Why Firms Form (Or do not Form) RJVS

Abstract

In this article, we examine why it is difficult to induce firms to form Research Joint Ventures (RJVs). We examine various incentives and disincentives for RJV formation by estimating an endogeneous switching model using data from the US National Cooperative Research Act. The empirical findings support hypotheses that firms of different sizes have disincentives to form RJVs and that cost-sharing is an important incentive for RJV participation. In the early 1980s, a shift in technology policy took place in both the US and Europe. This was seemingly motivated by increased international competition, particularly from Japan, in high-technology sectors. Many scholars, policy makers and industrialists identified Japan's more cooperative business environment as a competitive advantage (Jorde and Teece, 1990). Japan's 1961 Act on the Mining and Manufacturing Industry Technology Research Association and the proactive efforts of MITI that encouraged joint ventures were identified as policy tools by which the Japanese created such a cooperative atmosphere. The response by US policy makers was to enact the 1984 National Cooperative Research Act (NCRA) and to provide government support for ventures such as SEMATECH.' In Europe, a block exemption for Research Joint Ventures (RJVs) was provided under EU Competition Law. Furthermore, billions of euros have been spent in framework programmes in the EU, which subsidises R&D undertakings only when different organisations work in collaboration with one another. Despite these policy changes, the number of RJVs registered under the NCRA has been quite small, ranging from 17 in 1986 to 67 in 1993 and averaging about 40 RJVs per year from 1985 to 1994. Thus the question: why is it so difficult to induce firms to form an RJV? The goal of this article is to determine empirically what the incentives are for firms to participate or not participate together in RJVs. Our study is based on US data available through the NCRA. In order to obtain consistent evaluations of the hypotheses, we address two important empirical issues. Since much of the literature suggests a link between RJVs and R&D expenditures (see below), we must control for changes in R&D expenditures and its effect on RJV formation. As we can observe R&D changes at the time of RJV formation for only those firms that form an RJV, we have a missing data problem. To deal with this first issue, we need a consistent estimate of the expected effect on R&D expenses when RJVs are not formed. A second issue is one of simultaneity between the change in R&D expenditure and the decision to form an RJV. The decision to form an RJV may be determined by its impact on R&D expenses. R&D investments, however, are in turn also determined by RJV participation, which implies