Code Red: Two Economists Examine the U.S. Healthcare System

November 17, 2010

Hospital Price Dispersion

Filed under: Competition,Health services research — David Dranove and Craig Garthwaite (from Oct 11, 2013) @ 12:30 pm

A study by the Center for Studying Health System Change that will be released tomorrow shows that hospitals receive different prices for treating the same diseases. Center President Paul Ginsburg says this about the findings:

“The variation in hospital prices found in this study are (sic) inconsistent with highly competitive markets—at least for markets outside of health care,” said HSC President Paul B. Ginsburg, Ph.D.,

Hospital markets may not be highly competitive, but this argument is silly. One might as well say “The variation in automobile prices is (not “are”) inconsistent with highly competitive markets.” But one would be wrong in either case.

Vertical quality differentiation (i.e., some sellers are better than others) generates price dispersion in competitive markets. It is only in the most basic treatment of competition — in the first week of an intro economics course — that vertical differentiation is ignored. Observed price dispersion is not incompatible with competition.

Moreover, there is severity dispersion within diagnoses (e.g., some hospitals get the sickest patients within DRGs). Medicare ignores this when setting payment rates. But private insurers need not ignore this and can calibrate pricing accordingly. Hospitals getting sicker patients can get higher prices. Again, this is not incompatible with competition.

So what did we learn from this study? That different hospitals get different prices? We also see price dispersion for autors, and for dry cleaning and electric pencil sharpeners, for crying out loud. This isn’t news and there are no policy implications to be drawn from the study. But this is health care, so I predict this study will be front page in the New York Times, even if it fails to tell us anything we didn’t already know.


  1. […] This post was mentioned on Twitter by Kellogg School, Sher Graham. Sher Graham said: RT @KelloggSchool: Hospitals receive different prices treating same disease #Kellogg Prof David Dranove questions HSC … […]

    Pingback by Tweets that mention Hospital Price Dispersion « Code Red -- — November 19, 2010 @ 6:22 pm

  2. Professor Dranove,

    Do you have any reading suggestions about measuring differences between hospitals re: severity within DRGs? I’ve seen Geweke, Gowrisankaran, and Town (2003) but that’s a bit too technical for me right now. Basically I’m looking for a “good” version of this kind of study that makes sensible adjustments in order to make cross-sectional price comparisons.

    Comment by Matt — December 9, 2010 @ 10:16 pm

    • Matt,

      I think there are lots of studies that document this but none come to mind since most of these studies would have been written for some other purpose. (You are write about Geweke et al being nearly inpenetrable!) I

      Every year Medicare publishes its hospital “caseweights” for every hospital which document easily measurable differences in severity but again I don’t have a specific reference.

      You can get this data for yourself if you are so inclined by purchasing (for $200) the “nationwide inpatient sample” from a group called HCUP. If you are clever with Excel you can soon document the variation in severity!

      Sorry not to have any more immediate citations.


      Comment by dranove — December 9, 2010 @ 10:53 pm

      • Thanks for the HCUP reference. I will definitely look into that.

        Thinking this through here (so apologies for a bit of rambling), I’m reminded of a study I heard about that found that brushing your teeth lowered your risk of heart disease. They control for other heart disease risk factors, but I still can’t really help but think “this HAS to be spurious.” It just seems like there’s too much to control for, many things which may be difficult to measure.

        I get the same feeling when I start thinking about price dispersion among hospitals. We have CMI to control for discharge mix, IRB to control for other outputs (i.e. research and new doctors), etc. However, then we run into things like quality (patient perceived? are mortality rates meaningful and/or properly risk-adjusted? etc etc) which seem remarkably difficult to measure.

        With the above in mind, it then seems that, for the existing state of the world, it is difficult to say that A or B hospital charges supracompetitive prices (because dispersion itself is not meaningful and there are only so many adjustments we can make). However, with a merger, we have this nice little natural experiment where all of a sudden there’s a more highly concentrated market but quality, etc. can probably reasonably be assumed to be more or less constant. Is that the general logic behind the retrospective analysis of hospital mergers and associated price effects?

        I’m still working through the literature on all this, but I’m trying to get some intuition too (not always in the mood for academic papers on a Saturday morning).

        Comment by Matt — December 11, 2010 @ 10:39 am

RSS feed for comments on this post.

Blog at

%d bloggers like this: