Comprehensive Medicare Reform: Defined Benefit vs. Defined Contribution

Final Report
Prepared for:

National Coalition on Health Care
September 1, 1998

By:John F. Sheils
Andrea Fishman
The Lewin Group, Inc.

Table of Contents

Executive Summary *

The Voucher Model *

Defined Benefit vs. Defined Contribution *

Defined Contribution/Defined Benefit Savings Compared *

Safety Net Issues *

Medicare Benefits Package *

Closing *

A. Introduction *

B. Historical Perspective On Medicare Cost Containment *

C. Adequacy of the Medicare Benefits Package *

D. The Voucher Model *

1. The Federal Employees Health Benefits Program (FEHBP) Model *

2. The Aaron/Reischauer Proposal *

3. The Progressive Policy Institute (PPI) Voucher Proposal *

4. The Medical Savings Account (MSA) Model *

5. The Personal Retirement Insurance for Medical Expenses (PRIME) Program *

E. Key Issues In Voucher Program Design *

1. Benefits Package *

2. Setting the Voucher Amounts *

a) The Current HMO Payment Method *

b) Move to Competitive Bidding *

c) Potential Savings from Competitive Bidding *

d) Defined Contribution and Defined Benefit Savings Compared *

3. Selection Effects *

4. Safety Net Issues *

5. Quality of Care under a Voucher-Based System *

F. Conclusion *Executive Summary

The Medicare program is defined by its commitment to provide a specified set of health benefits to all eligible persons regardless of income. Although the Medicare benefits package is less comprehensive than most private health plans, it covers inpatient hospital care, physician services, outpatient hospital services, home health and skilled nursing facility care. All of these services are covered subject to specified deductibles and coinsurance requirements. The current Medicare program is considered to be a “defined benefit” program by virtue of its commitment to provide these specified services to all eligible persons.

The growing cost of Medicare has led to proposals that would replace Medicareís commitment to provide a defined level of services to all eligible persons with a program that provides a “defined contribution” voucher to Medicare beneficiaries that would be used to help purchase private health insurance coverage. The amount of the voucher would be determined by the Congress through the budgeting process based upon the amount of funds available for Medicare rather than the projected cost of a given level of coverage.

However, under the defined contribution model, there is no guarantee that the voucher amounts will be enough to purchase coverage comparable to the existing Medicare benefits package. Many beneficiaries would have to supplement the voucher amount with their own funds to maintain even the comparatively spartan Medicare benefits package that all beneficiaries now have. While Medicaid will provide a safety net for persons below the federal poverty level (FPL), other near-poor aged and disabled persons may not be able to afford to supplement the voucher amount to maintain at least the Medicare benefits package. Consequently, some form of premium subsidy will be needed for poor and near-poor persons to prevent these beneficiaries from becoming substantially underinsured.

Not all Medicare voucher proposals imply a shift to a defined contribution model, however. The amount of the voucher, for example, could be set equal to the average bid price in each market area for the current Medicare benefits package. This voucher model is still a defined benefits plan in that all beneficiaries are guaranteed that the voucher amount would be enough to purchase at least this defined set of benefits in their area. The voucher model becomes a defined contribution program only when the guarantee that beneficiaries can obtain a minimum level of services when the voucher amount is eliminated.The Voucher Model

Many voucher proposals are inspired by the managed competition model that was so prevalent during the heath reform movement earlier this decade. Under managed competition, individuals select from several health plans competing on the basis of price and quality. Health plans will typically vary in terms of their use of managed care, co-payments, types of services covered and price. Detailed information is provided to the consumer on the premiums charged and the characteristics and limitations of each plan. Consumers are also provided with summary comparisons of the plans on various measures of quality so that individuals can consider both price and quality in selecting the health plan that is right for them. The price competition among private health plans also would be a powerful force for controlling health care costs.

