The Portland Bureau of Transportation has released the long-awaited Enhanced Transit Corridors plan, a set of policies and strategies to increase the speed and reliability of Portland’s major bus lines. As in many cities, Portland’s bus lines have been slowing down about 1% a year, which is slowly eating away at people’s access to all kinds of opportunity.
The plan also points the way for the City of Portland to be a more active leader in transit policy, much as the City of Seattle has done so effectively.
The Portland Bureau of Transportation invited me to draft the Executive Summary, which offers some examples of ways to make this topic urgent for more people. I’m happy with how it came out, and people fighting similar battles in other cities may find it useful.
Here’s the Executive Summary. Here’s the rest of the report.
Portland residents: Have your say at or before the City Council hearing on June 20 at 2 pm, in the Council Chambers. You can also submit comments before June 20 by writing to cctestimony at
In Toronto, the TTC has been removing and ‘rationalizing’ stops.
What this means is that many riders have an addtional few minutes/blocks to walk, possibly missing their streetcar/bus. In return, service is no quicker. Often enough, I find myself looking out at the stop that used to be there, but of course I can’t get off because the next nearest stop is further along and the streetcar is stuck in traffic.
It’s amusing that the executive summary linked seems to write off trips under three miles as “faster to bicycle or even walk”. What kind of nonsense is that–it’s pretty much an hour’s walk.
At the same time as transit systems are made accessible, the stops get spaced out. Hey! Why not walk or ride your bike?
Jarret Walker has many good and logical ideas, but when I see the effects of greater stop spacing in the system that I ride regularly, I find the “space stops out more” to be an anti-transit idea.
In fact, my local streetcar stop was removed some years ago. I now ride the streetcar less. It’s a ten-minute walk to the streetcar stop. I can *drive* to many final destinations in the time it would take me just to walk to the stop. So guess what I choose to do?
Good point. I often read comments about how long distances between stops is some sort of panacea. Then I look at the most popular transit system in North America, the New York City Subway. Nine million rides a day and … (wait for it) … they have very little space between stops. On a freakin’ subway!
They get away with it, of course, by having a few express routes. But like most North American cities, not nearly as many as they should. People just have to wait a while if they are trying to get across town. Or they can take an express bus.
The point being that stop spacing really is a zero sum game. Compared to other improvements, it is controversial for good reason. You often make things worse. If a city like Portland suddenly decided to do away with on-board payment, then the buses would run much faster. Everyone — every single rider — would get to their destination faster. If they spent money on bus lanes, then it is basically the same thing. Build a bus tunnel, or a new subway? Wonderful.
But with stop spacing you have to be very careful. I’m all for limited stop express overlays. But if it is the *only* bus route along that corridor, and you make people walk a long distance, then it is quite possible that more people will come out behind than ahead. Let’s hope that Portland has done their homework, and knows what they are doing, instead of simply cutting out a few stops and then bragging that their buses are faster. Because, of course, their buses would be really fast if they simply avoided picking up riders.
The motivation behind calls for wider stop spacing is to get a faster form of transit that is more useful to them and can compete better with driving, and the feeling that more people are traveling between the urban village centers than at the minor stops, yet they have to sit through the minor stops that only get onesies and twosies. And some cities really do need a stop diet: stopping every two blocks is excessive.
In most American cities, subways and commuter rail are faster than parallel buses — or at least not slower — and they’re more reliable and frequent, especially off-peak (although commuter rail is often peak-only), and express buses get caught in traffic jams. New York basically inverts this model: the subways provide local transit for the masses, the express subways function as “limited-stop” intermediate transit, and premium express buses are the fastest way to go long distances from areas without commuter rail. Or at least I’m told that New York’s express buses from Jersey City and the far side of the outer boroughs is very good.
