August 9, 1994

August 9, 1994 | FYI on

Health Care Debate Talking Points #2: Why the Numbers Will Be Wrong

(Archived document, may contain errors)

August 9. 19 94



Daniel J. Mitchell Stuart M. Butler McKenna Senior Fellow Vice President and Director of in Political Economy Domestic Policy Studies


Lawmakers soon will have to cast fateful votes on the most sweeping reform of the health care system ever proposed. They will have to vote on legislative language that few lawmakers, or their staff, have even read. The bills themselves have been cobbled together by the majority during hasty markups and behind closed doors. Members often discover that a drafted bill differs from the legisla- tion they believe they crafted in committee-as Senators on the Finance Committee recently discov- ered. And changes will be added during the floor debate itself which will have significant impacts on the operation of provisions already voted on. As Congress tries to accomplish this daunting task, lawmakers will be guided by estimates of the impact of various provisions on government spending, on tax revenues, and on businesses and households. As each provision is debated, the Congressional Budget Office (CBO), as well as a small army of Administration and congressional staff, will rush out cost estimates. The price tag for the bill as a whole will undergo continual adjustment. Major decisions by Congress will hinge on these cost estimates. And lawmakers considering the implications of their votes on constituents and the economy will be forced to rely heavily on these numbers, provided to them with an air of authority and precision. There is just one problem: history shows that the numbers will prove to be wrong. They will not be wrong by a small margin. In all likelihood many of the most important numbers will be way off the mark. And invariably the actual cost to households, businesses, and the taxpayer will be much larger than forecast-not smaller. Why is this? The overarching reason is that Congress is trying to restructure one-seventh of the entire U.S. economy. Even small changes in the highly complex triHion-dollar health care economy can have huge and unexpected results as patients, physicians, and hospitals modify their behavior to adjust to the change. But the changes proposed in the leadership's health bills are far from small.

Explains CBO Director Robert Reischauer: What people have to recognize is that Congress is trying to make changes that are more complex than anything Congress or the executive branch Ihas considered since at least the New Deal and probably since the founding of the nation. The complexity of the bills has led CBO to urge caution in the use of their estimates. Discussing the Senate Finance Committee bill, for instance, on which the Mitchell bill is based, CBO declares its " estimates of the effects of this proposal are unavoidably uncertain. 9,2 Beyond the breathtaking scale of the changes under consideration, there are specific reasons why even with the latest computers and software official estimates are likely to be well off the mark. Ex- amples: Crucial cost assumptions may be wrong. Total cost estimates may hinge on guess- ing the cost of key elements. Reputable estimates of the standard benefits package in the Clinton bill, for instance, range from $5,565 for family coverage by CBO to $5,970 by Foster Higgins, one of the nation's foremost health benefits firms. A Heri- tage estimate of the cost of Majority Leader Richard Gephardt's bill using the Foster Higgins cost, rather than CBO's, raises the cost to business in 1999 by $38 billion. 3 Disagreements about who pays what Estimating the impact of a proposal on a seg- ment of the population depends to some degree on how costs are passed on from one group to another. For instance, many estimates of employer mandates assume that re- quiring employers to pay most of the cost is a "free lunch" to the employee. Other, more reliable, estimates assume that the employee actually bears much of the "ern- ployer's" cost in reduced future wages. An analysis of this "pass-through" wage effect by Lewin-VHI, a leading health care modelling firm, found that under the Clinton plan the mandate on employers would mean an average $1,243 per year reduction in wages for employees not currently insured. 4 Garbage in - garbage out. In addition to technical disagreements about the way to cost out components of a bill, hastily developed cost estimates for measures enacted amid a flurry of last-minute changes-the conditions now prevailing on Capitol Hill- are even more likely to lead to huge underestimates in the projected costs of programs made during congressional debate. The history of health care legislation is replete with rosy forecasts proved hopelessly wrong. Example: The Medicaid program was expected to cost $1 billion annually when it was en- acted in 1965. The actual cost last year was over $80 billion. Example: In 1965, the hospital insurance component of Medicare was projected to cost $9 bil- lion in 1990. It actually cost $67 billion. Even the 1965 estimate for 1970, just five years ahead, proved to be only half the actual cost.

