Paradise LOSt (Part I): How Long Will the City Keep Us Stuck in Our Cars?

Editor’s note: Today we begin Part I of our occasional series on LOS reform.

Bus_in_traffic.jpgTraffic engineers are reluctant to give exclusive lanes to buses (or bikes) for fear of the impact on cars

The Pseudo-Science of LOS

There’s a dirty little secret you should know about San Francisco: It’s engineered first and foremost for automobility and will never be able to shed this bias if the traffic engineers are in the driver’s seat wielding their traffic analysis tools like bibles. As long as the city continues prioritizing the use of transportation analysis known as Level of Service (LOS), you might as well burn our Transit First policy for warmth.

On the one hand, LOS is a very simple and blunt metric for understanding the speed that vehicles can move about the city. LOS measures the amount of vehicular delay at an intersection, with A through F grades assigned to increased delay. This measurement is taken during the peak 15 minutes of evening rush hour and if an intersection slips from LOS D to LOS E, traffic managers will try to mitigate the impact, which usually means widening the road, shrinking the sidewalks, removing crosswalks, softening turning angles, and adjusting signal timing to speed the movement of vehicles.

LOS_Graph.jpgLOS delay from Highway Capacity Manual

LOS analysis seems like science, free from political or ideological considerations, the perfect traffic-engineering tool to rationalize our cities, but the methodology behind it is far from precise. As Jason Henderson, professor of geography at San Francisco State University, said at a recent presentation, LOS is a very poor tool methodologically. In the early years of its development, the "science" was merely traffic engineers assuming what made motorists uncomfortable. He cited the fact that LOS F used to represent a delay of more than 60 seconds, but that in the 2000 Highway Capacity Manual it was revised to 80 seconds. And motorist behavior studies since have shown that inconvenience with delay can depend on numerous factors and differ dramatically between drivers.

Yet the result of relying on this poor methodology to shape the growth of cities has a profound affect on the politics of human mobility, privileging the movement of vehicles before the movement of anything else. Quite simply, LOS analysis has given us Phoenix and Atlanta, congestion and ever-longer commutes, and a whole host of ills that accompany reliance on the inefficient use of street space for our cars.

"I’ve been doing transit analyses in California for 20 years," said Jeffrey Tumlin, principal of Nelson Nygaard, a transportation and land use consulting firm. "In my practice the single greatest promoter of sprawl and the single greatest obstacle to transit oriented development (TOD) and infill development is the transportation analysis conventions under CEQA, the California Environmental Quality Act, LOS."

LOS has been used since the middle of the 20th century and has the weight of convention so thoroughly backing it up that it is one of the Traffic Ten Commandments, somewhere near the top. When the U.S. prioritized the expansion of roads everywhere and cities were seen merely as job centers to be driven to and from, LOS was the lubricant to facilitate the growth of suburban and exurban rings.

As San Franciscans began to see their city as a collection of neighborhoods, where livability and public space were more important that vehicle speeds on the streets, the political support for unfettered automobility declined and LOS became a target for reform.

"How can LOS have such a tremendous impact or power over shaping San Francisco and shaping cities when so much of LOS is anathema to cities?" asked Henderson.

The Clamor For Change

In the early 1990s, San Francisco bicycle advocates found out the hard way that any attempt to take space away from cars to give back to bikes (or transit and pedestrians) would be shut down by the engineers who asserted that such constriction of vehicular access violated the sacrosanct LOS rules, which, they were informed, were required under CEQA.

The advocates soon discovered that while CEQA requires that a project be analyzed for its significant environmental impacts the LOS "rule" was written after the fact by the state Office of Planning and Research (OPR).  It was an administrative guideline and didn’t carry the force of CEQA law.

"The words ‘traffic’ or ‘congestion’ or ‘parking’ appear nowhere in the CEQA legislation of the 1970s," said Tumlin.  "The
problem lies in the CEQA guidelines, issued administratively by the state OPR.  There are three lines in Appendix G, the environmental
checklist…[that] were later adopted administratively by the executive branch of government."

Further, CEQA specifically delegates planning decisions to the local level.

The adherence to LOS was a convention adopted by San Francisco and most other municipalities around the state to evaluate transportation impacts under CEQA, and over the course of nearly three decades its use had been upheld numerous times in court, but just as it had been adopted as a convention, advocates believed, so could it be overturned.

