflated to 1993 dollars using the medical component of the Consumer Price Index. within classes defined by firm size. Discussion Nationwide, about 60 percent of employees are covered under health insurance plans that are fully insured and fall under state regulation; the other 40 percent are covered under self-insured plans that are subject to federal oversight. There is considerable variation in this proportion among states, with federal regulation governing more than half of the market in some states. There is also variation by firm size. About 85 percent of insured workers in firms with fewer than 100 employees are in plans under state jurisdiction, whereas states regulate plans covering only about 40 percent of insured workers in firms with more than 500 employees. Despite the marked contrast between the state and federal regulatory regimes, we find a great deal of similarity between self-insured and fully insured plans. They serve similar populations, offer similar benefits, and have similar premium costs. Several factors may account for these similarities. The demographic

11 276 HEALTH AFFAIRS Summer 1996 similarities between enrollees may reflect the insurance industry s reliance on experience rating for large employers that buy traditional insurance. With experience rating, a firm s premium is based on the health risks of its own employees, even if it is fully insured. It does not have to self-insure to avoid pooling claims with firms whose workers have less favorable demographics that might raise premiums. The similarities in benefit design may suggest that employers choices with respect to plan design are constrained by the preferences of employees, regardless of whether or not the plan is self-insured. Our analysis shows that employees in fully insured and self-insured plans are similar in many characteristics that are related to the demand for insurance; hence, one would expect the plan benefits to be similar as well. The broad coverage standards set by the HMO Act of 1973 for federally qualified HMOs may serve to constrain the range of options as well, if consumers or employers view them as a federal statement of the minimum necessary coverage. The similarities between the different types of plans may also indicate that fully insured and self-insured plans continue to compete with each other to a significant extent, despite playing under different rules. As noted above, a sizable share of policies in large firms-which are in the best position to self-insure-remain fully insured. Some critics of the dual regulatory regime express concern about the lack of solvency standards for plans under federal regulation. Our results suggest that workers in self-insured plans may not be at greater risk than workers in fully insured plans, at least with respect to financing the uncommon event of one or more extremely high cost cases. Most self-insured plans include a reinsurance mechanism. In addition, most self-insured plans are sponsored by large employers and therefore enjoy the benefit of pooling risks over many covered lives. Our data, however, do not permit us to address a second dimension of solvency-the adequacy of reserves. The similarities among the self-insured and fully insured plans have a number of implications for state and national health policy. First, the similarity in benefits indicates that past state policies related to mandated benefits have not driven firms to self-insure in order to escape the cost of providing these benefits. Similarly, Gail Jensen and colleagues found that although state mandates may have been a factor leading to the decision to self-insure in the early 1980s they have not been a major factor encouraging self-insurance in more recent years.16 Moreover, the trend to selffunding among large firms may have leveled off in recent years as these employers have turned to managed care plans-still primarily purchasedto contain costs. Second, there are few differences in premiums and demographic characteristics, suggesting that self-insured plans do not enroll a more favorable

12 DATAWATCH 277 selection of health risks than fully insured plans enroll. Thus, premiums in state or national health alliances and purchasing cooperatives are unlikely to be affected by decisions to include or exclude self-insured plans. The similarities between the different types of plans also may make it easier to contemplate unified reforms that would apply to both types of plans. There do not appear to be major differences between the two types of plans that would justify special treatment for one or the other. Finally, our findings suggest that ERISA may not present significant obstacles to effective implementation of many of the current state initiatives that focus on insurance rating, underwriting, and purchasing reforms for small businesses. Although ERISA has led to a division in the employersponsored insurance market between plans governed by state regulation and those operated under federal rules, our data show that this division is largely related to firm size. Most policies offered by small businesses are fully insured and, therefore, are under the purview of state regulators. Although this suggests that state efforts to reform the small-group market are not limited at present, the picture could change if small businesses claim to self-fund by buying low-deductible reinsurance. This research was supported by separate awards from the Pension and Welfare Benefits Administration of the U.S. Department of Labor and The Robert Wood Johnson Foundation. Any views expressed herein do not necessarily represent the views of the above sponsors, RAND, or The Urban Institute. The authors thank Daniel Dowhan, Roald Euller, and Afshin Rastegar for computer programming support. NOTES 1. G.L. Atkins and K. Bass, ERISA Preemption, The Key to Market Innovation in Health Care (Washington: Corporate Health Care Coalition, 1995). 2. P.A. Butler, Roadblock to Reform: ERISA Implications for State Health Care Initiatives (Washington: National Governors Association, 1994). 3. However, a recent U.S. Supreme Court decision held that ERISA does not preempt state laws that only indirectly impose costs on self-insured plans, thus opening the door for some financing schemes that states have been hesitant to pursue. See P.A. Butler, Recent ERISA Developments: Implications of the Travelers Case fur State Health Policy Initiatives, Issue Brief (Washington: National Governors Association, 1995). 4. C.B. Sullivan et al., Employer-Sponsored Health Insurance in 1991, Health Affairs (Winter 1992): Tabulations based on the HIAA survey use employee-level weights and reflect only valid responses to survey questions. That is, we assumed that any nonresponse to specific survey items is random, and we calculated ratios and averages using only nonmissing information. 5. The states are Colorado, Florida, Minnesota, New Mexico, New York, North Dakota, Oklahoma, Oregon, Vermont, and Washington. 6. J.C. Cantor, S.H. Long, and M.S. Marquis, Private Employment-Based Health Insurance in Ten States, Health Affairs (Summer 1995): Response rates varied by establishment size and state; however, the results reported here are weighted to adjust

