Key Issues In Analyzing Major Health Insurance Proposals
CONGRESS OF THE UNITED STATES
CONGRESSIONAL BUDGET OFFICE
CBO
Key Issues in
Analyzing Major
Health Insurance
Proposals
DECEMBER 2008
© Royalty-free/bxp2598/Brand X
Pub. No. 3102
A
CBO
S T U D Y
Key Issues in Analyzing
Major Health Insurance Proposals
December 2008
The Congress of the United States O Congressional Budget Office
Notes
Numbers in the text and tables may not add up to totals because of rounding.
This document references a number of the Congressional Budget Office’s (CBO’s) cost
estimates and other publications, which are available at CBO’s Web site, www.cbo.gov.
Preface
Concerns about the number of people who lack health insurance and about the high
and rising costs of health insurance and health care have led to proposals that would substan-
tially modify the health insurance system in this country. Because the Medicare program
already provides nearly universal coverage to the elderly, those proposals generally focus on
options for providing coverage to and reducing costs for the nonelderly population. Because
most nonelderly people obtain their insurance coverage through an employer, proposals could
affect that coverage in some way. They could, for example, provide new federal subsidies to
pay some portion of health insurance premiums; impose mandates for individuals to purchase
coverage or for employers to offer it; encourage alternatives to employment-based insurance or
provide strong incentives to purchase coverage individually; or create new federally adminis-
tered options for obtaining health insurance (including a single-payer system in which all citi-
zens or residents would be offered coverage under Medicare).
This report describes some of the key assumptions that the Congressional Budget Office
(CBO) would use in estimating the effects of key elements of such proposals on federal costs,
insurance coverage, and other outcomes; the evidence on which those assumptions are based;
and—if the evidence points to a range of possible effects rather than a precise prediction—
the factors that would influence where a proposal falls within those ranges. In doing so, it also
reviews many of the major issues that arise in designing such proposals. This document does
not provide a comprehensive analysis of any specific proposal; rather, it identifies and dis-
cusses many of the critical factors that would affect estimates of various proposals. In accor-
dance with CBO’s mandate to provide objective and impartial analysis, the report makes no
recommendations.
This report is the product of an intense effort on the part of a large number of CBO analysts.
Philip Ellis and Janet Holtzblatt of CBO’s Health and Human Resources Division organized
and revised the final drafts of each chapter, under the supervision of James Baumgardner.
David Auerbach, Lyle Nelson, Ben Page, Lara Robillard, Rob Stewart, and Chapin White
contributed major sections of the report and drafted several of its chapters. Other significant
contributions came from Colin Baker, Paul Cullinan, Noelia Duchovny, Renee Fox,
Tim Gronniger (formerly of CBO), Stuart Hagen, Keisuke Nakagawa, Jean Hearne,
Melissa Merrell, Allison Percy, Lisa Ramirez-Branum, Bill Randolph, and David Weiner.
In addition, Tom Bradley, Pete Fontaine, Keith Fontenot, Holly Harvey, Kate Massey, and
Bruce Vavrichek provided important guidance on the report throughout its development. The
analysis also benefited greatly from comments by Joseph Newhouse of Harvard University and
Mark Pauly of the University of Pennsylvania. (The assistance of external reviewers implies no
responsibility for the final product, which rests solely with CBO.)
Christine Bogusz and Sherry Snyder edited the report. Maureen Costantino designed the
cover and prepared the report for publication. Lenny Skutnik printed the initial copies, Linda
Schimmel coordinated the print distribution, and Simone Thomas prepared the electronic
version for CBO’s Web site.
Finally, special thanks are due to CBO’s former Director, Dr. Peter R. Orszag, who conceived
the idea for this report and was instrumental in its development.
Robert A. Sunshine
Acting Director
December 2008
Contents
Summary
ix
1 Introduction and Background
1
Scope and Focus of This Report
2
Health Insurance Coverage
3
Health Care Spending
18
2 Approaches for Reducing the Number of Uninsured People
27
Methods of Subsidizing Premiums
28
Subsidizing Premiums Through the Tax System
29
Subsidizing Premiums Through Spending Programs
37
Effects of Premium Subsidies on Rates and Sources of
Insurance Coverage
43
Individual and Employer Mandates
48
Automatic Enrollment Provisions
54
3 Factors Affecting Insurance Premiums
59
Design of Benefits and Cost Sharing
60
Management of Benefits
63
Administrative Costs of Health Plans
69
Effects of Gaining Insurance Coverage on Health Care
Use and Spending
71
4 Proposals Affecting the Choice of an Insurance Plan
77
Regulating Insurance Premiums and Sales
77
Revealing the Relative Costs of Health Plans
84
Expanding Access to Federally Administered Plans
91
CBO
VI
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
5 Factors Affecting the Supply and Prices of Health Care Services
99
Background on the Supply of Health Care Providers
99
Payment Methods and Providers’ Incentives
102
Payment Rates
105
Responses to Changes in Payment Rates or Demand for Services
108
Uncompensated Care and Cost Shifting
112
6 Administrative Issues and Effects on Other Programs
117
Implementation Issues and Timelines
117
Maintenance-of-Effort Requirements
120
Effects on Other Federal Programs
123
Coverage of Certain Populations
125
7 Changes in Health Habits and Medical Practices
131
Modifying Health Habits
132
Expanding the Use of Clinical Preventive Services
136
Establishing a “Medical Home”
139
Adopting Disease Management Programs
141
Comparing the Effectiveness of Medical Treatments
144
Adopting Health Information Technology
147
Modifying Laws About Medical Malpractice
150
8 Effects on Total Health Care Spending, the
Scope of the Federal Budget, and the Economy
155
Effects on National Health Expenditures
156
Global Budgets for Health Care
157
Impact of Proposals on the Composition of Compensation and
Tax Revenues
160
Flow of Payments and Budgetary Treatment
160
Macroeconomic Effects
162
CBO
CONTENTS
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
VII
Tables
S-1.
Sources of Insurance Coverage and Insurance Status of the
Nonelderly Population, 2009
xiv
1-1.
Sources of Insurance Coverage and Insurance Status of the
Nonelderly Population, 2009
4
1-2.
Share of Employees Offered Health Insurance, by Size of Firm, 2009
7
1-3.
Health Care Expenditures in 2008, by Insurance Status
14
1-4.
National Health Expenditures, by Source of Payment and
Type of Service, 2009
19
1-5.
Persistence of Health Care Spending
24
2-1.
Distribution of the Nonelderly Population, by Insurance Status,
Family Income, and Family Structure, 2009
29
2-2.
Illustrative Tax Subsidy for Employment-Based Health Insurance for a
Single Worker Who Receives $40,000 in Total Compensation, 2009
31
2-3.
Effects on a Single Worker of Repealing the Tax Exclusion and
Replacing It with an Above-the-Line Deduction, 2009
33
2-4.
Effects of a Premium Subsidy on Offer Rates for
Employment-Based Coverage, by Size of Firm, 2009
44
2-5.
U.S. Experience with Enforcing Mandates
52
2-6.
Impact of Third-Party Data and Enforcement Methods on
Income Tax Compliance, 2001
54
3-1.
Administrative Costs for Private Health Plans, by Category, 2006
69
6-1.
Current Allocation of Major Administrative Responsibilities for
Financing and Regulating Health Care
118
7-1.
Medical Injuries, Negligence, and the Filing of Malpractice Claims, 2003
152
Figures
1-1.
Patterns of Health Insurance Coverage for Nonelderly People, by
Family Income Relative to the Federal Poverty Level, 2009
11
1-2.
Uninsurance Rates of Full-Time Workers, by Size of Firm and
Family Income Relative to the Poverty Level, 2009
12
1-3.
Projected Distribution of the Uninsured Nonelderly Population, by
Selected Characteristics, 2009
13
CBO
VIII
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
Figures (Continued)
1-4.
The Relationship Between Quality of Care and Medicare Spending, by
State, 2004
25
2-1.
Distribution of Marginal Tax Rates on Income for the
Nonelderly Uninsured Population, 2009
36
2-2.
Probability of Enrolling in an Individually Purchased
Insurance Plan with a Subsidy
47
3-1.
Average Annual Premiums for Covered Workers for Single and
Family Coverage, by Plan Type, 2008
66
7-1.
Trends in Smoking and Obesity
133
Boxes
1-1.
Regulation of Health Insurance and the Employee Retirement
Income Security Act
6
2-1.
Tax Exclusions, Tax Deductions, and Tax Credits
34
2-2.
Issues with Refundable Tax Credits
38
2-3.
Health Insurance Mandates in Hawaii and Massachusetts
50
3-1.
What Is Actuarial Value?
64
7-1.
Types of Tort Reforms
153
CBO
Summary
Much of the health care provided in the United the federal budget, the number of people who have health
States confers tremendous benefits, extending and
insurance coverage, and spending for health care. Some of
improving lives. But the high and rising costs for health
its main conclusions are as follows:
care in this country impose an increasing burden on indi-
viduals, businesses, states, and the federal government,
B The rising costs of health care and health insurance
and a substantial number of people may have trouble
pose a serious threat to the future fiscal condition of
paying for that care because they do not have health
the United States. Under current policies, CBO pro-
jects that federal spending on Medicare and Medicaid
insurance. Those issues are related: Rising costs for health
will increase from about 4 percent of gross domestic
care make health insurance policies more expensive and
product (GDP) in 2009 to nearly 6 percent in 2019
thus more difficult to afford. Lack of insurance can limit
and 12 percent by 2050. Most of that increase will
access to care, but having insurance can increase spending
result from growth in per capita costs rather than from
by encouraging the use of services that provide limited
the aging of the population.
health benefits. More generally, despite spending more
per capita than other countries, the United States lags
B Without changes in policy, a substantial and growing
behind lower-spending countries on several metrics,
number of nonelderly people (those younger than 65)
including life expectancy and infant mortality. Indeed,
are likely to be without health insurance. CBO esti-
evidence suggests that a substantial share of spending on
mates that the average number of nonelderly people
health care contributes little if anything to the overall
who are uninsured will rise from at least 45 million in
health of the nation, but finding ways to reduce such
2009 to about 54 million in 2019.
spending without also affecting services that improve
B
health will be difficult.
Those problems cannot be solved without making
major changes in the financing or provision of health
insurance and health care. In considering such
Main Conclusions
changes, policymakers face difficult trade-offs between
Concerns about the number of people who are uninsured
the objectives of expanding insurance coverage and
and about the rising costs of health insurance and health
controlling both federal and total costs for health care.
care have given rise to proposals that would substantially
B By themselves, premium subsidies or mandates to
modify the U.S. health insurance system. The complexi-
obtain health insurance would not achieve universal
ties of the health insurance and health care systems pose a
coverage. Proposals could, however, achieve near-
major challenge for the design of such proposals and
universal coverage using a combination of approaches.
inevitably raise questions about their likely impact. To
assist the Congress in its upcoming deliberations, this
• One option would establish an enforceable man-
report seeks to provide useful background information as
date for individuals to obtain insurance and would
well as insights into how the Congressional Budget Office
provide subsidies for lower-income households to
(CBO) would estimate the effects of such proposals on
help them pay their required premiums.
X
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
• Another option, under a voluntary system, would
estimates of a variety of proposals. In particular, it consid-
provide subsidies that cover a very large share of the
ers the types of issues that would arise in estimating the
expected costs of insurance for every enrollee and
effects of proposals to:
establish a process to facilitate enrollment (as is
done in Medicare).
B Provide tax credits or other types of subsidies to make
insurance less expensive to the purchaser.
Other policies could achieve substantial reductions in
the number of people who are uninsured at a lower
B Require individuals to purchase health insurance,
budgetary cost.
typically paired with a new system of government
subsidies.
B Serious concerns exist about the efficiency of the
health care system, but no simple solutions are avail-
able to reduce the level or control the growth of health
B Require firms to offer health insurance to their
care costs. Steps to restructure the insurance market
workers or pay into a fund that subsidizes insurance
and to encourage people to purchase less extensive
purchases.
coverage could reduce the use of treatments that pro-
vide minimal benefits, but enrollees would face higher
B Replace employment-based coverage with new pur-
cost sharing or tighter management of their care.
chasing arrangements or provide strong incentives for
people to shift toward individually purchased
B Other approaches—such as the wider adoption of
coverage.
health information technology or greater use of
preventive medical care—could improve people’s
B Provide individuals with coverage under, or access to,
health but would probably generate either modest
existing insurance plans such as the Medicare
reductions in the overall costs of health care or
program, either as an additional option or under a
increases in such spending within a 10-year budgetary
“Medicare-for-all” single-payer arrangement.
time frame.
Wherever possible, the analysis presented here describes
B In many cases, the current health care system does not
in quantitative terms how CBO would estimate the bud-
give doctors, hospitals, and other providers of health
getary and other effects of such proposals. In other cases,
care incentives to control costs. Significantly reducing
it describes the components that a proposal would have
the level or slowing the growth of health care spending
to specify in order to permit estimation of its effects on
would require substantial changes in those incentives.
federal spending or other outcomes. This report reflects
the current state of CBO’s analysis of and judgments
Scope and Focus of This Report
about the likely response of individuals, employers, insur-
In the near future, the Congress is expected to consider
ers, and providers to changes in the health insurance and
major proposals to modify the health insurance system.
health care systems. The details of particular policy speci-
This report describes the assumptions that CBO would
fications and the way in which they are combined, as well
use in estimating the effects of key elements of such pro-
as new evidence or analysis related to the issues discussed
posals on federal costs, insurance coverage, and other out-
here, could affect future CBO estimates of the effects of
comes; the evidence upon which those assumptions are
large-scale health insurance proposals.
based; and, if the evidence points to a range of possible
effects rather than a precise prediction, the factors that
Because such proposals could incorporate a number of
would influence where a proposal falls within those
the elements that are discussed in this report, they could
ranges.
have interactions that are difficult to predict. Those pro-
posals could also affect both tax revenues and outlays.
This document does not provide a comprehensive
Estimates of the impact on revenues of proposals to
analysis of any specific proposal; rather, it identifies and
change the federal tax code are prepared by the staff of
examines many of the critical factors that would affect
the Joint Committee on Taxation (JCT) and would be
CBO
SUMMARY
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
XI
incorporated into any formal CBO estimate of a pro-
ing spouses and dependents (see Summary Table 1). One
posal’s effects on the federal budget. In preparing this
fundamental reason such plans are popular is that they
report, CBO consulted with the JCT staff about the
are subsidized through the tax code—because nearly all
behavioral considerations that are incorporated into both
payments for employment-based insurance are excluded
agencies’ estimates.
from taxable compensation and thus avoid income and
payroll taxes. Another factor is the economies of scale that
The question of whether and how any net increases in
larger group purchasers enjoy, which reduce the average
federal spending for health care and health insurance
amount of administrative costs that are embedded in pol-
would be financed by policy changes outside the health
icy premiums; partly as a result, large employers are more
sector is beyond the scope of this report. Whether a pro-
likely than small employers to offer insurance to their
posal would make health insurance more affordable for a
workers. Overall, about three-fourths of workers are
given individual or family would depend not only on its
offered employment-based insurance and are eligible to
impact on the health insurance premiums that they face
enroll in it.
but also on the effect that its financing mechanisms
would have on the household’s budget. To the extent that
Another commonly cited reason for the popularity of
such proposals would be financed by policy changes that
employment-based policies is that employers offering
fall outside the health sector—through tax increases or
coverage usually pay most of the premium—a step they
reductions in spending for other federal programs—those
take partly to encourage broad enrollment in those plans,
effects are not addressed here.
which helps keep average costs stable. Ultimately, how-
ever, the costs of those employers’ payments are passed on
to employees as a group, mainly in the form of lower
Background on Spending and Coverage
wages.
Health care costs are an important issue, not just for indi-
viduals and families seeking insurance coverage but also
Other Sources of Coverage
for the federal budget and the economy as a whole.
Other significant sources of coverage include the individ-
Spending on health care and related activities will
ual insurance market and various public programs.
account for about 17 percent of gross domestic product
Roughly 10 million people are covered by individually
in 2009—an expected total of $2.6 trillion—and under
purchased plans, which have some advantages for enroll-
current law that share is projected to reach nearly 20 per-
ees; for example, they may be portable from job to job,
cent by 2017. Annual health expenditures per capita are
unlike employment-based insurance. Even so, individu-
projected to rise from about $8,300 to about $13,000
ally purchased policies generally do not receive favorable
over that period. Federal spending accounts for about
tax treatment; in most states, premiums may vary to
one-third of those totals, and federal outlays for the
reflect an applicant’s age or health status, and applicants
Medicare and Medicaid programs are projected to grow
with particularly high expected costs are generally denied
from $720 billion in 2009 to about $1.4 trillion in 2019.
coverage.
Over the longer term, rising costs for health care repre-
Another major source of coverage is the federal/state
sent the single greatest challenge to balancing the federal
Medicaid program and the related but smaller State
budget. The number of uninsured individuals is also
Children’s Health Insurance Program (SCHIP). Both
expected to increase because health insurance premiums
programs provide free or low-priced coverage for children
are likely to continue rising much faster than income,
in low-income families and (to a more limited degree)
which will make insurance more difficult to afford.
their parents; Medicaid also covers poor individuals who
are blind or disabled. On average, Medicaid and SCHIP
Employment-Based Insurance
are expected to cover about 43 million nonelderly people
For several reasons, most nonelderly individuals obtain
in 2009 (and there are also many people eligible for those
their insurance through an employer, and employment-
programs who have not enrolled in them). Medicare also
based plans now cover about 160 million people, includ-
CBO
XII
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
Summary Table 1.
B Subsidizing health insurance premiums, either
Sources of Insurance Coverage and
through the tax system or spending programs, which
Insurance Status of the Nonelderly
would make insurance less expensive for people who
are eligible; or
Population, 2009
B Establishing a mandate for health insurance, either
Number
by requiring individuals to obtain coverage or by
(Millions)
Percent
requiring employers to offer health insurance to their
Source of Coverage
workers.
Employment-Baseda
160
61
Individually Purchased
10
4
By themselves, premium subsidies or mandates to obtain
Medicare
7
3
health insurance would not achieve universal coverage.
Medicaidb
43
17
Those approaches could be combined and could be
Otherc
12
4
implemented along with provisions to facilitate enroll-
Insurance Status
ment in ways that could achieve near-universal coverage.
Insured, Any Sourced
216
83
Uninsured
45
17
Subsidizing Premiums
Whether new subsidies are delivered through the tax
Source:
Congressional Budget Office’s health insurance simulation
system or a spending program, several common issues
model.
arise. Trade-offs exist between the share of the premium
Note: The nonelderly population (those younger than 65) excludes
that is subsidized, the number of people who enroll in
people in institutions and residents of U.S. territories.
insurance as a result of the subsidies, and the total costs of
a.
Includes coverage obtained through local, state, and federal
the subsidies. As the subsidy rate increases, more people
employers.
will be inclined to take advantage of them, but the higher
b.
Includes the State Children’s Health Insurance Program.
subsidy payments will also benefit those who would have
c.
Includes military and other sources of coverage.
decided to obtain insurance anyway. Beyond a certain
d.
The sum of people by their sources of coverage exceeds the
point, therefore, the cost per newly insured person can
total number who are insured because about 14.5 million
grow sharply because a large share of the additional sub-
people are covered by more than one source at a time.
sidy payments is going to otherwise insured individuals.
covers about 7 million people younger than 65 who are
disabled or have severe kidney disease. About 12 million
To hold down the costs of subsidies, the government
people have insurance coverage from various other
could limit eligibility for subsidy payments to individuals
sources, including federal health programs for military
who are currently uninsured. That restriction, however,
personnel. The total number of nonelderly people with
would create incentives for insured individuals to drop
health insurance at any given point in 2009 is expected to
their coverage. Some proposals might try to distinguish
be about 216 million.
between people who become uninsured in response to
subsidies and those who would have been uninsured in
the absence of a government program (for example, by
Approaches for Reducing the
imposing waiting periods for individuals who were previ-
Number of Uninsured People
ously enrolled in an employment-based plan), but such
Concerns about the large number of people who lack
proposals could be very difficult to administer. In addi-
health insurance have generated proposals that seek to
tion, providing benefits only to the uninsured might be
increase coverage rates substantially or achieve universal
viewed as unfair by people with similar income and fam-
or near-universal coverage. Two basic approaches could
ily responsibilities who purchase health insurance and are
be used:
therefore ineligible for the subsidies.
CBO
SUMMARY
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
XIII
Another approach to limiting costs would target subsidies
ited if people do not receive refundable tax credits before
toward the lower-income groups, who are most likely to
their premium payments are due.
be uninsured otherwise, but such approaches can also
have unintended consequences that affect the costs of a
Providing Subsidies Through Spending Programs. The
proposal. If eligibility was limited to people with income
government could seek to increase coverage rates by
below a certain level, then those with income just above
spending funds to subsidize insurance premiums. New
the threshold would have strong incentives to work less or
subsidies could be provided implicitly by expanding eligi-
hide income in order to qualify for the subsidies or main-
bility for Medicare, Medicaid, or SCHIP or explicitly by
tain their eligibility. Phasing out subsidies gradually as
creating a new program. To hold costs down, benefits
income rises would reduce those incentives, but it would
could be targeted on the basis of income, assets, family
increase the amount of subsidy payments that go to indi-
responsibilities, and insurance status. Targeting benefits,
viduals and families who would have had insurance in
however, would require program administrators to certify
any event.
eligibility and enforce the program’s rules, which would
affect coverage and the program’s costs.
Restructuring the Existing Tax Subsidies. Tax subsidies
could be restructured to expand coverage in several ways.
The Effects of Subsidy Proposals. Proposals to subsidize
For example, the current tax exclusion for employment-
insurance coverage would affect decisions by both
based health insurance could be replaced with a deduc-
employers and individuals. Employers’ decisions to offer
tion or tax credit to offset the costs of insurance, and tax
insurance to their workers reflect the preferences of their
subsidies could be extended to include policies purchased
workers, the cost of the insurance that they can provide,
in the individual insurance market. That step would sever
and the costs of alternative sources of coverage that work-
the link between employment and tax subsidies for pri-
ers would have. Smaller firms appear to be more sensitive
vate health insurance and could give similar people the
to changes in the cost of insurance than are larger
same subsidy whether or not they were offered an
employers. Subsidies that reduce the cost of insurance
employment-based health plan.
offered outside the workplace would cause some firms to
drop coverage or reduce their contributions. When decid-
Deductions and credits differ, however, in their effective-
ing whether to enroll in employment-based plans, work-
ness at reaching the uninsured. An income tax deduction
ers would consider the share of the premium that they
might provide limited benefits to low-income individuals
pay as well as the price and attractiveness of alternatives.
because, like the existing exclusion, its value is less for
The available evidence indicates that a small share of the
those in lower tax brackets. In contrast, tax credits can be
population would be reluctant to purchase insurance
designed to provide lower- and moderate-income taxpay-
even if subsidies covered nearly all of the costs.
ers with larger benefits than they would receive from tax
deductions or exclusions. An important question regard-
Mandating Coverage
ing tax credits—particularly for lower-income people
In an effort to increase the number of people who have
who pay relatively little in income taxes and are also more
health insurance or to achieve universal or near-universal
likely to be uninsured—is whether the credits would be
coverage, the government could require individuals to
refundable and therefore fully available to individuals
obtain health insurance or employers to offer insurance
with little or no income tax liability.
plans. Employer mandates could include a requirement
that employers contribute a certain percentage of the pre-
For the same budgetary costs, a refundable tax credit
mium, which would encourage their workers to purchase
might be more effective at increasing insurance coverage,
coverage. To the extent that the required contributions
both because it can be designed to provide a larger benefit
exceeded the amounts that employers would have paid
to people who have low income than they receive under
under current law, offsetting reductions would ultimately
current law and because those recipients might be more
be made in wages and other forms of compensation.
responsive to a given subsidy than are people with higher
income. Still, the effect on coverage rates might be lim-
The impact of a mandate on the number of people
covered by insurance would depend on its scope, the
CBO
XIV
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
extent of enforcement, and the incentives to comply, as
encourages enrollees to use fewer services. An offsetting
well as the benefits that enrollees would receive. Individ-
factor is that average administrative costs are much higher
ual mandates, for example, could be applied broadly to
for individually purchased policies. The remainder of the
the entire population of the United States or to a specific
difference in premiums probably arises because people
group, such as children; employer mandates might vary
who purchase individual coverage have lower expected
by the size of the firm.
costs for health care to begin with.
Penalties would generally increase individuals’ incentives
The federal costs of providing premium subsidies, or the
to comply with mandates, but when deciding whether to
effects of those subsidies on the number of people who
obtain insurance, people would also consider the likeli-
are insured, would depend heavily on the premiums
hood of being caught if they did not comply. Data from
charged. Premiums reflect the average cost that any
the tax system and from other government programs,
insurer—public or private—incurs, and those costs are a
where overall rates of compliance range from roughly
function of several factors:
60 percent to 90 percent, indicate that mandates alone
would not achieve universal coverage, largely because
B The scope of the benefits the coverage includes and its
some people would still be unwilling or unable to pur-
cost-sharing requirements;
chase insurance.
B The degree of benefit management that is conducted;
Facilitating Enrollment
Simplifying the process of enrolling in health insurance
B The administrative costs the insurer incurs; and
plans or applying for subsidies could yield higher cover-
B The health status of the individuals who enroll.
age rates and could also increase compliance with a man-
date to obtain coverage. One approach would be to enroll
Insurers’ costs also depend on the mechanisms and rates
eligible individuals in health insurance plans automati-
used to pay providers and on other forces affecting the
cally, giving them the option to refuse that coverage or to
supply of health care services. Proposals could affect
switch to a different plan. Automatic enrollment proce-
many of those factors directly or indirectly. For example,
dures have been found to increase participation rates in
the government might specify a minimum level of bene-
retirement plans and government benefit programs.
fits that the coverage must provide in order to qualify for
Automatic enrollment requires the government, an
a subsidy or fulfill a mandate; such a requirement could
employer, or some other entity to determine the specific
have substantial effects on the proposal’s costs or its
plan into which people will be enrolled, however, and
impact on coverage rates.
those choices may not always be appropriate for every-
one.
Design of Benefits and Cost Sharing
Health insurance plans purchased in the private market
Factors Affecting Insurance Premiums
tend to vary only modestly in the scope of their bene-
Premiums for employment-based plans are expected to
fits—with virtually all plans covering hospital care, physi-
average about $5,000 per year for single coverage and
cians’ services, and prescription drugs—but they vary
about $13,000 per year for family coverage in 2009. Pre-
more substantially in their cost-sharing requirements. A
miums for policies purchased in the individual insurance
useful summary statistic for comparing plans with differ-
market are, on average, much lower—about one-third
ing designs is their “actuarial value,” which essentially
lower for single coverage and one-half lower for family
measures the share of health care spending for a given
policies. Those differences largely reflect the fact that pol-
population that each plan would cover. Actuarial values
icies purchased in the individual market generally cover a
for employment-based plans typically range between
smaller share of enrollees’ health care costs, which also
65 percent and 95 percent, with an average value between
80 percent and 85 percent. Cost-sharing requirements for
CBO
SUMMARY
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
XV
enrollees tend to be greater for policies purchased in the
10 percent lower than for plans using minimal manage-
individual insurance market, where actuarial values
ment. Conversely, proposals that restricted plans’ use of
generally range from 40 percent to 80 percent, with an
such management tools would result in higher health care
average value between 55 percent and 60 percent.
spending than proposals that did not impose such
restrictions.
Public programs also vary in the extent of the coverage
they provide. Medicaid requires only limited cost sharing
Administrative Costs
(reflecting the low income of its enrollees); cost sharing
Some proposals would affect the price of health insurance
under SCHIP may be higher but is capped as a share of
by changing insurers’ administrative costs. Some types of
family income. Medicare’s cost sharing varies substan-
administrative costs (such as those for customer service
tially by the type of service provided; for example, home
and claims processing) vary in proportion to the number
health care is free to enrollees, but most hospital admis-
of enrollees in a health plan, but others (such as those for
sions incur a deductible of about $1,000. In addition, the
sales and marketing efforts) are more fixed; that is, those
program does not cap the out-of-pocket costs that enroll-
costs are similar whether a policy covers 100 enrollees or
ees can incur. Overall, the actuarial value of Medicare’s
100,000. As a result of those economies of scale, the aver-
benefits for the nonelderly population is about 15 percent
age share of the policy premium that covers administra-
lower than that of a typical employment-based plan.
tive costs varies considerably—from about 7 percent for
Those considerations would affect CBO’s analysis of
employment-based plans with 1,000 or more enrollees to
proposals to expand enrollment in public programs.
nearly 30 percent for policies purchased by very small
firms (those with fewer than 25 employees) and by indi-
In general, the more comprehensive the coverage a health
viduals.
plan provided, the higher would be the premium or cost
per enrollee. Indeed, an increase in a health plan’s actuar-
Some administrative costs would be incurred under any
ial value would also lead enrollees to use more health care
system of health insurance, but proposals that shifted
services. Reflecting the available evidence, CBO estimates
enrollment away from the small-group and individual
that a 10 percent decrease in the out-of-pocket costs that
markets could avoid at least a portion of the added
enrollees have to pay would generally cause their use of
administrative costs per enrollee that are observed in
health care to increase by about 1 percent to 2 percent. A
those markets. In general, however, substantial reductions
similar analysis would be applied to proposals that
in administrative costs would probably require the role of
included subsidies to reduce the cost-sharing require-
insurance agents and brokers in marketing and selling
ments that lower-income enrollees face.
policies to be sharply curtailed and the services they pro-
vide to be rendered unnecessary.
Management of Benefits
Another factor affecting health insurance premiums and
Spending by Previously Uninsured People
thus the costs or effects of legislative proposals is the
The impact that the mix of enrollees has on health insur-
degree of benefit and cost management that insurers
ance premiums is also an important consideration, partic-
apply. Nearly all Americans with private health insurance
ularly for proposals that would reduce the number of
are enrolled in some type of “managed care” plan, but the
people who are uninsured. The reason is that the use of
extent to which specific management techniques are used
health care by the previously uninsured will generally
varies widely. Common techniques used to constrain
increase when they gain coverage. On average, CBO esti-
costs include negotiating lower fees with a network of
mates, the uninsured currently use about 60 percent as
providers, requiring that certain services be authorized in
much care as the insured population, after adjusting for
advance, monitoring the care of hospitalized patients, and
differences in demographic characteristics and health sta-
varying cost-sharing requirements to encourage the use of
tus between the two groups.
less expensive prescription drugs. Overall, CBO estimates
that premiums for plans making extensive use of such
On the basis of the research literature and an analysis of
cost-management techniques would be 5 percent to
survey data, CBO estimates that enrolling all people who
CBO
XVI
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
are currently uninsured in a typical employment-based
Insurance Market Regulations
plan would increase their use of services by 25 percent to
Proposals could seek to establish or alter regulations gov-
60 percent; that is, they would use between 75 percent
erning the range of premiums that insurers may charge or
and 95 percent as many services as a similar group of
the terms under which individuals and groups purchase
insured people. The remaining gap in the use of services
coverage. Purchases in the individual insurance market
reflects the expectation that, on average, people who are
and most policies for small employers are governed pri-
uninsured have a lower propensity to use health care, a
marily by state regulations. Those regulations differ in the
tendency that would persist even after they gained cover-
extent to which they limit variation in premiums, require
age. For more incremental increases in coverage rates,
insurers to offer coverage to applicants, permit exclusions
CBO would expect that people who chose to enroll in a
for preexisting health conditions, or mandate coverage
new program would be more likely to use medical care
of certain benefits. Roughly 20 percent of applicants for
than those who decided not to enroll.
coverage in the individual market have health problems
that raise their expected costs for health care substantially,
In addition, recent estimates indicate that about a third of
and in most states they may be charged a higher premium
the care that the uninsured receive is either uncompen-
or have their application denied; as a result, premiums are
sated or undercompensated—that is, they either pay
correspondingly lower in those states for the majority of
nothing for it or pay less than the amount that a provider
applicants.
would receive for treating an insured patient. To the
extent that such care became compensated under a pro-
Proposals may seek to modify the regulation of health
posal to expand coverage, health care spending for the
insurance markets in order to make insurance more
uninsured would increase, regardless of whether their use
affordable for people with health problems or to give con-
of care also rose.
sumers more choices, but those goals may conflict with
one another. For example, limiting the extent to which
Proposals Affecting the Choice of an
premiums for people in poor health can exceed those for
people in better health (as some states currently do)
Insurance Plan
would reduce premiums for those who have higher
The government could affect the options available to
expected costs for health care, but it would also raise pre-
individuals when choosing a health insurance plan—and
miums for healthier individuals and thus could reduce
the incentives they face when making that choice—in a
their coverage rates. Other proposals might counteract
number of ways. In particular, proposals could establish
such limits on variations in premiums—for example, by
or alter regulations governing insurance markets, seek to
allowing people to buy insurance in other states. That
reveal more fully the relative costs of different health
approach would enable younger and relatively healthy
insurance plans, or have the federal government offer new
individuals living in states with tight limits to purchase a
health insurance options.
cheaper policy in another state. Older and less healthy
residents who continued to purchase individual coverage
The effects of proposals on insurance markets would
in the tightly regulated states, however, would probably
depend on more than the impact they have on the premi-
face higher premiums as a result.
ums charged or on the share of the premium that enroll-
ees have to pay; those effects would also reflect the market
By themselves, changes in the regulation of the individual
dynamics that arise as individuals shift among coverage
and small-group insurance markets would generally have
options and as policy premiums adjust to those shifts. In
modest effects on the federal budget and on the total
particular, the risk that some plans would experience
number of people who are insured. Those budgetary
“adverse selection”—that is, that their enrollees will have
effects would primarily reflect modest shifts into or out of
above-average or higher-than-expected costs for health
Medicaid, SCHIP, or employment-based coverage as
care—has important implications for the operation of
those options became more or less attractive relative to
insurance markets and for proposals that would regulate
coverage in the individual market. Proposals to require
those markets or introduce new insurance options.
insurers to cover all applicants or to guarantee coverage of
CBO
SUMMARY
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
XVII
preexisting health conditions would benefit people whose
value, would have proportionally smaller effects on aver-
health care would not be covered otherwise, but insurers
age premiums.
would generally raise premiums to reflect the added costs.
The key features of a managed competition system
Another approach that has attracted attention recently
involve a sponsor, such as an employer or government
involves so-called high-risk pools. Most states have estab-
agency, offering a structured choice of health plans and
lished such pools to subsidize insurance for people who
making a fixed-dollar contribution toward the cost of
have high expected medical costs and have either been
that insurance. Enrollees would thus bear the cost of any
denied coverage in the individual insurance market or
difference in premiums across plans. CBO estimates that
been quoted a very high premium. Overall participation
a proposal requiring that approach would yield average
in high-risk pools is limited—there are currently about
premiums for health insurance that are about 5 percent
200,000 enrollees nationwide—but proposals could seek
lower than those chosen under current law. Proposals that
to expand the use of those pools by providing new federal
also adopted other features of managed competition, such
subsidies. The costs of such subsidies would depend pri-
as standardization of benefits across plans and adjust-
marily on the average health care costs of enrollees, the
ments of sponsors’ payments to those plans to reflect the
share of those costs covered by the pool, and the number
health risk of each enrollee, might yield more intense
of people who enrolled as a result.
competition among plans and help avoid problems of
adverse selection.
Steps to Reveal Relative Costs
Some proposals would seek to restructure the choices
Federally Administered Options
that individuals face—and expose more clearly the rela-
Under some proposals, the federal government would
tive costs of their health insurance options—either by
make available additional insurance options—for exam-
reducing or eliminating the current tax subsidy for
ple, by providing access to the private health plans that
employment-based insurance or by encouraging or
are offered through the Federal Employees Health Bene-
requiring the establishment of managed competition sys-
fits (FEHB) program. The effects of that approach would
tems. Both approaches would provide stronger incentives
depend critically on how the premiums for nonfederal
for enrollees to weigh the expected benefits and costs of
enrollees were set. If insurers could charge different
policies when making decisions about purchasing insur-
premiums to different applicants on the basis of their
ance. As a result, many enrollees would choose health
expected costs for health care, the option would resemble
insurance policies that were less extensive, more tightly
the current individual and small-group markets and thus
managed, or both, compared with the choices made
would have little impact. Alternatively, if new enrollees
under current law.
were all charged the same premium, the FEHB plans
would be most attractive to people who expected to have
The current tax exclusion for the premiums of employ-
above-average costs for health care. If no subsidies were
ment-based health plans provides a subsidy of about
provided, the total premiums charged to nonfederal
30 percent, on average, taking into account both the
enrollees would probably be much higher than those
income and payroll taxes that are avoided. Eliminating
observed in the program today—so the number of new
that exclusion, or replacing it with a fixed-dollar tax
enrollees would probably be limited. Depending on the
credit or deduction, would effectively require employees
specific features of such proposals, providing access to
to pay a larger share of the added costs of joining a more
FEHB plans might not prove to be financially viable
expensive plan; conversely, employees would capture
because of adverse selection into those plans.
more of the savings from choosing a cheaper plan. As a
result, CBO estimates that people would ultimately select
The government could also design an insurance option
plans with premiums that are between 15 percent and
based on Medicare that would be made more broadly
20 percent lower than the premiums they would pay
available, on a voluntary basis, to the nonelderly popula-
under current law. Less extensive changes, such as cap-
tion. The federal costs per enrollee would depend primar-
ping the amount that may be excluded at a certain dollar
ily on the benefits that system provided; the rates used to
CBO
XVIII
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
pay doctors, hospitals, and other providers of health care;
rates, affect the supply of health care services by influenc-
and the extent of any premium subsidies that were
ing the decisions that doctors, hospitals, and other pro-
offered to enrollees—all of which could differ from
viders of care make about how many patients to serve and
Medicare’s current design. As for whether such a plan
which treatments their patients will receive. Average
would be more or less costly than a private health insur-
payment rates for Medicare, Medicaid, and private insur-
ance plan that provides the same benefits to a representa-
ers also differ, which would affect the budgetary impact
tive group of enrollees, the answer would vary
of proposals that shifted enrollees—and their costs—
geographically. Assuming that Medicare’s current rules
from one source of coverage to another. Changes in pay-
applied, those costs would be comparable in many urban
ment rates for public programs or in the amount of
areas, but in other areas the cost of the government-run
uncompensated care provided to the uninsured could also
plan would be lower (as is evident in the current program
affect private payment rates.
through which Medicare beneficiaries may enroll in a pri-
vate health plan). At the same time, because Medicare
Payment Methods and Providers’ Incentives
currently provides broad access to doctors and hospitals
Most care provided by physicians in the United States is
and employs little benefit management, a Medicare-based
paid for on a fee-for-service basis, meaning that a separate
option might attract relatively unhealthy enrollees, which
payment is made for each procedure, each office visit, and
could drive up its premiums, federal costs, or both.
each ancillary service (such as a laboratory test). Hospitals
are generally paid a fixed amount per admission (a bun-
Many of the same considerations would arise in designing
dled payment to cover all of the services that the hospital
a single-payer, Medicare-for-all system, but that approach
provides during a stay) or an amount per day. Such pay-
might raise some unique issues as well—and the scale of
ments may encourage doctors and hospitals to limit their
its impact on federal costs could obviously be much larger
own costs when delivering a given service or bundle, but
if nearly all of the population was covered. Enrollees
they can also create an incentive to provide more services
could be offered a choice of plans under a single-payer
or more expensive bundles if the additional payments
system (as happens in Medicare). If, instead, only one
exceed the added costs.
design option was offered and all residents were required
to enroll in it, then concerns about adverse selection
Other arrangements, such as salaries for doctors or peri-
would not arise. That approach could also reduce the
odic capitation payments (fixed amounts per patient), do
administrative costs that doctors and hospitals currently
not provide financial incentives to deliver additional ser-
incur when dealing with multiple insurers. The lack of
vices. Those approaches raise concerns, however, about
alternatives with which to compare that program, how-
providers’ incentives to stint on care or avoid treating
ever, could make it more difficult to assess the system’s
sicker patients. One study randomly assigned enrollees to
performance. More generally, that approach would raise
different health plans and found that those in an inte-
important questions about the role of the government in
grated plan (which owns the hospitals used by enrollees
managing the delivery of health care.
and pays providers a salary) used 30 percent fewer ser-
vices than enrollees in a fee-for-service plan, but whether
Factors Affecting the Supply and
those results could be replicated more broadly is unclear.
Prices of Health Care Services
Proposals could seek to change payment methods either
The ultimate effects of proposals on the use of and spend-
indirectly or directly. They could change the payment
ing for health care depend not only on factors that affect
methods used by private health plans indirectly by
the demand for health care services, such as the number
encouraging shifts in enrollment toward plans that have
of people who are insured and the scope of their coverage,
lower-cost payment systems. For public programs, such as
but also on factors that affect the supply and prices of
Medicare and Medicaid, federal policymakers could
those services. The various methods used for setting
directly change payment methods. In either case, making
prices and paying for services, and the resulting payment
those changes could prove to be very difficult.
CBO
SUMMARY
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
XIX
Payment Rates
their demand for care but also on how that increase
The financial incentives created by different payment
affected the supply and prices of services. Within a
systems—and the spending amounts they yield—also
10-year budgetary time frame, the number of U.S.-
depend on the level at which payment rates or prices are
trained physicians is largely fixed, so adjustments would
set. Those rates depend partly on the methods that are
have to occur in other areas—which could include
used to set them. Private-sector payment rates are set by
changes in the number of hours doctors worked or in
negotiation, reflecting the underlying costs of the services
their productivity, inflows of foreign-trained physicians,
and the relative bargaining power of providers and health
or changes in doctors’ fees and patients’ waiting times.
plans; in turn, bargaining power depends on factors such
as the number of competing providers or provider groups
Whether and to what extent the supply of physicians
within a local market area. Fee-for-service payment rates
would become constrained also depends on the size of the
in Medicare and Medicaid are generally set administra-
increase in demand for their services and the amount of
tively. That method poses a number of challenges, includ-
time available for adjustments to occur. CBO’s analysis
ing how to determine providers’ costs—particularly for
indicates that providing the uninsured population with
services that require substantial training or that become
coverage that is similar to a typical employment-based
cheaper to provide when they are performed more
plan would increase total demand for physicians’ services
frequently. Additional issues that arise include how to
and hospital care by between 2 percent and 5 percent.
account for the quality of those services and their value to
If payment rates rose in response to that increase in
patients, and what impact rate setting might have on the
demand, the impact on spending could be larger. Spend-
development of new medical technology.
ing on behalf of previously uninsured people would also
increase to the extent that the uncompensated care they
On average, payment rates under Medicare and Medicaid
had received became compensated.
are lower than private payment rates. Specifically,
Medicare’s payment rates for physicians in 2006 were
Uncompensated Care and Cost Shifting
nearly 20 percent lower than private rates, on average,
Another issue that arises when analyzing payment rates is
and its payment rates for hospitals were as much as
whether relatively low rates for public programs or the
30 percent lower. As for Medicaid, recent studies indicate
costs of providing uncompensated or undercompensated
that its payment rates for physicians and hospitals were
care to the uninsured lead to higher payment rates for
about 40 percent and 35 percent lower, respectively, than
private insurers—a process known as cost shifting. To the
private rates. Within Medicare, and probably within
extent that such cost shifting occurs now, proposals that
Medicaid as well, those differentials vary geographically
reduced the uninsured population or that switched
and tend to be larger in rural areas and smaller in urban
enrollees from public to private insurance plans could
areas (where competition among providers is generally
affect private payment rates and thus alter insurance pre-
greater). Given those differences, proposals that shifted
miums. For that to occur, however, doctors and hospitals
enrollment between private and public plans could have a
would have to lower the fees they charged private health
large impact on payments to providers and on spending
plans in response to a decline in uncompensated care or
for health care. Depending on how providers responded
an increase in their revenues from insured patients.
to those changes, enrollees’ access to care could be
affected.
Overall, the effect of uncompensated care on private-
sector payment rates appears to be limited. According to
Responses to Changes in Demand or Payment Rates
one recent set of estimates, hospitals have provided about
Changes in payment rates could also have an indirect
$35 billion in uncompensated care in 2008, representing
effect on spending by altering the number of services that
roughly 5 percent of their total revenues. Roughly half of
providers would be willing to supply. Similarly, the bud-
those costs may be offset, however, by payments under
getary effects of covering previously uninsured individu-
Medicare and Medicaid to hospitals that treat a dispro-
als would depend not only on the resulting increase in
portionate share of low-income patients. Estimates of
uncompensated care provided by doctors are considerably
CBO
XX
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
smaller, amounting to a few billion dollars, so the costs of
rapidly because it largely built on the existing infrastruc-
providing such care do not appear to have a substantial
ture of the state-operated Medicaid program.
effect on private payment rates for physicians.
Maintenance-of-Effort Requirements
Whether and to what extent payments to hospitals under
A proposal that created new subsidies for health insur-
Medicare and Medicaid fall below the costs of treating
ance could lead employers or states to scale back the cov-
those patients is more difficult to determine. Recent stud-
erage that they sponsor, particularly if a new federally
ies indicate, however, that when payment rates change
funded program provided similar or more generous bene-
under those programs, hospitals shift only a small share of
fits. To prevent such responses or offset their effects on
the savings or costs to private insurers (the same logic
federal spending, proposals could include maintenance-
would apply for uncompensated care). Instead, lower
of-effort provisions. Monitoring and enforcing such
payment rates from public programs or large amounts of
requirements for private firms would be difficult,
uncompensated care may lead hospitals to reduce their
however, unless proposals specified effective reporting
costs, possibly by providing care that is less intensive or of
mechanisms and sufficient penalties for violations.
lower quality than would have been offered had payments
per patient been larger.
