Health Care and Markets
Prepared for Members and Committees of Congress
Health care spending is one of the most rapidly growing portions of the federal budget.
Projections suggest if the rapid growth in health care costs is not curtailed, governments at all
levels will face an uncomfortable choice between significant cuts in other spending priorities or
major tax increases. This report examines the economic justification for government intervention
and involvement in health care markets.
Many analysts claim market-oriented policies, in certain instances, could lower costs and enhance
efficiency in health care. This report discusses the Invisible Hand Theorem, which states that
when certain assumptions hold, market outcomes will be efficient. These assumptions require that
no one has an informational advantage over another, that no spillover effects exist in consumption
or production, that no one exerts market power, and that no scale economies exist in production.
Many characteristics of health care markets fail to satisfy the assumptions of the Invisible Hand
Theorem. Moreover, fundamental characteristics of health care (such as informational
asymmetries between patients and health care professionals and between payers and providers, as
well as ethical and distributional concerns) complicate efforts to expand the use of market or
market-like incentives in health care.
Rising health care costs in part reflect the cost of technological advances, whose benefits exceed
their costs, and the aging of the U.S. population. The growing role of third-party reimbursement
over the past half century weakened incentives to minimize costs and thus has also led to higher
health care costs. Many analysts have called for initiatives which would improve the functioning
of health care markets, such as improving consumer information and allowing greater use of
bargaining. These initiatives may help reduce or slow the growth of health care costs, but may
also have unintended negative consequences. Greater use of market-like incentives can improve
the efficiency of the health care system, but only if they take into account the special
characteristics of health care.
Introduc tion ..................................................................................................................................... 1
The U.S. Health Care System..........................................................................................................2
Health Care Costs and Public Spending....................................................................................4
Market Failure and Health Care......................................................................................................5
Externalities and Public Goods.................................................................................................5
Informational Advantages and Market Failure..........................................................................6
The Principal-Agent Problem.............................................................................................6
Adverse Selection and Splintering of Risk Pools...............................................................7
Pressures on the Employer-Provided Health Insurance....................................................13
Changing Incentives for Technological Innovation..........................................................15
Efficiency and Redistribution in Health Care.........................................................................17
The Federal Budget and Market-Oriented Health Care Reform...................................................18
Improve Health Consumer Information..................................................................................20
Information on Pricing......................................................................................................20
Information on Quality.....................................................................................................21
Make Extras Cost Extra..........................................................................................................22
Expand Use of Information Technology.................................................................................23
Author Contact Information..........................................................................................................26
Many analysts, policy makers, and politicians argue that the U.S. health care system would
perform better if market or market-like institutions played a larger role. This view is based on the
belief held by economists that markets generally work most efficiently when left alone, provided
that certain conditions are met. These conditions essentially state that consumption or production
by one person does not affect others, that no one has a privileged position in the market, and that
property rights are well-defined. If those conditions are violated, however, markets may function
inefficiently, which is what economists call “market failure.” Government intervention may
enhance economic efficiency, although in other cases government action may exacerbate market
failure. The special characteristics of health care often lead to market failures. The extent of
government activity vis-a-vis the health system is seen by many economists and health care
analysts as a policy response to the inequalities and inefficiencies associated with such market
Health care professionals often view a market orientation as a threat to their traditions, ethics, and
culture. A bottom-line mentality, they argue, cannot deliver the same high quality of care as the
so-called traditional approach based on professional ethics and responsibilities. Critics such as
Arnold Relman, editor emeritus of the New England Journal of Medicine, denounce the 1
“medical-industrial complex” for its dedication to profitability rather than patient well-being.
Nevertheless, health care institutions have always cared about their pecuniary interests as well as
their patients. As one historian noted, “in many respects [hospitals] have behaved as businesses.
But, ... hospitals have simultaneously carried symbolic and social significance as embodiments of 2
American hopes and ideals.”
Even if some view market forces and the use of the price system to allocate health care services
as an intrusion, consumers and providers of health care are strongly affected by economic
incentives. Hospitals cut the average length of an inpatient stay sharply after Medicare switched 3
from cost-based reimbursement to paying a flat diagnosis-related fee. Physicians increased the
volume (i.e., number of patient visits) and intensity of patient care (i.e., number or complexity of 4
services provided in an average visit) seen after reductions in Medicare Part B payments.
Consumers use less health care when they must pay a larger share of the cost. These behavioral
responses, although predicted by economic theory, do not necessarily enhance economic
efficiency. The challenge for those who wish to expand the use of market incentives in health care
is to design policies that align material incentives facing consumers and providers of health care
so that changes in behavior enhance economic efficiency.
Both friends and foes of the expanding role of markets have sometimes relied on crude ideas
about what markets can or cannot accomplish. This report explains what well-grounded economic
theory has to say about the limits and capabilities of the market in the health care sector. These
1 Arnold S. Relman, “The New Medical-Industrial Complex,” New England Journal of Medicine, vol. 303 (October
1980), pp. 963-970.
2 Rosemary Stevens, In Sickness and Wealth (New York: Basic Books, 1989), p. 6.
3 Louise B. Russell, Medicare’s New Hospital Payment System: Is It Working? (Washington, DC: Brookings, 1989),
4 Centers for Medicare and Medicaid Services, “Physician Volume and Intensity Response,” memorandum from
Volume-and-Intensity Response Team, Office of the Actuary, HCFA, to Richard S. Foster, Chief Actuary, August 13,
1998; available at http://www.cms.hhs.gov/ActuarialStudies/downloads/PhysicianResponse.pdf.
limits and capabilities then outline what the government can or cannot do to improve the health
care system’s performance. The report provides an overview of efforts to expand the use of
market or market-like institutions in health care and considers effects of these initiatives or
proposals on the federal budget. Reforms that are designed to address sources of market failure
have better chances of enhancing economic performance than those that are not. Nonetheless,
improving the performance and efficiency of the health system presents significant policy and
political challenges even to well-designed reforms.
The U.S. health care system is a complex mixture of public and private providers of care, paid for
by a mixture of public and private payers, staffed by dozens of health professions, working in
clinics, hospitals, nursing homes, private offices, health maintenance organizations, work sites,
and home settings, among other venues. This patchwork system provides patients with a broad set
of alternatives; gives health professionals in general, and physicians in particular, a substantial
degree of autonomy; and assigns separate, if often overlapping, responsibilities among various
levels of government. The complexity of the U.S. health system, which gives it considerable
flexibility, is also one of the principal causes of its inefficiencies. Dissatisfaction with the
performance of the U.S. health care system has spurred wider interest in using market or market-
like institutions to generate better results and lower costs.
While the United States is a leader in areas of medical technology, outcomes for several key
public health indicators, such as average longevity and infant mortality, are among the bottom
quarter of the 30 advanced industrial countries that comprise the Organization for Economic 5
Cooperation and Development (OECD). Comparisons of specific conditions or procedures find
the U.S. health system does better in some areas and worse in others. An OECD study of ischemic
heart care found that the United States used more intensive procedures and had lower mortality
rates for older patients, but had higher mortality rates for younger (40-64) patients compared with 6
other OECD countries. Another study used data from Australia, Canada, England, New Zealand,
and the United States to compare quality. Twenty-one quality of care indicators were selected on
the basis of comparability and importance. These researchers found that each country was best on
at least one indicator and worst on at least one indicator. Among these five countries, the United
States did worst on kidney transplant survival rates, second worst on liver transplants, and worst
on incidence of Hepatitis B, but did best for breast cancer survival, incidence of measles, and the 7
cervical cancer screening rate.
The United States spends far more on health care by any measure than any other country in the 8
world. National health expenditures per capita are expected to reach $7,498 in 2007. One of
5 Uwe E. Reinhardt, Peter S. Hussey, and Gerard F. Anderson, “Cross-national Comparisons of Health Systems Using
OECD Data, 1999,” Health Affairs, vol. 21, no. 3 (May/June 2002), pp. 169-181.
6 Pierre Moise, Stéphane Jacobzone, et al., OECD Study of Cross-national Differences in the Treatment, Costs and
Outcomes of Ischaemic Heart Disease, (Paris: OECD, May 2003), p. 8.
7 Peter S. Hussey et al., “How Does The Quality Of Care Compare In Five Countries?” Health Affairs, vol. 23, no. 3
(May/June 2004), pp. 89-99.
8 Gerard F. Anderson, Peter S. Hussey, Bianca K. Frogner and Hugh R. Waters, “Health Spending In The United States
every six dollars spent in the United States is spent on health care, and by 2016, projections 9
indicate that proportion will rise to one in five. According to the latest OECD figures, U.S. health
care spending in 2004 was about a third higher than Switzerland, which had the second highest 10
level of spending. This spending is not due to differences in the proportion of population aged 11
While the United States spends more on health care than other countries, fewer health care
resources per capita are available compared with many other advanced nations. The number of
physicians per capita in the United States is about two-thirds of the OECD average, and the
number of hospital beds per capita is lower than the OECD average, so greater availability of
medical inputs cannot explain the spending difference between the United States and other OECD
countries. Greater intensity and complexity of medical care in the United States may account for
some of the difference. A 1996 McKinsey research project found that health care providers in the
United States were more efficient, in the sense of producing more outputs for a given amount on
inputs, than those in Germany and to some extent than those in the United Kingdom. Many
analysts, however, emphasize other explanations for why medical care in the United States costs
more than in other advanced countries.
Prices of medical inputs and administrative costs are much higher in the United States than in 12
other OECD countries. For example, average remuneration for medical specialists in the United
States, as compared with per capita national income, is higher than for any other OECD country, 13
while the number of specialists per capita in the United States is below the OECD average. A
group of Harvard Medical School researchers estimated that administrative costs accounted for
about a quarter of health care expenditures in 1999. Administrative costs in Canada and Europe 14
appear to be much lower, although the extent of the difference is subject to controversy. Other
factors that help explain differences in health costs between the United States and other countries
And The Rest Of The Industrialized World,” Health Affairs, vol. 24, no. 4 (July/August 2005), pp. 903-914.