Example: When the Medicare Catastrophic legislation was passed in 1988, the cost of the new drug benefit was projected at $5.7 billion over five years. Just one year after passage, the govemment reestimated the cost at $11.8 billion over five years-triggering repeal of the legislation. Example: The skilled nursing care component of Medicare Catastrophic was projected during passage to cost $2.1 billion over five years. The next year it was reestimated at a stagger- ing $13.5 billion over five years. Lawmakers facing votes on health care reform legislation should bear in mind these hard lessons. Optimistic forecasts made by Administration and congressional staffers in August 1994 are likely to be proved grossly inaccurate just a few months after legislation has passed and analysts have had time to examine the full implications of detailed program changes. Meanwhile, constituents will face costs that are very different from staff estimates. This potential for error reinforces the argument for slower, cautious change, allowing the impact of one significant change to be assessed in the real world before another major step is taken. Whether reformers are liberal or conservative, wise poficy-making requires great caution.


The Administration-backed proposals on Capitol Hill offer very costly standard benefit packages, create several new open-ended entitlement programs, and may reduce incentives for many busi- nesses to control health care costs. The pmposals all contain large tax increases and spending cuts which supposedly will balance the new spending. But many of these cuts, such as deep projected re- ductions in Medicare spending, are unlikely to generate the promised savings even if they are en- acted. The central element of the Clinton-style plans before Congress is a standard benefits package. With this package, the govemment will dictate the level and type of health insumnce consumers are allowed to have. Often compared to a Fortune 500 insumnce policy, the standard benefits package in recent bills backed by the Administration is likely to be much more costly than projected for the Clinton bill. The Senate Labor bill, The Presidenies Health Care Reform Proposal for instance, Estimated National Average Premiums for 1994 adds such (dollars per policy) benefits to the CHnton Admin CBO EBRI Lewin-VI'll Wyatt Hewitt list as more ENROLLMENT CLASS services for Individual 1,932 2,100 2,202 2,260 2,285 2,"0 childmn and Couple 3,865 4,200 4,404 4,522 4,570 4,880 Single Parent 3,893 4,095 4,008 4,278 4,603 4,619 the disabled Dual Parent 4,360 S,565 6,210 4,9" 5,1S5 6,946 and more "conomy Furce: The E , the Budge4 and Health Cam Reform (Washington, D.C.: Committee generous for a Responsible Federal Budget, 1994), p. 86. mental health benefits, and the House leadership bill adds other services. Moreover, independent estimates sug- gest.that premium costs for the Clinton package could be much higher than the Administration's forecast for the Clinton bill. Higher premium packages would mean higher costs to business. More generous packages would likely prove even more costly. The calculation of the health benefit package is critical because it drives virtually all other costs in the bills. If the Administration or congressional leadership tum out to be low-balling their estimates,and the independent figures turn out to be more accurate, the financial consequences would be enor- mous. The bulk of the new spending in the various versions of the Clinton bill is for such things as premium subsidies for small businesses and low-income individuals and, in the case of the House leadership bill, to fund a new Medicare Part C program covering perhaps half the entire population. If White House and congressional estimates of premiums are too low, the cost of these subsidies grows dramatically. The cost to the private sector also would increase proportionately as employers, faced with a mandate to pay 80 percent of the premiums (in the House leadership bill), would see their costs rise sharply. Needless to say, this would boost job loss estimates even higher. Changes in the estimated cost of the standard benefit package mean widely different projected to- tal costs for any piece of legislation, even assuming no other component of the program is revised. For instance, an econometric simulation of the Senate Labor Committee's version of the Clinton bill, conducted by Heritage Foundation analysts, calculated the additional cost to business would have been $23.7 billion had the legislation been in effect in 1993. This simulation used CBO's esti- mate for the cost of the benefit package.5 If the cost of the package turned out to be 5 percent more than CBO assumed, however, the cost rises to $32.7 billion, or $9 billion more. If the package cost estimate calculated by Foster Higgins, the health benefits firm, is used the cost to business rises sharply to $45.6 billion-or virtually double the figure using CBO's estimate for the standard pack- age.