States’ maintenance-of-effort provisions are generally
structured in two ways: requiring states to maintain exist-
Administrative Issues and Effects on
ing programs at historical eligibility or benefit levels (as is
done under SCHIP); or requiring states to continue
Other Programs
spending funds at certain historical or projected levels or
The extent to which proposals would affect health insur-
to return some of their savings to the federal government
ance coverage or federal budgetary costs, and the timing
(as is done for the Medicare drug benefit). The effective-
of those effects, would depend partly on the administra-
ness of such requirements would depend on how they
tive responsibilities and costs that those proposals
were defined, the enforcement mechanisms that were
entailed and partly on their interactions with other gov-
specified, and the incentives for states to comply. The
ernment programs. Other factors would affect coverage
provisions for SCHIP and the Medicare drug benefit are
and cost, including the impact of any maintenance-of-
examples of effective approaches.
effort provisions that might be applied to states or
employers and the treatment of various segments of the
Effects on Other Federal Programs
population, including people who are ineligible for
Proposals could also have unintended effects on eligibility
current government health programs and those who—
for other federal programs that are not directly related to
although eligible—are generally difficult to reach and
health care. New subsidies for health insurance may be
enroll.
counted as income or assets when determining eligibility
for benefits in means-tested programs (such as the Sup-
Administrative Issues
plemental Nutrition Assistance Program, formerly known
Proposals could require both federal and state govern-
as the Food Stamp program) unless explicitly excluded by
ments to assume new administrative responsibilities and
law. Proposals that changed the employment-based health
could allocate those responsibilities to new or existing
insurance system could shift compensation between
agencies. How well agencies fulfilled new missions—and
wages and fringe benefits, thus affecting eligibility for
how long it would take them to do so—would depend on
government benefits (including Social Security) or tax
their scope and funding. Even with adequate funding,
credits (such as the earned income tax credit) that are
implementing a major initiative might take several years,
based on cash earnings. Temporary or aggregate adjust-
as illustrated by the experience with the new Medicare
ments could be made to benefit formulas in order to min-
drug benefit. Another way to ease the implementation of
imize any adverse effects, but some recipients might still
a new federal program would be to build on existing pro-
be made worse off.
grams; SCHIP, for example, was implemented relatively
CBO
SUMMARY
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
XXI
Treatment of Certain Populations
help discourage the use of less clinically effective or
The treatment of certain populations would present vari-
less cost-effective treatments;
ous administrative challenges for proposals to expand
coverage. Some individuals, including military personnel
B Expanding the use of health information technology,
and veterans, already receive health insurance from the
such as electronic medical records, which would make
federal government, and issues might arise regarding the
it easier to share information about patients’ condi-
tions and treatments; and
coordination of their current benefits with new federal
subsidies. In other cases, federal health programs cur-
B Modifying the system for determining and penalizing
rently deny benefits to certain populations, such as unau-
medical malpractice.
thorized immigrants or prison inmates, and proposals
would have to specify whether and how those restrictions
Some of those initiatives could improve individuals’
would apply to new programs. Other populations, such
health or enhance the quality of the care that they receive,
as the homeless, face challenges enrolling in existing pro-
but it is not clear that they also would reduce overall
grams, and similar issues might arise in designing new
health care spending or federal costs. In its analysis of
subsidies for health insurance. Those considerations
such initiatives, CBO considers the available studies that
would affect both the costs of proposals and their overall
have been done to assess particular approaches. In many
impact on rates of insurance coverage.
cases, those studies do not support claims of reductions in
health care spending or budgetary savings.
Changes in Health Habits and
Medical Practices
For several reasons, it may be difficult to generate reduc-
tions in health care spending from such initiatives. In
In addition to any broader changes they make in the
some cases, the problem is largely one of identifying and
health insurance and payment systems, proposals may
targeting the people whose participation would cause
include specific elements designed to induce individuals
health care spending to decline. Broad programs aimed at
to improve their own health or to encourage changes in
preventive medical care and disease management could
how diseases are treated. Through a combination of
reduce the need for expensive care for a portion of the
approaches, proposals could try to change the behavior of
recipients but could also provide additional services—and
both patients and providers by:
incur added costs—for many individuals who would not
have needed costly treatments anyway. To generate net
B Promoting healthy behavior, including measures
aimed at reducing rates of obesity and smoking;
reductions in spending, the savings that such interven-
tions generated for people who would have needed
B Expanding the use of preventive medical care, which
expensive care would therefore have to be large enough to
can either impede the development or spread of a dis-
offset the costs of serving much larger populations.
ease or detect its presence at an early stage;
A related issue is that many individuals or health plans
B Establishing a “medical home” for each enrollee, typi-
might already be taking the steps involved (or will in the
cally involving a primary care physician who would
future) even in the absence of a new requirement or
coordinate all of his or her care;
incentive. The effect of any proposal would have to be
measured against that trend, and a large share of any sub-
B Adopting “disease management” programs that seek to
sidies involved might go to people who (or health plans
coordinate care for and apply evidence-based treat-
ments to certain diseases, such as diabetes or coronary
that) would have taken those steps even if there were no
artery disease;
requirements or incentives to do so. For example, some
doctors and hospitals are already using electronic medical
B Funding research comparing the effectiveness of dif-
records, and more will adopt that technology in the
ferent treatment options, the results of which could
future under current law, so any subsidy payments those
CBO
XXII
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
providers might receive under a proposal would add to its
risk of being sued. CBO has not found consistent evi-
costs but would not affect adoption rates.
dence of such broader effects, but that may reflect the
difficulty of disentangling the impact of changes to the
In other cases, the effect on health care spending depends
medical malpractice system from other factors affecting
crucially on whether doctors and patients have incentives
medical costs.
to change their use of health care services. For example,
studies may find that a given treatment has fewer clinical
benefits or is less cost-effective (meaning that added costs
Effects on Total Health Care Spending,
are high relative to the incremental health benefits) for
the Scope of the Federal Budget, and
certain types of patients. But those results may not have a
the Economy
substantial effect on the use of that treatment unless the
Proposals that would substantially change the health
financial incentives facing doctors (through their pay-
insurance market could have an effect on total spending
ments) or patients (through their cost sharing) are aligned
on health care, the flow of payments between various sec-
with the findings. Similarly, proposals to establish a med-
tors of the economy, and the operation of the U.S. econ-
ical home may have little impact on spending if the pri-
omy. CBO will consider those effects in its analysis of
mary care physicians who would coordinate care were not
major health care proposals.
given financial incentives to limit their patients’ use of
specialists.
Effects on Total Spending and the Scope of the
Federal Budget
Other types of initiatives might ultimately yield substan-
Many health insurance proposals would have an impact
tial long-term health benefits but might not generate
on total spending for health care, and some might con-
much savings, at least in the short term. Even if success-
tain provisions that explicitly limit the level of or rate of
ful, measures to reduce smoking and obesity—two factors
growth in health care spending; such proposals might
linked to the development of chronic and acute health
impose a global budget or budgetary cap on all or a part
problems—might not have a substantial impact on health
of that spending. The effectiveness of such strategies
care spending for some time. In the long term, spending
would depend on several factors, including the scope of
on diseases caused by poor health habits could decline
the global budget, the targets selected for different cate-
substantially, but the impact on federal costs would also
gories of spending, and the mechanisms used to enforce
have to account for people living longer and receiving
the caps.
more in Medicare benefits (for the treatment of other dis-
eases and age-related ailments) as well as other govern-
In addition to their overall effects on federal spending
ment benefits that are not directly related to health care
and revenues, proposals that made substantial changes to
(including Social Security benefits). Similarly,
the health insurance system or its financing methods
investments in health information technology might
could raise a number of budgetary issues. Such proposals
require substantial start-up costs that would be difficult
could have substantial effects on the flows of payments
to recapture in the typical 5- and 10-year budgetary time
among households, employers, and federal and state gov-
frames used to evaluate legislative proposals.
ernments—even if the proposals were budget neutral
from a federal perspective. Some proposals might assign
Demonstrating savings might also be difficult because of
the federal government a more active role in the health
data limitations and methodological concerns. For
insurance market; for example, the government could be
example, studies have found that tort limits, by reducing
required to disburse subsidies covering the cost of health
malpractice awards, cause premiums for malpractice
insurance, collect health insurance premiums from poli-
insurance to fall and thus could have a very modest
cyholders, or make payments to insurers. Any of those
impact on doctors’ fees and health care spending. Some
changes might raise questions regarding who—the
observers argue that tort limits would yield larger reduc-
government, the insured, or the insurer—would bear
tions in that spending because doctors would stop order-
financial responsibility for any shortfalls in payments that
ing unnecessary tests and taking other steps to reduce the
might occur.
CBO
SUMMARY
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS XXIII
Other proposals might require that individuals or
average wages could affect the allocation of workers
businesses make payments directly to nongovernmental
among firms.
entities. Depending on the specific provisions of such
proposals, CBO might judge that payments resulting
B Proposals that required employers to provide health
from federal mandates should be recorded as part of the
insurance could adversely affect the hiring of
employees earning at or near the minimum wage,
federal budget even if the funds did not flow through the
because the total compensation of those workers could
Treasury. The extent of federal control and compulsion is
exceed their value to the firm.
a critical element in determining budgetary treatment. In
general, CBO believes that federally mandated pay-
Some observers have asserted that domestic firms provid-
ments—those resulting from the exercise of sovereign
ing health insurance to their workers incur higher costs
power—and the disbursement of those payments should
for compensation than do competitors based in countries
be recorded in the budget as federal transactions.
where insurance is not employment based and that fun-
damental changes to the health insurance system could
Effects on the Economy
reduce or eliminate that disadvantage. Although U.S.
Proposals that made large-scale changes affecting the pro-
employers may appear to pay most of the costs of their
vision and financing of health insurance could also have
workers’ health insurance, economists generally agree that
an impact on the broader economy. Because most health
workers ultimately bear those costs. That is, when firms
insurance is currently provided through employers, pro-
provide health insurance, wages and other forms of
posals could affect labor markets by changing individuals’
compensation are lower (by a corresponding amount)
decisions about whether and how much to work and
than they otherwise would be. As a result, the costs of
employers’ decisions to hire workers. Such effects could
providing health insurance to their workers are not a
arise in several ways:
source of competitive disadvantage for U.S.-based firms.
B Proposals that decreased the return from an additional
In addition to their effects on the labor market, proposals
hour of work, by imposing new taxes or phasing out
could also affect the size of the nation’s stock of
subsidies or credits for health insurance as earnings
productive capital, especially through their effects on
rise, could cause some people to work fewer hours or
government budgets. Those effects would depend partly
leave the labor force.
on how the costs of any insurance expansions or other
B Proposals that made health insurance less dependent
changes are financed, which is beyond the scope of this
on employment status could induce some people to
report. The net effect on the economy of a broad pro-
retire earlier and others to change jobs more often.
posal to restructure the health insurance system would,
not surprisingly, depend crucially on the details.
B Proposals that treated firms differently on the basis of
such characteristics as the number of employees or
CBO
C H A P T E R
1
Introduction and Background
The health care provided in the United States can outlays for the Medicare and Medicaid programs are pro-
confer tremendous benefits, extending and improving
jected to reach about $720 billion in 2009.
lives. But the high and rising costs for health care in this
country impose an increasing burden on individuals,
Over the longer term, rising health care costs represent
businesses, states, and the federal government, and a sub-
the single greatest challenge to balancing the federal bud-
stantial number of people may have trouble paying for
get. The Congressional Budget Office (CBO) projects
that care because they do not have health insurance.
that under current law, federal spending on Medicare and
Those issues are related: People seek insurance to protect
Medicaid alone will grow from about 4 percent of GDP
themselves against the risk of experiencing financial hard-
in 2009 to about 12 percent of GDP by 2050.
ship when they need expensive medical care, and yet ris-
The number of people who lack health insurance is also
ing costs for health care also make health insurance poli-
on the rise. Recent estimates indicate that about one in
cies more expensive and thus more difficult to afford.
four nonelderly residents (those age 64 or younger) is
Lack of insurance can limit access to care, but having
uninsured at some point during the year, and at any given
insurance can increase spending by encouraging the use
point in 2009 at least 45 million residents are expected to
lack insurance—a figure that has grown rather steadily
of services that have relatively low clinical value. More
over time and that CBO projects to reach about 54 mil-
generally, despite spending more per capita than other
lion by 2019.1 The nonelderly population will also
countries, the United States lags behind lower-spending
increase, so the share of that population projected to be
countries on several metrics, including life expectancy
uninsured at a given point during the year will grow more
and infant mortality. Indeed, evidence
slowly, from about 17 percent in 2009 to about 19 per-
suggests that a substantial share of spending on health
cent in 2019. (Those estimates for 2009 do not reflect
care contributes little if anything to the overall health of
the recent deterioration in economic conditions, which
the nation, but finding ways to reduce such spending
could result in a larger uninsured population.) The num-
without also affecting services that improve health will
ber of uninsured individuals is expected to increase
be difficult.
because health insurance premiums are likely to continue
Spending on health care services and related activities
rising considerably faster than income, which will make
accounts for about 17 percent of gross domestic product
insurance more difficult to afford.
(GDP)—an expected total of about $2.6 trillion in
2009—and under current law that share is projected to
1. Those figures are CBO estimates and differ somewhat from the
estimates of the uninsured population released annually by the
reach nearly 20 percent by 2017 (according to the Cen-
Census Bureau (the most recent of which indicated that about
ters for Medicare and Medicaid Services). Correspond-
46 million people were uninsured at any given point in 2007).
ingly, annual health expenditures per capita are projected
Like other estimates, CBO’s figures represent the civilian, non-
institutionalized population; they therefore include unauthorized
to rise from about $8,300 to about $13,000 in nominal
immigrants but exclude residents of nursing homes, people who
(current-dollar) terms over that period. Federal spending
are incarcerated, and U.S. citizens living abroad. See Chapter 6 for
accounts for about one-third of those totals, and federal
further discussion of such populations.
CBO
2
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
Scope and Focus of This Report
B Replace employment-based coverage with new pur-
The challenge of seeking to increase the number of
chasing arrangements or provide strong incentives for
people who have health insurance while attempting to
people to shift toward individually purchased
constrain the high and rising costs of health care has led
coverage.
to proposals that would substantially modify the health
B
insurance system in this country. Because the Medicare
Provide all citizens or residents with coverage under, or
program already provides nearly universal coverage to the
access to, existing insurance plans such as the Federal
elderly, those proposals generally focus on options for
Employees Health Benefits program or the Medicare
providing coverage to and reducing costs for the
program (or modified versions of such plans), either as
nonelderly population; even so, options that would
an additional option or under a “Medicare-for-all”
reduce Medicare spending could be used to offset the
single-payer arrangement.
government’s costs for insurance expansions and could
Wherever possible, the analysis presented here describes
have broad effects on spending for health care.
in quantitative terms how CBO would estimate the bud-
This report describes the assumptions that CBO would
getary and other effects of such proposals. In other cases,
use in estimating the effects of key elements of such
it describes the components that a proposal would have
proposals on federal costs, insurance coverage, and other
to specify in order to permit estimation of its effects on
outcomes; the evidence on which those assumptions are
federal spending or other outcomes. This report reflects
based; and, if the evidence points to a range of possible
the current state of CBO’s analysis of and judgments
effects rather than a precise prediction, the factors that
about the likely response of individuals, employers, insur-
would influence where a proposal falls within those
ers, and providers to changes in the health insurance and
ranges. (A companion report—Budget Options, Volume 1:
health care systems. The details of particular policy speci-
Health Care—describes CBO’s estimates of the effects on
fications and the way in which they are combined, as well
the federal budget of numerous specific options related to
as new evidence or analysis related to the issues discussed
health care and health insurance.)
here, could affect future CBO estimates of the effects of
large-scale health insurance proposals.
This document does not provide a comprehensive
analysis of any specific proposal; rather, it identifies and
Because such proposals could incorporate a number of
discusses many of the critical factors that would affect
the elements that are discussed in this report, they could
estimates of various proposals. In particular, it considers
have interactions that are difficult to predict. Those pro-
the types of issues that would arise in estimating the bud-
posals could also affect both tax revenues and outlays.
getary effects of proposals to:
Estimates of the impact on revenues of proposals to
change the federal tax code are prepared by the staff of
B Provide tax credits or other types of subsidies to make
the Joint Committee on Taxation (JCT) and would be
insurance less expensive to the purchaser.
incorporated into any formal CBO estimate of a pro-
posal’s effects on the federal budget. In preparing this
B Require individuals to purchase health insurance,
report, CBO consulted with the JCT staff on the behav-
typically paired with a new system of government
ioral considerations that are incorporated into both agen-
subsidies.2
cies’ estimates.
B Require firms to offer health insurance to their
CBO’s analysis of whether people obtain insurance
workers or pay into a fund that subsidizes insurance
focuses on the scope of the coverage that the available
purchases.
policies provide and the cost to them. All else being
equal, more comprehensive coverage will be more expen-
2. The issue of whether a policy proposal would be considered a
sive. The cost of an insurance policy also depends on who
mandate for purposes of the Unfunded Mandates Reform Act and
related questions regarding the contents of a formal cost estimate
enrolls in the plan; how much health care they seek; and
are discussed in Chapter 8.
how much doctors, hospitals, and other providers of care
CBO
CHAPTER ONE
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
3
are paid for their services. Although those considerations
The question of whether and how any net increases in
are closely related, this report analyzes the following
federal spending for health care and health insurance
questions:
would be financed by policy changes outside the health
sector is beyond the scope of this report. Whether a pro-
B For insurance policies with the same scope and total
posal makes health insurance more affordable for a given
cost, how does the share of that cost that individuals
individual or family would depend not only on its impact
have to pay affect whether they purchase insurance?
on the health insurance premiums that they face but also
How would various types of subsidies that reduce the
on the effect that its financing mechanisms have on the
cost to them directly or indirectly—or mandates to
household’s budget. To the extent that such proposals are
offer or purchase coverage—affect the rates and
financed by provisions that fall outside the health sec-
sources of insurance coverage?
tor—through increases in tax revenues or reductions in
spending for other federal programs—those effects are
B How does the cost of an insurance policy vary with the
not addressed in this report.
scope of its coverage, insurers’ use of various cost-
management techniques, and the types of people it
As background for the discussion of the broad policy
covers? How would health care spending and average
options presented in subsequent chapters of this report,
policy premiums be affected by extending coverage to
the remainder of this chapter describes the primary
people who are now uninsured?
sources of health insurance coverage, the reasons that
people lack coverage, the extent and nature of the cover-
B Taking the demand for insurance overall and the pre-
age that is currently purchased, and the main compo-
miums charged for various options as given, how are
nents and drivers of health care spending.
individuals’ decisions about which policy to choose
affected by the laws and regulations governing those
choices? How would consumers respond to changes in
Health Insurance Coverage
the structure of or incentives governing the insurance
The primary purpose of health insurance is to protect
market?
individuals against the risk of financial hardship when
they need expensive medical care. In principle, most peo-
B What impact do factors affecting the supply of health
ple would be willing to pay an insurance premium that
care services and the level and mechanism of payments
was somewhat higher than their own expected costs for
to providers have on the costs of health care and insur-
health care in order to avoid that risk, but in practice
ance premiums? How would changes in those supply
many people with low income or high expected costs
factors interact with demand to determine future
might consider the premiums they would face to be
spending on health care?
unaffordable.
Proposals to modify the health insurance system that
Over the years, various policies have been adopted that
include subsidies would probably have the most immedi-
subsidize insurance coverage for certain groups. Medicare
ate and direct impact on the federal budget. Their costs
provides highly subsidized coverage to the elderly and
would depend primarily on the nature and extent of
also insures several million people under the age of 65
those subsidies, the number of people who take advan-
who are disabled—two groups that have relatively high
tage of them, and the scope of insurance coverage that is
costs for health care. The Medicaid program and related
purchased or provided as a result. This report also consid-
initiatives offer free or low-priced coverage to many
ers other effects, including any federal administrative
children and (to a more limited degree) their parents;
costs and challenges that might be involved in imple-
Medicaid also covers many elderly and disabled individu-
menting a proposal; the effects on eligibility for and
als who have low income and few assets (and thus would
spending under other federal programs; the impact of
have difficulty paying for insurance). Most employers
provisions that seek to reduce spending on health care by
offer health insurance to their workers and most workers
encouraging consumers to make healthier choices and
enroll in a plan, motivated in part by a tax subsidy for
providers to change some of the ways in which they
employment-based insurance. People may also be able to
practice medicine; and other macroeconomic effects or
purchase coverage in the individual insurance market, but
budgetary implications that a proposal might have.
that coverage is not generally subsidized. Those sources of
CBO
4
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
Table 1-1.
The availability of health insurance affects not only who
Sources of Insurance Coverage and
enrolls but also how much health care people consume.
Insurance Status of the Nonelderly
People who are insured are likely to use more health care
Population, 2009
than they would if they had to pay the full costs of those
services—a phenomenon economists call “moral hazard.”
To offset that tendency toward increased use, health
Number
insurance policies typically feature some degree of cost
(Millions)
Percent
sharing by enrollees. Health plans may also seek to con-
Source of Coverage
trol their costs and premiums by using various methods
Employment-Baseda
160
61
of managing care and by varying the range of benefits
Individually Purchased
10
4
offered. Of course, those features also affect the premi-
Medicare
7
3
Medicaidb
43
17
ums for health insurance policies and the attractiveness of
Otherc
12
4
the coverage to enrollees.
Insurance Status
Sources of Insurance Coverage
Insured, Any Sourced
216
83
In the United States, most people obtain health insurance
Uninsured
45
17
coverage from either public or private sources, but about
17 percent of the nonelderly population will be unin-
Source:
Congressional Budget Office’s health insurance simulation
model.
sured in 2009 (see Table 1-1).3 Insurance obtained
Note: The nonelderly population excludes people in institutions
through an individual’s employment is the primary
and residents of U.S. territories.
source of coverage for the nonelderly.
a.
Includes coverage obtained through local, state, and federal
employers.
Employment-Based Insurance. In 2009, roughly 160 mil-
b.
Includes the State Children’s Health Insurance Program.
lion people under the age of 65—or about three out of
every five nonelderly Americans—are expected to have
c.
Includes military and other sources of coverage.
health insurance that is provided through an employer or
d.
The sum of people by their sources of coverage exceeds the
total number who are insured because about 14.5 million
other job-related arrangement, such as a plan offered
people are covered by more than one source at a time.
through a labor union. That figure includes active work-
ers, spouses and dependents who are covered by family
coverage also vary in the ease of enrollment, which affects
policies, and nonelderly retirees.
their attractiveness.
One prominent feature of employment-based insurance
Because health insurance provides more benefits to peo-
is that employers generally contribute a large share of the
ple who incur relatively high costs for health care, health
total premium; that is, the amount that is directly and
insurance coverage generally—or specific health insur-
ance plans—may attract enrollees with above-average
visibly deducted from workers’ paychecks for health
costs, a phenomenon known as “adverse selection.”
insurance (called the employees’ contribution) usually
Conversely, people with low expected costs for health care
represents a relatively small share of the average cost per
may be reluctant to pay an insurance premium that
enrollee. According to a survey of firms conducted in
reflects the average costs of all enrollees, or they might
2008, employers contribute 73 percent of the cost of a
prefer to wait until they develop a health problem to sign
family policy for their workers and 84 percent of the cost
up for coverage. To the extent that such adverse selection
occurs, average insurance premiums (or the costs of gov-
3. Estimates of health insurance coverage presented in this report are
ernment subsidies for insurance) would tend to rise to
derived from a simulation model that the Congressional Budget
reflect the higher spending per enrollee. The potential for
Office (CBO) developed in order to analyze the effects of various
policy options on coverage and spending for health care. For a
adverse selection exists with almost any health insurance
detailed description of that model and the data and evidence on
plan, but the manner in which it arises and the mecha-
which it is based, see CBO’s Health Insurance Simulation Model:
nisms used to address it differ across insurance markets.
A Technical Description, Background Paper (October 2007).
CBO
CHAPTER ONE
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
5
of single coverage, on average.4 One reason employers
would adjust to changes in those contributions is less
make those contributions is to encourage broad participa-
clear. In principle, workers who would obtain more bene-
tion by their employees, so as to limit the potential for
fits from health insurance coverage—such as older work-
adverse selection.
ers, who have higher average costs for health care—would
be willing to accept a greater reduction in their wages
Although employers may appear to pay most of the costs
than other workers would accept in return for that cover-
of their workers’ health insurance, economists generally
age. The extent to which that phenomenon occurs in
agree that workers ultimately bear those costs. Employers’
practice, however, is uncertain.7 Similarly, it could take
contributions are simply a form of compensation, and if
labor markets several years to adjust to unexpected
labor markets are competitive (which is generally the
changes in employers’ costs for health care. For purposes
case), an employee’s total compensation should equal his
of estimating the impact of proposed legislation, however,
or her contribution to the revenue of the firm. Thus,
CBO makes the simplifying assumption that total com-
when an employer offers to pay for health insurance, it
pensation is fixed and that changes in the costs of health
pays less in wages and other forms of compensation than
insurance translate immediately into offsetting changes in
it otherwise would, keeping total compensation about the
wages and other forms of compensation; the JCT staff
same.5
makes the same assumption when estimating the effects
That relationship can be difficult to observe and may not
of proposals on revenue collections.
hold perfectly for every worker at every instant. In partic-
ular, workers who turned down an employer’s offer of
Compared with the individual insurance market,
subsidized health insurance generally would not see an
employment-based coverage offers several advantages,
immediate or corresponding increase in their wages.
particularly for employees of larger firms. Unlike wages,
Moreover, firms offering health insurance actually tend to
the employer’s costs for providing that coverage are
pay higher wages than firms that do not do so, but those
excluded from the enrollee’s taxable income. As a result,
differences in total compensation reflect disparities in the
that portion of employees’ compensation is not subject to
skill and productivity of the workers, not a failure to pass
individual income and payroll taxes. In addition, most
on the costs of providing insurance. For their part, many
employees are also able to exclude the portion of the pre-
employers behave as though they do bear the costs of the
mium that they pay. For a typical worker, that favorable
insurance plans they offer (as reflected in their efforts to
tax treatment provides a subsidy from the government
control those costs). Nevertheless, the available evidence
that reduces the net cost of employment-based health
indicates that employees as a group ultimately bear the
insurance by about 30 percent.
costs of any payments an employer makes for health
insurance.6
That tax subsidy provides an incentive for workers to
obtain insurance through their employer and for their
How the costs of employers’ contributions are allocated
employer to provide it. Because out-of-pocket costs for
among different types of workers and how quickly wages
health care do not generally receive a tax subsidy, workers
also have an incentive to secure more extensive coverage,
4. Henry J. Kaiser Family Foundation and Health Research and
thereby increasing the share of spending for health care
Educational Trust (Kaiser/HRET), Employer Health Benefits: 2008
that is covered and decreasing the share that they pay out
Annual Survey (Washington, D.C.: Kaiser/HRET, September
of pocket. The value of the exclusion from taxation is
2008).
generally somewhat larger for workers with higher
5. Even if a given labor market was not competitive, firms operating
income because they face higher income tax rates
in that market would still be expected to hold total compensation
(although they may also face lower rates of payroll
fixed, so that other forms of compensation would be reduced to
offset the costs of providing health insurance. The allocation of
taxation).
compensation among wages, health insurance, and other fringe
benefits would reflect the preferences of workers and the firms’
7. One study examined the impact of a state mandate to cover
efforts to attract employees.
maternity benefits and found that reductions in the wages of
6. For a discussion of that evidence, see Jonathan Gruber, “Health
women of child-bearing age and their spouses roughly offset the
Insurance and the Labor Market,” in A.J. Culyer and
average costs of providing those benefits. See Jonathan Gruber,
J.P. Newhouse, eds., Handbook of Health Economics, vol. 1
“The Incidence of Mandated Maternity Benefits,” American
(Amsterdam: North Holland, 2006), pp. 645–706.
Economic Review, vol. 84, no. 3 (June 1994), pp. 622–641.
CBO
6
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
Box 1-1.
Regulation of Health Insurance and the Employee Retirement Income
Security Act
In the United States, some forms of private health
employer that bears the financial risk of providing it
insurance are subject to both state and federal regula-
and that 72 million will have coverage from an
tion, but others are exempt from state regulation.
insurer that is subject to state regulation. (Policies
That distinction, which is a common source of con-
covering another 10 million enrollees that are bought
fusion, stems from the treatment of employment-
in the individual insurance market are also regulated
based health plans under the Employee Retirement
by the states.) Large firms are more likely to bear
Income Security Act of 1974 (ERISA). Under that
insurance risk for their workers; according to one sur-
act, employers that bear the financial risk of covering
vey, 86 percent of workers at firms with 5,000 or
their workers’ health insurance claims—and thus
more employees were in such plans in 2007, com-
effectively serve as the insurer—are exempt from state
pared with 12 percent of workers at firms with fewer
insurance laws and regulations. If, instead, an
than 200 employees.1
employer contracts with an insurance company to
provide coverage and that company bears the associ-
Confusion about the implications of ERISA may
ated financial risk, then state insurance laws and
stem in part from the terminology that is used to
oversight apply.
describe its provisions and from subtle distinctions
about the roles of employers and insurers. Employers
The main practical effect of the difference in treat-
that bear insurance risk are referred to as having “self-
ment is that employers who serve as the insurer for
insured” or “self-funded” plans, whereas employers
their employees are exempt from the benefit man-
that contract with an insurer are said to have
dates and other insurance regulations that many
“insured” or “fully insured” plans. Many employers
states impose (such as requirements to cover certain
that bear insurance risk still use insurers to carry out
treatments, procedures, or types of providers). A
some functions, such as developing networks of pro-
rationale for that arrangement is that an employer
viders, negotiating payment rates, processing claims,
with operations in several states would otherwise be
and so forth. In those cases, the insurance company is
unable to offer the same coverage to all of its employ-
called a third-party administrator. Further, employers
ees, given the variation in state mandates and regula-
may qualify for ERISA’s exemptions even if they pur-
tions; similarly, complying with the differing require-
chase a separate insurance policy (known as reinsur-
ments in each state might be cumbersome for such an
ance or “stop loss” coverage) to protect themselves
employer.
against unusually high claims, so long as the
employer continues to bear sufficient financial risk.
Of the roughly 160 million people whose primary
insurance will come from an employment-based plan
1. William Pierron and Paul Fronstin, ERISA Pre-emption:
in 2009, the Congressional Budget Office estimates
Implications for Health Reform and Coverage, Issue Brief
that about 88 million will have coverage from an
No. 314 (Washington, D.C.: Employee Benefit Research
Institute, February 2008), www.ebri.org.
CBO
CHAPTER ONE
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
7
Table 1-2.
the premiums that insurers may charge, and other terms
Share of Employees Offered Health
of purchase. (Those regulations are discussed further in
Insurance, by Size of Firm, 2009
Chapter 4.) Policies provided in the large-group market,
by contrast, generally face few legal constraints regarding
their benefits and premiums. One exception is that,
Employees Offered
among workers who are similarly situated (that is, work-
Size of Firm
Total Employees
Health Insurance
ers who are in the same class of employment and work in
(Number of
Number
Number
employees)
(Millions) Percent
(Millions)
Percent
the same geographic location), employers may not vary
employees’ contributions to premiums on the basis of
Fewer than 25
31.0
22
14.9
48
their health.
25 to 99
17.6
13
12.7
72
100 to 999
27.2
19
21.0
77
Whether employers offer coverage largely reflects the
1,000 or More
63.9
46
54.9
86
aggregate preferences of their workers, but for several rea-
All
139.7
100
103.5
74
sons smaller firms are less likely to offer insurance than
larger firms. Overall, about half of the workers at very
Source:
Congressional Budget Office’s health insurance simulation
small firms (those that have fewer than 25 employees) are
model.
offered coverage and are eligible for it, compared with
Employment-based insurance offers a number of other
77 percent of the workers at firms with 100 to 999
advantages. For example, because sales and marketing
employees and 86 percent of the workers at firms with
costs for insurers are relatively fixed, as the number of
1,000 or more employees (see Table 1-2).9 One reason is
that households with lower income find it more difficult
enrollees covered by an employer’s policy increases, those
to accept lower wages in return for health insurance, and
fixed costs can be spread over a larger number of enroll-
smaller firms are more likely to employ low-wage work-
ees. As a result, the average premium needed to purchase
ers. Another reason is that policies purchased by smaller
a given amount of coverage is lower for employees of
firms incur higher administrative costs per enrollee, so
larger firms. Some analysts have suggested that employers
the share of the policy premium that covers medical costs
also act as employees’ agents, using their power to bargain
is lower, reducing the attractiveness of such policies.
for lower premiums, sorting out the employees’ options,
Because employees of larger firms constitute most of the
and making it easier for them to choose an insurance
total workforce, the percentage of all workers who are
plan.8 In particular, employers may take steps that sub-
offered coverage—about three out of four—is closer to
stantially simplify the process of enrolling in a health
the proportion for larger firms.
insurance plan, and the use of automatic payroll
The share of workers who are enrolled in employment-
deduction to pay for employees’ premiums may also
based coverage has varied somewhat over time, partly
encourage participation.
reflecting changes in the mix of employment and partly
tracking fluctuations in the business cycle. According to
Another important feature of employment-based insur-
recent surveys of employers, that share rose from 62 per-
ance is that policies offered by firms of all sizes are subject
cent in 1999 to 65 percent in 2001 but has fallen since
to certain federal requirements, but most policies offered
then and stands at 60 percent in 2008.10 The coverage
by larger firms are exempt from state insurance laws and
rate has been somewhat more volatile for smaller firms
regulations. That distinction stems from the provisions of
(those with fewer than 200 workers); that rate was
the Employee Retirement Income Security Act, which are
described in Box 1-1. As a result, policies offered by
9. Among firms that have similar numbers of workers, the share of
smaller employers generally must comply with require-
firms reporting that they offer coverage to their employees is
ments that vary by state regarding the benefits they cover,
generally larger than the share of employees reporting that they
have an offer, but that discrepancy simply reflects the fact that
some workers at firms that offer coverage are not eligible to enroll
8. Jeff Liebman and Richard Zeckhauser, Simple Humans, Complex
in it. For example, many part-time workers are ineligible.
Insurance, Subtle Subsidies, Working Paper No. 14330
(Cambridge, Mass.: National Bureau of Economic Research,
10. Kaiser/HRET, Employer Health Benefits: 2008 Annual Survey; and
September 2008).
Employer Health Benefits: 1999 Annual Survey (October 1999).
CBO
8
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
52 percent in 1996, rose to 58 percent in 2001, and fell
become disabled or too sick to keep their job may eventu-
back to 52 percent in 2008. Studies have attributed the
ally lose their employment-based coverage.
recent decline in enrollment to a combination of modest
reductions in the number of employers offering insur-
Individually Purchased Insurance. Overall, CBO esti-
ance, shifts in employment toward firms and industries
mates that about 10 million nonelderly individuals will
that are less likely to offer health insurance coverage, and
be covered by a policy purchased in the individual insur-
a reduction in enrollment rates among workers who are
ance market in 2009. In principle, anyone may purchase
offered coverage. The estimated impact of each of those
coverage in that market—to cover only themselves or
factors varies, however, depending on the specific years
their family as well—but in practice that option may be
examined, the data used, and the methodology
more attractive to some people than to others. (Such
employed.
coverage is sometimes called “nongroup” insurance to
distinguish it from group coverage, which is primarily
One source of employment-based health insurance that
employment based.)
has received considerable attention is the Federal
Employees Health Benefits (FEHB) program, which
The potential for adverse selection may be stronger in the
provides coverage to about 8 million active and retired
individual market than in the employment-based market,
federal employees in 2008. Under that program, several
partly because people can apply for individual insurance
private health insurance plans are available nationwide,
at any time and may therefore wait until a health problem
and in most regions employees have a range of local plans
arises before seeking coverage and partly because appli-
available to them as well. The federal government covers
cants do not have to be healthy enough to work. To
75 percent of the cost of each participating plan up to a
address those possibilities, insurers usually “underwrite”
limit set at 72 percent of the national average premium;
the policy—a process by which they assess the health risk
to purchase a policy more expensive than that, the
of applicants. Although most applicants end up being
enrollee has to pay the added costs (although those pay-
quoted a standard premium rate (which usually varies by
ments may also be excluded from taxable income).11 Like
age), underwriting can result in adjustments to premi-
employees of private firms that offer a choice of insurance
ums, adjustments to benefits (for example, to exclude
plans, federal workers may generally sign up for coverage
coverage of known health conditions), or denials of
or change plans only during an annual open-enrollment
coverage. As a result, individuals who have more health
season—a rule that limits their opportunities to wait
problems may face higher premiums when they apply for
until they develop a health problem to enroll or to switch
coverage. Some states, however, prohibit or limit those
plans for health reasons and thus limits the degree of
practices—which generally has the effect of reducing pre-
adverse selection that can occur.
miums charged to older or less healthy applicants and
raising premiums for younger and healthier applicants (as
Although employment-based insurance has certain
discussed further in Chapter 4).
advantages, the central role of employers in sponsoring
coverage also has disadvantages. Unlike federal workers,
Individual insurance products have some other advan-
many employees are not offered a choice of insurance
tages and disadvantages compared with employment-
plans, and others may have only a few plans from which
based coverage. Some applicants may be able to obtain
to select, so the plan in which they enroll might not fit
basic insurance protection (such as “catastrophic cover-
their preferences. Furthermore, employees and their
age” plans) in the individual market at a relatively low
dependents typically have to change plans when changing
cost. That market generally offers consumers a greater
jobs and could become uninsured if their new employer
choice of plans, and the coverage may be portable from
does not offer coverage—potentially making them reluc-
one job to another. Insurers incur greater administrative
tant to switch jobs in the first place (a phenomenon
costs for policies sold in the individual market, however,
known as “job lock”).12 In addition, employees who
12. Workers who previously held employment-based insurance may
11. For more information, see Mark Merlis, “The Federal Employees
seek coverage in the individual insurance market, and insurers
Health Benefits Program: Program Design, Recent Performance,
must generally offer them a policy if they apply, but some workers
and Implications for Medicare Reform” (briefing prepared for the
may find the terms of that coverage unattractive. See Chapter 4
Henry J. Kaiser Family Foundation, May 30, 2003).
for additional discussion.
CBO
CHAPTER ONE
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
9
and those costs are built into the policy premiums. Com-
When it was created, Medicare had two primary compo-
pared with the enrollment process for an employment-
nents: Part A, which generally covers hospital care and
based plan, the effort required of applicants to search for
other services provided by institutions; and Part B, which
a policy and sign up for coverage in the individual market
generally covers physicians’ services and various forms of
can be considerably greater. In general, individually pur-
outpatient care. Enrollment in Part A is free of charge and
chased coverage does not receive favorable tax treatment,
essentially automatic for individuals (and their spouses)
which also makes its effective price higher.13
who have sufficient earnings subject to payroll taxes to
qualify for Social Security benefits; certain others may
Reflecting those disadvantages, participation in the indi-
enroll but must pay a monthly premium. To participate
vidual insurance market is relatively low. Only about
in Part B, enrollees must pay a monthly premium that
1 percent of nonelderly adults who are offered
covers about 25 percent of the program’s average costs.
employment-based coverage (either by their own
Although participation is voluntary, seniors who choose
employer or through a spouse) elect to purchase individ-
not to participate in Part B when they are first eligible are
ual coverage. Even among people who lack other coverage
subject to penalties if they decide to enroll at a later
options, only about 20 percent elect to purchase a policy
date—penalties that are intended to discourage eligible
in the individual market; the rest are uninsured. In many
individuals from waiting to develop a health problem
before they enroll. As a result of those provisions, nearly
cases, individually purchased policies are held for rela-
95 percent of individuals who are eligible to enroll in
tively short periods of time—serving to cover individuals
Part B do so. Many of those who do not enroll have
between jobs, for a short period following college (a point
retiree coverage from a former employer that limits the
at which children may become ineligible for coverage
benefits they would receive from enrolling in Part B (and
under their parents’ plan), or between retirement and age
may also exempt them from the late-enrollment penalty).
65 (the age of eligibility for Medicare).
A voluntary outpatient prescription drug benefit—
Medicare. Medicare provides coverage for about 37 mil-
known as Part D—was added to Medicare in 2006; its
lion people who are age 65 or older, and it also covers
premium subsidy and penalty for late enrollment are sim-
about 7 million nonelderly people who are disabled (and
ilar to Part B’s. About 70 percent of the people who are
generally become eligible after a two-year waiting period)
eligible to participate in Part D have chosen to do so.15
or have severe kidney disease.14 In 2008, about 80 per-
Analysis by the Centers for Medicare and Medicaid Ser-
cent of Medicare’s beneficiaries are insured through the
vices (CMS) indicates that a majority of those non-
traditional fee-for-service program, which pays providers
enrollees have drug coverage from another source that is
for services directly using prices set administratively; the
at least as comprehensive as the Medicare benefit, but
rest have chosen to receive coverage through private
about 10 percent of the Medicare population appears to
insurers that contract with Medicare to provide program
lack substantial drug coverage.
benefits in return for a fixed monthly payment per
enrollee (known as the Medicare Advantage option).
Medicaid and the State Children’s Health Insurance
About 3 percent of people under age 65 are covered by
Program. Medicaid is the main source of health insurance
Medicare (see Table 1-1 on page 4), but their average
coverage for Americans who have very low income, and
costs to the program are substantial—more than $35,000
the smaller State Children’s Health Insurance Program
per person in 2007 for those with kidney failure and
(SCHIP) provides coverage for children in families that
roughly $8,000 per person for other disabled enrollees.
have somewhat higher income. Unlike the Medicare pro-
gram, which does not take into account income or assets
when determining eligibility and is federally financed,
13. Exceptions include self-employed individuals, who may deduct
Medicaid and SCHIP are needs-based assistance pro-
the costs of their health insurance from their taxable income, and
grams that are jointly financed by the federal government
individuals who claim itemized medical deductions in excess of
7.5 percent of their adjusted gross income. See Chapter 2 for
and state governments.
additional discussion.
14. According to the most recent estimates from the Census Bureau,
15. That figure includes retirees who continue to receive drug
about 700,000 elderly people, or roughly 2 percent of individuals
coverage from a former employer if that employer receives a
age 65 or older, were uninsured in 2007.
subsidy payment from Medicare on their behalf.
CBO
10
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
CBO estimates that at any given point in 2009, roughly
SCHIP was established in 1997 to provide coverage to
64 million nonelderly individuals will be eligible for
children whose family income is above the eligibility lev-
Medicaid or SCHIP coverage and that about 43 million
els for Medicaid. States generally cover children in fami-
will be enrolled.16 Eligibility for Medicaid was originally
lies that have income up to 200 percent of the federal
limited to very low income families with dependent chil-
poverty level (or about $44,000 for a family of four in
dren and to poor elderly or disabled individuals. Over the
2009), but some states have higher income limits and
past two decades, coverage has been extended to children
some cover parents as well as their children. Like Medic-
in families with somewhat higher income and to preg-
aid, SCHIP is jointly funded by the federal government
nant women. Nonelderly, nondisabled adults who have
and the states, but the federal share of costs is higher for
no children are generally ineligible for the program. Able-
SCHIP—covering 70 percent of health care claims, on
bodied parents and children represent about three-
average. States have a fair amount of discretion in design-
fourths of all Medicaid enrollees, but about 70 percent of
ing and implementing their programs: They may expand
the program’s spending is for the remaining enrollees who
Medicaid, create a new state system specifically for
are either elderly or disabled and have low income and
SCHIP, or use some combination of the two
few assets.
approaches.18
Subject to broad federal requirements governing eligibil-
ity and benefits, the Medicaid program is largely adminis-
SCHIP is currently authorized in law through March
tered by the states, and thus its specific features may vary
2009. Consistent with statutory guidelines, CBO
considerably from state to state. On average, the federal
assumes in its baseline spending projections that federal
government covers about 57 percent of the costs of the
funding for the program in later years will continue at
health care services received by enrollees (the share varies
$5.0 billion, the base amount provided for the first half of
among states and is higher for states with relatively low
fiscal year 2009. In fiscal year 2008, the program’s budget
per capita income). State Medicaid programs cover a
authority was $6 billion and its outlays were about
comprehensive set of services, including hospital care
$7 billion. Because average costs per enrollee are expected
(both inpatient and outpatient), physicians’ services,
to rise, CBO projects that average enrollment would
nursing home care, home health care, and certain addi-
decline from a peak of about 5.3 million in 2008 to about
tional services for children. States have the authority to
2 million in 2018 under that assumption about future
cover other services and populations and have used that
funding. (References to Medicaid in the remainder of this
authority extensively.17 They may also apply to the
chapter also include SCHIP.)
federal government for waivers from various federal
Medicaid rules.
Other Sources of Coverage. A significant number of peo-
ple obtain insurance coverage from various other sources
16. That figure represents average enrollment and excludes nonelderly
including the military, universities (for students), and
individuals living in institutions (such as nursing homes) and
other organizations. CBO estimates that roughly 12 mil-
people living in U.S. territories. CBO has also projected that the
lion people will be covered under such arrangements in
total number of individuals enrolled in Medicaid at any point
2009. Although military coverage could be considered
during 2009 (including elderly and institutionalized enrollees and
residents of territories) will be 65 million, of which about
a form of employment-based insurance, it is typically
59 million will be nonelderly. Many of those individuals will be
counted separately. The Department of Veterans Affairs
enrolled in the program for only part of the year.
provides some health care to military veterans, but its
17. According to one estimate, total spending on optional populations
programs are not considered a comprehensive health
and benefits accounted for about 60 percent of the program’s
insurance plan; similarly, the Indian Health Service pro-
expenditures in 2001. Of that total, 30 percent was spent to pro-
vide optional benefits to mandatory groups; 50 percent, to
vides some care to Native Americans and Alaska natives
provide mandatory benefits to optional groups; and 20 percent,
but is not counted as a source of health insurance (such
to provide optional benefits to optional groups. See Kaiser
programs are discussed more extensively in Chapter 6).