There is substantial evidence that competition has been successful in reducing the rate of growth in health care costs in the private sector. For example, between 1993 and 1997 the average annual rate of growth in per-capita premium costs was 2.7 percent for privately insured persons compared to an increase of 7.8 percent per Medicare beneficiary. The rate of growth in per-capita costs under the Federal Employees Health Benefits Program (FEHBP) program, which adopts several of the principles of managed competition, has actually been lower than in the private sector. In addition, several state Medicaid programs have reduced costs through competitive bidding among private managed care plans. Figure ES – 1 compares the annual rates of growth in per-capita health spending for Medicare and private coverage over the 1990 – 1997 period.

Figure ES – 1
Growth in Medicare Spending per Beneficiary and Private Health Insurance Spending per Enrollee

a/ Estimated.Source: Data from the Health Care Financing Administration (HCFA) for 1971 through 1996 and Congressional Budget Office (CBO) estimates for 1997. Private health insurance spending data are from the HCFA National Health Expenditure Accounts and include: employer sponsored-coverage, retiree coverage, individually purchased non-group coverage, and individually purchased Medigap coverage.

The crucial question in this analysis is whether competition will continue to slow the growth in health care costs in future years. In recent months, there have been widely publicized reports of large premium increasesñincluding an 8.5 percent average premium increase under the FEHBPñsuggesting that rapid health care inflation is returning to the private sector. However, two recent surveys of employers throughout the country reveal that average per-capita premium increases in 1998 were between only 3.3 percent to 3.5 percent, which is about half the projected rate of increase in per-capita spending under Medicare. Moreover, there are several studies of cost growth over the past two decades that show that the long-term rate of growth in health care costs is significantly reduced as the share of the population covered under managed care plans with selective contracting arrangements increases. Based upon these studies, the increase in managed care enrollment implied by a voucher program with competitive bidding would result in a long-term reduction in the long-term annual rate of growth in health spending of between 0.5 and 1.5 percentage points.

To illustrate the potential long-term impact of price competition on Medicare costs, we estimated the Hospital Insurance trust fund balance assuming that competition slows the long-term rate of growth in Medicare spending by one-half of one percentage point per year starting in 20051. Under this scenario, the trust fund deficit in 2030 would be reduced by about 25 percent (Figure ES – 2). If competition results in a reduction in the long term rate of growth in per-capita spending of 1.0 percentage points per year, the trust fund deficit in 2030 would be reduced by nearly half. The deficit in 2030 would be reduced by two-thirds if competition reduces cost growth by 1.5 percentage points per year. Although the recent cost performance of the private sector suggests that savings of these or even larger magnitudes could be achieved in todayís health care markets, it is unclear whether this level of savings could be achieved continuously in future years.
Figure ES – 2
Long-Range Hospital Insurance Trust Fund Income and Costs under Alternative Cost Growth Assumptions Expressed as a Percentage of Taxable Payroll

Source: 1998 Annual Report of the Board of Trustees of the Federal Hospital Insurance Trust Fund; Lewin Group estimates.

Defined Benefit vs. Defined Contribution

Voucher models can be setup as either defined benefit or defined contribution plans. The voucher model can be considered a defined benefit plan as long as it assures that the voucher amounts will be enough for each beneficiary to purchase at least the equivalent of the Medicare benefits package in their market area. While competition is expected to reduce cost growth under the voucher program, it is not expected to completely eliminate the Medicare funding shortfall in future years. Thus, under the defined benefit model, other provisions such as increasing the Medicare eligibility age or increasing revenues will still be needed to ensure the financial integrity of Medicare into the next century.

The voucher model becomes a defined contribution program only when the guarantee that the voucher will be enough to purchase a defined set of benefits is eliminated. Under most defined contribution voucher proposals, the voucher amount is envisioned to be enough to purchase the Medicare benefits package in at least the first year of the program. However, the voucher amounts are unlikely to keep pace with health care inflation in future years as the voucher amounts are constrained to stay within the spending limits set by Congress through the budgeting process. Thus, over time, beneficiaries will have to supplement their voucher amount with their own funds just to maintain the equivalent of the current Medicare benefits package. Thus, the defined contribution model would reduce the federal cost of Medicare by gradually shifting an ever-increasing portion of the cost of the Medicare benefits package to beneficiaries over time.

