Contract Structure and the Performance of Technology Transfer Agreements: Evidence from University Licenses
Dan Elfenbein (Harvard University and University of California, Berkeley)

FULL TEXT

--Summary by Dan Elfenbein

Summary

Contracts form the basis for many economic relationships. By their nature, contracts generate incentives for the parties that enter into them. University technology licensing contracts, which typically specify up-front fees, royalties, milestones, and/or firm equity as payment, are no different. In this paper I investigate the role that these incentives play in the process whereby new inventions developed at universities enter the market. I find that provisions for contingent payments such as milestones (where a payment is tied to reaching a predefined technical or market objective) and royalties reduce the likelihood that licensees will terminate a license. I am also able to identify a separate effect in small firms, where R&D budgets may be smaller and more closely monitored.  In these firms, contingent payments may have an opposite effect, increasing the likelihood of termination.

Background and Theoretical Insight

University technologies tend to be licensed at a very early stage in their development. It is not unusual for universities to license technologies to private firms before an application for the technology is known. Typically, the licensee must make significant investments to determine whether the technology works and whether it has value in the marketplace. Additionally, further input from the inventor may be important for the success of the project. The importance of the inventor’s input may result from tacit information (i.e., information that cannot be written down as part of the license) about either the function of the technology itself or about its potential applications.  

Two recent theoretical papers highlight the importance of the incentives that university technology licenses generate. Decheneaux, et al. (2003) focus on licensees’ decisions about whether to continue investing in a new technology or to abandon it and terminate their license. In this model, developing a new technology is risky. The licensing party does not know exactly how long it will take to make the project a technical success, nor does it know how valuable the technology might be once it enters the marketplace. As long as the technology remains under development, the licensee continues to receive new information about its potential commercial value. If this information is good enough—in other words, if the expected value (including option value) for the licensee of continuing to invest in the technology is positive—the firm will continue to invest in the technology. Otherwise, the firm will terminate the license. Because the expected value to the licensee depends on the size and nature of the future payments it must make to the licensor, these contract terms enter into the firm’s decision about whether to continue its project. Other things being equal, higher levels of royalties or other contingent payments reduce the licensee’s incentives to keep the project going. 

Jensen and Thursby (2001) focus on the importance of the inventor’s input in bringing a licensed technology to market. It is common for inventors to have information about the licensed technology, its potential applications, or both. Writing a licensing contract that forces the inventor to provide this information is extremely difficult, because it is impossible for a court to know exactly what kind of information is “in the inventor’s head.” Therefore, the licensee must offer the inventor incentives that induce him or her to provide this expertise. Jensen and Thursby suggest that royalties, equity shares, and other payments that are contingent on the success of the licensed technology will induce the inventor to put forth such effort, as long as the university shares a substantial fraction of these payments with the inventor.  

Analysis and Results

Using data on the structure and performance of exclusive patent licenses between Harvard University and for-profit firms, I test the predictions of these two models. Specifically, I analyze whether the presence and the levels of royalties and other contingent payments affect the likelihood that a firm will give up on a project and terminate its license. One major empirical challenge results from the fact that the data are not generated by an experiment. Contract terms are not randomly assigned; in fact, they may relate systematically to other factors that affect the likelihood of early termination. I account for this possibility using a common method in economics, instrumental variables. In addition, the likelihood of termination may also relate to the specific characteristics of a technology. I attempt to control for as many of the observable characteristics of the technology as possible.  

The analysis yields some support for the incentive effects postulated by both Jensen and Thursby (2001) and Decheneaux et al. (2003). In particular, when contracts specify higher royalties and milestone payments, they tend to reduce the likelihood of termination. This finding supports the idea that such contractual provisions induce inventors to provide important input to the commercialization process. For small firms, however, the royalties and milestones have a countervailing effect: they increase the likelihood of termination. These contract terms reduce the ultimate payoffs to the licensees; if firms adhere to strict economic calculation about whether to continue the project or not, these terms are likely to play an important role. The difference in behavior between small and large firms may arise because small firms may react more strongly to their own incentives. 

Implications

This paper highlights two major issues that are important for university technology transfer policy. First, the paper suggests that the likelihood that a licensed technology will be commercialized is related to the form of the licensing agreement. University technology licensing offices face a complex set of incentives. Typically they must balance the need for revenue from commercialization with the need to serve the faculty and the periodic threat of outside regulation. The relationship between license structure and the likelihood of early termination can influence each of these considerations. The findings of this paper also suggest that contingent payments increase the likelihood of commercialization by inducing inventors to provide input to the project. If this is indeed the case, then changes in how universities share revenues with inventors may have major effects on the efficiency of licensing operations.  

References

Dechenaux, Emmanuel, Brent Goldfarb, Scott A Shane, and Marie C. Thursby. 2003. "Appropriability and the timing of innovation: Evidence from MIT inventions." NBER Working Paper No. 9735. 

Jensen, Richard and Marie C. Thursby. 2001. "Proofs and prototypes for sale: The licensing of university inventions." American Economic Review, 91:1, pp. 240-59.

This work is licensed under a Creative Commons License.