Recently, a number of health policy experts have suggested that the best route to universal coverage might be to expand Medicare Advantage (MA) rather than enact Medicare for All—in other words, a private-sector instead of a government solution. Under some proposals, President Joe Biden’s Medicare-like public option would be replaced with private insurance options. Either MA plans would be allowed to compete for non-Medicare business on the insurance exchanges, or insurers would negotiate prepaid contracts with providers, using government funds.
At this point, with major health care reform barely on the legislative horizon, these ideas are purely theoretical. Nevertheless, they are serious policy proposals. Therefore, some hard questions about Medicare Advantage need to be asked and answered now. To reach any firm conclusions about the desirability of expanding the MA program, or patterning a new program after it, we first need to determine what we know and don’t know about MA.
What We Know About Medicare Advantage
We know that MA enrollment is growing much faster than that of traditional Medicare. By the end of 2020, more than 40 percent of Medicare beneficiaries were enrolled in MA plans (exhibit 1), and the Congressional Budget Office has forecast that 47 percent of beneficiaries will be enrolled in these plans by 2029. In short, we’re witnessing the rapid privatization of Medicare.
Exhibit 1: Medicare Advantage enrollment rates over time
Source: Authors’ analysis of Medicare Advantage State/County Penetration data as of December for each year.
Why is this happening? Simply put, MA plans offer beneficiaries a better deal than traditional Medicare. Instead of having to buy a Medigap plan (at an average cost of $172 per month) to cover costs that traditional Medicare doesn’t cover, as well as a Part D prescription drug plan ($46), MA enrollees pay only a small monthly premium ($40 on average in 2019), and 56 percent of enrollees pay no premium on top of their Medicare Part B premium. In many plans, they get some level of dental, vision, and hearing coverage, which is missing in traditional Medicare. Some plans offer additional benefits, such as transportation to the doctor and home-delivered meals for those recently discharged from the hospital. There’s also a statutory annual limit on out-of-pocket costs, which is not part of traditional Medicare.
On the other hand, MA networks are restricted, and some are very limited. In 2015, 35 percent of MA enrollees were in plans with narrow physician networks—although recent research shows that most MA members have broad access to primary care doctors. MA enrollees who become seriously ill may have to pay hefty fees to out-of-network providers.
We know that the penetration of MA plans varies widely across the country, ranging from 1 percent in Alaska (where no plans are directly offered) and 4 percent in Wyoming to 49 percent in Florida (exhibit 2). Nationally, 99 percent of Medicare enrollees have access to at least one MA plan, but there is significant variance in the number of insurers offering plans in each state, ranging from zero in Alaska and 1 in Wyoming to 10 each in Arizona, California, Florida, and Ohio.
Exhibit 2: Medicare Advantage penetration by state, 2020
Source: Authors’ analysis of Medicare Advantage State/County Penetration data from December 2020.
We know that enrollment in MA plans varies by demographics. Enrollment in MA plans tends to increase by the age of beneficiary up through the late 70s and then decreases beyond that point, and women are more likely than men to enroll in MA (exhibit 3). There are also significant differences among races; in 2018, 43.8 percent of Hispanic Medicare beneficiaries enrolled in MA plans compared to 39.2 percent of Blacks, 32.4 percent of Whites, and only 16.1 percent of Native Americans (source: authors’ analysis of 2018 Medicare Master Beneficiary Summary File). As the demographics of our country continue to shift, we need to: continue to measure these differences in enrollment by race, understand why the differences exist, and determine whether the differences matter.
Exhibit 3: Medicare Advantage enrollment by age and sex
Source: Authors’ analysis of 2018 Medicare Master Beneficiary Summary File.
We know that Medicare pays MA plans slightly more per enrollee than it pays for comparable traditional Medicare enrollees, after accounting for quality bonuses and diagnostic coding intensity. This intensity represents the additional diagnoses that many physicians document for MA members but not for traditional Medicare beneficiaries. MA plans have a financial incentive to ensure that their providers record all possible diagnoses, since higher enrollee risk scores result in higher payments to the plan. “On average across the nation, MA payments are about 2 percent higher than expected FFS [fee for service] expenditures for similar beneficiaries,” wrote the Medicare Payment Advisory Commission (MedPAC) in its March 2020 report to Congress.
The ratio between government payments to MA plans and traditional Medicare costs would have been even higher, MedPAC noted, had it not been for Medicare’s downward adjustment for MA coding intensity. In 2018, this adjustment lowered MA risk scores—which help determine plan payments—from more than 8 percent higher to only 2 percent to 3 percent higher than risk scores for similar beneficiaries in traditional Medicare.