This is not an unreasonable way of dealing with the impending crisis in Medicare financing, assuming that we want to ask beneficiaries to finance the growing cost of Medicare. For example, the impact of such an approach on families with aged persons would probably not be very different than that of an across-the-board increase in the Medicare Part-B premium that is calibrated to achieving the same level of program savings. The key issue here is whether we want to ask beneficiaries to shoulder the full cost of financing expected shortfalls in the Medicare program.Defined Contribution/Defined Benefit Savings Compared

As discussed above, the competitive bidding model is unlikely to save enough to eliminate the projected funding deficit for the Medicare program. This differs from the defined contribution model whereby definition expenditures under the program are constrained not to exceed the amount of funding available under the program. For example, if we were to adopt the defined contribution model beginning in 2005, Medicare spending would be reduced by $3.6 trillion over the 2005 through 2030 period (Figure ES – 3). By comparison, we estimate that adopting a defined benefit voucher program based upon competitive bidding would save between $1.0 trillion and $2.8 trillion over that same period, depending upon the extent to which competitive bidding slows the rate of growth in spending.

While the defined contribution model achieves greater savings than a defined benefit voucher approach, it does so by eliminating the guarantee that the voucher amount will be enough to purchase at least a given level of coverage. By comparison, the savings realized through the defined benefit voucher model are achieved through competitive bidding while maintaining the guarantee of a minimum level of coverage for all Medicare beneficiaries. Thus, even with an aggressive competitive bidding system, we will still need to make other changes in the program such as an increase in the Medicare eligibility age or increases in revenues to fully close the Medicare funding gap through 2030. Therefore, if we wish to maintain Medicareís benefits guarantee, competitive bidding would represent only one ingredient in a long-term funding solution.
Figure ES – 3
Potential Savings to Medicare under the Defined Contribution and the Defined Benefit Competitive Bidding Voucher Models Compared (in billions)

a/ Assumes that Medicare expenditures would be constrained to grow no faster than the growth in revenues in each year.

b/ Assumes that all Medicare beneficiaries would be covered under a voucher program using competitive bidding beginning in 2005. Savings are estimated under three alternative assumptions on the extent to which competitive bidding will slow the rate of growth in spending for the Medicare program.

Source: Lewin Group estimates.

Safety Net Issues

A particular concern with the defined contribution model is that its financial impact on beneficiaries generally will be the same regardless of the income of the beneficiary. This is because all beneficiaries will have to pay the cost of supplementing the voucher amount to maintain at least the Medicare benefits package. However, many poor and near-poor individuals will not be able to afford the cost of supplementing the voucher amount to maintain their current level of coverage, which could cause many Medicare beneficiaries to become substantially underinsured. Low-income aged and disabled persons who qualify for Medicaid generally will be protected because the Medicaid program covers the cost of Medicare-covered services that are not paid under Medicare (e.g., Medicare coinsurance). However, many near-poor Medicare beneficiaries who do not qualify for Medicaid will not be able to afford the cost of supplementing the voucher amount.

Some form of a subsidy program will be needed to help these near-poor beneficiaries pay the premium supplement required to maintain at least the Medicare benefits package if we are to avoid creating a substantial sub-population of lower-income underinsured Medicare beneficiaries. The amount of these subsidies and the number of persons requiring them would increase in future years as the difference between the voucher amount and the cost of the basic Medicare benefits package increases over-time. However, even with a subsidy program the defined contribution model could result in substantial net savings to the federal government. Moreover, the adoption of a subsidy program would help assure that the shift of costs to beneficiaries under the defined contribution model is concentrated among those who can most afford it.Medicare Benefits Package

Finally, there is a growing paradox in the Medicare reform debate. While there is heavy emphasis on doing something to avert the impending Medicare funding crisis, there is growing interest in expanding upon the services covered under Medicare. For example, even some of the proponents of the defined contribution model would add prescription drugs and a stop-loss limit on out-of-pocket spending to the Medicare benefits package. We estimate that even a modest version of these benefits (i.e., stop-loss limit of $5,000 and prescription drug coverage with a $500 deductible, 20 percent copay and $1,500 benefit cap) would increase Medicare program costs by $16.6 billion in 1999 (Figure ES – 4), which would represent an 7.6 percent increase in Medicare costs. This would substantially increase Medicareís long-term funding deficit unless an adequate source of funding is specified.
Figure ES – 4
Federal Cost of Providing Prescription Drug and Stop-Loss Coverage under Medicare for Alternative Provisions in 1999 (in billions) ^a/

a/ Assumes that the deductibles, benefit caps and stop-loss limits are indexed at the same rate as the Medicare Part A hospital deductible over time.

b/ Drug benefit valued at average acquisition cost.

