What Does Transit Do About Traffic Congestion? (updated)

A revised and improved version of this post, in response to excellent comments.

Now and then, someone mentions that a particular transit project did not reduce traffic congestion, as though that were evidence of failure.  In fact, the relationship between transit and congestion is indirect.  (In this post “congestion” means that volume/capacity ratio for motor vehicles on a roadway is high enough to substantially reduce average speeds.)   It is not always wise to claim congestion reduction as a likely result of your proposed transit project.

Road widening, however, is also not a very good way to relieve congestion, except in the short term.  In his 1992 book Stuck in Traffic Anthony Downs described the effect of widening an expressway in terms of a “triple convergence”:

In response, three types of convergence occur on the improved expressway: (1) many drivers who formerly used alternative routes during peak hours switch to the improved expressway (spatial convergence); (2) many drivers who formerly traveled just before or after the peak hours start traveling during those hours (time convergence); and (3) some commuters who used to take public transportation during peak hours now switch to driving, since it has become faster (modal convergence).

Today we call this phenomenon induced demand, and it’s the main reason that widening roads doesn’t reduce congestion (and why removing roads doesn’t always make congestion worse.)

And Downs is describing only the immediate effect of the road expansion.  Further increases in traffic will come from any new development that is attracted to the road’s catchment as a direct consequence of its expansion.

So the only way to make the congestion benefit of new road capacity permanent is to severely restrict development in the catchment area of the road — an impossible bar in most cases.  In fact, parties who will profit from further development in a corridor may be part of the political consensus in support of a road expansion, even as the same expansion is marketed to existing residents as a congestion reducing project.

Otherwise, there appear to be two broadly applicable ways to relieve congestion in a substantial and permanent way.

  1. Economic collapse.  Traffic congestion drops during economic slowdowns, because fewer people have jobs to commute to, or money to spend on discretionary travel.  A complete economic collapse, which causes people to move away from a city in droves, is always a lasting fix for congestion problems!
  2. Correct pricing of road space.  Fundamentally, congestion is the result of underpricing.  If you give away 500 free concert tickets to the first 500 people in line, you’ll get 500 people standing in line, some of them overnight.  These people are paying time to save money.  Current prevailing road pricing policy requires all motorists to act like these frugal concertgoers.  Motorists are required to pay for road use in time, rather than in money, even though some would rather do the opposite and our cities would be safer and more efficient if they could.  Current road pricing policy requires motorists to save money, a renewable resource, by expending time, the least renewable resource of all.

So if transit isn’t a cause of reduced congestion, what is its role?  Do transit advocates offer nothing in response to congestion problems that have many voters upset?  In fact, transit’s role is essential, but its effect is indirect.

