Please use this identifier to cite or link to this item: https://hdl.handle.net/10419/174342 
Year of Publication: 
2017
Series/Report no.: 
MAGKS Joint Discussion Paper Series in Economics No. 46-2017
Publisher: 
Philipps-University Marburg, School of Business and Economics, Marburg
Abstract: 
In the context of the ongoing debate about an innovation crisis in the pharmaceutical industry, we study the success rates of pharmaceutical R&D projects as a measure of innovative productivity. The empirical literature suggests success rates have been decreasing during recent decades. We critically review this literature and only find few studies with a focus on the development of success rates over time. Further, the empirical analysis of success rates imposes difficulties with respect to methodological aspects like data censoring, the definition of success, and the range of firms included in the samples. These difficulties are generally not discussed by the literature. We therefore discuss these issues when critically reviewing the empirical studies and complement this discussion with own empirical results. While most other studies use samples containing a small number of firms and cover just a short time period, we use a broad sample containing firms of different sizes over an observation period of more than 20 years (1989-2010). Descriptive results suggest a declining success rate of pharmaceutical projects during recent years. Correcting for censored observations shows there has been a stabilization of success rates, but at a lower level than before. The main underlying reason for a lower success rate is the start of many more projects in more recent time periods. Results from hazard rate models even suggest there has only been a temporary drop in the success rate for projects between 1995 and 2002.
Subjects: 
pharmaceutical R&D
drug development
innovation
success rates
JEL: 
O32
L65
Document Type: 
Working Paper

Files in This Item:
File
Size





Items in EconStor are protected by copyright, with all rights reserved, unless otherwise indicated.