Abstract:

We conduct a field experiment to examine whether offering free pre-peak usage effectively coordinates peak-time consumption. In this experiment, subway commuters on treatment receive a full rebate of their fares if they exit the fare gate before the start of morning peak time. In a leniency period afterwards, the full rebate reduces to a 50-cent discount. This fare structure is motivated by findings in Shampanier, Mazar, and Ariely (2007) that “free” is disproportionally more appealing than any other discounts for consumers. Different from their experiment, commuters in our experiment need to invest efforts (changing travel time) to receive benefits. A difference-in-difference analysis reveals that the full rebate is only marginally more effective than the discount in shifting peak-time consumption to pre-peak, but much less cost-effective. "Free" only appear to be special when it crowds out the attractiveness of the discount: In some treatments, commuters ignore the 50-cent discount if the full rebate is too difficult to achieve.