  • Transit raises the level of economic activity and prosperity at a fixed level of congestion.  Congestion appears to reach equilibrium at a level that is maddeningly high but that can’t be called “total gridlock.”  At that level, people just stop trying to travel.  If your city is car-dependent, that limit becomes the cap on the economic activity — and thus the prosperity — of your city.  To the extent that your city is dependent on transit, supported by walking and cycling, economic activity and prosperity can continue to grow while congestion remains constant.  For example, commenter Brent writes“Toronto achieved significant downtown employment growth without increasing road capacity after the 1960s, thanks first to increased subway ridership and later due to increased commuter rail ridership. Congestion is still bad on the roads and expressways into downtown, even with transit expansion, but (as you say) the expansion of transit has permitted the downtown to grow beyond what the road network would have supported.”  A similar pattern can be observed in many similar cities.
  • Transit enables people who don’t drive to participate in economic life.  Groups who don’t have the option to drive include many seniors and disabled persons, some youth, and a segment of the poor.  There are also many people who are just terrible drivers, and/or who hate driving.  Providing mobility to these groups is not merely a social service; it also expands prosperity and reduces emissions.  (The latter effect is because many of these trips would have gone by car — driven by another person — if transit where not available).  This benefit of transit should routinely be described in terms of economic efficiency, emissions, and vehicle trip reduction, rather than just by appeals to pity or “economic rights,” as social-service language often implicitly does.
  • Transit-dependent cities are generally more sustainable than car-dependent cities.  They cover less land and tend to have fewer emissions both per capita and per distance travelled.  The walking that they require is also better for public health, which produces further indirect economic benefits in reduced healthcare costs.
  • Intense transit service is essential for congestion pricing.  Congestion pricing appears to be the only effective and durable tool for ensuring free-flowing roads while maintaining or growing prosperity.  Congestion pricing always causes mode shift toward public transit, so quality public transit, with surplus capacity, must be there for a pricing plan to be credible.
  • Surface exclusive transit lanes (for buses, rail, and arguably two-wheelers and taxis) improve the performance of emergency services.  This argument should be much more prominent, because even the most ardent car-lover will understand it.  Few things are more distressing than to see an emergency vehicle stuck in traffic, sirens blaring.  When confronted with this, all motorists do their best to help.  But if the entire width of a street or highway is reserved for cars (moving or parked), and is therefore capable of being congested, it can be impossible to get out of the way of an emergency vehicle even if every motorist present has the best of intentions.  Emergency response should be one of the strongest and most obvious cases for surface transit lanes.  Motorists understand the need to drop to a low speed in school zones, to protect the life of every single child.  Why do we not accept come degree of delay to save a child who may be dying somewhere else, because the ambulance is stuck in traffic?

In the end, of course, “congestion” is not a good measure of the outcomes of transit.  In fact, the very notion of congestion presumes a motorist’s view of the world.  I agree with commenter Rodrigo Quijada when he writes:

What we’d like to do in a city is to reduce TRAVEL TIMES. Reducing congestion is a way to do that, but in no way the only one. Over the decades, in places where car transportation has become dominant, people have got used to see travel times and congestion as the same thing, thus orienting their thinking and their solutions to reduce congestion. But this is essentially a confusion.

Still, in real-world transit politics, selling transit projects to current motorists is a necessity, and the current motorist is likely to see her problem as one of congestion.  So it’s important to be clear on what transit can readily do for her.

  1. It can provide an alternative to driving which may be faster, more cost effective, and less stressful.  This argument can be put quite selfishly:  Good transit won’t eliminate congestion in your city, but it can eliminate it from your daily life.
  2. Transit helps reduce government spending on social services by enabling transit disadvantaged groups to participate in the economy.   This obviously has a range of health and wellness benefit apart from its economic role.
  3. It can increase the level of prosperity at a fixed level of congestion.
  4. Its exclusive lanes protect emergency vehicles from congestion-related delays, potentially saving lives.


21 Responses to What Does Transit Do About Traffic Congestion? (updated)

  1. Jonathan Rabinowitz July 30, 2010 at 7:42 am #

    Jarrett, great post. I would quibble however with the effect of congestion on emergency vehicles. I worked for several years driving an ambulance, and I can affirm that one benefit of having hospitals distributed around the city is that congestion on highways doesn’t become an issue for emergency patient transport. In New York City, nobody lives more than 10 minutes from a hospital, and a fire house is even closer.

  2. Peter Smith July 30, 2010 at 8:36 am #

    I’m not sure what the reluctance is to state plainly that transit reduces congestion. It’s been studied and proven, as previous commenters have noted, but still, it’s apparently ‘something which cannot be said’. Why?
    Transit reduces congestion by some amount for some period of time — about five years is what I’ve read before. This is significant.
    And we have to assume population/other growth — so even aside from the early decrease in congestion, there is the additional effect of slowing the growth of congestion.
    To restate, here are the simple, plain, truthful answers:
    1) Does transit reduce congestion.
    2) Does transit slow the growth of congestion, too?
    This has been another edition of Simple Answers to Simple Questions.

  3. rico July 30, 2010 at 9:09 am #

    Great post,
    I enjoyed the original post, but the updated post really seems to have crystalized the ideas. Especially the clarification of congestion vs travel time. Great job.