The Kennedy Bill: Why Assumptions Matter

Beneft Package Estimated By Total Additional Cost to Business

CBO $23.7 billion CBO + 5% $32.7 billion Foster Higgins $45.6 billion

A similar calculation for the House leadership bill leads to a similar difference in the net addi- tional cost to business in 1999 if different estimates am used as the basis for the standard benefits package.

The Gephardt Bill: Why Assumptions Matter Benefits Package Estimated By Total Additional Cost to Business

CBO $42.6 billion Foster Higgins $76.5 billion

Assumptions about the costs of services are just one of many crucial components that affect the to- tal cost of a health bill. Consider another element in the Clinton bill and other plans put forward by Democmts in Congress: spending limits (often called "global budgets"). There is much controversy over just how effective such spending controls would be. Using the assumption that such controls would be "75 percent effective," CBO estimated that the American Health Security Act (the "Cana- .dian" plan) would lead to national health spending being $57 billion more in 1998 than if controls proved to be "100 percent effective."


Compounding this general problem is the fact that the Administration is proposing to create sev- eral new entitlements for which cost estimates are, at best, blind guesses. In reality, these provisions( are almost certain to cost much more than is being forecast and will help drive up the cost of the standard benefit package, and in turn the total cost of the bills. Among the new entitlements which are expected to increase the deficit are:

Long-Term Care The Gephardt and Mitchell versions of the Clinton bill each contain a new long-term care pro- gram. The potential costs of any long-term care benefit program have been a concern for a number of years, and with very good reason. Approximately 7 million Americans over age 65 utilize some form of long-term care. The Congressional Research Service estimates that this number will grow to 14 million by the year 2030. Of that figure, approximately 9 million elderly Americans are expected to reside in nursing homes with the remaining 5 million staying in their own homes. Public expenditures for nursing home care alone, mostly through the Medicaid program, cost tax- payers over $32 billion in 1991. The cost of nursing home care in the United States was $74 billion in 1993. This is projected to reach $91 billion by 1995 and $138 billion by the year 2000. 7 Medicaid contributes just over 50 percent of the total, with other funds coming from Medicare, private out-of- pocket payments, and private insurance. Medicaid recipients are means-tested, meaning they must meet certain income and disability eligibility requirements. The Gephardt bill establishes a new Long-Term Care Program for Home and Community Based Services. Based upon the House Ways and Means Committee Report, the Secretary of Health and Human Services would "[e]stablish a new long-term care program to provide home and community based services for individuals ... without regard to age or income through approved State plans."8 While the goal of helping the frail elderly is well-intentioned, the creation of this new entitlement program could mean exploding future costs. Lawmakers do not need to look back far to realize why. Most home health care components of ex- isting government health programs already are exploding. Home care rose from 2 percent of total Medicare spending in 1980 to 4.8 percent on 1991. And the proportion of Medicaid dollars going to home care during that period shot up from 1.4 percent to 5.3 percent. Medicare spent about $12 bil- lion on home care in 1993, up from $5.4 billion just two years earlier. One reason for the growth is that it is difficult to monitor real need in a home setting. According to two scholars analyzing the program, "the average number of visits per beneficiary using home health care is driving overall growth."9 One can only imagine the potential for abuse and runaway cost in a universal home and community care entitlement.

Prescription Drugs The Gephardt and Mitchell bills expand the current Medicare Part B program to cover outpatient prescription drugs. Medicare Part C recipients are also eligible for this new prescription drug bene- fit. Estimates of the new prescription drug benefit have placed its cost at approximately $72 billion over a five-year period to serve the current Medicare population. Lawmakers should bear in mind the pitfalls of relying on such estimates. When the Congress passed the 1988 Medicare Catastrophic Coverage Act, the official Congressional Budget Office esti- mate for the new benefits was $30 billion over five years. Just one year later, the CBO reestimated the cost of the benefits at $43.8 billion, a $13.8 billion increase in less than a year. With respect to prescription drugs alone, the official CBO estimate in June 1988 put the new Medicare drug benefit at $5.7 billion over five years, which was the estimate used when Catastrophic was passed. But one year later, before the drug benefit was even implemented, CBO's estimate skyrocketed to $11.8 bil- lion. This represented a 207 percent increase in the estimate.