But most American cities aren’t ready for this model. They have slow local buses and no prewar subway or commuter rail, and if they build a subway network it’s a financial stretch, and adding faster express buses from the outer train stations on top of that is too much, or at least the public isn’t sure about adding those taxes on top of the trains, or it’s such a new idea it would take time to gather support. Most American cities are still grappling with the problem that a trip that takes ten minutes by car takes thirty or forty-five minutes on transit, and they need to address that before they get to local subways and express buses. So instead they position subways as better-than-local service (through higher performance and possibly limited stops) — an in-between level of transit the city has never enjoyed before (or at least not since the 1940s, which was before many people were born).
Limited-stop overlays (stops every 1-2 miles, or only at urban villages) provide this in-between level of service, which is sadly missing in most cities. There is a counterargument that the resources for it should go into making the local route more frequent instead, but that’s not very persuasive to those who facing forty-five minute trips on the local route. The best situation would be both: a frequent limited-stop route and a frequent local route. (Although in some cases people vote with their feet and take the limited, thus revealing there isn’t much demand for the local.) But if a city can afford only one or the other (a limited overlay or a more frequent local), then they’ll have to make that tradeoff. Making a limited-stop route the only route is dubious: the in-between stops probably need at least coverage service.
A lot of this has to do with density: in some cities the in-between stops have detached single-family houses and no businesses, and the majority of people in those houses drive, so that depresses ridership from those stops and makes urban-village riders feel like they’re having to sit through these onesies and twosies, whereas if the in-between areas were denser and had more businesses, there might be three or four people at the stop and that would make the stop more justified.
Off-board payment can help, but mainly in high-volume situations or where many people pay cash. Some of these routes have only two or three high-volume stops, and the other stops are usually onesies or twosies, so the time savings from tapping off-board and entering a rear door is little or nonexistent.
Premium buses are fast, but not faster than equivalent rail+bus; the premium mostly comes from the eliminated transfer penalty and the all-but-guaranteed seat (there is no standing permitted in express buses)
It might be worth noting that current bus stop spacing in Portland is often every 2-3 blocks. When there are riders waiting at every single stop (as there often are with the frequent service lines) it has an effect on transit times. Surely it’s not too much to ask riders to walk an extra 2 blocks.
Portland Streetcar also closed some stops and I can say it is much better. Again, these stops were every 2 blocks and ridership was high enough that there would always be people waiting at these stops. However, along with the stop closures was a lane reconfiguration. Now the streetcar, when it is downtown, runs mostly in its own lane (it’s a right-turn only lane for motorists). Even if it only saved a minute or two, the psychological benefits are much greater because it feels like we’re going somewhere.
From your description, it seems that the Toronto system would work if the traffic weren’t in the way.
I know I was annoyed by the fact that the TTC used to put stops on both sides of a traffic light or less than a building width’s away from each other. Most riders didn’t even know those stops existed, except for the few people “in the know” who actually made use of them to shave a few seconds off their walking times, wasting everyone else’s time in the process. I’m glad that those have been “rationalized” away.
We noted that some people ride transit <3 miles, but the fact is that transit at Portland's scale is less competitive at that distance. Need for frequency is inversely related to distance traveled, so for short distances, we need frequencies <10 minutes which, at this point, we can't afford. Cycling also picks up a much larger market share at these short distances.
Speaking of short distances, Jarrett – a thought for a future blog post might be the subject of dockless electric scooters. I’ve seen hot takes on them that run the gamut of opinion – it would be interesting to hear some thoughts on their impact on urban mobility from someone who knows about transportation networks. Anecdotally, I’ve seen far more people zipping around on them in my urban core (Miami) than I ever have with the bikeshare bikes, so that’s at least *one* totally unrepresentative data point.
I’d actually be very interested in a blog post about short distance transit vs long distance transit. I suspect that the two are very different beasts.
Particularly, does the possibility of short distance trips providing an alternative for walking (due to weather, time etc) form a meaningful part of the market for short transit trips, and how does this change the balance between speed, frequency, stop spacing?
In general, I feel like people traveling a short distance to use their feet instead of the bus is a good thing for transit. Short distance riders can cause a lot of crowding, and a lot of dwell time at bus stops, and are a big reason why transit trips from one end of downtown to the other end can take so ridiculously long. This, in turn, causes trips of longer distance (where walking is infeasible) to be glacially slow.