Source: Lewin Group estimates using the Medicare Benefits Simulation Model (MBSM).

The increase in Medicare benefits costs under this example benefit package could be offset by reducing other benefits provided under the program. For example, we estimate that increasing the annual Part-B deductible to $600 for all Medicare Part-B covered services in 1999 (i.e., excluding drugs) would reduce costs by about as much as is needed to pay for this additional prescription drug and stop-loss coverage in that year2. Despite the increase in the deductible, many individuals would be able to reduce or eliminate their Medigap coverage resulting in premium savings to the beneficiary. However, such a large increase in the deductible would be a hardship for many beneficiaries on fixed incomes, and is likely to be unpopular.

The cost of adding these benefits could be reduced if they were implemented as a part of the voucher program. For example, the cost of these benefits could be reduced through competitive bidding. Also, many plans are likely to offer these additional services at no charge to the beneficiary as an inducement to enroll. However, even under a competitive bidding voucher model, additional revenues would be needed to cover the cost of these additional benefits.

This calls to mind the lessons learned from the repeal of the Medicare Catastrophic Coverage act of 1988, which would have provided coverage for these services under Medicare. The Act was repealed largely due to financing provisions that would have financed the entire cost of the program with premium increases, some of which were income-related. Beneficiaries objected to this approach in part because many found themselves paying substantially more in premiums for coverage than they already had under an employer-sponsored retiree health plan or their Medigap insurance policy. Indeed, the fact that most beneficiaries already have coverage for these services suggests that it will be difficult to justify a program to cover these services that is financed by increased contributions from current beneficiaries.Closing

The voucher model could provide a vehicle for using market-based competition to control the growth in Medicare spending without actually retreating from Medicareís commitment to provide coverage for a defined set of services to all eligible beneficiaries. While there is evidence that this competitive model could result in substantial savings to Medicare, it is unlikely that savings will be great enough to eliminate Medicareís long-term funding deficit. Consequently, we would still need to adopt other Medicare program changes such as tax increases or increases in the Medicare eligibility age to fully close the Medicare funding deficit.

However, the defined contribution model has the potential to eliminate the Medicare funding shortfall on its own. This is because the voucher payments under the program could be set at levels that are equal to the amount of funds that Congress determines to be available for the program in each year. Under this model, there is never a funding shortfall for the program simply because payments are not permitted to exceed the amount of available revenues. However, this approach abandons Medicareís commitment to ensure access to a given level of services for all eligible beneficiaries and effectively shifts a large portion of the cost of providing these services to beneficiaries regardless of their ability to pay. Indeed, the fundamental issue in Medicare reform will be re-appraising our commitment to assure a minimum level of benefits for current and future beneficiaries.

1. Introduction
2. Historical Perspective on Medicare Cost Containment
3. Adequacy of the Medicare Benefits Package
4. Alternative Voucher Models
5. Issues in Voucher Design
6. Conclusions
7. Historical Perspective On Medicare Cost Containment
8. Adequacy of the Medicare Benefits Package
9. The Voucher Model
10. The Federal Employees Health Benefits Program (FEHBP) Model
11. The Aaron/Reischauer Proposal
12. The Progressive Policy Institute (PPI) Voucher Proposal
13. The Medical Savings Account (MSA) Model
14. The Personal Retirement Insurance for Medical Expenses (PRIME) Program
15. Key Issues In Voucher Program Design
16. Benefits Package
17. Setting the Voucher Amounts
18. The Current HMO Payment Method
19. Move to Competitive Bidding
20. Potential Savings from Competitive Bidding
21. Defined Contribution and Defined Benefit Savings Compared
22. Selection Effects
23. Safety Net Issues
24. Quality of Care under a Voucher-Based System
25. Conclusion

