We develop and test the thesis that corporate social performance (CSP) constitutes a socially constructed and shared strategic asset, which is not only influenced by factors specific to a firm, but also by the social performance of firms in its industry and inter-corporate network. Using variance decomposition, we analyze data from 130 large Japanese firms and find that both firm-specific and industrylevel factors account for significant variance in CSP, but network-level factors do not.Keywords Corporate social performance . Japan . Variance decomposition . Keiretsu . Networks . Industry-effects . Resource based view Over the past two decades, corporate social performance (CSP) has been the focus of significant attention from policy makers and the public at large and much theorizing and empirical investigation from the academic community. Wood (1991) defines CSP as the responsiveness a firm has towards its stakeholders. Similarly, Hillman and Keim (2001) consider CSP to combine a firm's attention to the needs of its stakeholders with its efforts to address social issues. Research has shown that firms exhibiting stronger CSP may gain advantages in attracting consumers (Brown & Asia Pacific J Manage (2007) 24:283-303