Discrete choice models applied to travel demand analysis: focus on risk and heterogeneity

Discrete choice methods model a decision-maker’s choice among a set of mutually exclusive and collectively exhaustive alternatives. They are used in a variety of disciplines as transportation, economics, psychology, and public policy in order to inform policy and marketing decisions and to better understand and test hypotheses of behavior. This dissertation is concerned with the enhancement of applications based on discrete choice methods, focusing on heterogeneity of decisions on transport markets and the evaluation of risk attitude of agents. The study of travel-related behavior has become one of the most flourishing research fields of the applied microeconomics since 1970. The multitude of articles published on this topic confirm the important role that the mobility of people and merchandises acquired within the panorama of modern economics – see, for instance, the richness of the literature concerning road regulation and the choice of route, the choice of travel mode and network models. Higher mobility aggravated the congestion and induced the need to understand how travel demand adjusts to economic policies such as the congestion pricing, the environmental taxation, the competition between route operators and among different travel modes. From a methodological viewpoint, this issue has raised new questions about how to formalize, evaluate and better interpret phenomena as: • network effects derived from congestion pricing and its properties under different welfare scenarios and on different transport markets (see, for example, the works of De Borger et. al. 2005, de Palma 2004, de Palma et. al. 2005, and 2008, de Palma and Lindsey 2000, Verhoef 1996); • questioning upon the deterministic or stochastic behavior of agents in the context of different transport networks and in relation to price dispersion (de Palma and Lindsey 2004, Borenstein 1989, Borenstein and Rose 1994); • questioning upon the individual or collective character of travel-related decision making processes (Chiappori, de Palma and Picard 2010, de Palma, Picard and Ziegelmeyer, 2005); • the analysis of attitudes towards risky outcomes (de Palma and Picard 2004, 2005, 2006, 2010) and under different information settings (symmetric versus asymmetric information). • the simultaneous effects of policies addressed to different markets coexisting in a common urban framework (De Borger and Proost 2001). This study aims to integrate two fundamental approaches to the study of regional economics and the transport policy: i) first, a theoretical approach based on analytic models supported by numerical simulation and ii) an empirical approach to test theoretical assumptions. The micro-economic perspective is a challenging and complex issue at the same time. Some of the open issues at the center of debate are: the taxation of roads, perfect and imperfect competition among the networks supporting private trips, congestion pricing and the attitude towards risky outcomes as uncertain travel times, the decision making process and resource sharing among different members of a household. The micro approach has been integrated with two empirical experiments in which we propose new approaches to study two traditional problems: i) modal choice, by introducing the joint decision making process applied to a sample of active couples from Ile-de-France, and ii) route choice in the Middle East - a context in which the travel time, as an input of the decision process, is characterized by uncertainty and depends on the political scenario at the macro level. The thesis is structured in five chapters. The first chapter presents the motivation, providing an overview of the related literature on discrete choice models. Particular emphasis is put on the studies related to the analysis of travel demand heterogeneity and the formalization of risky outcomes in the context of politically uncertain scenarios. In the second chapter, we study the modal choice problem to support the idea that the choice between private car and public transit represents the output of a joint decision-making process rather than an individual process. This hypothesis is tested on a sample of active couples from the French Census 2001, that use either the public transit or the private car for commuting trips in Ile-de-France. The results are significant and support the idea that mode choice is correlated for the male and the female of the same household, especially in the presence of children less than 6 years old. The third chapter studies the issue of congestion pricing in the context of airline networks. In particular, in this chapter we present the results obtained from the static comparative analysis of Nash tolls at the symmetric equilibrium in the presence of stochastic travel demand and congestion effects, in two types of network settings: with routes in parallel and with routes in series. Then the discussion is developed around the question of the agents' loyalty programs and switching costs charged to those agents that change operators at a given node in the network. The results show that optimal prices positively adjust to account for the heterogeneity of airline services. The model assents to the literature on air transport competition and loyalty programs in the airline industry. The fourth chapter presents a complete micro-macro analysis of the freight travel demand in a context of military occupation. Macroeconomic costs are derived from travel time uncertainty at the micro level based on data collected in the Middle East. A route choice model with risk averse and risk-neutral users is formalized to represent freight flows in “good” days and “bad” days depending on the political turmoil. In the conclusions, we provide interesting lines to open the way for the analysis of new arguments. The future research will rely on the underpinnings of this study and will focus on the formalization of agents' behavior on travel markets so as to assimilate uncertainty and information asymmetry effects. The analysis of economic policies' acceptability and the evaluation of sustainable development policies emerge as two of the most important issues to be discussed in the future context of an integrated urban and regional economy.