Honeymoons and liabilities: The relationship between age and performance in research and development alliances

The use of strategic alliances by technology ventures has increased dramatically over the last 20 years. During this period companies not only have increased the use of alliances but also have used them in more strategically important areas, particularly in research and development (R&D) and new product development. Thus, successful management of strategic alliances in high-technology industries has become critical to a firm's new product development and ultimately to firm performance. Yet little is known about what determines the performance of individual alliances. This article examines the relationship between the age of an alliance and the performance of the alliance. Two competing hypotheses regarding the form of the functional relationship between alliance age and alliance performance are developed and are tested. First, a liability of newness hypothesis, which posits that alliance performance increases in a linear fashion over time, is tested. Then a honeymoon hypothesis, which posits that the relationship between age and alliance performance is nonlinear with alliance performance decreasing initially but increasing over time, is tested. It is proposed further here that alliances that are more important to the focal firm exhibit longer honeymoon periods. A measure of individual alliance performance is developed based on our field study in the biotechnology industry. The competing hypotheses are tested using regression analysis on the sample of 115 R&D alliances. Then the analysis is extended by splitting the sample into high- and low-importance alliances to enhance the robustness of the findings. Further, such a split-sample approach enables testing for a potential moderating effect of alliance importance on the hypothesized relationship between alliance age and alliance performance. The results suggest that the relationship between age and alliance performance seems to be U-shaped curvilinear rather than linear, with the minimum point Of alliance performance occurring after approximately four and one-half years. Thus, the results indicate that strategic alliances appear to face a liability of adolescence rather than a liability of newness. Contrary to expectations, it also is found that important alliances exhibit generally shorter honeymoons.