9 These projections were calculated by the Centers for Medicare and Medicaid Services. For details, see John A. Poisal,
et al., “Health Spending Projections Through 2016: Modest Changes Obscure Part D’s Impact,” Health Affairs Web
Exclusive, February 21, 2007, available at [content.healthaffairs.org/cgi/reprint/hlthaff.26.2.w242v1].
10 The city-state Luxembourg, which spent $5,089 per person, is excluded. These data were obtained from OECD,
“OECD in Figures 2006-2007 - Health spending and resources from the OECD” (OECD: Paris, February 6, 2007),
available at [ocde.p4.siteinternet.com/publications/doifiles/012006061T02.xls]. Figures computed in U.S. dollars using
purchasing power parity (PPP).
11 Gerard F. Anderson et al., “Health Spending In The United States And The Rest Of The Industrialized World,”
Health Affairs, 24, no. 4 (July/August 2005), pp. 903-914 and Uwe E. Reinhardt, Peter S. Hussey, and Gerard F.
Anderson, “U.S. Health Care Spending In An International Context,” Health Affairs, vol. 23, no. 3 (May/June 2004),
12 McKinsey Global Institute, Healthcare Productivity, October 1996, available at http://www.mckinsey.com/mgi/
13 OECD, Health at a Glance - OECD Indicators 2005, (Paris, March 7, 2006), Chart 2.10, available at
14 S. Woolhandler, T. Campbell, and David U. Himmelstein, “Costs of Health Care Administration in the United States
and Canada,” New England Journal of Medicine, vol. 349, no. 8 (Aug 21, 2003), pp. 768-775. On the other hand, some
argue that the research by Woolhandler and colleagues overstates the magnitude of the difference in administrative
costs between the U.S. and Canada. See Henry Aaron, “The costs of health care administration in the United States and
Canada—questionable answers to a questionable question,” New England Journal of Medicine, vol. 349, no. 8 (August
21, 2003), pp. 801-803.
include differences in payment systems, average national income, average population age as well 15
as other factors.
Health care costs take up a large and growing part of the economy and of public budgets.
Medicare, the largest health expenditure in the federal budget, will cost in total $428 billion in
FY2007, according to CBO baseline projections, or 3.1% of gross domestic product (GDP), and 16
the federal portion of Medicaid is expected to cost $193 billion, or 1.4% of GDP. Medicare Part
A costs per beneficiary grew on average 4.66% a year between 1970 and 2005, and Part B costs 17
grew on average 8.76% a year over the same period. One Congressional Budget Office (CBO)
scenario anticipates that in 2050 Medicare spending will take up 8.6% of GDP, with Medicaid 18
taking up another 4.0% of GDP, even aside from state contributions to Medicaid. Medicaid is
one of the largest and fastest growing components of state spending budgets. In FY2004,
Medicaid accounted for 22.3% of total state spending, which was slightly larger than the 21.4%
spent on elementary and secondary education. The size and rapid growth of Medicaid spending
led the National Governors’ Association and the National Association of State Budget Officers to 19
call the program “the dominant force in state spending.”
Governments in general and the federal government in particular are deeply involved in health
care markets. Public spending in 2007 is expected to account for 47% of national health
expenditures, up from 38% in 1970. Federal spending alone was expected to reach 34% of 20
national health expenditures in 2007, up from 24% in 1970. Furthermore, federal involvement
in health care extends well beyond spending. Health care markets are extensively regulated, and
the federal tax code affects health care markets in important ways. In particular, the cost of the
federal tax exemption for employer-paid health insurance premiums exceeds that of any federal
health program other than Medicare and Medicaid. This exemption will decrease federal revenues 21
by an estimated $100 billion in FY2007.
The pace of health care costs and the expanding public role in health care have stoked interest in
using market incentives to slow or limit costs and to improve quality of outcomes. The nature of
health care, however, provides some inherent limits to the effectiveness of the market.
15 For one analysis, see McKinsey&Company,”Accounting for the Cost of Health Care in the United States,” 2003
available at http://www.mckinsey.com/mgi/publications/healthcare/images/MGI_Healthcare.swf.
16 Congressional Budget Office, The Budget and Economic Outlook: Fiscal Years 2007 to 2016,” January 2006, p. 52
et seq. Medicare figures exclude offsetting receipts, which are expected to total $60 billion in FY2007.
17 Author’s calculation using data from Table V.B1, “HI and SMI Average per Beneficiary Costs” in the 2006 Annual
Report of the Boards of Trustees of the Federal Hospital Insurance Trust Fund and the Federal Supplementary
Medical Insurance Trust Fund and GDP deflator data from the Bureau of Economic Analysis.
18 See the discussion of the intermediate spending/lower revenues scenario in the CBO Report, The Long Term
Outlook, December 2005.
19 National Governors Association and the National Association of State Budget Officers, The Fiscal Survey of States,
June 2006, pp.vii and 10.
20 Poisal, et al., op. cit.
21 U.S. Congress, Senate Committee on the Budget, Tax Expenditures: Compendium of Background Material on
Individual Provisions, committee print, 109th Cong., 2nd sess., S. Prt. 109-072, p. 661. For a comprehensive discussion
of tax benefits for health spending, see CRS Report RL33505, Tax Benefits for Health Insurance and Expenses:
Overview of Current Law and Legislation, by Bob Lyke and Julie M. Whittaker.
Economists’ belief in the efficiency of markets is based on the Invisible Hand Theorem, which 22
states that market outcomes are efficient, so long as certain conditions hold. These conditions
• No externalities. An externality, or spillover effect, exists when one’s
consumption or production affects the ability of another to consume or produce.
Public goods, defined as goods which more than a single person can enjoy at the 23
same time, are a special type of externality.
• Symmetric information. Everyone knows the same things. No one has an
informational advantage over others.
• No market power. No one acts as if he can influence prices through his actions.
• Voluntary trade. Property rights are well-defined and individuals can refuse 24
trades that make them worse off.
The Invisible Hand theorem takes the distribution of buying power as given. However, to the
extent society cares about fairness or the evenness of distribution, it may wish to use taxes and
transfers to alter the distribution of buying power. If transfers and taxes could be made without
economic distortions, then society could move to another distribution of buying power, and the 25
economy would run at full efficiency. In practice, however, all taxes and transfers will cause
some economic distortions. Thus, society alters the distribution of buying power to a more “fair”
allocation, but at some cost of economic efficiency.
An epidemic is a classic example of an externality or spillover effect. Without some form of
coordination, individuals will contribute inefficiently small amounts for the prevention of
contagious diseases. Therefore, an efficient government can improve public well-being by
imposing taxes and spending an appropriate amount on prevention and public health.
Information gained through medical research is an example of a public good, because the same
information can benefit many people simultaneously. Individual donations, however, would yield
an inefficiently low level of support for medical research. Charging those who benefit from better
medical technology or procedures can provide a partial solution; but, to the extent that some
patients will be priced out of the market, economic inefficiencies will persist. What economists
22 Among economists, the Invisible Hand Theorem is called the First Welfare Theorem. See Gerard Debreu, Theory of
Value, (New Haven: Yale, 1959), pp. 94-96 for a proof.
23 More precisely, a good is considered a pure public good if the cost of allowing one more person to enjoy its benefits
is zero. For example, a radio broadcast can be considered a pure public good because the cost of adding one listener is
24 A technical condition on consumer preferences (non-satiation) that rules out “bliss points,” is also required. This
discussion uses the assumption of symmetric information, rather that the assumption that individuals possess perfect
information used in standard proofs. A proof of the First Welfare theorem with symmetric information requires slightly
stronger conditions on consumer tastes.
25 This is the Second Welfare Theorem.
define as public goods are often financed by private individuals or groups. For example, privately
supported basic research is an example of a privately funded public good. Many goods provided
by governments are private goods, not public goods. Extending the benefits of police, fire, and
municipal trash collection to more people generally requires a proportionate increase in resources, 26
so such services would not be considered a public good by economists.
Redistribution of resources to the poor, according to some economists, can also be considered a 27
public good. In this view, each member of society would benefit by the knowledge that all other
members were kept from falling below some minimum standard of living. However, the “warm
glow” that each member might feel from redistribution that ensured that minimum standard often
falls short of an intensity sufficient to induce opening of one’s own pocketbook. Compulsory
taxation provides a way to finance redistribution by sharing the cost among all taxpayers. Many
people would derive some satisfaction from knowing that a health care system fulfilled the ethical
norm that access to medical care was available regardless of ability to pay. Supporting such a
system, as a practical matter, requires government intervention.
The lack of symmetric information, which occurs when some have informational advantages over
others, is the source of many of the key problems in health care. When consumers know the
quality and price of goods, they can search to find higher-quality and lower-priced goods. This
puts pressure on firms to increase quality and reduce price. When consumers cannot see prices or
quality, however, markets work less well. In general, when one party knows more than another
party, market failure can occur. Economists distinguish among several different types of 28
asymmetric information, which are considered as follows.
The principal-agent problem occurs when one person, the principal, must delegate some decision
or activity to an agent, who has special knowledge. If the principal can directly assess outcomes
and effort, then no problem exists. However, if the principal cannot observe how hard the agent
works or if the principal is unable to tell the difference between incompetence and bad luck, then
market failure can result. The interaction of a patient and a physician is an example of principal-
agent relationship. The patient goes to a physician because the physician has special knowledge.
The physician, however, faces different incentives that are not always aligned with the patient’s
best interests. For example, the physician may be paid more by ordering a test whose costs exceed
its benefit. Or the physician may have a patient make additional visits, whose benefits fall short of
their costs. Principal-agent problems can be mitigated in several ways.
A physician’s reputation can provide one solution to the principal-agent problem. If a good
reputation brings benefits to a physician, and if patients occasionally get some information of
26 Tom Means and Stephen Mehay, “Estimating the Publicness of Local Government Services: Alternate Congestion
Function Specifications,” Southern Economic Journal, vol. 61, no. 3 (January 1995), pp. 614-627.
