October 24, 2008 | WebMemo on Health Care
Presidential candidate Senator John McCain (R-AZ) has put forth an ambitious health care plan.The plan proposes:
Analyzing proposals based on campaign documents and media accounts is inherently difficult, as they lack the level of detail necessary for a rigorous econometric analysis. Nonetheless, several organizations have done so, using a variety of assumptions and methodologies. Most notable are The Lewin Group, Health Systems Innovations Network, The Urban Institute-Brookings Institution Tax Policy Center, and academic research published in Health Affairs.
The best independent research shows that the McCain plan would cover roughly half of the 45 million uninsured through an expansion of private coverage and cost-savings would be achieved by reforming perverse incentives in the health sector. However, as proposed, the plan would require considerable increases in federal expenditures.
Coverage. According to the Lewin Group, the McCain plan would reduce the number of uninsured by 21.1 million people in 2010 if fully implemented in that year. The plan would bring about significant shifts in sources of coverage. While 9.4 million people would lose employer coverage, 23.9 million would enroll in non-group coverage and 5.8 million would enroll in new high-risk pools. Public coverage would decrease by 5.4 million as people transitioned into private insurance. Altogether, according to Lewin, private coverage would increase by 26.5 million people.
Lewin applied a type of model known as a micro-simulation. Health Systems Innovations Network (HSI) conducted an analysis (funded by the McCain campaign) that used this type of model as well. HSI found that the McCain plan would reduce the uninsured by 27.5 million. In contrast to Lewin, HSI found that enrollment in employer coverage would actually rise, not fall, under the McCain plan.
Studies from the Tax Policy Center (TPC), as well as those published in Health Affairs, applied a different type of model known as an elasticity-based approach. The results were dramatically different. The TPC estimated that the McCain plan would reduce the number of uninsured by 1.3 million in 2009. In that year, 3.3 million people would lose employer coverage while 4.7 million would enroll in private non-group coverage.
The TPC did not incorporate key components of the McCain plan that would impact coverage, such as enrollment in new high-risk pools. Nor did the TPC model account for savings potential from such measures as the ability of insurers to sell plans across state lines, which would make insurance more affordable. These are significant omissions. The Congressional Budget Office, for example, estimates that state health insurance regulations increase premiums an average of 13 percent. A recent academic study found that allowing for interstate commerce in health insurance alone could reduce the uninsured by approximately 12 million.
Thomas Buchmueller of the University of Michigan and his colleagues published a widely cited analysis in Health Affairs estimating that elimination of the income tax exclusion for employer-sponsored insurance would cause 20 million Americans to lose such coverage. They also estimated that about 21 million would take up non-group coverage, resulting in a net increase in coverage of about 1 million people.
In order to arrive at these estimates, Buchmueller and colleagues took elasticity projections from previous studies and then made a "middle-range" assumption based on these earlier works. Like the TPC, the authors did not take into account key components of the McCain plan, including the ability to purchase insurance across state lines that would have direct implications for coverage.
In projecting coverage estimates, the authors also left out a significant effect of the McCain plan. As MIT economist Jonathan Gruber explains, health insurance expenditures come entirely from workers' wages:
Both economic theory and a large body of economic evidence show that there are no employer dollars: the money that employers spend on insurance would otherwise just be spent on worker wages. If MIT stopped offering insurance, over a several year period my wages would rise by $10,000 to offset the lost insurance compensation.
In other words, under normal labor market conditions, workers who lost employer coverage would see their wages rise, increasing their financial capacity to buy health insurance. Buchmueller and colleagues failed to account for these labor market dynamics.
Cost.According to the Lewin Group, under the McCain plan, health care system-wide savings over the 2010-19 period would be about $432.6 billion. Over half of these savings would result from converting the current income tax exclusion to a tax credit--which would offer new incentives to find more affordable coverage--and from allowing insurers to sell plans across state lines. As Professor Mark Pauly of the University of Pennsylvania explains, cost containment through tax reform--which is central to McCain's plan--is more realistic than cost containment through health care delivery reforms, such as the greater use of health IT and preventive care, common to Obama and McCain's plans:
The main problem is that these [common methods] are "if only" savings, which can be achieved "if only" certain events would occur, such as physicians' being willing to adopt health IT, consumers being willing to accept changes in diet and exercise. ... In contrast, cost savings from limiting the exclusion only require the government to take action.
