June 17, 2004 | Backgrounder on Health Care
Families lacking health insurance is a persistent problem in the United States. According to projections based on a sampling for the Kaiser Family Foundation, approximately 43 million non-elderly Americans were uninsured at any point during 2002.1 According to the Congressional Budget Office, based on survey figures for 1998, between 21 million and 31 million people lacked insurance for the entire year, while nearly 60 million were uninsured at some point during the year.2
For some of these people, a short spell without insurance poses no real hardship. Some are "voluntarily" uninsured, in that they consciously decide to forgo insurance that they can afford and take the financial risk. Many of these individuals pay directly for routine care and/or use the emergency room. But millions of others desire insurance, yet cannot afford it or otherwise obtain adequate coverage. For these Americans, a major illness or accident could mean financial ruin or going without necessary care.
According to the Kaiser survey, about two-thirds of the non-elderly uninsured are from low-income families (less than 200 percent of the poverty level, or approximately $29,000 for a family of three). Moreover, about 80 percent (including children) come from working families, and 70 percent have a family member working full-time.3
While most uninsured people are in working families, they are not spread evenly across the workplace. Instead, they are heavily concentrated in the small-business sector. The Kaiser survey4 indicates that:
The concentration of uninsurance in small-business and lower-income households helps to explain the high level of uninsurance among non-managers in such occupations as agriculture (42.7 percent of non-managers uninsured), construction (37.8 percent), and services (34.6 percent), where small firms and lower-income households are disproportionately represented.
There are certainly weaknesses in using the large-business sector to provide insurance, but it does function as a workable system; however, for close to a majority of workers in small firms, the system of health insurance in the small-business sector is practically dysfunctional.
To address the inherent weakness of employer-sponsored coverage in the small-business sector, policymakers should not try to force or induce small employers to act like large-firm sponsors of insurance. That will never be effective. Instead, they should empower employees of small firms to make the same choices as employees of large firms while enabling small employers to facilitate those choices. Specifically, Congress should:
Surveys indicate that working Americans generally prefer employer-based health coverage to other ways of acquiring health insurance, and many experts maintain that employment-based coverage has many advantages.8 However, most of the generic advantages of employment-based insurance apply far less, or not at all, to the self-employed and to workers in small firms.9
But there are two snags with this form of tax subsidy:
It favors high-income households over low-income households. For an insured family with an annual income over $100,000, the average value of the tax benefit in 2004 is estimated by Lewin Group analysts at $2,780. For lower-income but insured families, the tax benefit is a small fraction of that amount because their marginal tax rate is lower. Families with household incomes of from $20,000-$30,000 receive a tax benefit averaging just $725.11
If the employer does not offer insurance (or affordable insurance) for a particular worker, the family typically does not receive a tax break or other subsidy to help purchase insurance. If the employer does not offer insurance, or if the worker cannot afford to enroll in the available plan, there is of course no tax subsidy. But if such an uninsured person considers buying coverage for himself and his family, he normally receives no tax benefits at all. The whole cost is in after-tax income.
A small firm is far less likely to offer insurance, which means employees of such a firm would receive no tax break, and lower-income workers are more commonly employed in smaller firms. According to a recent survey by the Kaiser Family Foundation, while 98 percent of firms with at least 200 employees offered insurance in 2003, only 65 percent of firms with fewer than 200 offered insurance. Not surprisingly, in 2001, some 64 percent of uninsured workers were not even offered insurance through their own job.12 According to a survey of firms in 1999, only 55 percent of firms with fewer than 10 employees offered insurance.13
Similarly, when an employer provides health coverage, an employee can easily participate in the plan, assuming the worker can afford it. Premiums are paid directly by the employer, and the worker does not even have to apply for a tax exclusion. The W-2 form, which indicates the worker's income for tax purposes, simply makes no mention of the employer's contribution to the worker's health insurance. Moreover, if the worker has to pay something toward the cost of the plan, this is usually done through a convenient payroll deduction during each pay period.
This "automatic" way of obtaining health insurance works well for larger firms; but because smaller employers are far less likely to offer insurance, they are also far less likely to set up a payroll deduction system for employees who wish to arrange their own coverage.
Moreover, if a new employee poses particularly high--or low--insurance risk (and therefore incurs particularly higher or lower medical expenses), this will not significantly change the group's expected total cost, and an insurer can offer a group premium that will not change drastically over time (other than tracking the general growth rate for medical expenditures), despite possibly wide variations in medical risks among employees. Large companies also have the economies of scale and sophistication to provide insurance at a low administrative cost per employee.