Mental Health Benefits The Gephardt and Mitchell bills include mental health benefits as part of the standard package. But experience suggests that holding down the future cost of mental health benefits could prove dif- ficult, if not impossible. According to the Congressional Research Service, part of the problem in determining a cost for this benefit is the difficulty in identifying the mentally ill population and thus eligibility for serv- ices. 'This population is heterogeneous and ill-defined, with widely differing problems and needs," notes the CRS. 10 The government has just settled several lawsuits in which hospitals and psychia- trists fraudulently claimed Medicaid reimbursements for mental health services. A recent article in the Washington Post points out that neither the White House nor the Congres- sional Budget Office has estimated the cost of unlimited mental health and substance abuse cover- age. 11 "The truth of the matter is that no one really knows ... because that sort of coverage is generally not available," said Paul Van de Water, CBO deputy assistant director. Given the lack of any dependable cost data, including such an open-ended benefit in a standard package could be very costly to companies and the government. Abuse is all but certain. According to the director of employee benefits for McDonnell Douglas, William Proffitt, "If I had unlimited coverage, no one would ever get well. They would continue to go to a doctor for the rest of their life." Other corporations agree that such an open-ended benefit is fiscally unsound. A spokes- woman for Xerox adds, "If you combine the blank-check capability of a well-meaning federal gov- ernment with a system that is unmanaged and still out of control, the bill for mental health care will 9912 be just unbelievable.


Though it has received relatively little attention in discussions of the cost of reform, one of the biggest problems in estimating costs arises because of the magnitude and complexity of the proposal. The Administration and Hill leadership propose to restructure one-seventh of the U.S. econ- omy, an effort that would mean significantly altering how almost $1 trillion is allocated. But even more than the sheer scale of the undertaking is that Congress would have to change a wide array of complex relationships and behavior affecting doctors, hospitals, insurance companies, businesses, government, researchers, drug companies, and a myriad of other elements that comprise America's health care system. As the proposal becomes larger and more complex, the odds of making correct fiscal estimates rapidly become lower. Stated simply, say a system has only two components that influence each other. If a government program changes the way one component must behave, only that component and the other it influences will change their behavior. Thus, estimating the total impact is relatively simple (although by no means certain). But say, like in the health care economy, there are thousands of components and each affects the other in ways that are not always certain. And instead of a gov- ernment program changing the behavior of just one component, say it directly changes the behavior of hundreds of components, each of which causes other components to adjust their behavior. With literally thousands of complex changes possible, predictions of the net effect become subject to far higher degrees of error. Thus, not only must government forecasters somehow accurately assess the costs of significant new entitlements for things like mental health, long-term care, and prescription drugs-all of which present immense methodological problems-they must also grapple with the fact that the health care market is interrelated. When the government mandates changes in one portion of that market, by either subsidizing insurance, imposing price controls, raising taxes, altering reimbursement pro- cedures, rationing care, or any of the other major changes contemplated, that change may have a rip- ple effect on other portions of the health care system. A study by the Office of Technology Assessment @9TA), a special office of Congress, openly ad- mits that government forecasts are quite unreliable. I The study cites four critical areas where ade- quate information is lacking, meaning assumptions often are nothing more than guesses. The four areas are: Do Cost Controls Work? The study questions the validity of expenditure limits and raises concerns about the potential role of payment mechanisms, administrative systems, and controls on cost overruns.

Does Managed Competition Save Money? The study raises the issue of how much savings could be expected from managed competition approaches compared with traditional fee- for-service care. Moreover, OTA notes also that unless it is possible to estimate accurately how many consumers would choose managed care, it is very difficult to predict the tow impact of savings that might exist. How Much Does It Cost to Cover the Uninsured? While supporters of government-run care sometimes claim huge savings will be achieved by extending coverage to all, the OTA study reveals that the all-important question is how much additional care this popula- tion will request once they have insurance. Wide variations in this estimate mean wide dif- ferences in the total cost of a measure.

