Policy implications of work-trip mode choice using econometric modeling

Muhammad Irfan, Ahmed N. Khurshid, Muhammad B. Khurshid, Yasir Ali, Aemal Khattak

Research output: Contribution to journalArticlepeer-review

9 Scopus citations


Transportation agencies seek comprehensive policies and planning to overcome urban congestion and manage transportation demand with due cognizance of traveler behavior. This research develops a travel behavioral model for work-trip mode using revealed and stated choice data collected through a survey questionnaire in the city of Rawalpindi, Pakistan. The multinomial logit model specification is found best suited to develop a modal-split model and estimate travelers' perceived expectation utility functions. The model is used to calculate elasticities and demand response to the policies of improvement in transit/bus rapid transit (BRT) and implement congestion pricing on major arterials of an urban road network. Travel demand is found elastic with respect to congestion pricing and out-of-vehicle travel time. It is concluded that improvement in transit services by introducing BRT alone does not induce a major change in the share proportion of automobile demand; however, congestion pricing has a significant effect on the reduction of automobile demand. Furthermore, the combination of two policies induces more modal-split than does congestion pricing alone. This research highlights traffic congestion pricing as one of the means of traffic demand management by demonstrating its contribution to improving urban traffic congestion.

Original languageEnglish (US)
Article number04018035
JournalJournal of Transportation Engineering Part A: Systems
Issue number8
StatePublished - Aug 1 2018


  • Congestion pricing
  • Demand elasticities
  • Econometric modelling
  • Mode choice
  • Multinomial logit model
  • Stated preference

ASJC Scopus subject areas

  • Civil and Structural Engineering
  • Transportation


Dive into the research topics of 'Policy implications of work-trip mode choice using econometric modeling'. Together they form a unique fingerprint.

Cite this