Flexible estimation of transport demand functions: recommendations for public policy makers

In some situations, economic theory provides directions for constructing an econometric model, while in others (generalized) linear modeling is used only for convenience. The latter is especially true for modeling transport demand functions as economic theory is often limited to predicting whether a covariate is expected to have a positive or a negative impact. However, functional misspecification can have serious consequences for estimation, prediction, and further inferences. The authors introduce a semiparametric model and a corresponding estimation procedure that is easy to interpret and implement. Using an empirical study of university students' transport choices in the Bilbao area, we illustrate how this procedure can help to significantly improve statistical analysis. More specifically, our procedure allows the utility that individuals get from the mode of transport to be modeled and estimated in a flexible way, thus essentially mitigating the danger of misspecification errors.