Small firms, however, are by definition small insurance pools. A retail store with a handful of employees is a dismal pool for insurance purposes. Hiring a new employee with a disability, or the diagnosis of a chronic heart problem in an older worker, can dramatically change insurance costs for the employer from one year to the next. States and the federal government recognize this and are exploring various ways to group small firms together to form larger insurance pools. But the need for these efforts only underscores the fact that small firms are a poor basis for pooling employees' insurance risks.
Three criteria for risk pools. It is also important to recognize that the size of a risk pool is only one of three important criteria for a good insurance risk pool. The other two are randomness and stability, which also present problems for small employers and even groups of small employers. In other words, the group must be in line with the health risk associated with a random cross section of the population from which the employees are typically drawn, and the group's composition must not change frequently. Unfortunately, the employee turnover rate in small business is relatively high, as is the tendency of firm-owners to withdraw from multi-employer groups if they can obtain less expensive coverage somewhere else.
Again, this advantage does not exist with small firms. Small firms face relatively high administrative costs, and many small-business owners consequently do not see it as efficient to organize insurance. Precisely because they lack the economies of scale and the management resources of larger firms, small businesses tend to face high costs when administering plans. According to data collected by the Congressional Budget Office, overhead costs for providing insurance can be over 30 percent of premium costs for firms with fewer than 10 employees, compared with about 12 percent for firms with more than 500 employees.14
In addition to simple economies of scale, other things such as higher staff turnover contribute to this difference. Moreover, many small-business owners have little desire to engage in the demanding task of organizing health insurance to meet the often-varied needs of their employees.
A recent study by Jon Gabel and Jeremy Pickreign for the Commonwealth Fund underscores this disadvantage. The study used 2002-2003 survey data from the Kaiser Family Foundation and other sources and found that, although premiums charged to firms were comparable between firms of different sizes, the premiums bought fewer benefits for the workers and their families in small firms.16
For example, only 38 percent of workers in firms with fewer than 25 employees were offered dental benefits, compared with 87 percent in firms of 200 or more. Meanwhile, 100 percent of employees in the large firms had access to prenatal care benefits, compared with 93 percent in the smaller firms. Moreover, employees of the small firms faced far higher deductibles for single or family coverage.
With such a heavy concentration of the uninsured employed and their dependents in the small-business sector, it makes sense to focus efforts on addressing the obstacles facing families in that sector.
As noted earlier, many lower-paid employees in small firms face a subsidy double-whammy. Those who are offered insurance are paid less and thus get a much smaller tax benefit than upper-income employees through the exclusion from taxable income of employer-sponsored health benefits. Many have no employer-sponsored insurance at all, and if they purchase their own insurance, they typically receive no tax break.
A sensible reform would be to provide similar tax breaks or other assistance to families whether or not they obtained their insurance through the workplace. Rather than a tax exclusion or a tax decision, which gives the most help to those with the highest income, a more efficient and fairer approach would concentrate more help on lower-paid Americans, perhaps through a tax credit.
A sensible reform would be to permit workers in small firms to use any tax break or other subsidy available to them to purchase a plan of their own choice, not just the one (if any) selected by their employer. This reform would allow workers in small firms to obtain insurance through large pools or organizations equivalent in size and sophistication to large employers. It would also mean that employees could retain their chosen coverage if they changed employers.
Goal #3: While workers would continue to sign up for coverage in the workplace and obtain tax subsidies through the workplace, employers should not have to sponsor health insurance for workers in order to be eligible for tax subsidies.
Changing the nature of today's tax subsidy and widening the choice of insurance plans for workers in small firms means rethinking the role of small employers in the provision of health insurance. Large firms typically both sponsor insurance (i.e., select the plans or self-insure) and facilitate insurance (i.e., arrange for employees to sign up and pay for insurance).
Given the obstacles that make it very uneconomic for small firms to offer coverage, divorcing the sponsor and facilitator roles--and leaving smaller firms with only the facilitator role--would reduce the burden and risk for many small employers and make it more likely that they would help their employees to select and sign up for coverage.
The inherent weaknesses of small firms as sponsors of health insurance require policymakers to think differently about the role of small employers. Thinking of them as just small versions of large firms overlooks the different nature of the small-business workplace.