Short distance also tend to be not great for people taking them. As Jarrett indicated, the shorter the trip, the greater the percentage of total travel time is spent waiting, and, at some point, frequency has to improve beyond what the agency can afford, just to get the bus to be faster than walking.
When running a transit agency, it is best to focus on getting people out of their cars, not out of their feet, since it’s cars, not walkers, that cause congestion and pollute the air. Any trip in a downtown-like environment that is short enough to walk in a span of a few minutes is probably already faster walking than driving, simply because of the overhead of getting a car out of one parking garage and into another. So, we don’t need transit to compete with the car for those sorts of the trips. It is longer trips, where walking is less feasible, where transit needs to be competitive with cars. And, in order for transit to be competitive with cars, it needs to move, and not constantly stop.
In terms of getting people out of there cars, I think short distance trips are just as important as long distance. I know a lot of people who *only* take transit to work, for example. They wouldn’t dare take transit to the grocery store, or a restaurant or to visit a friend. For that they use a car (of course). They don’t rent a car, either, but own one. And if you own one, you are way more likely to use it.
A big reason for that is that like Portland, the system doesn’t work well for that (as Jarrett said). So what would a system geared towards the needs of short distant riders look like? I think the key elements are:
1) Short stop spacing.
2) Very frequent service.
3) Off board payment.
Of course I’ve just described the New York Subway system, which, by the way, has the highest number of people without cars. Yes, of course there are other reasons for that, and of course people walk a lot in that town. But the two go together. If a transit agency manages to solve the short distance trip problem, they probably have solved the long distance one at the same time. Or, like New York, they really haven’t solved it, but people muddle along anyway.
For very short distances, travel time is determined mostly by time stopped, not cruising speed. So, being able to travel faster than walking speed matters little, but not having to wait for a bus to show up, or sit there while people get on or off a bus matters a lot. That is why for distances under half a mile, walking is nearly always faster than riding a bus.
Similarly, when the distances start to increase a bit (maybe ~1/1.5 miles), bicycles and scooters (https://www.theverge.com/2018/6/8/17441904/uber-lyft-jump-electric-scooters-san-francisco) tend to get you where you’re going much more quickly than buses and trains. Only when distances increase further, to the point where you really need does traditional public transit start to become faster (and, even then, there are plenty of door-to-door trips where riding a bicycle for 5-10 miles is *still* faster than riding the bus, usually cases where the bus involves transfers, indirect routing, or horrendous traffic).
The NYC subway doesn’t stop every single block like most local buses in most of the US. My local route in Minneapolis, for example, stops approximately every 550 feet and during rush hour it makes every single stop. That’s pretty average. Getting rid of every stop and having 3-4 people using each instead of 1-2 would make the ride way more reliable and make it seem faster- which also matters. Even with a limited stop route, there’s no reason to stop a local bus every 500-600 feet.
Ross, I’m not sure I agree with you that NYC subway stop spacing is “short”, unless you’re comparing it to Sound Transit’s light rail, parts of Portland’s MAX system, or commuter bus or rail service anywhere else in the country. NYC subway lines pretty consistently have stops 1/3 to 1/2 miles apart (in Manhattan and Brooklyn) from one another, which I don’t think can be considered short when compared to how frequently most local buses stop. Compare this to King County Metro Routes 7 or 8, whose stop spacing is almost never wider than 1/4 mile and usually closer to 1/5 or 1/6 of a mile.
Given that, I do agree that NYC subway stop spacing is pretty optimal for short trips, and even long trips given the subway is insulated from (auto) congestion and has express options.
Subways (or grade-separated rail-based transit, in general), traditionally has relatively wide stop spacing, not so much because of a deliberate tradeoff of speed vs. accessibility, but because the construction cost of building all those extra stations to locate the stops closer together would be prohibitively expensive. This, by accident, leads to faster speeds, leaving the truly local transit service, to buses.
