- Issue: Medicare Advantage (MA) enrollment has grown significantly since 2009, despite legislation that reduced what Medicare pays these plans to provide care to enrollees. MA payments, on average, now approach parity with costs in traditional Medicare.
- Goal: Examine changes in per enrollee costs between 2009 and 2014 to better understand how MA plans have continued to thrive even as payments decreased.
- Methods: Analysis of Medicare data on MA plan bids, net of rebates.
- Findings: While spending per beneficiary in traditional Medicare rose 5.0 percent between 2009 and 2014, MA payment benchmarks rose 1.5 percent and payment to plans decreased by 0.7 percent. Plans’ expected per enrollee costs grew 2.6 percent. Plans where payment rates decreased generally had slower growth in their expected costs. HMOs, which saw their payments decline the most, had the slowest expected cost growth.
- Conclusions: In general, MA plans responded to lower payment by containing costs. By preserving most of the margin between Medicare payments and their bids in the form of rebates, they could continue to offer additional benefits to attract enrollees. The magnitude of this response varied by geographic area and plan type. Despite this slower growth in expected per enrollee costs, greater efficiencies by MA plans may still be achievable.
Proponents of a larger role for private plans in Medicare argue that they are likely to be more efficient than the government-sponsored program, and that competition among such plans will increase the choices available to beneficiaries and control costs. But for most of its history, Medicare payment policies have not provided private plans with strong incentives to achieve efficiency. Under the Medicare Modernization Act of 2003, payment levels for private plans in most counties exceeded per beneficiary costs in traditional Medicare.1
The Affordable Care Act of 2010 sharply reduced the growth of per enrollee payments to private plans, now known as Medicare Advantage (MA) plans, narrowing the discrepancy between those payment rates and per beneficiary spending in traditional Medicare.2 As a result, the ratio of per enrollee payments to MA plans to projected spending per beneficiary in traditional Medicare, which was 114 percent in 2009, fell to 100 percent by 2017 (Exhibit 1).3,4,5 This reduction is particularly striking in the context of slow growth in traditional Medicare; between 2009 and 2017, spending per beneficiary in traditional Medicare is estimated to have grown at an annual rate of only 1.6 percent.
The decline in payments to MA plans led to predictions of a decline in private-plan enrollment, as had occurred in the late 1990s.6,7 However, MA plan enrollment has continued to grow rapidly, from 10.5 million in 2009 to 18.5 million in 2017, with private plans serving about one-third of all beneficiaries.8
How have Medicare Advantage plans continued to thrive in recent years despite less generous payment? Little is known about how they have controlled their costs in response to decreased payment. In this issue brief, we examine changes in Medicare per enrollee payments to plans and their expected per enrollee costs across counties and types of plans, using data from the Centers for Medicare and Medicaid Services (CMS) from 2009 and 2014 (see How We Conducted This Study).9
Changes in Medicare Advantage Plan Payments, 2009–2014
Per enrollee payments to MA plans are determined by two parameters: the benchmark rate set for each county, determined by a statutory formula that reflects the county’s traditional Medicare spending per beneficiary, and the bid submitted by each plan. If a plan’s bid is less than the benchmark rate, it receives its bid plus an additional rebate (equal to a proportion of the difference between its bid and the benchmark) as total payment per enrollee. If a plan’s bid is greater than the benchmark rate, its total payment per enrollee is limited to the benchmark rate.10
These parameters changed considerably between 2009 and 2014. The benchmark rates are now set at 95 percent, 100 percent, 107.5 percent, or 115 percent of traditional Medicare spending per beneficiary, depending on the county’s costliness (i.e., Medicare spending per beneficiary). The rebate has been reduced from 75 percent to 50 percent of the difference between the benchmark and the plan’s bid, and the benchmark and/or rebate also may vary depending on the plan’s quality rating. These changes have reduced the growth of per enrollee payments to MA plans and shifted a greater share of payments to high-performing plans.11
How Have Plans Responded?
