Post-patent prescription pharmaceuticals

The three essays in this dissertation examine various economic and econometric issues in the market for post-patent prescription pharmaceuticals.;"Discrete Choice Estimation with Grouped Data: Further Results from Post-Patent Prescription Pharmaceuticals" examines three alternative estimators for the generic prescription discrete choice model described in the first essay, where the data are naturally grouped as multiple patient observations for a sample of physicians. The estimators differ in their underlying assumptions about the appropriate asymptotic approximation for the finite sample distribution of the data and the treatment of unobserved physician effects on the prescription decision. The empirical results of the three estimators are nevertheless all similar and concur with those reported in the first essay.;"Divisia Indexes: A Mixed Blessing for Pharmaceutical Price Indexes" is an empirical examination of the relative performance of Divisia price indexes and traditional fixed-weight indexes in characterizing price changes in the pharmaceutical industry. Two separate pharmaceutical data sets are used to construct Divisia, Laspeyres, and Paasche indexes, and the results of the various indexes are compared. The results from the first data set indicate that in most cases the Divisia index does do a better job than more traditional indexes at accurately portraying the long term price growth for various forms of cephalosporins. However, in the second data set, Divisia "drift," a form of bias in the index, is evident. Simply paying lip service to the possibility of the existence of drift is therefore not enough. Divisia drift can be a severe problem in certain cases, and it can dramatically and permanently alter the indexes. This should serve as a cautionary note to recent enthusiasts of Divisia indexes.;"The Demand for Post-Patent Prescription Pharmaceuticals" examines why physicians continue to prescribe trade-name drugs when less expensive generic substitutites are available. I utilize a data set on physicians, their patients, and the multisource drugs prescribed to study the prescription habits of physicians in prescribing generic and trade-name drugs. The results indicate that almost all physicians prescribe both types of drugs to their patients. There is, however, persistence in the prescription behavior of physicians, so that some physicians are more likely to prescribe trade-name drugs, while others more often prescribe generics. While most of this persistence cannot be explained by observable characteristics of the physician or the physician's patients, patients who are treated by physicians with large numbers of HMO or pre-paid patients are more likely to be prescribed generics, and there is wide regional variation in the propensity of physicians to prescribe generic drugs. The results are most consistent with an explanation of physicians' prescription behavior based on habit persistence.