That’s not to say there aren’t exceptions. By subway standards, NYC has very close stop spacings, at least in Manhattan. But, when every stop is an underground station costing hundreds of millions of dollars, rather than a pole in ground supporting a “bus stop” sign, the cost of frequent stop spacing adds up very fast.
Downtown areas benefit from short-distance circulation, even if purists think they “should” walk, and trains’ high capacity and high-volume boarding can absorb crowds in stride, including ballgames and Fourth of July and street-closing protests. In some cases a regular subway or ultra-frequent (5-10 minute) bus line can do double duty, both bringing people from the neighborhoods and providing downtown circulation. Portland went this way with a MAX free-ride area downtown, encouraging people to take the train to speed up the regular buses. I don’t think it’s free anymore but the idea of a double-duty train or BRT is still interesting. Another approach is a distinct downtown circulator, although I’m more skeptical of these because they don’t address trips just beyond their ends. Of course, relief runs (short-distance runs on part of a regular route) function the same as circulators, and don’t have the dilemma of dedicated circulators (e.g., a short-distance streetcar): if you’re going beyond the end of the circulator, take the regular route, and you’re not missing out on the fancy circulator ride.
Outside downtown, short-distance trips are especially important for low-income people without cars and the elderly, to get to the local supermarket and other everyday errands. Then there are unlimited passholders who take it because it’s free and why not. Both of these are worthwhile transit markets.
I can’t see short-distance trips alone as enticing able-bodied people to take transit, but if they start commuting on transit, then the availability of short-distance transit could make the difference between using it just for work and using it more generally (for both work, short-distance trips, and medium-distance errands).
As a daily local bus rider in Alexandria, Virginia, I see both sides of it. The Metrobus really is only slightly faster than walking, partly because it’s circuitous and partly because of so many stops. I could travel by bus all the way to work, but it requires taking two buses, so I do this only when it’s raining. It’s faster to ride only one bus and walk the remaining mile rather than wait for and ride another bus. And some of the stops do seem awfully close together. Some are less than 300 feet apart, but the bus doesn’t stop at both unless someone is waiting there or wants to get off. Sometimes, though, it does keep stopping every few hundred feet. I could see the benefit of closing some of those stops.
My take is that when closing stops on a bus route, the transit agency needs to close a bunch at the same time. If it closes just one or two, the travel time savings will be too small to outweigh the greater walking time to the route. But if a number of stops are closed at once, the travel time savings will outweigh the additional walking time. There are tools to speed up bus routes, it would be preferable if the agency used multiple strategies at once. And cities need to be helpful partners to transit agencies if they really want to improve travel time and transit service.
The short stop spacing totally torpedoes trip efficiency and speed, unfortunately. When dwell time makes up to 30-40% of a bus’s travel time, there’s really only so much off-board payment can do when buses still need to stop every block. Getting people to congregate and consolidate stops (within reason) makes the trips faster for everyone.
Now, if a downtown circulator makes stops every block, well, no one was using that bus anyway.
How short is short? It’s different in different contexts. Generally, stop spacing every 0.25-0.5 miles is the sweet spot for local service, and what you’d find in Germany. If it’s closer than that, then the time to stop, dwell, and go again becomes excessive. But if you increase local spacing to half a mile, then you start leaving out people. It’s not only people directly between the stops, who have a maximum 5-10 minute walk (depending on their walking speed and stoplights), but also people on the side edges of the diamond (if you draw a diamond from the two stops out to the sides), who face 10-20 minute walks. If people consider a five-minute walk “great”, a ten-minute walk “OK”, and a twenty-minute walk “only atheletes”, then you’re leaving people at the sides fo the diamond with poor service. They may think it’s worth it anyway if it’s an enhanced route (transit lanes, full-time frequent, etc), but not if it’s a typical local route.
Sometimes, interactions between a bus stop and ordinary traffic control can cause a single stop to have outsize impact. It is quite common to see signals timed in such a way that stopping to serve a bus stop – even if the bus only has its doors open for a few seconds – causes the bus to miss the upcoming light and wait multiple minutes.
