Assessing the Incorporation of Latent Variables in the Estimation of the Value of a Statistical Life 20-22

We explore the incorporation of latent variables (attitudes, beliefs, and perceptions) in the estimation of the value of a statistical life using a hybrid choice model framework. Latent
variables cannot be observed; therefore, one must indirectly infer information about them through questionnaires. We used two latent variables: risk controllability and concerns regarding traffic and cardiorespiratory risks. The use of a hybrid choice model allows us to explicitly model unobserved taste heterogeneity, to improve the behavioral realism of the model, to enhance the model efficiency (due to the incorporation of more information about the latent variables and an increase in the accuracy of predictions), and to help design effective policies related to risk reduction. For traffic risks, we estimate a value of a statistical life of US$ 4.58 million using the conditional logit model and US$ 4.73 million using the hybrid choice model. Our results provide new evidence on the interaction between self-perceived control and risks, stressing the inverse relationship between control and willingness to pay for risk reduction, and the positive relationship between risk concerns and willingness to pay