27 Mark V. Pauly, “Income Distribution as a Local Public Good,” Journal of Public Economics, vol. 2 (1973), pp. 35-
28 This discussion follows Kenneth Arrow, “Uncertainty and the Welfare Economics of Medical Care,” American
Economic Review, vol. 53, no. 5 (1963). For a modern analysis of market failure in health care markets see Peter
Zweifel and Friedrich Breyer, Health Economics, (New York: Oxford Univ. Press, 1997), ch. 5.
whether a physician is acting in their best interests or not, then physicians will have an incentive
to avoid actions which would damage their reputations. However, if patients rarely obtain
information on whether a physician acts in their best interest, or if the penalty of losing a good
reputation is too small, then the agents will have weak incentives to be faithful to their principal,
and the principal will use physicians less often.
Centralized tracking of physicians is another possible solution. Centralized mechanisms, such as
licensure requirements, exist to certify that physicians have received appropriate training. The
National Practitioner Data Bank (NPDB) contains information on malpractice claims and
disciplinary actions against health care professionals. Governments, hospitals, health plans, and
related organizations can access specific records, but individuals cannot. Some consumer groups
have attempted to compile their own databases that would allow consumers to assess the quality
of different physicians. So far, these databases are fragmentary and limited.
Third-party monitoring can also give incentives for health care providers to maintain and improve
quality. Accreditation or certification of health care providers gives consumers and payers a seal
of approval for institutions that meet certain standards. The National Committee on Quality
Assurance (NCQA) conducts research on managed care plans and performs accreditation audits,
and the Joint Commission on Accreditation of Healthcare Organizations (JCAHO) accredits
hospitals and related institutions.
Adverse selection occurs when some personal characteristics which affect health care costs are
hidden from others. Pricing of individual insurance contracts can only depend on characteristics
that the insurer knows or can discover, and which can be verified by a third party, such as a court
or an arbitrator. If sicker and healthier persons are indistinguishable, they will pay the same health
insurance premiums. While insurers may not know each individual’s characteristics, they will
have information, based on claims history and other data, about the aggregate characteristics of
various pools of persons seeking insurance. Furthermore, insurers understand that a carefully
designed and marketed menu of insurance plans can induce individuals to separate themselves
into different risk pools. For instance, a low-premium high-deductible plan is more likely to
attract healthier individuals, and a high-premium low-deductible plan is more likely to attract
individuals with greater health care needs. Thus, even if insurers cannot observe individual
characteristics, they can still achieve some separation of high- and low-risk enrollees.
Insurance companies in a competitive market face strong incentives to attract low-risk customers
and avoid high-risk customers, which they can do through underwriting. Underwriting is the
process of assessing levels of risk associated with different insurance contracts and setting terms,
conditions, and prices for those contracts. Underwriting splinters the insured population into
smaller and smaller risk groups, and reduces risk sharing across a broader population. For auto
and property insurance this is standard. Adverse selection is less of a problem because insurers
can predict risk using observable characteristics of drivers and their claims history. Pooling
careful and careless drivers in effect compels careful drivers to subsidize careless drivers. This
pooling would be inefficient if such implicit subsidies caused careless drivers to either drive more
or drive less well, or if resulting increases in premiums caused careful drivers to buy less
insurance. Underwriting by auto insurance companies raises no strong fairness issues to the extent
that the risk of having auto insurance claims is linked to personal choices, including the choice of
whether to drive or not.
Health care is often viewed differently than other kinds of goods, and health insurance has always
worked differently than other lines of insurance. Few would argue that anyone has a right to be a
careless driver. On the other hand, differences in health costs are often presumed to stem from 29
factors which are beyond the control of individuals. The idea that access to health care should
be equal has had enormous influence on health policy. The need to use health care is typically
viewed as a result of bad luck or genetics, rather than carelessness. To the extent that individual
demand for health care is unaffected by insurance status, the costs of providing health care can be
considered a fixed sum. In this case, the practice of medical underwriting, which consists of
offering better prices and conditions to the healthy, rearranges the cost burden of health care but
does not affect overall costs. That is, while an individual insurer earns higher profits by attracting
a healthier risk pool via medical underwriting, total costs are not reduced. Because underwriting
consumes real resources, administrative costs in a system with medical underwriting will be
higher than when all risks are in one pool.
Legal or administrative measures can require insurers to charge the same premiums to individuals
with differing characteristics, which in the context of health insurance is called “community
rating.” In a competitive market, community rating may be unsustainable. If sicker and healthier
persons face the same premiums, sicker persons face stronger incentives to enroll in insurance
plans and to choose more generous health plans. Health plans that attract a higher proportion of
sicker enrollees will have higher average costs. If health insurers are unwilling to sustain losses,
higher average costs lead to higher premiums, which gives healthier individuals incentive to
purchase less insurance relative to a situation in which adverse selection were absent.
A mixed insurance pool may be stable, however, if the proportion of sicker people in a pool is
small enough and if economies of scale make larger plans more efficient. If healthier persons
remain in the same plans as sicker people, then in effect they subsidize sicker persons. If the
proportion of sicker persons is sufficiently large, however, private insurance providers can earn a 30
profit by introducing plans that attract a preponderance of healthier people. As healthier people
leave the mixed insurance pools, average costs for remaining enrollees increase, leaving more
people unable to afford health insurance. This dynamic is often termed a “death spiral.”
The ability to organize large pools of diverse individuals is a central advantage of employer-based 31
health insurance, which has dominated the U.S. health care system for the past half century.
Blue Cross/Blue Shield organizations once adhered to the “community rating” principle that
spread risks across large, heterogenous pools. Community rating was vulnerable to for-profit
insurers’ pricing strategies that offered lower rates for firms with healthier employees, which cut
Blue Cross’s market share. Twelve states still require insurers to use community rating for certain 32
markets. Some see the shift of market share from Blue Cross plans that used community rating
29 For example, few employers offer plans which charge non-smokers lower premiums than smokers, suggesting a
strong reluctance to use experience rating within groups even when cost differences are strongly associated with
30 For a presentation of the economic theory of insurance in the presence of information asymmetries, see Michael
Rothschild and Joseph E. Stiglitz, “Equilibrium in Competitive Insurance Markets: An Essay on the Economics of
Imperfect Information,” Quarterly Journal of Economics, vol. 90, no. 4(1976), pp. 630-49.
31 For an overview of the U.S. health insurance market, see CRS Report RL32237, Health Insurance: A Primer, by
32 Mila Kofman and Karen Pollitz, “Health Insurance Regulation by States and the Federal Government: A Review of
Current Approaches and Proposals for Change,” March 2006, Health Policy Institute working paper, Georgetown
to for-profit insurance plans, which offer lower premiums to healthier groups, as an example of a 33
death spiral. Other researchers, however, contend the introduction of community rating need not 34
result in a death spiral. Furthermore, the financial problems of some Blue Cross organizations 35
may have had more to do with administrative problems than with adverse selection.
The tax exemption for employer-provided health care was a major factor in the expansion of
employee-based health insurance. While linking health insurance with employment has
advantages of low administrative costs and broad pooling of risks, the tax exemption gives the
largest subsidies to those with the most generous health plans, and for employees with a choice of
plans, encourages the choice of more generous plans. Tying health insurance to employment,
which the tax exemption encourages, can discourage employees from switching jobs, a problem
know as “job lock.” The President’s Advisory Panel on Tax Reform recommended capping this 36
Organizing government-sponsored risk pools is another way to ensure that the risks of incurring
major health care costs are spread across a large population. In single-payer systems, such as
Canada’s, all eligible persons are in the same pool. Putting all Canadians in a single pool and the
ban on private health insurance prevents any splintering of the health insurance market that could
trigger a death spiral, and ensures that the health care costs of the sickest patients are borne by the
A more limited approach is to set up risk pools, which allow those with serious medical problems 37
to obtain health insurance. As of 2005, 33 U.S. states had set up risk pools that offer insurance
to individuals denied insurance due to an existing medical condition. Despite subsidies from state
and federal funds, risk-pool premiums are much more expensive than premiums paid by healthy 38
individuals who have employer-provided insurance. In 2004, state subsidies totaled more than
$0.5 billion, and premiums typically cost 125% to 150% of comparable individual market
premiums. The number of enrollees in risk pools (about 180,000 in 2004) comprises only a small 39
fraction of the pool of uninsured.
Adjusting payments to plans and providers based on characteristics of enrollees, at least in theory,
can provide incentives for private insurers to treat a pool of patients more efficiently rather than
to focus on attracting a healthier pool of patients. If insurers were paid less for covering healthy
33 Paul Starr, The Social Transformation of American Medicine (New York: Basic, 1983), pp. 327-331.
34 Thomas C. Buchmueller and John E. DiNardo, “Did Community Rating Induce an Adverse Selection Death Spiral?
Evidence from New York, Pennsylvania and Connecticut,” American Economic Review, vol. 92, no. 1. (March 2002),
35 James C. Robinson, “The Curious Conversion of Empire Blue Cross,” Health Affairs, vol. 22, no. 4 (July/August
2003), pp. 100-118.
36 President’s Advisory Panel on Tax Reform, Simple, Fair, and Pro-Growth: Proposals to Fix America’s Tax System:
Report of the President’s Advisory Panel on Federal Tax Reform, November 2005, pp.80-81. In particular, both the
“Simplified Income Tax Plan” and the “Growth and Investment Tax Plan” recommend that this exemption be capped at
the average health premium, which the Panel estimated at $5,000 for individuals and $11,500 for families.
37 For state-level details, see CRS Report RL31745, Health Insurance: State High Risk Pools, by Bernadette Fernandez.
38 Mark Merlis, “Fundamentals of Underwriting in the Nongroup Health Insurance Market: Access to Coverage and
Options for Reform,” National Health Policy Forum Background Paper, April 13, 2005, pp. 19-21. A list of state risk
pools is available at http://www.healthinsurance.org/riskpoolinfo.html.