The close correlation of MA payments to traditional Medicare spending is built into the program. “Currently, all savings to the [Medicare] program that come from MA must be generated through FFS [fee for service] spending reductions,” noted MedPAC. This means that MA costs currently cannot be reduced independent of fee-for-service reductions.
MedPAC further commented that MA plans are expected to be more efficient than traditional, or fee-for-service, Medicare, because they actively manage care and have an incentive to keep costs down. However, MedPAC added, “Given the level of overutilization in FFS and other factors, we cannot conclude that achieving payment parity between MA and FFS Medicare would leverage any efficiency from the MA program.”
We know that per-enrollee costs for MA are rising much faster than per-enrollee costs in traditional Medicare. In 2019, per-enrollee MA costs grew 6.3 percent, compared to 2.4 percent for traditional Medicare. This was the fourth year in a row that MA cost growth per member exceeded that in the traditional program (exhibit 4). This finding is unexpected, as MA plans have more controls and incentives to manage costs and cost growth than traditional Medicare does.
Exhibit 4: Medicare per beneficiary cost growth rate
Source: Authors’ analysis of National Health Care Spending reports conducted by the Centers for Medicare and Medicaid Services for 2016, 2017, 2018, and 2019.
We know that MA is dominated by a few large companies, and that it is a profitable line of business. At the end of 2020, the four largest MA companies accounted for more than 60 percent of total enrollment, and the 15 largest accounted for more than 80 percent. MA has also been very lucrative for many private health insurers. In 2017, according to MedPAC, the average pretax margin of for-profit MA plans was 5.2 percent; in 2018, it was 4 percent. Researchers from the Commonwealth Fund found that between 2016 and 2018, annual gross margins in the MA market averaged $1,608 per covered person, roughly double the insurers’ margins in the individual and group commercial markets. Unsurprisingly, UnitedHealthcare and Anthem are quickly ramping up their MA enrollments. Humana derived two-thirds of its revenue from Medicare Advantage in 2019.
What We Don’t Know About Medicare Advantage
We don’t know how to explain the faster per-enrollee cost growth in MA compared to traditional Medicare. In their 2020 report on national health spending, the actuaries of the Centers for Medicare and Medicaid Services (CMS) attributed it to “faster spending growth among all medical goods and services.” But if overall cost growth were to blame, why didn’t per-beneficiary spending rise just as fast under traditional Medicare? Is the higher growth a function of program design that can be modified, or is it indicative of differences between the MA and traditional populations?
We don’t know whether or not MA offers better quality of care for beneficiaries than traditional Medicare. A number of studies suggest that it does. A 2015 paper, for example, found that MA plans beat traditional Medicare on four measures of diabetes care. A 2017 study of enrollees in three large states showed that MA outperformed traditional Medicare on all of 16 quality measures and four of six patient experience measures, although MA enrollees had less access to care. And a 2018 Avalere study found that MA enrollees with certain chronic conditions had 23 percent fewer inpatient stays and 33 percent fewer emergency department visits than did matched traditional Medicare beneficiaries.
Overall, however, a “definitive finding [that MA offers superior quality] is not possible with currently available data,” MedPAC stated in 2019. Except for patient experience data, the commission said, “the data needed to compare MA with FFS are lacking.” Researchers have suggested that assessing quality is difficult due to the lack of necessary data.
Another marker of quality is whether Medicare beneficiaries remain in MA plans when their health deteriorates. On this basis, there is reason to question whether MA plans offer higher quality. One paper found that MA members who had been hospitalized at least once had a higher rate of switching back to traditional Medicare than did other MA enrollees. The same was true for users of home care and long-term nursing home care.
We don’t know if MA saves money for Medicare. While MA plans save money for beneficiaries, we don’t know whether they create savings for Medicare. As noted, the plans are paid, on average, slightly more than traditional Medicare spends on similar beneficiaries in the same region. There is some evidence that where MA market penetration is high, there’s a spillover effect in which better care management in MA plans leads to lower costs of care in traditional Medicare. The spillover effect of MA on traditional Medicare has been studied, but the total impact has not been definitively shown. We compared MA penetration with risk-adjusted standardized traditional Medicare costs for 2018 and found that a 1 percent higher penetration of MA was associated with a decrease of $23 per beneficiary per year in traditional Medicare, although we don’t know if the relationship is causal.
