We examine the relationship between the knowledge capital and the stock market value of innovating firms in the pharmaceutical and semiconductor industries in the United States. We use unique patent-firms matched data in the two industries and estimate Tobin' s q equation on the R&D intensity, patent yield of firms, and citations to patents. We find the duration of patent citations received explains the variance of a firm's market value better than the conventional variable of knowledge capital stock. This relationship becomes stronger when the number of citations received 4 or more years after a patent is granted is used as a proxy for the knowledge stock. Use of number of citations received in long-term fixes heterogeneity problem of patent values, thus enables us to better study how firms organize their intangible assets, in particular knowledge stock, and how these are reflected in their future cashflows.
(ProQuest: ... denotes formulae omitted.)
The output of the R&D investment activity conducted by innovating private firms is an intangible asset to the firm, thus, the size and quality of this intangible asset should be reflected in the market value of the firm. Many studies use patent counts as a proxy for R&D success and estimate the market value of firm.
Since the study of patents (Scherer, 1 965), patent statistics have been widely used to answer questions about sources of economic growth, technological change, and innovation activities. Economists (Schmookler, 1966; Griliches, 1984) use patent counts as indicators of innovation output. However, major problems exist in using patents as output indicators. That is, i) not all inventions are patentable, ii) not all inventions are patented, and iii) inventions that are patented differ in quality. The first two problems can be taken care of by industry dummies, or by limiting the analysis to a particular sector or industry (Griliches, 1990). One study (Pakes and Schankerman, 1986) provides an evidence of the third problem using patent renewal data.
One of remedies on the third problem of patent quality issue is using citation-weighted patent counts (Trajtenberg, 1990; Lanjouw and Shankerman, 1999; Jaffe, 1993). Patent citations trace linkages between inventions, inventors, scientists, firms, etc. Hence, citation weighted patent counts allow us to create indicators of the importance of individual patents, thus able to capture the heterogeneity in the value of patents. A recent study (Hall, Jaffe, and Trajtenberg, 2005) utilizes information on the number of subsequent citations received by a firm's patents to cure this problem. They find the market value of a firm has higher correlation with patent citation counts than patent counts themselves using Tobin's Q for the firms.
However, patent and patent citation data may contain bias from firm's propensity to patent due from speed of technological change, size of market, degree of R&D competition, and importance of patent protection against imitation across technology, sectors, and industries. Patent citation counts come from two sources. The first source is the importance of underlying innovation of a patent. The second comes from firm's patenting activities and propensity to patent. A patent may receive greater number of citations where firms in the industry are involved in more active patenting activities, for example, although the value of underlying innovation of the patent is not significant. In a shorter term two effects are mixed so we cannot distinguish the sources. In a longer term the important patents will be continuously cited while the second source diminishes. Thus patent citations in a longer term better measure the ex-post value of patent than total number of citations.
In this paper, we test whether information of patent citations received long after a patent is granted is a better measure of firm's R&D …