Instead, policymakers need to construct a health insurance infrastructure for workers in the small-business sector that achieves--or exceeds--the insurance advantages of large firms by altering the role of the small employer and changing how benefits are subsidized. (For an overview of the proposed changes, see Figure 1 and Figure 2 in the Appendix.) Congress needs to take three steps to do this:
Unlike a tax deduction or tax exclusion, which favors upper-income workers, a tax credit provides either the same level of assistance to each recipient or even more help for lower-paid individuals. It can be designed in various ways. A credit can be in the form of a fixed dollar credit; a percentage of the premium and/or out-of-pocket, perhaps with a maximum credit amount; or a combination--a base fixed amount plus a percentage of the premium. Each has different effects and financial consequences.17
Making a credit refundable means that if the available credit exceeds the tax liability of an individual or family, the government would remit the difference. Hence, a refundable credit is in effect a health insurance voucher available through the tax system.
A refundable credit could be limited to workers who are not offered a plan by their employer. A criticism of this approach is that it might induce some small employers now offering insurance to end their plans in favor of allowing their employees to qualify for a credit. While, in most cases, this would actually make the employee better off, it remains a widely held criticism. On the other hand, giving the same tax credit to all workers in small firms--whether or not they are offered employer-sponsored coverage--means that workers with employer-sponsored coverage would enjoy "double-dip" tax relief in the form of the tax credit and the exclusion for their employer-directed compensation.
To avoid either situation, a "full" tax credit could be given to workers who are without sponsored insurance and a smaller tax credit to those who have a sponsored plan; the amount of the latter credit, when combined with the tax value of the exclusion, would be designed to be approximately equal to the full credit. In this way, the tax credit would not be biased either for or against an employer-sponsored plan.
Credit Through the Withholding System
The simplest way to deliver the subsidy to workers would be through an adjustment in tax withholdings, much as deductions (such as mortgage interest) or credits (such as the child care credit) are typically handled today with the employer remitting tax payments to the government that are net of the credits. This means that workers would receive the tax benefit in increments throughout the year when they receive their paychecks.
Employers could also institute a system of payroll deductions for health premiums, perhaps through the existing rules for flexible benefit plans, so that the money would be available when premiums were due. Employers could pay premiums directly from these accounts on behalf of employees. In this way, the credit-premium transaction would be relatively simple for both employer and employee.
Assigning the Credit to a Health Plan
Another option would be to permit families to assign the value of their credit to their insurance plan in return for a lower premium. With the assignment, the employee signs a document allowing the insurer to claim the credit on his behalf and the insurer agrees to reduce premiums by the same amount. Insurers would normally obtain the credit through an adjustment in their tax payments to the government. Thus, rather than deal with the withholding system, a family would only have to establish its eligibility for a fixed or simple percentage credit.
This alternative would be particularly attractive to those lower-income families that do not even file tax returns and would address the concern that the tax subsidy might not be available when premiums are due. The process would mirror the premium payment system in the Federal Employees Health Benefits Program (FEHBP), under which Members of Congress and other federal employees are quoted premiums net of the government contribution.
Coverage Through Automatic Enrollment
Whether or not they sponsored insurance, employers could institute an automatic enrollment and payment system to make health insurance premium payments and obtain health-related tax benefits. This means that employees would automatically be enrolled in a health plan unless they explicitly declined to enroll, perhaps by signing a document indicating that they understood the possible consequences of not enrolling in a plan. Alternatively, a state could establish a default bare-bones health plan in conjunction with a private insurer, to which anyone not otherwise choosing a plan would be assigned.
Evidence from pension plans indicates that an automatic enrollment system for health insurance could sharply increase sign-up rates.18
Proposals in Congress
Several recently introduced legislative proposals are based on the health care tax credit concept. With some variation, the proposals focus primarily on providing tax credits to lower-income individuals and families without coverage.
These proposals have also garnered bipartisan and even tripartisan support. For example, Senator Rick Santorum (R-PA) and Representatives Mark Kennedy (R-WI) and William Lipinski (D-IL) introduced similar legislation in the Fair Care for the Uninsured Act (S. 1570 and H.R. 583). Representatives Kay Granger (R-TX) and Albert Wynn (D-MD) introduced the Securing Access, Value, and Equality (SAVE) in Health Care Act (H.R. 1236). In 2001, Senator James Jeffords (I-VT) introduced S. 590, the Relief, Equity, Access, and Coverage for Health (REACH) Act, with the support of both Democrats and Republicans--including Senators Bill Frist (R-TN) and John Breaux (D-LA).
