Channel Integration: An Explanation According to David Teece’s Theory of the Boundaries of the Firm

Decisions about the degree of channel integration are a critical component of any manufacturer’s marketing channel strategy. At one extreme, the manufacturer can integrate forward and perform the distribution function itself, by establishing a sales subsidiary. At the other extreme, the manufacturer can choose not to perform this function, using, instead, independent distributors (Klein, Frazier, and Roth 1990; Aulakh and Kotabe 1997). Over the last 30 years, transaction cost theory (TCT), proposed by Oliver Williamson (1975, 1985, 1986, 1999), has emerged as one of the dominant theoretical lenses through which to explain channel integration decisions (cf. John and Reve 2010; Rindfleisch et al. 2010). Empirical studies have provided considerable support for TCT’s hypothesized effects on asset specificity; that is, the specificity of asset distribution has a positive impact on the degree of channel integration (e.g., Anderson and Schmittlein 1984; Anderson and Coughlan 1987; John and Weitz 1988; Klein, Frazier, and Roth 1990). Although the positive effects of asset specificity on channel integration are well documented in empirical studies, several unanswered questions remain that deserve further examination. Among them, the one pointed out by resource-based view (RBV) and capabilities theory (CT) scholars is as follows. Traditional TCT assumes, implicitly, that governance decisions are determined, largely, by the characteristics of a transaction (e.g., asset specificity). This assumption undervalues the role of the firm’s and the market’s idiosyncratic resources and capabilities of production (e.g., appropriability regime, interdependence of complementary activities, and thinness of the market). In addition to efficiency concerns about economizing transaction costs, governance decisions may also be influenced by strategic concerns about leveraging or protecting valuable resources and capabilities (Rindfleisch et al. 2010). Over the last 15 years, a growing number of scholars have examined governance decisions using TCT and RBV/CT. This study focuses on David Teece’s (1986, 2006, 2007a, b, 2010a, b) theory of the boundaries of the firm, and explains channel integration decisions using Teece’s theory. Teece’s theory argues that a firm not only economizes transaction costs but also develops and deploys capabilities for the creation and management of productive know-how. It builds a capabilities dimension into the Williamson theory (Teece 2010a). The characteristic argument is that interaction effects between the appropriability regime and asset specificity and between interdependence of complementary activities and thinness of the market impact the boundaries of the firm (Teece 1986, 2006; Chesbrough and Teece 1996). Based on the argument, we propose two hypotheses in this study to explain channel integration: H₁: The interaction between the weakness of the appropriability regime and the specificity of distribution assets enhances the degree of channel integration.H₂: The interaction between the interdependence of production and distribution activities and the thinness of the distribution market enhances the degree of channel integration. In empirically testing such interaction effects, in this study, two-stage least squares (2SLS) regression employing the SYSLIN procedure in SAS was used (cf. Bollen and Paxton 1998; Im, Hussain, and Sengupta 2008). Data for the study were collected through a survey of Japanese manufacturers. The sample included manufacturers of general machinery, electrical appliances, and transport machinery. In total, we received 105 usable surveys. The results from the 2SLS regression analysis supported the study’s two hypotheses. In the regression model, the adjusted R² statistic equaled 0.22, and it was statistically significant at the 0.01 level of overall significance. The coefficient of weakness of the appropriability regime × the specificity of distribution assets was positive and statistically significant at p < 0.01, indicating a positive interaction effect on channel integration. Also, the coefficient of interdependence of production and distribution activities × the thinness of the distribution market was positive and statistically significant at p < 0.01, also indicating a positive interaction effect on channel integration.