The industrial organization of health care markets

That’s the title of a new NBER working paper by Martin Gaynor, Kate Ho, and Robert Town. I confess to not reading the whole thing thoroughly, yet it’s worth knowing about if only for reference. Here are a few highlights:

The hospital sector, at $850.6 billion, represents 5.6% of US GDP, physician services constitute 3.6%, and health insurance is 1%. [Footnote: Health insurance is measured as the “net cost of health insurance,” the difference between premiums collected and benefits paid in a calendar year.] This makes the hospital and physician sectors some of the largest industries in the US economy, larger than construction (3.6% of GDP), mining and oil and gas extraction (1.95%), agriculture (1.37%), computer and electronic products (1.29%), broadcasting and telecommunications (2.5%), automobile manufacturing (0.33%), or even breweries (0.11%).

Even breweries!!! 

Employers pass through higher health care costs dollar for dollar to workers, either by reducing wages or fringe benefits, or even dropping health insurance coverage entirely (see, e.g., Gruber, 1994; Bhattacharya and Bundorf, 2005; Baicker and Chandra, 2006; Emanuel and Fuchs, 2008) [Links added.]

I know this well-supported fact is very hard for some people to swallow. Consider this a test of how scientific you are.

We structure our discussion of this literature around a multi-stage model of the market. In the first stage providers (hospitals and physicians) make investments that determine their quality. In addition to being influenced by horizontal competition, these investments may be affected by demand-side factors such as the amount of information on quality that is provided to consumers and the amount of choice they are offered when they need medical care. In the second stage, given their quality levels, providers negotiate with insurers to determine insurers’ provider networks and the prices paid to providers. This complex interaction has been analyzed in a growing number of papers; it has substantial implications for consumer welfare and for costs. Third, insurers choose their premiums to maximize their objective functions, taking into account their own characteristics and those of competing insurers. In the fourth stage consumers observe each insurer’s provider network and other characteristics, including premiums, and choose their insurers. Finally, when the enrollment process is complete, some consumers get sick and utilize providers either from within their insurers’ networks or (incurring a larger out-of-pocket payment) from outside the network.

Each stage of this model has an impact on the equilibrium outcome, and therefore on welfare. Clearly every stage is related to the others: optimal choices in one stage are functions of expectations regarding the rest.

It’s probably surprising to non-health economists how much we don’t know about health care markets, though we know some things. This passage suggests why. It’s extremely complex. The authors continue,

[V]ery few papers try to address more than one or two stages of the model, in part because of modeling issues (model complexity and the difficulty of identifying a complete model of all the stages), and in part due to lack of data required to estimate a complete model. Most instead focus on one of the stages.

The rest of the paper, which walks through what is known about the various linkages in health care markets, is worth at least a skim, just to know what’s in it.


Hidden information below


Email Address*