Commission on Medicaid and the Uninsured, Medicaid Enroll-
ment and Spending by “Mandatory” and “Optional” Eligibility and
Benefit Categories (Washington, D.C.: Henry J. Kaiser Family
18. For additional information, see Congressional Budget Office, The
Foundation, June 2005), p. 11.
State Children’s Health Insurance Program (May 2007).
CBO
CHAPTER ONE
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
11
Figure 1-1.
ance lasting four months or less and about one in six had
Patterns of Health Insurance
spells lasting two years or more.
Coverage for Nonelderly People, by
According to CBO’s projections, the average number of
Family Income Relative to the Federal
people who are uninsured at any one time will rise to
Poverty Level, 2009
about 54 million, or about 19 percent of the nonelderly
population, by 2019. The number of uninsured individu-
(Percent)
als is expected to increase because health insurance premi-
100
ums are likely to rise considerably faster than income,
90
Public Coverage
which will make insurance more difficult to afford.
80
Uninsured
Characteristics of the Uninsured. The purchase of health
70
Private Coverage
insurance in the United States is voluntary, so the main
60
reason that people are uninsured is that they are unwill-
50
ing or unable to purchase coverage. Several characteristics
40
are associated with insurance status—including income,
age, being offered insurance at work, or being eligible for
30
public coverage—but whether they are a causal factor or
20
are merely correlated with coverage rates is not always
10
clear.
0
Below 100
100–200
200–400
Above 400
Because the costs of health insurance can represent a sub-
Income as a Percentage of Poverty Level
stantial share of income for lower-income individuals and
families who are not eligible for subsidized public cover-
Source:
Congressional Budget Office’s health insurance simulation
age, it is not surprising that coverage patterns are strongly
model.
correlated with income. In particular, as income rises, the
share of nonelderly people who are uninsured or have
The Uninsured Population
public coverage declines and the share with private cover-
About 45 million people, or about 15 percent of the total
age rises (see Figure 1-1). In 2009, the highest rates of
U.S. population, will be uninsured at any given point
uninsurance—about 30 percent—will be found among
in 2009, by CBO’s most recent estimates. Because the
people whose family income is below 200 percent of the
elderly have near-universal coverage from Medicare,
federal poverty level. For people in that group that have
many analyses of the uninsured focus on the nonelderly
insurance, those with family income below the poverty
population, about 17 percent of which is expected to lack
line will be much more likely to have public coverage,
coverage in 2009. Those estimates for 2009 do not reflect
whereas those with income above the poverty line will be
the recent deterioration in economic conditions, which
more likely to have private insurance. Only about 12 per-
could result in a larger uninsured population.
cent of people below the poverty line will have private
In many cases, people’s insurance status varies over the
coverage; that rate rises to 40 percent for those between
course of a year. For example, CBO’s analysis of survey
100 percent and 200 percent of the poverty level. For
data showed that between 57 million and 59 million
people whose income is between 200 percent and
people—or roughly one-fourth of the nonelderly n popu-
400 percent of the poverty level, by contrast, 74 percent
lation—were uninsured at some point during 1998. The
have private coverage and 16 percent are uninsured. For
average number of people who were uninsured at a give
people with income above 400 percent of the poverty
point in 1998 was smaller—between 39 million and
level, 90 percent have private coverage and 4 percent are
44 million, of which 21 million to 31 million were
uninsured.
uninsured for all of that year.19 CBO also found that for
those who became uninsured at some point between July
19. Congressional Budget Office, How Many People Lack Health
1996 and June 1997, nearly half had spells of uninsur-
Insurance and For How Long? (May 2003).
CBO
12
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
Figure 1-2.
within a given income range, workers in relatively small
Uninsurance Rates of Full-Time
firms (which are less likely to offer coverage) are much
Workers, by Size of Firm and
more likely to be uninsured than workers in larger firms
(see Figure 1-2). For example, among full-time workers
Family Income Relative to the
whose income is between 100 percent and 200 percent
Poverty Level, 2009
of the federal poverty level, CBO projects that 56 percent
of those employed by very small firms (fewer than
(Percent)
25 employees) will be uninsured in 2009, compared
60
with 30 percent for those employed by larger firms (those
Employees in Firm
with 100 or more workers). Determining cause and effect
50
Fewer Than 25
is difficult, however, because workers with less of a desire
25–99
for insurance or who consider coverage unaffordable
40
would be more likely to join firms that do not offer
100–999
coverage and pay those workers higher wages instead.
30
1,000 or More
Looking at income levels and insurance options simul-
20
taneously may provide additional insights about the
uninsured population. For example, CBO projects that
among the uninsured in 2009, 17 percent will have fam-
10
ily income above 300 percent of the poverty level (about
$65,000 for a family of four); 18 percent will be eligible
0
for but not enrolled in Medicaid; and 30 percent will be
100–200
200–400
Above 400
offered, but will decline, coverage from an employer (see
Income as a Percentage of Poverty Level
Figure 1-3). Some people will be in more than one of
Source:
Congressional Budget Office’s health insurance simulation
those categories at the same time—so overall, about half
model.
of the uninsured will meet at least one of those three cri-
Another characteristic that is associated with the lack of
teria. Conversely, the rest of the uninsured are projected
health insurance, at least among adults, is age. Younger
to have relatively low income and to lack both an offer of
adults are particularly likely to be uninsured—about
employment-based coverage and eligibility for public
27 percent of those ages 18 to 34 lacked coverage, com-
coverage.
pared with about 14 percent of those ages 45 to 64 in
2007—possibly reflecting a lower perceived need for
The reasons people remain uninsured even though they
using health care services (younger people are generally
are offered employment-based coverage or are eligible for
healthier) as well as lower average income and assets.20
Medicaid are not always clear. In the case of employment-
Those younger adults make up about one-fourth of the
based coverage, the share of the premium that the
nonelderly population but represent about 40 percent of
employee must pay may be relatively high, or the
the uninsured. Children under the age of 18 account for
employee may simply place a low value on having insur-
about the same share of that population but are much less
ance. As for Medicaid, studies indicate a mixture of rea-
likely to be uninsured.
sons for failing to enroll. Some people may not be aware
that they are eligible; others may be deterred by the
Not surprisingly, rates of coverage are also associated with
application process or see some stigma associated with a
whether an individual (or a close family member) is
program for low-income families. An additional factor is
offered insurance at work. In part that correlation proba-
that people who are eligible for Medicaid may be enrolled
bly reflects differences in income—firms with more low-
when they are hospitalized and then may gain retroactive
wage workers are less likely to offer coverage—but even
coverage for recent medical expenses; thus, eligibility—
even without enrollment—gives them some degree of
20. U.S. Census Bureau, Income, Poverty, and Health Insurance
protection against high medical costs and may reduce the
Coverage in the United States: 2007, P60-235 (August 2008).
incentive to enroll sooner.
CBO
CHAPTER ONE
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
13
Figure 1-3.
Medicare and offer emergency services are required by
Projected Distribution of the
law to stabilize any patient who arrives, regardless of
Uninsured Nonelderly Population, by
whether he or she has insurance or is able to pay for that
care. In addition, most hospitals are nonprofit organiza-
Selected Characteristics, 2009
tions and thus have some obligation to provide care for
free or for a minimal charge to members of their commu-
Family Income Only
nity who could not afford it otherwise. For-profit hospi-
(9%)
Family Income and
tals also provide such charity or reduced-price care.21
Offer of EBI
(8%)
Estimates of how much uncompensated care the unin-
sured receive vary depending on the data sources and
methods used and the categories of spending that are
included in the analysis. Some measures of uncompen-
None of the
sated care compare the amount that providers are actually
Offer of EBI Only
Three Criteria
(16%)
paid for their services with their list prices or posted
(49%)
charges for those services. A more useful comparison,
however, is with the total payments that providers would
receive for the same service when treating a privately
Eligibility for
insured patient, because that amount (which is generally
Medicaid and
much lower than the list price) more closely resembles
Offer of EBI
their costs.
(6%)
Eligibility for
Medicaid Only
A recent study by Hadley and others, which used that
(12%)
analytic approach, examined a sample of medical claims
Source:
Congressional Budget Office.
for uninsured individuals and projected that they would
Note: This analysis categorizes uninsured nonelderly people
receive about $28 billion in uncompensated care in
according to whether they will meet any of the following cri-
2008.22 That study also examined reports by doctors and
teria in 2009: Their family income will be above 300 percent
hospitals and derived a higher estimate: Their gross costs
of the federal poverty level; they will have an offer of
of providing uncompensated care would be about
employment-based insurance (EBI); or they will be eligible
for Medicaid or the State Children’s Health Insurance Pro-
$43 billion in 2008, of which $8 billion would come
gram (SCHIP). The Congressional Budget Office estimates
from doctors and $35 billion would come from hospitals.
that a very small number of people will have family income
But as the study noted, at least a portion of those costs
above 300 percent of the federal poverty level and will be eli-
could be offset by added payments under Medicare and
gible for Medicaid or SCHIP.
Medicaid to hospitals that treat a disproportionate share
Use of Health Care by the Uninsured.
of low-income patients (and by similar dedicated pay-
How the uninsured
ments made under other federal and state programs).
obtain health care affects both their incentives to seek
Another recent study found that, as a group, office-based
insurance coverage and the impact that policies designed
to reduce the number of uninsured have on spending and
21. For a discussion, see Congressional Budget Office, Nonprofit
health. Many of the uninsured receive care from free
Hospitals and the Provision of Community Benefits (December
clinics and other community health centers, which are
2006).
funded by a combination of federal and state sources and
22. Jack Hadley and others, “Covering the Uninsured in 2008: Cur-
private donations. Others may use traditional health care
rent Costs, Sources of Payment, and Incremental Costs,” Health
providers—hospitals as well as physicians in private prac-
Affairs, Web Exclusive (August 25, 2008), pp. W399–W415.
tice—and pay all charges for the services they receive.
That study also reported that uncompensated care would total
about $56 billion in 2008 if all costs not paid out of pocket by the
In many cases, however, people who are uninsured receive
uninsured were included in the tally. But that amount would seem
to be an overestimate because the study found that, even though
treatments from traditional providers for which they
no payments were made by insurers, about half of those costs were
either do not pay or pay very little, which is known as
directly compensated by various third parties (such as workers’
“uncompensated care.” Hospitals that participate in
compensation programs).
CBO
14
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
Table 1-3.
Health Care Expenditures in 2008, by Insurance Status
Third-Party
Out-of-Pocket
Payments
Uncompensated
Insurance Status
Spending
Insurance
Othera
Care
Total
Dollars of Spending
Uninsured for Full Year
583
0
567
536
1,686
Insured for Part of the Year
550
2,030
260
145
2,983
Privately Insured for Full Year
681
3,018
215
0
3,915
Insured for Full Year
654
3,563
246
0
4,463
Shares of Spending (Percent)
Uninsured for Full Year
35
0
34
32
100
Insured for Part of the Year
18
68
9
5
100
Privately Insured for Full Year
17
77
5
0
100
Insured for Full Year
15
80
6
0
100
Source:
Congressional Budget Office based on data from Jack Hadley and others, “Covering the Uninsured in 2008: Current Costs, Sources
of Payment, and Incremental Costs,” Health Affairs, Web Exclusive (August 25, 2008), pp. W399–W415. The authors used data
from the Medical Expenditure Panel Survey, 2002–2004, and adjusted the data to 2008.
a.
Includes workers’ compensation, veterans’ benefits, and other payments not counted as health insurance.
physicians roughly “broke even” when treating uninsured
for those who are insured for part of the year (about
patients because some of those patients paid more than
$3,000) falls between those two points. According to
the doctors would have received for treating a privately
those estimates, average out-of-pocket payments are simi-
insured patient.23 (The issue of whether and to what
lar for each group, although those payments cover a
extent the net costs of providing uncompensated care are
higher share of total spending for the uninsured.
shifted to other payers in the health sector is discussed in
Chapter 5.)
Reflecting a range of other findings on that topic, CBO
estimates a somewhat smaller disparity in the use of
The uninsured generally use fewer health care services
health care services than the study by Hadley and others
than people who have insurance, although estimates
would indicate.24 According to several other studies and
regarding the magnitude of the difference also vary. The
CBO’s own analysis of data for the nonelderly popula-
study by Hadley and others estimated that an individual
tion, the uninsured do use fewer health care services than
who is uninsured for all of 2008 will use about $1,700
the insured, but the difference is generally in the range of
worth of care—including about $540 in uncompensated
30 percent to 50 percent. (See Chapter 3 for a more
care—or less than half as much as someone who is
extensive discussion of those estimates.) Studies compar-
privately insured all year would use (see Table 1-3). The
ing the insured and uninsured populations usually
disparity in the amount spent for care is even larger; sub-
account for any differences that are observed in the
tracting uncompensated care yields an estimate that
demographic characteristics and health status of those
spending incurred by and on behalf of people who are
populations, which would affect their use of health care.
uninsured for the entire year (about $1,160) is about
30 percent of the amount spent for people who are pri-
24. If the study by Hadley and others underestimated the number of
vately insured all year (about $3,900). Spending by and
services used by uninsured individuals, its estimate of uncompen-
sated care could also be correspondingly low. (That factor could
account for the higher estimate of uncompensated care that study
23. Jonathan Gruber and David Rodriguez, How Much Uncompen-
derived using reports by doctors and hospitals.) If, instead, the
sated Care Do Doctors Provide? Working Paper No. 13585
study overestimated the number of services used by insured indi-
(Cambridge, Mass.: National Bureau of Economic Research,
viduals, that would not necessarily affect the estimate of uncom-
November 2007).
pensated care.
CBO
CHAPTER ONE
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
15
Thus, CBO would expect an uninsured person to use
mortality is a relatively crude measure of the benefits con-
30 percent to 50 percent fewer health care services, on
ferred by medical care, but the ability to analyze other
average, than a person who is similar in other respects but
outcomes, such as quality of life, is constrained because
has typical private insurance coverage. Among people
those effects are more difficult to measure.
who have similar demographic characteristics and health
status, there are two possible reasons why those who are
Nature and Extent of Coverage
uninsured would use fewer services than those who are
In addition to differences in the sources of and financing
insured: First, some of the uninsured may simply be less
for health insurance and health care, coverage varies by
inclined to seek health care, resulting in less use of ser-
the type of health plan providing it, the scope of services
vices; and second, the prospect of having to pay the full
that are covered, and the cost-sharing requirements and
cost of the services they receive gives them an incentive to
limits that apply. That variation largely reflects different
use less medical care or less expensive services.
approaches to controlling costs for insured individuals
and can have substantial effects on the premiums charged
A related consideration is whether the lack of insurance
for an insurance policy (as discussed in Chapter 3).
has adverse effects on health. Some studies examining the
treatment of serious health conditions have found rela-
Types of Plans. Through the 1980s, private health insur-
tively clear links between insurance coverage and health
ance coverage in the United States typically took the form
outcomes.25 For example, uninsured individuals who
of an “indemnity” policy, which reimbursed enrollees for
develop cancer generally have poorer outcomes and die
their incurred costs, left it to them and their doctors to
more quickly than cancer patients who have private
determine what care to provide, and largely allowed
health insurance. That difference is attributed partly to
doctors and hospitals to set the prices for those services.
later diagnosis for the uninsured; broader analyses of the
As health care costs grew rapidly in the 1980s, however,
private insurance coverage began to shift from indemnity
uninsured population have found that they are less likely
policies toward other types of health plans, involving var-
to receive screening tests, such as mammograms. Simi-
ious degrees of managed care (as described below) and
larly, uninsured individuals who have heart disease are
negotiated pricing.
less likely to receive expensive treatments for it and also
have higher rates of mortality than those who have heart
One form of managed care plan that emerged was a pre-
disease but are privately insured.
ferred provider organization (PPO). PPOs establish lists
or networks of preferred doctors and hospitals and—to
For more routine care, however, disentangling the effects
give enrollees an incentive to use those providers—charge
on health of being uninsured from the impact of other
factors that are associated with lack of insurance is more
difficult. One recent and comprehensive review of the lit-
26. Helen Levy and David Meltzer, “The Impact of Health Insurance
on Health,” Annual Review of Public Health, vol. 29 (April 2008),
erature noted that most studies of such effects on health
pp. 399–409. One study that sheds some light on the impact of
simply compare insured and uninsured individuals and
health insurance on health is the RAND Health Insurance
thus do not account for underlying differences between
Experiment, which randomly assigned large groups of nonelderly
those populations.26 Some studies with a better design
individuals to different health insurance plans and tracked their
have examined the effects of expanding eligibility for
experience over several years. In general, the study found that par-
ticipants who faced cost sharing did not have worse health than
public insurance programs and have found specific health
those who got all of their care for free; one exception was lower-
benefits for the targeted populations, but broad health
income participants with prior health problems, who did not
improvements stemming from insurance coverage have
control their blood pressure as effectively when they faced cost
been difficult to identify. For example, one recent study
sharing. An important limitation of the study, however, is that no
participants lacked insurance. For additional discussion of those
found that the creation of Medicare had no discernible
findings, see Congressional Budget Office, Consumer-Directed
effect on the mortality rates of the elderly during the first
Health Plans: Potential Effects on Health Care Spending and
10 years of the program’s operation.27 Of course, reduced
Outcomes (December 2006), pp. 54–55.
27. Amy Finkelstein and Robin McKnight, “What Did Medicare Do?
25. For a summary of those studies, see Institute of Medicine, Care
The Initial Impact of Medicare on Mortality and Out of Pocket
Without Coverage: Too Little, Too Late (Washington, D.C.:
Medical Spending,” Journal of Public Economics, vol. 92, no. 7
National Academy Press, 2002), www.iom.edu.
(July 2008), pp. 1644–1668.
CBO
16
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
more for care received outside the plan’s network. The
about 8 percent of enrollment in employment-based cov-
preferred providers thus gain a higher volume of patients
erage; one form of consumer-directed plan (known as a
and, in return, usually accept lower negotiated payment
health savings account) can also be purchased in the
rates for each service from the health plan. According to a
individual insurance market.29
major survey of employers conducted by the Kaiser
Family Foundation, PPOs are the most common type of
Scope of Covered Services. Both public and private
managed care plan, accounting for about 58 percent of
health insurance plans generally cover hospitalizations,
enrollees in employment-based plans in 2008.28 (That
visits to doctors and other outpatient care, tests and
survey is the primary source of statistics about coverage
imaging services (such as X-rays), and prescription drugs.
and benefits cited in this subsection.)
Coverage varies to a greater extent for dental care and
vision-related services, particularly when care is discre-
At the same time, more stringent forms of managed care,
tionary (for example, laser surgery to correct vision prob-
such as health maintenance organizations (HMOs), also
lems is typically not covered). According to a 2004 survey
grew in prominence. Like PPOs, those plans establish
of employers, about 20 percent offered vision benefits
networks of providers; unlike PPOs, they offer no cover-
and two-thirds offered dental benefits (although nearly
age for services received outside their networks (except for
all firms with more than 500 employees offered dental
emergencies). HMOs have also instituted various mea-
benefits and about half of those firms offered vision bene-
sures to limit the use of certain services, such as requiring
fits).30 Another source of variation is government
patients to get a referral from a primary care physician in
requirements to cover certain types of benefits (such as
order to see a specialist or to obtain prior authorization
infertility treatments) or the services of specific providers
from the plan before using some types of specialty care.
(such as chiropractors), which some states impose and
Some HMOs are fully integrated; the plan owns the
others do not. Those mandates generally affect policies
hospitals, and doctors work on salary. A more common
offered in the individual market and by small employers.
arrangement, however, is to have a network of indepen-
Cost-Sharing Requirements. A more significant way in
dent hospitals and physicians’ practices in which provid-
which health insurance plans vary, even among the broad
ers either receive a fixed payment per patient (in the case
categories of plans noted above, is their cost-sharing
of some primary care physicians) or are paid negotiated
structure. Most plans include one or more of the follow-
rates on a fee-for-service basis. As a share of enrollment
ing provisions:
in employment-based plans, HMOs peaked at roughly
30 percent in the mid-1990s and then fell, reaching
B An annual deductible (expenses that enrollees must
about 20 percent in 2008.
pay out of pocket before the insurer begins paying for
services),
Point-of-service (POS) plans have emerged as a kind of
middle ground between PPOs and HMOs. Like PPOs
B Coinsurance (a specified percentage) or copayments (a
they allow enrollees to go outside a plan’s network for care
specified amount) that enrollees pay out of pocket to
(albeit at a higher charge), but like HMOs they typically
providers after satisfying any deductible, and
require enrollees to secure referrals for specialty care from
a primary care physician within the plan’s network. More
B An out-of-pocket maximum (a cap on the total
common among small firms, they accounted for 12 per-
amount that an individual or family pays out of
cent of enrollment in employment-based plans in 2008.
pocket in a given year).
Another design option that has arisen in recent years is a
Those features not only affect the share of health care
consumer-directed health plan, which combines a high-
costs covered by the insurance policy but also influence
deductible insurance policy with an account that enroll-
total spending for health care.
ees can use to finance their out-of-pocket payments on a
tax-preferred basis. (In other respects, those plans are usu-
29. For additional discussion of those plans, see Congressional Budget
ally similar to PPOs.) As of 2008, those plans account for
Office, Consumer-Directed Health Plans.
30. Mercer Human Resource Consulting, National Survey of
28. Kaiser/HRET, Employer Health Benefits: 2008 Annual Survey.
Employer-Sponsored Health Plans 2004 (New York: Mercer, 2004).
CBO
CHAPTER ONE
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
17
Cost-sharing requirements typically differ by type of
for them, require higher rates of cost sharing, or limit the
plan. According to the 2008 Kaiser/HRET survey of
maximum annual benefits that enrollees can receive.
employment-based health insurance plans, almost 20 per-
cent of HMO enrollees face a deductible in 2008, com-
Cost-sharing requirements tend to be higher in the indi-
pared with about 68 percent of PPO enrollees. Among
vidual insurance market, reflecting not only insurers’
PPO enrollees, deductibles for care received within the
efforts to control the health care spending of their enroll-
plan’s provider network average about $560 for single
ees but also enrollees’ desire for lower premiums (because
coverage and about $1,300 for family coverage in 2008.
those policies are generally not subsidized through the tax
For hospital care, some enrollees face separate deduct-
code). One survey of policies purchased in the individual
ibles, and most (about 69 percent) are subject to coinsur-
market in late 2006 and early 2007 found that about
ance or copayments.
70 percent of single policies had deductibles of more than
$1,000 and about two-thirds of family policies had
Most HMO and PPO plans that have a deductible
deductibles of more than $2,000.31 Largely because they
exempt visits to a physician’s office for care received
cover a smaller share of enrollees’ health care costs, the
within the network. Enrollees typically have a fixed
premiums for those policies are generally lower than the
copayment of around $20 for seeing a primary care phy-
average premiums observed for employment-based insur-
sician and around $25 for seeing a specialist physician
ance (even though the premiums for individually pur-
within their network. For visits outside the network, PPO
chased policies include higher administrative costs per
enrollees who have met the deductible typically pay
policy).
coinsurance in the range of 30 percent to 35 percent
(thus encouraging enrollees to use network providers and
Cost-sharing requirements in the Medicaid program tend
also limiting the plan’s liability for those costs). Most peo-
to be much lower than those in employment-based or
ple who have employment-based insurance must also pay
individually purchased plans—typically $1 to $3 for a
a portion of the costs for advanced diagnostic tests and
doctor’s visit or $2 to $3 for a brand-name drug prescrip-
outpatient surgery (coinsurance is more common) and
tion—reflecting the limited income of Medicaid recipi-
for emergency room and urgent care visits (copayments
ents. Cost-sharing requirements may be more substantial
are more common).
under SCHIP but are generally limited to about 5 per-
cent of enrollees’ family income.
Most plans also limit total out-of-pocket spending that
enrollees might incur in a given year. For PPO plans,
Cost sharing under the Medicare program varies widely
median levels of the out-of-pocket maximum are roughly
by service. In 2009, enrollees will face a deductible of
$2,000 for single coverage and $4,000 for family cover-
about $135 for physicians’ services and will be charged
age in 2008, although those limits vary considerably
20 percent coinsurance beyond that point. Some services,
across plans. Nearly half of HMOs do not have an out-
such as lab tests and home health care, are free to the
of-pocket limit, but those plans typically have no deduct-
enrollee. Most hospital admissions require a deductible of
ible and relatively low cost sharing for individual services,
about $1,070, however, and the effective coinsurance
so enrollees would be unlikely to incur very high out-of-
rates for some skilled nursing care and outpatient hospital
pocket costs in the aggregate.
services may exceed 30 percent. In addition, the program
does not cap annual out-of-pocket costs. To limit their
Many plans vary the amount of coinsurance by the type
financial exposure, most Medicare enrollees have some
of service or exempt some services from the general
form of supplemental insurance that covers most or all of
deductible in an attempt to create differing incentives for
their cost-sharing obligations. That supplemental
enrollees to use certain types of care. For example, pre-
coverage typically comes from a former employer, the
ventive services may have little or no cost sharing, either
Medicaid program, a Medicare Advantage plan, or an
because insurers want to encourage their use or because
individually purchased medigap policy.
those benefits are attractive to enrollees. Similarly, plans
typically exempt prescription drugs from their general
31. AHIP Center for Policy Research, Individual Health Insurance
deductible and require relatively low copayments for less
2006–2007: A Comprehensive Survey of Premiums, Availability, and
expensive generic drugs. Conversely, plans that cover den-
Benefits (Washington, D.C.: America’s Health Insurance Plans,
tal and vision services may charge a separate deductible
December 2007).
CBO
18
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
Health Care Spending
drugs for 10 percent, and the administrative costs of pub-
Both the amount and rate of growth of spending for
lic and private insurers for about 7 percent. (Administra-
health care have important implications for proposals
tive costs borne by doctors, hospitals, and other providers
that would seek to expand insurance coverage, reduce
are financed through the payments they receive for their
that spending, or do both simultaneously. The budgetary
services.) The remainder of personal health care expendi-
impact of subsidizing insurance coverage depends in part
tures is primarily for dental and other professional care,
on the health care costs that would be covered, and the
home health and nursing home care, and medical
effects of efforts to control costs depend on how those
equipment.
efforts influence the factors that cause cost growth. Other
key aspects of health care spending include its concentra-
Compared with other developed countries, the United
tion (a relatively small share of individuals account for the
States devotes a substantially larger share of its economy
bulk of expenditures in any given year), how much an
to health care and related expenditures. That share was
individual’s health care costs vary from year to year, and
about 16 percent of gross domestic product in 2006—up
the substantial variation in average spending that is
from about 8 percent in 1975. Under current law, that
observed from one region of the country to another.
share is projected to reach nearly 20 percent by 2017 (the
last year of the current CMS projections). By contrast,
Amount and Growth of Spending
spending for health care among the other countries that
For 2009, the Centers for Medicare and Medicaid Ser-
belong to the Organisation for Economic Co-operation
vices projects that national health expenditures will total
and Development (OECD) averaged about 9 percent of
$2.6 trillion. The bulk of that spending—about five out
GDP in 2006.33
of every six dollars spent in the health sector—is for
personal health expenditures. That category includes such
Comparisons of growth rates for health care spending
services and supplies as hospital care, physicians’ and
across countries can be sensitive to the time period used
clinical services, and prescription drugs, among others.
and to other factors included in the analysis (such as age,
The remaining expenditures are for broad categories of
average income, or overall rates of economic growth).
spending that support but do not provide health care,
Some comparisons indicate that real (inflation-adjusted)
including the administrative costs of private and public
growth rates have been similar across developed coun-
insurers; the outlays of public health departments and
tries, which might suggest that common forces are caus-
related activities; and investments in medical research,
ing spending to rise despite substantial differences in their
equipment, and structures.
health care systems; other studies conclude, however, that
the United States has experienced faster growth in the
Data on national health expenditures can be broken
share of GDP spent on health care than have other, com-
down in two basic ways: by the sources of payment and
parable nations.34
by the types of services provided (see Table 1-4). Private
spending will account for about 54 percent of the total in
Within the United States, growth rates in health care
2009, and public outlays—primarily for the Medicare
spending have varied over time but have generally out-
and Medicaid programs—will account for the remaining
paced those in the overall economy. An exception was the
46 percent.32 About 65 percent of private health care
period between 1993 and 2000, when the share of GDP
costs are covered by insurance, and the rest are paid out of
spent on health care held nearly constant at about 14 per-
pocket or from other sources (including philanthropy).
cent, but spending growth has accelerated since then.
As for the types of services provided, hospital care and
Over extended periods, the annual growth rate of health
physicians’ services combined will account for about half
of all health care expenditures, outpatient prescription
33. Organisation for Economic Co-operation and Development,
OECD Health Data 2008, www.oecd.org. Adjusting for differ-
ences in income would reduce that disparity somewhat because
32. If the cost of the tax expenditure from excluding premiums for
income in the United States is higher than the OECD average and
employment-based insurance (estimated to be roughly $250 bil-
higher income is correlated with higher spending on health care.
lion at the federal level in 2007) was included in public spending
rather than private spending, then the public share of spending
34. Chapin White, “Health Care Spending Growth: How Different Is
would be about 57 percent. See Chapter 2 for additional
the United States from the Rest of the OECD?” Health Affairs,
discussion.
vol. 26, no. 1 (January/February 2007), pp. 154–161.
CBO
CHAPTER ONE
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
19
Table 1-4.
National Health Expenditures, by Source of Payment and Type of Service, 2009
(Billions of dollars)
Private
Public
Total, NHE
Out-of-
Sub-
Medicaidb
Sub-
Billions of
Insured
Pocket
Othera
total
Medicareb
Federal
State
Otherc
total
Dollars
Percentage
Personal Health Care Expenditures
Hospital Care
291
27
36
355
230
79
61
75
445
800
31
Physicians' and
Clinical Services
264
55
34
353
108
22
16
34
180
533
21
Dental and Other
Professional Care
79
65
3
148
15
6
4
4
29
177
7
Prescription Drugs
113
56
0
170
52
14
10
19
95
264
10
Home Health and Nursing
Home Care
17
43
6
67
51
48
39
6
144
210
8
Medical Equipment and
Other Personal Care
3
51
7
61
10
34
27
14
85
146
6
___
___
__
____
___
___
___
___
___
____
__
Subtotal, Personal
Health Care
767
298
87
1,152
467
203
156
152
978
2,131
83
Other Expenditures
Administration and Net
Cost of Private Insurance
112
0
2
113
28 d
16 d
13 d
14
71
184
7
Public Health Activity
0
0
0
0
0
0
0
72
72
72
3
Research, Equipment, and
Structures
0
0
104
104
0
0
0
65
65
169
7
___
_
___
___
__
__
__
___
___
___
__
Subtotal, Other
112
0
105
217
28
16
13
151
207
424
17
Total, NHE
Bil ions of dollars
879
298
193
1,369
495
219
169
303
1,186
2,555
100
Percentage
34
12
8
54
19
9
7
12
46
100
Source:
Centers for Medicare and Medicaid Services.
Note: NHE = national health expenditures.
a.
Includes private philanthropy.
b.
Figures for Medicare and Medicaid differ from the Congressional Budget Office’s projections.
c.
Includes payments for workers’ compensation programs and for health care provided by the Departments of Defense and Veterans Affairs.
d.
Administrative costs for Medicare and Medicaid include costs incurred by private health plans to deliver their benefits.
CBO
20
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
care spending per capita in the United States has typically
ing the forces that are also causing private spending for
exceeded the growth rate of GDP per capita by 2 percent-
health care to rise.
age points or more, accounting for the substantially larger
share of the economy that spending for health care now
Sources of Spending Growth
represents. CBO projects that as the share of family and
The effects of proposals to reduce spending on health care
state budgets devoted to health care grows even larger,
depend in part on how they affect the factors that are
growth in health care expenditures will eventually moder-
driving the growth of that spending. The factor with the
ate even in the absence of changes in federal law. By
greatest impact on spending growth is probably the devel-
CBO’s estimates, spending per capita will nevertheless
opment and diffusion of medical technology (broadly
continue to grow more quickly than the economy as a
defined). Other influences include the aging of the popu-
whole—about 1.7 percentage points faster, on average—
lation; reductions in the share of costs paid out of pocket;
and total health care spending will reach nearly 40 per-
growth in the relative prices of health care services; and
cent of GDP by 2050.35
the growing prevalence of chronic health problems. (A
recent CBO report analyzed several of those factors and
Over the past 30 years, federal spending on Medicare and
provides additional information about the studies used to
Medicaid has nearly tripled as a share of GDP, rising from
estimate their effects.37) In addition, the manner in
about 1.5 percent in 1975 to about 4.0 percent in 2007.
which health care services are financed probably has an
According to CBO’s projections, such spending will reach
effect on the amount of spending and could also affect its
about 12 percent of GDP by 2050 under current policies,
growth rate.
but substantial uncertainty surrounds that estimate. If
Advances in Medical Technology.
spending per enrollee continued growing over the next
Many analysts attribute
four decades as quickly as it has over the past four—about
the bulk of the growth in health care spending to the
2.5 percentage points faster than per capita GDP—then
development and diffusion of new medical “technol-
federal spending on those programs would reach about
ogy”—a term that is defined broadly to include new pro-
cedures and treatments as well as new medicines and
17 percent of the economy. If, instead, spending per
devices. Some breakthrough developments permit the
enrollee grew at the same rate as GDP per capita, demo-
treatment of previously untreatable conditions; such
graphic changes alone would push those federal expendi-
innovations can confer substantial benefits, but they also
tures to about 6 percent of GDP in 2050.
add new sources of spending. Other advances may simply
As those figures suggest, the rate at which health care
improve medical outcomes (compared with those pro-
spending grows relative to the economy is the most
vided by older treatments) but at added costs. In some
important determinant of the country’s long-term fiscal
cases, however, new procedures and treatments—or
balance; it exerts a significantly larger influence on the
broader application of existing ones to new types of
budget over the long term than other commonly cited
patients—could add to spending without yielding better
factors, such as the coming retirement of the baby-boom
outcomes. Whatever the magnitude of the health benefits
may be, studies indicate that about half of the growth in
generation.36 Rising health care spending represents a
health care spending over the past several decades reflects
challenge not only for the federal government but also for
changes in medical care made possible by the develop-
private payers. Indeed, trends in both sectors reflect many
ment of new treatments and procedures.
of the same underlying forces, so controlling federal out-
lays over the long term will be difficult without address-
Improvements in medical technology do not have to
increase costs; technological innovation could reduce the
35. Congressional Budget Office, The Long-Term Outlook for Health
unit cost of treating a given health problem and could
Care Spending (November 2007). The figure for per capita cost
also reduce total spending. Under current arrangements,
growth reflects the projected rate of growth after accounting for
however, the nature of technological advances in medi-
the aging of the population, referred to as “excess” cost growth.
Those projections assume that no changes are made in federal
cine and the changes in clinical practice that have ensued
policies.
36. For a discussion, see Congressional Budget Office, The Long-Term
37. See Congressional Budget Office, Technological Change and the
Budget Outlook (December 2007).
Growth of Health Care Spending (January 2008).
CBO
CHAPTER ONE
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
21
in the United States have tended to raise total spending—
30 percent between 1965 and 2005—from 9.5 percent to
because the treatments themselves are expensive, because
12.4 percent. By itself, however, that change would cause
the number of people receiving them grows rapidly, or for
total spending on health care to rise by about 16 percent
both reasons. In the aggregate, that tendency may well
and thus accounts for only about 3 percent of the total
reflect the willingness of individuals to pay for the added
cost growth over that period. (After adjusting for general
spending through higher insurance premiums; as some
price inflation using the GDP implicit price deflator,
observers have noted, the demand for health plans offer-
per capita spending on health care grew by more than
ing “1960s medicine at 1960s prices” appears to be low.38
500 percent between 1965 and 2005.)
Decisions about whether to cover new technologies,
however, are ultimately made by public and private insur-
Aging has had a larger effect on federal spending for
ers, and the benefits and costs of those technologies may
health care, however, primarily because nearly all resi-
not be carefully evaluated in each case, in part because
dents become eligible for Medicare once they turn 65. In
the information needed to do so is lacking in many
particular, the impending eligibility of the baby-boom
situations.39
generation will have a substantial effect on the share of
GDP devoted to Medicare as a result of the increase in
In assessing the role of medical technology, analysts also
enrollment, but that effect pales in comparison with the
considered other sources of past spending growth
likely impact of continued increases in health care spend-
(including increases in income and rising administrative
ing per enrollee. According to CBO’s analysis, future
costs for insurers, as well as those listed above). Yet each
demographic changes will account for somewhere
of those factors individually has accounted for a relatively
between one-fifth and one-third of the increase in federal
small share of that growth, and collectively they can
spending on Medicare and Medicaid over the next 25 to
account for about half of total spending growth—even
75 years, and rising outlays per enrollee (over and above
using estimated effects toward the upper end of the range
demographic effects) will account for the remainder.41
for each factor. Analysts have thus attributed the large
Reductions in the Share of Costs Paid for Out of Pocket.
residual effect to technology because it is the one remain-
ing force that could be responsible for cost growth (and
Another important factor that both reflects and has con-
tributed to rising health care expenditures is the declining
because the effects of technology on spending are hard to
proportion of those costs that are paid out of pocket—
measure directly).40 Even if that conclusion is correct, it
and the corresponding increase in the share covered by
still leaves open the question of what underlying forces
insurance. According to the estimates of national health
are causing technological changes to be adopted or why
expenditures produced by CMS, out-of-pocket payments
those changes tend to yield net increases in spending.
accounted for 33 percent of all personal health care
Aging of the Population. One noteworthy finding from
expenditures in 1975; by 2000, that share had fallen to
studies that have analyzed past spending growth is that
17 percent, and it declined to 15 percent in 2006.
the impact of aging has been relatively small. The elderly
Reducing the share of costs that patients have to pay gen-
do use more health care than the nonelderly, and the
erally increases their demand for care, and studies have
share of the population that is elderly increased by about
concluded that more extensive insurance coverage is
responsible for about 10 percent of historical spending
38. Joseph P. Newhouse, “Medical Care Costs: How Much Welfare
growth. But that estimate does not account for the effect
Loss?” Journal of Economic Perspectives, vol. 6, no. 3 (Summer
that expanded insurance coverage has on the diffusion of
1992), pp. 3–21.
medical technology. By contrast, a recent study that
39. See Alan M. Garber, “Cost-Effectiveness and Evidence Evaluation
examined the effects of Medicare’s introduction found
as Criteria for Coverage Policy,” Health Affairs, Web Exclusive
that a broad expansion of insurance coverage had much
(May 19, 2004), pp. W4-284 to W4-296.
larger effects on spending. It attributed part of the impact
40. A more precise description would label the residual as the effect of
to more rapid and widespread adoption of existing
changes in medical technology that are not attributed to other
observed forces. For example, increases in income can account for
some growth in health care spending, and the mechanism through
41. See Congressional Budget Office, Accounting for Sources of
which that growth occurs might also be the greater use of medical
Projected Growth in Federal Spending on Medicare and Medicaid,
technology.
Issue Brief (May 28, 2008).
CBO
22
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
treatment methods (such as those provided by cardiac
additional services or more expensive bundles if the pay-
intensive care units), although some questions remain
ments exceed the costs of providing care.
about the precise magnitude of those effects.42
Fee-for-service payments may yield a higher quantity or
Out-of-pocket payments have continued to decline
a greater intensity of services at any given time, but
slightly as a share of health care spending in recent years,
whether that type of payment contributes to the rate
despite recent increases in cost-sharing levels. For exam-
of spending growth is less clear. Because that method of
ple, the average deductible for single coverage in an
financing has been in place for many years, it could have
employment-based PPO plan tripled between 2000 and
affected the amount of spending in a constant way
2008 (rising from $187 to $560).43 However, total out-
without changing the growth rate of spending. Consis-
of-pocket payments have not increased as quickly, and
tent with that view, an older study found that growth
spending covered by insurance has also risen substantially,
rates of spending in HMOs and fee-for-service plans did
roughly keeping pace with the increases in out-of-pocket
not differ substantially.45 Compared with other payment
costs. Indeed, the overall rise in health care spending
systems, fee-for-service payment could encourage or at
and the decline in the share paid out of pocket have had
least facilitate the adoption of newer, more costly services,
roughly offsetting effects on the share of GDP accounted
but whether that happens depends on how quickly fees
for by out-of-pocket costs, which has held steady over
are established for new treatments and on the level at
the past three decades at about 2 percent. Even so, such
which those fees are set. (See Chapter 5 for additional
increases in cost-sharing requirements have raised con-
discussion of fee-setting mechanisms.)
cerns that some people who have insurance coverage may
be underinsured. For example, a recent study estimated
Growth in the Relative Prices of Health Care Services.
that about 25 million insured adults faced relatively high
Growth in payment rates that exceeds general price infla-
out-of-pocket costs (as a share of their income) in 2007,
tion has probably contributed to the increase in the share
up from about 16 million in 2003.44
of GDP devoted to health care. Between 1975 and 2005,
the increase in the medical component of the consumer
Financing of Health Care Services. The way in which
price index was nearly twice as large as the increase in
health care services are financed also affects the amount
prices overall—which might suggest that price increases
of health care spending and could affect its growth rate as
for health care have played a large role in cost growth.
well. With the exception of some HMOs, most health
care provided by doctors in the United States is currently
Measuring price inflation in the health sector can be
paid for on a fee-for-service basis. In some cases (such as
difficult, however, both because it is hard to control for
hospital stays under Medicare), a fixed payment is made
changes over time in the quality or type of the products
for a bundle of related services. Such payments encourage
being compared (which can make historical price com-
doctors and hospitals to deliver a given service or bundle
parisons misleading) and because discounts negotiated by
efficiently, but they can also create an incentive to provide
private insurers are typically confidential. Such problems
can arise with any price index but may be particularly
42. Amy Finkelstein, “The Aggregate Effects of Health Insurance:
acute for health care because of the relatively large role
Evidence from the Introduction of Medicare,” Quarterly Journal of
played by technological advances and because the preva-
Economics, vol. 122, no. 1 (February 2007), pp. 1–37.
lence of insurance obscures the price of many trans-
43. Kaiser/HRET, Employer Health Benefits: 2008 Annual Survey; and
actions. Despite those challenges, some observers have
Employer Health Benefits: 2000 Annual Survey (September 2000).
suggested that prices for health care, when properly mea-
sured, have actually grown at rates comparable with or
44. See Cathy Schoen and others, “How Many Are Underinsured?
Trends Among U.S. Adults, 2003 and 2007,” Health Affairs, Web
lower than general inflation and that prices have not
Exclusive (June 10, 2008), pp. W298–W309. The study relied on
played a substantial role in the growth of U.S. health care
self-reported income and out-of-pocket health care costs of survey
spending over time. But other analyses (which are also
respondents; it defined individuals as underinsured if their health
cited in CBO’s January 2008 paper on the growth of
plan’s deductible exceeded 5 percent of their income or if their
out-of-pocket costs exceeded either 10 percent of their income
(for those with family income above 200 percent of the poverty
45. Joseph P. Newhouse and others, “Are Fee-for-Service Costs
level) or 5 percent of their income (for those with family income
Increasing Faster Than HMO Costs?” Medical Care, vol. 23, no. 8
below 200 percent of the poverty level).
(1985), pp. 960–966.
