A driver feeds one of the "smart meters" that make the SFpark program possible. (Photo courtesy of the San Francisco Municipal Transportation Agency) 

Terrance Heath wouldn’t think of driving to work. His law office in San Francisco overlooks Union Square, a shopping mecca where the daily snarl of honking taxis, open-top sightseeing buses, and double-parked delivery trucks make curbside parking nearly impossible. Parking in the garage under the square would cost more than $30 a day—too much for Heath, a single father of two.

That was before the launch of SFpark—one of the most far-reaching innovations in US parking policy since the invention of the parking meter in 1935. Under the program, San Francisco uses a “smart meter” system to vary the cost of parking in city-owned spots on the basis of location, time of day, and level of demand. City officials, drawing on carefully tracked usage data, adjust prices about every two months. Now, for the first time ever, they can calibrate parking rates to find a sweet spot where prices are neither so high that they discourage people from driving and shopping in the city nor so low that they encourage long stays that limit parking-space turnover.

Drivers, meanwhile, can use the SFpark website or an SFpark smartphone app to access color-coded maps that indicate how much it costs to park on a given street. “You can shop for parking!” says Heath, who notes that he’s now more willing to drive downtown for work or for a visit to the YMCA with his children. At present, the program covers only a select number of high-traffic areas within the city. But in those areas, Heath says, he is generally able to find a spot that he can afford.

The aim of SFpark is to tackle several problems that beset dense, popular cities like San Francisco. When there’s a glut of parked cars in a given area, drivers tend to circle endlessly until they find a free space. That behavior leads to traffic congestion, excessive fuel use, and air pollution. Congestion causes buses and streetcars to fall behind schedule, and it causes merchants to lose business and governments to lose tax revenue. “Our goal was to have one free parking space per block,” says Jay Primus, program manager of SFpark. “The way we used to park was senseless. It was a waste of time and money, and it contributed to greenhouse gases.”

SFpark launched in 2011, and since then it has generated considerable interest in city planning circles. A report by the International Parking Institute named SFpark the most innovative parking program in the United States. The program also caught the eye of scholars at Harvard University’s John F. Kennedy School of Government, which last year selected SFpark as one of the nation’s top 25 innovations in government. “SFpark is a breakthrough for parking, but even more for government efficiency,” says Stephen Goldsmith, director of the Innovations in American Government Program at the school.

Donald Shoup, Distinguished Professor of Urban Planning at UCLA and author of the book The High Cost of Free Parking, echoes that point. “Transparent, data-driven pricing can bypass the usual politics of parking,” he says. “In most cities, it’s the city council that has to debate and make every change in parking prices, and that makes it a very political process that leads to inertia. Cities all over the world are watching what San Francisco is doing to get the prices for on-street parking right.”

Metering Demand

The crux of the SFpark program involves moving away from the one-rate-fits-all model that has long characterized metered parking. Under the new variable pricing scheme, drivers have a choice that the traditional system couldn’t offer: Pay more for the convenience of parking in high-demand areas at peak times, or park in less popular areas at off-peak times and pay less. Today, the hourly rate to use a city-owned spot can be as low as 25 cents or as high as $7. (That’s how much a driver might pay to park within walking distance of AT&T Park during a Giants baseball game.) Time limits, which used to last just one or two hours, now last up to four hours.

Since 2011, SFpark has initiated 12 rate changes. Over that time, the average hourly rate at a curbside meter in San Francisco has gone down by 32 cents (it’s now $2.41). For 20 percent of the curbside meters covered by SFpark, the cost is now $1 per hour or less. At that rate, according to Primus, San Francisco now offers some of the least expensive metered parking to be found in any major US city. The cost of parking in public garages has dropped in a similar fashion. “We feared the rate changes would be a big issue, but we haven’t had one complaint [on that matter],” Primus says. That pattern matches the findings of behavioral research, he notes: “People don’t mind paying for parking, because what they truly value is their time.”