Using data from 2009 and 2014, we examine how Medicare per enrollee payments to plans and plans’ bids (which generally represent their expected total costs per enrollee of providing Medicare Part A and Part B benefits, including medical expenditures, administrative costs, and a predetermined profit rate) changed over that period.12,13 In the aggregate, while projected spending per beneficiary in traditional Medicare increased by 5.0 percent in the five years between 2009 and 2014, MA plan benchmarks increased only 1.5 percent (Exhibit 2). Plan bids, representing their expected costs per enrollee, increased only 2.6 percent — half as much as in traditional Medicare. Total per enrollee payments to plans, including rebates, actually decreased, by 0.7 percent. Nonetheless, MA plan enrollment increased across the board, growing 48 percent nationwide. These changes, however, varied across counties and by type of plan.
Changes by County-Level Medicare Spending Per Beneficiary
To examine changes at the county level, we ranked counties by traditional Medicare spending per beneficiary in 2009, from low to high, and grouped them into four quartiles, each with roughly equal numbers of counties. Not surprisingly, there was wide variation in traditional Medicare spending per beneficiary across the quartiles in 2014, ranging from an average of $7,764 in the quartile with the lowest spending to $10,499 in the quartile with the highest spending — a 35 percent spread. The counties with the highest traditional Medicare spending per beneficiary in 2009 tended to experience the smallest increases over the next five years, and vice versa.
There was a similar pattern in MA benchmarks and per enrollee payments to MA plans, with the plans in the highest-spending counties experiencing a decline of 0.5 percent in benchmarks and 2.4 percent in per enrollee payments between 2009 and 2014 (Exhibit 3).
The change in MA plan bids (i.e., expected per enrollee costs) also varied across the county quartiles, in roughly the same pattern as the change in Medicare per enrollee payment to plans (Exhibit 4). This indicates that MA plans may have responded to substantial payment pressure by reducing their own costs.14 Notably, though, even in the group of counties that experienced the largest declines, total Medicare payments to MA plans in 2014 were still 13 percent higher than their expected costs per enrollee of providing Medicare Part A and Part B benefits (Exhibit 5).
Changes by Type of Plan
We also grouped plans into three types: health maintenance organizations (HMOs), the most tightly structured type of plan in terms of managing care, and local preferred provider organizations (PPOs) and regional PPOs, each of which are less structured than HMOs.15 In terms of enrollment, HMOs have more than 70 percent of MA enrollees, while local and regional PPOs, with about 25 percent of all MA enrollees combined, are the fastest-growing plan type.
HMOs experienced essentially flat growth in benchmarks between 2009 and 2014 (0.1 percent) and the largest average decline in per enrollee payments of any plan type, with a decrease of 2.1 percent between 2009 and 2014 (Exhibit 6). They also had the smallest increase in their bids (1.5 percent) over the five years. Local and regional PPOs had benchmarks that grew by 1.4 percent and 3.8 percent, respectively; they experienced a small increase or slight decrease in per enrollee payments to them over the same period, and their bids grew faster than those of the HMOs.
Despite the slow growth in Medicare benchmarks faced by each type of plan, plans were able to keep their bids below those benchmarks rates, which enabled them to continue to receive rebates (Exhibit 7). HMOs, for whom benchmarks remained essentially flat, nonetheless were most successful, on average, in controlling their per enrollee costs. However, per enrollee payments to HMOs in 2014 were still 12 percent higher than their expected per enrollee costs of providing Medicare Part A and Part B benefits.
Our analysis indicates that, in the face of pressure from slow — and in some cases negative — benchmark growth, MA plans responded by controlling their cost growth. However, Medicare per enrollee payments to MA plans are still greater on average than plans’ bids, which means Medicare is paying substantially more than plans’ expected total costs of providing Medicare benefits to their enrollees. MA plans retain a portion of these extra payments as administrative fees and by law must use the remaining extra payments to offer additional benefits, which tend to attract new enrollees. This may help explain the continued enrollment growth in Medicare Advantage.
