Parking is a serious issue in many urban areas, especially those experiencing rapid population growth.  To address this problem, some cities have implemented demand-responsive pricing programs, where parking prices vary depending on the occupancy rate in a previous period. Yet, few empirical studies have rigorously evaluated these programs, though they have the potential to improve turnover and much needed parking availability in growing cities with limited infrastructure.  In this study, we investigate the impacts of SFpark, a demand-responsive pricing parking program in San Francisco that began in 2011.  We focus on metered, on-street parking and exploit the timing of SFpark as a natural experiment.  We observe effects on three important aspects of urban transportation: parking availability, transit bus ridership and congestion.  The timing of this program is plausibly exogenous to factors that affect these outcomes of interest since it is based on bureaucratic decision-making, so endogeneity is less of a concern.  We generate a novel panel data set by merging detailed parking occupancy and meter rate data with micro-level transit bus ridership information at the bus stop-bus shift level.  Results show that SFpark led to more areas meeting the target occupancy range of 60-80%.  We also find heterogeneous effects on transit ridership and show an increase in meter rates is associated with a modest increase in ridership, suggesting people are substituting between transit and non-transit travel and that meter rates factor into mode choice. Finally, we find SFpark reduced congestion, specifically decreasing lane occupancy and increasing vehicle speed.  These results have important implications for transportation policy as cities continue to expand and implement demand-responsive pricing programs globally.