Some proposals have integrated tax credits with other health care initiatives. Both President George W. Bush and Democratic presidential candidate Senator John Kerry (D-MA) have integrated tax credits into their overall health care proposals. Representative John Shadegg (R-AZ) has introduced the Small Business Access and Choice for Entrepreneurs Act (H.R. 3423), and Senator Jeff Bingaman (D-NM) and Representative Marcy Kaptur (D-OH) have introduced the Health Coverage, Affordability, Responsibility, and Equity Act (S. 1030 and H.R. 2402). All three combine the tax credit approach with an overall health reform proposal.
Step #2: Create alternative pools for the employees of small firms--including plans offered through churches, unions, and other intermediaries, as well as through the FEHBP--so that these workers and their families can access a wide range of affordable plans.
Providing a health insurance subsidy to employees who lack adequate help today is only one part of the solution to a lack of coverage. Affordable coverage that can be purchased with the help of a credit is the other part. For younger and healthier individuals and families, the individual insurance market offers affordable policies; but for many with poor health, obtaining affordable private coverage is difficult or impossible.
A solution to this problem is to construct forms of group insurance, in essence mimicking the large pools of employees available to the biggest employers. This would spread high risks across the pool so that sicker individuals and families would not face unaffordable premiums.
Stability of Groups
Creating such pools, however, poses a number of challenges that require careful design decisions. (Bringing several small employers together as a group poses similar challenges.) A major worry is the stability of voluntary insurance groups. The danger in bringing individuals together and establishing a group insurance premium--in effect, an average premium--is that healthier individuals would have the incentive to leave the group to get cheaper individual coverage reflecting their low risk. Meanwhile, sicker individuals would wish to join the group to get relatively inexpensive coverage. The group could then face a "death spiral" of ever-higher group rates.
Certain steps can reduce this problem to a degree. For example, some combination of higher rates, waiting periods, and pre-existing condition exclusions could be imposed on those seeking to join such a group who did not have prior coverage. Thus, individuals would be rewarded for buying and maintaining coverage when they are healthy and would be penalized if they sought coverage only when they needed medical care. Long-term contacts with penalties for dropping coverage could also make the group more stable.
Pool with a Risk Adjuster
Although steps can be taken to improve the stability of pools, for long-term success and for equity reasons, an effective risk adjustment mechanism needs to be incorporated into group coverage for families in the small-business sector. A risk adjuster can take different forms, but the basic idea is to ensure that there are appropriate cross-subsidies between high-risk and low-risk enrollees within the pool, regardless of which insurance plan individuals choose.
An example would be a mandatory reinsurance pool in a state or other area, in which all insurers would pay a percentage of their premiums into a reinsurance pool and the member insurers would receive payments from the pool according to whether they had an above-average or below-average share of high-cost enrollees relative to the other carriers in that market. In this way, an insurer attracting a disproportionate share of high risks (perhaps because of good coverage for cancer) would be subsidized through the reinsurance pool by an insurer that attracted a disproportionate share of low risks (perhaps by offering a leaner policy with a lower premium).
Of course, if all carriers in the market were attracting about the same share of high-risk enrollees, then little--if any--cross-subsidy would occur through the reinsurance pool, since no single carrier or group of carriers was being disadvantaged in the market.
Intermediaries in the Insurance System
Large firms are more effective than small firms in offering insurance not only because they can assemble large pools, but also because they are sophisticated negotiators and buyers of insurance. Insurance groups based on large organizations could achieve many of the marketing and administrative economies of scale that are normally available only to the employees of large firms. Typically, such organizations would not get into the business of insurance themselves, but would act much as a buyers club does by negotiating an arrangement with existing insurance companies. Organizations that might function in this way include groups of churches, trade unions, the American Association of Retired Persons (AARP), professional and trade associations, farm bureaus, and credit unions.
Some organizations (e.g., many farm bureaus) already offer plans, but working families who join these plans typically are not eligible for tax relief. Tax credits for health insurance would change that by making it more economical to offer insurance because far more potential enrollees would be able to afford premiums. For years, many African-American church congregations have organized various forms of insurance and other services for their members. Moreover, in many inner-city communities, these churches are typically larger, more stable, and more sophisticated--as well as more trusted--than the typical employer, making them a natural avenue through which many families armed with tax credits could obtain their health insurance.