  4. Dan July 30, 2010 at 10:40 am #

    Here’s another perspective on your first point:
    The benefit of transit operating in a largely exclusive right-of-way (not in mixed traffi) is that it fixes, or guarantees, transit travel times for the future.
    What that means is that while roadway congestion may worsen in a city and auto travel times increase over time, transit travel times will stay constant (and experience less variability), and as a result become increasingly competitive.
    Let’s be clear. That new LRT/BRT/Metro line doesn’t eliminate or ‘cure’ congestion. But what it does is act as a localized treatment or management strategy that frees a city’s economy and society to function and thrive while congestion occurs.
    Some degree of congestion is almost inevitable. But having your city suffer due to that congestion isn’t.

  5. EngineerScotty July 30, 2010 at 12:58 pm #

    One way that transit (in exclusive ROW) minimized congestion is by ensuring that volume of service never exceeds the capacity of the system–the volume/capacity ration is always well less than one.
    The analogous policy for automobiles is quotas–not tolls or congestion pricing, but limits on the number of autos which may be owned or operated at any given time. Freeway ramp meters are one way of placing a quota on a freeway; and roadway pricing does impose a de facto quota if the price is set sufficiently high, but transit solsves the problem by ensuring that consumption never exceeds capacity to being with.

  6. Steve Lax July 30, 2010 at 5:41 pm #

    @Dan and @Scotty – You both suggest in somewhat different ways that transit (in exclusive ROW) can guarantee travel time/not be above operational capacity, if I understand you correctly.
    However, this is not a given. Both operational breakdowns and too tight scheduling can cause the transit reliability model to breakdown rapidly. And it is often politically difficult to tell both riders and politicians that it is unwise to add to frequency (or to add additional stations) because the quality of the service will degrade. So, you end up with situations like NJ Transit and the LIRR into Penn Station, New York operating at the absolute peak frequency such that the slightest hiccup destroys reliability or you have a Port Authority Bus Terminal in New York that cannot handle the volume of buses scheduled.

  7. Ben Smith July 30, 2010 at 6:17 pm #

    In the last thread, the example I gave was an example of modal shift, and how in many cases transit cannot “cure” congestion while remaining competitive.
    With that said, while we COULD reduce congestion by widening roads to unimaginable sizes, we don’t because it would not produce a net benefit. This got me thinking, if we moved up the mode to compete with the roadway (example: frequent commuter rail to compete with highways rather than subways/metros), would the investment provide a net benefit, or is it better to use a lesser mode and deal with a level of congestion?

  8. alex July 30, 2010 at 8:25 pm #

    I agree with the analysis, but everyone seems to be missing the point — the primary benefit of transit is economic development not reduction of traffic congestion. This is why every real estate developer on the planet loves it when cities invest in more transit, particularly rail lines.
    Yet Mayor Villaraigosa, most every other politician, and many other transit advocates (particularly those vehemently for more rail) have been consistently selling transit to the public as the solution to our congestion problems.

  9. Jarrett at HumanTransit.org August 1, 2010 at 12:45 am #

    @Alex. I don’t think it’s wrong to tell a voter that transit is the solution to her congestion problems, in that it will give her an alternative to driving. But it won’t solve the congestion problems of people who continue to drive.

  10. samussas August 1, 2010 at 6:54 am #

    “What that means is that while roadway congestion may worsen in a city and auto travel times increase over time, transit travel times will stay constant (and experience less variability), and as a result become increasingly competitive.”
    Not completely true because transit like roads can get congested and its performances can be influenced by the surrounding road network congestion (in a case of tram/LRT/bus/BRT system). You can have signal priority but if the road is not cleared, you can’t cross the intersection. Also, if transit get congested you’ll see travel time increase and headways becoming variable. Exclusive ROW can make transit more reliable to a certain extent.

