Typically, full service health insurance companies will have several clusters of business. There is the low cost government cluster of Medicaid and SNP managed care. There is the medium cost government supported sector of Medicare Advantage, CHIP and low cost Exchange plans. And then there is the higher premium cluster of employer sponsored fully insured (ESI) plans and Administrative Services Only (ASO) self-insured employer plans. I worked at a full service firm. I was on the Medicaid geek team for the last three years.
Recently, I brought up the Rogers, Chernew and McWilliams Health Affairs paper on the impact of market power on local level provider and insurer pricing. They were only looking at employer sponsored insurance market power. They found what was to be expected. Entities with high relative market power got “better” rates from their point of view compared to entities with low market power. This was a good set of results as they were able to math up the trade-offs and attach some real numbers to the intuition.
My question though is how to account for the results if we are to assume that log-rolling in negotiations occurs?
An insurer might have a low market share for the high premium ESI/ASO market in a region. That same carrier could have a very high market share of the Medicare Advantage market. If that carrier is talking with a hospital that has never been in network to sign a comprehensive, all products contract, does the negotiation’s plausible agreement region get defined by each line of business’s relative market share or is the plausible agreement region defined solely by a blended dollar weighed marketshare?
More practically, does a hospital say that in order to get a stream of the Medicare Advantage money they’ll take lower than anticipated by RCM commercial rates or the carrier offer slightly higher Medicare Advantage rates to buy access for the employer side plans?
My intuition is that this type of log-rolling happens a lot. So how does it get measured and evaluated?
I don’t know.
bbleh
Haven’t read the paper, but since you’re talking about synergies across markets, the measure would be the difference between the observed aggregate activity (e.g., revenue, lives) and that predicted by a market-by-market analysis, assuming such predictions exist and are sufficiently accurate. Obvious problems include various sources of noise in the data, and actual dynamics in the markets (how long does a player’s position remain substantially the same in ALL relevant markets?). But for sure it exists — I’ve seen it firsthand in a closely related industry.
David Anderson
@bbleh: I think that type of analysis would be useful. I don’t think the data is good enough to do it on a national basis but it could be hacked together from states with good all claims payer databases
Aleta
OT: Lawyers and bankers reap $1.5 billion from failed health insurance mergers
randy khan
Unless both sides are bad at negotiating, I would expect some amount of log-rolling. Essentially, the hospital will be willing to take some things it wouldn’t want to get the more lucrative bits, and the insurer would be willing to pay a bit more for some things to get the hospital to take bits it wouldn’t want.
David Anderson
@randy khan: Same here — this is just a data limitation of a very useful study. It was not measuring logrolling
DLG
If this was truly and only a case of bilateral bargaining, then bargaining over a broader range of market segments/trades/issues/etc would be beneficial. Party A could get more of what it really values in segment 1 in exchange for more of what Party B cares about in segment 2. I.e. “log-rolling.”
The problem is that there are other interests involved — health-care consumers, employer-employee benefit agreements, government — interests that are our (the public; policy analysts) primary concern. Not to mention actual/potential competitors in both the insurance and healthcare provider markets.
I can think of two aspects where general (economic) industrial organization theory can shed some light. First, “multi-market contact” increases the likelihood of anti-competitive collusion. The key point being a game-theoretic argument that the more extensive the interactions, the larger the scope for “punishing” any party to the collusion who deviates from the (possibly implicit) collusive agreement. This logic is usually applied to oligopolistic competitors, but can also apply to vertically related businesses (e.g. insurers and providers) extracting rents from consumers and third-parties.
Which leads to the second area — “rent-seeking,” which is the creation and exploitation of market and regulatory inefficiencies (in the economists’ sense) in pursuit of unearned/unjustified excess profits (a type of “rent” in economic jargon). Log-rolling to the detriment of third parties would be a type of coordinated rent -seeking. Just as multi-market contact can increase the opportunities for mutually beneficial exchanges (see para 1), it can also increase the opportunities for coordinated rent-seeking. Specifics will depend on the particulars of markets, institutions, and regulations (beyond my expertise in the case of healthcare). But one pretty obvious opportunity is to shift accounting of costs between market segments in order to then increase the costs shifted to third-party burden sharers (e.g. co-pays, various types of govt subsidy).
Empirically, absent fine-grained information on agreements across market segments, a crude way to investigate the significance of these issues would be to condition on (e.g. in cross-tabulations or regressions) market concentration and share data for the other market segments the insurer and provider operate on. However, since opportunities for rent-seeking are sensitive to institutional and regulatory regime, we shouldn’t expect any regression effect to be stable when there are significant changes, such as implementation of the ACA.