According to health insurers, MA plans spend 20 percent to 40 percent less on care than traditional Medicare does. The question is, if these plans are being paid at above Medicare rates, where does the rest of the money go? Some of it is being invested in extra benefits and reduced or zero plan premiums. All of that comes out of the 50–70 percent of the difference between a plan’s bid and its CMS-determined benchmark that CMS rebates to the plan. In 2020, the average bid of MA plans was 88 percent of their county-level benchmarks, which are based on the prospective amount of fee-for-service spending, and the average benchmark was 107 percent of traditional Medicare costs. Fifty to 70 percent of the 19-point difference between these two figures is far less than what the plans claim to save. For many plans, a meaningful share of the savings consists of profits; there are also administrative and other overhead expenses, but it is unclear what factors explain the whole gap between rebates and plan savings.
We don’t know how MA plans lower the cost of care. Without considering favorable risk selection, there appear to be at least three possible explanations:
- MA plans negotiate favorable contracts with providers;
- MA plans skimp on care; or
- MA plans improve care coordination and incentivize primary care doctors to manage care better for their enrollees.
With regard to provider payments, studies have found that MA plans pay hospitals between 95 percent and 105 percent of Medicare rates. One of these papers hypothesized that hospitals accept these payments partly because the MA program prohibits them from accepting more than Medicare rates if they’re out of network.
Out-of-network physicians also must accept traditional Medicare rates. A study of payments to physicians for mid-level office visits and electrocardiograms found little difference between MA and traditional Medicare rates. What’s unclear is how much MA plans pay in-network doctors in quality bonuses and shared savings—not to mention risk contracts. Considering that MA network doctors have incentives to manage care and to code every potential health issue, they likely are receiving something beyond a flat traditional Medicare payment. Anecdotal evidence indicates that these bonuses can be quite substantial in some cases, but more information is needed to understand how prevalent these risk-based and incentive-laden payment models are for MA providers.
MA plans include health maintenance organizations and preferred provider organizations, both of which use standard managed care techniques to varying extents. Whether tools such as prior authorization and utilization review result in skimping on necessary care is unclear. However, if MA plans were really providing substandard care, it would show up in patient experience surveys, which generally give these plans high grades. MedPAC found few differences between MA and traditional Medicare in this respect. While there is evidence of lower use of services in MA plans, compared to traditional Medicare, that doesn’t mean that the omitted services are necessarily of high value.
So, we’re left with the third possibility, which is that MA plans and their network providers are actually doing a pretty good job. If this is the case, we’d expect to see better care management by primary care physicians that keeps patients out of the hospital and the emergency department and minimizes their use of specialists and expensive drugs. Beyond some case studies, however, we were not able to find evidence about how MA plans manage patients and if the care management is the primary mechanism they use to lower costs.
Future Research Directions
In researching this paper, the biggest conundrum we encountered is the fact that per-enrollee MA cost growth exceeds the rate of per-beneficiary cost growth in traditional Medicare. Perhaps, this is related to how MA plans raise their risk scores: When primary care physicians diagnose additional conditions in MA members, they’re obligated to provide more services to manage those conditions, pushing up costs in the short run. With MA expanding rapidly, many new enrollees with chronic conditions are entering that funnel, so per-enrollee costs are rising; but in the long run, those costs should plateau or decline because of avoided medical crises. Newly available encounter data from CMS might clarify whether this is true or whether there is another explanation.
It is also imperative to find out whether MA plans provide better quality overall than traditional Medicare does. MedPAC has suggested that comparing traditional Medicare claims data to analogous MA encounter data might answer this question. We also think that encounter data could be helpful.
To better measure the spillover effect from MA to traditional Medicare, researchers might want to compare per-enrollee costs for both programs in low- and high-MA penetration counties. Since the spillover effect has been observed mainly in the latter areas, the comparison with low-penetration areas would clarify whether there is a wider spillover effect.
We also need to better understand how shifting demographics are likely to impact the demand for MA and assess whether MA has differing results for different enrollee demographics.
In conclusion, it is not clear that Medicare Advantage for All would help bend society’s cost curve more than Medicare for All or other policy proposals. In fact, the current evidence suggests that MA plans have not saved Medicare any money relative to traditional Medicare. To the extent that they lower costs, the lion’s share of those savings seems to be flowing to insurance companies, partly in the form of profits. Policy makers should consider whether this is the direction in which they want health care financing to go.
Medicare Advantage may or may not be an ideal policy lever to pull. However, prior to any wholesale expansion of Medicare Advantage to non-Medicare individuals, we need to take some time and answer some basic questions about it.