In this dissertation, we develop two experimental methods for the problem of pricing or incentivizing use of a transportation service and apply them to the pricing of employee parking at the University of California, Berkeley.
The University of California, Berkeley, with 23,962 employees is the largest employer in the eastern half of the San Francisco Bay Area and has a problem with employee parking. The university wants to explore a daily parking cash-out program, named the FlexPass, to make employees more mindful of their parking consumption. We use a Randomized Controlled Trial(RCT) to reveal the causal power of the cash-out. The RCT is applied to 392 employees, representing 10% of the university employees driving alone and parking, over three months using an IT system able to collect daily parking consumption, weekly commute mode reports and location data. The FlexPass treatment reduced consumption by 6.1% with high significance.
Our second experiment is focused on measuring an incentive response curve. We use a repeated 2nd price reverse auction, in which 215 parking permit holders participate for 61 working days. Our method measures the incentive response curve for our subjects and we estimate the curve for the employee population using a quantile regression. We find the known and heavy overhead of repeated bidding can be removed by a lightweight IT system compressed of apps on iPhone and Android and a server in the cloud.
Finally, we build a two-stage signaling game and design a variable-rate daily incentive scheme, where the incentive changes based on weekday and weather. The variable-rate daily incentive outperforms the fixed-rate daily incentive on both parking cruising times and leftover parking spaces.