39 Bruce Abbe, “Overview—State High Risk Health Insurance Pools Today,” available at
patients and more for covering sicker patients, insurers would have a weaker incentive to attract
the healthy and avoid the sick. For instance, Medicare Advantage sets rates for its HMO
capitation program based on enrollee characteristics. So far, however, Medicare Advantage (MA)
uses crude rules of thumb to set reimbursement adjustments, which do not appear to have
changed incentives facing insurers in significant ways. Medicare administrators are currently
exploring more sophisticated ways of reimbursing MA providers.
Designing reimbursement schemes to change insurers’ incentives is difficult for two reasons.
First, health expenditures are highly skewed among the population: 5% of patients account for 40
over 50% of health care costs. Thus to alter insurers’ behavior, substantial pricing adjustments
would be needed. Second, insurers are likely to have better information about health
characteristics of enrollees and potential enrollees, as well as more sophisticated methods of
analyzing those characteristics. To the extent that payment adjustments misprice risk, private
insurers can profit by medical underwriting.
Providing people with the opportunity to extend or renew their health insurance coverage can
provide some protection against the effects of splintering risk pools. If insurance plans are subject
to frequent renewal decisions that may depend on past claims history, then insurance becomes
less of a shield against financial calamity and more of an installment plan. Congress and several
states have enacted reforms intended to preserve enrollees’ ability to renew coverage. The
Consolidated Omnibus Budget Reconciliation Act of 1985 (COBRA, P.L. 99-272) contains
provisions which allow retirees, former employees, and dependents to pay no more than 102% of
the full, normal premium to continue coverage. Terminated employees may generally obtain
COBRA coverage for 18 months. In some circumstances beneficiaries can obtain COBRA
benefits for longer periods. Because employers typically contribute a portion of the cost of
premiums, obtaining insurance under COBRA almost always costs former employees much more
than what current employees pay for health insurance. COBRA policies, however, cost less than 41
individual health insurance plans, with the possible exception of younger and healthier persons.
On average, employees pay 16% of premium costs for individual coverage and 27% for family 42
The Health Insurance Portability and Accountability Act of 1996 (HIPAA, P.L. 104-191), also
known as the Kassebaum-Kennedy bill, restricted preexisting condition exclusions and limited
insurers’ ability to deny coverage to people who recently had group coverage. HIPAA also
guaranteed eligible employees’ ability to renew coverage and to carry over coverage when 43
changing employers, a provision intended to reduce job lock. HIPAA, however, set no limits on
premium increases, so insurers could discourage HIPAA-eligible persons from enrolling by 44
setting premiums at high levels.
40 M. L. Berk and Alan C. Monheit, “The Concentration of Health Care Expenditures, Revisited,” Health Affairs, vol.
20, no. 2 (March/April 2001), pp. 9-18.
41 U.S. Dept of Labor, Employee Benefits Security Administration, “FAQs About COBRA Continuation Health
Coverage,” available at http://www.dol.gov/ebsa/faqs/faq_consumer_cobra.html.
42 Kaiser Family Foundation, Employer Health Benefits 2006 Annual Survey, September 2006, sec 6.
43 For an account of the passage of this bill, see Brian K. Atchinson and Daniel M. Fox, “The Politics Of The Health
Insurance Portability and Accountability Act,” Health Affairs, vol. 16, no. 3 (May/June 1997).
44 Merlis, op. cit., pp. 13-14.
Almost all state governments have enacted reforms to limit experience rating of small group
health insurance, and 21 states limit experience rating for nongroup health insurance. Some states
also require insurers to issue policies, and many more have extended COBRA and HIPAA 45
provisions regarding continuation of health insurance coverage.
Despite various federal and state measures intended to limit experience rating and expand health
insurance coverage, many people, especially those with preexisting conditions, have trouble 46
finding affordable health insurance. The health insurance trade association reported in 2002 that 47
Moral hazard occurs when a person’s actions are unobservable, so that changes in behavior 48
cannot be observed. In particular, moral hazard refers to changes in behavior that affect the risks
being insured. For example, after a person obtains health insurance she may take less care to
remain healthy or may visit her physician more often than necessary. Moral hazard in health
insurance may be limited by non-insurable costs. For instance, going to the physician takes time,
having procedures performed is painful, and neglecting one’s health can cause problems which
medical care cannot easily cure.
Copayments and deductibles are standard methods to limit moral hazard. According to economic
theory, insurance policies that divide risk-sharing between an insurer and the insured in an
efficient way require individuals to pay for relatively small, regularly occurring expenses, such as
eyeglasses and routine dental care and require insurers to pay a major part of large, unexpected
expenses. The value of insurance is higher for rare or unusual expenses which would cause
serious financial disruption for a household. Small and routine costs present lesser financial risks
to policyholders, so the benefit of insuring such expenses is minor. Thus economic theory
suggests that if consumers are informed and rational then insurance with “doughnut” provisions,
which cover small- and high-cost claims, but which fail to cover claims over some intermediate 49
range, is an inefficient means of sharing risks. The Medicare Part D drug coverage is one
example of health insurance with a doughnut provision. On the other hand, the Medicare
Catastrophic Coverage Act of 1988 (P.L. 100-360), which introduced changes in the Medicare
system arguably more in line with standard economic theory, was subsequently repealed in 1990 50
after vigorous protests by some Medicare beneficiaries.
45 Merlis, op. cit., pp. 15-16.
46 Merlis, op. cit., pp. 10-12.
47 Thomas Musco and Thomas Wildsmith, “Individual Health Insurance: Access and Affordability,” Health Insurance
Assoc. of America Brief Analysis, October 2002, available at http://www.heartland.org/Article.cfm?artId=15320.
48 Mark V. Pauly, “Overinsurance and Public Provision of Insurance: The Roles of Moral Hazard and Adverse
Selection.” Quarterly Journal of Economics, vol. 88 (1974), pp. 44-62.
49 Arrow (1963), op. cit., Appendix. Price distortions introduced by the tax exemption for employer-provided health
insurance or the impression that individuals are unable to compare accurately the costs and benefits of health care may
explain why many policies cover routine or preventative care.
50 Thomas Rice, Katherine Desmond, and Jon Gabel, “The Medicare Catastrophic Coverage Act: a Post-mortem,”
Health Affairs, vol. 9, no. 3(Fall 1990), pp. 75-87. The act ensured that all costs of a hospital stay beyond an initial
deductible of about $500 and all reasonable physicians’ bills beyond an annual limit of about $1,400 would be covered
for Medicare beneficiaries.
Empirical evidence shows that patients who do not pay copayments and deductibles receive more
health care. The RAND Health Insurance Experiment, which randomly assigned families to
insurance plans, found that free care encouraged health care use. Families that paid nothing for
their health care were about 10% more likely to use medical care, and incurred about 25% more
medical expenses than families with 25% copay plans. Families with 50% and 95% copay plans 51
were less likely to use medical care and incurred fewer medical expenses. Health status for
those receiving free care was little different than health status for those in copay plans, with two
exceptions. First, low-income people with poor vision and free care had slightly improved vision
compared with others. Second, low-income populations with free care had fewer problems with
high blood pressure. Within that high-risk group, better control of high blood pressure appeared to 52
lessen mortality risks. Another study, using data from the Medical Outcomes Study, found that
older patients with chronic conditions who had zero or low copayments consumed more medical 53
care than those with high copays.
Requiring families to pay a portion of their health care costs, according to the usual theory of
supply and demand, reduces demand for low-benefit care. To the extent that costs of some types
of care exceed their benefits, excluding such items is a more efficient way to design insurance
plans. Demand for health care items whose benefits greatly exceed their costs should be affected
to a lesser degree, if patient behavior conformed to standard economic assumptions. There is
mixed evidence, however, that higher copayments have different effects on low-benefit and high-
benefit care. The RAND Health Insurance Experiment and the Medical Outcomes Study found
that higher patient copays and deductibles reduced use of low-benefit items, but also reduced
consumption of some high-benefit items as well. That suggests demand for health care responds
to monetary incentives, but that patients often have difficulty in distinguishing high-benefit and
low-benefit care, or that some portion of the patient pool responds to monetary incentives and
another portion does not. In some instances, changes in what patients pay can induce reductions
in low-benefit care without significantly affecting high-benefit care. For instance, a 1996 study
found that requiring Kaiser-Permanente HMO enrollees to pay $25 to $35 for emergency room
visits reduced overall emergency room visits by 15%, while the reduction in emergency room 54
visits for conditions classed as “always an emergency” was small and statistically insignificant.
The extraordinary cost of a major medical intervention presents a difficult dilemma to those who
design cost-sharing provisions in health insurance plans. While few patients incur huge charges,
that small percentage of cases accounts for a large proportion of total health care costs. On one
hand, a health insurance plan has limited value if it does not prevent a health emergency from
becoming a financial calamity. Forcing families to pay even a fraction of the costs of an
expensive medical episode could strain the finances of most families, even to the point of 55
bankruptcy. On the other hand, if insurance pays all, or nearly all, of the charges associated with
a major health episode, then patients, their families, and their physicians have little incentive to
51 W. G. Manning et al, “Health Insurance and the Demand for Medical Care,” American Economic Review, vol. 77,
no. 3 (June 1987).
52 R.H. Brook et al., “Does Free Health Care Improve Adults Health?” New England Journal of Medicine, vol. 309
(1983), pp. 1426-34.
53 Mitchel D. Wong et al., “Effects of Cost Sharing on Care Seeking and Health Status: Results from the Medical
Outcomes Study,” American Journal of Public Health, vol. 91, no.11 (November 2001), pp. 1889-1894.
54 Joe V. Selby et al., “Effect of a Copayment on Use of the Emergency Department in a Health Maintenance
Organization.” New England Journal of Medicine, vol. 334, no. 10 (March 7, 1996), pp. 635-642.
