RETIREMENT RESEARCH June 2011, Number 11-8 DOES MEDICARE PART D PROTECT THE ELDERLY FROM FINANCIAL RISK? By Gary V. Engelhardt and Jonathan Gruber* Introduction The Medicare Modernization Act of 2003 added the ates Part D’s impact using the 2002-5 and 2007 waves Part D prescription drug benefit to the Medicare of the Medical Expenditure Panel Survey (MEPS) before program. This addition, which became effective in and right after the program’s implementation. 2006, increased Medicare program costs by more The brief is organized as follows. The first section than 10 percent in order to provide, for the first time, presents background on Part D. The second section prescription drug coverage to enrollees. Part D has describes the MEPS data. The third section presents since enrolled a sizeable share of elders and now pays the results of Part D’s effect on prescription drug cov- for a large percentage of their prescriptions. Despite erage and expenditures and offers a tentative assess- the program’s size and importance, however, little is ment of the program’s overall social impact. The final known about its effectiveness. One way to measure section concludes that Part D has resulted in substan- its success is to determine to what extent it provides tial crowd out of both coverage and expenditures and, financial security to elders. If Part D covers prescrip- as of 2007, has produced only modest benefits. tion drug spending that was putting older Americans at financial risk, it may result in large social gains. If it simply substitutes for – or “crowds out” – exist- Medicare Part D Program ing insurance arrangements, the social gains may be much smaller. Beyond a crowd-out analysis, a Medicare, which provides universal health insurance full evaluation of Part D also needs to consider other coverage to people over age 65 and to those on the social benefits and costs, such as the potential health federal Disability Insurance program, was established benefits and the efficiency costs of subsidizing drug in 1965. The original program covered most medi- coverage. The study summarized in this brief evalu- cal needs for the elderly and disabled, but excluded * Gary Engelhardt is a professor of economics at Syracuse University. Jonathan Gruber is a professor of economics at MIT. Both authors are research associates of the Center for Retirement Research at Boston College. This brief was adapted from a longer paper (Engelhardt and Gruber, 2011 forthcoming). The research reported herein was supported by the CRR pursu- ant to a grant from the U.S. Social Security Administration funded as part of the Retirement Research Consortium. Gruber also acknowledges financial support from the National Institute on Aging. The opinions and conclusions are solely those of the authors and should not be construed as representing the opinions or policy of the Social Security Administration, any agency of the federal government, the CRR, MIT, NBER, or Syracuse University. The authors are grateful to John Graves for excellent research assistance. 2 Center for Retirement Research prescription drug coverage. However, in the 1990s, Impact of Part D on Coverage the advancement of prescription drug treatments for common illnesses among the elderly drew attention and Expenditures to this coverage gap. In 2003, the Bush Administra- tion and Congress agreed on a far-reaching prescrip- The analysis addresses three separate questions. tion drug benefit package at a projected cost to the First, did Part D help reduce financial risk by increas- federal government of $40 billion per year for its first ing prescription drug coverage among the elderly? 10 years.1 Second, did Part D increase total spending on pre- The resulting benefit – known as Medicare Part scription drugs and help ease the spending burden D – is delivered by private insurers under contract on individuals? Third, considering both benefits and with the government. Beneficiaries can choose from costs, what is the initial net social impact of Medicare three types of insurance: 1) stand-alone plans, called Part D? Medicare Prescription Drug Plans, that just offer prescription drug benefits; 2) Medicare Advantage Coverage plans, which provide all Medicare benefits (including prescription drugs); or 3) the beneficiaries’ current The first step is to examine trends in prescription employer/union plan (for those offered such cover- drug insurance coverage for older Americans. Figure age), as long as coverage is at least as generous as the 1, which compares coverage before and after Part D standard Part D plan.2 enactment, shows the age profile of coverage from The small literature that has emerged on the any source for 50-80 year-olds. Prior to Part D enact- Medicare Part D program has primarily investigated ment, coverage rates from any source were constant two issues. The first is the determinants and efficacy at about 75 percent until age 65, before dropping by of decisions to enroll in the program and which plan about 5 percentage points. After Part D, the age pro- to choose. The second evaluates the impacts of the file is similar through the early 60s before diverging plan on prescription drug utilization. Few studies sharply at age 65 as coverage rates jump to as high address the issue of how Part D has affected financial as 90 percent. This shift over only a one- to two-year security.3 period is remarkable. The Data Figure 1. Prescription Drug Coverage Rates from any Source Before and After Part D Enactment,* This study uses the 2002-5 and 2007 waves of the Individuals Ages 50-80 MEPS, which is comprised of a nationally represen- tative set of respondents drawn from the National Health Interview Survey (NHIS). The MEPS is a two- 100% year overlapping panel focused on health insurance coverage, health care utilization, and expenditure, 75% and is used to construct data for the National Health Accounts. For each calendar year of the survey, the sample 50% is a combination of individuals in their first year of the panel and individuals in their final year of the