The fact that MA plans, in the aggregate, have responded to declining payment rates by controlling costs is encouraging news and lends credence to the theory that plans would respond to competitive market conditions by increasing efficiencies while also serving the needs of their enrollees. These results also undercut arguments that increased payment pressure would lead to the demise of Medicare Advantage; in fact, enrollment has continued to grow. Moreover, there may be further opportunities for MA plans to show how efficient they can be as their payments continue to exceed their costs and traditional Medicare spending per beneficiary in many counties — and while traditional Medicare moves from volume-based to value-based payments.16
Various proposals have called for different roles for private plans as Medicare enters its second 50 years. Some would increase the role of private plans, but critics worry that beneficiaries lack the information needed to choose among plans, and that many of those proposals might end up costing beneficiaries more and concentrating older and sicker beneficiaries in traditional Medicare.17,18 Alternatively, improvements in traditional Medicare’s structure and benefit package have been proposed that could enable it to compete on a more level playing field with private plans.19
In any case, more attention needs to be devoted to how private plans are paid. Until recently, Medicare payment policy did not provide strong incentives for private plans to perform effectively and efficiently or link payment to performance. More data should enable better analysis of these issues and — particularly given the growth of Medicare Advantage enrollment and calls for increased competition between traditional Medicare and private plans — the focus should be on improving both MA plans and traditional Medicare, to the benefit of all Medicare beneficiaries.
How We Conducted This Study
This analysis uses data from the Centers for Medicare and Medicaid Services (CMS) containing Medicare Advantage (MA) plan costs and payments (adjusted to reflect the average level of risk represented by plan enrollees),20 county-level MA enrollment,21 and county-level predicted traditional Medicare spending for 2009 and 2014.22 In these analyses, we include health maintenance organization (HMO), local preferred provider organization (PPO), private fee-for-service (PFFS), and regional PPO plans that are available to all Medicare enrollees, including Special Needs Plans (SNPs). MA plans reimbursed on a cost basis have been excluded, as have employer group plans. The resulting dataset includes 2,797 MA plans in 2009 and 2,329 MA plans in 2014.23 All estimates in this study are weighted by enrollment.
Before completing our analyses, we adjusted the projected traditional Medicare spending data from CMS to account for changes in CMS’ approach to Sustainable Growth Rate cuts. In 2009, the CMS Office of the Actuary did not assume that Sustainable Growth Rate cuts would be prevented from taking effect when developing the 2009 projected traditional Medicare spending that underlies MA plan benchmarks. We therefore adjusted the 2009 projected traditional Medicare spending upward to reflect that those cuts did not take place to provide a more accurate reflection of actual traditional Medicare spending that year.24,25 This adjustment was not necessary in 2014 because the CMS Office of the Actuary assumed that Sustainable Growth Rate cuts would not go into effect that year and calculated projected traditional Medicare spending accordingly.26
1 B. Biles, J. Pozen, and S. Guterman, The Continuing Cost of Privatization: Extra Payments to Medicare Advantage Plans Jump to $11.4 Billion in 2009 (The Commonwealth Fund, May 2009).
2 B. Biles, G. Casillas, G. Arnold et al., The Impact of Health Reform on the Medicare Advantage Program: Realigning Payment with Performance (The Commonwealth Fund, Oct. 2012).
3 Medicare Payment Advisory Commission, Report to the Congress: Medicare Payment Policy (MedPAC, March 2009); and Medicare Payment Advisory Commission, Report to the Congress: Medicare Payment Policy (MedPAC, March 2017).
4 Taking into account the effect of coding practices by MA plans, the ratio of payments to MA plans to per beneficiary spending in traditional Medicare was between 102 percent and 104 percent in 2017.
5 The analyses in this issue brief cover 2009 to 2014, the most recent years for which detailed, comparable data are available.
6 Congressional Budget Office, “Comparison of Projected Enrollment in Medicare Advantage Plans and Subsidies for Extra Benefits Not Covered by Medicare Under Current Law and Under Reconciliation Legislation Combined with H.R. 3590 as Passed by the Senate (based on draft legislative language and modifications discussed with staff),” in Selected CBO Publications Related to Health Care Legislation, 2009–2010 (CBO, Dec. 2010), pp. 66–68.