Subsequently, advocates spent years lobbying the Board of Supervisors to make changes to the rule, though with mixed results.  After the board finally called for a study of the impacts of LOS analysis, the San Francisco County Transportation Authority (TA) issued a significant report (PDF) at the end of 2003, which asserted that "existing LOS measures and standards ostensibly favor preserving auto level of service at the expense of improving transit, bicycle, and pedestrian conditions" and argued that new measures of analysis besides LOS should be developed.

Within a couple years the Board of Supervisors was convinced that LOS analysis contravened San Francisco’s Transit First policy and issued a resolution (PDF) stating that "automobile analysis alone is not an appropriate metric for assessing environmental impacts and for analyzing projects that may improve overall environmental quality in conformance with [the Transit First policy]."

Not long after the resolution was adopted in early 2006, advocates saw firsthand how low the interpretation of LOS could stoop, when Rob Anderson and his Coalition for Adequate Review (CAR, a suitable acronym) sued the city to prevent the Bicycle Plan from being implemented without the full environmental review process, banking his argument on the possible negative LOS implications of removing vehicle lanes to add bicycle lanes.

Though it might have seemed ludicrous to the lay observer that adding bike lanes would have a negative environmental impact, the rules of LOS dictated that it was so and the onus was on the city and the bicycle community to show in excruciating–and expensive–detail that promoting non-polluting transportation might significantly effect the environment.  But that’s exactly what the injunction mandated and why San Francisco hasn’t added even a brush stroke of paint to its roads to build out its bicycle plan for nearly three years.

Paradise LOSt (Part II): Turning Automobility on Its Head

Paradise LOSt (Part III): California’s Revolutionary Plan to Overhaul Transportation Analysis

Flickr photo: cbcastro

  • Thank you for bringing this topic to light with such a great article. As the economic situation worsens and cars are reposessed by the millions, we will hopefully see more pressure to improve transit options that are less expensive than individual car ownership. Changing the LOS metric to allow this is a good first step.

  • Jon

    Thanks for this informative piece. I’ve been a transit worker, for several different public and private properties, in three different cities. It’s been apparent to me that traffic engineering is based on some faulty premises, but I didn’t know how to understand the problem. I’ve understood that our cities were being sacrificed to cars but the mechanisms are sometimes hard to see.

  • this is really informative, thank you. do you have a reference of a different measure that is being used by other cities to better measure the impact of changing street development/positioning/..?

    is LOS the issue, or how it is implemented? or do you want to toss LOS?

  • Thanks, Omar. Stay tuned for tomorrow’s article for more answers.

  • I sure hope part two on LOS has a happy ending; but as a pedestrian, Muni rider and cyclist in SF I see the tragic ending every day. And it’s not happy for anyone.

  • Jake Wegmann

    Terrific concise, focused piece on a hugely important issue. Keep the heat on!

  • Brooke

    Terrific article Matthew! Way to make LOS standards a compelling and readable topic for Streetsblog readers.

  • Jarvis

    I agree that the LOS analysis is much like “hoop jumping”, but to largely attribute our transportation/traffic problems to these analyses overlooks one key point: sprawl happens because most people (planners included) simply cannot afford to live in many of our central cities. If you want to get to the root of the traffic problem, start with the lack of quality affordable housing in the central cities.

  • Sprague T

    Thank you for shedding light on why improvement seems so slow-paced in San Francisco. I recently returned from living in Vienna, Austria for more than ten years and it is clear to me that if a city’s leadership truly endorses transit improvements then great things can happen (providing the laws and funding allow it). The San Francisco Bay Area is a wonderful place but there is so much that could be emulated from other urban areas to improve the quality of life here.

  • Great article I didn’t know that the problem was the CEQA Guidelines, not the CEQA law – and that cities can arguably overturn those guidelines.

  • Kenny

    It makes some sense that a bike lane, etc. that dramatically slowed traffic might actually be a net negative in terms of pollution. It also makes some sense that you’d want to investigate the impact on car traffic when planning changes to the street-scape. Of course, we should be considering pedestrian, bike, and transit traffic as well. It sounds like LOS is not an ideal metric, and some accounting for other modes of transport need to be made. But in general, I’d say this article overstates things and hasn’t made too strong of a case.

  • This is a very important message to spread. Thanks for giving it its due attention.

  • “sprawl happens because most people (planners included) simply cannot afford to live in many of our central cities. If you want to get to the root of the traffic problem, start with the lack of quality affordable housing in the central cities”

    Auto oriented development has a lot to do with this. Parking garages and parking lots are not housing for people, they are housing for cars. Which is more important?