panel. Interviews are conducted three times per year. 25% Before enactment, 2002-2005 The analysis summarized in this brief uses variables After enactment, 2007 measured as of the end of each calendar year (i.e., 0% from the last interview of the year). It excludes 2006 50 55 60 65 70 75 80 because that was a transition year between private Age coverage and public coverage for many.4 * “Before” covers 2002-2005 and “after” covers 2007. Source: Engelhardt and Gruber (2011 forthcoming). Issue in Brief 3 Figure 2 shows how the source of coverage age for those individuals age 65 and over. Putting the changes from private to public by age after Part D two estimates together (12.3/40.8) implies very large enactment.5 The public coverage rate in 2007 was less crowd out of 70 percent, which means that the vast than 10 percent for individuals under age 65. Public majority of those who signed up for Part D moved coverage then jumps dramatically to 70 percent or over from another source of coverage (see Figure 3).7 more for those 65 and older. This age-related in- crease in public coverage is much larger than the total increase in insurance coverage (as shown in Figure 1, Figure 3. Estimate of the Crowd-Out Effect of a gap between the before-and-after lines of roughly 15 Public Prescription Drug Coverage of the percentage points), suggesting significant crowd out Elderly, Age 65 and Over of existing coverage by the Part D expansion. Gained coverage, Figure 2. Percentage of Individuals Ages 50-80 30.1% with Public Prescription Drug Coverage After Part D Enactment, 2007 100% 75% Previously covered, 50% 69.9% 25% Source: Engelhardt and Gruber (2011 forthcoming). 0% 50 55 60 65 70 75 80 Age Expenditures Source: Engelhardt and Gruber (2011 forthcoming). The next step is to examine the impact of Part D on prescription drug spending, which is interesting for two reasons. First, it extends the crowd-out analy- To confirm that other factors that could affect sis to look more specifically at the dollar reduction drug coverage during this period are not responsible in spending covered by private insurance relative to for the apparent crowd out, regression analysis is the dollar increase in public spending. Second, using required. The analysis summarized here compares the spending data allows for a direct assessment of prescription drug coverage and utilization between the extent to which public insurance increased the the near-elderly (ages 60-64) and the elderly (age 65 financial protection of the elderly through reduced and over). The results are similar across numerous out-of-pocket costs. specifications, including defining near elderly as ages The study again turns to regression analysis to 50-64, defining elderly as ages 65-70, and controlling estimate how changes in public prescription drug for demographics, region, health status, and income.6 spending for the elderly influence changes in total The estimates show a very large increase in drug spending. The resulting estimates indicate that prescription drug coverage from 2002-05 to 2007 for Part D crowds out substantial spending by both pri- those individuals age 65 and over with only a moder- vate insurers and individuals (see Figure 4 on the next ate corresponding increase for 60-64-year-olds. Over- page). Private insurance spending falls by more than all, Part D was associated with a 12.3-percentage-point 42 cents for every dollar increase in public spending. rise in prescription drug coverage among the elderly. Out-of-pocket spending by the elderly falls by about The corresponding calculation for public coverage 34 cents per public dollar spent.8 Thus, the estimates shows a rise of 40.8 percentage points in public cover- suggest that each dollar of public expenditure raises 4 Center for Retirement Research Figure 4. Estimate of the Crowd-Out Effect of Overall Social Impact of Part D Public Prescription Drug Spending of the Elderly, Age 65 and Over While the regression results indicate a very high degree of crowd out for both participation and expen- ditures, a full evaluation of the program requires a New public Reduced private broad assessment of its net social benefits. On the spending, 24.5% insurance benefit side, as noted in the previous section, Part D spending, 41.9% could help beneficiaries by reducing the risk associ- ated with prescription drug spending. It could also, perhaps, reduce spending needs on health care ser- vices other than prescription drugs. On the cost side, Part D – like any tax-funded program – involves a loss in economic efficiency from redirecting resources from the private market to a public activity. Another cost is the potential moral hazard associated with Reduced out-of-pocket the program; as drugs become relatively cheaper for spending, 33.6% individuals, they could be less careful with how much they consume so that they end up spending more on Source: Engelhardt and Gruber (2011 forthcoming). drugs that may not offer compelling health benefits. While calculating the costs and benefits is difficult and requires many simplifying assumptions, this total expenditure by 25 cents, or that there is about 75 study’s findings, which are detailed in the full paper, percent crowd out. This result is strikingly similar to suggest that both the potential direct gains and the the coverage crowd-out estimate above. costs are relatively modest.9 Despite the large crowd-out estimate, Part D may still provide valuable risk protection to the elderly. However, a closer examination of the data shows that Conclusion – for the median participant – Part D is associated with a reduction in only $180 in out-of-pocket spend- This study finds that the introduction of Medicare ing. For those with very high out-of-pocket spending Part D is associated with substantial crowd out of – the 90th percentile – the reduction is $800. To put both prescription drug coverage and expenditures. these numbers into context, the baseline value for An initial assessment of the broader social impact of out-of-pocket spending is $2,500. the program suggests that both its benefits and costs are modest. However, a comprehensive evaluation of the program requires further research, including an analysis of whether Part D is associated with gains in health that would both improve individuals’ quality of life and potentially lower health spending on non- prescription drug services. Issue in Brief 5 Endnotes 1 Congressional Budget Office (2002). 