Data Show Bay Bridge Crossing Speeds Not Affected by Variable Tolls

Graphic: Eric Fischer
Weekday crossings. Graphic: Eric Fischer

When the tolls on Bay Area bridges were increased on July 1, the Bay Bridge was given a higher toll at the times of its greatest usage in an attempt to reduce congestion by discouraging drivers from using the bridge at peak times.  Crossing the bridge into San Francisco costs $6 from 5:00 to 10:00 am and 3:00 to 7:00 pm on weekdays, $4 at other times on weekdays, and $5 on weekends.  However, the toll structure does not seem to have had the desired effect.

Using data for about 50,000 bridge crossings from Stamen Design’s Cabspotting and from NextBus for AC Transit, I calculated and plotted the time required to cross the bridge at different times of day before and after the toll increase.

The graphs are noisy because of the small sample size, but the time required to cross the bridge by car at the morning and afternoon weekday peak times seems basically unchanged since 2008.  The off-peak weekday crossing is a little slower than it used to be, perhaps because of the S-curve detour for construction of the new east span.  On weekends, the off-peak crossing time seems to be unchanged since 2008.

The weekend peak data from 2008 is especially noisy so it is hard to tell exactly what has changed there, but if anything, it takes longer to cross the bridge now than it did before.  It seems that either the demand for the bridge is so inelastic that the variable tolls are not an effective way of shifting or reducing demand, or that the price differential has not been made large enough to have an effect.

What does seem to be effective is the exclusive bus lane that allows AC Transit buses to cross the bridge nearly as quickly during the morning weekday peak as at off-peak times.  I don’t understand why it doesn’t provide the same benefit during the afternoon peak.

Graphic: Eric Fischer
Weekend crossings. Graphic: Eric Fischer

John Goodwin, a spokesperson for the Metropolitan Transportation Commission, said it’s too early to draw any conclusions but a group of UC Berkeley professors, including Robert Cervero, was hired to do a comprehensive analysis of the new toll schedule, including the impact on the Bay Bridge. Some preliminary results are expected to be released next week.

“It’s still far too early to draw any meaningful conclusions but we’re very much interested in getting as much data as we can. We’re certainly looking forward to receiving the initial findings,” said Goodwin.

A report [pdf] released by the MTC in October showed that Bay Bridge traffic did not significantly decline in 2010 when compared to 2009.

  • James

    Does this account for changes in overall traffic volume between 2008 and 2010?

  • Eric Fischer

    No, it doesn’t. I wish I had data from just before the toll increase for a more direct comparison, but unfortunately I don’t.

  • pceasy

    You are forgetting that the economy was much worst in 2008 and gas prices were higher in the first part of the year so more people took BART. Prices were lower in 2009 and 2010 – although at the end of the year gas prices were climbing to the 2008 levels.

    Also what is the dedicated AC Transit lane you are speaking of? Are you talking about the HOV/Bus lanes?

  • James

    That makes it hard to know whether or not an increase in overall traffic affected the numbers, though I’d guess numbers of vehicles per day is a number that’s recorded somewhere.

    Interesting, though. I wouldn’t be surprised if it ultimately had no effect, as $1 isn’t much on the whole. I wouldn’t reschedule my day around $1. I start thinking about riding my bike somewhere when MUNI roundtrip is $4, though.

  • Eric Fischer

    Yes, the bus lane through the toll plaza. Sorry to be confusing.

    I know 2008 is not a perfect baseline, but it’s what I have. If I can get additional data, I will do more comparisons.

  • Jon

    Can you define what a crossing time is? If it’s from the time the toll is paid to, say, the Folsom/Embaradero exits then I wouldn’t expect a difference. Wait time is a big factor and could help deduce an overall quantity based on your limited sample size.

    Really though, isn’t volume or volume per time increment the key metric you’d be after to judge whether congestion pricing works (in discouraging driving)?

  • Thanks for sharing this research, Eric. Can you explain what the second graph is? I don’t see it mentioned in the article.
    I think it’d be helpful if you’d label the axes and give each graph a title, too — I think the x axis is time of day in hours and the y axis is crossing time in minutes, right?
    Thanks again!

  • Eric Fischer

    I was defining it as the time from either the 80/580 merge point or the West Grand/Maritime intersection to where the bridge crosses over Beale street just before the first SF exit. It would probably be good to break it down into time spent before and after the toll booths too, and also to account for backup further back into Oakland and Berkeley.

    I was interested in whether it made crossing the bridge faster, since that was the benefit that was claimed to the public, although that and vehicle count ought to be related. I hope the MTC study will have detailed vehicle count information that I don’t have.

    James, various MTC publications seem to claim that traffic on the bridge has been falling every year since 2003. But I can’t find one yet that gives numbers for the whole series.