CBO
CHAPTER ONE
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
23
health care spending) suggest that rising relative prices for
cult. Trends in the incidence and prevalence of chronic
medical care may have accounted for as much as 10 per-
and acute health problems have varied—some rates have
cent to 20 percent of past spending growth.
increased, some have decreased, and some have held
steady. For example, cancer is a leading cause of death
Whether or not they have contributed to the growth in
and a major source of health care spending, but the inci-
spending, price levels affect total spending, so the meth-
dence of cancers has declined slightly since 1990. In other
ods used to set those levels can also play an important
cases, the analysis is complicated by the fact that reported
role. In some cases, private insurers may have difficulty
rates of disease prevalence may rise when new treatments
negotiating low prices for health care items and services,
for the disease become available. Moreover, increases in
whereas public purchasers have sometimes intervened to
the intensity of treatment may also increase the likelihood
obtain relatively low prices. Limited competition among
of diagnosing a disease (even if the true prevalence of the
doctors and hospitals in some parts of the country
disease has not changed).
hampers the ability of private insurers to negotiate lower
payment rates for their services. In the case of prescrip-
For example, obesity is associated with many serious
tion drugs, public policy (through patents) gives manu-
medical conditions, including diabetes, heart disease, and
facturers monopoly power, which leads to higher prices
high blood pressure. According to one government sur-
when drugs are introduced but also encourages those
vey, the share of adults with diabetes grew by about 2 per-
drugs to be developed in the first place. Federal and state
centage points between 1988 and 2004, from about
purchasers have established payment systems that yield
8 percent to about 10 percent.48 The share of the popula-
lower prices for drugs (under Medicaid and the health
tion being treated for diabetes grew even faster—by more
program for military veterans) and for doctors and hospi-
than 50 percent among those with private insurance,
tals (under Medicare and Medicaid), although many doc-
according to one study—partly because the probability
tors are unwilling to accept Medicaid’s payment rates.
that someone with diabetes would be diagnosed also
(See Chapter 5 for additional discussion.)
increased by about 10 percent.49 Over that same period,
however, the fraction of adults who have high blood pres-
Rising Prevalence of Health Problems. Spending on
sure held constant at about 18 percent (in part because an
health care would also be expected to grow if Americans
increasing share of patients were taking medications to
were developing more health problems or were becoming
lower their blood pressure). Meanwhile, the percentage of
more likely to contract diseases, but the evidence on those
adults with high cholesterol has fallen steadily and is now
points is mixed. Perhaps the most alarming trend has
about half what it was in the early 1960s, partly because
been the growth in obesity over the past several decades.
of the development and use of cholesterol-lowering
According to one set of surveys, the share of the adult
drugs. Overall, it is not clear what role changes in the
population that is obese grew from about 23 percent in
prevalence of disease—as opposed to increases in the rate
1988 to about 34 percent in 2004, and the share that is
at which existing diseases are diagnosed and in the inten-
either obese or overweight increased from 56 percent to
sity of their treatment—are playing in the growth of
67 percent over that period.46 CBO’s analysis indicates
health care spending.
that the share of spending growth attributable to rising
weight over a similar period is between 4 percent and
Individual and Regional Variation in
12 percent, depending on the methodology used.47
Health Care Spending
In addition to the overall level and growth of health care
More generally, determining whether spending on health
costs, three other significant aspects of spending for
care is rising because Americans are getting sicker is diffi-
health care are the concentration and the persistence of
46. National Center for Health Statistics, Health, United States, 2007,
48. National Center for Health Statistics, Health, United States, 2007.
DHHS Publication No. 2007-1232 (November 2007). Recent
About 3 percent of the population was estimated to have undiag-
data on obesity rates suggest that those rates may have leveled off,
nosed diabetes in both years (which was determined by conduct-
but it is probably too early to tell whether that development is
ing medical tests on survey participants).
temporary or is likely to endure.
49. Kenneth E. Thorpe and others, “The Rising Prevalence of Treated
47. See Congressional Budget Office, Technological Change and the
Disease: Effects on Private Health Insurance Spending,” Health
Growth of Health Care Spending, Box 1, p. 10.
Affairs, Web Exclusive (June 27, 2005), pp. W5-317 to W5-325.
CBO
24
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
Table 1-5.
Persistence of Health Care Spending
Percentage of 2003 Subgroup with
Percentage of
Total 2004 Spending in Range
Spending in 2004
Total 2003
Population in
$1,000 to
$5,000 or
All
(Dollars)
Spending Range (Dollars)
Range
0 to $1,000
$5,000
More
Ranges
Mean
Median
Zero to 1,000
50
78
17
5
100
1,214
279
1,000 to 5,000
35
41
47
13
100
2,597
1,313
5,000 or More
15
24
39
37
100
7,765
3,316
Source:
Congressional Budget Office based on merged data for 2003 and 2004 from the Medical Expenditure Panel Survey conducted by the
Department of Health and Human Services, Agency for Healthcare Research and Quality.
Notes: The figures include only individuals who were under the age of 65 in 2003 and were privately insured for all of that year.
CBO increased total health care spending for 2003 to 2004 dollars by using the growth in health care spending per capita as estimated
from the national health expenditures compiled by the Department of Health and Human Services.
individuals’ spending and the substantial geographic vari-
higher spending increases to 30 percent, and the share
ation in average spending levels across the United States.
of health care spending attributable to those individuals
In any given year, the vast majority of spending on health
rises to 86 percent. By contrast, about 55 percent of
care is generated by the relatively small share of individu-
that population used less than $1,000 worth of care in
als who use extensive or expensive services. Furthermore,
2004, and their collective spending amounted to only
people with high health care costs in one year tend to
6 percent of the total (with average spending of about
have above-average costs the next year; below-average
$300). Among the Medicare population, similar degrees
costs for health care are also likely to persist.
of concentration are observed. In 2001, the most expen-
sive 5 percent of enrollees accounted for about 43 percent
In addition to individual variation, average health care
of program spending in one year, and the top 25 percent
spending varies sharply from one region of the country to
accounted for 85 percent of spending.50
another in ways that are not explained by regional differ-
ences in age or measures of sickness and that do not
Protecting themselves against the relatively low risk of
appear to yield better health overall in the high-spending
incurring substantial costs is the main reason most people
regions. Reducing those differences in spending without
seek health insurance, and the uncertainty about those
harming health would improve the efficiency of the
costs is large enough that most people who can afford to
health sector, but steps to achieve that goal would
purchase insurance do so. Even so, a predictable element
undoubtedly prove quite challenging and complex to
of health care spending also affects the type and extent
implement.
of insurance coverage that people seek. CBO’s analysis
found that nonelderly insured individuals who used
Concentration and Persistence of Individuals’ Health
less than $1,000 worth of care in 2003 had a 78 percent
Care Spending. The concentration of annual health care
chance of using less than $1,000 worth of care in 2004
spending among a relatively small share of the population
but only a 5 percent chance of using more than $5,000;
has been well documented, both among the nonelderly
their average costs in 2004 were about $1,200 (see
and in the Medicare program. For example, CBO ana-
Table 1-5). By contrast, individuals who used more than
lyzed spending by nonelderly individuals who had health
$5,000 worth of care in 2003 had a 37 percent chance of
insurance and found that 13 percent of them used
using more than $5,000 worth of care in 2004 and only a
more than $5,000 worth of care in 2004. That high-
24 percent chance of using less than $1,000 worth; their
spending subgroup (with average costs of about $15,000)
average costs in 2004 were about $7,800. In some cases,
accounted for about 68 percent of the health care costs
for that population. If the threshold is lowered to $2,000
50. Congressional Budget Office, High-Cost Medicare Beneficiaries
worth of care, the share of nonelderly insured people with
(May 2005).
CBO
CHAPTER ONE
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
25
Figure 1-4.
decline further, reflecting a common statistical phenome-
The Relationship Between Quality of
non known as “regression to the mean.” People with very
Care and Medicare Spending, by State,
high initial spending were probably hospitalized, which is
unlikely to happen year after year (even though they may
2004
have an above-average chance of being hospitalized
again). Conversely, some of those who had low initial
(Composite measure of quality of care, 100 = maximum)
spending may develop a chronic or acute health problem
90
that generates higher costs. As a result, when examined
over longer periods of time, health spending appears to be
less concentrated. For example, looking at Medicare
spending over a five-year period (from 1997 to 2001),
85
CBO found that the most expensive 5 percent of
Medicare beneficiaries accounted for 27 percent of
total Medicare spending (compared with 43 percent in
80
2001 alone) and that the top 25 percent of beneficiaries
accounted for 68 percent of total five-year spending
(compared with 85 percent in 2001). Analysis of younger
populations has been limited by lack of data, but one
75
study simulated expenditure patterns for workers from
0
age 25 to age 60 and suggested that the most expensive
0
5
6
7
8
9
10
25 percent of employees would account for roughly half
Annual Spending per Beneficiary
of expenditures over that 35-year period.51
(Thousands of dollars)
Geographic Variation in Health Care Spending. Spending
Source:
Congressional Budget Office based on data from the
Centers for Medicare and Medicaid Services and from the
on health care varies not only from person to person
Department of Health and Human Services, Agency for
(because of differences in their health and in the treat-
Healthcare Research and Quality, National Healthcare
ments they receive) but also from region to region in the
Quality Report, 2005 (December 2005), Data Tables
United States. In particular, per capita health care spend-
Appendix, www.ahrq.gov/qual/nhqr05/index.html.
ing varies widely within the Medicare program, and yet
Notes: The composite measure of the quality of care, based on
that variation is not correlated with measures of the qual-
Medicare beneficiaries in the fee-for-service program who
ity of care that beneficiaries receive or with available met-
were hospitalized in 2004, conveys the percentage who
received recommended care for myocardial infarction, heart
rics of overall health outcomes. In 2004, for example,
failure, or pneumonia.
Medicare spending per beneficiary ranged from about
Spending figures convey average amounts by state.
$5,600 in South Dakota to about $8,700 in Louisiana.
Yet a comparison of composite quality scores for medical
those correlations are an artifact of the calendar year and
centers and average Medicare spending per beneficiary
simply reflect treatments begun in late 2003 that contin-
shows that facilities in states with high average spending
ued into early 2004. But in other cases, those raw year-to-
are no more likely to provide recommended care for some
year correlations may understate the extent to which indi-
common health problems than are facilities in states with
viduals can anticipate their likely needs for health care in
lower spending (see Figure 1-4). Health care spending per
the near future; even if they have similar spending ini-
capita also varies widely when examined for the entire
tially, people who have had health problems that are
population—ranging from roughly $4,000 in Utah to
unlikely to recur and those who have conditions that are
$6,700 in Massachusetts in 2004—but the connection
more chronically costly can use that information in
between that variation and health outcomes has not been
choosing a health insurance plan.
51. Matthew J. Eichner, Mark B. McClellan, and David A. Wise,
As more time passes, the difference in average spending
Insurance or Self-Insurance? Variation, Persistence, and Individual
between those who initially had high expenditures and
Health Accounts, Working Paper No. 5640 (Cambridge, Mass.:
those who initially had low expenditures would tend to
National Bureau of Economic Research, June 1996).
CBO
26
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
examined as closely. Medicaid spending per enrollee also
What factors contribute to geographic variation? Some of
varies considerably among states (partly reflecting differ-
the differences in spending reflect varying rates of illness
ences in the population covered and benefits provided).
as well as differences in the prices that Medicare pays for
the same service (those prices are adjusted on the basis of
The observed variations in Medicare spending per
local costs for labor and equipment in the health sector).
enrollee are even greater when examined by the area in
But according to researchers at Dartmouth, differences in
which enrollees generally receive their hospital care, but a
illness rates account for less than 30 percent of the varia-
link between higher spending and better health in that
tion in spending among areas, and differences in prices
population is still hard to discern. In 2005, average costs
can explain another 10 percent—indicating that more
ranged from about $5,200 in the regions with the lowest
than 60 percent of the variation is due to other factors.53
spending to nearly $14,000 in those with the highest
Differences in income or the preferences of individuals
spending (those averages were adjusted to account for dif-
for specific types of care appear to explain little of the
ferences in the age, sex, and race of Medicare beneficiaries
variation in spending. Unmeasured differences in the
in the various regions). According to one study, higher-
demand for care could be important, but some of the
spending regions did not have lower mortality rates than
variation in medical practice probably is attributable to
lower-spending regions, even after adjustments were
regional differences in the supply of medical resources
made to control for different rates of illness among
(specialist physicians or health care facilities, for example)
patients and across regions.52 That study also found that
and the propensity to take advantage of the financial
higher spending did not slow the rate at which the elderly
incentives provided by Medicare or other payers in
developed functional limitations (a measure of their diffi-
developing and using those resources. Overall, patterns
culties in taking care of themselves).
of treatment in high-spending areas tend to be more
intensive than those in low-spending areas; that is, in
Other studies of spending variation reach somewhat
high-spending areas, a broader array of patients will
different conclusions, even though they also suggest
receive more costly treatments.54
opportunities to improve the efficiency of the health
sector. Some research suggests that health overall might
In sum, the evidence about variation in spending suggests
that efficiency gains in the health care system are possible:
not suffer if medical practice in higher-spending regions
Expenditures in high-spending regions could probably be
changed to match that of lower-spending regions.
lowered without producing worse outcomes, on average,
Patients who would benefit most from more expensive
or reducing the overall quality of care. But if policies that
treatments, however, might be made worse off as a result,
reduced expenditures in high-spending areas did not suc-
whereas patients who would do better with less expensive
cessfully target ineffective or harmful treatments—a chal-
treatments would gain. Other, older studies of geographic
lenging task—they might not lead to increased efficiency
variation indicate that there may be room to reduce
and could result in worse health outcomes.
spending without harming health in both high-use and
low-use regions of the country, because a large share of
certain surgeries were performed in both types of regions
53. See John E. Wennberg, Elliott S. Fisher, and Jonathan S. Skinner,
“Geography and the Debate Over Medicare Reform,” Health
even though they were found to be clinically inappropri-
Affairs, Web Exclusive (February 13, 2002), pp. W96–W114; and
ate or of equivocal value.
Center for the Evaluative Clinical Sciences, Dartmouth Medical
School, The Dartmouth Atlas of Health Care 1999 (Lebanon,
N.H.: Health Forum, Inc., 1999), pp. 22–23.
52. Elliott S. Fisher and others, “The Implications of Regional Varia-
tions in Medicare Spending, Part 2: Health Outcomes and Satis-
54. For a more extensive discussion of this topic, see Congressional
faction with Care,” Annals of Internal Medicine, vol. 138, no. 4
Budget Office, Geographic Variation in Health Care Spending
(February 18, 2003), pp. 288–298.
(February 2008).
CBO
C H A P T E R
2
Approaches for Reducing
the Number of Uninsured People
About one in six nonelderly people in the United At the federal level, subsidies for health insurance premi-
States will be without health insurance at any given time
ums have been provided through spending programs and
during 2009. Those without insurance will include nearly
tax provisions. Millions of low-income children and their
10 million children, over 14 million adults living in
parents receive subsidized health insurance coverage
families with children, and another 21 million adults who
through Medicaid and the State Children’s Health Insur-
do not reside with children. Nearly two-thirds of the
ance Program; tax subsidies, such as the exemption of
uninsured are in families whose income is less than
employer-paid premiums from taxation, encourage
200 percent of the federal poverty level.
middle- and higher-income taxpayers to purchase private
health insurance (primarily through their employer).
Concerns about the number of people who lack health
Those subsidies, however, are distributed unevenly. Some
insurance have generated proposals that seek to increase
low-income adults—particularly those who are under the
coverage rates substantially or to achieve universal or
age of 65, childless, and able-bodied—are generally not
near-universal coverage. Coverage could be expanded by:
eligible for Medicaid or SCHIP. Taxpayers who do not
work for a firm that offers coverage may not receive any
B Subsidizing health insurance premiums, either
tax subsidies for purchasing private health insurance.
through the tax system or spending programs, which
would make insurance less expensive for people who
Coverage could be expanded by restructuring tax subsi-
are eligible.
dies, spending programs, or both. However, redesigning
existing subsidies or creating new benefits raises several
B Mandating health insurance coverage, either by
issues. First, the form of the subsidy can determine who
requiring individuals to obtain coverage or by requir-
would benefit. Tax preferences, such as the current-law
ing employers to offer health insurance to their work-
exclusion or a tax deduction, reduce taxes but do not pro-
ers. If effective penalties were imposed on those who
vide benefits to those who do not have any income tax
did not comply, a mandate would increase insurance
liability. A refundable tax credit would provide full bene-
coverage by making it more costly for individuals to be
fits to individuals, regardless of whether they have any
uninsured and for employers not to offer coverage to
income tax liability, but might require some people to file
their employees.
returns solely to obtain the subsidy. A second consider-
ation is costs, which could be high depending on the
numbers of uninsured receiving the subsidies and the
B Automatically enrolling individuals in health plans,
giving them the option to refuse coverage or switch
amounts necessary to encourage them to enroll in health
plans. Recent studies suggest that automatic enroll-
plans. Targeting benefits toward specific segments of the
ment in plans that subsidize savings for retirement
population would reduce costs but could also add to the
substantially increases participation rates, especially
burden of administering a program. A third consideration
among young and low-income workers.
is the impact of the subsidies on people who already have
coverage; although subsidies would probably increase
The three approaches could also be used in combination
coverage on net, some subsidies would go to people who
to reduce the number of people who are uninsured.
would have coverage anyway, and the availability of subsi-
CBO
28
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
dies might cause some people to lose coverage because
on the basis of income, family structure, availability of
firms might drop existing plans if their workers could
insurance, or other factors. By lowering the costs of
obtain comparable health insurance elsewhere at equal or
health insurance to enrollees, subsidies encourage unin-
lower cost.
sured individuals to obtain coverage. The design of the
subsidies, however, may involve trade-offs with other
Because subsidies may not be sufficient to achieve univer-
policy goals that could affect their costs.
sal coverage, some analysts have suggested imposing a
mandate on individuals to obtain health insurance or on
A basic set of trade-offs arises between the share of the
employers to offer plans. The effectiveness of a mandate
premium that is subsidized, the number of people who
in expanding coverage would depend on its scope, the
enroll in an insurance plan as a result, and the total costs
incentives to comply, and the ease of enforcement. Many
of the subsidies. As the rate of the subsidies increases,
factors affect compliance with a mandate, and the
more people will be inclined to take advantage of them,
Congressional Budget Office will consider the specifics of
but the higher subsidy payments also go to those who
each proposal requiring individuals or employers to pur-
would have purchased insurance anyway. Beyond a cer-
chase health insurance in determining the proposal’s
tain point, therefore, the cost per newly insured person
effect on coverage.
can grow sharply because a large share of the additional
subsidy payments is going to individuals who are other-
Coverage could also be increased by automatically enroll-
wise uninsured.
ing individuals in health insurance plans, giving them the
option to refuse coverage. Automatic enrollment has
To hold down the costs of subsidies, proposals could seek
increased participation in employer-sponsored 401(k)
to limit eligibility for subsidy payments to individuals
plans and certain government programs. Firms that offer
who are currently uninsured. That restriction, however,
health insurance might be encouraged or required to
increases incentives for people who have insurance to
automatically enroll employees in a basic plan, unless the
drop their coverage (or for their employers to stop offer-
employees chose to opt out of coverage or signed up for a
ing coverage) in order to qualify for the new subsidies.
more comprehensive plan. Automatic enrollment, how-
Some proposals may try to distinguish between people
ever, might be more difficult to implement in settings
who become uninsured in response to subsidies and those
other than the workplace or public programs because of
who would have been uninsured in the absence of a gov-
the complexities of determining eligibility, collecting
ernment program, but such proposals raise significant
premium payments, and other factors.
administrative challenges. In addition, providing benefits
only to the uninsured may be viewed as unfair by people
All three approaches to expanding coverage could affect
with similar income and family responsibilities who pur-
participation in employment-based health plans. To the
chase health insurance and are therefore ineligible for the
extent that employers’ payments for health insurance
subsidies.
increased as a consequence, firms would respond over the
long term by paying lower wages and providing fewer
Another way to limit costs would be to target subsidies
fringe benefits than they otherwise would in order to
toward lower-income groups, who are most likely to be
maintain the same level of compensation. Because
uninsured otherwise, but such approaches can also have
employers’ contributions for health insurance (unlike
unintended consequences that affect the costs of the pro-
wages) are exempt from income and payroll taxes, such
posal. If eligibility was limited to people with income
changes could have substantial effects on the federal
below a certain level, then those with income just above
budget.
the threshold would have strong incentives to work less or
hide income in order to qualify for the subsidies or to
maintain their eligibility. Phasing out subsidies gradually
Methods of Subsidizing Premiums
Proposals that are designed to increase substantially the
number of people who have health insurance typically
1. This chapter focuses on subsidies to individuals or their employ-
include federal subsidies to cover some portion of the pre-
ers. Another approach would be a federal reinsurance program
mium for that coverage.1 In addition, proposals may set
that would seek to reduce premiums by covering a portion of the
eligibility for subsidies or the size of the subsidy payment
spending insurers incur for their high-cost cases.
CBO
CHAPTER TWO
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
29
Table 2-1.
Distribution of the Nonelderly Population, by Insurance Status,
Family Income, and Family Structure, 2009
(Millions)
Family Income
Relative to
Adults
Poverty Level
Children
With Children
Without Children
Total
(Percent)
Insured
Uninsured
Insured
Uninsured
Insured
Uninsured
Insured
Uninsured
Below 100
17.2
2.7
6.5
3.5
7.3
6.9
30.9
13.1
100 to 200
15.5
3.4
10.8
5.8
9.1
6.6
35.3
15.8
200 to 300
11.9
1.8
12.9
2.9
10.9
3.6
35.7
8.3
300 to 400
9.6
0.9
12.3
1.3
10.3
2.0
32.2
4.2
Above 400
17.2
0.8
28.9
1.0
36.1
1.9
82.2
3.7
____
___
____
____
____
____
______
____
Total
71.3
9.6
71.3
14.5
73.7
20.9
216.3
45.1
Source:
Congressional Budget Office’s health insurance simulation model.
Note: Children are age 22 or younger.
as income increases would reduce, but would not elimi-
The design issues raised by various subsidy systems and
nate, those incentives. At the same time, the more gradu-
their implications for the federal budget can be illustrated
ally the subsidies were phased out, the greater the number
by examining more closely the two largest subsidies cur-
of people who would be eligible for them—and the more
rently provided to the nonelderly population: the tax
likely that subsidy payments would go to those who
exclusion for employment-based insurance and the
would have had insurance in any event. The number of
Medicaid program (along with the smaller SCHIP pro-
uninsured—regardless of the individual’s age or the pres-
gram). Both the tax exclusion and Medicaid also illustrate
ence of children in his or her home—gradually declines
the many challenges involved in providing subsidies to
as family income rises above 200 percent of the federal
lower-income individuals and families, who typically
poverty level. Still, nearly 4 million uninsured individuals
have limited tax liabilities—and thus might derive little
have family income that is greater than 400 percent of the
benefit from certain types of tax-based subsidies—but
federal poverty level; however, over 80 million insured
may find it burdensome to apply for programs like
individuals have income that exceeds that level (see
Medicaid or SCHIP or may be ineligible for those two
Table 2-1).
programs under current rules.
Whatever eligibility rules are applied, subsidy systems
generally need to establish methods for determining who
Subsidizing Premiums Through the
is eligible, how much of a subsidy each person receives,
Tax System
and how the subsidy will be delivered. In particular, bas-
Most workers receive a subsidy through the tax system
ing subsidies on income requires a system for measuring
when they purchase private health insurance through
and verifying income, and trade-offs can arise between
their employer. Employers’ payments for health insurance
the timeliness and accuracy of that information. Verifying
are a form of compensation, but those payments are
eligibility could impose costs not only on the agencies
exempt from income and payroll taxes (as are most
that administer the programs but also on the individuals
employees’ payments for their share of health insurance
applying for subsidies who might choose to forgo benefits
premiums). Changes to those subsidies could have sub-
rather than bother with administrative hassles and the
stantial effects on coverage rates and the federal budget.
perceived stigma of participating in such programs. Sub-
sidy payments could go directly to individuals or could
Current Tax-Based Subsidies
instead be channeled through insurers, employers, state
The favorable tax treatment currently provided for health
governments, or other intermediaries.
insurance purchased through an employer represents the
CBO
30
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
largest single source of federal premium subsidies for the
tions for payroll taxes, and fringe benefits—takes differ-
nonelderly population. Employers may compensate their
ent forms at the two firms:
employees by paying health insurance premiums in lieu
of cash wages, but the two types of compensation receive
B Employee A works for a firm that does not offer health
very different tax treatment.2 Employers may deduct the
insurance. He receives about $37,160 in cash wages,
costs of providing that coverage as a business expense—
and his employer pays the remaining compensation—
just as they deduct employees’ wages and other forms of
about $2,840—to the government in the form of
compensation—and thus those payments are not subject
payroll taxes. Employee A pays $5,000 for a health
to corporate income taxes. But unlike wages, the costs
insurance plan in the individual market.
that employers pay for health insurance are also excluded
from the taxable income and earnings of the covered
B Employee B works for a firm that offers health insur-
employees. That portion of employees’ compensation is
ance. She receives about $32,500 in wages, and her
therefore exempt from individual income and payroll
employer pays nearly $2,500 in payroll taxes on her
taxes.
wages. In addition, she has an employment-based
health insurance plan valued at $5,000 per enrollee.
Partly as a result of that favorable tax treatment, employ-
ers that offer health insurance to their workers typically
For simplicity, assume that both workers have no other
pay a substantial share of the premium for that coverage;
sources of income and are in the 15 percent income tax
that is, the amount that employees pay directly usually
bracket; that the employee’s and the employer’s portions
covers a relatively small fraction of the total premium.
of the Social Security and Medicare payroll taxes (which
Many firms also offer their workers a “cafeteria plan,”
have a combined rate of 15.3 percent) are ultimately paid
which allows employees to choose cash or other taxable
by the workers; and that the costs of the second firm’s
benefits in lieu of receiving nontaxable benefits. (Such
health plan are borne evenly across its workforce.4
plans are referred to as Section 125 plans, after the section
of the tax code that authorizes them.) Under that
Although the two workers receive the same total compen-
arrangement, employees are able to exclude the portion of
sation and have comparable health insurance coverage,
the health insurance premium that they pay from their
their tax liabilities differ. Employee A, who purchases
taxable income—so for most workers, the full cost of the
health insurance in the individual market, pays $9,439 in
employer-sponsored plan receives favorable tax treat-
income and payroll taxes, or $1,407 more than the
ment.3
worker who receives part of her compensation in the
form of health insurance premiums. For Employee B,
The subsidy provided by the current tax exclusion shows
federal taxes have effectively reduced the cost of insurance
up as a reduction in taxes (commonly referred to as a tax
by more than 28 percent, to $3,593 (see Table 2-2). The
expenditure) rather than as an overt payment. The man-
effective subsidy rate increases by several percentage
ner in which the tax exclusion subsidizes health insurance
points if the employee lives in one of the 41 states (or the
can be seen by comparing the tax liabilities of two other-
District of Columbia) that have an individual income tax;
wise identical workers employed at different firms. Both
those states generally follow federal definitions of
workers receive $40,000 in compensation from their
earnings and other income and thus exclude employers’
respective employers in 2009, but that compensation—
contributions for health insurance from their calculation
which is a combination of wages, employers’ contribu-
of taxable income.5
2. See Chapter 1 for further discussion of the incidence of employ-
4. Although considered part of compensation, employers’ contribu-
ers’ contributions for health insurance.
tions for payroll taxes are not subject to income taxes or the
employees’ portion of payroll taxes.
3. Employees of a firm that does not offer cafeteria plans cannot
exclude their share of health insurance premiums from taxable
5. An offsetting consideration is that excluding health insurance
income for income and payroll tax purposes. However, they may
premiums from taxable wages reduces future Social Security bene-
be able to claim those premiums as an itemized deduction on their
fits, which are based on average earnings, at the same time that
income tax return if their total medical expenses exceed 7.5 per-
it reduces payroll tax payments (see Chapter 6 for further
cent of adjusted gross income.
discussion).
CBO
CHAPTER TWO
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
31
Table 2-2.
Illustrative Tax Subsidy for Employment-Based Health Insurance for a
Single Worker Who Receives $40,000 in Total Compensation, 2009
(Dollars)
Employee A:
Employee B:
Pays $5,000 for
Receives $5,000 of
Individual Health
Employment-Based
Insurance
Health Insurance
Difference
Compensation
Cash wages
37,157
32,513
4,644
Premiums for employment-based health insurance
0
5,000
-5,000
Employers' contribution for payroll taxes
2,843
2,487
355
______
_______
_____
Total compensation
40,000
40,000
0
Premiums for Health Insurance in Individual Market
5,000
0
5,000
Income Tax
Adjusted gross income
37,157
32,513
4,644
Minus personal exemption
3,650
3,650
0
Minus standard deduction
5,700
5,700
0
______
______
_____
Taxable income
27,807
23,163
4,644
Income tax
3,754
3,057
697
Payroll Tax
Employee's contribution at 7.65 percent
2,843
2,487
355
Employer's contribution at 7.65 percent
2,843
2,487
355
_____
_____
____
Total payroll tax
5,685
4,974
711
Total Income and Payroll Taxes
9,439
8,031
1,407
After-Tax Cost of Health Insurance
5,000
3,593
1,407
Subsidy as a Percentage of Costs of Health Insurance
0
28
-28
Source:
Congressional Budget Office.
Note: To simplify the example, both workers are assumed to be unmarried, to have no dependents, to receive $40,000 in total compensa-
tion, and to have no sources of income other than wages and salaries.
The aggregate effects of that exclusion on the federal bud-
annually by allowing self-employed individuals to deduct
get are large, exceeding federal spending on Medicaid.
the costs of health insurance from their taxable income
The Joint Committee on Taxation has estimated that the
(but health insurance costs for the self-employed are not
total federal tax expenditure associated with the exclusion
deductible for purposes of payroll taxes). However, the
for employment-based health insurance was $246 billion
magnitude of the estimated tax expenditures is not the
in 2007, consisting of $145 billion in individual income
same as the increase in revenues that would result from
taxes and $101 billion in payroll taxes.6 (By comparison,
repealing the current exclusion or the deduction for the
the federal government spent over $195 billion on
self-employed, because the calculation of the tax expendi-
Medicaid in 2007.) In addition, the federal government
tures does not account for any changes in taxpayers’
incurs an additional tax expenditure of about $5 billion
behavior that would result if the exclusion was repealed.
(The revenue gain from repeal would be less than the esti-
mated tax expenditures because some individuals would
6. Joint Committee on Taxation, Tax Expenditures for Health Care,
JCX-66-08 (July 30, 2008).
find other ways to reduce their tax liabilities if the exclu-
CBO
32
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
sion was repealed; some individuals, for example, might
Taxpayers who itemize deductions on their income tax
claim their health insurance premiums as an itemized
return may deduct unreimbursed medical expenses,
deduction.)
including any premiums and out-of-pocket expenses that
they paid out of after-tax income. The deduction is
The current tax treatment of health insurance premiums
generally limited to expenses in excess of 7.5 percent of
encourages employers to offer health insurance to their
adjusted gross income; for example, a taxpayer with
employees and encourages employees to enroll in those
$50,000 in adjusted gross income could deduct medical
plans, but it has also raised several concerns. In particular,
costs in excess of $3,750. Furthermore, the total amount
the exclusion does not provide benefits for health insur-
of itemized deductions is gradually reduced for taxpayers
ance evenly. Individuals with the same income and
with adjusted gross income above $166,800 in 2009. In
similar family responsibilities can receive very different
2007, the tax expenditure for the itemized deduction for
tax benefits for medical costs. Employees who can
medical expenses was about $9 billion.
exclude premiums for employment-based insurance from
payroll taxation, as well as from individual income taxes,
The health coverage tax credit covers up to 65 percent of
typically receive more generous tax subsidies than do self-
the cost of health insurance for certain dislocated work-
employed individuals. Employees who work for firms
ers. Because the credit is refundable, individuals can
that do not offer insurance do not benefit from the
claim the full benefit even if its value exceeds their
exclusion.
income tax liabilities. To be eligible, individuals must be
receiving either trade adjustment assistance or payments
In addition, the current system provides different tax sub-
from the Pension Benefit Guaranty Corporation (which
sidies to people at different income levels. Because the
pays at least a portion of the pension benefit promised to
rate structure of the income tax is progressive—that is, as
retired workers if their company goes out of business or
income rises, each additional dollar of income may be
otherwise defaults on its obligations). The credit is not
taxed at a higher rate—the value of the exclusion gener-
available to people receiving certain other government
ally grows as income increases. If, in the example above,
health benefits, including Medicare. In 2007, the tax
the single employee with an employer-sponsored health
expenditure for the credit was about $100 million.7
insurance policy worth $5,000 had earned $70,000 in
total compensation instead of $40,000, that individual
Options to Modify Tax Subsidies for Health
would probably be in the 25 percent rate bracket; being
Insurance
in that higher bracket would increase the total tax savings
Tax subsidies could be redesigned in several ways to
by $465 (from $1,407 to $1,872) and raise the federal tax
expand coverage. One option would be to replace the
subsidy to over 37 percent. The share of the premiums
current exclusion of premiums for employment-based
that the federal exclusion offsets can be somewhat lower
health insurance with a tax deduction or a tax credit.
at higher levels of income if taxpayers reach the wage ceil-
Another option would be to provide new subsidies to
ing for Social Security payroll taxes ($106,800 in 2009).
employers, in the form of tax credits, to encourage them
The value of the exclusion represents a larger percentage of
to offer health insurance and to pay a portion of their
income for middle-income households than for high-
employees’ premiums. (Such an option could replace or
income households, however, largely because average
supplement the current-law exclusion or be combined
premiums for health insurance do not vary substantially
with new credits or deductions for individuals.)
with income and therefore decline as a share of income as
income rises.
Replacing the Exclusion with a Tax Deduction or a
Tax Credit. The exclusion of premiums for employment-
Although the exclusion of employer-paid premiums is by
based insurance could be replaced with a deduction or a
far the largest tax expenditure related to health care, two
tax credit that is designed to encourage coverage. In addi-
others worth noting are the itemized deduction for medi-
tion, eligibility for those tax deductions or credits could
cal expenses and the health coverage tax credit that is
be extended to all taxpayers who purchase health insur-
available for workers displaced from their jobs by interna-
tional trade. (For a general discussion of the key differ-
7. The estimate of the tax expenditure includes the amounts (or out-
ences between tax exclusions, tax deductions, and tax
lays) paid to taxpayers in excess of their income tax liability, which
credits, see Box 2-1.)
result from the refundable nature of the health coverage tax credit.
CBO
CHAPTER TWO
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
33
Table 2-3.
Effects on a Single Worker of Repealing the Tax Exclusion and
Replacing It with an Above-the-Line Deduction, 2009
(Dollars)
Change from Current Law
Effect of
Effect of Repealing
Above-the-Line
Exclusion for
Deduction for
Current-Law
Employment-Based
All Health
Combined
Taxes After
Type of Tax
Taxes
Health Insurance
Insurance
Effect
Change
Employee A: Pays $5,000 for Individual Health Insurance
Individual
3,754
0
-750
-750
3,004
Payroll
5,685
0
0
0
5,685
_______
_
_____
_____
______
Total taxes
9,439
0
-750
-750
8,689
Employee B: Receives $5,000 of Employment-Based Health Insurance
Individual
3,057
697
-750
-53
3,004
Payroll
4,974
711
0
711
5,685
______
______
_____
____
______
Total taxes
8,031
1,407
-750
657
8,689
Source:
Congressional Budget Office.
Notes: To simplify the example, both workers are assumed to be unmarried, to have no dependents, to receive $40,000 in total compensa-
tion, and to have no sources of income other than wages and salaries. For Employee B, repealing the exclusion causes the employer’s
contributions for payroll taxes to rise by $355 and cash wages to fall by an offsetting amount.
An above-the-line deduction is subtracted from total income to derive adjusted gross income. Taxpayers can claim both an above-the-
line deduction and the standard deduction.
ance, including those who purchase policies in the indi-
itemized deduction for medical expenses), taxpayers can
vidual market. Providing tax preferences for individually
claim the deduction only if they itemize instead of claim-
purchased health insurance, however, could cause some
ing the standard deduction. In contrast, taxpayers can
employers to drop plans because they realize their workers
claim “above-the-line” deductions along with their stan-
have alternative tax-preferred options. In response, some
dard deduction.
of those workers may switch to individually purchased
insurance, and others may become uninsured. (The likely
Consider, again, the two single workers who each earn
magnitudes of those responses are discussed later in this
$40,000 in total compensation but one receives health
chapter.)
insurance at work and the other purchases a comparable
policy in the individual market. Employee A, who pur-
The different structure of tax deductions and tax credits
chases a health insurance policy for $5,000 in the individ-
affects not only the value that various types of individuals
ual market and does not itemize deductions, receives no
and families will derive from them but also the impact
that those subsidies will have on insurance purchases.
Like the current exclusion for health insurance premi-
8. In its 2008 and 2009 fiscal year budgets, the Bush Administration
ums, a deduction reduces taxable income, causing the
proposed a deduction that applied to both the individual income
value of the deduction to increase as income and mar-
and payroll tax bases. However, exempting premiums for individ-
ually purchased health insurance from payroll taxes presents
ginal tax rates rise. A deduction is subtracted from total
administrative challenges; employers cannot easily adjust with-
income solely for purposes of computing the income tax
holding for such premiums because they do not have independent
and thus may have no impact on payroll taxes—unlike
information regarding how much each of their workers paid for
the existing exclusion.8 In some cases (as with the current
health insurance in the individual market.
CBO
34
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
Box 2-1.
Tax Exclusions, Tax Deductions, and Tax Credits
Several types of subsidies for health care costs are
retirement savings plans are not counted as income
embedded in the current structure of the individual
for employees, and employees’ contributions are
income tax. Proposed tax subsidies can take the form
subtracted from their earnings when determining the
of exclusions, deductions, or credits, each of which
amount that is reported as taxable. (Contributions to
has a different structure and different effects on indi-
401(k) plans are still subject to payroll taxes, how-
vidual income tax liabilities. Briefly,
ever.) Similarly, the amounts that employers pay for
employees’ health insurance are not counted as
B A tax exclusion reduces the amount that tax filers
taxable income for employees, thus subsidizing the
report as their total, or gross, income.
purchase of employment-based health insurance.
B A tax deduction is an expense that is subtracted
Tax Deductions
from total income when calculating taxable
There are several types of income tax deductions. All
income. It reduces tax liability in proportion to an
taxpayers may subtract certain types of income or
individual’s tax bracket.
expenses—commonly referred to as above-the-line
deductions—from total income to derive their
B A tax credit is the direct dollar-for-dollar reduc-
adjusted gross income. Those deductions may try to
tion of an individual’s tax liability. If the tax credit
adjust for differences among taxpayers in terms of
is refundable, individuals can receive its full
family or other personal characteristics or to meet
amount even if they do not have any income tax
other goals of tax and social policy. For example,
to offset.
people who move more than a specified distance may
deduct their moving expenses, and contributions to
Tax Exclusions
individual retirement accounts may also qualify (up
Certain forms of compensation are excluded from
to an annual limit) for an above-the-line deduction.
taxable income, effectively providing a subsidy for the
Similarly, self-employed individuals may deduct the
excluded amount. Some types of income are excluded
full cost of their health insurance. (However, the self-
because they are difficult to measure. Other types of
employed are not allowed to exclude health insurance
income are excluded to reflect policy choices to
premiums from their income for purposes of payroll
encourage taxpayers to engage in a particular activity.
taxes.)
For example, employers’ contributions to 401(k)
tax benefit for his or her premiums under current law; an
In contrast, tax credits can be designed to provide lower-
above-the-line deduction for the costs of health insurance
and moderate-income taxpayers with larger subsidies
would lower that worker’s taxes by $750 (see Table 2-3).
than they would receive from tax deductions or exclu-
That same proposal would increase taxes by $657 for
sions. A credit could reduce income tax liabilities by a
Employee B, who receives $5,000 in employment-based
fixed amount, or it could have a progressive rate schedule,
health insurance; that worker’s taxes would rise because
thereby reducing the dollar value of the tax credit as
the amount spent on employment-based health insurance
income rises.
would no longer be exempt from payroll taxes. In con-
trast to current law, both workers would pay the same
An important issue with tax credits—particularly for
total amount of taxes—$8,689—if the above-the-line
lower-income individuals and couples that pay relatively
deduction replaced the current exclusion.
little in income taxes and are more likely to be unin-
CBO
CHAPTER TWO
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
35
Box 2-1.
Continued
Tax Exclusions, Tax Deductions, and Tax Credits
Starting from adjusted gross income, taxable income
and deducts the cost of a $5,000 health insurance
is computed by subtracting personal exemptions and
policy reduces his or her taxes by $1,250; in the
either a standard deduction amount or the total
35 percent bracket, the tax savings is $1,750.
amount of itemized deductions, and it is generally to
taxpayers’ advantage to subtract the larger of the two.
Tax Credits
In 2009, the standard deduction ranges from $5,700
Tax liabilities can be reduced by tax credits. For
for single filers to $11,400 for married couples filing
example, a portion of the costs that working parents
jointly. Expenses that are allowed as itemized deduc-
incur for child care can be taken as a tax credit. An
tions include property taxes and mortgage interest,
important distinction between tax credits, on the one
state and local income taxes, and charitable contribu-
hand, and exclusions and deductions, on the other, is
tions; medical expenses not covered by insurance are
that a tax credit can be designed so that its dollar
also allowed, but only to the extent that those
value does not depend on one’s tax bracket.
expenses exceed 7.5 percent of adjusted gross income.
The value to taxpayers of allowing itemized deduc-
Most tax credits are nonrefundable, however, mean-
tions for certain expenses thus depends in part on
ing that the actual credit that taxpayers receive cannot
what other expenses they have that can be itemized
exceed their income tax liability. Because lower-
and how those expenses compare with their standard
income individuals and families generally owe less in
deduction. In general, higher-income households are
income taxes than those with higher income, they are
more likely to itemize their deductions, although the
less likely to benefit from nonrefundable tax credits.
total amount of itemized deductions that can be
taken is gradually reduced for taxpayers whose
Some tax credits are refundable, however, allowing
adjusted gross income exceeds $166,800.
individuals to receive the entire credit amount
regardless of their income tax liability. The only
Tax liabilities are next determined by applying the
example of a tax credit related to health care is a
statutory tax rates, currently ranging from 10 percent
refundable one for workers who lost their job as a
to 35 percent, to taxable income. The value of tax
result of international trade and are receiving trade
exclusions and deductions generally depends on an
adjustment assistance (certain other workers are also
individual’s marginal tax rate—the rate that applies to
eligible); they may be eligible for a tax credit for 65
the last dollar of income. For example, a self-
percent of the costs of their health insurance.
employed person who is in the 25 percent tax bracket
sured—is whether the credits are refundable. If they are
the tax credit would be limited by the amount of the sec-
not, the value of the credit may not exceed a taxpayer’s
ond worker’s income tax liability,
income tax liability. Compare two workers who each
effectively reducing the costs of his or her health insur-
purchase a health insurance policy in the individual mar-
ance by $1,180 (to $820). If, instead, the credit was made
ket that costs $2,000; however, one worker earns $40,000
refundable, the second worker would receive the full
and the second earns $20,000. Under current law, the
amount of the tax credit—providing the lower-earning
worker who earns $40,000 pays $4,180 in income taxes.
worker with the same subsidy for health insurance as the
Because that worker has more than $2,000 of income tax
higher earner.
liability, he or she would be entitled to the full $2,000 tax
credit; the credit thus effectively reduces the costs of
Recent expansions in child-related tax credits have
health insurance to zero. In contrast, the worker with
increased the amount of income a taxpayer with children
lower earnings owes $1,180 in income taxes. The value of
must have before he or she owes any individual income
CBO
36
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
Figure 2-1.
effectiveness of a tax credit, however, could be lessened if
Distribution of Marginal Tax Rates on
low-income individuals with limited resources had to
Income for the Nonelderly Uninsured
wait until they filed tax returns to claim the benefit or if it
was difficult to reach eligible individuals because they do
Population, 2009
not file tax returns. For those reasons, some proposals
would make tax credits payable in advance as well as
(Percent)
refundable, but doing so would raise a number of addi-
30 Percent or
tional administrative issues (see Box 2-2).
Higher
No Income Tax
Liability
Providing Tax Credits for Employers. To encourage firms
Rate of
9%
to offer health insurance to their employees, some pro-
20 Percent to
21%
30 Percent
posals would provide subsidies to businesses that contrib-
ute toward their employees’ health insurance, with the
18%
expectation that those subsidies would ultimately benefit
workers. Employers could receive a tax credit to cover a
12%
Rate Lower
Than
specified percentage of the per-worker cost of health
10 Percent
insurance or a fixed-dollar amount per worker. As with
40%
tax credits for individuals, tax credits for businesses can
be designed so that their value does not depend on the
business’s marginal tax rates. In a competitive labor mar-
Rate of
ket, such subsidies would be passed on to employees in
10 Percent to
the form of higher wages or lower premiums for health
20 Percent
insurance.
Source:
Congressional Budget Office’s health insurance simulation
model.
Providing subsidies to businesses entails many of the
same issues and trade-offs that arise in providing subsidies
Note: A dependent is assigned the marginal tax rate for the tax-
payer who claims him or her as a dependent.
to individuals. Small firms are less likely to offer health
insurance to their workers, particularly if a large share
tax, making it more difficult to target assistance toward
of those workers has low income. Thus, in firms with
lower-income families through the tax system unless
fewer than 25 employees, more than half of full-time
credits are made refundable. In 2009, a married couple
workers who have income between 100 percent and
with two children may not owe any income taxes unless
200 percent of the federal poverty level lack insurance
their adjusted gross income is about 200 percent of the
(see Figure 1-2). Partly as a result, small businesses are
federal poverty level (approximately $44,000 for a family
more likely to respond to a subsidy for insurance than are
of four); in contrast, for workers without children, the
larger firms. To reflect those relationships, proposals
threshold for owing income taxes is close to the poverty
might target tax credits toward smaller firms or to firms
level. In 2009, over 20 percent of the people who are pro-
that do not currently offer health insurance.
jected to be uninsured at any given time either do not
have any income tax liability or are claimed as a depen-
Such targeting strategies could reduce the cost of a sub-
dent by others who do not owe any income taxes. Nearly
sidy proposal but would also lead some employers to
half as many are in the 10 percent income tax bracket and
respond in ways that would diminish the budgetary bene-
thus may not have sufficient income tax liability to
fits of targeting. Basing subsidies on the size of a firm’s
receive the full benefit of a nonrefundable tax credit (see
workforce might discourage some businesses from
Figure 2-1).
expanding and encourage others to reorganize into
smaller entities in order to take advantage of the subsi-
Because low-income individuals are also more likely to be
dies. Phasing out subsidies as the size of firms increases
uninsured, a refundable tax credit could be more effective
would reduce those incentives but would also make more
in increasing health insurance coverage than other forms
firms eligible for the subsidies than would a strict cutoff
of tax-based subsidies for the same budgetary costs. The
based on size. As with subsidies provided directly to indi-
CBO
CHAPTER TWO
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
37
viduals, requiring that recipients had not previously
receiving payments from the government for the differ-
offered insurance to their workers could raise concerns
ence between the average cost of providing coverage and
about equitable treatment of similar firms—some of
the premium (if any) that enrollees are charged. The
whom had already offered coverage—and would also
primary determinants of a proposal’s cost would be the
create an incentive for firms to drop coverage in order to
amount of the subsidy per enrollee and the number of
qualify.
participants.
