TY - JOUR AU - Cockburn, Iain AU - Henderson, Rebecca AU - Stern, Scott TI - The Diffusion of Science-Driven Drug Discovery: Organizational Change in Pharmaceutical Research JF - National Bureau of Economic Research Working Paper Series VL - No. 7359 PY - 1999 Y2 - September 1999 DO - 10.3386/w7359 UR - http://www.nber.org/papers/w7359 L1 - http://www.nber.org/papers/w7359.pdf N1 - Author contact info: Iain M. Cockburn Questrom School of Business Boston University 595 Commonwealth Ave Boston, MA 02215 Tel: 617/588-1486 E-Mail: cockburn@bu.edu Rebecca Henderson Heinz Professor of Environmental Management Harvard Business School Morgan 445 Soldiers Field Boston, MA 02163 Tel: 617/495-8014 Fax: 617/496-4072 E-Mail: rhenderson@hbs.edu Scott Stern MIT Sloan School of Management 100 Main Street, E62-476 Cambridge, MA 02142 Tel: 617/253-3053 Fax: 617/253-2660 E-Mail: sstern@mit.edu AB - Recent work linking the adoption of key organizational practices to productivity raises an important question: if adoption increases productivity so dramatically, why does adoption across an industry take so long? This paper explores this question in the context of one particularly interesting practice, the adoption of science driven drug discovery by the modern pharmaceutical industry. Over the past two decades, the established pharmaceutical industry has slowly shifted towards a more science-oriented drug discovery: (a) adopters experienced substantially higher rates of R&D after the late 1970s and (b) the rate of adoption across the industry was extremely slow. Motivated by the apparent contradiction between large boosts in performance and slow rates of adoption, this paper characterizes the sources of differences in rates of adoption between 1980 and 1993. The principal finding is that adoption of a science-oriented research approach was a function of initial conditions, or subject to 'state dependence': some firms simply began the sample period at a much higher level of science orientation. Moreover, while these effects attenuated over time, our empirical results suggest that it took more than ten years before adoption was unrelated to initial conditions. In addition, consistent with theories developed in the context of technology adoption, we find that relative diffusion rates depend on the product market positioning of firms. More surprisingly, adoption rates are seperately driven by the composition of sales within the firm. This latter finding suggests the potential importance of differences among firms in terms of the internal structure of power and attention, an area which has received only a small amount of theoretical attention. ER -