Operationalizing Corporate Social Performance through Exploration and Structuring of Proxy Data

Abstract

Strategy scholars have long been interested in how firms employ corporate social practices (CSP), but the CSP construct still lacks consensus on operationalization. Our research seeks to understand how CSP practices can better be framed and operationalized using the popular KLD data as proxy. Scholars will often aggregate the 70-plus categories of KLD data into a single index score without weighting or imposing a latent structure, leading to a potential misapplication of the data. Our paper reframes how the KLD data may be used for CSP studies by exploring the KLD data, using partial least squares path modeling techniques to evaluate different structures. Using a sample of 938 firm-years, we conducted a series of analyses to determine an optimal structure and predictive weighting for the CSP relationship with financial performance. This approach yielded a six-fold increase in explanatory power over conventional methods, and the signs and magnitudes of the weights in the construct offer a nuanced view of how CSP and financial performance are related.