The FEHBP System
as a Possible Model
The Federal Employees Health Benefits Program could be another model of an alternative intermediary for workers in small firms. The FEHBP is an example of an insurance arrangement that offers group rates for individuals and also incorporates plans offered through voluntary associations, primarily employee organizations and unions. An FEHBP "look-alike"--organized by states and that has a separate risk pool--could be one way to provide an insurance infrastructure.
The FEHBP provides federal workers and their dependents (nearly 10 million covered individuals) with a wide choice of plans.19 There have been many proposals in recent years to open it up to non-federal workers under various conditions, typically using a separate insurance pool. To make the FEHBP available to non-federal workers using tax credits, Congress would need to amend federal law governing the FEHBP to permit a separate insurance pool for non-federal employees (so that premiums for federal employees would not be affected), with the exact structure in each state negotiated between the state and the federal government. Plans currently available in the FEHBP might be allowed to market to the new state pool if they wished, and other plans could market exclusively to the new pool provided they met the general requirements of the state-based version of the FEHBP.
Unions organize several of the leading FEHBP plans. For example, the Mail Handlers even offers associate membership to non-union members who wish to gain access to the health plan. These unions do not carry the insurance risk themselves; instead, they organize a group and negotiate an insurance package from an insurer for a fee. CNA Insurance organizes the Mail Handlers Benefit Plan, which has roughly 10 times as many enrollees as the union has regular union members. This "friendly society" role of unions has a long history in this and other countries.
Many union-sponsored plans also operate under the Taft-Hartley Act, where union-sponsored plans are a rational way to provide coverage when there is only a weak relationship between employer and worker. They flourish in markets that have fewer tax and regulatory obstacles to union-sponsored plans and where enrollees can receive tax or other subsidies--such as the FEHBP.
State governments could also charter FEHBP-style purchasing groups to act as intermediaries in their states, and a number of states are already experimenting along those lines with various purchasing group designs.
Most Americans pay their taxes through the workplace. This is a convenient system under which employers withhold income and Social Security taxes and send the money to the government. In addition, employees typically adjust their withholdings to take advantage of any tax breaks for which they may be eligible (e.g., the mortgage interest deduction). In a sense, the employers are actually operating the basic income tax system, but they do not in any sense design the tax code for their employees or "sponsor" the tax system. They could more appropriately be considered a clearinghouse for tax payments.
The place of employment is likewise particularly convenient and efficient for handling health insurance payments. Workers with employer-sponsored health insurance benefits typically sign up for the firm's plan when they take a job and arrange for a payroll deduction to cover premium costs for them and their families.
With individual tax credits for employees available, a small employer who is reluctant to sponsor coverage could instead carry out the critical clearinghouse role for the plan choices of his or her employees, making tax adjustments and premium payments. The employer might also decide to make a cash payment toward the plan chosen by the employee, as a fringe benefit.
Commonly, the payroll firm handling wages and benefits for the small firm would conduct these transactions. In this way, smaller employers could either directly or indirectly take responsibility for the mechanical aspects of arranging for payroll deductions and premium payments (similar to their role in the tax collection system) without having to sponsor a plan.
With tax credits, in principle, eligible employees could join any plan available in their area, not just one sponsored by their employer, and still obtain tax benefits. Thus, a small employer could play an important role in facilitating coverage without having to organize coverage by such things as providing information and making sign-up simple, instituting a payroll deduction and payment system (as many small firms do today for employee-directed savings plans), and making withholding adjustments to reflect available credits.
The government could spur the "facilitator" role of small firms that are disinclined to sponsor coverage themselves by clarifying the status of employer contributions to plans that are chosen by the employee and not sponsored by the employer. An employer wishing to set up such an arrangement today faces a dilemma. If the employer helps to pay for coverage chosen by the worker, that coverage is deemed to be an employer-sponsored plan under federal employee benefit law, and both the employer and the coverage issued to the worker by the insurer become subject to federal employer-sponsored plan regulations. To avoid such regulation, the employer must pay the worker taxable cash, which the worker can then use to purchase coverage. But that means the worker must forgo the tax benefit derived from his employer's making pre-tax contributions toward the cost of his coverage.
If the plan is interpreted as an employer-sponsored plan, the employer could face a regulatory nightmare under state insurance rules or the Employee Retirement Income Security Act (the federal law affecting certain employers) since any plan chosen by the employee would embroil the employer in complex insurance rules. However, if the arrangement is not considered an employer-sponsored plan, both employer and employee lose favorable tax benefits.