  11. jaded August 1, 2010 at 2:57 pm #

    Good post. Personally, I like to take transit. And choose transit when driving is stressful or parking is annoying. Unfortunately, I live in an area where transit service is targeted at commuters, so my weekend and evening options are limited. As a result, I transit to work, and drive the rest of the time.
    In my area, I constantly hear “drivers” complaining that transit doesn’t pay for itself, etc. etc. It is clear they have no concept of how transit helps them by reducing travel times.
    Economic participation is critical as well, and I wish this was used as an argument more frequently. A few of my friends are 100% transit because they choose not to drive. For one, a car would be completely unaffordable. For the second, using transit offers them a better quality of life and lower cost of living (more $$ for fun.) Who can argue with these benefits. I am hoping to move into a more transit friendly neighborhood, and trash my car when it dies in the next 12 months.

  12. Chris August 2, 2010 at 2:09 pm #

    The three convergences are interesting. Has there been any study quantifying the relative effect of each of the three? Many people who live in an affected area would like spatial convergence, for example, and would probably think of it as a positive. Unfortunately it seems as though spatial convergence – changing from a local route to a rail line – is the most common in transit.

  13. Nathanael August 3, 2010 at 11:50 pm #

    “These people are paying time to save money.”
    Oh God — you just explained something really clearly.
    *This is why wealthy, busy people often like grade-separated trains, but have little use for mixed-traffic buses.*
    They would rather spend money to save time. They also prefer trains they can work (or read) on to wasting their time driving. (Some can work on buses, many can’t; everyone can work on trains.)

  14. EngineerScotty August 4, 2010 at 12:08 am #

    Everyone can work on trains
    While it’s easier to work on a train then a bus, as you generally aren’t subject to either lateral motion (of the bus pulling into/out of stops), or the bouncy suspensions necessary to deal with typical uneven road surfaces–doing meaningful work on a train that is crushloaded tends to be difficult. Even if you are lucky and have a front-facing window seat, it’s not a terribly conducive environment to productivity.
    One reason that commuter rail services and express busses are popular with the working set is that they often have “coach-style” seating–everyone gets a seat, no standing in the aisles.

  15. Jarrett at HumanTransit.org August 4, 2010 at 12:46 pm #

    Nathanael.  You know, however, that trains are not always grade separated, and buses are not always mixed with traffic.  The rail-bus distinction is not the same as the exclusive/mixed distinction.  It's the latter that determines reliability.

  16. M1EK August 5, 2010 at 9:13 am #

    But we also know that a train is far more likely to be grade-separated than a bus, and far more likely to have reserved guideway than a bus. In the beautiful abstract world, there’s no difference between the two, but in the real world, there almost always is.

  17. Jarrett at HumanTransit.org August 5, 2010 at 11:10 am #

    M1.  If we allowed ourselves to be constrained by current habit and cultural history, as you propose, no progress would be possible.  It is only by breaking out of past cultural associations that any progress happens at all.

  18. Alon Levy August 5, 2010 at 11:27 am #

    American LRT systems are rarely grade-separated, and even more rarely grade-separated for reasonable cost. It creates serious problems for cities that have constrained ROW or NIMBYs who think the ROW is constrained, because it forces a lot of expensive underground construction.

  19. M1EK August 5, 2010 at 11:59 am #

    Jarrett, again that ignores the quite likely possibility that the reason so few BRT or Better Bus proposals really deliver “like rail” is because there’s something fundamental about the technology that leads it to be applied in that way, despite what would happen in the perfect abstract world. (Can you tell I was reading one of your old postings about LA today?)