  • So, the question remains, how do you make it so that all modes are equal under the law? So that they all can enjoy a safe street. Does it mean that we have to bring the car down a notch and thus give the other modes better access? Do the cars have way too much space?

  • Our streets – one third of the public space in SF – are primarily designed to keep motorists from feeling uncomfortable. I think that is a gross inequality. I’m glad the discussion to drop LOS is catching steam.

  • greg

    Until we dedicate some road space to cyclists, delays to bike traffic won’t measure in seconds– they’ll measure in years.

  • ian

    it is entirely true that the current methodology for calculating roadway performance is auto-centric. in fact, it looks at peds, bikes, and transit as impedences to traffic flow). that said, i find it sad that matthew roth demonises the traffic engineer as some horrible devil who wants to destroy all that is good in the city. this is more of the same nonesense that divides the planning community and makes people with AICP after their name look like propagandists rather than community focused..and it is entirely untrue. here’s the facts: laws and regulations are made by politicians. at one time, when cars were mostly everyone’s priority, lawmakers asked trarffic engineers to develop criteria for traffic operations. they did so, with quite accuracy, and that was it for ~50 years. now comes the new era when people’s priorities shift as the car overwhelms us and the idea of biking and walking become trendy (argue if yo ulike, but until we’re walking and biking for decades it’s just as much a fad as the prius is to hollywood). traffic engineers are just as capable of developing “complete street” criteria to address all modes and uses if asked. some have even offered such methodologies without being prompted by the pols this time. yes, i am a traffic engineer, but i ride my bike to work all the time, i prefer to walk for my local chores in the dense city i chose to move to, and i advocate reducing auto use as much as possible in my work every day. i am not a devil (although i was trained in those devilish techniques at school) and when the politicians are convinced that LOS should be calculated differently, then i’m sure the older-school in my profession will adopt the new methodology (albeit just as grudgingly as auto-workers would adopt a new profession).

    point is, traffic engineers are not the scapegoat for the sorry state of america’s streets; we all are. and we all need to do something about it to make a change, not just point the finger.

  • Evan

    If CAR was really concerned with LOS, they would ask that an environmental review process be required before any new car can enter the transportation system (paid for by the prospective owner, of course). After all, every new car just adds more delay for all the other cars…and takes up a heck of a lot more space than a bicycle does.

  • Gary Toth

    Matthew, has SF done anything in response to the Council’s 2006 proclamation calling for more balanced LOS measures?

    Have you seen the new NCHRP report 616, which provides a methodology for a balanced LOS calcuation? Google it and you can download the pdf.

    Finally, the Project for Public Spaces wrote an article on traffic projections and LOS, which is posted on the Federal Highway Administrations Context Sensitive Solutions website. See


  • Asst. Trans. Spec.

    As another person/cyclist applying LOS standards but hating it, I agree entirely with Ian.

    The worst thing we can continue to do is allow automobility to run unimpeded.

    Congestion is what we need.

    LOS F = Good

    Then when it’s bad. People will say ‘eff’ this, give me some transit, some bike lanes. LOS can remain, but new CEQA revisions, should rate the expanded lanes, emissions and carnage by cars as significant impacts; with the only potential mitigation being bus/bike lanes and wider sidewalks.

    /@work rant

  • Prolly

    I know ATS, and he’s a pretty solid dude. Been riding bikes for a long time, for what it’s worth.

  • Derek

    “sprawl happens because most people (planners included) simply cannot afford to live in large houses on big lots in many of our central cities.”

    There, fixed that for you. 🙂

  • Gary,
    Thanks for pointing to NCHRP report 616.
    See today’s post for more on what the city is doing in response to the 2006 proclamation for more balanced LOS.

  • Matthew Ridgway

    Good post. Another good source for information on CEQA and LOS is here

  • Joe Student

    More info on how LOS standards are required by the State can be found in the California Government Code 65089 (Congestion Management Plans).

    From San Francisco County’s website (and plan):
    “The California Government Code mandates the development of a Congestion Management Program (CMP) for each county in the state to manage the effects of land use decisions on the transportation system, and vice versa. It requires that all elements of the CMP be monitored at least biennially by the designated Congestion Management Agency (CMA) to determine if the county and city governments, known as Member Agencies, conform to the CMP. The San Francisco County Transportation Authority (SFCTA) is the designated CMA for San Francisco County, and therefore is responsible for CMP monitoring.”