Thus, to encourage smaller employers to play the role of insurance facilitator--with or without a tax credit available to employees--federal and/or state employee benefit law needs to make clear that favorable tax benefits are available at least to the employee for an employer's contribution toward coverage whether or not the insurance is deemed an employer-sponsored plan.
High rates of uninsurance among working families in small firms are a testament to the limitations of the employment-based health system in the small-business sector. Yet both the tax system and government insurance rules discourage other insurance arrangements for these uninsured working families.
Proposals for individual tax credits for health coverage would help to remove this barrier to alternative insurance arrangements. In addition, taking steps to build an insurance infrastructure with affordable choices would enable these families to have coverage that is similar to--or even better than--the insurance available to employees of large firms.
With these reforms in place, new forms of coverage--including plans offered through churches, large corporations, and the FEHBP--would become available to working Americans in the small-business sector. For this to occur, however, Congress must recognize that an important distinction exists between using the workplace as a convenient location to obtain insurance and making tax relief to families contingent upon employer sponsorship of their health insurance.
Stuart M. Butler, Ph.D. , is Vice President for Domestic and Economic Policy Studies at The Heritage Foundation.
2. Douglas Holtz-Eakin, "The Uninsured and Rising Health Insurance Premiums," Congressional Budget Office testimony before the Subcommittee on Health, Committee on Ways and Means, U.S. House of Representatives, March 9, 2004. See also Congressional Budget Office, How Many People Lack Health Insurance and for How Long, 2003.
5. "Sources of Health Insurance and the Characteristics of the Uninsured: Analysis of the March 2003 Current Population Survey," Employee Benefit Research Institute Issue Brief No. 264, December 2003.
8. For a summary of the advantages of employer-sponsored coverage, see William S. Custer, Charles N. Kahn III, and Thomas F. Wildsmith IV, "Why We Should Keep the Employment-Based Health Insurance System," Health Affairs, Vol. 18, No. 6 (November/December 1999), pp. 115-122.
9. For a summary of the pros and cons of employer-sponsored coverage, see Uwe E. Reinhardt, "Employer-Based Insurance: A Balance Sheet," Health Affairs, Vol. 18, No. 6 (November/December 1999), pp. 124-132.
11. This figure averages workers with and without insurance, so the tax subsidy for an insured lower-income worker would be higher than this. Nonetheless, a worker in the lowest federal tax bracket would receive only just over half the subsidy for insurance received by an upper-income worker.
12. Kaiser Commission on Medicaid and the Uninsured, The Uninsured: A Primer (Washington, D.C.: Kaiser Family Foundation, 2003), p. 13, at www.kff.org/uninsured/loader.cfm?url=/commonspot/security/getfile.cfm&PageID=29345.
13. Kaiser Commission on Medicaid and the Uninsured, Uninsured in America: A Chart Book (Washington, D.C.: Kaiser Family Foundation, 2000), p. 41, at www.kff.org/uninsured/loader.cfm?url=/commonspot/security/getfile.cfm&PageID=14629.
15. Kaiser Family Foundation/Health Research and Educational Trust, Employer Health Benefits, 2003 (Washington, D.C.: Kaiser Family Foundation, 2004), at www.kff.org/insurance/ehbs2003-6-set.cfm. See also "Exhibit 2.6: Percentage of Employers Providing a Choice of Health Plans, by Firm Size, 2003," in Trends and Indicators in the Changing Health Care Marketplace, 2004 Update, Kaiser Fami
18. A recent study found that automatic enrollment for 401(k) plans boosted participation rates from 37 percent to 86 percent for such voluntary pensions, with even sharper increases for young and lower-paid employees. See Brigitte Madrian and Dennis Shea, "The Power of Suggestion: Inertia in 401(k) Participation and Savings Behavior," National Bureau of Economic Research Working Paper No. 7682, May 2000, p. 51.
19. For descriptions of the FEHBP, see Harry Cain, "Moving Medicare to the FEHBP, or How to Make an Elephant Fly," Health Affairs, Vol. 18, No. 4 (July/August 1999), pp. 25-39; Stuart Butler and Robert Moffit, "The FEHBP as a Model for a New Medicare Program," Health Affairs, Vol. 14, No. 4 (Winter 1995); and Craig Caplan and Lisa Foley, Structuring Health Care Benefits: A Comparison of Medicare and the FEHBP (Washington, D.C.: AARP Public Policy Institute, May 2000).