ENDNOTES

1	We assume that it would take five to six years for a Medicare voucher program to be fully implemented nationwide.
2	Estimate assumes that the Medicare Part-B deductible is increased annually at the annual rate of growth in per-capita Part-B spending.
3	Medicare enrollment grew by about 2.5 percent per year between 1968 and 1994. See: “Background material and data on programs within the jurisdiction of the Committee on Ways and Means,” Congressional Research Service (CRS), The Green Book. Program enrollment is now growing at an annual rate of about 1.2 percent.
4	Ball, Robert M., “What Medicare’s Architects had in Mind,” Health Affairs, Winter 1995.
5	“Health Benefits in 1996,” KPMG, Peat Marwick, LLP, 1997.
6	Starting in the early 1970s, beneficiaries could be covered under HMOs that were reimbursed on the basis of costs incurred for these beneficiaries. Medicare conducted risk-based HMO demonstrations starting in the late 1970s and the HMO option was made available to all participants in 1982 under TEFRA.
7	Congressional Budget Office (CBO), “The Economic and Budget Outlook: Fiscal Years 1999 – 2008,” January 1998.
8	Prospective Payment Commission, “Medicare and the American Health Care System,” June 1995.
9	Data from the Health Care Financing Administration (HCFA) for 1971 through 1996 and Congressional Budget Office (CBO) estimates for 1997. Private health insurance spending is from the HCFA National Health Expenditure Accounts and includes: employer-sponsored coverage, retiree coverage, individually purchased non-group coverage, and individually purchased Medigap coverage.
10	Health Insurance Association of America (HIAA); and “Health Benefits in 1996,” KPMG Peat Marwick, 1992 through 1997. <
11	One study indicates that much of the gap between Medicare and private insurance growth rates prior to 1994 is narrowed but not eliminated when growth rates are used for only those services that are typically covered under both Medicare and private insurance plans (i.e., long-term care services are removed from Medicare spending and prescription drugs are removed from private plans). See: Moon, Marilyn and Zuckerman, Stephen, “Are Private Insurers Really Controlling Spending Better than Medicare?” (Report to the Henry J. Kaiser Family Foundation), Urban Institute, July 1995.
12	Hay Group, “Hay Benefits Report,” Hay Group 1990 through 1997.
13	Under the BBA, rates are set at the highest of three amounts calculated for each county: 1) rate calculated as a blend of an area-specific (local) rate and a national rate, 2) A minimum floor rate, and 3) A rate reflecting a minimum increase from the previous year’s rate.
14	Lewin Group estimates using the Medicare Benefits Simulation Model (MBSM) based upon the Medicare Current Beneficiary Survey (MCBS) data.
15	Administration estimates.
16	Eppig, Franklin J. and Chulis, George S., “Trends in Medicare Supplementary Insurance: 1992-96,” Health Care Financing Review, Fall 1997.
17	Foster Higgins, “National Survey of Employer-Sponsored Health Plans,” Report 1996.
18	Lewin Group estimates using the Medicare Current Beneficiary Survey (MCBS) data.
19	Chulis, George S., Eppig, Franklin P., Hogan, Mary O., Waldo, Daniel, and Arnett, Ross H., III, “Health Insurance and the Elderly,” Health Affairs; and data provided by the Health Insurance Association of America (HIAA).
20	Beneficiaries pay a copayment of $191 per day for hospitalizations of between 60 and 90 days. Beneficiaries also face coinsurance of $382 per day for an additional lifetime allowance of 60 days for inpatient care. Coinsurance for nursing facility care is $95.50 per day.
21	Christensen, Sandra and Shinogle, Judy, “Effects of Supplemental Coverage on Use of Services by Medicare Enrollees,” Health Care Financing Review, Fall 1997.
22	Manning, W. G., et al., “Health Insurance and the Demand for Medical Care: Evidence from a Randomized Experiment,” The American Economic Review, June 1987.
23	“Medicare Supplemental Loss Ratios in 1994,” National Association of Insurance Commissioners, October 1995.