Proposals to provide tax credits to employers would also
Many of the factors that would affect eligibility for and
need to address the issue of whether the credits are
participation in a publicly funded program—and thus
refundable or payable in advance. As is the case with tax
the federal costs involved—can be illustrated by examin-
credits for individuals, some employers—particularly
ing the current rules for Medicaid and SCHIP. Whether
smaller employers—may not have sufficient income tax
new subsidies would be provided by expanding those pro-
liability to take full advantage of a credit. Similarly,
grams or creating a new program, similar design issues
smaller employers may have liquidity problems, making
would arise. In particular, CBO’s estimates of program
it difficult for them to cover the costs of an insurance pol-
participation and costs would depend heavily on such
icy if they have to wait until they file their tax return to
factors as:
receive a tax credit for health insurance. Those consider-
ations would affect whether firms took advantage of the
B Whether and how family and personal characteristics
subsidies that they are offered.
affect eligibility,
Trying to target subsidies according to the size of a firm
B Whether and how income is counted and used to
and the characteristics of workers would also raise admin-
determine eligibility or the size of the subsidy (or
istrative challenges. For example, the Internal Revenue
both),
Service currently does not have sufficient information to
target subsidies on the basis of the size of a business’s
B Whether and how asset holdings will be taken into
workforce. Quarterly and annual reports on withholding
account,
taxes contain some information on the number of
employees, although those reports do not specify whether
B Whether and how eligibility will be targeted toward
employees work full or part time or on a temporary or
individuals who are otherwise uninsured,
permanent basis. Tax subsidies could, instead, be targeted
toward businesses on the basis of gross receipts (as
B The process for determining and recertifying eligibil-
reported to the IRS), but some small firms might not
ity, and
qualify as a consequence. Basing subsidies on workers’
characteristics (such as their earnings or the number of
B The incentives for states to participate if some state
hours they work each week) would raise additional
financing is required.
complexities.
Family and Personal Characteristics
Subsidizing Premiums Through
As with tax-based subsidies, spending programs that are
designed to provide subsidies to or coverage for families
Spending Programs
or children need to define rules for determining who
Proposals designed to increase coverage rates may use
counts as a family member and who is responsible for the
spending programs rather than tax provisions to subsidize
health care of children. In many cases, those relationships
insurance premiums. One reason for taking that
will be straightforward, but a variety of challenges may
approach is the difficulty of reaching many uninsured
arise. For example, determining who may claim a
individuals through the income or payroll tax systems in
dependent for a benefit can be difficult, particularly when
a timely fashion. Subsidies provided through spending
children live with a divorced parent or in an extended
programs could take the form of direct payments to indi-
family. One option is to provide the benefit to the
viduals for purchasing health insurance. Alternatively,
custodial parent, who is generally the recipient of other
individuals could receive the subsidy indirectly through
child-related benefits. In many instances, however, health
reductions (or elimination) of premiums, with the insurer
insurance is obtained by noncustodial parents (who, in
CBO
38
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
Box 2-2.
Issues with Refundable Tax Credits
The value of a tax credit to people who owe little or
employment taxes. Others may have filed to obtain a
no income tax depends crucially on whether the
refund of overwithheld amounts or to claim a refund-
credit is refundable, because the value of a
able tax credit like the earned income tax credit
nonrefundable credit cannot exceed the recipient’s
(EITC). Nearly all of those “nontaxable filers” were
income tax liability. Implementing refundable tax
employed. In contrast, nearly all individuals with
credits in a way that helps low-income households
legitimate reasons for not filing a return had low
purchase health insurance, however, presents at least
income and little or no attachment to the workforce.
two challenges:
For the estimated 28 million nonfilers with no
B Reaching all eligible individuals, many of whom
income tax liabilities, obtaining subsidies through the
may not file income tax returns; and
income tax system could require them to file a return.
Experience with the recent economic stimulus pack-
B Making the funds available in a timely manner—
age illustrates that many such individuals may not
that is, before the premium payments are due.
file, even though they would gain financially by
doing so. In 2008, individuals were eligible for stimu-
Reaching Eligible Individuals, Including
lus payments if they had at least $3,000 of income
Nonfilers
from earnings, Social Security, or veterans’ benefits,
Making tax credits refundable allows people who
even if they had no income tax liabilities. As of
have little or no income tax liability to receive a bene-
September 2008, however, about 4.2 million retirees
fit, but many may have to file forms with the Internal
and disabled veterans who normally do not file a tax
Revenue Service (IRS) solely for the purpose of
return but were eligible for the stimulus payments
obtaining the subsidy. Filing behavior differs among
had not yet claimed the one-time benefit. (Individu-
those who have no income tax liability, largely
als who were induced to file a return by the stimulus
depending on whether they work or not. In early
package were not included in JCT’s analysis cited
2008, the Joint Committee on Taxation (JCT) esti-
above, which was prepared before the enactment of
mated that more than 66 million potential tax filing
that legislation.)
units (primarily individuals or married couples) did
not have any income tax liability.1 Of those, more
1. Joint Committee on Taxation, Overview of Past Tax Legisla-
than half were expected to file a tax return for 2007.
tion Providing Fiscal Stimulus and Issues in Designing and
Many were required to file because their income was
Delivering a Cash Rebate to Individuals, JCX-4-08R
above the IRS filing threshold or they owed self-
(February 13, 2008).
fact, may be required to do so under provisions for child
or pregnant and to families with children; able-bodied
support). In the Medicaid program, custodial parents
adults who do not have children are generally ineligible.
may apply for benefits for themselves and their children,
Individuals are presumed to be eligible for Medicaid in
but the state is authorized to take cost-effective measures
most states if they receive Supplemental Security Income
to determine the legal liability of noncustodial parents to
(SSI) benefits (which go to elderly and disabled individu-
contribute toward the costs of the children’s benefits.
als) or would have met the eligibility criteria for Aid to
Families with Dependent Children (AFDC) that were in
Eligibility could be narrowed on the basis of other fac-
effect when welfare reform was enacted in 1996. SCHIP
tors, including age, disability, or participation in another
primarily covers children under the age of 19. Medicaid
public assistance program. For example, Medicaid eligi-
and SCHIP both provide mechanisms for states to extend
bility is limited to individuals who are elderly, disabled,
eligibility to other groups by requesting waivers of
CBO
CHAPTER TWO
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
39
Box 2-2.
Continued
Issues with Refundable Tax Credits
Paying Credits in a Timely Manner
Tax credits could also be paid in advance. For exam-
A tax benefit meant to encourage the purchase of
ple, eligible individuals can claim advance payments
health insurance is less effective if beneficiaries must
of the EITC and the health coverage tax credit. One
wait a long time to receive the subsidy. Some low-
challenge of providing advance payments of tax cred-
income individuals will lack the resources to purchase
its is verifying eligibility before the end of the tax
health insurance until they receive the subsidy. Sev-
year—particularly if eligibility or payment amounts
eral studies have found that individuals procrastinate
depend on total income received during the year.
when faced with large up-front costs and complicated
Workers may claim advance payments of the EITC
choices.2 Long lags between the premium’s due dates
by giving a form to their employer, who then pro-
and the receipt of tax subsidies may result in low
vides them with a prorated amount of the credit in
enrollment rates.
their paycheck based on their projected earnings. A
change in circumstances during the year (for exam-
Generally, however, taxpayers do not claim tax credits
ple, a spouse’s entering the workforce) could cause
until they file their tax return at the end of the year,
the couple’s income to rise, reducing the amount of
so they might not receive any benefit from reduced
the EITC to which they are entitled for the tax year.
tax payments or larger refunds until after their insur-
As a consequence, they would be required to repay
ance premiums are due. In principle, individuals can
the overpayment when they filed their tax return.
adjust the amount of taxes withheld from their pay-
Some analysts believe that the risk of such overpay-
checks in order to accelerate the receipt of tax bene-
ment discourages eligible individuals from claiming
fits, but many low-income individuals have little or
the EITC in advance. One recent study found that
no income tax withholdings to adjust and cannot
only 3 percent of eligible taxpayers claim the EITC in
reduce their withholding below zero.
advance, and that among those who do claim
advance payments, erroneous payments are prevalent
and difficult to recapture.3
2. See, for example, Janet Currie, The Take-Up of Social Benefits
3. Government Accountability Office, Advance Earned Income
(paper prepared for the Conference in Honor of Eugene
Tax Credit: Low Use and Small Dollars Paid Impede IRS’s
Smolensky, Berkeley, Calif., December 2003; revised June
Efforts to Reduce High Noncompliance, GAO-07-1110
2004).
(August 2007).
existing rules from the Department of Health and
abled enrollees must generally have low income as well.
Human Services. For example, some states have been
SCHIP covers children in families with income up to 200
granted waivers to cover able-bodied individuals who
percent of the poverty level, although some states use
have no dependents under Medicaid or to cover parents
higher income limits. Those income cutoffs provide
of eligible children under SCHIP.
incentives for potential enrollees to reduce or hide their
income in order to maintain their eligibility. Those incen-
Income
tives may be reduced (but not eliminated) by phasing out
Subsidies can be targeted toward lower-income individu-
subsidies more gradually as income rises—for example,
als and families by reducing or eliminating the subsidy as
by charging premiums to enrollees on a sliding scale—
income rises. Medicaid covers pregnant women and
but that approach tends to increase the number of people
children under age 6 whose family income is below
who are eligible and thus raises the program’s costs.
133 percent of the federal poverty level. Older children
(younger than 19) are covered if their family income is
Tying subsidies to income requires rules for determining
below 100 percent of the poverty level. Elderly and dis-
how to count the income of individuals and families.
CBO
40
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
That requirement raises several issues: whose income to
for individuals or $3,000 for couples, but the program
include, what types of income to include, and the period
does not include the value of a person’s home, burial sites,
over which to measure income.
or, in most cases, car. Asset tests, however, can be difficult
to administer and may discourage saving. Beneficiaries
Whose Income to Include. Eligibility for Medicaid and
and program administrators may have to assess the value
SCHIP is based on family income, including income
of property and other investments for which complete
received by the head of a family, his or her spouse, and
and reliable information may not be available. Asset tests
possibly their children. States, however, have some discre-
also create incentives for individuals to divest themselves
tion in defining whose income to include in the family
of current investments and avoid accumulating new assets
unit.
in order to qualify for assistance.9
What Types of Income to Include. Some sources of
Under current law, states may restrict eligibility for
income may be excluded because they are difficult to
Medicaid or SCHIP on the basis of assets but are not
measure. Public programs such as Medicaid or SCHIP
required to do so. As of January 2008, 45 states plus the
generally reduce gross income by deducting the amount
District of Columbia did not consider assets when deter-
of the costs that individuals incur in earning income;
mining whether children qualify for Medicaid or SCHIP;
such “income disregards” include costs for child care and
21 states and the District of Columbia disregarded assets
other work-related expenses. In addition, individuals with
for determining parents’ eligibility for those programs.
higher income may qualify for the program if they incur
For disabled and elderly enrollees, however, eligibility for
high medical expenses, which may be deducted from
Medicaid can be linked to SSI, so the same asset tests may
their income so that they “spend down” to become eligi-
apply.
ble. Not surprisingly, broader measures of income would
tend to reduce the number of people who could meet a
Insurance Status
given income standard, whereas greater use of income
One of the more vexing issues that arise when providing
disregards would tend to expand eligibility.
assistance for health insurance is whether applicants have
to be currently uninsured to receive a subsidy. On the one
Choosing a Period Over Which to Measure Income. Cur-
hand, it may be considered inequitable to subsidize
rent income best reflects a potential beneficiary’s ability to
uninsured individuals and families while providing no
pay for health insurance, but income may fluctuate and
assistance to those with similar income and family
may be more difficult to measure accurately. Many fed-
responsibilities who have incurred the costs of purchasing
eral programs use relatively short accounting periods to
health insurance. On the other hand, offering subsidies to
measure income (for example, the current month). For
those who are already covered by insurance—referred to
any given income standard, using shorter accounting
as “buying out the base”—can substantially increase a
periods is likely to increase eligibility because monthly
program’s costs. As a result, many policy proposals, par-
income tends to vary more than annual income. The
ticularly those that would expand Medicaid or SCHIP,
income tax system bases eligibility for tax subsidies on all
seek to contain federal costs by targeting subsidies toward
income received during a year, but that information is
the uninsured alone. However, determining who is unin-
not available until the following year, and data from a
sured and thus qualified for subsidies raises a number of
previous year’s tax return may overstate or understate cur-
policy and administrative concerns that could affect a
rent resources.
program’s costs.
Assets
One concern is that determining who is uninsured can be
Proposals may further target subsidies through asset tests.
difficult because of the dynamic nature of insurance
Such tests measure the value of certain investments that
status. Many people are uninsured for short periods of
families could liquidate to pay for their needs (in this
time—for example, they may lose coverage for a few
case, health insurance) in order to determine eligibility.
months while between jobs. During the period that they
Some types of assets may be excluded because they are
difficult to sell or are viewed as a necessity. For example,
to be eligible for the Supplemental Security Income pro-
9. See Chapter 8 for a discussion of the impact of asset tests on
gram, applicants’ assets must be worth less than $2,000
savings.
CBO
CHAPTER TWO
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
41
are uninsured, they may qualify for subsidies; however,
The trade-off between strict policies to prevent crowd-out
unless program administrators can certify eligibility
and efforts to encourage maximum enrollment among
continuously—which would raise administrative costs
the targeted population has led many states to allow
substantially—they may continue to receive the subsidies
exceptions to waiting periods (for example, if a family
for some time after they start their new job and health
loses insurance coverage because of a job loss). However,
plan.
states’ strategies to prevent crowd-out have apparently
met with limited success. A number of studies analyzing
Another concern is that public programs that are targeted
crowd-out in the Medicaid and SCHIP programs have
toward the uninsured increase incentives for individuals
varied widely in their methodologies and in their find-
(or their employers) to drop their current coverage and
ings. On the basis of a review of that empirical literature,
switch to the new subsidized program.10 To the extent
CBO has estimated that for every 100 children who gain
that such crowd-out occurs—with public insurance
coverage as a result of SCHIP, there is a corresponding
replacing private insurance—program costs rise without
reduction in private coverage of between 25 and 50 chil-
reducing the number of people who are uninsured.
dren.12 The extent to which that finding would apply to
Determining whether applicants were previously unin-
new proposals is difficult to predict, but as a general mat-
sured may not be difficult; most private insurance is pro-
ter the probability of crowd-out increases as eligibility for
vided through employers, and program administrators
a public program is extended to people at higher income
can check with them about an applicant and do not have
levels, simply because a greater share of the population
to rely on self-reporting. Yet program administrators can-
already has private insurance. Because incomes, the costs
not easily determine the reason an applicant is uninsured
of health insurance, and coverage rates vary geographi-
or distinguish between applicants who would have been
cally, the extent of crowd-out under a given subsidy
uninsured in the absence of the subsidized program and
schedule is also likely to vary from region to region.
those who lack insurance because of incentives to drop
coverage.11
Crowd-out may occur even if eligibility for a public
program is not limited to the uninsured. If a new public
States’ experiences with SCHIP provide some insight into
program provides a larger effective subsidy than the one
the challenges of minimizing crowd-out. One common
that privately insured individuals receive, those individu-
approach has been to impose a waiting period, specifying
als will have a financial incentive to shift to the public
the length of time that previously insured children must
program. That incentive can be blunted if the subsidies
be uninsured before they are permitted to enroll in the
are actively extended to privately insured individuals who
program. As of July 2007, 37 states imposed a waiting
would otherwise qualify, as was done for the Medicare
period (usually three to six months). Although that
drug benefit. Extending new subsidies to individuals who
approach discourages privately insured individuals from
already have private insurance tends to increase total
dropping their coverage, it also imposes a hardship on
program costs, however, and also generates crowd-out of
affected applicants by making them wait longer to receive
another type—substituting public spending for private
coverage.
spending without changing the source of insurance for
those individuals.
Certifying and Verifying Eligibility
10. Such a change in coverage need not result from previously insured
individuals dropping their current policies to enroll in public
Targeted benefits require that program administrators
plans, although that is one possibility. Another possibility, for
certify eligibility and enforce the program’s rules. Benefit
example, is for people who lost their coverage to decline coverage
programs differ in their approaches toward certification
at a new job if they now qualify for a public program.
and enforcement, and those differences have effects on
11. Efforts to minimize crowd-out become even more challenging
coverage and program costs. Two common approaches
over time, as some firms go out of business and other firms are
are the self-assessment model (generally with third-party
created. A new firm, which previously might have offered its
verification) and the caseworker model, each of which has
employees health insurance, may decide not to offer coverage,
knowing that many of its workers can obtain public subsidies
instead. The likelihood of that effect goes up as the share of the
12. Congressional Budget Office, The State Children’s Health Insurance
firm’s workforce that qualifies for subsidies increases.
Program (May 2007), pp. 11–12.
CBO
42
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
different implications for accuracy rates and for adminis-
worker model uses in-office visits, up-front requests for
trative burdens on the programs and applicants. A related
documentation (such as pay stubs), and phone calls to
issue is the time between recertifications of eligibility for
third parties to validate nearly all statements made by
enrollees, which raises the same trade-offs between
every applicant. As a result, the application process may
accuracy and administrative burdens.
be time-consuming for program administrators and for
applicants. That factor, and any stigma that applicants
Self-Assessment Model. Under this model, individuals
may associate with being in the program, tends to reduce
assert their eligibility on an application form, and agen-
participation.
cies try to verify at least some applicants’ statements with
information from reliable third parties. That approach is
The differences in the verification processes used by vari-
used to verify eligibility for many tax subsidies, including
ous programs affect participation, compliance with the
the earned income tax credit, as well as some spending
programs’ stated rules, and administrative costs. For
programs (such as the school lunch program).
example, when some EITC claimants were asked to
obtain and submit documentation showing that they met
The self-assessment model appears to work best when
the credit’s child residency requirements, compliance
third-party information is readily available and can be
improved but the number of claims by eligible taxpayers
matched to the applications (or tax returns) automati-
dropped and administrative costs increased. Studies of
cally—and worst in the opposite case. For example, stud-
Medicaid and federal nutrition programs also find that
ies of compliance with the EITC have shown errors to be
participation decreases as the complexity of the applica-
most prevalent for family-related eligibility criteria (for
tion process increases.15 In response to those concerns,
example, the taxpayer’s marital status or whether a child
most states have eliminated face-to-face interviews for
claimed as a dependent actually resides in the taxpayer’s
determining eligibility for families with children under
home), for which the IRS has lacked administrative
Medicaid and SCHIP and have simplified their applica-
data.13 Conversely, rates of misreported income are gen-
tion processes.
erally lower because the IRS receives information on most
earnings from employers. For other administering agen-
Exceptions for Retroactive and Presumptive Eligibility.
cies, which have different monitoring capabilities, the
Another issue concerns the timing of enrollment in a
sources of error may be reversed. For example, misre-
public program. Some people may be eligible to partici-
ported income constitutes the largest source of error in
pate in a public program but do not apply for benefits
the school lunch program; schools generally rely on par-
until the onset of a medical emergency. Some programs
ents’ reports of income to certify eligibility for subsi-
would deem those individuals covered—at least for a
dies.14 Schools may be in a stronger position to monitor a
short period before their application for benefits. Medic-
child’s living arrangements than to monitor his or her
aid coverage, for example, may be retroactive: Once an
family income.
application is approved, the plan may cover costs
incurred in any or all of the three months before
Caseworker Model. This model is typically used to deter-
application.16
mine eligibility for the Supplemental Nutrition Assis-
tance Program (SNAP), which was formerly known as
the Food Stamp program, and for the Temporary Assis-
tance for Needy Families (TANF) program. The case-
15. See Jennifer Stuber and others, Beyond Stigma: What Barriers
Actually Affect the Decisions of Low-Income Families to Enroll in
Medicaid? George Washington University Medical Center Issue
13. Janet Holtzblatt and Janet McCubbin, “Issues Affecting Low-
Brief (July 2000); Janet M. Currie and Jeffrey Grogger, “Explain-
Income Filers,” in Henry J. Aaron and Joel Slemrod, eds., The
ing Recent Declines in Food Stamp Program Participation,” in
Crisis in Tax Administration (Washington, D.C.: Brookings Insti-
William G. Gale and Janet Rothenberg-Pack, eds., Brookings-
tution, 2004), pp. 148–200.
Wharton Papers on Urban Affairs, 2001 (2001), pp. 203–244; and
Marianne P. Bitler, Janet Currie, and John Karl Scholz, “WIC Eli-
14. Michael Ponza and others, NSLP/SBP Access, Participation, Eligi-
gibility and Participation,” Journal of Human Resources, vol. 38
bility, and Certification Study: Erroneous Payments in the NSLP and
(Supplement 2003), pp. 1139–1179.
SBP, vol. 1, Study Findings (Report No. CN-07-APEC, submitted
by Mathematica Policy Research, Inc., to the Department of Agri-
16. The counts of the insured population, however, do not reflect
culture, November 2007).
those who have such “provisional” insurance.
CBO
CHAPTER TWO
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
43
Another way that a public plan may provide coverage to
same time, increases in the share of costs borne by the
those who have not completed the application process is
federal government not only raise federal spending but
through presumptive eligibility. States have the option of
also reduce the incentives that states have to manage
allowing qualified entities—including Medicaid provid-
those funds in a prudent manner.
ers and certain other programs serving low-income
children (for example, Head Start)—to deem that a child
is temporarily eligible for Medicaid; if the child’s parent
Effects of Premium Subsidies on
or guardian does not submit a completed application
Rates and Sources of Insurance
during the time period established by the state, the child’s
Coverage
presumptive eligibility ends. States can also extend pre-
Because of the central role played by employment-based
sumptive eligibility to women who are pregnant or who
insurance, the effects of any proposal to offer new pre-
need treatment for breast or cervical cancer.
mium subsidies or to modify existing ones depend not
only on how individuals respond to those provisions but
Time Between Recertifications. The length of time
also on how firms respond in their decisions about offer-
between certification periods affects participation, cover-
ing coverage to their employees and about subsidizing
age, and program costs. Basing eligibility on current
that coverage. To capture those complex interactions,
resources provides the best measure of ability to pay, but
CBO has developed a microsimulation model to estimate
continuously monitoring eligibility is costly. Beneficiaries
how rates of coverage and sources of insurance would
also may be more likely to drop out of a program if they
change from those currently projected as a result of pro-
are frequently asked to recertify their eligibility. Most
posals that alter the subsidies for—and thus the net cost
states have a 12-month renewal period for SCHIP, which
of—various insurance options. That model is based on
enables children to remain enrolled unless their family
survey data and includes a wide range of information
reports a change in income or other circumstances—
about a representative sample of individuals and families,
something they may be reluctant to do if it would cause
including their income, employment, health status, and
them to lose eligibility. For SNAP, states have the option
health insurance coverage. This section discusses some of
to allow households with earned income to retain the
the key parameters and assumptions that CBO uses to
same benefits for six months even if their income fluctu-
estimate coverage rates and sources of insurance.17
ates over that certification period, as long as their total
income remains below 130 percent of the federal poverty
Employers’ Decisions to Offer Health Insurance
level.
Most nonelderly Americans obtain health insurance
through their employers, but before individuals can
Providing Incentives for States
enroll in such private group coverage, employers must
The effects of any proposal that requires action by the
offer it to their employees. In general, businesses compete
states—and, in particular, funding from them—depends
for workers by offering wage and benefit packages that
on the incentives states have to participate. States cur-
will attract and retain employees. Employers offer health
rently have substantial latitude to expand Medicaid;
insurance (and other benefits) if they believe their
indeed, over half of current Medicaid spending is for cov-
employees want such coverage enough, in effect, to
erage of populations or benefits beyond those required by
trade cash wages for it. Consequently, an employer’s
federal law. States could expand their programs even
response to a policy will be a function of how that policy
more (with approval from the Department of Health and
affects its workforce, on average. Those effects could arise
Human Services), but the fact that they have not pursued
from proposals that would change the subsidies for
substantial further expansions of Medicaid suggests that
employment-based health insurance, but they could also
the current federal matching rate—which averages
stem from changes to employees’ other health insurance
57 percent for medical costs—is not a sufficient
inducement to do so. Indeed, states’ budgetary pressures
sometimes lead them to limit eligibility or benefits, but
17. A more complete description of the model, the key parameters
those changes have generally been modest. Proposals that
and assumptions it uses, and the academic literature on which it is
would expand insurance coverage by, in effect, subsidiz-
based can be found in Congressional Budget Office, CBO’s Health
ing states’ efforts would probably require a higher federal
Insurance Simulation Model: A Technical Description, Background
subsidy rate to generate a substantial response. At the
Paper (October 2007).
CBO
44
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
Table 2-4.
Effects of a Premium Subsidy on Offer Rates for Employment-Based Coverage, by
Size of Firm, 2009
Employees Offered and Eligible for Coverage
Size of Firm
Employees Number
(Number of employees)
(Millions)
(Millions)
Percent
Elasticity of Offera
Fewer Than 25
30.6
14.7
48.0
-1.14
25 to 99
17.4
12.5
72.0
-0.38
100 to 999
26.9
20.8
77.2
-0.15
1,000 or More
62.9
54.0
85.7
-0.07
All
137.8
102.0
74.0
-0.28
Source:
Congressional Budget Office’s health insurance simulation model.
a.
Elasticity of offer is a factor that measures how employers’ offers of health insurance respond to changes in price. It is based on estimates
from several studies: Jonathan Gruber and Michael Lettau, “How Elastic Is the Firm’s Demand for Health Insurance?” Journal of Public
Economics, vol. 88, nos. 7–8 (July 2004), pp. 1273–1293; Jack Hadley and James D. Reschovsky, “Small Firms’ Demand for Health Insur-
ance: The Decision to Offer Insurance,” Inquiry, vol. 39, no. 2 (2002), pp. 118–137; and Len Nichols and Linda J. Blumberg, “Estimating
Employer Elasticity of Demand for Health Insurance” (paper presented at Academy Health Annual Research Meeting, Washington, D.C.,
1999).
options—either in the individual insurance market or
example, for firms with between 25 and 99 employees,
from a public program such as Medicaid.
CBO estimates the elasticity of offer to be -0.38; thus, a
10 percent increase in the premiums for firms of that size
Changes in Subsidies for Employment-Based Insurance.
would cause a 4 percent decline in the number of
For several reasons, large employers are more likely than
employees offered coverage. Consistent with the available
small employers to offer health insurance. Reflecting that
evidence, the relevant “price” in that calculation is the
fact, the response of firms to changes in the subsidies for
total cost of insurance to the employer and employee
employment-based insurance would depend not only on
combined—net of any federal (or state) subsidies—not
the impact of those changes on the net price of insurance
just the portion that the employer pays directly.
but also on the size of their workforce. To estimate the
likelihood that firms would offer (or drop) health insur-
Consider, for example, how firms of different sizes would
ance in response to a change in price, CBO multiplies the
respond to a subsidy proposal that reduced the net price
average change in price for a firm’s employees by an “elas-
of employment-based insurance by 20 percent. CBO
ticity of offer”—a factor that measures how employers’
anticipates that such a subsidy would increase the avail-
offers of insurance respond to changes in price. The elas-
ability of health insurance at very small firms (those with
ticity of offer varies with the size of the firm and is based
fewer than 25 employees) by about 23 percent (the
on estimates from several studies (see Table 2-4).18 For
20 percent reduction in price multiplied by -1.14, the
elasticity of offer). The share of employees at such firms
18. See Jonathan Gruber and Michael Lettau, “How Elastic Is the
that are currently offered insurance is estimated to be
Firm’s Demand for Health Insurance?” Journal of Public Econom-
48 percent, so the proposal would be expected to increase
ics, vol. 88, nos. 7–8 (2004), pp. 1273–1293; Jack Hadley and
that share by 23 percent, to 59 percent, or a gain of
James D. Reschovsky, “Small Firms’ Demand for Health Insur-
3.4 million workers. Larger firms would be less respon-
ance: The Decision to Offer Insurance,” Inquiry, vol. 39, no. 2
sive to the subsidy, and analogous calculations for all
(2002), pp. 118–137; and Len Nichols and Linda J. Blumberg,
firms would yield an overall increase in the offer rate of
“Estimating Employer Elasticity of Demand for Health Insur-
ance” (paper presented at Academy Health Annual Research
about 6 percent (an average elasticity of offer of -0.28
Meeting, Washington, D.C., 1999).
times -20 percent); that translates into an increase of
CBO
CHAPTER TWO
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
45
about 6 million in the number of workers offered cover-
Individuals’ Decisions to Enroll in an
age, roughly half of whom would have been uninsured
Insurance Plan
before receiving the new offer.19
Individuals’ decisions to enroll in a plan are affected by
the price of insurance, their income, and the options
Changes in Employees’ Other Insurance Options.
available to them. In the private market, those options
Employers’ decisions about offering coverage are also
may include enrolling in an employment-based plan (if
affected by changes in the relative attractiveness of their
they work for a firm that offers insurance) and purchasing
employees’ other insurance options, such as individually
insurance in the individual market. Individuals’ decisions
purchased plans or Medicaid coverage. For example,
to enroll in a private plan will also be affected by whether
legislation that expanded eligibility for Medicaid could
they are eligible to participate in a public program.
make some of a firm’s employees (or their dependents)
eligible for that program; alternatively, a proposal could
Purchasing Employment-Based Insurance. After firms
provide a new tax credit for policies that are purchased in
make their decisions about offering and subsidizing
the individual market. Other factors held equal, a firm
insurance—reflecting the average effects of the policy
would be less likely to offer coverage if the relative
proposal on their workers and the aggregate cost of insur-
attractiveness of its employees’ other options increased.20
ance—workers choose whether to enroll and where to
The magnitude of that effect is estimated to be roughly
obtain their coverage. As with employers’ decisions to
one-third as large as the direct effect of changes in the
offer health insurance, the choice to enroll in a plan will
price of employment-based insurance; that is, a proposal
depend on the price of employment-based health insur-
that provided a 10 percent subsidy for policies purchased
ance and alternative options. In this case, however, the
in the individual insurance market would have about the
key factor affecting enrollment rates is the premium that
same effect on employers’ offers as a proposal that
employees themselves pay, not the total cost to the
increased the net price of insurance purchased through an
employer and employee combined. Even though workers
employer by 3 percent to 4 percent. (Both proposals
ultimately “pay” for employers’ contributions toward
would increase the relative attractiveness of individually
their health insurance, primarily through reduced wages,
purchased insurance.)
studies have found that employees’ decisions about
enrollment are not sensitive to the amount the employer
In addition to a firm’s size, the other factors that would
pays. That finding reflects the fact that once an individ-
affect whether employers drop coverage include the
ual has decided to work for an employer that offers insur-
average tax rate its employees face (lower rates mean that
ance, it is generally infeasible for that worker to recoup
employees obtain less of a benefit from the current exclu-
the employer’s contribution by declining coverage.
sion for employment-based insurance) and the relative
value of the alternative insurance. Some firms that con-
Several studies have attempted to estimate employees’
tinue to offer coverage may also change the amount they
responses to changes in the amount they have to pay for
contribute toward premiums in response to changes in
employment-based insurance by comparing the behavior
the attractiveness of outside insurance options. On the
of workers who face lower insurance premiums with that
basis of the limited evidence that is available, CBO
of workers who face higher premiums.22 The results of
estimates that firms would increase the share of the total
premium employees pay by about 2 to 3 percentage
21. See Thomas Buchmueller and others, “The Effect of SCHIP
points if the share of their workers eligible for Medicaid
Expansions on Health Insurance Decisions by Employers,”
increased from 20 percent to 40 percent.21
Inquiry, vol. 42, no. 3 (2005), pp. 218–231; and M. Susan
Marquis, “The Role of the Safety Net in Employer Health Benefit
Decisions,” Medical Care Research and Review, vol. 62, no. 4
(2005), pp. 435–457.
22. See Linda J. Blumberg, Jessica Banthin, and Len Nichols,
19. CBO generally assumes a “linear” relationship between the change
“Worker Decisions to Purchase Health Insurance,” Journal of
in the price of insurance and the elasticity of offer; that is, the elas-
Health Care Finance and Economics, vol. 1, nos. 3–4 (2001),
ticity of offer does not vary with the magnitude of the price
pp. 305–325; and Jonathan Gruber and Ebonya Washington,
change.
“Subsidies to Employee Health Insurance Premiums and the
20. For a more extensive discussion, see Congressional Budget Office,
Health Insurance Market,” Journal of Health Economics, vol. 24,
The State Children’s Health Insurance Program.
no. 2 (2005), pp. 253–276.
CBO
46
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
those studies can be illustrated by examining what would
in the individual market are affected by its price, their
happen to enrollment in employment-based plans as the
income, and the availability of other insurance options.
share of the premium that the employee pays is reduced
from current levels to zero; for simplicity, the illustration
Estimates of the response to changes in the price of indi-
focuses on the 149 million nonelderly individuals,
vidually purchased insurance are most reliable if subsidy
including workers and their dependents, who work for
rates are low, because that situation is similar to current
employers that offer health insurance but who are not
experience. The available studies suggest that a new
eligible for public coverage.
25 percent subsidy for individually purchased coverage
would cause 2 percent to 6 percent of the uninsured pop-
According to a survey of employers conducted by the
ulation to buy that coverage. The academic literature is
Kaiser Family Foundation, employees’ premiums cur-
not very informative, however, when prices are close to
rently average about 16 percent of the cost of single cov-
zero (that is, when subsidies approach 100 percent),
erage and about 27 percent of the cost of family coverage,
because that situation is currently not observed in the
and at those subsidy rates approximately 137 million peo-
individual market. CBO therefore estimates the effects of
ple are covered through an employment-based plan in
high subsidies using evidence about participation rates in
2009. Another 1 million choose to purchase insurance in
existing public programs (which are, in general, highly
the individual market. Of the 11 million who are not
subsidized).
covered, CBO estimates that about 3 million would
obtain coverage if their contribution was cut in half, and
On the basis of that evidence, CBO estimates that people
about 4 million more would do so if that contribution
would gradually become more responsive to changes in
was reduced to zero. In other words, even if they were
the price of individually purchased insurance as subsidy
offered insurance for free, the remaining 4 million—or
rates increased; moreover, they would become increas-
3 percent of the individuals who could be covered
ingly likely to obtain coverage when subsidy rates
through their employer—would decline that coverage
exceeded 70 percent (see Figure 2-2). Of the roughly
and remain uninsured.
45 million nonelderly individuals who do not work for
employers that offer health insurance or who are not eli-
Those enrollment rates are average measures of the
gible for public coverage, about 20 percent are covered by
expected response by employees to a change in the price
an individually purchased policy in 2009; the current
of insurance, but other factors would also play a role. For
subsidy rate is close to zero. Adding a 25 percent subsidy
example, the probability of uninsured individuals’ enroll-
for individually purchased coverage would increase the
ing in their firm’s plan is affected by their income and by
participation rate for that population by 3 percentage
the availability and attractiveness of other coverage
points, to 23 percent, CBO estimates, reducing the unin-
options. Workers with higher income tend to have higher
sured by about 1.4 million people. Increasing the subsidy
enrollment rates than those with lower income, although
rate to 50 percent would roughly double the impact, but
those rates would be expected to converge as subsidies
increasing the participation rates above 50 percent would
approached 100 percent. By contrast, individuals with
require the subsidy rate to exceed 80 percent (holding
access to Medicaid or other public coverage or whose
other factors equal). Such subsidies, moreover, would
children have public coverage would be less likely to
make insurance in the individual market more attractive
enroll in family coverage offered by an employer, and the
relative to employment-based plans, causing some
likelihood would be reduced in proportion to the per-
employers to decide not to offer coverage to their
centage of children covered by a public program.
employees.
Purchasing Insurance in the Individual Market. Proposals
Impact of Eligibility for Public Programs. People who are
could seek to expand coverage through the individual
eligible to participate in public programs, such as Medic-
insurance market—for example, by equalizing the tax
aid and SCHIP, will also be affected by proposals that
treatment of employment-based and individually pur-
lower the costs of employment-based or individually pur-
chased coverage or by subsidizing individually purchased
chased plans. Their response will be influenced by many
insurance through new tax credits and tax deductions. As
of the same factors that affect the decisions of those who
with their choices regarding employment-based coverage,
are not eligible for public programs to purchase health
individuals’ decisions about whether to purchase coverage
insurance in the private market—the price of health
CBO
CHAPTER TWO
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
47
Figure 2-2.
applying for a public program until the onset of a medical
Probability of Enrolling in an
emergency.
Individually Purchased Insurance
Those factors may help explain why many people who are
Plan with a Subsidy
eligible for Medicaid or SCHIP purchase coverage in the
private market. In 2009, about 23 percent of nonelderly
(Percentage taking subsidy)
people who are eligible for Medicaid or SCHIP will have
employment-based insurance coverage and about 2 per-
100
cent will have individually purchased coverage. Reflecting
that evidence, CBO models the purchase of private cover-
80
age among people who are eligible for or enrolled in
Medicaid in a manner that is similar to enrollment rates
60
among the uninsured but is reduced by a factor (generally
more than half ) to reflect that those people would be giv-
ing up free or low-cost public coverage to take private
40
coverage.
20
Decisions to Switch or Drop Coverage
Changes in the relative attractiveness of health insurance
0
options could affect whether individuals who already
0
10
20
30
40
50
60
70
80
90
100
have insurance would remain in their current plans. A
proposal to expand coverage could cause some people to
Subsidy (Percent)
switch from one type of plan to another and still others to
Source:
Congressional Budget Office’s health insurance simulation
become uninsured, even though the proposal would, on
model.
net, result in an increase in coverage.
Note: Roughly 45 million nonelderly U.S. residents lack access to
an offer of employment-based insurance and are ineligible
A proposal, for example, might increase the attractiveness
for Medicaid or the State Children’s Health Insurance Pro-
of individually purchased coverage through a deduction
gram. The curve displays the estimated probability of that
or tax credit for health insurance premiums. Because their
population enrolling in an individually purchased insurance
plan as a function of a subsidy set at a percentage of the
employees could receive tax benefits for purchasing
insurance premium.
health insurance elsewhere, some employers would drop
existing plans. In response, some of their workers would
insurance, their income, and the availability of other
obtain coverage in the individual market; others might
insurance options.
(if able) obtain employment-based insurance through a
spouse’s job, enroll (if eligible) in Medicaid, or become
In addition, other nonmonetary factors appear to affect
uninsured. A number of workers at firms that continued
whether an individual who is eligible for a public plan
to provide coverage might switch to an individually pur-
purchases private insurance—and, more generally,
chased plan. Workers in relatively good health would be
whether that person instead enrolls in a public plan or
the most likely to prefer individually purchased insurance
does not obtain any coverage. Applying for a public pro-
because they might pay lower premiums in that market
gram may impose burdens that people do not encounter
(in which premiums usually vary to reflect enrollees’ age
when they apply for private health insurance. Those bur-
and health status) than in the employment-based insur-
dens include learning about the program and its eligibil-
ance market (in which healthier workers are pooled with
ity rules, the additional expenses incurred when applying
less healthy coworkers).
for benefits (for example, the amount of time it takes to
complete an application, collect documentation proving
Suppose, for example, that a subsidy for individually pur-
eligibility, and—if required—visit a government office
chased insurance was provided to low-income individuals
for an interview); and concerns that enrolling in a public
and families in the amounts of $1,500 for single coverage
program will somehow stigmatize the participant. Retro-
and $3,000 for a family policy. Under that scenario,
active and presumptive eligibility rules may also affect the
CBO estimates that about 2.3 million people who would
timing of a decision to enroll, by allowing people to delay
have been uninsured would instead use the voucher to
CBO
48
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
obtain coverage. In addition, 100,000 people who would
workers responded to the subsidies (and to the ensuing
have otherwise had employment-based insurance would
offsetting changes in other aspects of their compensa-
be covered in the individual market; roughly half of them
tion). As a result, determining the degree of employers’
would no longer be offered insurance by their employer
compliance with mandates is only the first step in esti-
as a result of the new policy, and half of them would have
mating the impact of the mandate on rates of insurance
individually purchased coverage despite being offered
coverage.
coverage by their employer. About 100,000 people would
become uninsured because their employer elected not to
Experience with Mandates
offer coverage and they neither purchased an individual
The federal government does not have any experience
plan nor enrolled in a public program.23
administering a health insurance mandate, but Hawaii
and Massachusetts, which have enacted such require-
Individual and Employer Mandates
ments, provide some insights. In addition, it is useful to
examine the impact of other mandates that may be
Premium subsidies would increase the number of people
analogous, such as requirements to pay taxes, purchase
with insurance but would not be sufficient to achieve
automobile insurance, and vaccinate children.
universal coverage, even if the subsidies covered a very
large share of policy premiums. To increase coverage rates
Existing State Mandates for Health Insurance. Since
further, policymakers could impose a mandate that indi-
1974, Hawaii has required employers to provide health
viduals obtain insurance or that employers offer coverage.
insurance for their full-time, permanent employees.
Individual mandates can be applied broadly (to the entire
Employers also must pay at least half of the premium for
population of the United States) or to a specific group
single coverage for each eligible employee (with the
(for example, children). Employer mandates target a
exception of part-time employees, government employ-
more specific subpopulation of the uninsured: workers
ees, and seasonal workers).
without insurance and, depending on the scope of the
mandate, their spouses and dependents. Under an
In 2007, Massachusetts began phasing in a mandate that
employer mandate, firms could be required to provide
all adult residents have health insurance. By 2009, all
health insurance for their workers or contribute to a
adults must have a policy that covers a basic set of bene-
fund—an approach commonly referred to as “play or
fits or they will have to pay a penalty—for each month
pay.” If effective penalties were imposed on individuals
not covered—equal to half the premium for the cheapest
and firms that did not comply, mandates would increase
health insurance plan that provides the minimum benefit
coverage rates by raising the cost of remaining uninsured
package. Penalties are waived, however, for people who
or of not offering coverage—that is, by using a stick
are deemed unable to afford insurance. Massachusetts
rather than a carrot. However, the ultimate impact of a
also requires all but the smallest employers to provide
mandate would depend largely on its scope, the extent of
health insurance for their workers or pay $295 per year
its enforcement provisions, and the resulting incentives to
for each uncovered worker to a state fund. (For additional
comply.
details about the mandates in each state, see Box 2-3.)
An important distinction between the two types of man-
Several studies have examined the effects of the employer
dates is that individual mandates generally require people
mandate in Hawaii. Although Hawaii has relatively high
to have insurance coverage, whereas employer mandates
rates of insurance coverage compared with other states,
generally require employers to offer coverage. Employer
determining how much of that difference is due to the
mandates could also require that employers subsidize a
mandate itself and how much reflects other factors, such
certain percentage of the premium, which would encour-
as population characteristics, is a challenge. One study
age their workers to purchase coverage, but the ultimate
found that the number of uninsured fell by roughly
effect on coverage rates would depend on how those
1 percentage point—from about 11 percent to about
10 percent—after Hawaii implemented the mandate,
23. For further discussion of that option, see Congressional Budget
attributing that relatively small effect to the exclusion of
Office, Budget Options, Volume 1: Health Care (December 2008).
dependents and certain classes of workers as well as
CBO
CHAPTER TWO
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
49
limited enforcement.24 Other studies have examined the
legal requirements—which may reflect the intrinsic
effects of the mandate on exempted sectors, particularly
value of a mandated item (such as auto insurance), subsi-
part-time workers. Although part-time employment in
dies (such as free or low-cost vaccinations), or social
Hawaii increased after the mandate relative to part-time
norms (such as a perceived obligation to pay taxes). For
employment in the rest of the country, the evidence is
example, studies indicate that about 77 percent of young
mixed regarding the mandate’s effects on rates of insur-
children were vaccinated against chicken pox in states
ance coverage in the exempt sectors.25
without immunization requirements and that mandates
increased vaccination rates by about 8 percentage points
Although it is still too soon to evaluate the full effect of
(to 85 percent).
the mandates in Massachusetts, preliminary data suggest
that the number of uninsured has fallen substantially
Factors Affecting Compliance with Mandates
since the state implemented its health reform plan.
The observed variation in compliance with existing man-
According to one recent study, the share of nonelderly
dates reflects the fact that their impact depends largely on
adults who were uninsured declined from 13 percent in
their scope and on the expected costs of noncompliance
the fall of 2006 to 7 percent in the fall of 2007.26 How-
for people who are subject to them. In the case of a man-
ever, the study did not attempt to isolate the effects of the
date to have health insurance, individuals would generally
coverage mandates from other aspects of the state’s initia-
weigh the benefits of that coverage against those expected
tive, which included reforms in the insurance market and
costs when determining whether to comply.
new premium subsidies for lower-income individuals and
families.
Scope of a Mandate. An individual mandate could apply
to the entire population or be limited to certain subpopu-
Other Types of Mandates. Although the U.S. experience
lations. Requiring all residents to have health insurance
with enforcing health insurance mandates is limited,
coverage would encompass the largest number of unin-
some lessons can be drawn from other types of mandates.
sured but could be more difficult to implement quickly
For example, both federal and state governments require
than a more narrowly targeted mandate. For example, it
individuals and businesses to pay taxes; many states have
might take government agencies longer to develop the
imposed mandates for drivers to have auto insurance and
administrative apparatus necessary to enforce mandates
to wear seat belts; and many school districts require chil-
covering the entire population. Alternatively, a mandate
dren to be immunized in order to attend public schools.
could be limited, at least initially, to adults, children, or
For those mandates, national compliance rates range
another subpopulation; that approach would allow
from 63 percent to 86 percent (see Table 2-5). Those
insurers, providers, and government administrators to
rates, however, do not clearly identify the effect of the
adapt more gradually to new responsibilities but could
mandate itself because they include people who might
potentially reduce or slow the impact on the uninsured
have acted in the desired manner even if there were no
population.
Proposals could also limit the scope of individual man-
24. Andrew Dick, “Will Employer Mandates Really Work? Another
Look at Hawaii,” Health Affairs, vol. 13, no. 1 (1994),
dates by allowing for certain exemptions. In particular,
pp. 343–349.
individuals who are not able to afford health insurance
could be exempted from the mandate. In Massachusetts,
25. See Norman Thurston, “Labor Market Effects of Hawaii’s Manda-
tory Employer-Provided Health Insurance,” Industrial and Labor
for example, uninsured individuals whose income is
Relations Review, vol. 51, no. 1 (1997), pp. 117–135; and Sang-
below 150 percent of the federal poverty level are not
Hyop Lee and others, The Effect of Mandatory Employer-Sponsored
penalized if they fail to obtain health insurance.27
Insurance (ESI) on Health Insurance Coverage and Labor Force Uti-
lization in Hawaii: Evidence from the Current Population Survey
(CPS) 1994–2004, Working Paper No. 05-12 (Honolulu, Hawaii:
University of Hawaii at Manoa, Department of Economics, July
27. Massachusetts’s affordability standards use information from tax
6, 2005), www.economics.hawaii.edu/research/workingpapers/
returns to verify eligibility for income-related exemptions and also
WP_05-12.pdf.
take into account the cost of premiums. In 2008, individuals with
26. Sharon K. Long, “On the Road to Universal Coverage: Impacts of
annual income of up to $52,500 were exempt from the mandate if
Reform in Massachusetts at One Year,” Health Affairs, vol. 27,
they could not find health insurance with a premium of less than
no. 4 (July/August 2008), pp. W270–W284 (published online as
$330 a month. For families, the corresponding income and pre-
a Web Exclusive, June 3, 2008).
mium amounts were, respectively, $110,000 and $792.
CBO
50
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
Box 2-3.
Health Insurance Mandates in Hawaii and Massachusetts
Both Hawaii and Massachusetts have enacted health
Employers must generally pay at least half of the pre-
insurance mandates. In Hawaii, the mandate is
mium for individual coverage, subject to limits on
imposed on employers, requiring them to offer cover-
how much the employee must pay. Employers may
age to most of their workers and to subsidize the pre-
demonstrate compliance with the mandate by pur-
mium; in Massachusetts, both adults and employers
chasing an approved plan from an insurer or by
are subject to mandates.
obtaining certification for their own health care
plan.1 The minimum penalty for failure to comply
Hawaii
with the employer mandate is $25, or $1 per
The state’s Prepaid Health Care system, as it is
employee for each day of noncompliance, whichever
known, was enacted in 1974 and requires all employ-
is greater. The business may be closed if the employer
ers in Hawaii to offer employees a health insurance
fails to provide health insurance for more than 30
plan that covers at least a minimum set of specified
days. Additional penalties are imposed for willful
benefits. Most employers with at least one employee
noncompliance. The mandate is administered by the
are subject to the requirement, but they do not have
state’s Department of Labor and Industrial Relations.
to offer coverage to employees who work less than 20
hours a week or whose monthly wages fall below a
1. Shortly after Hawaii enacted the Prepaid Health Care Act,
specified amount (about $600 in 2008). In addition,
the Congress passed the Employee Retirement Income Secu-
the mandate can be waived for employees covered by
rity Act (ERISA), which exempts many employers from state
Medicare, Medicaid, or their own (or their parents’)
health insurance mandates and other regulations. (For a dis-
health plan and for members of certain religious
cussion of ERISA, see Box 1-1.) The Supreme Court ruled in
1981 that ERISA preempted Hawaii’s legislation, but in
groups.
1983 the Congress granted Hawaii a waiver from the ERISA
provisions.
Establishing such exemptions would involve defining a
on the basis of the number of full-time permanent
standard for affordability—and, depending on the
employees, although the administering agency might find
definition, such provisions could vary in their ease of
it difficult to determine who meets those qualifications
administration. Massachusetts also allows exemptions for
without on-site visits (an administratively burdensome
religious beliefs and personal hardships; verifying whether
and costly activity). Depending on the definition used,
hardship exemptions are valid would require developing
businesses might find it advantageous to reorganize to
review and appeals procedures, adding to administrative
avoid mandates or to qualify for subsidies on the basis of
costs.
the number of employees.
Exemptions and subsidies could reduce the costs of com-
Expected Penalty for Noncompliance. An individual’s
plying with an employer mandate for certain types of
decision to purchase health insurance or an employer’s
employers, such as small businesses. Massachusetts
choice to offer health insurance will largely depend on the
exempts employers with 10 or fewer employees from its
costs of policies and the availability of subsidized alterna-
play-or-pay requirements. In addition, employers that
tives (such as Medicaid). Mandates could affect those
have 50 or fewer employees receive subsidies for each
decisions by imposing penalties on individuals who
employee whose family income is at or below 300 percent
remain uninsured or on employers who fail to offer
of the federal poverty level. Definitions of a “small”
health insurance.
business vary, however, in terms of the threshold for
determining who counts as an employee and the number
As is the case with other legal requirements, individuals
of employees involved. A small business could be defined
and employers will generally comply with mandates when
CBO
CHAPTER TWO
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
51
Box 2-3.
Continued
Health Insurance Mandates in Hawaii and Massachusetts
Massachusetts
tered by the state’s Department of Revenue; residents
As a result of a law enacted in 2006, every Massachu-
receive a form from their insurer that they must
setts resident who is age 18 or older is now required
submit along with their tax return to verify that they
to have health insurance. (New subsidies were also
have qualified coverage, and insurers and state agen-
offered to encourage families to obtain coverage, but
cies also report health insurance coverage to that
insurance is not mandatory for children.) The state’s
agency.
individual mandate is being phased in over three
years. By the end of 2007, individuals had to have
Employers with 11 or more full-time-equivalent
health insurance or they would be subject to income
employees must offer health insurance to their work-
tax penalties. In 2008, they must have coverage each
ers and must also establish a “cafeteria plan” that
month. In 2009, they must have coverage under a
allows employees to exclude payments for their por-
health plan containing specified benefits.
tion of the premium from income and payroll taxes.
The state has also established an insurance “connec-
Beginning in 2008, the penalty for noncompliance
tor” agency through which small businesses (as well
equals half of the premium for the minimum afford-
as individuals) may purchase qualified insurance
able health insurance plan (a fixed amount deter-
plans. In addition, employers must either make a “fair
mined by Massachusetts officials at the beginning of
and reasonable contribution” to their employees’
each year) for each month not covered. Subsidies are
health insurance or pay the state up to $295 per
provided to lower-income households to help them
employee. Employers are also subject to a surcharge if
pay for insurance, and individuals with income under
employees and their dependents use more than
150 percent of the federal poverty level are exempted
$50,000 of care a year from a state-funded pool that
from the mandate (and those with somewhat higher
has been established to finance free care for the unin-
income may also be exempt if they cannot find
sured. The amount of the surcharge varies with such
affordable insurance). Individuals may also apply for
factors as the number of employees and the amount
exemptions on the basis of religious beliefs or per-
of free care used.
sonal hardship. The individual mandate is adminis-
the costs of noncompliance exceed the net costs of com-
lowest-cost qualifying plan. Although that penalty repre-
pliance. Thus, a mandate can become more effective in
sents a significant portion of the insurance premium, it is
expanding coverage by either increasing the penalties for
still less than the cost of paying for coverage—at least for
noncompliance or raising the probability of being
higher-income individuals who do not receive premium
“caught”—either of which would raise the expected value
subsidies—and individuals who place a small value on
of the penalty.
insurance coverage may not be induced to purchase it.
For such individuals, the lower-cost option is to simply
Health insurance mandates differ from many other
pay the penalty. For others, however, the value of the
requirements, such as payment of taxes, because individu-
insurance coverage gained plus the penalty avoided by
als receive a tangible item that has some value in exchange
complying may be sufficient to induce enrollment.
for their compliance. Thus, individuals’ responsiveness to
mandates would depend, in part, on the amount of the
The degree to which individuals who are subject to a
penalty relative to the net value of insurance. For exam-
mandate believe that their noncompliance would be
ple, the penalty for noncompliance with the individual
detected, and that fines would be levied as a result, also
mandate in Massachusetts is equal to half the cost of the
greatly affects a mandate’s impact on coverage. The
CBO
52
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
Table 2-5.
U.S. Experience with Enforcing Mandates
Administering
Availability of
Methods of
Mandate
Agency
Data
Verification
Penalties
Evidence of Compliance
Individual Mandate
Immunization
School districts
Parents must
School officials
Exclusion from
Immunization rates for chicken
provide proof of
review
schools.
pox among children ages 19 to
immunization.
documentation
35 months range from
provided by
76.8 percent in states without
parents.
mandates to 84.9 percent in
states with mandates.a
Auto Insurance
State
Self-reporting;
Vary by states.
Vary by states,
85.4 percent of drivers were
Departments of
some states
ranging from
insured in 2004, but rates
Motor Vehicles
receive
none to $50 to
range from 74 percent in
information
$5,000.
Mississippi to 96 percent in
from insurers.
Maine;b compliance is better in
states with third- party
reporting.c
Seat Belts
State and local
None.
Vary by states;
Citations, with
Seat belt use was 82 percent in
law
26 states and
cash penalties.
2007, but rates vary by states
enforcement
the District of
with scope and enforcement,
agencies
Columbia allow
ranging from 63.8 percent in
police to stop
New Hampshire (where adults
cars solely to
are not required to wear seat
check if seat
belts) to 97.6 percent in
belts are worn.
Hawaii.d
Income Taxes
Internal
Self-reporting;
Matching;
Back amount due
86.3 percent of tax liabilities
Revenue
third-party data
audits.
plus interest; civil
was collected in tax year
Service
for many (but
and criminal
2001.e
not all) types of
penalties.
income.
Continued
probability of detection is higher when enforcement
reflecting differences in the IRS’s ability to detect report-
agencies have access to accurate and timely information
ing errors. Tax compliance is relatively high when the
regarding individuals’ coverage status and have sufficient
agency can match information in reports from third par-
resources to use the information collected to identify
ties (such as employers and financial institutions) to
noncompliance.
income tax returns and send notices to taxpayers when
discrepancies are found. The net misreporting rate for
Evidence from the individual income tax system illus-
income subject to third-party reporting was about 5 per-
trates how compliance can vary with the likelihood of
cent in 2001 (see Table 2-6). In contrast, the IRS in
detection and enforcement. Taxpayers are generally sub-
many cases cannot verify individuals’ reports of their
ject to the same penalties for misreported income and
self-employment income (including net income from
deductions, regardless of the source of the error. However,
both nonfarm proprietors and farms) on tax returns
compliance rates vary substantially across income types,
because third-party data are not independently reported
CBO
CHAPTER TWO
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
53
Table 2-5.
Continued
U.S. Experience with Enforcing Mandates
Administering
Availability of
Methods of
Mandate
Agency
Data
Verification
Penalties
Evidence of Compliance
Employer Mandate
Minimum Wage
Department of
Complaints filed
DOL’s
Back pay; criminal
63 percent to 75 percent in
Labor (DOL)
by workers.
investigations
prosecution and
1973.f
and employees’
fines if there is
lawsuits.
evidence of
egregious abuse.
Source:
Congressional Budget Office based on data from sources listed below.
a.
Matthew M. Davis and Michael A. Gaglia, “Associations of Daycare and School Entry Vaccination Requirements with Varicella Immuniza-
tion Rates,” Vaccine, vol. 23, no. 23 (2005), pp. 3053–3060.
b.
Insurance Research Council, Uninsured Motorists, 2006 Edition (Malvern, Pa.: Insurance Research Council, June 2006). The IRC com-
putes the uninsurance rate as the probability that an at-fault driver in an accident was uninsured or unable to meet the liability for some-
one else’s injury caused by the accident. The estimate is derived from insurance claims for 11 insurers (representing 58 percent of
premiums for private passenger auto liability insurance in the United States). The insurance rate is equal to one minus the uninsurance
rate.
c.
Yu-Luen Ma and Joan T. Schmit, “Factors Affecting the Relative Incidence of Uninsured Motorists Claims,” Journal of Risk and Insurance,
vol. 67, no. 2 (2000), pp. 281–294.
d.
Department of Transportation, National Highway Transportation Safety Administration, National Center for Statistics and Analysis, Seat
Belt Use in 2007—Use Rates in the States and Territories, DOT HS 810 949 (May 2008).
e.
Internal Revenue Service, Reducing the Federal Tax Gap: A Report on Improving Voluntary Compliance, IR-2007-137 (August 2, 2007).
f.
Orley Ashenfelter and Robert S. Smith, “Compliance with the Minimum Wage Law,” Journal of Political Economy, vol. 87, no. 2 (1979),
pp. 333–350.
to the IRS and resources for audits are limited.28 Largely
are correct, and individuals may be more vigorous in
as a consequence, the net misreporting rate for self-
defending themselves against the charge of noncompli-
employment income and other forms of income that are
ance so that they can avoid the penalty.30
not subject to third-party reporting exceeded 50 percent
in 2001.
Personal Values and Social Norms. Economic models
that focus solely on monetary benefits and costs would
Although compliance generally improves as penalties
probably overstate noncompliance with a mandate. In a
increase, there may be diminishing returns beyond a cer-
number of areas, program administrators and the judicial
tain point. As penalties increase, some individuals may
system lack complete information to monitor every
respond by taking more aggressive action to avoid detec-
individual’s compliance with government rules and regu-
tion; at the same time, program administrators may be
lations. Yet some compliance is generally observed, even
reluctant to impose penalties that seem excessive.29
when there is little or no enforcement of mandates.
Another issue is that penalties are not costless to impose.
If penalties are increased, administering agencies may
Compliance, then, is probably affected by an individual’s
devote more resources to ensure that their determinations
personal values and by social norms. Many individuals
and employers would comply with a mandate, even in the
absence of penalties, because they believe in abiding by
28. For individual income tax returns, the audit rate is about 1 per-
the nation’s laws. However, such compliance may also be
cent, with audit rates somewhat higher for returns reporting busi-
ness income.
30. Janet G. McCubbin, Optimal Tax Enforcement: A Review of the
29. James Andreoni, “Criminal Deterrence in the Reduced Form: A
Literature and Practical Implications, Working Paper 90 (Washing-
New Perspective on Ehrlich’s Seminal Study,” Economic Inquiry,
ton D.C.: Department of the Treasury, Office of Tax Analysis,
vol. 33, no. 3 (July 1995), pp. 476–483.
December 2004).
CBO
54
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
Table 2-6.
Impact of Third-Party Data and Enforcement Methods on
Income Tax Compliance, 2001
Net Misreporting Ratea
Income or Tax Preference
Third-Party Data
Enforcement
(Percent)
Wages and salaries
Employers file W-2s with IRS.
W-2s are matched to tax
1
returns, generating notices to
taxpayers when discrepancies
are detected.
Interest, dividends, Social
Payers file 1099s with IRS.
1099s are matched to tax
5
Security benefits, pensions,
returns, generating notices to
and unemployment
taxpayers when discrepancies
compensation
are detected.
Partnerships, S corporations,
Some, but not complete,
Some matching, but largely
9
deductions, exemptions, capital
information from third parties.
audits.
gains, and alimony
Nonfarm proprietor income,
None.
Audits.
54
farm income, other income,
rents and royalties, Form 4797
income, and adjustments
Source:
Congressional Budget Office based on data from Internal Revenue Service, Reducing the Federal Tax Gap: A Report on Improving
Voluntary Compliance, IR-2007-137 (August 2, 2007).
Note: IRS = Internal Revenue Service.
a.
The net misreporting rate is the ratio of the amount misreported (including errors in both the taxpayers’ and government’s favor) to the
amount that taxpayers should have reported. The estimates do not include amounts that should have been reported by individuals who
failed to file an individual income tax return.
moderated by perceptions of fairness; individuals may
Finally, a small number of individuals may refuse to com-
comply more readily if they believe that a mandate is fair
ply with mandates on religious or philosophical grounds.
and is consistently enforced. If enforcement efforts
Thus, even in the face of strict penalties and rigorous
appear to be unevenly applied, compliance may dimin-
detection and enforcement mechanisms, it is unlikely
ish.31 Social psychologists find that compliance could be
that everyone targeted by mandates would comply.
affected not only by personal values but also by individu-
als’ perceptions of how others will act. Such studies find
that many people want to take the popular—as well as
Automatic Enrollment Provisions
the moral—course of action.32
Some experts have suggested that proposals to facilitate
automatic enrollment in health insurance plans could
achieve coverage goals similar to those of a mandate but
31. Studies find conflicting evidence regarding the impact of
without requiring a complicated administrative system.
perceptions of fairness on tax compliance. In one experiment,
Under those proposals, individuals would be automati-
participants were more likely to engage in tax evasion when told
that their tax burden was high relative to that of others; compli-
cally enrolled in insurance plans for which they qualify,
ance improved if, instead, they were told that their tax burden was
but they could “opt out” if they chose to refuse coverage.
relatively small. See Michael W. Spicer and Lee A. Becker, “Fiscal
A person could begin receiving health insurance benefits
Inequity and Tax Evasion: An Experimental Approach,” National
without completing any additional forms, simply by vir-
Tax Journal, vol. 33, no. 2 (June 1980), pp. 171–175. However, a
second study found that perceptions of fairness had no impact on
compliance. See Paul Webley and others, Tax Evasion: An Experi-
32. Robert B. Cialdini, “Crafting Normative Messages to Protect the
mental Approach (Cambridge, England: Cambridge University
Environment,” Current Directions in Psychological Science, vol. 12,
Press, 1991).
no. 4 (2003), pp. 105–109.
CBO
CHAPTER TWO
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
55
tue of his or her participation in another public program
Until 1996, recipients of Aid to Families with Dependent
or employment with a firm offering coverage. People who
Children were automatically enrolled in Medicaid and
chose to opt out of insurance coverage would be required
the Food Stamp program. After the enactment of welfare
to initiate action and complete the necessary forms.
legislation, which replaced AFDC with a new program
called Temporary Assistance for Needy Families, some
Evidence of Effects
former AFDC recipients did not qualify for TANF but
Some evidence indicates that provisions for automatic
still retained eligibility for Medicaid and the Food Stamp
enrollment encourage higher participation in public pro-
program. Nonetheless, participation in the two programs
grams (Medicare and welfare programs, for example) and
fell among those families, and studies largely attribute
in employment-based benefits such as 401(k) plans.
that effect to the fact that the families were no longer
enrolled automatically.33
Public Programs. Most eligible individuals are enrolled
automatically in Part B of Medicare (which covers physi-
Employment-Based Plans. Other recent studies of auto-
cian and outpatient services) when they turn 65;
matic enrollment have focused on decisions about saving,
although they have the option of declining that coverage,
especially the choice to participate in 401(k) plans offered
they must send in a form to do so. Enrollees also receive
by employers. Those plans allow workers to defer pay-
a substantial premium subsidy—covering about 75 per-
ment of income taxes on money they save for retirement
cent of average program costs—and face a penalty for late
(and on the interest accumulated on those savings); in
enrollment, both of which encourage prompt sign-up for
many cases, employers also contribute to their employees’
Part B. Nearly 95 percent of those eligible are enrolled in
accounts, so workers who do not participate are leaving
Part B, and many of those who are eligible but are not
some money “on the table.” Until recently, the vast
enrolled have other coverage that substantially reduces
majority of employers who offered 401(k) plans required
the benefits of enrolling in Part B.
their workers to take action to enroll; that is, the default
option was to not participate in the plan. With such
By contrast, the new Medicare prescription drug benefit
“opt-in” arrangements, participation in 401(k) plans
(known as Part D) features a similar premium subsidy
tends to be low for newly hired workers, although it
and late-enrollment penalty but does not feature auto-
increases substantially with tenure at a firm.
matic enrollment (except in the case of some low-income
enrollees). Overall, roughly 70 percent of eligible individ-
Several case studies have found that participation in
uals are now enrolled in the Medicare drug benefit pro-
401(k) plans increases substantially, particularly among
gram in some way, including those receiving subsidized
new hires, as a consequence of automatic enrollment—
coverage through a former employer. Although many
that is, when the default case is to participate in the plan,
individuals who are eligible for Part D have not enrolled
with the choice to opt out. A study of one Fortune 500
because they have other qualified drug coverage, recent
company found that participation in a 401(k) plan
estimates indicate that about 10 percent of the Medicare
increased from 37 percent to 86 percent among employ-
population has not enrolled and does not have other
ees with 3 to 15 months of tenure with the firm after the
insurance for their drugs, even though it would generally
plan switched to automatic enrollment. The effects of
be financially beneficial for them to enroll.
automatic enrollment are greatest among workers who
are least likely to participate under opt-in systems,
Although the differences in participation rates between
including younger and lower-income workers.34 Three
Part B and Part D suggest that automatic enrollment
strategies increase participation, other differences
between the two programs may also have played a role.
33. Leighton Ku and Bowen Garrett, How Welfare Reform and Eco-
nomic Factors Affected Medicaid Participation: 1984–96, Assessing
Both programs provide relatively high premium subsidies
the New Federalism Discussion Papers 00-01 (Washington, D.C.:
and conduct extensive outreach campaigns; automatic
Urban Institute, February 2000); Currie and Grogger, “Explain-
enrollment could have a larger effect if combined with
ing Recent Declines in Food Stamp Program Participation.”
less generous subsidies or less extensive outreach.
34. Brigitte C. Madrian and Dennis F. Shea, “The Power of Sugges-
tion: Inertia in 401(k) Participation and Savings Behavior,”
Another example of the impact of automatic enrollment
Quarterly Journal of Economics, vol. 116, no. 4 (November 2001),
stems from recent changes to federal welfare programs.
pp. 1149–1187.
CBO
56
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
years later, participation in plans with automatic
benefits surrounding the two decisions are not identical.
enrollment was higher than in plans requiring active
The benefits of saving for retirement may seem remote,
enrollment; about half of the participants who were
particularly for younger or newly hired workers, but con-
automatically enrolled also continued to accept both the
tributions can be withdrawn at any time if the worker is
default contribution rate and the choice of investment
willing to pay income taxes and a 10 percent penalty on
fund.35 Reflecting those findings, many firms have
the withdrawal. For younger workers in good health,
switched to automatic enrollment for their 401(k) plans
having health insurance may not provide immediate
in the past few years.
financial returns, but enrollees gain protection against
unpredictable risks right away. At the same time, pre-
Implications for Enrollment Rates
mium payments (or forgone wages) cannot be recouped
The results from public and private programs suggest that
at a later date by withdrawing past contributions, so
inertia is a factor in decisionmaking and that the framing
enrollees might see their payments as wasted if they do
of choices affects behavior—aspects that have been a
not end up needing expensive care.
focus of recent research on behavioral economics. Many
individuals are likely to remain with the default option,
Options for Implementing Automatic Enrollment
even when they can choose a better option with little
As with mandates, the effectiveness of an automatic
effort. The studies of retirement savings also suggest rea-
enrollment provision would depend partly on its scope.
sons for such inertia. It may be a response to complexity;
Employers, for example, could be encouraged or required
that is, faced with complicated choices, people may opt
to provide a “default” health insurance plan for their
for the path of least resistance. Or they may procrastinate
employees if they offer insurance. Employees’ shares of
if the costs associated with a decision are certain and
premiums for a default plan, as with other employment-
immediate and the benefits do not accrue for some time
based plans, would be collected through payroll deduc-
or depend on the outcome of an uncertain event. People
tions. However, unless employers were also required to
may also stay with the default option because they view
provide health insurance, the effects of an automatic
the choice as “endorsed” by their employer or the firm’s
enrollment provision would extend only to uninsured
benefit advisers.36
workers at firms offering insurance. On the basis of
recent studies, CBO estimates that about 30 percent of
Decisions regarding health insurance share some features
the uninsured are workers (and their dependents) who
with choices about retirement saving. Retirement plans
turn down insurance offered by their employers (see
may vary considerably in the riskiness of their invest-
Chapter 1). Whether or not they would simply opt out
ments, their expected returns, and their fees. Similarly,
again if automatically enrolled would depend heavily on
health insurance plans may differ in their selection of
what premiums they would have to pay (net of any subsi-
benefits, provider networks, and costs. In both cases, an
dies). If automatic enrollment requirements increased
individual’s decision to participate involves trade-offs that
participation in employment-based health plans, wages or
may be difficult to weigh.
employers’ contributions toward health insurance premi-
ums might be lower than they otherwise would be, so
The two decisions differ, however, in some important
that the total amount of compensation paid to workers
ways. For example, the timing and nature of the costs and
would not change as a result.
Another approach might be to automatically enroll indi-
35. James Choi and others, “For Better or For Worse: Default Effects
and 401(k) Savings Behavior,” in David A. Wise, ed., Perspectives
viduals in health plans if they participate in any public
in the Economics of Aging (Chicago.: University of Chicago Press,
assistance program. Under current law, individuals receiv-
2004).
ing Supplemental Security Income are automatically
36. John Beshears and others, The Importance of Default Options for
enrolled in Medicaid in most states. Similarly, individuals
Retirement Saving Outcomes: Evidence from the United States,
who are dually eligible for full benefits under the Medi-
Working Paper No. 12009 (Cambridge, Mass.: National Bureau
care and Medicaid programs are enrolled by default in
of Economic Research, January 2006).
CBO
CHAPTER TWO
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
57
one of the lower-cost drug plans available to them (for
the number of individuals who are uninsured and partici-
which they pay no premiums) if they do not select a plan.
pating in those programs. In contexts other than the
Automatic enrollment could be extended to participants
workplace or public programs, automatic enrollment
in other needs-based programs such as TANF or SNAP
provisions might be more difficult to implement—
or to programs like unemployment compensation. The
depending on the complexities involved in determining
effects of such a strategy would depend, in large part, on
eligibility, collecting premium payments, or other factors.
CBO
C H A P T E R
3
Factors Affecting Insurance Premiums
The effect of proposals to increase coverage would who purchase individual coverage have lower expected
depend in part on the premiums charged and the value of
costs for health care to begin with. In other words, if both
the coverage provided. In particular, the costs of a subsidy
employment-based and individually purchased policies
that covers a specified percentage of policy premiums
covered the same enrollees, the difference in their
would be affected by the amount of those premiums,
premiums would be smaller.
whereas the impact of a fixed-dollar subsidy on coverage
rates would depend on the share of the premiums it
Those premiums could change under proposals that
covers. Thus, the factors that determine premiums also
would modify the health insurance market or extend cov-
affect the impact that a proposal has on insurance cover-
erage to individuals who are currently uninsured. Some
age and the federal budget.
proposals could affect premiums by requiring that indi-
viduals enroll in plans that meet certain design specifica-
In general, the premium charged for a private health
tions in order to qualify for subsidies or comply with a
insurance policy is equal to the sum of two components:
mandate. For example, proposals could require that plans
the average amount that an insurer expects to pay for
cover certain services, limit the amount of cost sharing
services covered under the plan; and a loading factor that
that would be required of enrollees, or be “actuarially
reflects the insurer’s costs of operating the plan (including
equivalent” to an existing plan.2 The more comprehen-
administrative expenses and a return on investment). An
sive the insurance coverage, the higher the premium
insurer’s costs for covered services in turn reflect the scope
would be. Because of the resulting increase in the use of
of benefits that are included, the plan’s cost-sharing
health care services, total spending also would be greater
requirements, and the health status of the plan’s enrollees.
under proposals that reduced cost sharing. CBO has
concluded that a 10 percent decrease in enrollees’ out-of-
The aggregate effects of those factors are illustrated by
pocket costs would typically cause average spending on
examining current premium levels. Reflecting the choices
health care to increase by 1 percent to 2 percent.
that individuals and families currently make, premiums
for employment-based plans are expected to average
1. See Didem Bernard and Jessica Banthin, Premiums in the Individ-
about $5,000 per year for single coverage and about
ual Insurance Market for Policyholders Under Age 65: 2002 and
$13,000 per year for family coverage in 2009. Premiums
2005, AHRQ Statistical Brief No. 2002 (Agency for Healthcare
for policies purchased in the individual insurance market
Research and Quality, April 2008); AHIP Center for Policy
are much lower—about one-third lower for single cover-
Research, Individual Health Insurance, 2006–2007: A Comprehen-
age and half that level for family policies.1 In large
sive Survey of Premiums, Availability, and Benefits (Washington,
D.C.: America’s Health Insurance Plans, December 2007); and
part, those differences reflect the fact that policies pur-
the Kaiser Family Foundation and Health Research and
chased in the individual market cover a lower share of
Educational Trust, Employer Health Benefits: 2005 Annual Survey
enrollees’ health care costs, on average, which also
(Washington, D.C.: Kaiser/HRET, September 2005) and
encourages enrollees to use somewhat fewer services.
Employer Health Benefits: 2007 Annual Survey (September 2007).
At the same time, average administrative costs are higher
2. Actuarially equivalent plans cover the same share of health care
for individually purchased policies. The remainder of the
spending for a given population; see Box 3-1 on page 64 for a
difference in premiums probably arises because people
more detailed explanation.
CBO
60
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
In addition, premiums could be affected by proposals
On the basis of a review of the research literature and
that changed insurers’ management of covered benefits.
original data analysis, CBO concludes that if all people
Most people who have private health insurance are
who are currently uninsured were enrolled in health
enrolled in some form of managed care plan. Those plans
insurance coverage that is similar in design to a typical
use various techniques to contain health care spending,
employment-based plan, they would use between
including negotiating lower fees with a network of pro-
75 percent and 95 percent as much care as the previously
viders, requiring that certain services be authorized in
insured. The remaining gap reflects CBO’s assessment
advance by the plan or by the patient’s primary care phy-
that, on average, people without insurance have a some-
sician, monitoring the care of hospitalized patients, and
what lower propensity to use health care services—a ten-
varying cost-sharing requirements to encourage the use of
dency that would persist if they became covered under a
less expensive prescription drugs. Proposals that restricted
new program. Providing all uninsured people with such
plans’ use of such management tools would tend to yield
coverage would thus cause total demand for health care
higher premiums and health care spending.
services to increase by 2 percent to 5 percent.
Another factor affecting the level of premiums is the cost
Those estimates are sensitive to the effects of proposals on
of administering a health plan. Some administrative costs
the extent of coverage that people receive, however; that
(such as those for customer service) vary with the number
is, more extensive coverage would have a larger effect. In
of enrollees in a plan, but others (such as those for sales
addition, how proposals that do not achieve universal or
and marketing efforts) are more fixed—that is, those
near-universal coverage would affect people’s health care
costs are similar whether a policy covers 100 enrollees or
spending depends on the extent to which the uninsured
100,000. As a result of those economies of scale, the aver-
would be covered under a plan and on assumptions about
age share of the policy premium that covers administra-
the underlying demand for health care among people
tive costs varies from about 7 percent for employment-
who would become insured. For more incremental
based plans with 1,000 or more enrollees to nearly 30
increases in insurance coverage rates, CBO would assume
percent for policies purchased by very small firms and by
that people who enrolled under a new program would
individuals. Some administrative costs are unavoidable,
have a greater propensity to use medical care than those
but proposals that shift enrollment away from the small-
who did not enroll. Depending on the design of such a
group and individual markets have the potential to avoid
proposal, those newly covered individuals might use
the added administrative costs per enrollee that are
health care services at a rate comparable with—or even
observed in those markets. In other cases, however, trade-
greater than—that of people with similar demographic
offs may arise between reducing administrative costs and
characteristics and health status who are currently
limiting overall health costs and insurance policy premi-
insured.
ums because some administrative costs are incurred when
using management tools designed to limit health care
In addition, studies indicate that about one-third of the
spending.
services the uninsured population uses either are provided
for free or yield lower total payments to providers than if
Proposals could change the types of coverage in which
the same services were provided to privately insured indi-
many people are enrolled as well as expand coverage to
viduals. To the extent that uncompensated care became
include people who would have otherwise been unin-
compensated, spending for the currently uninsured pop-
sured. The greatest effects on health care spending are
ulation would rise even if they did not use more services.
likely to be for the latter group because their use of health
care services could increase substantially once they
became insured. After accounting for differences in the
Design of Benefits and Cost Sharing
demographic characteristics and health status of the two
A health insurance plan is essentially a contract between
populations, CBO estimates that the uninsured use about
an insurer and an enrollee. In exchange for premium pay-
60 percent as much care as similar people who are
ments, the insurer agrees to cover certain medical services
insured.
that are specified in the plan. The plan also details the
share of costs that both the insurer and the enrollee will
bear for each of those services. Thus, two key design ele-
CBO
CHAPTER THREE
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
61
ments of a health insurance plan are its scope of covered
the country as a whole, existing state benefit mandates
benefits and its cost-sharing requirements.
increase premiums in the individual and small-group
markets by approximately 2 percent to 3 percent.4
Covered Benefits
Nearly all health insurance policies cover hospitalization,
Cost Sharing
physicians’ services, and prescription drugs—the three
Cost-sharing requirements—the amount that consumers
largest categories of spending on health care—but greater
are required to pay out of pocket when they use health
variation in coverage exists for dental care and more spe-
care services—can take the form of deductibles, co-
cialized medical services (such as infertility treatments).
insurance, or copayments. Deductibles are the amount of
Legislative proposals to increase the number of insured
spending an enrollee must incur before coverage begins;
people could require that health insurance plans cover
coinsurance and copayments are a portion of spending an
certain types of medical services. Under such proposals,
enrollee pays at the time of service. Most private insur-
individuals (or their employers) might not qualify for
ance plans also limit enrollees’ financial exposure through
subsidies or fulfill a mandate unless they were covered by
an annual cap on out-of-pocket spending. (See Chapter 1
plans that included those benefits.
for additional discussion of cost-sharing requirements.)
Benefit mandates ensure that enrollees who may need
A proposal to increase health care coverage could specify
those services will have coverage for them, but they also
either minimum or maximum levels of cost sharing that
tend to raise insurance premiums in order to cover the
would be allowed in order for an insurance policy to
added costs of the services. The extent of the premium
qualify for a subsidy or fulfill a mandate. For example, in
increase resulting from a mandate would depend not only
order to contribute to a health savings account (which
on the costs of the services involved and the likelihood
allows enrollees to pay many of their out-of-pocket costs
they would be used by enrollees but also on whether
using tax-preferred funds), an individual must be enrolled
health insurance policies would have covered those ser-
in a health insurance policy that in 2009 has an annual
vices in the absence of a mandate. Moreover, because
deductible of at least $1,150 for single coverage or $2,300
many states already require coverage of various benefits,
for family coverage and has an annual limit on out-of-
the impact of any federal mandates would depend on
pocket spending that does not exceed $5,800 or $11,600,
their scope relative to those existing state requirements
respectively.
and their applicability to plans that fall outside the pur-
view of state regulation.
Some proposals might also include additional subsidies to
reduce or eliminate cost sharing for lower-income indi-
Empirical evidence on the effect of benefit mandates on
viduals and families—to reflect their more limited ability
premiums and coverage is limited. A recent study spon-
to pay for services. The Medicaid program fills that role
sored by the state of Maryland found that the total cost of
for low-income Medicare enrollees by offering to cover
services covered by the state’s benefit mandates equaled
their cost-sharing requirements under Part A and Part B
15 percent of all covered claims.3 That figure overstates
of that program. About 12 million Medicare enrollees
the extent to which benefit mandates raise health insur-
with low income and few assets are entitled to subsidies
ance premiums nationally, for two reasons: first, because
that reduce or eliminate the deductible or other cost-
Maryland mandates more benefits than most other states;
sharing requirements under the Medicare drug benefit.
and second, because some insurers would have covered
Such subsidies would also raise many of the same issues
the mandated benefits even if they had not been required
about identifying and enrolling eligible individuals that
to do so (a factor noted in the study). On the basis of data
arise in designing and implementing premium subsidies.
on mandated benefits in other states and evidence on the
(See Chapter 2 for a discussion of the issues that arise
extent to which insurers cover such benefits in the
absence of mandates, CBO assumes that, averaged across
4. For evidence on the extent to which insurers would cover
mandated benefits in the absence of the mandates, see Maryland
3. Maryland Health Care Commission, Study of Mandated Health
Health Care Commission, Study of Mandated Health Insurance
Insurance Services: A Comparative Evaluation (January 1, 2008),
Services; and Jonathan Gruber, “State-Mandated Benefits and
http://mhcc.maryland.gov/health_insurance/mandated_1207.
Employer-Provided Health Insurance,” Journal of Public
pdf.
Economics, vol. 55 (1994), pp. 433–464.
CBO
62
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
in targeting such assistance toward lower-income
although some adverse effects were observed for low-
individuals.)
income people in poor health.
Changes in cost-sharing requirements primarily affect
Even though the RAND study was conducted more than
premiums by shifting the share of spending that is cov-
25 years ago, its findings continue to be widely used by
ered by the policy between the insurer and the enrollee.
analysts because of the strength of its experimental design
Those changes can also affect premiums, however, by
and the limited availability of more recent evidence.
causing total health care spending to increase or decrease.
Although the provision of health care and the design of
The best available evidence about the effects of cost shar-
insurance plans have changed considerably since the
ing on spending for health care comes from the RAND
study was conducted, the implications of those changes
Health Insurance Experiment, a large-scale study that was
for the impact of cost sharing are not obvious. Most of
conducted between 1974 and 1982.5 The RAND study
the plans examined in the RAND experiment were
measured the effects of cost sharing on the use of services,
indemnity plans, which essentially reimbursed enrollees
expenditures for health care, and health outcomes by
for the health care costs that they incurred; since that
randomly assigning nonelderly people to several different
time, the spread of managed care has provided insurers
types of health insurance plans and tracking their experi-
with tools to contain health care spending. To try to cap-
ence over time. A major advantage of using random
ture those effects in its analysis, CBO supplements the
assignment is that differences in outcomes across plans
results from the RAND study with more recent findings
can be attributed to the design features of each of the
and advice from experts in the health care industry.
plans rather than to the characteristics of the people
who were enrolled in them.
Based on its review of the literature and discussions with
outside experts, CBO assumes that a 10 percent decrease
The RAND study found that, compared with a plan in
in enrollees’ out-of-pocket costs will generally cause the
which all care was free, a plan with no deductible and a
total spending on their health care to increase by 1 per-
coinsurance rate of 25 percent reduced spending on cov-
cent to 2 percent. CBO uses that assumption for all types
ered services by about 20 percent; with a coinsurance rate
of plans and applies an additional factor to account for
of 95 percent, spending fell by about 30 percent.6 (The
the effect of benefit management on spending. The effect
differences in health care costs that would be covered by
of cost sharing on spending varies by type of service. For
the plan were even larger; compared with the free-care
example, the use of hospital care is less sensitive to cost-
plan, covered costs were about 40 percent lower with
sharing requirements than is the use of prescription
25 percent coinsurance and about two-thirds lower with
drugs. A similar analysis would be applied to proposals
95 percent coinsurance.) The RAND study also found
that provided subsidies to reduce cost sharing for lower-
that the effect of cost-sharing requirements varied with
income enrollees.
the type of services provided. Overall, though, cost-
sharing requirements resulted in less use of health care
Actuarially Equivalent Plans
services. Compared with study participants who received
One useful way to compare health insurance plans with
free care, those with cost-sharing requirements made, on
different design features is by examining their actuarial
average, one to two fewer visits to their doctors and had
value. That summary statistic measures the share of
20 percent fewer hospitalizations during a year. The
health care spending for a given population that would be
reduction in the use of health care services that resulted
covered by each plan and thus reflects both covered
from cost-sharing requirements did not have a substantial
services and cost-sharing requirements (see Box 3-1).
impact on health outcomes for the general population,
Although actuarial value provides one measure of the
comprehensiveness of benefits offered by an insurance
plan, it does not capture all features of a plan—such as
5. For a description of the RAND experiment, see Joseph P.
the utilization controls and size of the provider net-
Newhouse and the Insurance Experiment Group, Free for All?:
Lessons from the RAND Health Insurance Experiment (Cambridge,
work—that affect the benefits that are delivered. Because
Mass.: Harvard University Press, 1993).
of differences in those and other features, such as plans’
administrative costs and the populations they cover, plans
6. The plans that required coinsurance also featured an annual limit
on enrollees’ out-of-pocket costs that was equal to the lesser of
with the same actuarial value may charge different
$1,000 (in nominal dollars) or a percentage of family income.
premiums.
CBO
CHAPTER THREE
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
63
The actuarial values of current insurance plans vary across
Update Factors
employers and between the group market and the indi-
Whether a required level of coverage was defined using
vidual market. For employment-based plans, actuarial
the actuarial value of a benchmark plan or with reference
values—expressed as the share of a given population’s
to specific cost-sharing requirements, the way in which
medical claims that would be covered by the plan—are
those values were updated over time would have impor-
typically between 65 percent and 95 percent, with an
tant implications for a proposal’s effects on the federal
average value that is between 80 percent and 85 percent.
budget and on coverage rates. If a requirement regarding
Deductibles and other cost-sharing requirements are
the actuarial value of plans was fixed in nominal dollars
typically larger for policies purchased in the individual
(that is, not adjusted for inflation), plans would cover a
insurance market, where actuarial values generally range
declining share of health care costs as those costs rose.
from 40 percent to 80 percent, with an average value that
Alternatively, if plans were required to cover a specified
is between 55 percent and 60 percent.
percentage of health care costs, their actuarial value in
dollar terms would rise along with those costs. Similar
Actuarial-value calculations could be incorporated into
issues would arise if requirements were imposed on cost
legislative proposals in various ways. Proposals that speci-
sharing. If deductibles or other cost-sharing requirements
fied a particular benefit design could allow plans to devi-
were fixed in nominal dollars, the share of costs covered
ate from that design so long as they provided actuarially
by an insurance plan would increase over time as health
equivalent benefits. For example, the Medicare drug ben-
care costs rose—making the coverage more valuable but
efit specifies a standard benefit with a specific deductible,
also increasing its premium. (An example is the deduct-
coinsurance rate, and catastrophic threshold (above
ible under Part B of Medicare, which remained at $100
which enrollees pay about 5 percent of their drug costs).7
from 1991 to 2004.)
But drug plans are allowed to reduce the deductible, vary
the coinsurance rate, or use tiered copayments for differ-
Those issues can be addressed by indexing a plan’s
ent types of drugs so long as the plan’s overall actuarial
parameters, but the choice of index can significantly
value remains the same and certain other actuarial tests
affect the cost of a new program and the scope of cover-
are met. Drug plans are not, however, allowed to increase
age provided. If the required actuarial value of plans was
the deductible or change the catastrophic threshold.
specified in dollar terms and updated using a general
inflation index (such as the consumer price index) rather
In a similar manner, proposals could require that a quali-
than a health-specific index (such as growth in per capita
fied plan be actuarially equivalent to an existing plan. For
health expenditures), that value would probably decline
example, the standard Blue Cross and Blue Shield plan
in future years relative to the cost of health care because
available in the Federal Employees Health Benefits pro-
health care spending is expected to grow more rapidly
gram could be used as a model. The impact of that
than general price levels.
approach would depend partly on the extent to which the
model plan differed from the insurance plans that indi-
viduals currently have. According to CBO’s analysis, the
Management of Benefits
standard Blue Cross and Blue Shield plan has an actuarial
Over the past 30 years, private insurance coverage in the
value that is slightly above the national average for
United States has largely shifted away from indemnity
employment-based plans. In evaluating a proposal that
policies and toward managed care plans. Such plans vary
established the actuarial value of that plan as the mini-
considerably, but all use management techniques to try to
mum requirement for all qualified plans, CBO would
contain health care spending. Provisions that would
account for the increased spending that would result as
restrict plans’ use of such techniques would probably
lower-value plans were enhanced to meet that higher
increase premiums and health care spending, although
standard.
the amount of the increase would depend on the details
of the provisions.
Most people who have private health insurance are
7. The standard drug benefit also specifies an initial coverage limit;
enrolled in some type of managed care plan. Among
enrollees who do not have additional coverage are responsible for
all of their drug costs between that point and the catastrophic
workers covered by employment-based insurance, about
threshold (in the so-called doughnut hole).
58 percent are enrolled in preferred provider organiza-
CBO
64
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
Box 3-1.
What Is Actuarial Value?
Actuarial values provide a useful yardstick for com-
designs—but by holding medical claims fixed, they
paring insurance plans with different benefit designs.
do not take into account how enrollees might
Actuarial values may be calculated holding medical
respond to those different designs. The lower deduct-
claims constant, or they may take into account the
ible in Plan A could lead some enrollees to use more
effects of different designs on enrollees’ use of ser-
medical services, whereas the higher coinsurance rate
vices. Because those two methods can affect the com-
in that plan could have the opposite effect. If the
parisons of plans, it may be important for policy pro-
aggregate effects of those behavioral responses to cost-
posals that incorporate actuarial-value tests to specify
sharing requirements differed, the two plans would
which method to use.
have different expected costs and would be expected
to charge different premiums as a result.
The calculation of actuarial values is easiest to illus-
trate if enrollees’ medical claims are taken as constant.