  • Eric Fischer

    Sasha, the second graph is the crossing time on weekends. Sorry for the poor labeling. Yes, the horizontal axis is time of day in hours and the vertical axis is time to cross the bridge in minutes.

  • Grahm

    Eric, Nice work! This is getting nit-picky, but congestion pricing is when the toll or fee is set to a level that has influence on demand. The bay bridge is implementing variable tolling in which the toll amount changes, but doesn’t seem to influence demand. Essentially MTC or BATA is skimming a few extra dollars without really improving travel time between your start and end points.

  • The bus lane doesn’t help in the afternoon because it is only on the non-bridge part of I-80 and at the toll plaza. Since there is no toll plaza to get on the bridge going east in the afternoon, there is no metering of traffic onto the bridge like there is at the plaza. The carpool/bus lane doesn’t come into the picture until the buses are off the bridge in the afternoon. After that, it does help reduce the commute time compared to not having the special lane.

    On possibility for the lack in change of traffic is that the bridge toll is still less than a roundtrip BART ticket from most East Bay locations to the City. I.e. it might not be at the right price point.

  • Eric Fischer

    Goldenrail, I should have made it clear that this data is entirely for westbound bridge traffic, in both the morning and the afternoon. But I should have also calculated the speed of eastbound traffic to see whether there was any change in that over time even though there was never any charge for it.

    Grahm, thanks for the clarification about the meaning of congestion pricing. Sorry to muddy the issue.

  • Ah, thank you. I just assumed the commute for people working in the City. Didn’t think about reverse commute.

  • tth22

    Caltrans meters the bridge at peak times, so unless they increased the metered flow rate, then wouldn’t travel time from one side of the bridge to the other be relatively unchanged unless there was some change in vehicle speeds on the bridge itself? You sort of indicate this, since you discuss increased travel times attributable to the S-curve. Given the limited boundaries of the crossing time (Grand to Beale), unless there was a large decrease in peak period queuing at the toll plaza, you seem to be measuring a fixed amount of congestion before and after the toll was implemented. That is, the vehicle queues at the toll plaza generally still extend beyond the 80/880/580 maze during the commute peaks, so the only way to get a true before and after crossing time would be to observe time spent in queue from the point the queue begins (before and after) until Beale. For example, if the queue on the Grand approach to the toll plaza extended back to Maritime both before and after the toll increase was implemented and the ramp meters were kept at 9,000 vph, then you wouldn’t expect to see a change in travel time to cross because of the metered constraint. Rather, if you measured the length of the queue (and maybe the time spent in it) before and after the toll was implemented, then you could (perhaps) get a better sense of the affect of the toll increase.

  • Eric Fischer

    Tth22, the time here includes time spent waiting at the toll booths and metering lights, not strictly on the bridge itself, so I think it should be reasonably representative of the total experience. But you are right that delays can back up further so it would be good to calculate that as well. The tricky part is that the further back you go, the harder it is to distinguish between waiting for the bridge backup and being stopped for some other reason.

  • =v= Fascinating stuff, overall. This is a 6-mile stretch and the speed limit is 40mph, so I am shocked, SHOCKED! that anyone’s getting across in less than 9 minutes. Naturally these are during the less-congested hours.

  • Jeffrey W. Baker

    Here’s a great side-effect of the variable toll. At 9:59AM, drivers quite literally stop dead in their lanes, or pull over on the shoulder, to wait for the lower toll at 10AM.

  • Jon

    Eric, since the data does include an aggregated wait time, I think there’s still some interesting trends in the graph. If the MTC flaunted reduced wait times from congestion pricing then this claim isn’t coming to fruition. It’s not unheard of to wrap a grab at higher revenues under the guise of benefits that aren’t likely.

    As for the bus transit times, I think the increased crossing times in the afternoon is an effect of downstream factors. There’s less wait at the toll plaza so the buses don’t really benefit from the HOV lane. Any afternoon congestion is often due to the 80/101 split and traffic getting out of the city headed south on 101. This often slows traffic as early as the bridge (or the toll plaza on weekends). Compared to weekday mornings where most traffic has exited off the freeway by 5th.

    I’d still like to see a higher sample size or a good distillation of MTC data. I do think that your graphs show what any drivers are probably seeing on the roads – that there’s not much of a difference from before the increased toll.

    +1 to Jeffrey. The CHP should be issuing tickets for the increased danger posed by people pulling off to the shoulder to wait. I had a pretty close call once getting cut off in front of me and it was after 10 (clearly they didn’t check the clock).

  • Eric Fischer

    Jon, thanks for the explanation of what is going on in the afternoon! I should redo the analysis in a way that identifies where the bottleneck is at different times.

    By the way, the MTC presentation linked above claims that the CHP is ticketing people who stop and wait for the toll to change and that fewer people are doing it as a result.

  • EL

    Further proof that the extra $1 makes no real difference in commute patterns. Kinda makes you wonder about transit models, where you’re talking about fare changes of 25 cents, doesn’t it?

