We highlight several additional analyses that strengthen our primary results. First, we confirm that the link between ICWs and relationship duration is robust to propensity score matching on ICW disclosure, mitigating endogeneity concerns arising from potential correlated omitted variables. Our remediation analysis is also robust to propensity score matching. Second, we find that the influence of ICWs on relationship termination varies by a customer’s dependency on the supplier. Specifically, the association between ICWs and relationship termination is increasing in customer dependency until a critical point is reached where the
5
customer is highly dependent on the supplier. This non-linear effect of customer dependency suggests customers are more likely to terminate relationships as the importance of a supplier increases; however, customers perceive the costs of terminating a relationship with their most important suppliers as too costly even if that supplier discloses an ICW. Third, we analyze whether customer reaction to a supplier ICW varies with its severity. Using the number of individual SOX 404 and 302 material weaknesses to proxy for severity, we find evidence that customers are more likely to terminate relationships when a supplier discloses multiple ICWs.