Making better foreign friendships: the effects of increased cultural diversity in alliance portfolios and portfolio configuration decisions on firm performance

Abstract

Purpose This paper aims to examine the effect of increases in alliance portfolio cultural diversity (IAPCD) on a firm’s performance and how portfolio configuration characteristics moderate this effect, aiming to enable managers to make better partner choice and portfolio configuration decisions to improve performance. Design/methodology/approach The sample includes 2,326 focal firms from 93 countries that formed 7,616 alliances between the years 1992 and 2006. This study uses generalized method of moments estimation to examine the effects of portfolio changes on next year’s firm sales performance. Findings Results reveal an inverted-U relationship between IAPCD and firm performance. Data limitations led to examining moderating effects only on the upslope portion of the inverted-U, indicating that an increasing percentage of joint ventures in a firm’s alliance portfolio strengthens IAPCD’s contribution to performance. Further, increased numbers of marketing alliances or research and development alliances and increased percentage of horizontal alliances in an alliance portfolio have a negative moderating effect. Research limitations/implications The sample mostly covers large companies. The data indicate that nearly all firms are on the upslope of an inverted-U IAPCD–to–performance relationship, allowing testing of moderating effects pre-inflection point only. Practical implications Firms can leverage the additions of culturally diverse partners toward improved performance through astute configuration decisions in alliance portfolio composition. Originality/value This paper uses the knowledge-based view to contribute to the alliance portfolio literature. This study asserts that capacity constraints affect firms’ ability to realize performance gains when taking on culturally diverse partners, an effect moderated by portfolio configurations. This paper tests hypothesis with longitudinal data.