During their lifetime individuals form rankings concerning the quality of healthcare services of hospitals located in their neighborhood as well as, especially when in need of specialized treatments, of those located nation-wise. Most of the recent theoretical models dealing with the quality of healthcare of competing hospitals use a horizontal differentiated setting à la Hotelling (1929) with regulated prices and fixed locations to explain their strategic decisions about quality of hospitals (see, for instance, Brekke et al., 2011 and the literature cited therein). However, often hospitals are vertically differentiated quality-wise, i.e. individuals naturally rank them unanimously vertically and not horizontally on the quality ladder. In addition, the transmission of information about the quality of local and national hospitals does not usually spread instantaneously among consumers. In reference to these conceptual points, the current project aims to firstly develop a model in which hospitals compete in quantities and qualities to provide vertically differentiated healthcare services and where individuals are not immediately aware of the quality of the existing hospitals, initially knowing only the quality of the hospitals located nearby. However, over time individuals meet other people and exchange information on the quality of existing hospitals, and this affects progressively their demands of healthcare services. As a second objective we plan to use an extensive panel data of NHS English hospitals elaborated by the research team covering the last twenty years to investigate the quality and quantity choices made by hospitals, with the aim to develop measures of the quality healthcare services.
References:
Hotelling, H. (1929) "Stability in Competition", Economic Journal, 39, 153, 41-57.
Brekke, K.R., Siciliani, L., Straume, O.R., 2011. Hospital competition and quality with regulated prices. Scandinavian Journal of Economics, 113, 444-469.
As mentioned above, the project plans to fill two major gaps of the existing theoretical literature as well to test empirically the obtained findings by means of an extensive twenty-year panel data collected for the NHS English hospitals.
In terms of theoretical contribution, two major innovations will be introduced by the model: (i) Individuals are heterogeneous in their willingness to pay for quality healthcare although unanimously recognizing an initial ranking between two or more than two hospital units (as in the Mussa and Rosen's, 1978) original model); (ii) concerning the quality of national hospitals, the knowledge of consumers is assumed initially incomplete among patients, and individuals can update their information through a process of knowledge transmission which occurs at every period through random encounters as in Gabszewicz, Marini and Zanaj (2021).
We expect to produce interesting and innovative results concerning the quality equilibrium of competing hospitals given that their demands progress dynamically at every period in accordance to the information transmission process occurring between consumers and as determined by the initial clusters of individuals distributed around the existing local hospitals: this directly descends from the different masses of consumers initially distributed in each local community. The effect of competition will, therefore, be analysed dynamically through a simple dynamic setting with vertically differentiated services and, hence, adopted to study in detail the problem of the effect of competition into the quality of healthcare in competing hospitals which, as explained above, is a major problem faced by the existing literature.
The empirical analysis will be extremely useful to validate the theoretical findings of the model. We will use the flexible conditional DID approach (Dettmann et al., 2020), a modification of the matching and DID approach of Heckman et al. (1998) for the staggered treatment adoption design (as in Callaway and Sant'Anna, 2020), where units that are treated once in the observation time are regarded as treated units from that date onwards and where time is defined in relation to the treatment start. Flexibility is gained in three ways: including individual treatment time information from the panel into the matching process; introducing a combined statistical distance function for matching; and incorporating a flexible observation durations into the DID estimation (Dettmann et al., 2020, p. 1). Flexibility ensures a variation in treatment timing so that variable treatment effects can be accounted for and at every point in time every individual (treated) is compared to his/her matched counterpart (untreated), even when treatment is administered in a staggered way.
As the flexible conditional DID approach is a combination of propensity score matching (PSM) and DID methodology, the conditional independence assumption for matching and the common trend assumption for DID are replaced by the conditional parallel trend assumption (as proposed by Callaway and Sant'Anna, 2020): unobservable individual characteristics must be invariant over time for units with the same observed characteristics (Dettmann et al., 2020, p. 9). Also, as for PSM, the common support condition must be satisfied. Additionally, the approach assumes no spillover effects (due to the assumption of constant value of unit treatment for matching), and that potential carryover effects do not influence the matching variables at the matching time (see also Imai et al. 2020).
Callaway, B., P. Sant'Anna. (2020), "Difference-in-differences with multiple time periods". SSRN Research Paper.
Dettmann, E., A. Giebler, and A. Weyh. Flexpaneldid: (2020), "A Stata toolbox for causal analysis with varying treatment time and duration". IWH Discussion Papers, 3.
Gabszewicz, J. J., Marini, M. A., Zanaj, S. (2021), "Random Encounters and Information Diffusion about Product Quality", FEEM Working Papers, 2021/002
Imai, K., I. Kim, and E. Wang, (2020), "Matching methods for causal inference with time-series cross-sectional data". Harvard University Working Paper, 2020.
Mussa, M. and S. Rosen (1978), "Monopoly and Product Quality", Journal of Economic Theory, 18, 2, 301-317.