Abstract
This dissertation investigates the influence of corporate social responsibility (CSR) on firm production by using a neoclassical production function and modern estimation techniques. Although the impact of CSR on firm performance has been studied in the past, the results have been mixed, with some studies showing CSR having a positive influence, others a negative influence, and still others showing it has no influence. This brings to question the best estimation approach to investigate the relationship between CSR and production. This dissertation utilizes a proven theoretical model of firm production to drive the empirical estimation approach. Combining theoretical underpinnings with modern estimation techniques, I argue the results unveiled in my analyses are sound and provide assurance to the true directional relationship. The risk of selection bias is mitigated by using a dynamic panel of diverse publicly traded U.S. firms from 2011 to 2016 to ensure all firms are represented, regardless of industry type or market exit or entry. By using the community social capital of the firm to instrument CSR, the impact of endogeneity overestimation of the influence of CSR on firm production is mitigated. When estimated in a neoclassical production function, with a panel of dynamic data, properly instrumented to reduce estimation bias, this analysis shows the impact of corporate social responsibility is, in fact, overestimated in traditional analyses. These results highlight the inherent risk in studying CSR without addressing known estimation error concerns, and provide a path forward for continued research into how neoclassical theory and modern methods can be utilized to determine how a firm’s socially conscious activities ultimately impact the firm. Given these results, firms are also better positioned to make CSR policy decisions based on both the neutral impact of CSR on output, and the understanding that community social capital can play a role in firm level performance measurements.