Emotion-Driven Risk Preferences

We hypothesize that anxiety is an important determinant of investment decisions. By explicitly accounting for the role played by anxiety, we seek to shed light on systematic differences between performance in static and dynamic investment tasks, to explain the effect of incentives or stake size on performance, and to understand experimental subjects' propensity to self-select into risky dynamic investment tasks. In our experiment, we intend to go beyond documenting behavior that is consistent with the presence of anxiety: we seek to link departures from the predictions of standard models to the presence of emotions by measuring subjects' physiological arousal (through skin conductance and heart beats) during the experimental task. This promises to yield high quality evidence on the causal link between anxiety and outcomes in investment tasks.

In our experiment, subjects complete a static and a dynamic investment task, with either high or low stakes. In the dynamic investment task, subjects attend to a signal that lasts a random period of time. They know that the signal stopping time is uniformly distributed between a lower and an upper bound. Subjects press a button to stop the game with the aim of doing so before the signal stops. If the signal has stopped before they press the button, they receive nothing. If the signal is still continuing, the payoff they receive is increasing in the length of time they have waited. The only difference between the high stakes and the low stakes treatments is the return per second waited. In the static investment task, subjects are asked to pre-commit to a stopping time of their choice that the computer program.

The experiment has been implemented with around 70 subjects. The cleaning of physiological data is currently taking place and should be ready for analyses at the end of August 2015.