55 David Himmelstein, Elizabeth Warren, Deborah Thorne, and Steffie Woolhandler, “Illness and Injury As
Contributors To Bankruptcy,” Health Affairs, vol. 24 (2005), pp. 63-73.
control costs, even if the resulting benefits are minimal.56 One study found no correlation between
spending on care for terminal patients and regional mortality rates, suggesting that either higher
spending yields little or no benefit in mortality or that data aggregated by region are too crude to 57
identify effects. Many plans have out-of-pocket limits that commit the insurer to pay all charges
above a certain level, which provides families with substantial protection against financial
calamity at the cost of eliminating monetary incentives to avoid items or procedures with only
low or speculative benefits. In some cases, administrative and financial incentives may induce
terminal patients to be treated in high-cost settings, even when patients have expressed a desire to 58
be treated in a low-cost hospice or home setting.
Imposing tighter controls on access to medical care by imposing gatekeeper requirements is
another way to limit moral hazard. Most managed care plans require a primary care physician or
insurance plan representative to approve hospital visits, procedures, and specialist visits.
Managed care plans substantially increased their share of the health insurance market in the late
1980s and early 1990s, a development driven in part by the belief that managed health care and
health maintenance plans could control costs. Some research suggested that cost growth 59
moderated in markets where managed care plans had larger market shares. Other evidence 60
suggests that managed care plans do not deliver lower costs. In the late 1990s, consumer
dissatisfaction slowed or even reversed the expansion of managed care, although HMO and
managed care plans avoided losing market share by loosening controls on access to care and by 61
expanding into the Medicare population.
Employer-based health benefits, the bulwark of the American health insurance system, is
increasingly under strain, threatening the health insurance system’s ability to spread the financial
risks associated with significant medical problems. The number of people covered by employer-
provided health insurance has been dropping since 2000, as insurers have been willing to trade
higher margins for lower enrollments. More people are buying individual health plans, but this 62
increase is dwarfed by the decrease in group coverage. Many firms have sought to push a
greater share of health cost increases onto employees and retirees, which has become a major
56 One solution that preserves the value of insurance while introducing price incentives would be to allow families to
receive cash in lieu of aggressive treatment. See Margaret M. Byrne and Peter Thompson, “Death and Dignity:
Terminal Illness and the Market for Non-treatment,” Journal of Public Economics, vol. 76, no. 2 (May 2000), pp. 263-
57 J. Skinner and John E. Wennberg, “How Much is Enough? Efficiency and Medicare Spending in the Last Six Months
of Life,” National Bureau of Economic Research Working paper 6513, 1998.
58 Melinda Beeuwkes Buntin and Haiden Huskamp, “What Is Known About the Economics of End-of-Life Care for
Medicare Beneficiaries?” The Gerontologist, vol. 42, special issue III (2002), pp. 40-48.
59 Jack Zwanziger and Glenn A. Melnick, “Can Managed Care Plans Control Health Care Costs?” Health Affairs, vol.
15, no. 2 (summer 1996), pp. 185-199.
60 J. Sung, M. Wessel, S.F. Gallagher, J. Marcet, M.M. Murr. “Failure of Medicare Health Maintenance Organizations
to Control the Cost of Colon Resections in Elderly Patients,” Archives of Surgery, vol. 139, no. 12 (December 2004),
61 M. Susan Marquis, Jeannette A. Rogowski, and José J. Escarce, “The Managed Care Backlash: Did Consumers Vote
with Their Feet?” Inquiry, vol. 41, no. 4, pp. 376-390.
62 James C. Robinson, “The Commercial Health Insurance Industry in an Era of Eroding Employer Coverage,” Health
Affairs, vol. 25, no. 6 (November-December 2006), pp. 1475-1486.
source of labor-management conflict.63 Health benefits remain a standard feature of benefit
packages for nearly all businesses with more than 200 employees and for firms that tend to 64
employ high-wage employees. High-wage employees are more likely to fall within higher
marginal tax brackets, and therefore gain more from the tax exemption of employer-provided
health insurance than low-wage employees. For instance, a high-wage employee whose income
puts her in the 35% marginal tax bracket saves 35¢ in taxes for every dollar shifted from post-tax
compensation to health insurance paid in pre-tax dollars. A low-wage employee who pays no
federal income tax gains no federal income tax advantage from shifting compensation to health 65
benefits. Without some curtailment of the employer-provided health insurance tax exemption,
medium and large businesses and firms that tend to hire high-wage individuals will continue to
offer health benefits.
Smaller businesses and firms that employ large numbers of low-wage workers, who gain less
from tax exemptions, face a stronger temptation to drop or curtail health insurance benefits as
health premiums rise. The percentage of small businesses that offer health benefits dropped from 66
69% in 2000 to 60% in 2005. Low-wage workers may be more willing to take the risk of going
without health insurance in exchange for higher take-home pay, although low-wage individuals
are more likely to have poor health status.
For example, in 2005 fewer than half of Wal-Mart employees were in its health insurance plan.
While some worked too few hours or had not worked long enough to be eligible, many others 67
passed up coverage because of the high cost of benefits relative to their earnings. An internal
2005 Wal-Mart memorandum stated that significant proportion of workers (27%) obtained
coverage for their children through public insurance programs such as Medicaid and State
Children’s Health Insurance Program (SCHIP), and that an additional 19% had children who were 68
uninsured. Walmart employees and their children were less likely to have employer-provided
insurance and more likely to be covered by public insurance than the national average for all
workers, but were more likely to have employer-provided insurance and less likely to be covered
by public insurance than the average retail worker. In April 2006, in the face of pressure from
unions and several state legislatures, Wal-Mart announced changes in its benefits package
intended to increase the attractiveness of its health insurance plan, which shortened waiting 69
periods and expanded availability of health benefits.
63 Margaret Ann Cross, “Rising Costs Strike Unions As Being Cause for Unrest,” Managed Care, May 2003.
64 Kaiser Family Foundation/Health Research and Educational Trust 2005 Annual Employer Health Benefits Survey
(Kaiser/HRET) available at http://www.kff.org/insurance/chcm091405nr.cfm.
65 Shifting compensation from take-home pay to employer-provided health care would reduce payroll taxes for Social
Security and Medicare, but would also reduce future Social Security benefits in most cases.
67 Steven Greenhouse and Michael Barbaro, “Wal-Mart Memo Suggests Ways to Cut Employee Benefit Costs,” New
York Times, October 26, 2005. See “Wal-Mart Announces Additional Health Benefits Improvements and Timeline,”
available at http://www.walmartfacts.com/articles/1650.aspx. Under Wal-Mart’s benefit plan introduced in April 2006,
full-time associates become eligible for health benefits in six months, and part-time associates becomes eligible in one
68 Susan Chambers, “Reviewing and Revising Wal-Mart’s Benefits: Strategy Memorandum to the Board of Directors,”
2006, Exhibit 3, available at http://www.nytimes.com/packages/pdf/business/26walmart.pdf. Wal-Mart claims that
about 80% of their employees have health insurance coverage, either through family members or public insurance
programs such as Medicaid or Veterans’ Health Administration. See http://www.walmartfacts.com/FactSheets/
69 Wal-Mart Fact Sheet, “What’s the Story about Wal-Mart Health Care Benefits?” available at
Public and private health insurance systems do not smoothly conjoin to offer low-income
individuals and their families stable and predictable financial protection against medical costs. In
addition, the interaction of public and private insurance programs can distort incentives for low-
wage individuals and the firms that employ them. Although enrollments in Medicaid and SCHIP
increased in the 1990s, in recent years some states have tightened eligibility standards while other 70
states have expanded coverage. Many low-income families lose their eligibility for Medicaid
when their incomes rise above state-specific earnings thresholds, which imposes a high implicit 71
marginal tax rate on earnings of those families, serving as a deterrent to work. Some firms that
tend to hire low-wage workers have trouble offering insurance plans as attractive to low-income
families as public insurance programs supported by tax dollars. Firms that offer low-wage
workers health benefits find that through the tax system they pay health insurance costs of their
competitors’ employees as well as their own. In addition, waiting periods for benefits and other
eligibility hurdles impose barriers to insurance benefits and health care access for low-wage
workers, who change jobs more frequently and are more subject to economic and social
disruptions than workers with greater financial resources.
Most health economists and practicing physicians, according to one survey, believe that “the
primary reason for the increase in the health sector’s share of GDP over the past 30 years is 72
technological change in medicine.” In other sectors of the economy, most notably those
involving computers and information technology, technological advances brought better and
cheaper products. Asking why technology causes higher costs in health care but lower costs
elsewhere is natural. Part of the answer lies in how incentives facing developers of new medical
technologies interact with the structure of health care finance.
The direction and pace of advances in medical technology depend on incentives facing engineers
and researchers, which are affected by how health care is financed. Inventors and developers, if
motivated by profits, must consider who will pay for new technologies and innovative products.
The dominant role of insurers and governments in health care finance implies that converting
breakthroughs in medical technology into financial success generally depends on private and
public insurers’ decisions about what their plans cover. Defining coverage limits of health
insurance presents special challenges. While health insurance contracts in a given year specify
what is and what is not covered in considerable detail, over time coverage limits evolve as
medical practice patterns and medical technology progresses. As medical knowledge and
technology advance, what was considered experimental medicine yesterday becomes standard
70 Donna C. Ross and Laura Cox, “In a Time of Growing Need: State Choices Influence Health Coverage Access for
Children and Families,” Kaiser Commission on Medicaid and the Uninsured Report, October 2005, available at
h t t p :/ / www. k ff. o rg/ med i cai d / u p l oad / In -a-Ti me-o f- Gro wi n g-Need - S t at e-Ch o i ces-In fl u en ce-Heal t h -Co verage- Acce ss-
71 Medicaid eligibility rules are complex and vary among states. For a description of Medicaid eligibility rules for the
non-elderly, see CRS Report RL33019, Medicaid Eligibility for Adults and Children, by Jean Hearne.
72 Victor R. Fuchs, “Economics, Values and Health Care Reform,” American Economic Review, vol. 86, no.1 (March
1996), p. 8. A minority of economic theorists, however, believed technological change was to blame for increasing
today.73 The knowledge that what health insurance covers shifts over time has an important effect
on incentives facing those working to develop new drugs, devices, and procedures.
