Hospital Market Power Drives Private Payer Prices
The cost of private health insurance, the main mechanism by which people access and pay for health care, is high and rising. In fact, the average family spends more than $17,000 annually on health insurance premiums (KFF). That is, even when a family consumes no health care, they are buying the equivalent of a base-model Toyota Corolla in health insurance every year.
Why does private insurance cost so much? Until recently, we knew very little about the prices that insurers pay providers for services. In fact, with a few exceptions, almost everything we knew about health care spending came from analyzing Medicare data, a setting in which prices are administratively set.
My recent NBER Working Paper with co-authors Zack Cooper, Marty Gaynor, and John Van Reenen, fills this gap by using a big dataset containing all of the claims from three large health insurers to provide a first look into the market for privately insured hospital care.
Why hospitals? Because that’s where the money is
Recently, prices for prescription drugs have received a large amount of media attention. To be sure, this issue is an important one, and deserves much attention. But prescription drugs represent a small fraction of national health expenditures. In 2014, spending on hospital care was more than three times the amount spent on prescription drug coverage, and more than 50% higher than the amount spent on clinics and physician payments.
Why so expensive?
Health care spending varies dramatically across the country. For the privately insured, spending per person ranges from roughly $1,700 to over $5,500 per year. Perhaps one of our most surprising findings was that the highest spending places for Medicare beneficiaries are not the same as those that are high spending for the privately insured. And in fact, some of the regions held up as policy exemplars using the Medicare data—places like Grand Junction, CO and Lacrosse, WI—are exactly the places that are most expensive for privately insured care. By contrast, Philadelphia, which is the 41st (out of 306 hospital referral regions) highest spending region under Medicare, is actually slightly below average when we look at private spending.
So, what is so different between Medicare and the privately insured? One big difference is that under Medicare, prices are administratively set, so the price varies little from place-to-place, and most of the variation in spending is driven by the amount of care people receive. In the private domain, however, we find that higher prices lead to higher spending.
We also find that prices for narrowly defined services vary substantially across the country, even within small geographic regions. For example, in the Philadelphia area, the prices of a knee MRI can range from roughly $500 to almost $3,000. This is worrisome since do not expect the quality of an MRI to vary substantially across providers.
Finally, we find that the strongest predictor of higher prices is the market power of hospitals. In particular, monopoly hospitals (hospitals without competitors within a 15 mile radius) negotiate prices with insurers that are 15% higher on average when compared to those with three or more competitors. We also find that clinical quality plays a very small role in determining price.
What to do?
Our work demonstrates the importance of studying the privately insured, and particularly the prices paid. We can no longer assume that Medicare data provides insight into the workings of private insurance. We see three immediate implications of our work.
First, instead of trying to replicate health systems that are cheap for Medicare and expensive for private payers, we ought to highlight places that are low spending for both.
Second, as my co-author Marty Gaynor notes, our work was possible only because three large insurers shared their data with us. The availability of all-payer claims databases (APCDs) would greatly enhance our ability to understand market dynamics in the health care industry. However, a recent Supreme Court ruling (Gobeille v. Liberty Mutual) held that self-insured large employers were exempt from a state mandate to participate in an APCD.
Third, we have to start thinking about hospital care as a market, where market principles apply; and in particular, our work highlights the importance of limiting market power via anti-trust enforcement or constraining providers’ ability to set prices.
Stuart Craig is a doctoral student in the Health Care Management Department at Wharton.