If, instead, the calculation of actuarial value considers
In a simplified example, suppose Plan A has a $400
those behavioral responses to cost-sharing require-
deductible and charges a 25 percent coinsurance rate
ments, then the comparison of plans could be
beyond that point, up to an annual limit on out-of-
affected. For example, suppose Plan B yielded lower
pocket costs of $2,500, whereas Plan B has a $650
overall spending on health care and thus had lower
deductible, a 20 percent coinsurance rate, and a
covered costs. If that effect was taken into account,
$2,000 limit on out-of-pocket costs. An analysis of
the plan would have to reduce its cost-sharing
those plans’ actuarial values would simply take the
requirements (perhaps by lowering its coinsurance
same set of claims, apply the two benefit designs, and
rate or out-of-pocket limit) in order to have the same
determine the average amount of spending covered
actuarial value as Plan A.
by each policy (see the table for a stylized analysis in
which the actuarial values of those two plans are con-
Regardless of whether calculations of actuarial values
structed to be equal). Actuarial values are sometimes
account for enrollees’ responses to cost sharing, there
expressed as the percentage of health claims that an
are several other reasons that health insurance plans
insurance plan will cover and sometimes as the aver-
with the same actuarial value might charge different
age dollar value of covered claims; if the universe of
premiums:
claims is held constant, then those two measures are
interchangeable.
B The total price paid for each service may differ.
Calculations of actuarial value hold the prices of
Those calculations of actuarial value provide a simple
services constant, but in practice, one plan may
metric by which to compare differing benefit
negotiate lower payment rates for its providers.
tions, about 20 percent are enrolled in health mainte-
with an account that enrollees can use to help finance
nance organizations, and about 12 percent are covered by
their out-of-pocket costs. Those plans generally feature
point-of-service plans.8 Those plans vary in their use of
provider networks and other requirements that are similar
provider networks, authorization requirements for more
to PPO plans but are not included in the count of PPO
expensive treatments, and other features. About 8 percent
enrollees.
of workers are enrolled in policies known as consumer-
directed health plans, which combine a high deductible
Average premiums for those plans also vary (see
Figure 3-1); that variation reflects not only differences
in benefit design and management but also differences
8. The Kaiser Family Foundation and Health Research and Educa-
in administrative costs, the populations covered by the
tional Trust, Employer Health Benefits: 2008 Annual Survey. Only
about 2 percent of workers are enrolled in indemnity plans that do
plans, and the geographic areas that those plans serve.
not use managed care tools.
Consequently, those premiums do not measure the
CBO
CHAPTER THREE
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
65
Box 3-1.
Continued
What Is Actuarial Value?
Illustrative Calculation of Actuarial Values
(Dollars)
Plan A
Plan B
Total
Share
Share
Spending for
Out-of-Pocket
Covered
Covered
Out-of-Pocket
Covered
Covered
Health Care
Costs
Costs
(Percent)
Costs
Costs
(Percent)
Enrollee 1
600
450
150
25
600
0
0
Enrollee 2
1,000
550
450
45
720
280
28
Enrollee 3
1,500
675
825
55
820
680
45
Enrollee 4
3,700
1,225
2,475
67
1,260
2,440
66
Enrollee 5
18,200
2,500
15,700
86
2,000
16,200
89
Average
5,000
1,080
3,920
78
1,080
3,920
78
Source:
Congressional Budget Office.
Note: The actuarial value of both plans is $3,920, or 78 percent of enrollees’ average spending on health care.
B The degree to which benefits are managed may
tend to use more health care and may use a differ-
differ. One plan might use more utilization con-
ent mix of services.)
trols, such as requirements to receive authorization
for certain surgeries, or may contract with doctors
An additional consideration arises when evaluating
and hospitals that have less expensive patterns of
the actuarial value of consumer-directed health plans.
practice.
Such plans generally combine a high-deductible
health insurance policy with an account that enrollees
B Administrative costs may differ. Comparisons of
may use to help finance their out-of-pocket costs
actuarial value look only at covered benefits under
(and which may accumulate balances over time).
each plan, but insurance premiums also reflect the
By design, the high-deductible insurance policy will
administrative costs that insurers incur.
generally have a lower actuarial value than conven-
tional insurance policies. But the actuarial value of
B The enrollees in each plan may differ. Calculations
consumer-directed plans would include the expected
of actuarial values use a broadly representative
value of any contributions that an insurer or
population, but plans that attract enrollees who
employer sponsoring the plan would make to an
use above-average amounts of health care will tend
enrollee’s account—so that contribution could be set
to charge higher premiums. (Correspondingly, the
to make the overall actuarial value of the consumer-
actuarial value of a given benefit design would dif-
directed plan equal to the value of a conventional
fer if provided to an elderly population rather than
health plan.
a younger population, simply because the elderly
CBO
66
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
Figure 3-1.
Managed care plans differ in their coverage for services
Average Annual Premiums for
that enrollees receive from providers who are not in a
Covered Workers for Single and
plan’s network. That coverage affects plans’ costs directly
as well as their leverage in negotiations with providers.
Family Coverage, by Plan Type, 2008
PPOs cover services received from any licensed provider,
but they encourage enrollees to receive care from provid-
(Thousands of dollars)
ers in their network by charging them less for such care.
16
HMOs do not provide any coverage for services received
Single
Family
outside their provider network (except in emergencies).
14
POS plans offer a middle ground: They cover services
12
received outside a plan’s network (usually for a higher
charge), but they typically apply some of the same man-
10
agement tools that HMOs use to limit costs within the
8
network.
6
The differences between PPOs and HMOs have generally
4
narrowed in recent years. In the past, HMOs typically
had smaller provider networks than PPOs. In response to
2
consumer demand for broader networks, however,
HMOs have generally increased the size of their net-
0
works, and many are now similar to PPOs in that respect;
All Types
CDHP
POS
PPO
HMO
indeed, some insurance carriers use the same provider
Source:
Congressional Budget Office based on data from Kaiser
network for their HMO and PPO products.
Family Foundation and Health Research and Educational
Trust, Employer Health Benefits: 2008 Annual Survey
Use of Other Cost-Containment Practices
(Washington, D.C.: Kaiser/HRET, September 2008).
Note: CDHP = consumer-directed health plan; POS = point-of-
Health plans use a variety of other practices to contain
service plan; PPO = preferred provider organization;
health care costs. One approach is to manage access to
HMO = health maintenance organization.
expensive medical benefits by requiring prior authoriza-
tion before the services will be covered. A second
average costs of different types of plans in delivering a
approach is to use price signals—that is, variations in
particular benefit package to a particular enrolled
cost-sharing requirements—to encourage enrollees to use
population.
less expensive medical care. Managed care plans also may
use evaluations of providers on both price and quality
Use of Provider Networks
terms to give feedback to those providers or to structure
One important way that managed care plans seek to con-
the information and incentives given to enrollees.
trol costs is by developing networks of providers—includ-
ing hospitals, physicians, and laboratories—and negotiat-
Managing Access and Use. Controlling enrollees’ access to
ing discounts within those networks. Many providers are
more specialized (and expensive) medical services can
willing to charge lower fees to a plan’s enrollees in the
help health plans manage their costs. Many plans require
expectation that the plan will direct patients to them
prior authorization for nonemergency hospital admis-
(and to avoid the loss of patients that could occur if the
sions and other selected services. To limit access to expen-
provider was excluded from a plan’s network). Despite
sive drugs, many plans use “step therapy”—a process in
those efforts, the rates that health plans pay hospitals have
which patients are required to begin treatment with less
risen substantially in many market areas in recent years,
expensive alternatives (such as generic drugs) and then
as hospitals have achieved stronger bargaining positions
switch to a more expensive drug only if necessary. Enroll-
in their dealings with plans. The enhanced bargaining
position of hospitals has been attributed in part to con-
9. See Justin S. White, Robert E. Hurley, and Bradley C. Strunk,
solidation among hospitals and also to strong consumer
Getting Along or Going Along? Health Plan-Provider Contract
preferences that hospitals not be excluded from a plan’s
Showdowns Subside, Issue Brief No. 74 (Washington, D.C.: Center
network.9
for Studying Health System Change, January 2004).
CBO
CHAPTER THREE
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
67
ees in some HMO and POS plans must select a primary
discounts or rebates from drug manufacturers in return
care physician who is responsible for approving referrals
for giving their drugs preferred status on the formulary—
to specialists, but that approach is less common than in
and with it, lower copayments for enrollees. Plans also
the past. More generally, differences between HMOs and
encourage enrollees to use lower-cost generic versions of
PPOs have diminished as many managed care plans elim-
drugs when they are available, by setting the lowest
inated or relaxed some of their cost-control procedures in
copayment amounts for those drugs.
response to widespread complaints in the late 1990s from
consumers and providers. However, some of those cost-
More recently, some plans have begun using the informa-
containment procedures were subsequently reinstated or
tion collected from provider profiling to designate a
replaced by new procedures to limit spending.10
preferred “tier” of providers based on quality and cost
standards. For example, an insurer might rank providers
Some managed care plans also seek to monitor the perfor-
on the basis of the total cost per medical “episode,” which
mance of providers and the use of services more actively.
accounts for complications, readmissions to the hospital,
To limit the length of hospital stays, plans may conduct
and other factors that go beyond a consideration of the
“concurrent reviews” of the care and medical condition of
provider’s fees. In addition, enrollees may be given finan-
hospitalized patients; that is, the insurer may review the
cial incentives—such as lower cost-sharing require-
medical necessity of a treatment upon or shortly after a
ments—to receive their care from higher-tier providers.
patient’s hospital admission and continue to monitor the
services provided during the course of his or her treat-
Effects of Managed Care on Premiums and Spending
ment. Some insurers also apply such reviews to patients
Determining the effects of the various cost-containment
receiving other types of care, such as rehabilitation and
tools can be difficult because health plans use different
skilled nursing care.11 Insurers sometimes use informa-
combinations of them, and plans vary along a number of
tion from medical claims to obtain detailed information
other dimensions. Consequently, much of the published
about the care furnished by individual providers (a proce-
research has focused on comparing HMOs (which have
dure known as provider profiling). That information is
traditionally used more stringent cost-containment meth-
then used to give feedback to providers on how their
ods) with other types of plans. The evidence suggests that
practice patterns compare with those of their peers and to
HMOs deliver a given package of benefits at a lower cost
identify providers who are furnishing inappropriate or
than PPOs and other plans. In particular, studies have
excessive care (and who might be removed from the plan’s
found that HMOs reduce the use of hospital services and
network as a result).
other expensive services.12 Because those studies rely
largely on data that are more than a decade old, however,
Varying Cost Sharing. Plans encourage enrollees to use
they probably overstate the differences that exist today
providers within their network by requiring lower cost
between HMOs and other types of plans. On the basis of
sharing for in-network care. In some cases, plans also use
the available evidence, CBO estimates that plans making
differences in cost-sharing requirements or other tech-
more extensive use of benefit-management techniques
niques to influence consumers’ choices within their
would have premiums that are 5 percent to 10 percent
approved networks or range of covered treatments and
lower than plans using minimal management techniques.
services. For example, plans generally establish a drug for-
mulary or list of drugs that the plan covers (which is akin
The slowdown in health care spending that occurred in
to a provider network). In addition, plans typically try to
the 1990s—as private insurance coverage shifted away
limit spending on prescription drugs by negotiating price
from indemnity policies and toward various forms of
managed care—provides additional evidence on the
potential effects of managed care on spending. Before
10. Glen P. Mays, Gary Claxton, and Justin White, “Managed Care
1993, health care spending generally grew at a faster rate
Rebound? Recent Changes in Health Plans’ Cost Containment
than gross domestic product. From 1993 to 2000, the
Strategies,” Health Affairs, Web Exclusive (August 11, 2004),
pp. W4-427 to W4-436.
share of workers with private health insurance who were
11. Plans use various methods to pay providers, some of which are
designed to give providers incentives to limit spending. The
12. See Robert H. Miller and Harold S. Luft, “HMO Plan Perfor-
payment methods that plans use and their effects are examined in
mance Update: An Analysis of the Literature, 1997–2001,” Health
Chapter 5.
Affairs, vol. 21, no. 4 (July/August 2002), pp. 63–86.
CBO
68
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
enrolled in some kind of managed care plan rose from
enrollees are another area of concern, with legislative
54 percent to 92 percent. During that period, total
proposals addressing how information about a plan is
spending for health care remained nearly constant as a
presented to enrollees or specifying rules for the prompt
share of the economy, at about 13.8 percent of GDP.
payment of claims to providers.
Many analysts believe that the growth in managed care
plans contributed significantly to the slowdown in the
Other provisions could also regulate insurers’ networks of
growth of health care spending during that period.
providers. Any-willing-provider laws require that health
plans include in their network any provider who agrees to
By the end of the 1990s, opposition to the restrictions
abide by the terms and conditions of the plan’s contract.
imposed by managed care plans was growing among con-
Many states enacted such laws in the 1990s, but those
sumers and providers. The plans responded by relaxing
laws do not apply to employment-based plans that are
those restrictions, and enrollment shifted to more loosely
exempt from state regulation. Network-adequacy require-
managed PPO plans. Health care spending also began to
ments would establish rules about the number of differ-
increase at a faster rate, rising to 16.0 percent of GDP in
ent types of providers that plans must have in their net-
2006. Other factors, however, have undoubtedly contrib-
work, and restrictions on provider profiling would limit
uted to the growth in health care spending relative to the
plans’ ability to use their analysis of medical claims and
size of the economy since 2000; hospital mergers became
other factors to exclude providers from their networks or
more pervasive, for example, enhancing hospitals’ lever-
to develop tiered networks.
age in negotiating with health plans.
Effects on Health Insurance Premiums. In its previous
Regulating the Operations of Health Plans
analyses of proposals to create a Patients’ Bill of Rights in
Proposals to change the health insurance market or to
1999 and 2001, CBO generally determined that many of
subsidize insurance purchases might include provisions
their provisions—which are similar to those described
affecting the management of health plans. During the
above—would increase spending on health care.13 Since
past decade, for example, the Congress has considered
then, however, many health plans have dropped certain
several versions of legislative proposals—commonly
cost-containment procedures or replaced them with other
referred to as a “Patients’ Bill of Rights”—that would
techniques; to the extent that such changes were not
have restricted insurers’ management of health benefits.
anticipated, the magnitude of CBO’s estimates of the
Although lawmakers did not enact those proposals, some
effects of new proposals that affect plans’ management
states adopted similar provisions restricting health insur-
techniques may differ from its previous findings.
ers that operate in their jurisdiction. (As discussed in
Chapter 1, plans purchased in the individual insurance
For certain provisions that CBO analyzed, the effects
market and most plans purchased by smaller employers
today would most likely differ from what the agency
are subject to state regulations, whereas the majority of
previously estimated. For example, CBO estimated that a
plans offered by larger employers are exempt.) In model-
federal any-willing-provider law or federal network-
ing the effects of such proposals, CBO considers the
adequacy requirements and proposals requiring plans to
nature of any provisions governing the plan’s structure,
cover certain types of care—including visits to specialists
utilization management, and provider networks and their
without prior authorization, visits to an emergency room
interaction with existing state requirements.
if a “prudent layperson” would have regarded the patient’s
Types of Provisions. Proposals like the Patients’ Bill of
13. See the Congressional Budget Office’s cost estimates for S. 6,
Rights could change how health insurers interact with
Patients’ Bill of Rights Act of 1999 (June 16, 1999); for H.R.
enrollees, in several ways. Under some proposals, insurers
2315, Patients’ Bill of Rights Act of 2001 (July 20, 2001); for
would be required to cover certain types of care, such as
S. 1052, Bipartisan Patients’ Bill of Rights Act (July 20, 2001);
visits to specialists, without a referral from an enrollee’s
and for H.R. 2563, Bipartisan Patient Protection Act (July 26,
primary care physician. Past proposals also would have
2001). All of CBO’s estimates reflect the marginal effect of the
federal legislation on costs. That is, they compare the projected
granted enrollees rights of redress, allowing those who
costs under the new law with the estimated costs that would have
had been denied coverage for a particular service to
prevailed in the absence of the new law (recognizing that some
appeal the decision or pursue other remedies in civil
health plans may have made the proposed changes even without
courts. Transactions among insurers, providers, and
the new law).
CBO
CHAPTER THREE
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
69
Table 3-1.
disputes with insurers would probably be similar to the
Administrative Costs for Private
effects that were estimated in 2001 because the expecta-
tion in the original estimates that the legal environment
Health Plans, by Category, 2006
would not change substantially has, so far, proved to be
accurate. At that time, CBO estimated that the combined
(Billions of dollars)
effect of various provisions creating new civil remedies
Costs
and establishing grievance processes would have been to
Marketing Costs
increase premiums by 1.1 percent to 1.7 percent.
Sales and marketing
20
Customer service
3
Actuarial and underwriting services
1
__
Administrative Costs of Health Plans
Subtotal
24
Proposals to change the regulation of insurance mar-
kets—as well as many other types of proposals—could
Medical Activities
affect the costs of health insurance by changing the
Medical management
7
administrative costs of health plans (sometimes referred
Claims processing
7
to as “administrative load”). In this discussion, adminis-
Regulatory compliance
*
__
trative costs refer to any expenses insurers incur that are
Subtotal
14
not payments for health care services, including the
General Administrative Costs
profits retained by private insurers and the taxes paid on
General management and overhead
10
those profits.
Information technology
12
Taxes
9
Types of Administrative Costs
After-tax profits
21
__
Administrative costs can be divided into three categories:
Subtotal
52
B
Total Administrative Costs
90
Marketing costs include expenses for advertising,
sales, enrollment processing, customer service, billing,
Source:
Congressional Budget Office based on Diana Farrell and
and actuarial and underwriting activities. (Underwrit-
others, Accounting for the Cost of U.S. Health Care, 2008:
ing involves an assessment of an applicant’s health and
A New Look at Why Americans Spend More (San Fran-
expected use of health care in order to determine what
cisco: McKinsey Global Institute, December 2008).
premium to charge.)
Note: * = between zero and $500 million.
B Costs associated with medical activities include
condition as an emergency, and the routine costs of
expenses for claims review and processing, medical
enrollment in approved clinical trials—would, in combi-
management (such as utilization review, case manage-
nation, have increased private health insurance premiums
ment, quality assurance, and regulatory compliance),
by amounts ranging from 1.2 percent to 1.7 percent. If
and provider and network management (contracting
reintroduced today, however, similar provisions would
with doctors and hospitals and maintaining relations
probably have a smaller impact on premiums; to an
with providers).
extent not anticipated in CBO’s original estimates, many
health plans have acceded to consumers’ preferences for
B General administrative costs are difficult to allocate
broader access to care by expanding the size of their pro-
to a specific function; they include expenses for infor-
vider networks and eliminating or reducing some of their
mation technology, general management overhead,
restrictions on the use of covered services.
profits, and taxes.
For other provisions, CBO’s estimates of the effects
According to a recent analysis, administrative costs for
would be similar to the agency’s previous estimates. For
private health insurance totaled $90 billion in 2006 (see
instance, the effects of proposals to expand enrollees’
Table 3-1), of which about $24 billion was for marketing
access to the courts for pursuing civil remedies to settle
and related costs, roughly $14 billion was for medical
CBO
70
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
activities, and about $52 billion was for general expenses
latter loading factor is comparable with the one seen in
(including $9 billion in tax payments and $21 billion in
the individual insurance market, where administrative
after-tax profits).14 Overall, those costs accounted for
costs account for nearly 30 percent of premiums.
about 12 percent of private insurance premiums.
To a large extent, the variation in administrative costs
A common metric that is used to assess an insurer’s
among private plans reflects economies of scale. Some
administrative costs is the ratio of claims payments to the
types of administrative costs, such as sales and marketing
total premium, referred to as the “medical loss ratio.”
expenses, are relatively fixed for the group being insured;
(The difference between the medical loss ratio and 100
thus, the larger the group, the smaller the cost per
percent is the share of the premium devoted to adminis-
enrollee. In particular, plans that are sold to individuals
trative expenses.) When comparing two plans that are
and small groups are more likely to incur fees for insur-
equivalent on other dimensions, such as the total pre-
ance agents and brokers to handle the responsibilities that
mium and quality of service, a low loss ratio could indi-
larger firms generally delegate to their human resources
cate a plan that is run less efficiently. But a loss ratio is
departments—such as finding plans and negotiating pre-
not always indicative of a plan’s efficiency or value.15 For
miums, providing information about the selected plans,
example, a health plan that devotes more resources to
and processing enrollees. Because large firms can spread
managing the use of health care services might have a rel-
those costs over a greater number of enrollees, their aver-
atively low loss ratio but also a lower overall premium. In
age administrative costs per enrollee are lower.
contrast, a more lightly managed plan might have a high
loss ratio but a correspondingly higher overall premium
Other factors appear to play a lesser role in the variation
and might be covering more services that provide limited
of average administrative costs across markets. One com-
health benefits. The former plan, despite its low loss
monly cited difference is that underwriting is used in the
ratio, may well be preferable because of its lower overall
individual and small-group markets, but those efforts
premium for the package of services that it provides.
appear to account for a relatively small share of insurers’
Thus, a loss ratio provides just one way of evaluating a
administrative costs and thus seem unlikely to explain the
health plan’s administrative expenses.
higher administrative costs per enrollee that are observed
in those markets. Plans sold in the individual and small-
Variation of Administrative Costs
group markets are also generally subject to state taxes on
Administrative costs typically vary not only by the type of
the premiums they collect, whereas the plans offered by
insurance plan but also by the size and nature of the
large employers are generally exempt from such require-
group being insured. Among employment-based plans,
ments. Other expenses—such as the costs of responding
the share of the premium that pays for administrative
to telephone calls from enrollees and providers with ques-
costs varies significantly by the size of firms, from about
tions regarding coverage and payments—are roughly pro-
7 percent for firms with at least 1,000 employees to
portional to the number of enrollees (at least for broadly
26 percent for firms with 25 or fewer employees.16 The
similar populations) and thus would probably constitute
a similar share of the premiums for groups of different
sizes.
14. See Diana Farrell and others, Accounting for the Cost of U.S. Health
Care, 2008: A New Look at Why Americans Spend More (San
Francisco: McKinsey Global Institute, December 2008). National
Potential Effects of Proposals on
health expenditure data show a comparable estimate of about
Administrative Costs
$89 billion for 2006.
Depending on their design, proposals could have a signif-
15. For a discussion, see James C. Robinson, “Use and Abuse of the
icant impact on the administrative costs involved in pro-
Medical Loss Ratio to Measure Health Plan Performance,” Health
viding health insurance—which, in turn, could have a
Affairs, vol. 16, no. 4 (July/August 1997), pp. 176–187.
substantial effect on policy premiums. Administrative
16. Lower estimates of administrative costs for large-employer groups
costs would probably be affected indirectly by proposals
may reflect only the fees paid to insurers who act as third-party
that altered the number of insurers, the size of purchasing
administrators but who do not assume financial risk for operating
pools, and insurers’ responsibilities. Some proposals
an employer’s plan; when the employer is acting as the insurer,
some administrative costs are borne by the employer but may not
might seek to limit the amount spent on administrative
be included in the estimates.
costs by specifying a minimum loss ratio, but the net
CBO
CHAPTER THREE
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
71
effect of such proposals on insurance premiums or health
care. The extent to which the demand for care would
care spending is uncertain.
increase depends partly on the number and characteristics
of the newly enrolled individuals—including their health
Trade-offs are likely to arise between the number of
status and their preferences for medical care—and partly
insurance plans that are offered to consumers and the
on the scope of the coverage that they obtain. Estimating
total administrative costs incurred by all insurers. Because
that likely impact presents a number of challenges.
some administrative costs are largely fixed, duplication of
functions would arise in proportion to the number of
Based on a review of the research literature and original
insurers participating in the market. Greater competition
analysis, CBO concludes that if all people who are
among insurers, however, would also tend to provide
currently uninsured were enrolled in insurance coverage
stronger incentives to control costs and thus could yield
equivalent to a typical employment-based plan, they
lower total premiums despite causing aggregate adminis-
would use about 75 percent to 95 percent as much
trative costs to increase.
medical care as people who are currently insured (and
also have the same demographic characteristics and
Proposals that would organize insurance purchasers into
health status). Those figures provide a benchmark for
larger groups could avoid some of the high administrative
analyzing the impact of various coverage expansions.
costs observed in the individual and small-group markets.
Depending on their design, proposals for more incremen-
In the extreme, if a proposal established a purchasing
tal coverage expansions could provide coverage to a group
system under which all insurers incurred administrative
of people who would use at least as much health care as
costs that were comparable with the costs of large
similar people who are currently insured.
employment-based plans, average policy premiums
would be about 3 percent lower than they would be if
Estimates of Demand for Health Care by the
administrative costs for individual and small-group pur-
Uninsured
chasers remained at their current levels. Administrative
How much more care the uninsured would seek and the
savings, however, might be smaller if plans still had to rely
impact that such an increase would have on premiums
on insurance agents and brokers to enroll workers who
and spending depend in part on how much care they now
were not employed by large firms or if other entities had
receive. According to several studies and CBO’s own
to perform similar functions.
analysis of the nonelderly population, the uninsured use
Some proposals would try to directly limit administrative
about 50 percent to 70 percent as many health care ser-
costs by mandating minimum loss ratios—that is, by
vices as the insured.17 A key challenge in estimating the
specifying that the amounts spent on benefits should be
impact of a coverage expansion is sorting out the extent
at least some specified percentage of the premium. That
to which that disparity stems from the uninsured’s lack of
strategy could be problematic, however, because a high
coverage, how much reflects other observable differences
loss ratio may not imply greater efficiency on the part of
between the insured and the uninsured, and what role is
an insurer. Moreover, whether insurers serving the indi-
played by differences that researchers cannot easily
vidual and small-group markets could increase their loss
observe.
ratios simply because they were required to do so is not
clear, so the effects of such requirements on those markets
Although there are substantial demographic differences
are hard to predict. If the requirement was set too high,
between the insured and the uninsured, some of those
insurers would probably exit the market.
differences have offsetting effects on their relative use of
services. For example, younger adults are represented dis-
proportionately in the uninsured population, whereas the
Effects of Gaining Insurance
insured population is more likely to contain children
Coverage on Health Care Use and
Spending
17. For example, see M. Susan Marquis and Stephen H. Long, “The
Uninsured Access Gap and the Cost of Universal Coverage,”
Proposals that expand coverage to people who currently
Health Affairs (Spring 1994), pp. 211–220; and Brenda C. Spill-
lack insurance would lead to an increase in their use of
man, Stephen Zuckerman, and Bowen Garrett, Does the Health
medical services, which in turn would affect the costs of
Care Safety Net Narrow the Access Gap? Discussion Paper 03-02
those proposals and their impact on spending for health
(Washington, D.C.: Urban Institute, April 2003).
CBO
72
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
(who tend to use fewer health care services than average)
which people who are uninsured would increase their use
and older adults (who have above-average use). As a
of services if they were provided with coverage:
result, differences in age do not appear to explain much
of the overall disparity in use of services between the
B Simulations based on findings from the RAND
insured and the uninsured. Differences in health status
Health Insurance Experiment, which randomly
may play a larger role. CBO’s analysis of survey data
assigned individuals to different insurance plans;
indicates that the share of the nonelderly population
reporting their health as fair or poor is higher among the
B Analysis of so-called natural experiments, in which
uninsured (10 percent) than among the privately insured
coverage under Medicaid or Medicare has been
(5 percent).
extended to individuals who previously lacked insur-
ance; and
A more difficult factor to assess is whether the uninsured
differ from those with insurance in other less observable
B Studies that have compared the use of services by
ways that affect their demand for health care services.
people who are insured with that of people who are
Understanding the reasons that the uninsured currently
uninsured, taking into account various differences
do not have insurance could also provide some insight
among the two populations.
into how they would respond to an increase in coverage.
None of those approaches resolves all of the methodolog-
The uninsured are not a monolithic group, however, and
ical issues that arise when trying to estimate an uninsured
there are many reasons that they lack coverage. Some
individual’s likely use of health care if provided with
uninsured individuals may have a strong preference for
insurance. Reflecting the different strengths and weak-
health insurance but lack coverage because of limited
nesses of the approaches—as well as their differing
financial resources. If those financial constraints were
sources of data and analytic techniques—studies based on
relaxed, their use of health services might become compa-
those approaches have yielded a wide range of estimates
rable with that of otherwise similar people who have
of those effects.
insurance. Other people may not purchase insurance
because they place a relatively low value on health care or
The studies also differ in what they examine. Some stud-
think they will not need to use it. Still others may be will-
ies focus on people’s use of services (primarily doctors’
ing to accept more risk than those who enroll in health
visits and hospitalizations), and others analyze spending.
insurance plans or may believe that they will be able to
The impact that covering uninsured individuals has on
obtain the care they need without insurance. Such indi-
spending for health care and health insurance premiums
viduals may not substantially increase their use of health
depends both on the quantity of services that they use
care services even if they become insured.
and on the amount per service that is paid to the provid-
ers of their care. Analyzing those elements separately can
Both because individuals’ preferences for health care vary
be useful, however, because differences in payment rates
and those preferences are not easy to observe, estimating
can complicate comparisons between insured and unin-
the impact of gaining health insurance coverage on the
sured individuals. In particular, a substantial minority of
use of medical services is difficult. If individuals who are
the care that the uninsured receive is uncompensated or
more likely to use health care are also more likely to have
undercompensated—that is, they either pay nothing for
insurance, simple comparisons of the insured and unin-
it or pay less than the amount that a provider would
sured populations would overstate the impact of becom-
receive for treating an insured patient. To the extent that
ing insured. An ideal research strategy would randomly
such care became compensated under a proposal to
assign individuals to an insured or uninsured group and
expand coverage, health care spending for the uninsured
see how much care they use—but people would be
would increase, regardless of whether their use of care also
understandably reluctant to participate in such an experi-
rose. (Other factors that might affect the impact of an
ment. Short of that, researchers have used three broad
insurance expansion on spending, including any con-
methodological approaches to examine the extent to
straints on the supply of health care services and any
effects on other payment rates from reductions in
uncompensated care, are discussed in Chapter 5.)
CBO
CHAPTER THREE
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
73
Simulations Based on Experimental Evidence. One study
that people gain insurance coverage for reasons that are
used the results of the RAND Health Insurance Experi-
unrelated to their health or their preferences about health
ment to simulate spending for individuals enrolled in
care. One study examined the impact of Medicaid expan-
hypothetical insurance plans as well as for individuals
sions between 1984 and 1992, finding that children who
with no insurance.18 Although that study used older
became eligible for the program increased their likelihood
data, its results can be adjusted so that they reflect current
of visiting a physician at least once during the year by
levels of health care spending. The simulations indicate
about 10 percent and roughly doubled their probability
that, on average, people enrolled in a plan with a $400
of being hospitalized.19 A more recent study on expand-
deductible, 25 percent cost sharing, and a $4,000 maxi-
ing eligibility for Medicaid found similar effects on visits
mum on out-of-pocket spending (a plan that is roughly
to physicians.20
equivalent to typical employment-based plans) would
incur about 13 percent more in health care expenditures
Two other studies examined health care use by previously
than similar people who are uninsured. Because the prices
uninsured individuals shortly before and after they
paid for services were standardized in that analysis, that
became eligible for Medicare at age 65—a natural experi-
difference in spending also reflects the projected differ-
ment that is similar in many respects to a proposal that
ences in use of services.
yields near-universal coverage. One study, which focused
on a subset of clinical services, found that the use of pre-
Basing simulations on the findings from the RAND
ventive care by the previously uninsured rose substantially
experiment builds on both the strengths and weaknesses
once they were eligible for Medicare but remained below
of that study. Because random assignment ensured that
the levels seen for individuals who had been insured
enrollees in different plans were comparable, its design
before age 65.21 In addition, visits for arthritis treatments
allowed researchers to isolate responses to variations in
not only increased substantially among those who lacked
the extent of insurance coverage. Yet the RAND study
insurance before becoming eligible for Medicare but also
did not include any individuals who were uninsured; the
reached a higher level than was seen for the continuously
closest design resembled a high-deductible plan, which
insured. The other study compared overall numbers of
covered more than half of its enrollees’ health care costs.
physicians’ visits and hospital admissions and found that,
The researchers therefore had to extrapolate from the
among the near elderly, use of care was about 15 percent
RAND results to simulate spending among the unin-
lower for those who lacked insurance coverage compared
sured. One concern with that approach is that uninsured
with those who were insured.22 Once they enrolled in
individuals may be more reluctant to seek treatment than
Medicare, previously uninsured individuals increased
a comparison of enrollees in low-deductible and high-
deductible plans would indicate. Extrapolating from
19. Janet Currie and Jonathan Gruber, “Health Insurance Eligibility,
those results may also fail to reflect certain constraints on
Utilization of Medical Care, and Child Health,” Quarterly Journal
the medical care available to the uninsured. For example,
of Economics, vol. 111, no. 2 (May 1996), pp. 431–466. Because
some physicians do not accept new patients who are
actual enrollment in Medicaid reflects preferences about insurance
uninsured. As a result, simulations based on the RAND
and may be triggered by a health problem, the study compared
groups who gained eligibility for the program (regardless of
study’s results may underestimate the increased use of ser-
whether they actually enrolled) to groups whose eligibility did not
vices that would occur if insurance coverage was extended
change.
to people who are uninsured.
20. Jessica S. Banthin and Thomas M. Selden, “The ABCs of
Children’s Health Care: How the Medicaid Expansions Affected
Studies of Expansions in Medicaid and Medicare. Some
Access, Burdens, and Coverage Between 1987 and 1996,” Inquiry,
studies have examined the change in service use that
vol. 40, no. 2 (Summer 2003), pp. 133–145.
occurs when uninsured people become eligible for Med-
21. J. Michael McWilliams and others, “Impact of Medicare Coverage
icaid or Medicare. The creation or expansion of such pro-
on Basic Clinical Services for Previously Uninsured Adults,”
grams can provide useful insights, but only to the extent
Journal of the American Medical Association, vol. 290, no. 6
(August 13, 2003), pp. 757–764.
18. Joan L. Buchanan and others, “Simulating Health Expenditures
22. J. Michael McWilliams and others, “Use of Health Services by
Under Alternative Insurance Plans,” Management Science, vol. 37,
Previously Uninsured Medicare Beneficiaries,” New England
no. 9 (September 1991), pp. 1067–1090.
Journal of Medicine, vol. 357, no. 2 (July 12, 2007), pp. 143–153.
CBO
74
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
their health care use by 30 percent to 40 percent and
populations and then sought to project what expendi-
ended up with higher levels of use than were observed for
tures for the uninsured would be if they gained a typical
Medicare enrollees who had been insured before becom-
level of coverage.23 The study used data on the
ing eligible for Medicare at age 65 (although the differ-
nonelderly population from the Medical Expenditure
ences were not always statistically significant).
Panel Survey (MEPS), a large-scale survey that collects
information on individuals’ insurance coverage and use of
One advantage of the studies of Medicaid and Medicare
health care services. The analysis controlled for differ-
is that they clearly isolate the effects of gaining insurance
ences in demographic, health, and socioeconomic charac-
coverage; one limitation is that their results may not be
teristics between the insured and uninsured populations.
applicable to the entire uninsured population. To some
To account for the fact that a large share of services
extent, those studies may reflect the responsiveness of
received by the uninsured are uncompensated, the study
people who are uninsured primarily because of financial
also sought to adjust their spending figures upward so
constraints—and who are thus more likely to increase
that they would reflect payment rates for the privately
their use of health services once they receive coverage.
insured. The study estimated that spending on health
More specifically, the findings related to Medicaid expan-
care for the uninsured—and, by implication, their use of
sions may apply only to similar proposals. On the one
services—would increase by about 70 percent if they
hand, Medicaid coverage has relatively low cost sharing,
became continuously insured; the estimated impact was
so it could have stimulated demand for care to a greater
larger for individuals who had been uninsured all year
extent than a typical insurance policy; on the other hand,
and smaller for those who spent only part of the year
the observed impact may have been dampened because
uninsured. The resulting amount of spending per
some of the people who gained eligibility for Medicaid
person was similar to that observed for privately
would have otherwise had private coverage and because
insured individuals.
some doctors do not accept Medicaid’s relatively low pay-
ment rates.
Other studies have tracked individuals from one year to
the next to see what happens when they gain or lose
Comparisons of Insured and Uninsured Populations.
insurance coverage. One recent study used MEPS data to
Other studies have used statistical methods to try to iso-
analyze the health care expenditures of insured nonelderly
late the effects of insurance when comparing the use of
adults.24 It found that individuals who were also insured
medical services by people who are insured and those
in the previous year had expenditures that were similar to
who are uninsured. Those studies attempt to identify and
those who had been previously uninsured—suggesting
adjust for other differences between the two populations
that spending for the uninsured will rise to the level seen
(such as income and health status) that would be
for the insured once they gain coverage. That approach
expected to influence their use of services. To the extent
has the advantage of avoiding the need to estimate how
that the insured and the uninsured differ in ways that are
much uncompensated care the uninsured receive. An
not observed in the data, however, those studies may not
important limitation of that approach, however, is that
have isolated the effect of insurance on the use of medical
individuals who became insured in the second year may
services. The studies themselves also vary along several
not be representative of the entire uninsured population;
dimensions. Some studies examine the use of services,
that is, they may have obtained coverage partly because of
and others analyze spending; some look at a cross-section
of insured and uninsured people, and others focus on
changes in coverage over time. In addition, the studies
23. Jack Hadley and others, “Covering the Uninsured in 2008: Cur-
may analyze different subsets of the nonelderly popula-
rent Costs, Sources of Payment, and Incremental Costs,” Health
Affairs, Web Exclusive (August 25, 2008), pp. W399–W415.
tion. (Because a large number of studies have made such
comparisons, this discussion highlights only a few of
24. Lisa Ward and Peter Franks, “Changes in Health Care Expendi-
them.)
ture Associated with Gaining or Losing Health Insurance,” Annals
of Internal Medicine, vol. 146, no. 11 (June 2007), pp. 768–774.
That study also found that individuals who lost insurance cover-
One recent example of a cross-sectional study compared
age had expenditures while uninsured that were comparable with
spending on health care for the insured and uninsured
those of continuously uninsured individuals.
CBO
CHAPTER THREE
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
75
a greater preference for medical care (or a greater need for
used 26 percent more health care services during that year
care) than those who remained uninsured in the second
than those who were uninsured for all of the second year.
year. In particular, some people may have been enrolled
More specifically, people in that group who were insured
in Medicaid when they were hospitalized or treated in an
for all of the second year used 88 percent as much care in
emergency room—in which case it was their higher
that year as those who were continuously insured in both
health care spending that caused them to become insured
years, whereas those who were uninsured for all of the
rather than their insurance causing them to use more
second year used 70 percent as much care as the continu-
health care.
ously insured. The possibility that some of those who
became insured in the second year obtained coverage
More generally, the results of such comparisons are sensi-
partly because of a greater preference for medical care (or
tive to the methodologies used and to the types of people
a greater perceived need for care) than those who were
included in the analysis. To examine that sensitivity,
uninsured in the second year also remains an issue with
CBO conducted its own analysis of MEPS data to track
that analytic approach. To the extent that such sorting
changes in insurance coverage from year to year. The
occurred, however, it would mean that the true effect of
analysis measured use of services rather than expendi-
gaining insurance coverage was smaller than the 26 per-
tures, included children and nonelderly adults, and
cent estimate.
excluded people who had public coverage.25 CBO con-
ducted the analysis separately for two groups of people—
Among people who were uninsured throughout the
those who were insured for only part of the first year and
entire first year, CBO found that those who were insured
those who were uninsured throughout the entire first
during all of the second year used 29 percent more ser-
year—and then compared the use of services in the fol-
vices during that year than people who were uninsured
lowing year by people in each group who were insured or
throughout the year. That is, those who gained coverage
uninsured during all of that year. Grouping people on the
in the second year used 67 percent as much care as people
basis of their coverage in the first year was designed to
who were continuously insured in both years, whereas
limit the extent to which people gaining or losing cover-
those who remained uninsured in the second year used
age in the second year differed with respect to their atti-
52 percent as much care as the continuously insured.
tudes toward health insurance and medical care. Compar-
Thus, in both analyses, people who gained coverage in
ing the use of services rather than expenditures holds
the second year increased their use of services, but they
aside any differences in the prices paid for the same ser-
did not use the same amount of care as people who had
vices—and the challenges of measuring uncompensated
been continuously insured in both years.
care—that could affect the results.26 One downside of
that approach, however, is that it could miss differences
Synthesizing the Evidence
in the types of services provided to insured and uninsured
Although the wide range of estimates generated by
individuals during a given visit to a physician or a hospi-
different studies makes it difficult to be certain about the
tal stay.
effects of gaining health insurance coverage on health care
use and spending, some central tendencies can be
Among people who were insured for only part of the first
observed. A 2005 review of the research literature in that
year, those who were insured for all of the second year
area analyzed studies using the strongest methodologies
and concluded that extending insurance coverage to the
uninsured would increase the number of physicians’ visits
25. CBO used MEPS data for 1997 to 2005 rather than data for a
by 30 percent to 50 percent for children and by 60 per-
single two-year period to expand the size of the samples and thus
cent to 100 percent for adults.27 The studies that were
increase the precision of the estimates. The data were adjusted for
reviewed obtained a broader range of estimates for the
age, sex, health status, education, and income so that the analysis
effects on children’s use of inpatient hospital care; among
compared individuals who appear similar in observed characteris-
tics other than insurance status.
adults, the estimated increases in hospital use ranged
26. CBO measured the use of health care services using a single index
that reflected the use of physicians’ and hospital services, weighted
27. Thomas C. Buchmueller and others, “The Effect of Health Insur-
by average expenditures for each kind of use. Thus, reported
ance on Medical Care Utilization and Implications for Insurance
payments for individual services did not affect the comparisons
Expansion: A Review of the Literature,” Medical Care Research and
between insured and uninsured individuals.
Review, vol. 62, no. 1 (February 2005), pp. 3–30.
CBO
76
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
from about 40 percent to 80 percent. That review
Those figures provide a framework for analysis, but the
encompassed studies evaluating Medicaid expansions and
effects of specific proposals will depend in part on the
research comparing insured and uninsured populations; it
extent of coverage that the uninsured receive; that is,
did not include the findings based on the RAND study
more extensive coverage would have a larger effect on
or some of the more recent research on Medicare.
health care use and spending. In addition, the impact of
proposals that did not achieve universal or near-universal
Examining the full range of studies also highlights the
coverage would depend on the extent to which the unin-
potential biases in each type of analysis. The findings
sured would be covered under a plan and on assumptions
based on the RAND experiment may have understated
about the underlying demand for health care among
the dampening effect of being uninsured on expenditures
those who would become insured. For more incremental
and thus underestimated the increase that would result
increases in insurance coverage rates, CBO would assume
from gaining insurance coverage. Conversely, studies rely-
that people who enrolled under a new program would
ing on comparisons of currently insured and uninsured
have a greater propensity to use medical services than
people, or on changes in their insurance coverage from
those who did not enroll. Depending on the design of
one year to the next, may overstate the true effect of gain-
such a proposal, those newly covered individuals might
ing coverage because those studies may not fully account
use health care services at a rate comparable with—or
for differences between the two groups in their attitudes
even greater than—that of people who have similar
toward health care and health insurance. Studies using
demographic characteristics and health status and are
natural experiments resulting from policy changes often
currently insured.
yield intermediate results, but those findings generally
reflect how the target populations would respond and
All else being equal, the increase in use of services by
thus may not apply to proposals that would achieve near-
previously uninsured individuals would also yield a corre-
universal coverage.
sponding increase in health care spending. In addition,
measured spending for the uninsured population would
Based on its review of the literature and analysis of health
rise as uncompensated care that they had received was
care data, CBO has adopted an intermediate range of
compensated by their insurance plan. Even so, the assess-
estimates. Specifically, the agency expects that providing
ment that the entire pool of uninsured individuals would
all of the uninsured with health insurance coverage
have somewhat lower use of services under a typical
equivalent to a typical employment-based plan would
employment-based policy means that they would have
increase their demand for medical services to a level that
somewhat lower total spending per person than those
is between 75 percent and 95 percent of the level of simi-
who are currently insured. The expected effect on spend-
lar people who are currently insured. (To the extent that
ing also depends on several other factors, however; in par-
the insured and uninsured populations differ in age and
ticular, the rates used to pay providers are an important
health status, CBO would make additional adjustments
consideration, and limits on the ability and willingness of
to account for the effects of those differences on health
health care providers to meet an increase in demand
care use and spending.) Relative to current amounts of
could affect both utilization rates and payment rates (see
health care use by the uninsured—for which estimates
Chapter 5 for more details).
average around 60 percent of the amount seen for insured
individuals—those assumptions reflect an increase of
between 25 percent and 60 percent that would result
from gaining insurance coverage. Compared with cur-
28. The uninsured currently account for about 8 percent of health
rently projected levels of health care use for the popula-
care use nationwide; a 25 percent increase in their use of services
tion as a whole, that rise in the demand for services would
would thus translate into a 2 percent increase for the country as a
constitute an increase that is between 2 percent and
whole, and a 60 percent increase would translate into a rise of
5 percent.28
nearly 5 percent in total use.
CBO
C H A P T E R
4
Proposals Affecting the
Choice of an Insurance Plan
Proposals could affect the options available to individ- Many of the considerations that arise in designing a new
uals when choosing a health insurance plan—and the
option for individuals to enroll in Medicare would also
incentives they face when making that choice—in a num-
affect the analysis of proposals to replace the current mix
ber of ways. For example, proposals could:
of private insurance and public programs for the
nonelderly population with a single-payer, Medicare-
B Establish or alter regulations governing the range of
for-all system.
premiums that may be charged or the terms under
which insurers may sell or renew coverage.
