• How market concentration can be good

    Last Friday, the House Ways and Means Subcommittee on Health focused on the implications of hospital mergers and acquisitions. A helpful summary of the meeting can be found at The Advisory Board Daily Briefing.  Very briefly, increased hospital market concentration pushes up prices, something we don’t need more of in health care. That market concentration leads to higher prices is a general principal of economics. So, we tend to hear “market concentration” as “something we ought not have more of.” However, when it comes to health care, not all market concentration is bad. Sometimes, more is better.

    Take the health insurance market,* for example. It’s famously concentrated. The way the story is typically told is that a few big insurers with massive market clout are demanding excessive premium increases. Wouldn’t we all be better off if the insurance market were less concentrated? Wouldn’t greater competition lead to lower prices?

    Not necessarily. One reason it would not is offered by the work of Glenn Melnick, Yu-Chu Shen, and Vivian Wu in their recent Health Affairs paper “The Increased Concentration Of Health Plan Markets Can Benefit Consumers Through Lower Hospital Prices.” Here’s what the authors found:

    • “[Sixty-four] percent of hospitals operate in markets where health plans are not very concentrated, and only 7 percent are in markets that are dominated by a few health plans.” Translation: Be skeptical of statements by the hospital industry that most hospitals are suffering at the hands of dominant insurers.
    • [I]n most markets, hospital market concentration exceeds health plan concentration.” Translation: In fact, due to their relative market clout, most hospitals have the upper hand in negotiation over price.
    • “[G]reater hospital market concentration leads to higher hospital prices.” Translation: Do I really need to translate this?
    • “[H]ospital prices in the most concentrated health plan markets are approximately 12 percent lower than in more competitive health plan markets.” Translation: Insurer market power can be a good thing, leading to lower hospital prices.

    Consistent with a large body of prior work, some of which they cite (see also this (pdf) and a related FAQ entry), their results

    show that more concentrated health plan markets can counteract the price-increasing effects of concentrated hospital markets, and that—contrary to conventional wisdom—increased health plan concentration benefits consumers through lower hospital prices as long as health plan markets remain competitive.

    That last bit–“as long as health plan markets remain competitive”–needs some explanation. A question not addressed in the paper is whether insurers that are able to negotiate lower prices due to their relatively higher market concentration pass any of the savings along to consumers in the form of lower premiums. That is, do consumers really benefit if insurers get good prices? Or do insurers just turn around and stick it to the consumer anyway, raking in higher profit? A monopoly insurer would. Insurers in a relatively competitive market would not, or not as much.

    This all implies there is a sweet spot, a degree of market concentration of insurers relative to hospitals that leads to the lowest premiums. That’s what my figure illustrates:

    However, health economists do not know where that sweet spot is. Sarah Kliff also raised this point–that it is not clear from the paper by Melnick and colleagues precisely how much hospital and insurer market concentration affect premiums. A reason neither this, nor any other study I’m aware of, simultaneously relates insurers’ and hospitals’ market power to premiums is that there are some very challenging econometric issues in doing so and because comprehensive, historical data on commercial market health insurance premiums (as well as benefits needed to standardize them) are not readily available to researchers at relatively low prices. In fact, even the theory has not been fully worked out beyond its qualitative contours.

    One final point about lowering private prices to hospitals: it’s not only potentially good for consumers, it’s good for taxpayers. One of the constraints on Medicare’s ability to lower prices it pays to providers is the price level of the private market. Medicare can only push its payments so far below private payers before providers threaten to see fewer Medicare patients, creating a political problem for policymakers.

    So, paradoxically, policies that would increase competition among health insurers too much could actually harm consumers and taxpayers. Policies that would increase competition among hospitals, on the other hand, would probably do some good.

    (A key caveat to the paper by Melnick and colleagues is that their findings are descriptive and associative. This is acknowledged by the authors and is due to the methods employed. Having said that, the direction of causality between market power and price is accepted by economists based on theory and other empirical work.)

    * There is not one health insurance market, but many. For simplicity, unless stated otherwise, I will write as if there were just one.

    • Is there any research on economies of scale in health care in paying for potentially fixed costs (at least until patient volume becomes too high) e.g. paying for the construction costs of new facilities and the purchase of equipment like MRIs? Are fixed costs important in health care?

      • There are

        few firm conclusions about whether hospitals exhibit increasing, decreasing, or constant returns to scale, although the most recent and methodologically sophisticated work does suggest signifi cant scale economies. Those recent fi ndings provide some support for the claim that hospital facilities consolidation achieve savings by producing economies of scale. There is, however, little evidence that hospitals achieve signifi cant cost savings via ownership consolidation alone.


    • I have come to believe that Uwe Reinhardt (sp?) is correct. It is private costs that drive Medicare costs up. Medicare needs to pay just enough so that not too many physicians stop accepting it for payment. We concentrate on Medicare because of its budget implications, but a lot more attention should be paid to private insurance costs. (That figure is a favorite of mine. It really captures the effect of market power.)


      • This phenomenon–that private prices keep public ones higher than they otherwise would be–is not widely appreciated. I once spent some time trying to think of what to call this situation. At first, and only briefly, I thought it was a “reverse cost shift,” but it isn’t. It’s more like an additional tax. We all pay higher taxes because of higher private payments. I have no doubt Medicare prices would be lower if private ones were as well.

    • Every hospital and insurer consolidation that I am aware of, or participated in, was sold to the FTC/JD on the basis of lowering operating costs and premiums respectively. These goals fall by the wayside during rate negotiations between providers and insurers.

      Has there ever been a retrospective analysis of an insurer or hospital consolidation to see what it actually produced?

    • Not to disagree completely, but I would like to expand the commentary and some more color to your analysis

      “Our findings suggest that if policymakers are interested in lowering costs, they should find a way to restore competition among hospitals, in addition to assuring competition among health plans. There has been great consolidation among hospitals and physician practices, and that consolidation has allowed hospitals and doctors to raise prices.”


      “Translation: Be skeptical of statements by the hospital industry that most hospitals are suffering at the hands of dominant insurers.”

      The AHA always plays the victim.

      “One final point about lowering private prices to hospitals: it’s not only potentially good for consumers, it’s good for taxpayers.”

      Totally true, lack of coverage is the result of rising prices. We have an inflation problem, not a coverage problem.

    • At the risk of mixing politics and science…

      The market concentration curve posits that premiums will rise if either providers or insurers have too much concentration, but there is a difference, There is a well established doctrine for regulating private market insurance prices, and that task gets easier and more efficient as insurers concentrate. Witness Rhode Island and Vermont. Such regulation can force insurers to pass monopoly benefits back to consumers.

      There is no comparable doctrine for regulating providers.

      • Regulators forcing insurers in concentrated markets to return monopoly benefits to consumers may well be within your experience. You are writing from the West Coast, as you not?However, many insurers on the East Coast in highly concentrated markets with sophisticated regulators (ex Pennsylvania) more than doubled, some tripled their assets during the period 1996/97 through 2004+/-. Those increases in assets did not come primarily from investment income. Their assets are, as you probably know, are invested in bonds.

        After many years in New York, I know live in Rhode Island. Frankly, I don’t know of many business leaders who are happy with the RI premium experience. Vermont has invested years of work in reforming healthcare and they are now moving toward a single payer model because premiums and healthcare costs have risen substantially in spite of all the good work that they have done.

        For what it is worth, it seems to me that in concentrated markets, providers demand more and insurers pass those costs on while increasing their own share of premium.