There is little evidence that there are known methods to cause the "if only" behavior to occur...There is evidence that high and growing premiums fueled by tax subsidies to the upper middle class affect the premiums for all private insurance, and thus make it unaffordable for lower-income people.
Lewin, HSI, and TPC all found that spending by the federal government would, on net, have to increase significantly in order to implement McCain's plan. Lewin projected that the McCain proposal would increase federal spending by about $2.05 trillion over the 2010-19 period. Most of the federal spending increase would result from the tax credits being more generous than the income tax exclusion they would be replacing--most people would receive much more of a tax benefit from the tax credits than they currently receive under the income tax exclusion. In fact, Lewin projects an average savings of $1,411 per family under the McCain plan. The credits would also make federal tax breaks for health insurance less regressive and more equitable, in sharp contrast to the current system, which strongly favors higher-income groups. According to Lewin, the credits would cost the government $4.15 trillion while the elimination of the income tax exclusion would result in only $1.89 trillion in revenue.
Other estimates reached different conclusions. HSI projected a one-year cost impact of $287 billion, while the Tax Policy Center estimated that the McCain plan would cost $1.31 trillion over the 2009-18 period. However, as previously stated, the TPC did not include savings measures in their modeling.
McCain's tax credit proposal is not budget neutral. Nevertheless, several adjustments could move it closer to budget neutrality, as the campaign states it would be. Lewin projected that if the payroll tax exclusion (as well as the income tax exclusion) for insurance benefits were eliminated, the net federal cost over 2010-19 would decrease to $935 billion, but the total reduction in the number of uninsured would also drop by 3.4 million. Another option would be to restructure the credits from a flat to a progressive tax credit or voucher-type system, with more assistance going to lower-income persons and less going to upper-income persons. Like Obama, McCain has proposed cost-saving efficiencies within Medicare and Medicaid to offset federal expenditures while preserving promised levels of benefits for individuals in the programs.
Expanding Coverage and Realigning Perverse Incentives
The McCain plan would reduce the number of uninsured and help control costs, yet it would, as currently designed, require considerable increases in federal expenditures. The coverage expansion would be driven by enrollment in individual private plans chosen by the consumer that could be taken from job to job. Cost-savings measures would result from an effective realignment of financial incentives. Primarily because the new tax credit would be much larger than the existing tax break from the income tax exclusion that it would be replacing, the plan would amount to a significant tax reduction, particularly for the middle class. Costs of the McCain plan could be controlled by replacing or capping the federal payroll tax exclusion for health insurance or by changing the structure of the tax credits, moving it toward budget neutrality.
Greg D'Angelo is Policy Analyst in the Center for Health Policy Studies and Paul L. Winfree is a Policy Analyst in the Center for Data Analysis at The Heritage Foundation. Jeet Guram, a Heritage health policy intern from the University of South Carolina, contributed to the research in this paper.
McCain-Palin 2008, "The Truth about the McCain-Palin Health Care
Plan," at http://www.johnmccain.com/Informing/Issues/19ba2f1c
-c03f-4ac2-8cd5-5cf2edb527cf.htm (October 23, 2008); for an analytical discussion of the McCain health plan, see Robert E. Moffit, Ph.D., and Nina Owcharenko, "The McCain Health Plan: More Power to Families," Heritage Foundation Backgrounder No. 2198, October 15, 2008, at www.heritage.org/research/healthcare/bg2198.cfm.
Kevin Sack, "On Health Plans, the Numbers Fly," The New York
Times, October 21, 2008, at http://www.nytimes.com
/2008/10/22/us/politics/22health.html?hp (October 23, 2008).
Lewin Group, "McCain and Obama Health Care Policies: Costs and
Coverage Compared," October 8, 2008, at http://www.lewin.com/content/Files/The_Lewin_Group_McCain-
Obama_Health_Reform_Report_and_Appendix.pdf (October 23, 2008)
 Roger Feldman, Lisa Tomai, and Sally Duran, "Impact of Barack Obama 2008 Health Reform Proposal," HSI Network, August 21, 2008, at /static/reportimages/A104B72267B9BFC1D364935BD69CA3EC.pdf (October 23, 2008).