Because health insurance makes patients and physicians less sensitive to price, developers have
stronger incentives to invent new treatments or technologies which do something new or better,
rather than invent cheaper ways of doing existing things. Innovation in computer hardware
technology is mainly focused on making cheaper computers that run faster. If firms and
individuals had “computer insurance,” computers might perform more esoteric tasks, but at a
much higher price.
One proposed strategy is to design health insurance plans that will pay an amount equivalent to
the cost of an existing technology. If a patient wanted a newer and better technology, then the
patient would have to pay out of her pocket. If the new technology’s benefits warranted its higher
costs, then presumably the patient would be willing to pay the extra amount. For example, an
insurer could set reimbursement for pharmaceuticals aimed at a specific condition equal to the
cost of an existing drug of known efficacy. If a drug company developed a more effective drug
for that condition that was more expensive, then the patient would pay the difference. If the gain
in efficacy was large compared to the increase in price the consumer would presumably be more
willing to choose that drug. This reimbursement policy would provide developers of new medical
technologies with a powerful incentive to make newer products better and cheaper.
While fixing reimbursement levels for new pharmaceuticals at the level of existing approved
drugs would give drug developers strong incentives to consider the costs and prices during the
R&D process, it could also create pricing anomalies in the short run. For instance, a study based
on data from the Clinical Antipsychotic Trials in Intervention Effectiveness (CATIE) found no
significant differences in quality-of-life measures for second-generation drugs and a first-74
generation drug (perphenazine), even though newer drugs cost $300 to $600 more per month.
Thus, for some patients, switching to an older antipsychotic could yield large cost savings.
However, different patients with the same condition often respond differently to the same drug,
and patients with schizophrenia often must try several different drugs to find one that is clinically
effective and does not create serious side effects. Some patients will do better with a second-
generation drug, while other patients with the same condition will do better with a first-generation
drug. If reimbursement levels of second-generation drugs were set at the level of first-generation
drugs, either some patients would pay substantially more (or someone would pay more on their
behalf) for their treatment using a second-generation drug, or they would have to use a first-
generation drug which, for them, works less well.
Health care providers often have substantial market power. Shopping around for the most
attractive health provider when sick or injured is difficult or impossible. Switching physicians or
health plans is costly and inconvenient for patients, and changing health insurers is costly and
time-consuming for businesses. Prices for individual health services are difficult to find. Drug
73 Burton A. Weisbrod, “The Health Care Quadrilemma: An Essay on Technological Change, Insurance, Quality of
Care, and Cost Containment,” Journal of Economic Literature, vol. 29, no. 2 (June 1991), pp. 523-552.
74 Robert A. Rosenheck, Douglas L. Leslie, Jody Sindelar, et al., “Cost-Effectiveness of Second-Generation
Antipsychotics and Perphenazine in a Randomized Trial of Treatment for Chronic Schizophrenia,” American Journal
of Psychiatry, vol. 163 (December 2006), pp. 2080-2089.
and device manufacturers have legal monopoly powers due to patent protection. While major
buyers, such as governments, have bargaining power which can allow them to buy at lower
prices, individuals have little or no bargaining power.
In other markets consumers can, in effect, hire a firm or organization to bargain on their behalf. A
Wal-Mart shopper enjoys the benefits of the company’s bargaining power with manufacturers. A
union member enjoys the benefits of collective bargaining, such as better working conditions and
higher pay. Similarly, a patient benefits from the bargaining power of his employer with his 75
insurer, as well as from the insurer’s bargaining power with health care providers. Likewise, a
retiree can benefit from the government’s bargaining power. Lower prices for health care in other
OECD countries are, to a large extent, due to the willingness of governments to use their 76
bargaining power with providers. If consumers have a wide choice of organizations or firms that
can negotiate on their behalf, then competition will ensure that consumers will reap most of the
benefits. If consumers cannot easily choose or switch among such organizations, however, then
the middlemen will capture a larger portion of those bargaining benefits for themselves.
Moreover, while the bargaining power of employers, insurers, and governments benefits
consumers, as compared to a situation in which consumers face providers directly, bargaining
interactions among employers, insurers, and providers create economic distortions which can 77
Economists tend to separate questions of efficiency from questions of redistribution. If market
failure occurs, market outcomes are inefficient in the sense that other outcomes exist that would
make some people better off without making anyone else worse off. Even if markets are efficient,
society may decide to redistribute resources, despite efficiency losses. Designing policy,
according to mainstream view of public economics, is a matter of achieving a given set of 78
distributional goals with minimal loss of efficiency.
The concept of social insurance, which is embedded in the design of Social Security and much of
federal health policy, provides a way to redefine the distinction between efficiency and
redistribution. For instance, Social Security benefits resemble annuities offered by private
insurers. A retiree who purchases an annuity pays for the right to receive regular payments until
death. Annuities therefore, insure retirees against the risk of having their assets run out before
they die, and allow retirees to reallocate income to match over time financial resources to needs.
An actuarially fair annuity (i.e., one whose expected benefits match its cost for each category of
risk) cannot be properly said to redistribute income, even though annuity holders who live longer
receive more benefits than those who do not.
75 Henry J. Aaron, “A Funny Thing Happened on the Way to Managed Competition,” Journal of Health Politics, Policy
and Law, vol. 27, no. 1, February 2002.
76 Gerard F. Anderson, Uwe E. Reinhardt, Peter S. Hussey, and Varduhi Petrosyan, “It’s The Prices, Stupid: Why The
United States Is So Different From Other Countries,” Health Affairs, vol. 22, no. 3 (May/June 2003), pp. 89-105.
77 Mark V. Pauly, “Managed Care, Market Power, and Monopsony—Examining the Role of Regulation in an Evolving
Healthcare Marketplace,” Health Services Research, vol. 33, no. 5 pt. 2 (December 1998), pp. 1439-1460.
78 In some cases, treating individuals more equally may enhance economic efficiency. If incomes depend in some part
on luck or other random causes, then a social insurance policy that reduces income inequalities can enhance economic
efficiency. For details see Gareth D. Myles, Public Economics (New York: Cambridge), pp. 6-7, 470-484.
Unlike private annuities, Social Security payments are determined by a progressive benefit
schedule, so that retirees who had low wages receive more relative to the payroll taxes they paid
than what retirees who had high wages receive. Thus, the progressive structure of Social Security
benefits tends to transfer income from richer to poorer retirees, in a way that a private insurer
would not. While this could be viewed as pure redistribution, the progressive structure of Social
Security benefits can be seen as implicitly providing insurance against the risk of having low
earnings. To the extent that this socially provided insurance reduces the riskiness of individuals’ 79
incomes over their lives, economic efficiency is enhanced.
A broader interpretation of the social insurance concept would be that redistributional programs
aimed at the poor, act as an insurance program that all members of society join before birth. Some
part of regular taxes constitutes the premium for this social insurance, and individuals collect
benefits if they suffer bad luck of some sort that makes them poor. Social insurance provides
protection against some consequences of becoming poor, at the cost of higher taxes for others.
The introduction of Medicare and Medicaid in 1965 reflected both social insurance and more
purely distributional concerns. Medicare, which reduced the share of the elderly living in poverty,
added medical benefits to the Social Security program. By including nearly all of the working
population in Medicare, its original proponents sought to distinguish it from more redistributive
“welfare” programs and emphasize its social insurance aspects. In contrast, while Medicaid also
provides substantial social insurance benefits, its design continues to reflect its origin as a
redistributive program. Medicaid stemmed from various poor-relief programs aimed at low-
income groups and those impoverished by poor health, which were extended and consolidated in 80
a multi-billion-dollar federal-state program.
The rapid rise in health care costs is straining public budgets at all levels of government, and 81
projected increases in health care costs promise to intensify pressures on public budgets. If
health care costs continue to grow at past rates, cutbacks in other types of consumption will be
inevitable. Even if health care costs moderate, a substantial portion of the gains from economic
growth will be directed towards the health care system. Rising costs have stemmed in part from 82
the introduction of medical advances, which have increased longevity and reduced morbidity.
The increasing proportion of the elderly population is another cause of rising medical costs.
Finally, third-party reimbursement, by shielding consumers from health care costs, makes
consumers less sensitive to price signals. This, in turn, gives providers incentives to expand the
79 For a detailed analysis of the social insurance character of Social Security, see Peter Diamond, “Social Security,”
American Economic Review, vol. 94, no. 1 (March 2004), pp. 1-24.
80 See Paul Starr, The Social Transformation of American Medicine (New York: Basic Books, 1983) for a description
of the enactment of Medicare and Medicaid.
81 See point 10 of the United States of America—2006 Article IV Consultation, Concluding Statement of the IMF
Mission, May 31, 2006, available at http://www.imf.org/external/np/ms/2006/053106.htm and the GAO report, 21st
Century: Addressing Long-Term Fiscal Challenges Must Include a Re-examination of Mandatory Spending, GAO-06-
456T, February 15, 2006.
82 David M. Cutler, Your Money or Your Life, (New York: Oxford Univ. Press, 2004).
supply of health care and to change practice patterns to a more complex and intensive style of 83
Rising health care costs have cut into the growth of other types of consumption for most
households. In the decades following World War II, productivity and incomes grew fast enough
relative to health care costs to allow steady increases in health and non-health spending. In more
recent decades, real incomes for most households grew more slowly while health costs continued
to rise rapidly and cut into the growth of non-health expenditures. Since 1972, real incomes for
the lower 99% of household grew on average only 1.2% per year to 2002. Over the same period 84
real national health expenditures rose 4.9% a year. Had health care spending in the U.S. been
held to 10% of GDP, a proportion similar that in Canada, France, Germany, and Switzerland, non-85
health consumption would have grown about 25% faster.
To the extent that choices of rational consumers, or of rational voters and politicians, drive this
expansion of the health care system, rising health costs are not necessarily a cause for concern. In
all advanced industrial countries the fraction of the economy devoted to health care has been
rising. If consumers prefer to buy more technologically advanced medical care rather than more
advanced cars or refrigerators, then higher medical costs are a natural consequence of rising
standards of living. A 2001 review of studies regarding medical technology for heart attacks, low-
birthweight babies, depression, and cataracts indicated that increased benefits of better treatment
options far outweigh costs. For breast cancer, according to that review, increased costs and 86
benefits of new technologies were roughly of the same magnitude.
