Maximizing Innovation in Alliances

Should birds of a feather flock together? Not if the goal is to promote innovation, says Rachelle C. Sampson, assistant professor of logistics, business and public policy at the University of Maryland’s Robert H. Smith School of Business. After studying hundreds of research and development (R&D) alliances, Sampson reports in her 2004 working paper R&D Alliances and Firm Performance: The Impact of Technological Diversity and Alliance Organization on Innovation that alliances work best when there are moderate differences in technological capabilities between the partners. What’s more, she finds, the organizational form of an alliance can have a significant effect on the amount of innovation it yields.

Alliances have become increasingly popular as companies look for a higher return on their R&D investments. Alliances provide access to complementary capabilities, as well as economies of scope and scale. They often offer a nimbler alternative to that of in-house development, while allowing companies to share risks and costs. But many alliances fall short of their founders’ expectations. Knowledge transfer is a tricky business, particularly when knowledge is complex or tacit. Moreover, mutual distrust often hinders learning within alliances. Today’s partner, after all, may turn out to be tomorrow’s competitor.

In order for alliances to succeed, Sampson observes, the partners must be both able and willing to share knowledge-based capabilities. Two factors, in turn, influence the odds that these conditions will hold. First, the level of technological diversity shapes an alliance’s potential payoff — if the partners are too different, they’ll have little to say to one another; if they’re too similar, they’ll have little to learn. Second, an alliance’s organizational form helps determine the likelihood that its potential will be achieved. Drawing on two complementary theoretical perspectives — transaction-cost economics and the “knowledge-based” view — Sampson argues that knowledge transfer is easier within more hierarchical organizational forms, such as equity joint ventures, which benefit from clear lines of authority and common routines. In addition, cooperation may be more likely in these types of alliance. Managers sometimes hesitate to invest in an alliance out of concern that it might be “captured” by the other side. Giving alliances greater organizational independence and more formal governance can help allay these fears.

These advantages are likely to be especially important when technological diversity is high. When knowledge sharing is relatively straightforward, equity joint ventures may not have a significant advantage over looser organizational forms, such as alliances that are based on bilateral contracts. But once the barriers to cooperation start to rise, mechanisms that facilitate and encourage knowledge transfer can really pay off.

To test these hypotheses, Sampson collected data on all 463 R&D alliances formed by firms in the telecommunications equipment industry between 1991 and the end of 1993. She identified the level of technological diversity in each alliance by calculating the degree to which the partners’ patent portfolios overlapped. She also constructed a measure of postalliance innovative performance for each firm by counting the patents the company applied for in the four years following an alliance’s formation and weighting each patent by the number of citations it received in later patent applications.

Within this sample, Sampson found that firm patenting initially increased with the level of technological diversity in an alliance but then decreased. For the median firm, participating in an alliance with moderate diversity contributed nearly 14 times more to postalliance patenting than participating in an alliance with low diversity. Similarly, moderate-diversity alliances outperformed high-diversity alliances by a factor of 4. (This analysis controlled for a variety of factors, including pre-alliance patenting, alliance scope, experience with alliances and prior linkages between the partners.)

Splitting the sample by organizational form produced similar results. Of the 463 alliances Sampson studied, 398 were based on bilateral contracts, and 65 were organized as equity joint ventures. In both subsamples, postalliance patenting initially rose and then fell with technological diversity. However, equity joint ventures had a significant advantage over bilateral contracts once an alliance’s diversity exceeded 0.53, on a scale of 0 to 1. (The values observed in this sample ranged from 0.13 to 1.) At a moderate level of diversity, equity joint ventures delivered 35 times more innovation than contract-based alliances. When diversity was high, the median firm was found to have benefited over 100 times more from a joint venture.

Sampson is wary of basing out-of-sample predictions on these findings — in other words, your mileage may vary — or even of endorsing equity joint ventures outright. After all, all forms of organization have disadvantages (including higher setup costs and more bureaucracy in this particular case). Instead, she says, managers should draw two broad conclusions from her research. First, there is an optimal level of technological diversity within an alliance. And second, organization matters. “It’s not an afterthought,” says Sampson. “It’s not something you throw to the in-house counsel after you’ve closed the deal.” It’s a critical strategic variable that can help determine how much bang an alliance delivers for each R&D buck.

For more information, contact the author at rsampson@rhsmith.umd.edu