Patterns of Public-Sector and Private-Sector Patenting in Agricultural Biotechnology

Over the past two or more decades, the structure of agricultural input industries has changed very rapidly. Private-sector investment in agricultural and food research and development (R&D) has grown dramatically, while public-sector investment has remained relatively constant. Private-sector plant breeding has been the fastest growing segment of the private research portfolio. Mergers, acquisitions, strategic alliances, and some divestiture in recent years have characterized the sector (Shoemaker et al., 2001). The use of intellectual property (IP), in the form of plant variety protection certificates and utility patents, has also expanded very rapidly over roughly the same time period. More than any other component of agricultural R&D, agricultural biotechnology exemplifies these trends. In the literature on R&D, patents have often been used as a measure of inventive activity, because patent data are a relatively easy measure to acquire (see also Brennan, Pray, Naseem, & Oehmke, this issue). This article reports preliminary and descriptive data from the Agricultural Biotechnology Intellectual Property (ABIP) database (United States Department of Agriculture Economic Research Service [USDA ERS], 2004), in which agricultural biotechnology patents are grouped in several ways: by the type of institution (US or nonUS firm, university, or government) to whom the patent is assigned, and by the technology type under which the patent is classified in the ABIP database.1 These data are used to explore the question of what kind of institutions have been performing particular kinds of research in agricultural biotechnology. We will also consider whether patent data alone are sufficient to answer this question. Use of patent data is not without its difficulties. Patents are sometimes used as a measure of output of R&D activities (Griliches, 1990) and sometimes as a measure of inputs into the production of goods and services (Schmookler, 1954). Even when the analysis is primarily restricted to private firms, it is difficult to relate patent data to economically relevant industry or product groupings. Many patents turn out to have very little economic value, while a few may prove to be extremely valuable (Griliches, 1990). Moreover, the “economic value” of a patent may be hard to define, as firms may use patents for a variety of reasons—for example, to block competitors’ products, as bargaining chips in cross-licensing negotiations, or to prevent infringement suits—besides protecting returns to specific inventions (Cohen, Nelson, & Walsh, 2000). Public-sector institutions, such as universities or government agencies that patent, may do so for reasons differing from those motivating private firms (Henderson, Jaffe, and Trajtenberg, 1998; Jaffe & Lerner, 2001; Jensen & Thursby, 2001; Maredia et al., 1999). Finally, patent data are only one kind of indicator—other research input related data might include R&D expenditures, and other research