Are There Administrative Savings? While this is more of an issue when addressing propos- als for a Canadian-style single-payer system, the OTA study notes that there are major questions regarding whether administrative costs can be reduced if government becomes more involved.


These issues raised by the OTA are important because there is ample reason to believe that sup- porters are greatly understating the cost of Clinton-style health reform. While OTA notes that many of the questions it raised have no answers, history does offer a guide to the potential consequences of forecasting errors. The governments track record leaves much to be desired and suggests that in- creased government intervention will cause costs to skyrocket. Consider, for instance, the two biggest government health care programs:

Medicare This program for senior citizens is now the third largest portion of the federal budget, behind only Social Security and Defense. What is most noteworthy about Medicare is not so much its size but how quickly it has grown. Back in 1965, government estimators projected that the hospital costs for the program would reach $9 billion by 1990. The actual cost of Medicare hospital insurance (Part A) that year had risen to $67 billion-hardly a ringing endorsement for government accuracy. 14 The program was more than seven times more costly by 1990 than government forecasters pre- dicted. The other portion of Medicare, Part B Supplementary Medical Insurance, also has grown in part because of the very price controls and restrictions placed on hospital services and intended to cut costs. Among other things, these steps encouraged doctors and hospitals to shift from inpatient serv- ices covered by Part A to outpatient services covered by Part B. This cost-shifting is an example of the budget errors and accounting mistakes which inevitably will follow major health care reform.. Also driving costs beyond expectations is the fact that politi- cians expand programs. While it is not possible to predict precisely how a government-run health care bill enacted this year would be expanded in the future because of this effect, it can be counted on to happen. Critics of government-run health care note that once health care spending becomes politicized, interest groups and beneficiaries will lobby to have certain services added to the stand- ard insurance package employers and workers must buy. There is every reason to assume that tomor- rows politicians will respond similarly to yesterday's and today's lawmakers and vote to expand benefits. Another example is the Medicare home care program. Six years ago, estimators predicted 1993 costs would be $4 billion. The actual costs were $10 billion. 15

Medicaid Also created in 1965, the Medicaid program's costs have spiralled upwards. This program to fund health care for the poor originally was seen as a very limited program, with annual costs of $1 bil- lion. 16 Actual federal outlays this year will top $80 billion, and the cost of the program is growing at over 12 percent per year. Even more than with Medicare, the reason for the explosion of Medi- caid spending has been the propensity of politicians to expand the program. The Washington Post ran a series earlier this year detailing how one House Subcommittee Chairman, Henry Waxman of California, has manipulated the system to expand Medicaid. 17

But these maneuvers by congressional chairman often backfire as state and local officials figure out how to manipulate the change. Waxman's purpose was to channel more money to states to help hospitals serving a disproportionate share of uninsured and Medicaid patients. But state officials found loopholes to permit them to use the extra federal payments for often unrelated purposes, lead- ing to huge unanticipated cost increases to the federal government. The result: federal payments to states for the program shot up from $500 million in 1990 to $10.8 billion in 1992. According to the former health director for Louisiana, the Waxman change unwittingly was "God-given to Louisi- ana," and the state's goal became to "leverage every federal dollar we could get our hands on."18 White House officials at the time were taken completely unawares by the hemorrhaging of federal dollars. "Nobody knew what happened to the money," says former OMB official Thomas Scully, "It was money going out the back door when nobody was watching. 1119 This was the result of one small technical change in an existing health care program by a very knowledgeable House chairman. Lawmakers should ponder the potential for an avalanche of unin- tended and costly manipulations in the complex, often poorly drafted and sweeping changes in- volved in the major bills now before Congress.