24	The conversion of retiree benefits to a cash payment for retirees would have tax consequences for affected retirees because employer expenditures for health benefits are excluded from taxation while direct cash payments are taxable to the recipient.
25	Fronstein, Paul and Copeland, Craig, “Medicare on Life Support: Will it Survive?” Employee Benefits Research Institute, September 1997.
26	The FEHBP premium contribution is capped not to exceed 75 percent of the cost of a plan for any individual beneficiary.
27	Smith, W., editor, “Federal Health Benefits Information and Open Season Guide,” National Association of Retired Federal Employees, 1994.
28	Congressional Research Service (CRS), Federal Employees Health Benefits Program, May 1989.
29	Francis, W., “The Political Economy of the Federal Employees Health Benefits Program,” Health Policy Reform, Helms, B., American Enterprise Institute Press, Washington, D.C., 1993.
30	McArdle, F.B., “Opening up the FEHBP,” Health Affairs, Summer 1995.
31	Butler, Stuart M. and Moffit, Robert, “The FEHBP as a Model for a New Medicare Program,” Health Affairs Winter 1995.
32	Aaron, Henry J. and Reischauer, Robert D., “The Medicare Reform Debate: What is the Next Step?” Health Affairs, Winter 1995.
33	Sheils, John F., Claxton, Gary J., and Haught, Randall A., “Medicare Reform Options: An Analysis of Private Coverage, Beneficiary Spending and the Budget,” (Report to The Progressive Policy Institute), The Lewin Group, Inc., April 1996.
34	Pauly, Mark V. and Goodman, John C., “Tax Credits for Health Insurance and Savings Accounts,” Health Affairs, Spring 1995.
35	Lewin Group estimates using the Medicare Benefits Simulation Model (MBSM).
36	The BBA authorizes the sale of MSA coverage on a trial basis for a maximum of 390,000 persons.
37	Sheils, John F., Claxton, Gray J., and Haught, Randall A., “Changes in Medicare Program Spending under Alternative Medical Savings Account Models,” (Report to The National Committee to Preserve Social Security and Medicare), The Lewin Group, Inc., September 22, 1995.
38	Gramm, Phil, “Medicare Policy for Future Generations: A Search for a Permanent Solution,” April 30, 1998.
39	Savings, Thomas R. and Rettenmaer, Andrew J., “Financing Medicare,” (Statement to the Restructuring Taskforce of the National Bipartisan Commission on the Future of Medicare), Private Enterprise Research Center, Texas A&M; University, June 23, 1998.
40	Under the traditional Medicare program, hospitals that provide indigent care receive funds under the Disproportionate Share Hospital (DSH) program. These funds are used to cover the expenses of providing uncompensated care to uninsured individuals. Hospitals also currently receive payments for Graduate Medical Education (GME), which are intended to cover the costs associated with teaching and health care research.
41	Alternatively, the voucher amounts could be indexed to the growth in the consumer price index (CPI) or an index of wage growth. However, because health care costs typically grow faster than the CPI and wages, this approach guarantees that beneficiaries will have to supplement the voucher payment with their own resources if they are to maintain some minimal standard of coverage.
42	Lewin Group estimates using the Medicare Benefits Simulation Model (MBSM) based upon the Medicare Current Beneficiary Survey (MCBS) data.
43	Estimate assumes that the Medicare Part-B deductible is increased annually at the annual rate of growth in per-capita Part-B spending.
44	The BBA permits HCFA to set statewide rates for rural areas with small numbers of beneficiaries.
45	Hogan, Christopher and Cox, Donald F., “Risk Selection and Risk Adjustment in Medicare,” 1996 Annual Report to Congress, Physician Payment Review Commission, June 1996.
46	Brown, Randall S. and Hill, Jerrold W., “The Effects of Medicare Risk HMOs on Medicare Costs and Service Utilization,” HMOs and the Elderly, Luft, Harold S. (ed.), Health Administration Press, 1994; Riley, Gerald, et al., “Health Status of Medicare Enrollees in HMOs and Fee-For-Service in 1994,” Health Care Financing Review, Summer 1996.