  • I think the ATG described in the second LOS post would satisfy the requirement to have a Congestion Management Plan. It is another way of mitigating congestion, by assessing a fee to support transit rather than by adding road capacity:

    “ATG avoids intersection-specific analysis, instead evaluating new developments based on the number of car trips they would add to the aggregate traffic picture and assessing a transit mitigation fee based on the total number of additional trips.”

  • Tom Rubin

    Let’s not throw out the baby with the bathwater.

    First, while I am certainly not a transportation engineer PE, I am a member of ITE, use many transportation engineering concepts in my work as a transportation consultant, and would like to think I do make some understanding of some of the basic concepts involved.

    First, LOS is a good and useful tool, but it should never be utilized in autopilot, as in, OK, the objective is to produce the specified LOS, do what is necessary to make that happen.

    LOS analysis — properly performed, which should go without saying, but, in this day and age, does have to be said — is a very useful tool in analyzing impacts of proposed changes in transportation systems, land use, etc.

    However, it should never be the ONLY analysis performed.

    Also, keep in mind that there are entire families of LOS standards. For example, for auto use of roads, there is a standard for freeway lanes, for arterial lanes, for signalized intersections, etc. BUT, there are also LOS standards for transit and pedestrial transportation — for example, one transit standard relates to the extent of standees.

    So, one very simple and logical change is simply to ensure that where LOS is being studied for autos on roads, the impacts on other modes is also included, were applicable (there are situations, of course, where some modes of transit are not directly applicable, such as pedestrian and bicylist travel on the Bay Bridge, although there are often good reasons to include some comparable analysis even in these cases, such as, how about a way for cyclists to take their bikes on BART during rush hour or, if that is not desired by BART due to impacts on people carrying capacity, how about having bike lockers at ALL [well, OK, most] BART stations and allowing cyclists to have two lockers at different stations, so someone who wanted to could keep an oldie in SF for that last mile-and-half to the office).

    In fairness, however, you do have to reflect the quantities involved in such analysis. In the core CBD, the volume of pedestrian traffic is very large and the weighting given to pedestrian movement should reflect this; on a residential street with good sidewalks, the impacts of many changes will have little or no impact on pedestrial movement. Also, don’t forget that almost all goods movement to specific SF destinations is via truck (I’m excluding thru shipments).

    Transportation engineering is not a fixed science; it both constantly evolving and, when done properlty, has a significant portion of art — that cannot always be quantified.

    Finally, in regard to CEQA, keep in mind that this is, more than anything else, a checklist of steps to be performed, analyses to be performed, and information to be disclosed to the decision-makers so that they can make intelligent decisions — and to the public, so they can comment on both the recommendations and the decisions. CEQA, in and of itself, does NOT mandage specific actions — AND the decision-making agency has the option to make a decision on the basis of overriding concerns, and this is constantly done.

    It is very unusual to see a court overturn a decision made by a governmental body after an EIR et al on the grounds that the decision was not the best one; EIR’s are tossed out by the courts on the grounds that the PROCESS was not properly conducted, that something wasn’t studied that should have been, or that an analysis was faulty, or that a specific alternative should have been considered and wasn’t.

    Technically, if there was a statement in an EIR that one undesirable aspect of proceeding with a project was that it would lead to the end of life on Earth, and this was properly disclosed, and the government body decided to go ahead anyway, it could be very difficult to challenge the results on CEQA grounds because the process was followed completely and correrctly.

    This type of result may not be what any of us are looking for, but the implication is, even if the auto/truck LOS results of a specific decision are not good, the governmental decision-making body generally has the ability to go ahead with that decision anyway, as long as it can show that the matter was studied by competent experts, the process was follows, and the body has reason to make a finding of overriding concerns, after evaluating the possible mitigations. This has to be the way it works, because even the “best” projects generally have some negative impacts, and the elected officials and their delegates have to have the legal authority to make difficult decisions — AFTER following the required process and informing everyone what’s going on and giving everyong the opportunity to be heard.

  • There’s a lot of obfuscation and pseudo-technical jargon in this discussion that does nothing but obscure the whole point of LOS “reform”: to make it possible for the bike people to screw up traffic for motor vehicles with impunity. The SFBC and its many enablers in city government simply resent the fact that the Bicycle Plan project—and it is a huge project—should have to do any environmental/traffic analysis of their great planet-saving plans for city streets.