  20. EngineerScotty August 5, 2010 at 3:14 pm #

    Jarrett, again that ignores the quite likely possibility that the reason so few BRT or Better Bus proposals really deliver “like rail” is because there’s something fundamental about the technology that leads it to be applied in that way, despite what would happen in the perfect abstract world.
    There are a few bus/rail differences off the top of my head, that may lead to this dichotomy. I’m ignoring things which are entirely cultural factors (“Americans hate busses”), but am including some relevant political facts (things which reflect current law, common existing practice in public transit, and the like) which aren’t strictly technical differences:
    * Bus, especially “rapid” bus, often gets moved into mixed traffic because it CAN be–high-speed, high-weight rail (anything faster or bigger than a streetcar) simply is incompatible with mixed-traffic operation in all but the most constrained circumstances. (MAX and auto traffic used to share lanes on Portland’s Steel Bridge; though nowadays the cars have been kicked out of that lane and it’s exclusively for trains). Sometimes this is good (open BRT branching off in the suburbs) or practical (sharing a lane on a bridge to avoid constructing a new one); other times its bad (frequent mixing of traffic simply to build on the cheap; where cost/benefit analysis might suggest otherwise). High-speed rail HAS to be in its own ROW, for the most part–rail vehicles longer than streetcars simply have stopping distances too long for safe mixed-traffic operation, other than when travelling at very low speeds.
    * Rail, by being virtue of rail, can demand absolute signal priority at crossings, at least under US law. A transit authority building a rail line has the right to install grade crossings (gates and such), regardless of complaints from the roadway authority. Bus, on the other hand, has to negotiate this with state DOTs, city public works departments; many of whom refuse to cooperate, or place limits on signal priority schemes, often to the advantage of auto traffic (and the detriment of busses).
    * Fixed guideways make platform approaches easier and more reliable. There are some technical solutions for busses being developed, but rail has an advantage here.
    * Many rapid transit projects are also done with significant environmental benefits in mind; few US transit authorities operate trolleybusses (or are interested in doing so). As a result, the choice between electric and combustion-powered traction is, for all intents and purposes, a choice between bus and rail. If a key project criterion is eliminating greenhouse gasses or other pollutants, or reducing dependence on petroleum, this often produces an advantage for a rail-based solution–especially in locales where electricity is produced by means other than burning dead dinosaurs. (And even where fossil fuels are burned for electricity; a large point-source powerplant is generally more environmentally friendly then large numbers of mobile internal combustion engines; especially ones hooked to vehicular drivetrains).
    And as discussed in numerous other articles on the subject, there’s a distinction between “open” and “closed” BRT. A more accurate statement is that there’s a continuum with relevant tradeoffs between how “fixed” the guideway is. The more you “fix” things–the more closed the system becomes–the more operational efficiencies you can achieve. The more you limit yourself to street-legal, local-traffic-optimized vehicles (the more open the system becomes), the more flexibility you gain, but you lose opportunities for optimization. Rail rapid transit lies at one end of this spectrum for rapid transit; a BRT system consisting only of route improvements but which uses standard vehicles, platforms, and fare collection policies, lies at the other end.

  21. tomtakt October 20, 2011 at 5:05 pm #

    Obviously, I’m a little late to the party on this one, but you’ve forgotten a major point!
    Travel time reliability.
    This is a major reason why vehicle congestion is a problem. As you mentioned, congestion rarely stops traffic movement completely, but the problem is that congested or near-congested roadways become unpredictable because of the numerous opportunities for chance occurrences (crashes, sudden braking, debris) to induce major capacity reductions. Many people’s effective travel time is thus even longer than their actual average travel time because they may need to leave earlier on a regular basis to ensure being on time just in case they get stuck in traffic.
    Quality transit usually runs much more reliably because there are fewer opportunities for human error or chance occurrence to disrupt a mostly closed system like with rail or BRT. Obviously, traditional local bus service is an exception to this, although sometimes its delays are so much more related to passenger levels (boarding time) rather than traffic conditions that it may still be more predictable than driving times (in that ridership at a given time of day is more predictable than the likelihood of a crash or other chance event).
    But, the most notable difference in reliability is probably during extreme weather events, when transit usually vastly outperforms vehicle travel (congestion or no congestion). And that’s true even though many drivers simply avoid bad weather completely. This is another huge economic boon provided by transit in that work is less likely to stop due to weather issues.