Organizational Risk-Taking in the Context of Ambiguity

How do firms adjust their exposure to local market risks when those risks suddenly become ambiguous? Behavioral theories of risk-taking (e.g. Audia and Greve 2006, March and Shapira 1992) offer little guidance. They implicitly assume that the interpretations of organizational performance problems that trigger changes in risk- taking are unambiguous. This paper argues that firms that are confronted with ambiguity produced by rare shocks ignore historic performance feedback when adjusting their risk exposure to the affected region. Instead, firms rely on firm-wide insolvency thresholds, local social performance comparisons, and local market signals to guide their risk adjustments. The statistical analysis of an original panel dataset of homeowner insurance underwriting behavior in the U.S. between 1992 and 2012 supports this argument. Critically, however, it also demonstrates that a greater reliance on social learning in such circumstance is contingent on the unavailability of legitimatized means of forecasting future risk exposure. The institutionalization of hurricane forecasting models in insurance markets reduced reliance on such social learning and directly altered risk-taking. The paper contributes to efforts to better understand the scope condition and dynamics of problemistic search and learning processes and their impact on organizational risk-taking and adaptive behavior."