The role of environmental instability in detecting and responding to signals of impending regime shifts

Regime shifts caused by disruptive technologies, legislative changes, pandemics, climate change and terrorist threats affect our lives considerably. In the face of impending change, we are often required to both make accurate predictions about when a regime shift occurs and to change our decisions in response to these shifts. Past research has suggested that people exhibit a systematic bias when it comes to estimating the probability of a regime shift, which is known as the system neglect hypothesis (Massey & Wu, 2005, Management Science, 51, 932–947): decision makers primarily react to signals they observe rather than the underlying environmental system that generated the signal. As such, in unstable environments with precise signals, decision makers have been found to exhibit relatively more underreaction in their probability judgments, whereas in stable environments with noisy signals, they tend to exhibit relatively more overreaction. More recently, Seifert et al. (2023, Management Science, in press) demonstrated that under- and overreaction in decisions when there is an impending regime shift can be accurately described as the result of an interplay between prospect theory and system neglect. Specifically, under- and overreaction in pricing decisions mimics under-and overreaction in probability judgments predicted by system neglect only when the decision maker’s prior probability of a regime shift is low. In contrast, the effect of system neglect is dampened by changes in the decision makers’ risk attitude when prior probabilities are high. In the present chapter we study the extent to which these previously found patterns of under- and overreaction in probability judgments and pricing decisions hold for various levels of environmental instability which go beyond the ones studied in past research. In particular, the authors conduct a laboratory experiment, in which regime shifts occur with various levels of transition probabilities that may vary anywhere between 2% and 98%. The study reveals that underreaction in probability judgments significantly worsens until regime shifts occur with a 50:50 chance, after which the system neglect pattern is mirrored. Furthermore, a series of local regression analyses indicate that under- and overreaction in pricing decisions for low versus high prior probabilities held by the decision makers seems to be generally consistent with previous research as the behavioral pattern persists across various levels of environmental instability.