ELICITING RISK ATTITUDES FROM ROUTE CHOICES OVER SIMULATED RIVING CHOICES

摘要

Transportation planners and engineers alike have identified that drivers’ risk attitudes have asignificant effect on their route and departure time choices. We utilize methods fromexperimental economics to elicit risk attitudes through controlled incentivized experiments indriving simulators with actual monetary consequences. This last feature of our design caneliminate hypothetical biases observed in other studies that utilize purely hypothetical questions.We identify risk attitudes by defining simulated route choices as lotteries that may pay somemoney for sure or different monetary rewards with some uncertainty. We use structuralestimation to estimate risk attitudes in our sample and allow for both Expected Utility theory andRank-Dependent Utility theory. Our econometric approach applies the “contextual utility”correction by Wilcox (2011) to control for “size effects” that have been identified in theestimation of value of time and reliability. Our experimental design allows us to compare riskattitudes across two different regions in the US and study the demographic determinants of riskaversion. We find that, both under Expected Utility and Rank-Dependent Utility, risk attitudesare heterogeneous in our sample and largely explained by subjects’ age and accumulated wealthduring the experiment. Finally, we find no evidence of a structural difference between the twotraffic regions used in this study.