Rising medical costs threaten to price a growing number of Americans out of the insurance
market. In the past few decades, the number of uninsured has hovered around 40-46 million.
Depending on the definitions used, tens of millions more are underinsured against the risk of
having a health emergency become a financial catastrophe. As health care becomes more
expensive, the logic of supply and demand suggests the pool of uninsured persons will grow. Two
health economists projected that the number of uninsured persons will grow from 45 million in 87
83 Amy Finkelstein, “The Aggregate Effects of Health Insurance: Evidence from the Introduction of Medicare”
Quarterly Journal of Economics, vol. 122, no. 1 (2007).
84 Income data from Table A1 in the April 2006 update of tables for Emmanuel Saez and Thomas Piketty, “Income
Inequality in the United States, 1913-1998,” Quarterly Journal of Economics, vol. 118, no. 1 (2003), pp. 1-39,
available at http://elsa.berkeley.edu/~saez/TabFig2004prel.xls. According to the Saez and Piketty data, real gross
income (i.e., income before individual income taxes and individual payroll taxes but after employers’ payroll taxes and
corporate income taxes and excluding transfers) per tax unit was roughly the same in 1972 and in 2002. However, the
number of tax units per household rose, the Consumer Price Index (CPI) used to adjust for inflation overstates price
changes, and transfers to households rose. Combining these effects gives a 40% increase in real consumption per
household over this period. National health expenditures from the Centers for Medicare & Medicaid Services, Office of
the Actuary. Prices adjusted using the Bureau of Economic Analysis’s implicit GDP price deflator, which is less
subject to the biases of the CPI.
85 Calculations based on the following data: real income growth for households outside of the top 1% was 40% between
1972 and 2002, according to Piketty and Saez, and the proportion of national health care was 7.5% in 1972 and 15.3%
in 2002, according to the Centers for Medicare and Medicaid Services.
86 David M. Cutler and Mark McClellan, “Is Technological Change in Medicine Worth It?” Health Affairs, vol. 20, no.
5 (September/October 2001), pp. 11-29.
87 Todd Gilmer and Richard Kronick, “It’s The Premiums, Stupid: Projections of the Uninsured Through 2013,” Health
Affairs, Web Exclusive, April 5, 2005.
Absent the political will to increase taxes significantly, rising federal health care expenditures will 88
force major cuts in benefits or fundamental changes in the health delivery system. Economic
theory suggests that addressing the root causes of market failure provides the best chance for
enhancing system performance. The following discussion examines the potential of market-based
solutions, which address sources of market failure noted above, to help transform the health care
system in ways that would lead to better performance and lower costs.
Consumers shopping for cars have many sources of information on quality and price of various
models. To the contrary, consumers shopping for health care have trouble finding basic
information about quality and price. Certainly comparing health care alternatives, given their
complexity, uncertainty, made-to-order nature, will be harder than comparing mass-produced
goods. However, providing health care consumers with better information can stimulate
competition, which in turn can deliver better performance and prices.
More transparent pricing of health care could help consumers make better decisions. However,
transparent pricing is likely to be effective only when combined with other measures. Often, price 89
information is of little value without accompanying information on quality. In addition,
consumers will be sensitive to prices only if they share a non-trivial portion of their health care
costs. However, the bulk of medical costs stem from a small proportion of high-cost episodes, 90
often occurring in the last few days of life. Insured patients in those episodes are well above
out-of-pocket limits, above which the insurance plan pays until some very high limit of coverage
is reached. For this reason, insurance plans—not consumers—will be in the best position to put
pressure on providers to lower prices and improve quality for the most expensive types of health
care. That is, better price information for consumers can spur competition among providers of
eyeglasses, teeth cleaning, and routine check-ups, but better price information is unlikely to
sharpen competition among heart surgeons.
Hospitals, in general, have been reluctant to provide a transparent set of prices for their services.
In part, this is a consequence of their cost structure. Hospitals, for their part, must pay large fixed
costs for items such as buildings, maintenance, equipment, and computer systems. At the same
time, hospitals provide a perishable service: an empty hospital bed cannot be saved for tomorrow.
Economic theory suggests that industries that have high fixed costs, and which sell perishable
goods or services, face strong pressures to charge different customers different prices and 91
compete in markets subject to unstable prices. In addition, many hospitals provide services,
such as indigent care and graduate medical education, for which they may not be not wholly
88 For an analysis of the role of health care spending in the federal government’s future fiscal situation, see CRS Report
RL33623, Long-Term Measures of Fiscal Imbalance, by D. Andrew Austin.
89 Testimony of Sara R. Collins and Karen Davis entitled “Transparency in Health Care: the Time Has Come,” in U.S.
Congress, House Committee on Energy and Commerce, Subcommittee on Health, “What’s The Cost? Proposals to thnd
Provide Consumers with Better Information About Healthcare Service Costs,” hearings, 109 Cong., 2 sess., March
90 Of course, at the beginning of an episode of care, whether a patient will survive is usually unknown. Therefore,
distinguishing between life-saving care that saves a patient and futile care for a patient in his last days is difficult.
91 Lester G. Telser, “Competition and the Core.” Journal of Political Economy, 1996, vol. 104, no. 1, pp. 85-107.
compensated. Such hospitals must find other ways to finance these services, which often involves
cross-subsidies. In these conditions, a simple flat-rate price system may not be a viable strategy
for hospitals. Therefore, imposing greater transparency of health care prices may require closer
attention to cross-subsidies and uncompensated training and care.
Health care consumers have access to little useful information about quality. In part this is due to
the inherent complexity of medical care and the difficulty of defining and measuring quality.
However, the development of large electronic databases has opened the possibility of creating
quality indices based on sophisticated statistical methods. Large corporations, insurance
companies, and government agencies have developed extensive databases which contain
information reflecting the quality of health care. These data, aside from some limited exceptions,
are unavailable to consumers.
Medicare pays an average of more than $400 million per year to run 53 Quality Improvement
Organizations (QIOs), which work with hospitals, physicians, nursing homes and other providers
to improve quality of care. Research by academics and the Institute of Medicine has cast doubt on 92
the efficacy of QIOs. Other research, however, suggests that interventions by QIOs are 93
associated with quality improvements. Hospital professional organizations in 2002 created the
Hospital Quality Alliance, which provides comparative data. For each hospital, 20 indicators
measuring the proportion of patients who receive specific treatments recognized to constitute
“best practice” in the areas of heart attacks, heart failure, pneumonia, and prevention of surgical 94
infections are reported. For example, one item reports what percentage of heart attack victims
received aspirin upon arrival at the hospital. Critics say this type of reporting focuses on what a
hospital did, rather than what happened to patients. They note that if food critics operated
according to similar principles, perhaps their reviews would report which restaurants remembered
to include important ingredients of meals or how sophisticated the restaurant stoves were, while
failing to report how meals tasted.
Traditional approaches to quality monitoring in health care focused on “zero/one” indicators that
provide no information on gradations of ability or competence. Physicians were licensed, and
hospitals were accredited, and those who were not could not legally engage in medical care.
Providers were certified for Medicare reimbursement. Such measures, however, only served to set
lower bounds. The board certification of physicians is a partial exception, which provides
consumers an opportunity to select physicians who have passed a more rigorous set of standards.
92 Claire Snyder and Gerard Anderson, “Do Quality Improvement Organizations Improve the Quality of Hospital Care
for Medicare Beneficiaries?” Journal of the American Medical Association, vol. 293, no. 23 (June 15, 2005), pp. 2900-
2907; and the Institute of Medicine of the National Academies, Medicare’s Quality Improvement Organization
Program: Maximizing Potential, March 2006.
93 William Rollow, et al., “Assessment of the Medicare Quality Improvement Organization Program,” Annals of
Internal Medicine, vol. 145, no. 5 (September 5, 2006), pp. 342-353. For the period analyzed by this study QIOs were
allowed to select which providers they would work with. It is difficult, therefore, to infer whether QIOs were good at
helping providers improve quality of care or whether QIOs were good at selecting providers with higher potential for
quality improvement. Future studies by these authors promise more careful controls.
94 These data are available at http://www.hospitalcompare.hhs.gov/.
Providing consumers with more useful data on outcomes can improve health care quality.95 Of
course, outcome data must include risk adjustments, so that statistics reflect the fact that healthier
patients will on average have better outcomes. For example, the United Network for Organ
Sharing, established by Congress in 1984, collects data on all transplant operations in the United
States. Risk-adjusted outcome data for each transplant center are available at
http://www.unos.org. Public availability of risk-adjusted outcome data puts pressure on surgeons
and transplant centers to improve performance. New York State has published risk-adjusted
average mortality rates for cardiac surgery since 1991. After starting this program, the mortality
rate among cardiac patients treated in top-performing hospitals or by top-performing surgeons
was about half the mortality rate for patients treated by a hospital or surgeon rated in the bottom 96
25% of the rankings. A 2003 study, however, contended that publication of performance data
gave providers incentives to avoid difficult cases, leading to worse health outcomes for sicker 97
patients and higher resource usage.
In the absence of effective quality-control programs, malpractice and the tort system act as a 98
rough substitute for quality control. Expanding consumer access to useful medical outcome data
would take pressure off the tort system. For instance, once New York State started publishing
cardiac outcome data, surgeons with high reported risk-adjusted average mortality rates were
more likely to retire or stop performing operations. Giving patients information that allows them
to avoid surgeons with high mortality rates is preferable to having their estates sue.