    It would be interesting to see if the bridge revenues changed significantly too, because if there’s no traffic difference, then….

  • Eric Fischer

    In case anybody is still wondering, I found the table of traffic counts by year:

    Still too early for 2010-2011 totals, but the 2009-2010 paid traffic was down 0.8% from 2008-2009.

  • Patrick M. Mitchell

    I’ve crossed that bridge 5 or 6 times a week, every week, throughout the sampling period. Absolutely nothing has changed except the weight of my wallet. FasTrak lines are longer than ever (though they seem shorter in cash lanes…which leads to dangerous lane cutting, but that’s another post…)

    @Jim Dyer: the speed limit is only 40mph at the S curve; the remainder of the bridge is still 50mph.

  • So what’s this variable rate “congestion” toll pricing about?

    It’s simply a way for MTC’s inexplicably unindicted Steve “$5 billion Bay Bridge cost overrun” to ease in (inevitably fictitious, as in all MTC scams) HOT lane revenues as the Big Revenue Idea for their Regional Transportation Plan.

    With toll bridge variable tolls a “success” (just like TransLink(sm), BART “operates at a profit” to Millbrae, FasTrak(c), BART to Warm Springs, widened 101 and widended Caldecott Tunnel), HOT funding a fictitious set of regional capital projects will be approved by the captive/brainless/on-the-take MTC board.

    But — and here’s the BIG SURPRISE — only two of the non-highway projects in the RTP will go forward (PBQD’s Central Subway abd PBQD’s BART to San Jose) and everything else will be defunded when, in just a few short years, it “unexpectedly” turns out that due to “unforseen economic conditions” that the massive regional HOT revenues “failed to materialize”.

    That’s the way it’s worked for the last 20 years, it’s been very very profitable for those involved, and there’s no reason to change things at this point.

    So no, don’t expect variable rate bridge tolls to have any effect on aggregate throughput or on spreading peaks. That’s simply not the point — no more than adding a single new transit rider is the point of MTC diverting $300 million of your tax dollars to Cubic, Inc, for TransLink(c)(sm)(c)(r)Clipper(tm) was.

    Data? Who needs those?

  • Jon, I redid the plots to show the travel time as contours instead of just total time so it is possible to see which the slow parts are:

    You are right that what slows AC Transit down in the afternoon is a general slowdown of the bridge as it approaches San Francisco. The buses still make it through the toll plaza quickly.

  • David

    Exactly right. The “congestion pricing” was just a fig leaf for higher tolls to pay for cost overruns and mass transit that no one will use (BART to Livermore?! etc).

  • Lewis

    Three points:

    On weekdays, $6 vs. $4 is a very small difference in tolls given the extreme difference in demand at rush hour.

    The curve does show less variation in 2010 between peak and off-peak times.

    The best measure is the number of people moving across the bridge. If people are now carpooling more often, then it’s an improvement even if flow and speed are the same, because more people are moving.

  • Jon

    Eric – Nice shout out in the Gate today. Clearly more data is needed. Interesting though, and for Lewis’ point above, that carpools on the bay bridge are down. Backs up the unscientific observation that at least casual carpools are down. In the other gate article though, it appears that peak morning volume is the same despite the toll. I’ll still look out for the Berkeley and MTC report.

    Regrettably, I’ll be leaving SF for Oakland (saving $600/mo on rent) so I’ll be seeing a lot more of morning bridge traffic about 2-3 time a week that I bring a car into SF (and only in the carpool lane).

  • Goddess

    When is the less congested hours…?? just want to know so I can plan my trip to the bridge from Santa Rosa… im new.. thanks


Reporters, Upset Over Bridge Toll Increase, Get Weird And Whiny

Photo: Mrs. Gemstone The Metropolitan Transportation Commission (MTC), the Bay Area’s transportation planning body and the administrator of bridge tolls, managed a feat very nearly impossible today: They got the Bay Guardian and the Chronicle to agree on something. Namely, writers for both papers hate the idea of increased tolls on the region’s bridges starting […]

Bay Area Toll Authority Mulls Toll Increase Scenarios, Seeks Public Input

Photo: kpmarek As the Bay Area Toll Authority (BATA) continues its regional public hearings, including one tonight in San Francisco, the various options the agency is proposing for increasing bridge tolls are generating a number of debates and proposals, including the funding of the long-discussed pedestrian/bicycle/maintenance paths over the west span of the Bay Bridge. […]

Bridge the Gap!

Photo: Matthew Roth As I climbed the steps out of the Lake Merritt BART station this morning I heard loud chanting. "Wow," I thought, "those bicyclists have really pulled out the troops!" But the demonstrators that greeted me across 8th Street in Oakland were pile drivers, iron workers, carpenters and other trades workers, chanting "Jobs […]