Travels in Crowded Subways: Utility, Constitutive Factors and Policy Implications

Travels in Crowded Subways: Utility, Constitutive Factors and Policy Implications


Koning Martin, East Paris University - IFSTTAR - SPLOTT, Monchambert Guillaume, ENS Cachan, Haywood Luke, Deutsches Institut Für Wirtschaftsforschung - Berlin


Based on a survey collected in Paris subways, this paper focuses on attitudes towards crowding. We look at the factors explaining travelers' satisfaction, the channels by which crowding affects them and their ability to adjust schedules.


Over the last decade, a growing body of research has focused on the qualitative attributes offered by the public transportation (PT, see Litman, 2008 for a review). Whether the objective is to attract car drivers or to avoid passangers from switching away from PT, operators generally implement policies that improve the “travel experience” of users.

Among the numerous relevant qualitative attributes of PT, in-vehicle crowding is often singled out as one of the most desirable dimension (Eboli and Mazulla, 2008, Baker et al., 2007). This has led psychologists and economists to investigate the effect of an excessive usage of PT on the “time perception” (Li, 2003, Cox et al., 2006, Mahudin, 2012) and on the “generalized cost of travels” (Li and Hensher, 2011, Wardman and Whelan, 2011, Tirachini, 2012) respectively, with important implications for health, safety or stress, as well as route/mode choices and PT pricing.

Within such a framework, the objectives of this article are threefold:
- First, we want to understand the influence of the in-vehicle crowding on PT travelers’ stated satisfaction. In particular, we explore the respective weights of the objective vs. subjective crowding to explain the variation of individuals’ utility. Also, we try to depict the way this dimension of travels interacts with the in-vehicle time.
- Second, we characterize the channels by which the in-vehicle crowding deteriorates the travel experience. In fact, excessive PT patronage may reduce travelers’ utility through: the lack of seating, potential falls, bad smells, non-polychromic use of the travel time, risk of pick-pockets… We use stated nuisance scores to order these dimensions of crowding in terms of dissatisfaction. Because different people are sensitive to different features of PT crowding, we also propose a sort of “market segmentation”.
- Finally, we explore the potential for a policy which would allow the PT users to adjust their schedules. Switching from highly crowded to more comfortable time periods would improve utilities of individuals who can change their travel times, but also those of people obliged to arrive/depart at fixed hours. Here again, individuals’ characteristics play a major role to make such a measure feasible and we try to identify the PT users, as their respective weight into the total PT patronage, who could adapt their schedules.

This work is made possible thanks to an original field survey collected late 2010 on 1,000 users of the Paris subway network. Whereas the survey was initially designed to value crowded travel times (Haywood and Koning, 2013), it offers additional information about travel satisfaction, crowding perception, the channels by which discomfort affects the travel experience and the ability of PT users to adjust their schedules. Importantly, we do not have the same number of observations for each sub-question because the main survey was designed deliberately to be answered quickly to ensure representativity. People willing to let a train pass in order to answer the full questionnaire (which includes the qualitative questions) may have different preferences for crowding and/or abilities to adapt their travel times. As a consequence, we pay a great attention to the selection mechanisms in order to make our conclusions robust for the entire population of Paris subways users.


Association for European Transport