Abstract This paper aims to assess five urban transportation pricing policies (including petrol price, parking fee, taxi fare, bus fare, and metro fare). Using stated preference panel data, we estimate a mixed logit (ML) model and a latent class (LC) model, which account for individual unobserved heterogeneity. The estimated policy implications resulting from these models are compared to get a deeper understanding of the differences in policy recommendations when using different treatments of unobserved heterogeneity. In this regard, we focus on two criteria, which are the changes in mode share and the changes in expected consumer surplus in response to the changes in the level of pricing policies. Results by both ML and LC models reveal that in some comparisons, the ML and LC models yield similar results, as both indicate the policy of increasing petrol price is more effective than increasing parking fee on reducing car congestion. However, we can also find some differences in the behavioral outcomes of these models. Overall, the ML model estimated substantially greater choice elasticities and changes in mode share, in comparison to the LC model. Some of these differences may lead to different policy recommendations as the ML model suggests avoiding the policy of increasing bus fare due to its large negative impact on car share, while the LC model predicts a relatively small reduction in car share after increasing bus fare and therefore, can motivate policymakers toward this policy. This paper also provides new evidence on one of the key assumptions in discrete choice analysis, namely the assumption of an equal (generic) cost coefficient over transport modes is not satisfied when the travelers do not have a true perception about their travel costs (e.g. fuel cost and parking cost).
Highlights Assess five urban transportation pricing policies in order to find their implications in managing traffic congestion in Tehran. Using mixed logit model (ML) and a latent class model (LC) to account for individual unobserved heterogeneity. Increasing petrol price is considerably more effective than parking fee on decreasing car share. ML and LC models estimated relatively similar values of travel time saving and expected consumer surplus. The estimated cost parameter for the car is significantly smaller than the parameter of other alternatives.
Assessment of urban transportation pricing policies with incorporation of unobserved heterogeneity
Transport Policy ; 99 ; 12-19
2020-08-11
8 pages
Article (Journal)
Electronic Resource
English
Incorporating Unobserved Heterogeneity in Pavement Deterioration Modeling
British Library Conference Proceedings | 2005
|Incorporating Observed and Unobserved Heterogeneity in Urban Work Travel Mode Choice Modeling
Online Contents | 2000
|Incorporating Observed and Unobserved Heterogeneity in Urban Work Travel Mode Choice Modeling
British Library Online Contents | 2000
|Cities and Satellites: Spatial Effects and Unobserved Heterogeneity in Modeling Urban Growth
Online Contents | 2008
|