- This chapter is adapted from a March 2001 article in Weltwirtschaftliches Archiv, 137(1):58-79. It is a substantial revision of a chapter of Yang's doctoral dissertation at the University of Colorado at Boulder. We are grateful to James R. Markusen, Robert McNown, Thomas Rutherford, Marie Thursby, and an anonymous referee from Weltwirtschaftliches Archiv for valuable suggestions.
How international differences in intellectual property rights (IPRs) affect decisions to license technology is an important question that has attracted virtually no econometric study. Licensing is a crucial component of international technology transfers, but we have little systematic evidence about whether it is much influenced by the strength of local patent regimes. In pushing for the recent introduction of global minimum standards in IPRs through the World Trade Organization, governments in technology-exporting nations argued that stronger IPRs would encourage technology transfer and local adaptive innovation, allowing all regions of the world to benefit. However, some developing countries argued that tighter protection would only strengthen the monopoly power of potential licenser firms, which are largely based in industrial countries, to the detriment of technology-importing nations.
As we discuss later, the theoretical links between IPRs and technology trade are numerous and depend on varying circumstances. No unambiguous analytical Page 112 prediction may be made, making the issue inherently empirical. On the one hand, stronger IPRs reduce imitation risk faced by the licenser, reduce licensing cost, and increase the licenser's rent share. These effects increase economic returns to the licenser and raise its incentives to innovate and license. On the other hand, the licenser has more monopoly power, thanks to tighter protection, and its incentive to innovate and license is correspondingly reduced.
Licensing has attracted little attention in the empirical literature. Much of it focused on operations of foreign subsidiaries of U.S. multinational enterprises (MNEs). Ferrantino (1993) found little effect from IPRs, as measured by membership in international conventions, on overseas affiliate sales of U.S. firms to their various trade partners in 1982. Similarly, Maskus and Konan (1994) could not find any relationship between an index of patent rights and the international distribution of foreign direct investment (FDI). In contrast, Lee and Mansfield (1996) examined the relationship between business perceptions of the strength of a country's system of IPRs and the volume and composition of U.S. FDI in that country. They found that countries in which the IPR regimes were perceived to be weak tended to attract lower FDI volumes.1
Regarding arm's-length licensing, three earlier studies are relevant. Survey evidence in Mansfield (1994) indicated that U.S. MNEs were less likely to transfer advanced technologies to unaffiliated firms in countries with weak patent rights. In terms of econometric work, Contractor (1980) examined a sample of 102 contracts for technology licenses and showed that total returns on licensing were higher in patented technologies than in others. Ferrantino (1993) used 1982 cross-country data to show that membership in the Paris Convention, which stipulates that patents will be awarded to foreign applicants without discrimination, stimulated flows of U.S. receipts of unaffiliated royalties and license fees from the host country. However, this effect pertained only if the host country's domestic IPR regime was sufficiently strong, as measured by patent duration.
Licensing to unaffiliated firms is a significant activity. For example, U.S. receipts of unaffiliated royalties and license fees were 21 percent of the country's total royalties and license fees received from all the countries in the world in 1995. In 1996, royalties and license fees from unaffiliated foreigners increased by 9 percent, reflecting an increase in both fees for the use of industrial processes and for the right to sell products under a particular trademark, brand name, or signature (U.S. Department of Commerce ).
Our purpose here is to investigate the influence of international variations in patent regimes on global flows of technology, both within MNEs and at arm's length. In comparison with earlier studies, we use more recent data, assembled for several countries and over time, and a more accurate measure of IPRs to investigate Page 113 the link between patents and licensing. Numerous industrial and developing countries substantially strengthened their patent regimes between 1986 and 1995. We exploit this fact by developing a panel data set for the years 1985, 1990, and 1995 in 23 recipient nations to examine the effects of patent strength on the flow of unaffiliated royalties and license fees by U.S. firms in both absolute and relative terms.
We begin in the next section by describing a theoretical model that analyzes the role of IPRs in encouraging licensing. The essential point is that licensing could rise or fall with the imposition of stronger IPRs. Given this ambiguity, we go on in the third section to specify an econometric model of technology contracting. Data are discussed in the fourth section, while the fifth section presents the empirical results on licensing. The key findings are as follows: First, stronger patent laws have positive and significant effects on both the absolute flows and the relative flows (relative to trade volume) of U.S. receipts of unaffiliated royalties and license fees when the initial degree of patent protection in the technology recipient country is higher than a critical value. Second, the results indicate that patent rights are more important in promoting arm's-length technology trade relative to licensing through FDI. Concluding remarks are provided in the final section.
In principle, IPRs play an important role in technology trade. Patent protection is perhaps the most important means of safeguarding proprietary technology. Patent laws vary markedly across countries in terms of coverage, membership in international agreements, loss of protection, enforcement mechanisms, and duration (Primo Braga 1996). These differences in patent laws in host countries might influence licensing through several channels.
First, the degree of IPR protection could influence the decision to earn returns on a new technology by licensing it rather than exporting it through trade in products. Tighter patent protection should reduce imitation risk, uncertainty, and transaction costs involved in technology contracts, thereby encouraging licensing relative to commodity trade.
Second, the degree of IPR protection could influence the choice the firm makes between licensing and FDI (Horstmann and Markusen 1987). Strong IPRs could favor licensing by creating a legal framework for the enforcement of licensing and royalty contracts. In the presence of weak patents, problems of transacting information with licensing, such as the nonexcludability property of new knowledge, informational asymmetry, imitation risk, and transfer costs, could provide an internalization motive for FDI (Markusen 1995).2 Among these problems in licensing, Rugman (1986), in particular, views imitation risk faced by the firm from licensees as a cornerstone of internalization theory. If IPRs are weak in the Page 114 host country, the licensee could learn the technology quickly and imitate it to start a new domestic firm in competition with the MNE. In response, the MNE might exploit its firm-specific assets through internalization (Markusen 2001; Taylor 1994). Seen in this light, stronger patents reduce the imitation risk faced by the multinational firm and create a legal framework for the enforcement of licensing contracts, thereby encouraging licensing. Stronger patents also favor licensing because they reduce the legal costs associated with establishing and policing an arm's-length relationship.
Third, the strength of IPRs affects the sharing of rents between the licenser and licensee. Rent sharing is one of the salient features commonly observed in licensing contracts. Caves, Crookell, and Killing (1983) indicated that licensers earned, on average, 40 percent of the rents from innovation. One important strategic factor is that license rents are used by the licenser to deter imitation. Stronger patent protection makes it harder for the licensee to imitate, causing it to commit not to imitate at a lower rent share. But to preclude imitation, lax IPRs require higher rents for the licensee. Gallini and Wright (1990) showed that when imitation is possible and there is asymmetric information, the licenser sacrifices some rents though its share rises with imitation costs. Accordingly, the rent share accruing to the licenser rises with patent strength, raising the returns to licensing.
These effects suggest that unaffiliated licensing should rise as countries strengthen their patent regimes because of lower licensing cost and higher rent share. We call this possibility the economic returns effect. However, the ability to license depends on the pace at which new innovations are introduced by potential licensers. It is conceivable in theory that long-run innovation could slow down because the monopoly effect of stronger patents would reduce research and development (R&D) effort.3 In turn, licensing could be reduced. We call this possibility the monopoly power effect.
These ideas have been formalized in a dynamic general equilibrium model with endogenous innovation and licensing to study the effects of stronger IPRs in the south on the incentives of firms in the north to innovate and to license...