9 These findings are subject to a number of caveats. On the one hand, these welfare calculations may 2 In this case, the plan sponsor would receive a overstate the gains from Part D if other consumption- Retiree Drug Subsidy from the government. smoothing mechanisms are available to the elderly, such as private income transfers, own savings, or 3 Two studies that do address this issue are Lichten- uncompensated medical care. On the other hand, the berg and Sun (2007); and Levy and Weir (2009). gains from Part D may be understated because the calculations were based on an annual, rather than a 4 For a full discussion of the specific MEPS data used lifetime, measure of expenditure risk. In particular, in this analysis, see Engelhardt and Gruber (2011 there is some evidence that lifetime medical spending forthcoming). risk is greater than annual risk, because out-of-pocket expenditures are highly persistent over time (see 5 “Public” means either through Medicare or Med- Feenberg and Skinner, 1994; and French and Jones, icaid. The definition of public coverage and expen- 2004). Finally, the welfare calculations were predi- diture used in this study includes two important cated on the assumption that individuals do not value caveats. The first is how to classify prescription drug any improvements in health associated with increased coverage through Medicare HMOs. Before the imple- prescription drug spending. To the extent that Part D mentation of Part D in 2006, many, but not all, indi- is associated with health gains and they are valued, the viduals who were enrolled in Medicare HMO plans estimates will understate the true gains from Part D. received prescription drug coverage, which actually was a mix of private and public coverage. This study classifies Medicare HMOs as private in the pre-period and public in the post-period. The second is that the Medicare Modernization Act that created Part D gives subsidies to employers/unions to keep coverage under the Retiree Drug Subsidy program. Therefore, some government-funded insurance will be classified as “private” coverage under the definition used in this study. 6 See the full paper for details (Engelhardt and Gruber, 2011 forthcoming). 7 If coverage through a Medicare HMO prior to 2006 is treated as public coverage in the pre-period, then the estimate of the crowd-out effect drops to 70 percent. 8 The results also showed that there is little change in the odds that individuals fill a prescription. On the other hand, the number of prescriptions filled per enrollee went up astronomically, by four prescriptions per new enrollee, or 14 percent of the pre-period aver- age for those over age 65. 6 Center for Retirement Research References Congressional Budget Office. 2002. Issues in Designing a Prescription Drug Benefit for Medicare. Washing- ton, DC: U.S. Government Printing Office. Engelhardt, Gary V. and Jonathan Gruber. 2011 (forthcoming). “Medicare Part D and the Finan- cial Protection of the Elderly.” American Economic Journal: Economic Policy. Feenberg, Daniel and Jonathan Skinner. 1994. “The Risk and Duration of Catastrophic Health Expen- ditures.” Review of Economics and Statistics 76: 663-647. French, Eric and John Bailey Jones. 2004. “On the Distribution and Dynamics of Health Care Costs.” Journal of Applied Econometrics 19: 705-721. Levy, Helen and David Weir. 2009. “Take-up of Medi- care Part D: Results from the Health and Retire- ment Study.” Working Paper #14692. Cambridge, MA: National Bureau of Economic Research. Lichtenberg, Frank and Shawn Sun. 2007. “The Im- pact of Medicare Part D on Prescription Drug Use by the Elderly.” Health Affairs 26: 1735-1744. Centers for Disease Control and Prevention. National Health Interview Survey, Medical Expenditure Panel Survey, 2002-2005 and 2007. Washington, DC: U.S. Department of Health and Human Services. RETIREMENT RESEARCH About the Center Affiliated Institutions The Center for Retirement Research at Boston The Brookings Institution College was established in 1998 through a grant from Massachusetts Institute of Technology the Social Security Administration. The Center’s Syracuse University mission is to produce first-class research and forge Urban Institute a strong link between the academic community and decision-makers in the public and private sectors around an issue of critical importance to the nation’s Contact Information Center for Retirement Research future. To achieve this mission, the Center sponsors Boston College a wide variety of research projects, transmits new Hovey House findings to a broad audience, trains new scholars, and 140 Commonwealth Avenue broadens access to valuable data sources. Since its Chestnut Hill, MA 02467-3808 inception, the Center has established a reputation as Phone: (617) 552-1762 an authoritative source of information on all major Fax: (617) 552-0191 aspects of the retirement income debate. E-mail: crr@bc.edu Website: http://crr.bc.edu © 2011, by Trustees of Boston College, Center for Retire- The research reported herein was performed pursuant to ment Research. All rights reserved. Short sections of text, a grant from the U.S. Social Security Administration (SSA) not to exceed two paragraphs, may be quoted without ex- funded as part of the Retirement Research Consortium. The plicit permission provided that the authors are identified and opinions and conclusions expressed are solely those of the full credit, including copyright notice, is given to Trustees of authors and do not represent the opinions or policy of SSA, Boston College, Center for Retirement Research. any agency of the federal government, or the Center for Retirement Research at Boston College.