Economics Letters 41 (1993) 447-450 01651765/93/$06.00 0 1993 Elsevier Science Publishers B.V. All rights reserved 447 How well do patents measure new product activity? Timothy M. Devinney * zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA Australian Graduate School of M anagement, University of New South W ales, P.O. Box 1, Kensington, NSW 2033, Australia Received 6 October 1992 Final revision received ll.March 1993 Accepted 5 April 1993 Abstract The present research provides a preliminary examination of the relationship between patent intensity and new product development. Although patent intensity is related to commercial innovation at the industry level, patent intensity is not the best predictor of commercial innovation at the firm level. The major determinant of commercial innovation appears to be the patent intensity of the industry in which the firm is operating. zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLK 1. Introduction The research on the role and impact of innovation has grown dramatically in the last ten years but has not been without controversy [see Griliches (1990) for a survey]. One area of concern to innovation researchers is the quality of the surrogates used to measure the innovative output of firms. Historically, the two main measures of a firm’s innovativeness have been its research and development expenditures, either suitably deflated by a size measure and/or converted from an expense into an asset, and its patent intensity. More recently two new data sources have been developed. The first source uses announced innovations in all major publications for the year 1982 [Acs and Audretsch (1988)]. The second source, labeled C/D/W, uses all announced innovations in the Wall Street Journal for the 14-year period 1975-1988 [Chaney and Devinney (1992)]. Acs and Audretsch were able to show that the factors affecting innovation by large firms were different from the factors affecting innovation by small firms. More to our point here, they highlighted the fact that traditional published data sources biased the innovation/structure literature towards too strong a support for the positive firm size/innovation intensity relationship. Chaney and Devinney (1992) show confirmation of the general findings of Acs and Audretsch with respect to innovation - innovation tends to be done by larger firms, in less concentrated industries and is strongly affected by joint investments in advertising and research and development. In related research, Acs and Audretsch (1989) show that the innovation production function found for innovations appears to hold when patents are used to measure innovation intensity. One of the major weaknesses of the Acs and Audretsch research is the reliance on aggregate *Acknowledgement is given to the numerous comments received on this and related work by participants at recent seminars at UCLA, INSEAD, Wissenschaftzentrum Berlin, the University of Melbourne, and the AGSM.