Burman et al., "An Updated Analysis of the 2008 Presidential
Candidates' Tax Plans: Revised August 15, 2008," Tax Policy Center,
updated September 12, 2008, at http://www.taxpolicycenter.org/Uplo
adedPDF/411749_updated_candidates.pdf (October 23, 2008).
Joseph Antos, Gail Wilensky, and Hanns Kuttner,
"The Obama Plan: More Regulation, Unsustainable Spending," Health
Affairs, September 16, 2008, at http://content.healthaffairs.org/cgi/content/fu
ll/hlthaff.27.6.w462/DC1 (October 23, 2008).
 For an analysis of the non-group market, see Mark V. Pauly and Bradley Herring, "Risk Pooling and Regulation: Policy and Reality in Today's Individual Health Insurance Market," Health Affairs, Vol. 26, No. 3 (2007), pp. 770-779, at http://content.healthaffairs.org/cgi/reprint/26/3/770 (October 23, 2008); Mark V. Pauly and Robert D. Lieberthal, "How Risky Is Individual Health Insurance?" Health Affairs, Vol. 27, No. 3 (2008), w242-w249, at http://content.healthaffairs.org/cgi/reprint/27/3/w242 (October 23, 2008).
 Also, since the McCain plan would permit states to combine Medicaid funds with the new tax credit in order to purchase private insurance for program enrollees, 11.2 million people would drop traditional Medicaid/SCHIP coverage and 12.0 million would enroll in private coverage, bringing the net effect of the plan to a decrease in the number of uninsured of 21.1 million people.
 A micro-simulation is based on a utility-maximizing model of how individuals choose insurance plans, allowing for variation on a range of variables, including price and income. This type of modeling combines individual characteristics and behavior to estimate how each person, or subgroup, within a population would react to a policy change. However, micro-simulations face the possibility of selection bias, as they make a series of assumptions based on a small sub-population and then apply them to a larger group. Lewin used their Health Benefit Simulation Model (HBSM), a micro-simulation of the U.S. health care system developed by Lewin in 1989 and continuously applied and refined ever since. Lewin's work has been deemed "the gold standard of independent health-care analysis." See "A Liberal Supermajority," The Wall Street Journal, October 17, 2008, at http://online.wsj.com/article/SB122420205889842989.html (October 23, 2008). For a detailed description of the HBSM, see the Lewin Group, "Summary Description of the Health Benefits Simulation Model (HBSM)," January 29, 2007, at /static/reportimages/C0F8E652DE7821F78863A19E775D77DB.pdf (October 24, 2008).
 For a detailed description of the HSI micro-simulation model, see Roger Feldman et al., "Health Savings Accounts: Early Estimates of National Take-Up," Health Affairs, Vol. 24, No. 6 (2005), pp. 1582-1591, at http://content.healthaffairs.org/cgi/reprint/24/6/1582 (October 24, 2008).
 An elasticity model applies the same measure of responsiveness (how consumers respond to changes in variables, such as insurance cost) to an entire population, but different groups within that population may have different levels of responsiveness to changes in the variable(s) of interest. A micro-simulation allows for variation among different individuals and groups. Like a micro-simulation, an elasticity-based approach also faces the possibility of selection bias, as the population to which the elasticity is applied may vary systematically from the population within which it is established. For example, there may be significant differences between consumers in the non-group market today and individuals who would lose group coverage and enter this market under a policy change.
 E-mail communication with Surachai Khitatrakun, an author of the Tax Policy Center report, on October 15, 2008.
 Congressional Budget Office, Increasing Small-Firm Health Insurance Coverage Through Association Health Plans And Healthmarts, January 2000, at /static/reportimages/D7E782457B2237ECAB9984D7269FDB6E.pdf (October 24, 2008).
