• A revealing study of Medicare Advantage competitive bidding

    In Competitive Bidding in Medicare: Who Benefits From Competition? (AJMC, ungated), Zirui Song, Mary Beth Landrum, and Michael Chernew examine how plan bids in the current Medicare Advantage program respond to changes in the government’s maximum payment to plans (i.e., the maximum allowable subsidy plans receive to cover beneficiaries, also known as the “benchmark”).

    The conventional wisdom (or, perhaps, naivete) is that plans bid their costs. Indeed, that’s what they’re supposed to do. A plan bid is interpreted as its estimate of the cost of providing the standard Medicare benefit. When bids are below the benchmark, as they usually are, plans receive 75% of the difference in a rebate, which they are obligated to use to provide extra benefits or reduced cost-sharing. If a plan bids above the benchmark, it receives the benchmark as a subsidy and must charge the difference as premium.* If this is unfamiliar to you, it’s all explained in the paper. See, in particular, Figure 1. (You won’t find Figure 1 online. However, if you click to the end of the paper as it appears in HTML online, you will see a link to a PDF, which I believe you can obtain with free registration.)

    If plans actually bid their costs of providing the standard Medicare benefit, then that cost ought not change in response to an exogenous change in benchmark (i.e., assuming the benchmark is not itself responding to cost changes). The authors make an argument for why benchmarks are, in fact, disconnected from plan costs, and they toss in a bunch of controls just in case that’s not a perfect assumption. You can either trust me or read the literature on this: their argument and approach are standard and widely accepted.

    Contrary to simple models of competition that suggest that bids should not respond to changes in benchmarks (holding costs constant), we found that bids increased by about $0.33 to $0.50 for every $1 increase in benchmarks. This left only $0.34 to $0.38 for rebates to beneficiaries. Our results reject the notion that the private Medicare market is perfectly competitive. This has implications for current debates as well as prior literature on Medicare Advantage. […]

    Our findings have potential implications for proposals to expand the role of competitive bidding in Medicare such as the Ryan-Wyden plan. They suggest that markets for health insurance for Medicare beneficiaries may not be perfectly competitive and thus simple bidding systems may not drive bids down to plan costs. Thus, these findings should temper the enthusiasm for replacing Medicare with a bidding system. Of course, other models of bidding, including those in which the benchmark is not known in advance, may yield better results and thus attention to the rules surrounding bidding is important. Moreover, any bidding system must be compared with alternative strategies and administratively set prices have their own faults. While administratively set prices can mitigate concerns about plan market power, they may be subject to political manipulation that sets prices too high and therefore result in excessive spending (as many felt was the case in the pre-PPACA Medicare Advantage program). Alternatively, administratively set prices could be set too low, impeding access to care. Thus careful attention to the mix of markets and regulation will be needed as the debate over Medicare reform escalates. Any reliance on, and design of, market-based bidding mechanisms should balance the concerns over imperfect competition with the imperfections of administratively set prices.

    That’s a reasonable interpretation of the findings. At a minimum, they strongly suggest that the current Medicare Advantage market is far from perfectly competitive. Nobody expects perfect competition, but losing as much as 50 cents on the dollar to market imperfections is a large deviation. It does suggest that premium support without a strategy for increasing the degree of competition might come with a large dead weight loss. However, the very fact that the second lowest bid would set the subsidy level in Ryan-Wyden type premium support should increase price competition. By how much? Nobody knows.

    Still, there’s not much to love about today’s administrative pricing system. We’re wasting a lot of money on Medicare Advantage. A heap of it is not helping beneficiaries.

    * Read all this as applying to an beneficiary of average health. Actual plan subsidies are risk-adjusted.


    • Austin
      For clarification, when you state plans bid their costs, I assume you mean plans are incorporating their desired profit margin as well, yes?

    • Nice analysis, but of course the paper uses data and MA guidelines that were effective BEFORE the ACA was implemented. Post-ACA, plans receive a smaller cut of the bid/benchmark difference (both in absolute terms and also relative to the plan quality rating, i.e., < 75%). Additionally, the ACA imposes maximum MLR criteria which are designed to drive more of the revenue into plan benefits.

      It will be interesting to see how these results change once analysis is performed on post-ACA data.

      • It would also be interesting to see a comparison with dead weight loss in Medicare Part D. Although costs and benefits are structured differently than in MA, Part D employs a competitive bidding process. Plans have pushed the envelope when it comes to bidding rules in recent years, underscoring your point that a competitive bidding process requires strict attention to the rules.

        • Given the structure, it’s unclear to me how to find an exogenous change in the Part D subsidy. Got ideas?

          • You’re right – the part D subsidy is endogenous given its relationship to the average bid. Part D bidding is also complicated by the fact that Plans often use Part D to drive membership to private plans (MA and Med Sup) or to shape the MA bid in the case of MAPD plans. So of course, Part D is not perfectly competitive either.

    • I don’t see it explicitly stated anywhere whether the benchmark rate is known to the bidders before-hand. Is it? If it does, I’m perplexed as to why the government would do that.

      I also wonder if the rebates could be problematic. It might be difficult to reconstruct the particulars of the plan to distribute the rebates (I don’t think the plans can just directly refund the money). If so, it may act as a source of friction on the bids to keep them close to the benchmarks.

      Finally, I would be curious to know what the result is for the most competitive markets. The authors give the $0.33 figure when all types of plans are included, but that’s just an average, I think. What is the distribution across markets?

      • I don’t know if the benchmark is explicitly released prior to when bids are due, but I will find out. Even if it isn’t, it is based on a (very complicated) formula derived from public data. Insurers could crunch it.

        Plans may refund the Part B premium (up to about $100/month), as well as shift their copays, deductibles, or offer additional benefits.

        Good question on the distribution. I recommend emailing the authors.

        • Thanks for checking.

          If the insurers have no knowledge or estimate of the benchmarks, then I think that completely invalidates the study. It would mean that all bid changes were unaffected by benchmarks. If the benchmarks aren’t announced before-hand, I agree, it’s probably not too difficult for insurers to predict what they’ll be, but I think the onus is on the authors to prove that the companies can and do.

          I’m not really sure where the rebate figures in the paper come from, they’re not listed as independent variables. But if the insurers are rebating some of the money through copayments or deductibles, is that caught somewhere in their analysis?

          • The rebate is a straight computation, per figure 1. It’s not all rebated to benes in dollars. It’s rebated to the plans that way, and then they do as I wrote in my prior reply.

            The study is fine. See prior comment from Pedro. (The authors know this stuff very, very well. Chernew serves on MedPAC and advises CBO. He’s not going to get this stuff wrong.)

      • The MA benchmarks are known ahead of time. MA rates are based on county-level FFS costs, and a plan’s benchmark is an average of county rates weighted on its membership and adjusted for plan risk (to mitigate adverse pselection based on population profiles). The rebates explicitly have to pay for additional benefits, but the advantage is more attractive benefits for a lower premium. Yes, margin varies considerably across plans and markets.

        I should disclaim that I do analysis in the Medicare business of a very large insurer. My views and opinions are mine I course and don’t represent hose of my employer.

    • I’m not sure how prevalent it is outside of California, but I know that it is common practice in California to tie capitation payments directly to what CMS pays a Medicare Advantage plan. For example, if CMS pays an MA plan $1000 per month for a particular member, the physician group will get 35% of that as their capitation payment.

      As that knowledge is factored into the bid, that would most definitely lead to a correlation between bids and benchmarks.