47	Sheils, John F. and Haught, Randall A., “Risk Selection in Florida Medicare Risk HMOs,” (Report to the American Association of Health Plans (AAHP), The Lewin Group, Inc., July 16, 1997.
48	CBO Memorandum, “Predicting How Changes in Medicare’s Payment Rates would Affect Risk-Sector Enrollment and Costs,” March 1997.
49	Over a five-year period payment amounts in each area would become a 50/50 blend of national and local average FFS costs.
50	Medicare Payment Advisory Commission, “Report to the Congress: Medicare Payment Policy,” March 1998.
51	Estimates provided by the Congressional Budget Office (CBO).
52	Some sparsely populated counties would be combined on a statewide or regional basis to form areas that are large enough to attract bidders.
53	The General Accounting Office (GAO), “States Turn to Managed Care to Improve Access and Control Costs” (GAO/HRD-93-46), 1993.
54	Information provided by state program staff.
55	Lewin Group estimate based upon data from the Health Care Financing Administration (HCFA).
56	Welch, W.P., “HMO Market Share and its Effect on Local Medicare Costs,” HMOs and the Elderly, Health Administration Press, Ann Arbor Michigan 1994.
57	Robinson, J.C., “HMO Market Penetration and Hospital Cost Inflation in California,” Journal of the American Medical Association, 20 November 1991.
58	Zwanziger and Melnick, “Costs and Price Competition in California Hospitals, 1980-90,” Health Affairs, Fall 1994.
59	The Lewin Group, Inc., “The Cost of Legislative Restrictions on Contracting Practices: The Cost to Governments, Employers and Families,” June 1995.
60	Assumes that Medicare expenditures are constrained to grow no faster than the growth in revenues in each year.
61	Assumes that all Medicare beneficiaries would be covered under a voucher program using competitive bidding beginning in 2005. Estimate assumes this would reduce the rate of growth in Medicare spending by between 0.5 percent and 1.5 percentage points per year.
62	Lewin Group estimates using the Medicare Benefits Simulation Model (MBSM) based upon the Medicare Current Beneficiary Survey (MCBS) data.
63	Lewin Group estimates using the Medicare Benefits Simulation Model (MBSM) based upon the Medicare Current Beneficiary Survey (MCBS) data for 1995. These estimates are similar to those developed by Marylin Moon and Karen Davis using data from the Health Care Financing Administration (HCFA), presented in “Medicare: A Profile,” US Government Printing Office, February 1995.
64	Newhouse, J. P., “Patients at Risk: Health Reform and Risk Adjustment,” Health Affairs, Spring I 1994.
65	California Managed Risk Medical Insurance Board “Health Insurance Plan of California: Methods for Calculating and Applying Risk Assessment and Risk Adjustment Measures, Results of Simulation #1,” May 25, 1995.
66	National Association of Insurance Commissioners (NAIC) “Small Employer; Individual Health Insurance Availability Model Act,” Washington, D.C., 1996.
67	Savings would be greater if state Medicaid plans are permitted to require dually eligible beneficiaries to enroll in a Medicare HMO as a condition of receiving Medicaid benefits. States are not permitted to do this under current law, which has stymied state efforts to cover the dually eligible through HMOs.
68	Office of Inspector General, “Beneficiary Perspectives of Risk HMOs,” March 1995; and Tudor, C. G., Riley, G., and Ingber, M., “Satisfaction with Care: Do Medicare HMOs Make a Difference,” Health Affairs, Spring 1998.
69	General Accounting Office, Report to the Special Committee on Aging, US Senate, “Many HMOs Experience High Rates of Beneficiary Disenrollment,” April 1998.
70	Chassin, Mark R., “Assessing Strategies for Quality Improvement,” Health Affairs, May 1997.
71	Wennberg, John E. and Cooper, M.M., “Dartmouth Atlas of Health Care,” American Hospital Association Publishing, 1996.
72	Marmor, Theodore and Oberlander, Johnathan, “Rethinking Medicare Reform,” Health Affairs, January/February 1998.
73	Moon, Marilyn and Davis, Karen, “Preserving And Strengthening Medicare,” Health Affairs, Winter 1995.