Mergers and Changing International Information Flows
Paul N. Isely and Gerald P. W. Simons, Grand Valley State University
This paper examines how merger activity between companies based in different countries affects international information flows. When companies merge, they gain the knowledge previously acquired by each other. Jaffe, Trajtenberg, and Henderson [1993] find that citations to domestic patents are more likely to be domestic themselves, and to come from the same state and the same metropolitan statistical area, relative to a control. Along the same line, Jaffe and Trajtenberg [1999] find that patents assigned to the same firm are more likely to cite each other, and inventors are 30-80% more likely to cite colleagues from the same country. There thus appears to be a significant spatial aspect to knowledge flows. When companies merge they then have access to multiple geographic areas. This should lead them to better utilize information from the countries in which the companies are based.
In this paper, the manner in which mergers affect information flows is investigated using data about mergers in manufacturing. The U.S. PTO and the Lexis-Nexis patent databases are used to count international patent citations as a measure of information flows between countries. The actual measurement of knowledge flows is clearly an important aspect of research on spillovers. Jaffe, Fogarty, and Banks [1998] and Jaffe, Trajtenberg, and Fogarty [2000] find patent citations to be a valid, albeit “noisy,” measure of technology spillovers, and conclude that “aggregate citation flows can be used as proxies for knowledge-spillover intensity…between countries” [Jaffe, Trajtenberg, and Fogarty, 2000, p. 218].
Additional information on firms is found using COMPUSTAT, manufacturers’ annual reports, and discussions with industry consultants. The NBER Trade Database, the National Trade Databank, the U.S. Bureau of the Census, and the National Science Foundation are used to get industry-level data. Controls are created using this data for R&D spending, international trade, and government R&D budget authority.
The creation of a new patent is influenced by many factors. However, in models of learning-by-doing, the production of patents is often specified generally as:
(1) P = A f(Z)
where
P = output measured by patent production;
A = stock of existing knowledge;
Z = factors of production employed in research.
New knowledge is created using factors such as labor and capital in combination with existing knowledge.
In this study the simple linear model using fixed effects is adapted to capture the effects of international spillovers. R&D spending proxies for the quantity of research inputs employed by the firm. The base of existing knowledge, A, is proxied by the number of existing patents owned by the firms included in the study. International spillovers are measured by citations to foreign patents in patent applications. Federal government expenditure on industry R&D and the volume of trade among the countries may also influence the rate of innovation.
Preliminary results suggest several interesting findings. As theory indicates, after an international merger firms do appear to increase their citations from each other's countries. Interestingly, firms also appear to increase the number of citations from other countries, thereby widening the geographic area from which they get technological insights. However, firms also seem to decrease their overall patenting activity for several years after the merger takes place. Therefore, the short-run effect does not unambiguously result in greater information flows, but there is an increase in information flows in the long-run.