Association for Public Policy Analysis & Management

Well-known as a costly process, biopharmaceutical firms invested an estimated $51.2 billion in R&D in the U.S. in 2014, an amount representing 23.4% of their domestic sales. Meanwhile, most research-intensive small biotech firms are financially constrained, where internal funding often fails to cover R&D cost and external funding cannot be guaranteed. In addition, the separation between firm owner and managers create potential agency problem in the innovation decision (i.e., new drug choice, resource allocation across drugs in different stages). Therefore, substantial inefficiency may arise, if the product manager manipulates the new product choices in order to maximize short run objectives (such as gaining external funding from investors, building future alliances), instead of maximizing the expected NPV of new projects for the firm. Consequently, under-investment can happen for certain therapeutic categories, slowing down the overall innovation to an ex ante non-optimal level. One widely employed business strategy is co-opetition: once a new drug showed some promise in very early stage, firms seek to build drug development alliances to share resources within the alliance, even though they are still competitors in other projects. Admittedly, co-opetition may be useful in mitigate the early stage inefficiency, but the overall effect remains unclear, due to ambiguous effect of incentives vs cost sharing across organizational forms.

In this paper, I characterize the agency problem in pharmaceutical innovation with a motivational incomplete contract model, and test whether project co-opetition can mitigate the inefficiency. Empirically, I use the project-level data from PharmaProject during 1989-2004. Firm-specific product choices and Phase-specific survival rates are documented for big pharma and small biotech, in terms of new products’ degree of innovation vs lucrativeness. I construct non-parametric innovation index based on drugs’ world development status, relative difficulties across therapeutic categories, and bio-similarity; and define lucrativeness based on sales data and market shares. I define co-opetition as the different project integration status: whether a drug is developed with full ownership, in a joint venture, or being acquired. Using both parametric and non-parametric logit regression of drug’s therapeutic-specific survival rates on firm characteristics and across co-opetition status, I demonstrate reduced-form evidence of heterogeneous impact of co-opetition on pharmaceutical companies' innovation behavior across firm types and degree of innovation. I further estimate a dynamic optimization problem for managers’ pre-clinical drug portfolio choice, anticipating potential co-opetition in later stage. I will try counterfactual policy analysis on certain firm-specific regulations or therapeutic-specific subsidies (e.g., subsidize antibiotic development vs. cultivating global joint venture), to explore the potential welfare improvement in drug development intervention through a market channel.