Where the Parking Is Easy

Jay Primus, manager of SFPark, stands next to a variable-price, pay-by-phone parking meter.

Jay Primus, manager of SFPark, stands next to a variable-price, pay-by-phone parking meter.

San Francisco’s smart parking experiment — setting prices for on-street parking based on supply and demand — has brought lower average charges, made it easier to find a space and reduced parking-related anxiety.

By James A. Bacon

San Francisco has a reputation in many parts of the country as a bastion of left-wing politics. Haight-Ashbury and hippies. Harvie Milk and gay pride. The People’s Republic of Berkeley. (OK, Berkeley is across the Bay but it’s nearby). Between the municipal labor unions, high-speed trains and high taxes, if the city were a country, some might say, it would be more socialist than Sweden.

But there’s at least one government function in which the City of San Francisco has led the way in applying market-based principles — on-street parking. The city was one of the first in the world to scrap the time-honored practice of charging drivers a flat rate for a parking space regardless of time or location. San Francisco has put into place a state-of-the-art system that varies the charge for a parking space in response to supply and demand.

Imagine that. While cities and states that supposedly believe in the efficacy of markets are stuck in the socialist mindset of charging everyone the same rate for on-street parking, California’s citadel of liberalism installed a system of sensors, pay-by-phone billing and advanced algorithms to allocate roughly 25% of the city’s on-street parking using the price mechanism. Two years into the experiment, San Francisco officials and outside scholars are pronouncing the system a success. The program still needs tweaking, but the San Francisco Municipal Transportation Authority is considering how to expand the system throughout the city.

“We have a much better sense of our parking supply, and of demand for that finite supply. We can make more informed decisions,” says Jay Primus, manager of SFPark. While charges in high-demand blocks went up, the price of many parking spaces declined. Overall, San Franciscans pay less for on-street parking than they did before, he says, and they enjoy greater convenience because open spaces are almost always available.

Confirming the positive outcomes in an in-depth March 2014 study, “Is the Curb 80% or 20% Empty?”, three academic economists concluded that SFpark had cut the number of people clogging streets while cruising around looking for empty spaces. “[We] conclude that rate changes have helped achieve the City’s [60% to 80%] occupancy goal and reduced cruising by 50%.”

For seven decades San Francisco managed its on-street parking like every other American city – mechanical meters charged flat rates for short time limits to meet vague, oft-conflicting political goals. The result, as it was most places across North America, was confusion. Some spaces were always full, prompting motorists to drive around in circles looking for an open spot. Sometimes, on-street parking was cheaper than in nearby garages, a discrepancy that flushed drivers out of the garages and into street parking. Other spaces remained empty, costing the city revenue.

Economists have long argued that such problems could be solved by varying the pricing for on-street parking based on local supply and demand. UCLA urban planning professor Donald Shoup, arguably the Albert Einstein of parking, has long contended that cities should set prices at the block level to keep parking spaces occupied 60% to 80% of the time. If parking utilization falls beneath that range, city officials should cut rates to make the spaces more attractive and fill up more of them. If spaces are full more than 80% of the time, cities should raise rates and create more vacancies. But the tools to administer such a scheme cost-effectively did not exist and the theory remained just that, a theory.

In recent years, however, the advent of new technologies – sensors, wireless communications, cheap storage for Big Data and advanced algorithms, all part of what is known as the Internet of Things – has driven down costs associated with on-street parking. Inexpensive sensors can tell when parking spaces are empty or full and can transmit that data via wireless to the data warehouse. Drivers can download apps to find open parking spaces and pay with their smart phones. It is a simple matter to notify drivers when their time is about to expire and let them renew by phone. That option cuts down on the anxiety of people worrying about their time running out and getting a ticket.

In 2007 San Francisco was selected as one of five cities across the country to participate in a federal program to test the idea of demand-responsive pricing. The transportation authority won a $19.8 million federal grant, to be supplemented by nearly $5 million in local funds, to set up the program. The purpose wasn’t to maximize parking revenue but to reduce congestion and optimize convenience, thus encouraging people to live and do more business in the city.

There is no handbook for introducing variable-priced parking. San Francisco had to feel its way. One important decision was to convert just one-quarter of the city’s parking spaces in the hope of learning from the experience and fine-tuning the system before rolling it out city-wide. Another was to let rates vary – but to install dampers to preventing them from fluctuating wildly and creating unpredictability for drivers. Rates stayed in effect for four to six weeks, and they were not adjusted by more than $.25 per hour at a time.

Critical to the program’s success, says Primus, was a decision to make outreach a priority. SFPark met with every neighborhood group that expressed an interest in the program. “Everyone agreed parking was a problem. We came to them with a solution,” he says. The end result: “We had zero complaints – and that’s in a vocal town where people are not shy about letting their feelings be known.”

San Francisco is not the only city to experiment with variable-price parking – Los Angeles is another U.S. pioneer, while Moscow and Nice, France, have launched ambitious programs in Europe. Dozens of companies in the emerging “smart cities” industry are peddling sensors, meters, billing systems, algorithms and integration services catering to the surge in interest, and cities are announcing new initiatives with some regularity. But San Francisco was one of the first, it set up the program to collect and analyze data scientifically and it has shared its findings widely.

“The big, big winner is market pricing,” says Bern Grush, founder of Pay By Sky, a Toronto developer of a meterless parking payment system who has tracked the San Francisco experiment closely. “For downtown commerce, entertainment and meals, you don’t want people circling in cars. … We can actually minimize circling. It’s possible to set the prices properly.”

Another winner is the idea of parking guidance – signage, parking apps and other techniques to guide people to open parking spaces. Says Grush: “That works.”

In Grush’s estimation, yet another important finding is that creating payment options increases compliance. People are happy to pay by credit card or pay by phone instead of fishing around their pockets for coins. When people can pay more easily, they are less likely to park without paying and risk getting tickets. While San Francisco may have suffered a loss in ticketing revenue, it also cut down in enforcement costs.

The bigger picture story is that people feel better about doing business in San Francisco. Primus, the SFPark manager, equates the parking initiative to an economic development strategy. “No one likes to shop where they get a $60 ticket.”

Primus says the system also provides more parking data than the city ever had before. SFPark monitors the sensors to make sure they’re working properly. The contractor that installed them is penalized if the sensors fail to meet performance standards. “That’s the way government should operate – Government 101,’ says Primus. The data also is helpful when working with developers on parking requirements — “You can instantly make a map of an area. You can save the city a lot of administrative costs” – and when planning transportation improvements such as bicycle or bus-only lanes. Measures that affect parking can get emotional, he says. “We can inject more hard data into the decision-making process.”

One disappointment has been the sensors embedded in the asphalt that detected the presence of cars in parking spaces. The batteries were designed for a five-year life span but started dying within 18 months – and the only way to replace them was to dig them out of the street. As a result, the sensor data was spottier than expected. On the other hand, says Grush, the payment information from the parking meters filled in the missing occupancy numbers. In the end, he suggests, the sensors proved to be a redundant and unnecessary expense. (Grush’s company develops an automobile-based payment system that can provide occupancy data as well.)

SFPark expects to complete in June its analysis of the parking system and how to extend it city-wide, says Primus. He sounds confident that the project will move forward. For most people the biggest parking hassles are the uncertainty of finding a space and the difficulty paying. They don’t mind paying a supply and demand-based price for parking. SFPark has a winning formula, he says. “Make it easy to pay and reduce anxiety.”

(Cross posted from Datamorphosis)