7 M. Gold, “Medicare+Choice: An Interim Report Card,” Health Affairs, July/Aug. 2001 20(4):120–38.
8 Medicare Payment Advisory Commission, Health Care Spending and the Medicare Program: A Data Book (MedPAC, June 2017).
9 Although data are available for 2015, those data reflect changes in the calculation of Medicare Advantage benchmarks and changes to the normalization factor for risk adjustment that create a break in trend for time-series analysis. For more detail about these changes, see Centers for Medicare and Medicaid Services, Announcement of Calendar Year (CY) 2015 Medicare Advantage Capitation Rates and Medicare Advantage and Part D Payment Policies and Final Call Letter (CMS, April 7, 2014).
10 For more detail, see Medicare Payment Advisory Commission, Payment Basics: Medicare Advantage Program Payment System (MedPAC, Oct. 2015).
11 B. Biles, G. Casillas, G. Arnold et al., The Impact of Health Reform on the Medicare Advantage Program: Realigning Payment with Performance (The Commonwealth Fund, Oct. 2012).
12 Each MA plan is permitted to include in their bid a profit rate equal to the rate they receive on their commercial business.
13 For this analysis, we have excluded Medicare Advantage employer plans and plans reimbursed on the basis of their costs because the payment rules and markets for those plans are different than for other MA plans.
14 In addition, higher-cost plans may have left the market, or enrollees may have disproportionately joined lower-cost plans.
15 Although we also examined private fee-for-service (PFFS) plans, which are the least tightly structured of any type of MA plan, they are not discussed here because significant changes affecting PFFS plan enrollment may hinder the ability to compare them to other types of plans.
16 S. M. Burwell, “Setting Value-Based Payment Goals — HHS Efforts to Improve U.S. Health Care,” New England Journal of Medicine, March 5, 2015 372(10):897–99.
17 G. R. Wilensky, “Reforming Medicare — Toward a Modified Ryan Plan,” New England Journal of Medicine, May 19, 2011 364(20):1890–92.
18 H. J. Aaron and A. B. Frakt, “Why Now Is Not the Time for Premium Support,” New England Journal of Medicine, March 8, 2012 366(10):877–79.
19 K. Davis, C. Schoen, and S. Guterman, “Medicare Essential: An Option to Promote Better Care and Curb Spending Growth,” Health Affairs, May 2013 32(5):900–9.
20 Centers for Medicare and Medicaid Services, Plan Payment Data for 2014 (CMS, 2015).
21 Centers for Medicare and Medicaid Services, Medicare Advantage/Part D Contract and Enrollment Data (CMS, n.d.).
22 Centers for Medicare and Medicaid Services, 2014 Rate Calculation Data (CMS, 2013).
23 Most of the decline in the number of plans was among PFFS plans. There were 505 PFFS plans in our dataset in 2009 and only 114 PFFS plans in our dataset in 2014. Also, we note that our dataset does not include employer group plans, plans paid on the basis of their costs, and any plans with fewer than 11 enrollees.
24 For this adjustment, we multiplied the projected traditional Medicare spending in each county by 1.024 to reflect that the 10.6 percent cut to 22.5 percent of beneficiary spending was not implemented (10.6 * 22.5 = 0.024).
25 This methodology also was used in prior work comparing Medicare Advantage costs and payments across geographic areas. See B. Biles, G. Casillas, G. Arnold et al., The Impact of Health Reform on the Medicare Advantage Program: Realigning Payment with Performance (The Commonwealth Fund, Oct. 2012).
26 Centers for Medicare and Medicaid Services, Announcement of Calendar Year (CY) 2014 Medicare Advantage Capitation Rates and Medicare Advantage and Part D Payment Policies and Final Call Letter (CMS, April 1, 2013).
The authors thank Shawn Bishop and Eric Schneider of the Commonwealth Fund for their review and comments on various versions of this paper, and Brian Biles for his involvement in initial stages of this research.