One strategy for slowing the growth of health care is to make patients pay more when they
choose health care which is more expensive to provide. For example, many employers tie their
contribution to an employee’s health insurance plan to the cost of the cheapest plan. Employees
then face a choice of paying for the incremental cost of a more generous insurance plan out of
their own pocket or spending that money on other things. However, some employees can pay for
more generous insurance out of pre-tax income via premium conversion plans, so that the tax
exemption for employer-provided health care still tilts those employees towards buying more 99
Limiting the income tax exemption for employer-provided health care at some fixed level would
help ensure that those who benefit from more generous health plans pay for their added cost 100
themselves. To the extent that consumers make rational decisions between health and non-
health expenditures, the tax exemption creates an economic distortion that lowers allocational
95 For a more extensive analysis of the potential of providing consumers with useful outcome data see Michael E.
Porter and Elizabeth O. Teisberg, Redefining Health Care: Creating Value-Based Competition on Results (Allston,
Mass: Harvard Business School Publishing, May 2006).
96 Ashish K. Jha and Arnold M. Epstein, “The Predictive Accuracy of the New York State Coronary Artery Bypass
Surgery Report-Card System,” Health Affairs, vol. 25, no. 3 (2006), pp. 844-855.
97 David Dranove, Daniel Kessler, Mark McClellan, and Mark Satterthwaite, “Is More Information Better? The Effects
of ‘Report Cards,’” Journal of Political Economy, vol. 111 (2003), pp. 555-588.
98 Michelle Mello and Troyen Brennan, “Deterrence of Medical Errors: Theory and Evidence for Malpractice Reform,”
Texas Law Review, vol. 80 (2002), pp. 1595-1637.
99 Mark V. Pauly, “Taxation, Health Insurance, and Market Failure in the Medical Economy,” Journal of Economic
Literature, vol. 24, no. 2 (June 1986), pp. 629-75.
efficiency. Limiting this exemption would make employees with more generous health plans
more sensitive to the cost of health insurance, which would put downward pressure on health care
costs. As noted above, the President’s Advisory Panel on Tax Reform called for capping this
exemption for those reasons.
The Weisbrod proposal, described above, which would tie reimbursement for new drugs to the
level of an existing drug of recognized efficacy, is another application of the “make extras cost 101
extra” principle. Providing drug developers with a strong incentive to find new drugs that
would be cheaper but just as effective as currently available drugs could help constrain the
growth of health care costs. In effect, some insurers have partially implemented this approach in
their efforts to promote use of generic drugs by waiving copays or deductibles that would apply 102
for non-generic drugs. Patients who want more expensive drugs, rather than cheaper generics,
must pay some portion of the difference themselves.
The same “pay more for extras” approach appears in various forms in several health care reform
proposals. For example, this idea is central to the rationale for the favorable tax treatment of high-
deductible insurance plans linked to health savings accounts (HSAs), which have been an
important part of the President’s health reform agenda. Those who choose these plans pay for
routine expenses from HSAs, or beyond some point, out of pocket, but retain insurance protection
against the financial risks connected to health care.
Many health analysts contend that more extensive use of information technology (IT) could
improve quality and increase efficiency of medical care. The Veterans’ Health Administration
(VHA), which treats about 5 million patients per year, began an organizational transformation in
1995 that featured a centralized health information system. This system not only simplified record
keeping, giving physicians instant access to records on all previous visits made by a patient, but
also provided researchers and managers access to treatment and outcome data that provided hard 103
evidence on clinical effectiveness. The VHA is now considered a leader in the application of IT 104
to health care administration. Several studies have concluded that quality of care in the VHA 105
exceeds that in comparable private health care providers.
Centralized clinical IT systems can also allow payers to link physician reimbursement with
measures of quality of care. The British National Health Service (NHS) initiated pay-for-
performance contracts with family practitioners in 2004, which tied physician payments to 146
101 Burton A. Weisbrod, “The Health Care Quadrilemma: An Essay on Technological Change, Insurance, Quality of
Care, and Cost Containment,” Journal of Economic Literature, vol. 29, no. 2 (June 1991), pp. 523-552.
102 Henry Kaiser Family Foundation, “Insurance Companies Promote Use of Generic Drugs in Effort To Cut Costs,”
Daily Health Policy Report, September 19, 2006.
103 Kenneth W. Kizer, John G. Demakis, John R. Feussner, “Reinventing VA Health Care: Systematizing Quality
Improvement and Quality Innovation: VA’s Quality Enhancement Research Initiative,” Medical Care 38(6),
Supplement I:I-7-I-16, June 2000.
104 M.W. Morgan, “The VA Advantage: The Gold Standard in Clinical Informatics,” Healthcare Papers, vol.5, no. 4
(2005), pp. 26-9.
105 Steven M. Asch, et al., “Comparison of Quality of Care for Patients in the Veterans Health Administration and
Patients in a National Sample,” Annals of Internal Medicine, vol. 141, no. 12 (December 2004), pp. 938-945.
indicators of the quality of clinical care for 10 chronic diseases.106 In the first year of this
program, almost 97% of U.K. physicians met quality targets. While analysis of these initial data
cannot determine whether initial clinical quality targets were set too low or whether clinical
quality improved, the information collected on quality indicators provides the NHS with a
powerful tool to monitor the quality of patient care and to push for further improvements.
The federal government has taken preliminary steps in the same direction. The federal
government has funded demonstration projects that make extra payments to physicians and 107
hospitals that report certain quality indicators. The Medicare Prescription Drug, Improvement,
and Modernization Act of 2003 (P.L. 108-173) cuts payments by 0.4% to hospitals which do not
report on 10 quality indicators for acute myocardial infarction, congestive heart failure, and
pneumonia. The Deficit Reduction Act of 2005 (P.L. 109-171) increased that penalty to 2% of
payments and authorized the Department of Health and Human Services to modify the set of
quality indicators. The Tax Relief and Health Care Act of 2006 (P.L. 109-432) ties increases in
physician reimbursement under Medicare Part B to reporting of quality measures selected by the
Centers for Medicare and Medicaid Services’ Physician Voluntary Reporting Program. At present,
this program collects data on 16 quality measures. In future years, the Secretary of Health and
Human Services may select other quality measures. While the data supplied by U.S. hospitals and
physicians is much less detailed than information supplied by U.K. physicians participating in
pay-for-performance programs, U.S. efforts to collect a basic set of quality indicators in a
systematic way could underpin future efforts to link payments to quality of care.
Market competition has brought rapid technological change, higher quality, and lower prices to
many parts of the economy, leading many to ask why broader application of market principles to
the health care system could not reduce costs and improve performance. As many have noted,
health care is not a standard commodity. Unless institutions and reforms are designed to reflect
the unique characteristics of the health care market, failures will continue to be the norm.
Certainly past experience indicates that patients, insurers, and providers all react to financial
incentives. Unfortunately, designing incentives that align behavior of consumers and health care
payers and providers is difficult because of the nature of medical care. Broadening financial
protection against unexpected health care costs, enhancing access to routine and preventative
care, constraining costs, and setting up economic incentives that induce efficient behavior among
individuals and firms are all worthwhile goals. Designing policies with a reasonable expectation
of reaching each of those goals is difficult.
Information asymmetries between patients and physicians as well as between providers and
payers stem from the unavoidable need for highly specialized roles. The specialized expertise of
medical personnel gives patients access to sophisticated therapeutic measures, many of which
could hardly have been imagined a generation ago. This specialization of expertise also
necessitates an informational asymmetry that requires patients to depend on their physicians to
guide them through the system. Thus, the principal-agent relationship lies at the heart of the
106 T. Doran, C. Fullwood, H. Gravelle, et al., “Pay-for-Performance Programs in Family Practices in the United
Kingdom,” New England Journal of Medicine, vol. 355 (July 27, 2006), pp. 375-384.
107 Arnold M. Epstein, “Paying for Performance in the United States and Abroad,” New England Journal of Medicine,
vol. 355, no. 4 (July 27, 2006), pp. 406-408.
health care experience. Similarly, because physicians see patients while payers only see paper
claims, physicians possess an important informational advantage over third-party payers.
Expanded use of performance measurements and information technology has the potential to
reduce these problems.
The health insurance system’s ability to spread risks across a broad pool of beneficiaries will
continue to face challenges as insurers seek to avoid losses through medical underwriting and
related practices. Extending coverage to individuals in poor health, from the point of view of an
insurer, is more akin to providing a subsidy than insuring a risk. Insuring such individuals may
require either subsidies or pricing schemes that induce private insurers to compete on grounds of
efficiency of service rather than ability to avoid those with greater needs for health care.
Alternatively, health insurance for those in poorer health could be guaranteed by policies that
would create larger and broader pools, such as the Canadian single-payer system. Another
alternative is the United Kingdom’s two-tier system, in which all residents are enrolled in the
National Health Service, but those who wish access to more convenient services may buy
supplemental policies, such as provided by the British United Provident Association (BUPA).
This system gives all U.K. residents access to a reasonable standard of care, although this may
involve some level of inconvenience and waiting in some cases, and restricted access to some
care regarded by health planning authorities as of low benefit. Those wishing to bypass such
inconveniences can pay for access to more extensive and comfortable care.
Some state governments have introduced other reform strategies. After July 2007, Massachusetts
will require all residents to obtain health insurance. This approach parallels the typical automobile
insurance system, which requires citizens to have insurance, but allows them to choose among
many approved and regulated insurance providers. Low-income individuals will be eligible for 108
subsidized access to insurance on a sliding scale. Governor Schwarzenegger has proposed a
health insurance reform plan that also includes a requirement that all individuals obtain health 109
Even if the design of health care policy requires balancing of goals that present conflicting
requirements, policy innovations can deliver better results. Policies that harness market incentives
and recognize the nature of health care can enhance efficiency and improve performance.
Because asymmetric information is a cause of market failure, policies that provide better
information to patients and payers may improve performance. Because market power often serves
as a cause of market failure, policies that either reduce or counterbalance market power can help
improve performance. Policies that either fail to reflect the nature of the health care or that fail to
address causes of market failure are unlikely to lead to lasting improvements.
108 Henry J. Kaiser Foundation, “Massachusetts Health Care Reform Plan - Fact Sheet,” April 2006, available at
109 For details, see Arnold Schwarzenegger, “Governor’s Health Care Proposal,” press release, available at
D. Andrew Austin
Analyst in Economic Policy