Regulating Insurance Premiums and
B Reveal more fully the relative costs of different health
Sales
insurance options by reducing or eliminating the cur-
Proposals may seek to create, remove, or modify regula-
rent tax subsidy for employment-based insurance,
tions governing health insurance markets in order to
encouraging or requiring the establishment of man-
make insurance more affordable for people with chronic
aged competition systems, or providing more readily
health problems or to provide consumers with more
accessible information about those costs.
choices, but those goals may conflict with one another.
For example, proposals could limit the extent to which
B Have the federal government create additional insur-
premiums for people in poor health can exceed those for
ance options, either by offering a new health insurance
people in better health (as some states currently do). Such
plan through the Medicare program or by providing
provisions would reduce premiums for individuals who
access to the health plans that are available to federal
have higher expected costs for health care, but they would
employees.
also raise premiums for healthier individuals and thus
could reduce their coverage rates. Other proposals might
The effects of those options would not just depend on the
give people greater choice among insurance plans—for
factors that affect the premium for a given insurance pol-
example, by allowing them to buy insurance across state
icy or on the share of the premium that enrollees have to
lines. That approach would counteract tight limits on
pay; those effects would also reflect the market dynamics
variations in premiums; that is, younger and relatively
that arise as individuals shift among coverage options and
healthy individuals living in states with such limits could
as policy premiums adjust to those shifts. In particular,
purchase a cheaper policy in another state that does not
the risk that some plans would experience “adverse selec-
regulate premiums, but older and less healthy residents
tion”—that is, that their enrollees will have above-average
who continued to purchase individual coverage in the
or higher-than-expected costs for health care—can have
tightly regulated states would probably face higher premi-
important implications for the operation of insurance
ums as a result. Those trade-offs stem from the proposals’
markets and for proposals that would regulate those oper-
differing effects on the composition of insurance pools.
ations or introduce new insurance options. To the extent
that proposals had an impact on average premiums, they
As those examples suggest, federal efforts to alter the reg-
would also affect the federal costs of any premium subsi-
ulation of insurance markets would also have to take into
dies as well as coverage rates and spending on health care.
account states’ current regulatory practices. Although
CBO
78
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
existing state regulations might limit the impact of any
stable, with each one willing to pay premiums to protect
federal initiatives, they also provide insights about the
against the risk of an accident or other incident someday.
manner and magnitude of effects that similar proposals at
the federal level would have on both premiums and insur-
A challenge facing health insurance markets is that indi-
ance coverage. The effects of existing regulations also
viduals have some ability to predict their future use of
indicate the likely impact of federal proposals that would
health care.2 In particular, a substantial minority of peo-
override or restrict state-level provisions.1
ple have at least one chronic health condition and are
likely to incur higher health care costs year after year,
Although federal legislation regulating insurance markets
whereas others can expect to have lower costs. According
could have substantial consequences for the operation of
to one recent study, about 30 percent of people ages 18 to
those markets, by itself that legislation would tend to
64 have at least one of seven chronic health problems
have a relatively small impact on the federal budget. An
(including heart disease, high blood pressure, diabetes,
exception could arise with proposals to expand high-risk
and asthma).3 Among that population, average health
pools, which subsidize premiums for individuals who are
care costs also tend to rise with age, at least partly reflect-
denied insurance coverage or face very high premiums
ing a higher prevalence of chronic conditions. In some
cases, those conditions may have only a limited impact
because of their health problems. The federal costs of
on the expected use of health care; for example, high
such proposals would depend on the scope of those
blood pressure was the most common chronic condition
subsidies, the way they were financed, and the number
examined in that study, and another recent study indi-
of people who took advantage of them.
cates that the average annual costs of treating that condi-
Background on Insurance Pooling
tion are relatively low.4 Other chronic conditions such as
Arrangements and Regulatory Structure
diabetes and heart disease are more expensive to treat,
however, and a smaller subset of the population has sev-
People purchase insurance policies to protect themselves
eral conditions simultaneously. Determining what share
financially against the risk of an expensive adverse event,
of health care spending is truly predictable from one year
such as a car accident, house fire, or serious health
to the next is difficult, but some experts suggest that the
problem. Insurance markets that work well do not simply
overall share probably exceeds 20 percent to 25 percent.5
shift that financial risk from individuals to insurers, how-
ever; such markets actually reduce that risk by pooling
The large share of health care spending that is unpredict-
policyholders together. Even if the cost of insuring a
able provides a strong incentive for most people to pur-
given individual may vary widely, the average cost of
chase health insurance. To the extent that health care
insuring a large group can be fairly stable and predict-
spending varies in predictable ways, however, insurance
able—making it less risky (and thus less costly) to offer
insurance. For example, if there are 100,000 subscribers
2. That challenge is not unique to health insurance; the likelihood of
to a given homeowners’ insurance policy, only a few will
an automobile accident or a homeowners’ insurance claim can also
submit claims in a given year; if the frequency of inci-
vary across policyholders in predictable and observable ways. See
Chapter 1 for additional discussion about the predictability of
dents that result in a claim remains stable, the growth in
health care spending.
the average claim and the policy premium over time will
reflect only the rising costs of repairing or replacing the
3. Catherine Hoffman and Karyn Schwartz, “Eroding Access Among
Nonelderly U.S. Adults with Chronic Conditions: Ten Years of
lost, stolen, or damaged goods that are covered by the
Change,” Health Affairs, Web Exclusive (July 22, 2008),
policy. Because homeowners are unlikely to know before
pp. W340–W348. The other chronic conditions examined were
purchasing insurance whether they will need to submit a
stroke, emphysema, and cancer.
major claim, the pool of policyholders is also likely to be
4. Kenneth E. Thorpe and others, “The Rising Prevalence of Treated
Disease: Effects on Private Health Insurance Spending,” Health
Affairs, Web Exclusive (June 27, 2005), pp. W5-317 to W5-325.
1. For additional analysis of issues related to the regulation of insur-
ance markets, see Congressional Budget Office, CBO’s Health
5. Joseph P. Newhouse, Melinda Beeuwkes Buntin, and John D.
Insurance Simulation Model: A Technical Description (October
Chapman, “Risk Adjustment and Medicare: Taking a Closer
2007); and The Price Sensitivity of Demand for Nongroup Health
Look,” Health Affairs, vol. 16, no. 5 (September/October 1997),
Insurance (August 2005).
pp. 26–43.
CBO
CHAPTER FOUR
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
79
coverage in general—or plans that offer more extensive
Though open to all employers, the option of bearing
benefits in particular—could attract enrollees with above-
insurance risk is generally taken by larger firms—those
average costs for health care. Those who expect to have
with enough employees to form a more certain estimate
lower expenditures, meanwhile, might wait until they
of the cost per enrollee. As a result of that distinction,
develop a health problem before purchasing coverage, or
policies for individuals and most small employers must
they might prefer a plan that offers less extensive coverage
comply with requirements that vary by state regarding the
and lower premiums. If a health plan experiences such
benefits they cover, the premiums that insurers may
adverse selection, its costs may exceed the premiums it
charge, and other terms of purchase. Insurance coverage
has charged; premiums could be raised, but that might
provided through larger firms, by contrast, typically faces
encourage relatively healthy enrollees to switch to other
few regulatory or legal constraints regarding its benefits
plans. In extreme cases, adverse selection could trigger a
and premiums.
spiral of rising premiums and declining enrollment that
leads to the plan’s demise.
One exception is that employers, regardless of size, can-
not charge different premiums for similarly situated
In practice, such spirals are rarely observed—in part
workers on the basis of health-related factors—that is, for
because insurers take steps to avoid them—but the
workers who are in the same class of employment and
potential problems caused by adverse selection are a more
work in the same geographic location, the employees’
common concern. The available evidence suggests that
contribution for health insurance cannot vary on the
roughly 20 percent of applicants for individually pur-
basis of such factors as health status, medical condition,
chased health insurance have expected costs for health
claims experience, or medical history. Reflecting a con-
cern that having a uniform contribution could lead to
care that are substantially above the average for their age
adverse selection, however, insurers have typically
group. Similarly, employers seeking to offer insurance
required that a substantial share of employees participate
coverage to their workers may differ substantially in terms
in an employer’s plan in order to help stabilize the average
of those workers’ average costs for health care, and they
cost of providing that coverage. One way employers seek
may be concerned that workers with above-average costs
to achieve that goal is by contributing a large share of the
will be more likely to enroll. For applicants with higher
total premium. In addition, employers typically allow
expected costs, health insurers face competitive pressures
their employees to sign up for health insurance only at
to charge higher premiums, limit coverage of preexisting
selected times—when they are first hired or during an
health problems, or deny coverage altogether to keep pre-
annual open-enrollment period—in order to mitigate
miums down while remaining profitable.
selection pressures.
To address concerns about the operations of markets for
Types of Regulations
private health insurance or to achieve other policy goals,
Many state and federal regulations have an impact on
policymakers have adopted various laws and regulations
insurance premiums, either directly or indirectly. Some
governing those markets, and over time some important
regulations affect premiums directly by restricting the
distinctions have arisen in that regulatory structure. State
amount by which they can vary or the factors that may be
governments are generally responsible for regulating the
used to adjust them. (Many states require certain benefits
business of insurance; as a result, any policy that individ-
to be covered by health insurance, which also affects
uals and firms buy from insurance companies is regulated
insurance premiums.) Other regulations can affect the
at the state level. In some cases, however, federal legisla-
cost of health insurance indirectly, through their impact
tion has established provisions that supersede or limit
on the composition of the insured population. Some
states’ efforts. In particular, federal law exempts from
regulations aim to guarantee the offer or renewal of insur-
state regulation any coverage that is offered by an
ance policies, a step that primarily affects people who
employer who chooses to bear the financial risk of
might not otherwise be offered coverage. In addition,
providing health insurance to its employees and their
provisions may limit or prohibit insurers from excluding
dependents; in those cases, the employer effectively serves
coverage for preexisting medical conditions—health
as the insurer. (For additional discussion of the Employee
problems that are present at the time of application.
Retirement Income Security Act, which established that
Many states have also established high-risk pools, which
exemption, see Box 1-1 on page 6.)
offer subsidized insurance to people who have been
CBO
80
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
denied coverage in the private market because of their
employer’s group plan, but under most circumstances the
health problems.6
new policy may not limit coverage for preexisting condi-
tions. The law, however, does not restrict the premium
Guaranteed Issue and Renewal. The federal government
that insurers may charge for new policies purchased in
and many states have taken various steps to require that
the individual market.
insurers offer coverage to applicants (a practice known as
guaranteed issue) and that they renew policies that are
HIPAA allows states to take additional steps to regulate
not delinquent (guaranteed renewal). The existing provi-
the portability of insurance, and many states have done
sions differ between the individual and small-group
so. For individuals who were not previously insured,
markets, however. The Health Insurance Portability and
however, states generally give insurers broad latitude to
Accountability Act (HIPAA) requires insurers that offer
exclude certain benefits or services from coverage in the
coverage to small businesses (those who have fewer than
individual market. Currently, 38 states permit health care
50 employees) to accept all applicants; before the enact-
services that are related to preexisting conditions to be
ment of that federal legislation in 1996, most states had
excluded from coverage permanently, and most states also
the same or similar requirements.
allow insurers to determine whether a condition was in
fact preexisting by examining more closely the medical
By contrast, only a handful of states currently require
history of enrollees when they submit a claim. Proposals
insurers in the individual insurance market to offer poli-
that limit the ability of insurers to exclude high-risk indi-
cies to all individuals and families who apply for coverage,
viduals and preexisting conditions from coverage might
and federal legislation does not generally mandate that
benefit less healthy individuals, who might not be offered
such offers be made. HIPAA prohibits insurers from
coverage otherwise, but the effects of those proposals on
failing to renew policies for health reasons, however,
insurance premiums would depend on the rules that
whether those policies are purchased in the individual
apply in each state.
market or by employers. Insurers may still terminate poli-
cies for fraud or failure to pay premiums, and they may
Direct Regulation of Premiums. All insurers—whether
also require that plans purchased by employers meet a
they cover health care, property, automobiles and their
participation requirement (for example, that a specified
drivers, or another type of risk—seek to set premiums so
percentage of employees remain enrolled in the plan).
that the aggregate payments will at least cover the
expected payouts for the policies they sell as well as the
Federal legislation has addressed in a more limited way
administrative and other costs they incur in providing
the question of guaranteed offers of coverage in the
insurance. Other things being equal, expected costs for
individual market and the related issue of whether new
health insurance are higher for older people and for
policies may exclude coverage for preexisting medical
people with more, or more serious, health problems. In
conditions—steps designed to increase the portability of
theory, that relationship could yield premiums for indi-
insurance coverage. Specifically, HIPAA essentially
vidually purchased coverage that vary widely, with some
requires insurers to offer coverage to anyone who had
enrollees paying many multiples of the average quote for
held insurance through a previous job but was losing or
a given policy to reflect their higher expected costs for
had recently lost that coverage (for example, because he
health care.
or she changed jobs). The requirements differ somewhat
depending on whether the new coverage is purchased in
In practice, however, premiums in the individual insur-
the individual market or comes through the new
ance market do not vary as widely as do individuals’
expected costs for health care, for several reasons. First,
6. Many other laws and regulations govern health insurance but are
insurers may find it difficult or costly to obtain informa-
beyond the scope of this report. State insurance agencies are
tion about each applicant’s health status, so assessments
generally charged with monitoring the financial health of insur-
of the applicant’s expected costs (a practice known as
ance firms to ensure that they will be able to meet their promises
“medical underwriting”) are far from perfect. Second, to
to pay claims. Furthermore, many of those agencies regulate the
the extent that underwriting efforts are successful, insur-
sales practices of insurers. Federal law also establishes reporting
ers tend to limit coverage for or screen out applicants who
and disclosure requirements and fiduciary standards for the plans’
administrators. All of those regulations can also affect insurance
have preexisting health problems that are costly to treat.
premiums and coverage.
According to a 2005 study, about 70 percent of appli-
CBO
CHAPTER FOUR
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
81
cants for individual coverage are quoted a standard rate
Regulations may also affect the extent to which premi-
based only on their age; about 20 percent are either
ums can be changed over time. In the individual market,
charged a higher premium (generally not exceeding twice
states generally preclude the practice—sometimes called
the standard rate for their age group) or are sold a modi-
“re-underwriting” or experience rating—of adjusting a
fied package that does not cover treatments for their pre-
particular enrollee’s premium on the basis of his or her
existing health conditions (at least for some period of
insurance claims or changes in health status after purchas-
ing the policy. Thus, premiums for a given policy would
time); and about 10 percent are denied coverage.7 Some
generally increase over time to reflect higher expected
applicants are charged a premium that is only modestly
costs for health care on average, but they do not vary
higher than the standard rate, so the share of applicants
across individuals to reflect updated estimates of each
that are either charged a substantially higher premium or
one’s expected health costs. Insurers could circumvent
denied coverage is probably on the order of 20 percent.
those restrictions, however, by raising premiums for all
enrollees in an existing policy and simultaneously offer-
A third reason that premiums in the individual market
ing a new, cheaper product whose applicants would be
vary less than do enrollees’ expected health care costs is
subject to underwriting. That practice would tend to
the states’ regulation of those premiums, which takes var-
discourage individuals who had developed expensive
ious forms. Many states restrict premium “rating”—that
health conditions after enrolling in the original policy
is, they directly limit the extent to which premiums are
from changing plans, so they would pay the new, higher
allowed to vary according to the age or health status of
premium for that policy. It is not clear how common that
enrollees. The specific restrictions vary widely, however,
practice is, however.
in ways that differ between the individual and small-
group markets. According to one survey of states’ prac-
Premiums charged to small employers may be somewhat
less volatile than are premiums in the individual market,
tices in the individual insurance market, three states
for several reasons. First, those premiums reflect the
require pure community rating of premiums, meaning
average costs of their enrollees, so high expected costs for
that insurers may vary premiums for a given policy only
one person would be spread across all enrollees. Second,
by the size of the enrolling family and their place of resi-
insurance is regulated more extensively in the small-group
dence within the state.8 Six other states allow adjusted
market than in the individual market. According to a
community rating, meaning that health insurance
2003 survey, 35 states employed rating bands in the
premiums are allowed to vary by family size and residence
small-group market, 10 used adjusted community rating,
as well as by age and sex—but not by health status.
2 used pure community rating, and only 3 states and the
Twelve states apply rating bands that allow premiums to
District of Columbia chose not to regulate rates offered
vary on the basis of age and sex but prohibit insurers from
to small firms.9 Some states also limit the degree to which
deviating from the standard rate by more than a specified
premiums for small employers can increase from one year
percentage for reasons relating to health.
to the next to reflect enrollees’ costs or changes in their
health status (for example, permitting no more than a
15 percent adjustment for those reasons). In other states,
7. See Mark Merlis, Fundamentals of Underwriting in the Nongroup
however, high health care costs for an employee or a
Health Insurance Market: Access to Coverage and Options for Reform,
dependent in one year can lead to substantial increases in
NHPF Background Paper (Washington, D.C.: National Health
Policy Forum, April 13, 2005). In principle, insurers could charge
the average premium charged to the employer in the
a higher premium to applicants who have very high expected
following year, and lower-than-expected claims can lead
costs, but in practice they appear to assume that individuals who
to corresponding reductions in premiums.
would be willing to pay premiums exceeding twice the standard
rate would be likely to have even higher covered costs for health
The overall effect of those state regulations is generally to
care—so rather than charge a very high premium, insurers gener-
compress the range of premiums offered. Although insur-
ally deny coverage to such applicants instead.
ers could comply with a rating band by reducing the
8. Ibid. A recent analysis also found that in three states, a dominant
insurer used community rating even though the state did not
require all insurers to adopt that practice; see Congressional
9. General Accounting Office, Private Health Insurance: Federal and
Budget Office, The Price Sensitivity of Demand for Nongroup
State Requirements Affecting Coverage Offered by Small Businesses,
Health Insurance, Background Paper (August 2005).
GAO-03-1133 (September 2003).
CBO
82
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
premiums charged to the least healthy enrollees or
As of 2007, about 200,000 people were enrolled in high-
groups, they could also satisfy those regulations by raising
risk pools nationwide—about half of that total came
their standard rates. In practice, they appear to do some
from five states—so those enrollees account for about
of both, and rating restrictions have been found to
2 percent of the approximately 10 million nonelderly
increase premiums for healthier enrollees, decrease them
people who purchase health insurance in the individual
for sicker enrollees, and to raise average premiums (pri-
market.
marily because of the resulting increase in enrollment of
predictably higher-cost individuals).10 The net impact of
High-risk pools obviously reduce the health insurance
regulation of premiums on the number of people who
premiums that their enrollees pay, but covering those
have insurance coverage is difficult to predict in the
high-cost individuals separately could also lower premi-
abstract because some people face increases in premiums
ums for other purchasers because it would reduce the
and others face decreases.
average costs of the remaining enrollees. The strength of
that ripple effect on premiums depends on the extent to
High-Risk Pools. Another approach to reducing health
which premiums are allowed to vary within the state. At
insurance premiums is to separate people with the highest
one extreme, if no rating restrictions were in place and all
health risks from the rest of the pool and partially subsi-
enrollees were charged a premium exactly in accordance
dize their coverage. High-risk pools, as they are called, are
with their own expected expenses—or if high-risk appli-
a mechanism employed in varied forms by more than 30
cants had been denied coverage—then establishing a new
states, primarily to assist individuals who are unable to
pool for those with the highest expected costs would have
obtain health insurance for medical reasons. Typically,
no effect on the premiums of other policyholders. In a
community-rated state, by contrast, separating high risks
such individuals must apply for private insurance and be
could reduce premiums for the remaining enrollees in
denied coverage or be quoted a high premium before they
rough proportion to the share of covered costs that high-
can enroll in the pool. Enrollees are then charged a pre-
risk enrollees had generated. In states with rating bands,
mium that usually ranges between 125 percent and
the likely effect would fall between those extremes; reduc-
150 percent of the standard rate for their age group.
tions in the costs of covering high-risk enrollees could
make the bands less constraining and thus could lead
Those premiums are generally insufficient to cover those
insurers to reduce their standard rates.
enrollees’ costs for health care, however, so high-risk
pools require subsidies to remain solvent (typically aver-
Effects of Proposals on Insurance Markets
aging several thousand dollars per enrollee). To limit the
Proposals to change the regulations governing insurance
cost of those subsidies, states may cap enrollment in high-
markets would generally have modest effects on the fed-
risk pools. As of 2007, however, all states with pools but
eral budget, and many of them would entail trade-offs
one (Florida) appeared to be accepting new applicants.11
between reducing average policy premiums and making
In many cases, the costs of subsidizing high-risk pools are
insurance less expensive for individuals with health prob-
financed by an assessment or tax on other health insur-
lems. Although generalizing about the precise effects of
ance policies sold in the state; in recent years, the federal
such proposals is difficult because their content might
government has also provided some financial assistance to
vary substantially, some indication of the likely magni-
defray the costs of starting and operating high-risk pools.
tudes of budgetary effects and changes in insurance pre-
miums and coverage can be gleaned from the Congressio-
10. See M. Susan Marquis and Stephen H. Long, “Effects of ‘Second
nal Budget Office’s recent analysis of legislative proposals
Generation’ Small Group Health Insurance Market Reforms,
to modify state regulations or to allow individuals to buy
1993 to 1997,” Inquiry, vol. 38, no. 4 (Winter 2001/2002),
insurance across state lines. In addition, some quantita-
pp. 365–380; and Amy Davidoff, Linda Blumberg, and Len
tive or qualitative information can be provided to help
Nichols, “State Health Insurance Market Reforms and Access to
Insurance for High Risk Employees,” Journal of Health Economics,
illustrate the potential effects of or key considerations sur-
vol. 24, no. 4 (July 2005), pp. 725–750.
rounding proposals for which CBO has not previously
generated a cost estimate.
11. Information on the status of high-risk pools comes from
www.statehealthfacts.org. See also Bernadette Fernandez, Health
Insurance: State High-Risk Pools, RL31745 (Congressional
The Health Insurance Marketplace Modernization and
Research Service, October 1, 2008).
Affordability Act of 2006 is one example of a proposal
CBO
CHAPTER FOUR
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
83
affecting the regulation of insurance markets that CBO
employment-based plans, which would occur because
has analyzed.12 That legislation would have created a
individually purchased insurance would become relatively
more uniform set of regulatory standards for the individ-
attractive (especially to people with lower expected health
ual and small-group health insurance markets—standards
care costs). The increase in Medicaid spending would
that would have fallen somewhere between the strictest
reflect the net impact of an increase in spending for
and most lenient state regulations currently in place.
people who would lose private coverage and a decrease in
CBO estimated that those changes would decrease the
spending for those who would gain it. Overall, CBO esti-
average premium paid by policyholders in those markets
mated that the legislation would not have a substantial
by 2 percent to 3 percent, primarily by overriding some
effect on the number of people who have health insur-
benefit mandates and reducing costs that insurers incur in
ance because the number who would gain coverage
complying with varying state rules. The legislation would
(including previously uninsured people who would pur-
have increased insurance coverage by about 600,000 peo-
chase coverage in the individual market) would roughly
ple, on net, but it would have tended to increase premi-
offset the number who lost it.
ums (and thus reduce coverage) for people with health
problems.
CBO’s previous estimates of federal proposals to add new
regulatory requirements also indicate the important influ-
CBO also estimated the budgetary impact of that legisla-
ence that existing state practices have on those estimates.
tion, concluding that it would increase federal revenues
For example, the effect of the requirement under HIPAA
by about $3 billion over 10 years and would reduce fed-
to guarantee renewal of insurance policies was judged to
eral spending for Medicaid by about $1 billion over that
be limited because nearly all states already had such a
period. The increase in revenues would reflect a net
requirement in place. Similarly, CBO estimated that
reduction in spending on employment-based health
HIPAA’s requirement for portability of insurance from
insurance (stemming from the decline in average premi-
group to individual coverage would have a relatively small
ums). Reflecting CBO’s assumption that total compensa-
effect on insurance premiums in the individual market.
tion would not change, that development would shift
Although insurers would have to offer coverage to rela-
some compensation from a form that is tax-preferred
tively unhealthy individuals who would otherwise have
(health insurance premiums) to a form that is taxable
been turned down, CBO estimated that in most cases the
(wages and salaries). Because employment-based insur-
premiums for those policies could be set to reflect the
ance would become somewhat less expensive under the
expected costs for health care for those enrollees and thus
proposal, some people who would be covered by Medic-
would not have a substantial effect on premiums for
aid under current law would switch to private coverage
other enrollees.15
and federal Medicaid spending would decline.
Rather than add or remove regulations, the federal gov-
Alternatively, proposals could allow individuals to avoid
ernment could seek to affect the operation of insurance
the requirements set in their home state by purchasing
markets by offering additional subsidies for high-risk
insurance across state lines. In particular, that approach
would allow individuals who are relatively healthy and
13. A similar approach would facilitate the formation of association
live in states that regulate insurance more extensively to
health plans, which can be offered by trade, industry, or profes-
purchase a less expensive policy.13 CBO analyzed one
sional associations to their member firms. That option would be
proposal to allow cross-state purchasing of insurance—
attractive for smaller firms with relatively healthy workers that are
located in states that regulate premiums more extensively or have
the Health Care Choice Act of 2005—and concluded
more extensive benefit mandates. For an analysis of a recent
that over 10 years it would increase federal revenues by
legislative proposal, see Congressional Budget Office, cost esti-
about $13 billion and federal spending for Medicaid
mate for H.R. 525, Small Business Health Fairness Act of 2005
by about $1 billion.14 The increase in revenues would
(April 8, 2005).
result largely from a reduction of about 1 million in the
14. Congressional Budget Office, cost estimate for H.R. 2355, Health
number of people who receive health insurance through
Care Choice Act of 2005 (September 12, 2005).
15. See Statement of Joseph Antos, Assistant Director for Health and
12. Congressional Budget Office, cost estimate for S. 1955, the
Human Resources, Congressional Budget Office, before the Sub-
Health Insurance Marketplace Modernization and Affordability
committee on Civil Service, House Committee on Government
Act of 2006 (May 3, 2006).
Reform and Oversight, October 8, 1997.
CBO
84
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
pools. The costs of such proposals and their effects on
Revealing the Relative Costs of
coverage rates and premiums would depend primarily on
Health Plans
the following factors:
Most Americans with health insurance are shielded
from—or may not be aware of—the price of their cover-
B The number of individuals who would be eligible for
age, either in absolute terms or relative to other options.
and enrolled in those pools;
Many employers pay a large share of the premium for
their workers; even though employees as a group ulti-
B The scope of the insurance coverage they would
mately bear that cost, they may not know its magnitude.
receive;
Moreover, the tax code subsidizes employment-based
health insurance by excluding the employer’s contribu-
B The premiums they would have to pay themselves;
and
tions to the premium from the employee’s taxable wages
and income; in most cases, the employee’s contribution is
B The mechanism used to subsidize the difference
also excluded. Those features encourage people to have
between enrollees’ costs for covered health care ser-
insurance coverage, but they also lead workers to buy
vices and those premium payments.
more extensive insurance than they would if they faced
the full price of their policy; those features also may limit
Because nearly all states with high-risk pools are accept-
the extent of price competition in the insurance market.
ing new applicants, there may not be substantial unmet
Some proposals would make consumers bear the cost of
demand in those states given the coverage and premiums
their health insurance more directly, either by paying the
they currently feature (although additional subsidies
full cost themselves or by paying the added cost of more
could encourage more active efforts by states to enroll
expensive policies. Proposals could achieve that goal by:
eligible individuals). Lower premiums for enrollees and
more extensive coverage would generate higher enroll-
B Reducing or eliminating the current tax subsidy for
ment but would also increase subsidy payments and make
employment-based insurance, perhaps replacing it
it more likely that individuals who would have been
with a tax credit or some other fixed-dollar subsidy (an
insured otherwise would switch into the high-risk pool.
approach discussed in Chapter 2); or
The financing of subsidies for high-risk pools raises a
B Establishing a managed competition system, in which
number of issues. Larger federal subsidies could lead
a range of plans is offered and the employer’s or the
more states to create high-risk pools and could encourage
government’s contribution to the premium is a fixed
states to expand existing pools, but they could also cause
amount—for example, the premium of the average
some substitution of federal funds for existing state funds.
plan or the least expensive plan available—thus requir-
Proposals might also address whether payments would be
ing consumers to pay the additional cost of more
made to states that currently require guaranteed issue and
expensive plans.
use community rating or narrow rating bands in the indi-
Those approaches—taken separately or in combina-
vidual market; residents of those states might never meet
tion—would provide stronger incentives for enrollees to
the eligibility terms for a high-risk pool. Payments could
weigh the expected benefits and costs of policies when
be made to those states in an effort to reduce premiums
making their decisions about purchasing insurance. As a
in the individual market, but doing so would raise the
result, enrollees would generally choose health insurance
cost of the proposal. More generally, the impact of a pro-
policies that were less extensive, less expensive, or both,
posal on the federal budget would depend on whether
compared with the choices made under current law. A
and to what extent the costs of the subsidy payments were
related option would be to give workers more readily
shared between the federal and state governments; a
accessible information about the full costs of their cover-
higher federal share would encourage states to participate
age, including the employer’s contribution. Whether and
but would also reduce the incentive for them to control
how that information might affect their choice of a health
the pool’s costs.
plan is less clear, however.
CBO
CHAPTER FOUR
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
85
Reducing or Eliminating the Tax Exclusion
and payroll tax rates are largely determined at the federal
The current tax treatment of health insurance premiums
level—so the rules are similar across all states at any given
constitutes a relatively large subsidy—known as a tax
time. Although federal tax rates have changed over time,
expenditure—for the purchase of employment-based
many other aspects of the health care system and the
insurance, amounting to $145 billion in forgone federal
national economy have simultaneously changed, making
income taxes and $101 billion in forgone federal payroll
it difficult to separate cause and effect when comparing
taxes in 2007.16 Individuals living in states that have
one period with another. As a consequence of those
income taxes receive an additional subsidy because those
methodological challenges, the findings of older studies
states generally follow federal definitions of taxable
using aggregate data on tax rates and insurance premiums
income and thus exclude the costs of employment-based
vary widely, depending on the period they examined and
health insurance as well. The total tax subsidy averages
the assumptions they made.
about 30 percent and generally ranges from about 20 per-
cent to 40 percent of the premium for most workers,
Two recent studies have attempted to address those meth-
depending on their tax bracket and state of residence.17
odological issues more carefully, but some concerns
remain about using their results to estimate the impact of
Although the subsidy provides an incentive to purchase
eliminating the tax exclusion. A 2004 study by Gruber
insurance—and to do so through one’s employer—it also
and Lettau examined how employers’ spending on health
encourages people to buy policies that are more extensive
insurance varied across states with different tax structures,
or more expensive than they would purchase otherwise.
exploiting the fact that state income tax rates changed at
Reducing or eliminating that exclusion thus could have a
different times (and did so in ways that were not caused
large effect on insurance premiums and coverage because
by trends in health insurance).19 Extrapolating from
those results, they estimated that eliminating the tax
it could substantially increase the effective price of any
exclusion for health insurance premiums—which in the
given policy—by 25 percent for someone who had been
sample that they studied would increase the effective
receiving a 20 percent subsidy and by two-thirds for
price of health insurance by 58 percent, on average—
someone who had been receiving a 40 percent subsidy.18
would yield a 29 percent reduction in health care spend-
(The impact of such changes on whether people purchase
ing by employers who continued to offer coverage. In
insurance is discussed in Chapter 2.)
other words, the reduction in those employers’ contribu-
Relevant Studies.
tions would be about half as large (in percentage terms) as
Several studies have attempted to quan-
the increase in the effective price facing enrollees.
tify how removing or limiting the favorable tax treatment
for employment-based insurance would affect insurance
Gruber and Lettau’s paper improved substantially on ear-
coverage, insurance premiums, and total spending on
lier work by better isolating the effect of the net price of
health care. Ideally, a study would compare systemwide
health insurance on premiums, but it still has limitations.
outcomes with and without those tax preferences, hold-
In particular, their estimate is based on relatively small
ing all other factors equal. In practice, however, that type
differences in state tax rates, and extrapolating the effects
of comparison cannot be readily made because income
of those differences could overstate the impact of larger
changes. One way that employers could reduce premiums
16. Joint Committee on Taxation, Tax Expenditures for Health Care,
would be to limit the extent of the coverage they offer
JCX-66-08 (July 30, 2008).
(for example, by increasing cost-sharing requirements).
17. One offsetting consideration is that excluding health insurance
But that approach would also heighten the variability of
premiums from taxable wages reduces future Social Security bene-
health costs for employees, and workers might become
fits, which are based on average earnings, at the same time that it
increasingly reluctant to accept higher levels of cost
reduces payroll tax payments.
sharing as their degree of financial risk grew. At the same
18. Assume, for example, that an insurance policy has a total premium
time, more rigorous management efforts by health plans
of $5,000. Someone receiving a 20 percent tax subsidy would thus
(or shifts in enrollment toward more tightly managed
pay $4,000 on net. If the tax subsidy was eliminated, that person
would pay $5,000, or 25 percent more. Someone receiving a
40 percent tax subsidy would currently pay $3,000 for that policy.
19. Jonathan Gruber and Michael Lettau, “How Elastic Is the Firm’s
If the tax subsidy was eliminated, that person would pay $5,000,
Demand for Health Insurance?” Journal of Public Economics,
or 67 percent more.
vol. 88, no. 7 (July 2004), pp. 1273–1294.
CBO
86
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
plans) would yield somewhat lower premiums, but more
in a firm; to the extent that their study did not fully
substantial reductions might become increasingly diffi-
account for those differences, caution must be used in
cult to achieve. In other words, existing differences in
extrapolating their results to a broader population.
employers’ contributions across states could largely reflect
the use of cost-control options that represent the “low-
CBO’s Assessment. Reflecting the limitations of those two
hanging fruit.”
studies, CBO’s assessment is that removing the tax prefer-
ence would have a smaller effect on the level of premiums
Another limitation of the study is that it includes the
that individuals choose. Specifically, CBO estimates that
impact of employers changing the share of the premium
a 50 percent increase in the price of health insurance, all
they pay in response to different tax rates. In that case,
else being equal, would lead people to select plans with
employees would see their contributions rise but the total
premiums that are between 15 percent and 20 percent
premium for their coverage would not change. Even with
lower than the premiums they would pay under current
that effect included, the impact of changes in tax rates
law. Reaching that point would probably take several
that the study found barely meets the standard threshold
years, as health plans, employers, and enrollees adjusted
for statistical significance—that is, the odds of getting
their offerings and choices. A portion of that ultimate
their results by pure chance (assuming that the true effect
decrease in premiums would come from reductions in the
of the tax exclusion was zero) were only slightly less than
extent of coverage that enrollees purchased (that is, fewer
one in twenty. Gruber and Lettau estimated, on the basis
benefits covered or higher cost-sharing requirements),
of other studies, that reductions in the share of the
and the remainder would come from choosing plans
premium that employers cover would account for about
that exercise tighter management over the use of health
one-fourth of the effect on employers’ spending that they
care (that is, plans might have more features typical of
report. But if that component was removed, the remain-
health maintenance organizations such as utilization
ing effect they found might not meet a test of statistical
review, restricted provider networks, or gatekeeper
significance.
requirements).
A more recent study by Heim and Lurie avoided some of
The effect of a specific policy proposal would depend pri-
those methodological problems but was based on a rela-
marily on what changes it made in the tax treatment of
tively small segment of the population that may not be
health insurance. Removing the exclusion of premiums
representative. The study analyzed spending on health
from income and payroll taxation would increase the
insurance premiums for self-employed individuals, who
after-tax price of health insurance by roughly 50 percent,
were able to deduct a growing proportion of their premi-
on average, for people currently covered by employment-
ums from their taxable income over time.20 Their results,
based insurance. Removing the exclusion only for income
which were similar to Gruber and Lettau’s estimate,
tax purposes (keeping the payroll tax exclusion in place)
imply that the reduction in premiums that would result
would raise the average price by roughly 30 percent,
from scaling back the tax exclusion for health insurance
which would ultimately yield health insurance premiums
would be about half as large as the resulting price
that are 9 percent to 12 percent lower. In both cases, the
increase; that is, an increase of about 50 percent in the
reduction in overall spending on health care would be
net price of health insurance would lead people to choose
smaller than the reduction in premiums because some
policies with premiums that were about 25 percent lower
costs would be shifted from covered spending to out-of-
than otherwise. An advantage of their study is that it
pocket spending.
accounts for the full effect on insurance premiums rather
than the impact on employers’ contributions, because in
Alternatively, proposals could cap the amount of pre-
their study the employer and the employee are the same
mium payments that may be excluded from workers’ tax-
person. The self-employed, however, may differ in both
able income—the effects of which would depend criti-
observable and unobservable ways from people who work
cally on the level at which the cap was set. Workers whose
premiums exceeded the cap by a substantial margin
would have strong incentives to switch to a less expensive
20. Bradley T. Heim and Ithai Lurie, “Do Increased Premium Subsi-
plan. Workers whose premiums fell below the cap,
dies Affect How Much Health Insurance Is Purchased? Evidence
from the Self-Employed” (draft, Department of Treasury, Office
however, would not be affected, so the overall impact on
of Tax Analysis, January 7, 2008).
premiums would generally be smaller. One objective of
CBO
CHAPTER FOUR
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
87
capping the exclusion might be to target employees who
difference in premiums across plans (although that effect
have relatively extensive insurance coverage and, as a
would be muted if enrollees could continue to exclude
result, above-average premiums. Workers who reside in
their own premium payments from taxation). Sponsors
areas with higher-than-average medical costs or whose
would give enrollees comparative information about their
firms have higher premiums because their covered work-
options. Some versions of managed competition would
force is older or in poorer health could also be affected by
also involve standardizing the benefits offered—to a
a fixed-dollar cap, however, even if the generosity of their
greater or lesser degree—in order to foster stronger price
health plan was not above average.
competition. In addition, sponsors could adjust pay-
ments to health plans to account for differences in the
The effects of reducing, eliminating, or capping the
health status of their enrollees (in an effort to limit the
exclusion for employment-based insurance would also
impact of those differences on the plans’ premiums).
depend on a number of issues relating to implementa-
tion. Insurers and employers would have to report to
Background. Most employers do not use the principles of
both employees and the Internal Revenue Service the
managed competition to purchase health insurance bene-
amount of premiums subject to tax. However, calculating
fits for their employees. Indeed, surveys indicate that
the average premium and allocating those costs among
most firms that offer health insurance do not give their
employees could be difficult, particularly for large
employees a choice of health plans. That statistic is some-
employers whose plans cover employees’ expenses for
what misleading, however, because most firms have few
health care as they are incurred (in which case timely data
employees. Large firms are much more likely than small
may not be available). Limiting or eliminating the exclu-
firms to offer a choice of plans, and they also account for
sion would also create incentives for employers to misrep-
the majority of workers. Consequently, about 57 percent
resent benefits as company overhead or to reallocate costs
of workers who are offered insurance have a choice of
among subsidiaries so as to reduce their employees’ tax
plans. In the case of firms that do not offer their workers
liability. (Those considerations would affect the pro-
a choice of plans, health plans still compete on the basis
posal’s impact on revenues as well as the incentives for
of their price and value but do so in an effort to be chosen
workers to choose less expensive policies.)
by the employer. For small employers in particular, the
administrative costs of offering several competing plans
Another source of uncertainty is whether the 41 states
and the potential problems of adverse selection that could
(and the District of Columbia) that have their own
arise may outweigh the benefits of giving their employees
income tax would continue to follow the federal lead in
more options.
the tax treatment of premiums for employment-based
coverage. If, instead, some states took action to maintain
Even among firms offering a choice of plans, fixed-dollar
the full exclusion of premiums from taxable income, the
contributions to employees’ insurance premiums—
incentive for workers to choose a less expensive plan
another key feature of managed competition—are less
would be smaller. The extent of that difference would
common than fixed-percentage contributions. A 2002
depend on the number of states that did not conform
survey found that among Fortune 500 companies (which
their tax systems to mirror the federal tax change and on
generally offer their employees a choice of plans), only
the tax rate structure in those states.
about one-quarter took the fixed-dollar approach.22 The
following example illustrates the incentives created by
Establishing a Managed Competition System
each approach. Suppose that an employer makes two
The term “managed competition” refers to a purchasing
plans available—one with a total premium of $4,000 per
strategy that seeks to create stronger incentives for con-
sumers to be cost-conscious in their choice of health
21. See Alain C. Enthoven, “The History and Principles of
plans and for plans to compete more intensely on the
Managed Competition,” Health Affairs, vol. 12 (Supplement
basis of premiums and quality of care.21 Under that
1993), pp. 24–48.
approach, a sponsor—such as an employer or govern-
22. James Maxwell and Peter Temin, “Managed Competition Versus
ment agency—would offer a choice of health plans and
Industrial Purchasing of Health Care Among the Fortune 500,”
would make a fixed-dollar contribution toward the cost
Journal of Health Politics, Policy, and Law, vol. 27, no. 1 (2002),
of insurance. Enrollees would thus bear the cost of any
pp. 5–30.
CBO
88
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
year and one with a premium of $5,000. If that employer
tion systems, all enrollees in a given health plan would
pays 80 percent of the total premium for each plan, an
typically pay the same premium—so if payments to plans
employee who chooses the more costly plan pays an
were not adjusted, plans that attracted less healthy mem-
additional $200 (20 percent of the $1,000 difference in
bers would have higher premiums as a result.24 Because
premiums between the two plans). Under a managed
enrollees would have strong financial incentives to switch
competition system, however, the employer would con-
out of those plans, the adoption of managed competition
tribute the same amount to both plans (for example, 80
could trigger an “adverse selection spiral” for plans offer-
percent of the average premium, or $3,600). Employees
ing the most extensive coverage or doing little to manage
would face the full $1,000 price difference between the
benefits. In fact, some employers that implemented a
two plans and would therefore have a much stronger
managed competition system dropped such plans as their
incentive to choose the lower-cost plan. Making employ-
premiums skyrocketed and their enrollments plum-
ees pay the full difference in premiums could also stimu-
meted.25 (Health plans might also drop out of a managed
late greater competition among insurance plans to keep
competition system for other reasons that make them
their premiums down. (Whether enrollees actually faced
broadly unpopular with enrollees, such as being poorly
that full difference would also depend on whether their
run.)
premium payments were tax-preferred.)
In principle, adjusting the sponsors’ payments to plans to
Some proposals that are based on the principles of man-
account for expected differences in their enrollees’ health
aged competition would require health plans to offer a
care costs would limit the impact of adverse selection. If
standard benefit package. In principle, standardizing
those adjustments worked well, the premiums that
benefits would promote competition among health plans
enrollees faced would vary across plans because of differ-
by making it easier for consumers to compare their
ences in the value of their benefits or the efficiency of
options; that step would also help prevent plans from
their operation, but not because of differences in their
structuring their benefit packages to attract enrollees who
mix of enrollees. Government programs currently use risk
are less likely to use medical care (which could in turn
adjustment in cases in which private health plans com-
reduce the plan’s premiums and thus distort the compari-
pete against a government-administered option (as with
son of plans). In practice, however, some aspects of health
Medicare Advantage plans or Medicaid HMOs) and
benefits are easier to standardize than others. For exam-
against one another to deliver program benefits (as with
ple, specifying uniform levels of cost sharing is relatively
the prescription drug plans in Medicare).
straightforward, but other aspects—such as definitions of
covered services and utilization review procedures—can
In practice, however, risk-adjustment methods are impre-
affect a consumer’s ability to use certain benefits and are
cise, so fully offsetting the effects of enrollees’ characteris-
more to difficult to standardize.23 Moreover, having stan-
tics on a plan’s premium may not be feasible. Those
dard benefits has two disadvantages. First, by limiting
methods do not need to account for all differences in
consumers’ options, standardization would make some
health care spending across enrollees to be effective;
people worse off (specifically, those who would prefer a
indeed, comparisons of predicted spending using risk-
different design). Second, rigid standardization could
adjustment models with actual spending will inevitably
prevent health plans from developing innovative designs
find some enrollees who used more care than was
that might lead to more efficient delivery of care.
expected and some who used less. What matters is
Another important design issue is whether the sponsor’s
24. Under a managed competition system, insurers could be allowed
payments to insurers would vary to reflect differences in
to vary individuals’ premiums so that the premiums reflected each
expected health care costs for different enrollees—a pro-
enrollee’s expected costs for health care, in which case those premi-
cess known as risk adjustment. Under managed competi-
ums would already be adjusted for risk. In many respects, such an
arrangement would resemble the current market for individually
purchased insurance.
23. For a discussion of this issue, see Mark McClellan and Sontine
Kalba, “Benefit Diversity in Medicare: Choice, Competition, and
25. David M. Cutler and Sarah J. Reber, “Paying for Health Insur-
Selection,” in Richard Kronick and Joy de Beyer, eds., Medicare
ance: The Trade-Off Between Competition and Adverse Selec-
HMOs: Making Them Work for the Chronically Ill (Chicago:
tion,” Quarterly Journal of Economics, vol. 113, no. 2 (May 1998),
Health Administration Press, 1999), pp. 133–160.
pp. 433–466.
CBO
CHAPTER FOUR
KEY ISSUES IN ANALYZING MAJOR HEALTH INSURANCE PROPOSALS
89
accounting for the predictable differences in spending
total health insurance premiums (the amount paid by
that might affect an enrollee’s choice of a health plan or a
employers and employees combined) by 5 per