Stephen T. Parente et al., "Consumer Response to a National
Marketplace for Individual Insurance," Carlson School of
Management, University of Minnesota, June 28, 2008, at /static/reportimages/F9690BFFF0D90D97E780B092DA07E3DD.pdf
(October 23, 2008). Additionally, as John Goodman explains (in
response to the claim that the McCain plan would have no real
effect on the uninsured): "The authors are claiming that 40 million
people will turn down the opportunity to have at least $5,000 worth
of private insurance at no cost to themselves! Even if a $5,000
plan is less generous that what is normal in the group market, the
claim that 40 million people would choose to be uninsured instead
is prima facie not plausible." John Goodman, "Response to a
Critique of the McCain Health Care Plan," John Goodman's Health
Policy Blog, September 24, 2008, at http://www.john-goodman-blog.com/response
-to-a-critique-of-the-mccain-health-care-plan/ (October 24, 2008).
Thomas Buchmueller et al., "Cost and Coverage Implications of the
McCain Plan to Restructure Health Insurance," Health
Affairs, September 16, 2008, at http://content.healthaffairs.org/cgi/cont
ent/abstract/hlthaff.27.6.w472v1 (October 24, 2008).
Jonathan Gruber, Statement Before the Finance Committee, U.S.
Senate, July 31, 2008, at http://finance.senate.gov/
hearings/testimony/2008test/073108jgtest.pdf (October 23, 2008). Emphasis in original.
Other analysts have explained that the vast swing from the employer
system that the authors predict is not likely to happen quickly or
at that magnitude. For example, see Thomas Miller, "Health Plan
Scoring That Runs Out of Bounds," at http://healthaffairs.org/blog/2008/09/25/health-plan-
scoring-that-runs-out-of-bounds/ (October 23, 2008).
Mark V. Pauly, "Blending Better Ingredients
for Health Reform," Health Affairs, Vol. 27, No. 6 (2008), pp.
w482-w491, at http://content.he
althaffairs.org/cgi/content/full/hlthaff.27.6.w482/DC1 (October 23, 2008).
 According to Lewin, the tax credits would cost $4.2 trillion, the Medicaid/SCHIP opt-out would cost $151 billion, the Guaranteed Access Plans would cost $235 billion, and there would be a total of $2.5 trillion in offsets.
 The conclusion that the McCain health care tax credits would be, overall, more generous than the income tax exclusion they would be replacing is a key point of agreement among Lewin, TPC, and HSI. TPC estimates an average tax savings of $1,241, and HSI estimates that less than 5 percent of Americans not in public insurance programs would possibly face a tax increase. Also, see Anna Wilde Matthews, "Health-Care Fixes: Plan vs. Plan," The Wall Street Journal, October 23, 2008, at http://online.wsj.com/article/SB122470511751259445.html (October 24, 2008).
Joint Committee on Taxation, Tax
Expenditures for Health Care, July 30, 2008, p. 5, at /static/reportimages/AB758D1AFD79AF63B4122F21505DF3E8.pdf
(October 23, 2008). See also Thomas M. Selden and Bradley M. Gray,
"Tax Subsidies for Employment-Related Health Insurance: Estimates
for 2006," Health Affairs, Vol. 25, No. 6 (November/December
2006), pp. 1568-1579, at http://conte
nt.healthaffairs.org/cgi/content/abstract/25/6/1568 (October 23, 2008); John Shiels and Randall Haught, "The Cost of Tax-Exempt Health Benefits in 2004," Health Affairs, February 25, 2004, pp. W4-106-W4-112, at http://content.health
affairs.org/cgi/content/full/hlthaff.w4.106v1/DC1 (October 13, 2008).
 HSI estimated the one-year cost of the Guaranteed Access Plans to be $19 billion, while TPC estimates the 10-year cost to be on the order of$1 trillion.
"Transcript: The Vice-Presidential Debate," The New York
Times, October 2, 2008, at http://elections.nytimes.com/20
presidential-debate.html (October 23, 2008).
Stuart M. Butler, "New Research Both Candidates Should Study,"
The New York Times, October 17, 2008, at http://campaignstops.blog
study/#more-535 (October 23, 2008).
As McCain's senior policy advisor Douglas Holtz-Eakin explains,
"It's about giving them [entitlement enrollees] the benefit package
that has been promised to them by law at lower cost." Laura
Meckler, "McCain Plans Federal Health Cuts," The Wall Street
Journal, October 6, 2008, at http://online.wsj.com/article/SB122315
505846605217.html?mod=relevancy (October 23, 2008).