Members of Congress very often assume that forecasting is a precise science. In markups on ma- jor bills, staff will telephone analysts at the Office of Management and Budget or the Congressional Budget Office and come back within a few minutes with seemingly exact cost estimates of new pro- visions. And even before bills are fully drafted, detailed five-year projections are circulated. But forecasting the cost of complicated legislation is actually about as exact a science as predict- ing the weather two Tuesdays from now. Painful experience shows that apparently minor changes in the law can trigger huge new costs. The health care legislation prepared by the majority in Congress will mean radical changes in the nation's health care system. In the trillion-dollar economy, potentially hundreds of billions of dol- lars will be redirected from one set of hands to another, depending on how physicians, hospitals, businesses, and patients respond to the new system. The legislation will be combed for loopholes to be exploited. Huge new entitlement programs will be created. No CBO official can tell any lawmaker with even the slightest degree of certainty what a reform of the magnitude contemplated will actually cost. Nor can any Member of Congress know how his or her constituents will be affected and who in their district or state will consider themselves a loser. All lawmakers can take for granted is that the cost will be much higher than they are now being told and that a new health system will work very differently from the way its proponents claim.

About the Author

Daniel J. Mitchell, Ph.D. McKenna Senior Fellow in Political Economy

Show references in this report

I Quoted in Robert Pear, "Second Thoughts on Health Data," The New York Times, May 9, 1994.

2 Congressional Budget Office, "A Preliminary Analysis of the Health Security Act as Reported by the Senate Finance Committee," July 28, 1994.

3 Stuart M. Butler, with David H. Winston and Christine L. Olson, "Health Care Debate Talking Points #1: Cost to Business of the Gephardt Bill," Heritage Foundation FYI No. 21, August 3, 1994.

4 See Stuart M. Butler, "How the Clinton and Nickles-Steams Health Bills Would Affect American Workers," Heritage Foundation Issue Bulletin No. 188, April 11, 1994.Sourm: John C. Uu, with David H. Winston and Christine L. Olson, "Clinton Heavy: The Kennedy Health Bill," Hevitage Foundation Issue Bulletin No. 197, July 21,1994. Source. Stuart M. Butler, with David H. Winston and Christine L Olson, 'Health Care Debate Talking Points #1: Cost to Business of the Gephardt Bill,9 Heritage Foundation FYI No. 21, August 3, 1994 [updated August 9, 19941.

5 See John C. Liu, with David Winston and Christine L. Olson, "Clinton Heavy: The Kennedy Health Bill," Heritage Foundation Issue Bulletin No. 197, July 21, 1994.

6 U.S. Congress, Office of Technology Assessment, Understanding Estimates ofNational Health Expenditures Under Reform (Washington, D.C.: U.S. Government Printing Office, May 1994), p. 11.

7 Congressional Budget Office, "CB0 MEMORANDUM, Projections of National Health Expenditures: 1993 Update," p. 11, October, 1993.

8 Health Security Act, "Report of the Committee on Ways and Means, House of Representatives on H.R. 3600," 103rd Congress, 2nd Session, Rept. 103-601, p. 560, Part 1.

9 Christine Bishop and Kathleen Skwara, "Recent Growth of Medicare Home Health Care," Health Affairs, Vol. 12, No. 3 (Fall 1993), pp. 95-110. See also LTC News & Comment, December 1993, p. 5.

10 Edward R. Klebe, "Medicaid Services for the Mentally III," CRS Reportfor Congress, Congressional Research Service, July 20,1993,p.l.

11 David S. Hilzenrath, "Ibe Quandary Over Mental Health Care Costs' Clinton, Others Call for Unlimited Coverage; But Corporate America Sets Caps on Benefits," The Washington Post, July 25, 1994.

12 Ibid.

13 Understanding Estimates of National Health Expenditures Under Reform, p. 11.

14 "Health Care Fact Sheet: Original Medicare Cost Estimates," CRS Reportfor Congress, September 23, 1993.

15 "Remember Cost Control," Newsweek, July 25, 1994, p. 2 1.

16. "Health Care Costs a Long-Term Headache," 7he Washington Post, October 15, 1994. p. 1.

17 A four-part series of articles by Dan Morgan, running January 28, January 3 I-February 2, 1993.

18 "Louisiana Took 'Every Federal Dollar We Could Get Our Hands On"', The Washington Post, January 31, 1994, p. A9.

19 "Snudl Provision Turns Into a Golden Goose" The Washington Post, January 3 1, 1M, p-8.