  • Reform of LOS would have great benefits to MUNI and pedestrian improvements. If they pass we would not count a single bus as one vehicle, but one with lots of people on it. Plus, pedestrians might actually be considered in the mix too.

  • As Tom points out, LOS should be used a tool – one of many. And there are LOS indicators for other modes of transportation, but they’re not very good and don’t take context into account. For example, a sidewalk that is not very congested (few people) is given a grade of A. Perhaps in heavily urban areas with lots of pedestrian traffic, you want to decrease congestion (through widening, etc), but usually you *want* people on the sidewalk.

    There is reform coming though, as Gary mentioned, in NCHRP 616. That report is part of NCHRP 03-70 ( and will be in the Highway Capacity Manual update. Furthermore, cities are reforming their LOS calculations without national or statewide guidance. Seattle has been developing their method for a while now, and it sounds quite promising.

    I apologize if I’m repeating what others have already mentioned above. I’m just quite excited by multimodal LOS!

  • ian

    ron anderson – i think your concerns are off mark. bike/ped advocates, i believe, are not in it, as you say, “to screw up traffic for motor vehicles with impunity”. the argument, which is entirely valid, is that in the spectrum of roadway classifications, from urban city street to limited access highway interstate, there is also a spectrum of users. one might say that the closer you get to a city center, and away from the interstate, less priority should be given to cars, and more to buses, peds, and bikes. in a rich urban environment with multiple modal choices, this is a rational argument because these alternative ways of getting around are much more efficient, serve many more citizens, and provide more choices. so on highways, cars reign supreme, as they should. but in city centers, driving a car on scarce public right-of-way is a privelege shared to a significant extent with the many others on bike, foot, rail, or bus. and in these areas, the LOS methodology should be able to handle the mix of modes so that all users are taken into cosideration when an analysis of roadway operations is performed. that said, it is true that at the current time, many municipalities are implementing bicycle and pedestrian improvements without a practical justification for how the effects to auto traffic are justified. this could be done by developing a modal priority matrix for various roadways in a city, and designing the streetspace to achieve a balanced LOS that quantitatively meets this priority ranking. so on a highway, peds get little or no service and cars get all the capacity, but on a shared street, cars are down to just 10% of total capacity, with the rest split amongst all others.

    the talk in some traffic engineering circles is similar to what someone said above, that of replacing the 20th century VMT “vehicle miles traveled” with “person miles traveled”. doing so would allow for buses to take priority over private autos. but i think this does not go all the way to address the spectrum described above. something that better handles the full mix of modes for a given street is most applicable.

  • Kevin

    LOS. LOL.

  • Don

    Two Rob Anderson Haikus:

    Hates Water and Air
    Likes the Prison of His Car
    Please, Move to L.A.

    Rob Anderson’s Gaffe
    Fought For Your Oil in Iraq?
    I Didn’t Think So

  • aaron goodman

    The 19th Ave. Transit Corridor Survey is available on the SF Gov. Planning Dept. website, the hearing on this is Feb. 24th (one meeting!) at SFSU in regards to Parkmerced’s Vision Project and the impacts on cumalative development in District 7. It is CRUCIAL to address the concerns of SFSU/CSU’s masterplan and Parkmerced’s vision projects that do not address the traffic/parking issues and mandate for direct regional transit connections of the M-Line to Daly City Bart… 19th ave is a parking lot, and the great highway, sunset blvd., juniperro serra, brotherhood way, and lake merced blvd. are getting there quickly.


Paradise LOSt (Part II): Turning Automobility on Its Head

One of the unintended consequences of San Francisco’s bicycle injunction, which Rob Anderson and fellow NIMBYs will likely rue for some time to come, is the arduous thought and labor that advocates and professional planners have invested in doing away with LOS all together.   Two arguments in the debate over LOS have emerged. One […]

Beyond “Level of Service” — New Methods for Evaluating Streets

Streetsblog reported earlier this month that transportation agencies are increasingly aware of the insidious consequences of using “Level of Service” as the primary metric for their projects. Because Level of Service only rewards the movement of motor vehicles, it promotes dangerous, high-speed streets and sprawling land use. The question remains: How should streets and development projects […]

SoCal Regional Government Wants Exemptions, to Keep Planning for Cars

SCAG sent a last-minute letter attempting to delay progressive updates to California’s outdated environmental standards. In the letter [PDF], Southern California Association of Governments (SCAG)—the regional transportation planning organization for much of southern California—requested exemptions for highway expansion projects and freight corridors from proposed state rules that could show their true environmental impact in a […]