| Summary: | My dissertation consists of three essays in Industrial Organization and Health Economics, each paper applies the tools of empirical industrial organization to a health economics topic. The first essay, Exclusive Territories and Entry: Welfare Effects Estimation (co-authored with Joan-Ramon Borrell), is an empirical estimation of the welfare effects of changing geographical entry restrictions in pharmacies in Pamplona, Spain using a structural model. The second and third
essays are related in that they both tackle issues within the Dialysis Industry. Economies of Density and Economies of Scale in the New England Dialysis Market (co-authored with He Wang), presents an estimation of the economies of density and scale in the dialysis industry using a trans-log cost function. The third essay, The Effect of the Five-Star Quality Rating in the ESRD Industry, estimates the effect of an one-star quality ratings increase on the number of treatments offered at
dialysis facilities using principal component analysis to control for the underlying quality level of the dialysis providers. The first chapter, titled "Exclusive Territories and Entry: Welfare Effects Estimation" and co-authored with Joan-Ramon Borrell, is a study on the welfare effect of a change in government- mandated exclusive territories of retail competitors. Exclusive territories for pharmacies are a common practice in the European Union; they give pharmacies the exclusive
rights to sell retail drugs in an area around the drugstore's location. Despite the popularity of these exclusive rights, very little empirical evidence is available about the welfare effects of this type of policies. We estimate the parameters of a structural entry model in a retail market where consumers have spatial preferences and pharmacies' fixed cost varies over space. We use a two-stage estimation approach that is consistent with the presence of multiple equilibria. In the first
stage we estimate a binary response model of entry in reduced form that ignores the model's strategic components. Then, using the estimation in the first stage as an approximation of firms' beliefs, we proceed to estimate the structural parameters of the model via maximum-likelihood. This methodology allows the recovery of the parameters of the demand function and to approximate the effect of consumers' reduction in transportation cost caused by the entry of new pharmacies and the
change in firms' profits due to business stealing effect of entry. To verify the sensitivity of the of the counterfactual analysis' results to the estimation methodology we use four different methods to estimate the first stage choice probabilities; two parametric methods (logit and probit) and two non-parametric techniques (local polynomial and kernel smoothing) and found that the results are not sensitive to the technique of choice in the first stage. Additionally, we compute the
model equilibria under different, and often conceptually opposite, equilibria selection criteria to show that the results don't dependent on selecting a particular equilibrium. Despite having very limited information about the individual drugstores' and demand characteristics, we found that the firms strategy profiles at the different equilibria are consistent. This allow us to conclude that the reduction in the exclusive territory parameter was socially inefficient. Entry occurs in
areas that are close to existing firms, and the estimated changes to consumer surplus are relatively small compared to the reduction in variable profits caused by entry. The second essay is titled "Economies of Density and Economies of Scale in the New England Dialysis Market" and co-authored with He Wang. Economies of density or cost sharing among neighboring production units of the same firm, is a mechanism, similar to economies of scale that potentially can give firms a cost
advantage over the competition. In the presence of economies of density, firms with a large network of clustered facilities will tend to have a lower average cost of production than firms with smaller or more sparse networks. The cost reduction may be caused because as firms expand their networks they also gain market power in the market for inputs or they share inputs among neighboring facilities. In a sample of dialysis facilities in New England from 2012 to 2016, we found that chain
organizations tend to create geographical clusters of facilities at a higher rate than independent providers. In this sample, in 2016 the two largest chain organizations, Davita and Fresenius, have a combined 77% of market share, measured by using the number of hemodialysis treatment. Out of the 190 facility centers, Davita owns 45 facilities, and, Fresenius owns 77. This high degree of concentration, paired with the fact that chain organizations tend to create geographical clusters,
suggests that there may be some structural aspect in the production process of ESRD care that allows larger firms to thrive and limit the degree of competition in the market. The goal of this paper is to test the hypothesis that this unusual spatial concentration from chain organizations, can be explained in part by the presence of economies of density and economies of scale. To estimate economies of density and scale we follow a similar methodology to the literature of estimating
economies of scale using a trans-log transformation of the CES production function. We start with a CES production function for each facility and, under the assumption that the total factor productivity depends on the density of facilities for a given firm, we are able to estimate a linear approximation of the average total cost function for each facility, then we recover the parameters of the production function to determine the degrees of economies of scale and density simultaneously.
Due to concerns about the endogeneity of location choices and quantity, we use 2SLS and facilities and year fixed effects. We could not reject the hypothesis that economies of scale were present in the production process, but were able to reject the hypothesis of economies of density. The third essay, titled "The Effect of the Five-Star Quality Rating in the ESRD Industry" study the effects of ratings in the Dialysis industry. In 2015 the Centers for Medicare and Medicaid Services (CMS)
started publishing a five-star ratings system for dialysis facilities via the Dialysis Facility Compare (DFC) website. The goal was to create an easy to understand and intuitive measure of quality that patients can check when deciding what dialysis facility to visit. In this paper I attempt to estimate the impact of the 5-star ratings in the hemodialysis industry. If consumers' choice are very sensitive to the facilities' quality ratings then we should expect providers to have a
stronger incentive to invest in quality. Then, the introduction of a quality ratings system can serve as a mechanism to intensify competition among providers in quality. The five-star quality ratings summarizes the information of 9 different quality of care clinical measures of a dialysis facility into a single quality indicator. The ratings system is produced using principal component analysis (PCA) on a national sample of dialysis facilities, including quality measures for
hemodialysis, peritoneal dialysis, and pediatric hemodialysis. The methodology for creating the five-star ratings is described by the CMS and the clinical data is publicly available via the DFC website. This makes it possible to construct a continuous quality index. Using such index we can compare the effect of having an additional start for facilities that have similar quality. I found that a star difference among facilities with similar quality levels causes an approximate increase of
about 3 to 6 percent in the annual revenues of the typical dialysis facility. I found that firms' short term strategies remain unchanged after the introduction of the index, suggesting that the gains in revenues were not sufficient to justify new investment in quality of care.--Author's abstract
|
|---|