13 278 HEALTH AFFAIRS Summer 1996 for this differential nonresponse. The unit interviewed was the establishment; however, here we classify employers and workers by the size of the firm, not the establishment, for comparison with the HIAA and NMES results. 7. The respondent s report was used to classify plans as self-insured or fully insured. Reinsurance and its extent was not taken into account. For example, a reinsurance policy that pays when claims exceed $1,000 may be called self-insured with reinsurance by a business that seeks exemption from state regulation under ERISA, while state regulators would view the reinsurance policy as primary coverage. 8. The HIAA survey limited questions about plan characteristics to the largest plan offered among four different types (conventional, HMO, PPO, and point-of-service). Thus, if the self-insured plan was not the largest of its type, we might undercount the number of employees in self-insured firms. On the other hand, we may overstate enrollment in other firms. We attribute all of the enrollment for a particular type of plan to the largest plan of that type. Thus, enrollment in a self-insured plan would be overstated if the employer offered a smaller but fully insured plan of the same type. 9. C. Emmons and C. Hill, Questionnaires and Data Collection Methods for the Health insurance Plans Survey, NMES Methods 5, AHCPR Pub. no (Rockville, Md.: Agency for Health Care Policy and Research, 1994). 10. NMES estimates show this population at 5.4 million. Only 11 percent hold an employment-based policy, so their omission will have little effect on our estimates. See A.C. Monheit and P.H. Harvey, Sources of Health Insurance for the Self-Employed: Does Differential Taxation Make a Difference? Inquiry (Fall 1993): Dependents and retirees account for a somewhat higher proportion of the total enrollment in self-insured plans than in fully insured plans, based on NMES tabulations not shown here. Thus, self-insured plans encompass a larger share of the covered lives in employer-sponsored plans (47 percent) than their share of active workers. 12. For example, see G.A. Jensen, K.D. Cotter, and M.A. Morrisey, State Insurance Regulation and Employers Decisions to Self-Insure, Journal of Risk and Insurance (June 1995): The similarity of demographic characteristics extends to all enrollees-active workers, dependents, and former workers-based on NMES tabulations not shown here. 14. Deductibles and out-of-pocket maximums for employees in self-insured plans are lower than those for workers in fully insured indemnity and PPO plans. Those in self-insured plans are more likely to have coverage for special services such as prescription drugs and dental care. Our survey data do not reveal differences in administrative practices that deny coverage under special circumstances. In cases of dispute, consumers have recourse to state insurance regulatory processes when covered by a fully insured plan, but there is no corresponding relief under self-insured plans. 15. Premiums for self-insured plans are difficult to measure because there is not an observed transaction between the employer and the insurer. Surveys use different approaches to measure these premiums. The RWJF survey asked respondents to report a premium equivalent. The premium equivalent includes administrative costs. The administrative cost is imputed for self-insured plans that report a premium equivalent that reflects only claims payment. The HIAA premiums for self-insured plans are based on the subset of plans for which employers regularly put aside monthly allotments. Administrative costs are included if they are part of this allotment. The NMES premiums were imputed from the total funding (including administrative costs) of self-insured plans that provided this information. 16. Jensen et al., State Insurance Regulation.