A central tenet of industrial organization is that market concentration leads to higher prices. At the same time, there is a growing awareness that the market for health care is unique due to its complexity, and it is often difficult for consumers to make fully informed decisions. Might information frictions exacerbate market power in concentrated markets and lead to higher prices? While a growing literature seeks to address this question in a number of contexts, this dissertation focuses on the lack of price transparency in health care. I argue that the inability of consumers to compare prices is a major factor leading to high prices when health care is provided by the private market.
In Chapter 1, I use a dataset covering all private medical claims in a state to examine the introduction of a state-run website providing detailed information about out-of-pocket prices for a subset of medical procedures. Exploiting plausibly exogenous variation across procedures available on the website as well as the timing of the introduction, I use a difference-in-difference approach and find significant savings for both consumers and insurers. Part of the effect is due to consumers switching to lower cost providers. However, there is a small but significant supply-side effects in the long-run, i.e. there are lower negotiated prices. These lower prices benefit all insured individuals including those that do not use the website. Supply-side effects reduce price dispersion and are especially relevant when medical providers operate in concentrated markets.
A relatively small fraction of consumers actually used the price transparency website when it was available. Therefore, it is important to understand why more consumers aren't using the price transparency tool and what would happen if more consumers were informed about prices. Answering this question requires a structural model, which is the focus of Chapter 2 and Chapter 3.
In Chapter 2, I study demand for health care services when at least some consumers lack full information about prices. By exploiting the variation from the introduction of the website, I am able to separately identify consumer price sensitivity and the degree of uncertainty about prices. I also explicitly model the decision to use the price transparency website when it is available. This structural approach yields two main advantages over the reduced-form approach. First, the model can be used to examine what would happen if more consumers were incentivized to use the price transparency website. Second, the model provides insight into the welfare effects of price information.
Finally, the reduced-form evidence that there is a supply-side effect of the website when even a small fraction of consumers are informed motivates a more in depth analysis of the supply-side. Chapter 3 combines the demand model of Chapter 2 with a model of bargaining between medical providers and insurers to examine how price transparency affects equilibrium prices. Model estimates and difference-in-differences estimates both imply that the website reduces health care spending by 3 to 4 percent. I then use the model to examine the effects of price transparency more generally. In counterfactual simulations, I find that price transparency would generate a substantial reduction in equilibrium prices if a larger fraction of consumers in the market were informed. Combining the price transparency website with high cost sharing would give individuals more incentive to use the price transparency tool, reducing health care spending by 18 percent.
My research is intended to inform the policy debate surrounding the value of health care price transparency tools. In sum, I argue that while the value of price transparency tools is modest when only a small fraction of consumers are incentivized to use the tools, the savings become quite substantial when enough consumers are informed about prices.
Identifer | oai:union.ndltd.org:columbia.edu/oai:academiccommons.columbia.edu:10.7916/D8FJ2V14 |
Date | January 2017 |
Creators | Brown, Zachary Y. |
Source Sets | Columbia University |
Language | English |
Detected Language | English |
Type | Theses |
Page generated in 0.0022 seconds