Parking Pricing As a TDM Strategy Wei-Shiuen Ng Postdoctoral Scholar Precourt Energy Efficiency Center Stanford University ACT Northern California Transportation Research Symposium April 30, 2015
Parking Pricing Policies Applications Commuter Non-commuter Residential parking Objectives Financial - Revenue for operators Social - Maintain residential quality of life Economic - Support commercial success Environmental - Decrease vehicular emissions by managing travel demand, reducing congestion and travel time
Parking services are often offered at a subsidized fixed rate, which neither reflects the true cost of parking nor actual parking demand.
The High Cost of Parking Construction costs are affected by Size per space Size and shape of site Number of levels Topography Design Geographic location Images by SPUR
Construction Cost of a Parking Space
Projected Parking Structure Costs Source: UCB Parking TDM Master Plan (2011)
Current Studies on Parking Pricing Increasing parking pricing decreases parking demand San Francisco (Kulash, 1974) Portland (Dueker et al., 1998) Toronto (Gillen, 1977) Dublin (Kelly and Clinch, 2009) Sydney (Hensher and King, 2001) Removing parking subsidies decreases solo driving trips Los Angeles (Willson & Shoup, 1990) 15-38% Portland (Bianco, 2000; Hess, 2001) 60% Free parking reduces financial incentives to drive less and increases congestion from increased traffic flow and cruising.
Employee Parking Pricing Effect on Parking Demand
Case Study UC Berkeley Campus Diverse Employment Type Wide range of employment types, income levels and residential locations Varying work schedules Leading to different transportation demand Well-Served by Transit Located in a region with several transportation alternatives For example, AC Transit, BART, Amtrak etc. Physical and Financial Constraints Scarce land resources High parking capital and operation costs Fixed cost annual parking permits
Transportation and Parking Survey Three Main Sections Revealed preference actual behavior (e.g. mode choice, parking location, arrival and departure time) Stated preference behavior under hypothetical scenarios (e.g. mode choice, parking preferences) Socioeconomic and vehicle ownership questions Sample Population UC Berkeley Employees - faculty and staff only Approximately 30% response rate, n = 4,188
Transportation Mode Share Mode Choice from Survey Bike 8% Walk Only 8% Train (e.g. BART) 17% Car, Truck, or Van (Drive Alone Only) 51% Bus (e.g. AC Transit) 8% Motorcycle, Moped, or Scooter 1% Carpool or Vanpool 7%
Parking Preferences On-street, in residential parking zone with residential parking permit 1% Unmetered on-street parking space without time limit enforcement 5% Parking Location Unmetered on-street parking space with time limit enforcement 4% Private off-street parking space 2% Metered on-street parking space 4% Other, please specify 8% Public off-street parking garages or lot 5% Campus parking garage or lot 71% The Other category (eight percent) includes parking at BART stations, the Lawrence Berkeley National Laboratory, parking with disabled person placards or plates either on or off campus, private parking lots under contract with UC Berkeley, and parking on campus Nobel laureate (NL) parking space.
SP Parking Choice Question Example (1)
SP Parking Choice Question Example (2) Given the parking option you have chosen in the above question, how would you now travel to campus? Please select one mode of transportation for each day of the week.
Findings from SP Parking Choice Model (1): Value of Walking Time Value of Walking Time = Marginal Rate of Substitution (MRS) of Walking Time from Parking Location to Primary Workplace Value of Walking Time for Full Sample = 44% of Average Wage Rate
Findings from SP Parking Choice Model (2): Price Elasticity of Parking Demand Parking Option A has the lowest price elasticity Parking Option B has the second lowest elasticity estimate Parking Options C and D have higher elasticities compared to Parking Options A and B Employees are more sensitive to changes in the pricing of flexible parking options
Findings from SP Parking Choice Model (3): Transit and Pricing Incentives Significant Attributes in Choice Set (p = 0.00) Parking fee refund for Parking Option A (0.09) Free transit pass for Parking Options A & B (0.28 & 0.47) BART pass dummy (0.14)
Findings from SP Parking Choice Model (4): Socioeconomic Factors Heterogeneity of Individuals University affiliation - Staff members are more likely to choose monthly parking options than faculty Income - Higher income households prefer monthly and daily parking options, i.e. on-campus parking Age - Older employees are more likely to choose unlimited monthly parking options than hourly parking option
Findings from SP Parking Choice Model (5): Scheduling Factors Work Schedule Factors Arrival Time only significant for monthly parking options (0.31, p = 0.02; 0.27, p = 0.03 ) Departure Time only significant for monthly parking options (-0.38, p = 0.00; -0.34, p = 0.01) Having a second office decreases utilities for all parking options The longer the time spent on campus, the more likely employees will choose to park monthly parking options over daily parking option
Parking Pricing Scenarios
Percentage Changes in Mode Share
Implications for Parking Policies Parking pricing is a powerful TDM strategy Changes in pricing have to be coupled with other incentives Flexible parking permits are the most efficient Free off-campus parking locations serve as alternatives can influence impact of parking pricing Differences in value of walking time provide insights to optimal parking locations Frequency of commute trip and duration of stay on campus affect parking location type
Wei-Shiuen Ng wsn@stanford.edu
Additional Slides
Daily Parking Hangtags Source: Permit Rule Book, Department of Parking and Transportation, UC Berkeley, 2014.
Current UC Berkeley Parking Permits Source: Permit Rule Book, Department of Parking and Transportation, UC Berkeley, 2014.
More Parking Permits Source: Permit Rule Book, Department of Parking and Transportation, UC Berkeley, 2014.
28 The Ultimate Parking Permit UC Berkeley Nobel Laureates Randy Schekman (Physiology or Medicine, 2013) and Saul Perlmutter (Physics, 2011). Sources: gettyimages and Graduate Division, UC Berkeley (2014).
SP Choice Experiment Design Full factorial design = 8 2 *3*2 = 384 profiles
Discrete Choice Analysis: Multinomial Logit Model Utility Function U in = utility of the ith alternative for the nth individual β i = vector of unknown parameters (estimated from data) X in = vector of known variables (include attributes and characteristics) ε n = random utility component Example U PA = utility of Parking Option A α PA = alternative specific constant for Parking Option A β Cost = parameter for the cost of Parking Option A β WKTM = parameter for walking time
Random Utility Model: Notation
Choice Probability Vni = β'xnj, where Xnj is a vector of observed variables relating to alternative j
Estimation Results of Restricted Parking Choice Model
Estimation Results of Full Parking Choice Model (1)
Estimation Results of Full Parking Choice Model (2)
Estimation Results of Full Parking Choice Model (3)