The origins of SFpark go back to 2007, when San Francisco won a $19.4 million grant from the US Department of Transportation to launch a pilot program. City officials used a chunk of that money to meet a core challenge: To make variable pricing work, they needed to know how demand for parking actually varies—street by street, hour by hour. And so, in 2009, San Francisco conducted the nation’s first-ever parking census. Planners fanned out across the city to compile an inventory of every public curbside and garage parking space within its borders. They counted 440,246 spaces. According to their tally, 28,800 of those spaces are curbside spots with adjoining meters. SFpark covers 7,000 of these metered spaces, along with 13,000 spots in parking garages owned by the city.

In the next phase of SFpark, the city converted 13 public garages to a system that gathers real-time parking data from mechanical gate arms. In addition, the federal grant enabled the city to bury hockey-puck-sized sensors in the asphalt under thousands of curbside spaces. The sensors beamed updates wirelessly to a data center whenever a spot became empty or occupied. With those data in hand, SFpark officials were able to bring a high degree of precision to their system for adjusting meter costs.

During the first three years of the program, SFpark shared real-time data from the sensors with drivers through the SFpark smartphone app and through third-party apps as well. Using that information, drivers could steer toward open spots and thereby reduce the time that they spent circling for a space.

In January of this year, SFpark officials shut down the sensors, along with the data feed that the sensors made possible. (The program continues to provide real-time data on parking availability in certain public garages.) Batteries in the sensors were starting to die, and federal support for that capability was winding down, too. Officials knew that this element of SFpark would be temporary, Primus explains: “It’s a very new technology. The point was to learn about it.” Meanwhile, the sensors had served an essential purpose, which was to give planners a firm base of knowledge about parking usage. “Getting the prices right for parking is the primary strategy,” Primus says. “Real-time data is an enhancement.” For now, drivers will have to make do without that enhancement.

Making a Point

SFpark has attracted its share of controversy. Although meter fees have gone down in some places, they have risen in others. Some residents have complained that SFpark is “classist”—that it allows more-affluent drivers to take the most convenient spaces while forcing others to take more remote spots. “There is a huge disconnect between the theoretical goals of the city and how this directly affects the lives of residents,” says John Lum, an architect who helps lead the Eastern Neighborhoods United Front, a group that opposes SFpark. “The wealthy don’t ever have to worry about parking costs. It’s the blue-collar people who commute into the city to their carpentry and auto mechanic jobs who suffer.”

Opponents even created an anti-SFpark website (www.sfpark.info) whose name purposely resembles the name of the program’s official site (www.sfpark.org). There, people have posted various complaints. One common charge is that SFpark is just a money grab by the city.

City planners counter that SFpark isn’t meant to generate income. In fact, data from SFpark show that the program has been essentially revenue-neutral. The year before SFpark went into effect, parking meters generated $39 million for the city, and meter-related tickets brought in $125 million. In fiscal year 2012, parking meter revenue increased by $4.4 million, but parking meter citation revenue dropped by about $5 million. “The number of meter-related tickets has gone down,” Primus says. “But parking meter revenue is way up. As soon as we made it easy for people to pay, people purchased more time at the meter.”

The pilot phase of SFpark ends in June, when its federal funding runs out. But the smart meters will stay in place, and SFpark officials plan to retain the variable pricing system. Although they no longer have access to real-time information about parking in curbside spots, they’re confident that they can base future rate changes on data from meter and garage payments. Today, their main goal is to install smart meters throughout the city by the end of 2014. Doing so will allow planners to institute demand-based pricing in more and more San Francisco neighborhoods.

How ambitious SFpark officials can be on that front is an open question. But they hope that an impact study now under way will help make the case for extending the program. City workers are quizzing drivers about parking availability. They’re monitoring public transit times. They’re checking sales tax revenue to see if economic vitality has increased in SFpark zones. And they’re calculating the effect that reduced circling times have had on exhaust emissions. (The impact study is slated for publication in May.)

The results of that evaluation process are likely to reach beyond the streets of San Francisco. “This approach to managing parking has been broadly accepted as ‘best practice’ but not widely implemented,” Primus notes. “So it’s a very useful reference point in a high-profile city.”

Tracker Pixel for Entry