Rethinking performance based parking pricing: A case study of SFpark

Abstract

In an effort to reduce circling and cruising in cities’ central business districts (CBDs), a number of cities have begun implementing pricing programs that modify parking rates based on observed occupancy levels. We improve on this pricing mechanism by developing a forward-looking policy instrument. The instrument employs a two-stage panel data regression and optimization model that influences demand for parking spaces by changing parking rates via computed price elasticities of parking demand measures. Coefficient estimates that include the elasticity measures from the panel data regression are used to fit a linear prediction model that is the primary input to the optimization model.An application of the approach is presented using SFpark, a federal government-funded demonstration program in San Francisco as a case study. We evaluate the effectiveness of the modified pricing mechanism by comparing actual occupancy and parking rate tuples with the optimized result to ascertain the potential improvement in SFpark’s performance. Policy scenarios are subsequently explored by carrying out sensitivity analysis primarily through SFpark pricing rules. Relative to SFpark’s figures, our model yielded approximately 16% improvement in systems performance when measured by the number of blocks that deviate from the 60 to 80% occupancy target. Our findings highlight the importance of moving towards a predictive regime that allows for proactively managing the parking program compared to a reactive approach based on observed parking occupancy.