The major issue with bus stops is that the mere issue of stopping the bus has an inordinate impact on both speed and reliability. It can take between 20 and 30 seconds for a bus to speed up and slow down to make the stop. By comparison, dwell time is relatively quick, 1.5 – 3 seconds per boarding (assuming alightings all exit the back door). 1.5 seconds is for an account-based tap system like Portland or Chicago, 3.0 seconds is the average for cash. The big issue with stops is variability of stopping and usage. I pulled the APC information for Portland’s bus stops. It has 6,570 stops. 396 stops had no activity (boarding + alighting) at all in Fall 2017. 397 had 1 activity per day–all day–over the course of 15 to 24 hours a day. In fact, 49.1% (3,229) of the stops were responsible for only 5% of the ridership activity. This random act of riding adds untold additional time in stopping. And since ridership such as this tends to be stochastic not deterministic, it can’t be shown in the schedule. It only takes a few of these outlier stops to contribute as much to delay as traffic.Consolidating stops gives much better performance to the vehicles and the great majority of the riders of those vehicles and negatively affects only a very small proportion of those riders.
“Consolidating stops gives much better performance to the vehicles and the great majority of the riders of those vehicles and negatively affects only a very small proportion of those riders.”
How is that possible to say this? Consolidating stops will mean that the majority of people who walk to the stops will have to walk further. *Maybe* the travel time savings will offset this, but there’s no guarantee.
For people concerned that “it’s the stopping time, not the boarding”, there are removed stops on the TTC where the streetcar often has to stop for a red light anyway. But you can’t get off (even though that street may be your destination). The TTC also removed a number of stops at Bloor Street for buses or streetcars coming to the subway stations on the Bloor line. The theory is that you will get off at the station and walk back out. But then if there are delays or bunching, you might be sitting on the streetcar for ten minutes while the loop at the station clears of the vehicles ahead of you. Again, you can’t get off anymore.
Finally, for the people who say “transit has to be fast to get people out of their cars”. The removal of the stop closest to my house has the opposite effect. When I consider that my minimum walk to the “efficiently spaced” stop is not much less than my time to actually drive right to a local destination, this is not going to get me to use transit instead of driving. As someone else pointed out, wide stop spacing is okay if you think of transit as the way to go to work downtown and that’s all that you use transit for.
In Toronto, some of the worst traffic isn’t for employment, it’s people driving their kids to school and picking them up after school. Transit can only be an alternative if it provides fine-grained local service.
As for rideshare bicycles or scooters or whatever, again, the problem is how fine-grained the depots are going to be. How far you have to walk does not affect just transit, it affects bike sharing schemes as well.
The idea is that the stops that are consolidated away are not chosen randomly. They are picked deliberately to be stops that don’t get used often, or impose an unusually high time penalty on everybody else, in order to serve. Yes, there will always be a few people that will give up the bus because the stop is removed in front of their house. But, for every one of them, there will be many more, who enjoy faster service because of the removed stops.
“Transit can only be an alternative if it provides fine-grained local service.”
Once you get fine-grained enough, people’s trips start to diverge to the point where fine-grained local service is impossible without either making the trips too slow (due to the time to serve everybody else’s front door), or too expensive to operate (like a taxi). If transit operates as slow as walking, it doesn’t matter how close it goes – it’s going to be unattractive. Whereas, if the stops are further apart, allowing it to move faster, at least some people whose origin and destination points are near the stops that exist, will find it fast enough to be worth using. And, if the stop locations are chosen carefully, that “some” ends up being a large majority of the people who live and work along the time.
“As for rideshare bicycles or scooters or whatever, again, the problem is how fine-grained the depots are going to be.” Absolutely true. And, it’s a huge reason why dockless bikeshare has been so successful. Remove the stations, and allow people to pick up and return bikes anywhere, you get the advantages of fine-grained local service, but without the drawbacks of having transit provide it. In particular, when you’re on a bike, you can ride straight to your destination, unencumbered by where others around you need to go. Granted, you’re limited by how fast your can pedal, but for short enough distances, cruising speed doesn’t matter so much as the consistency of always moving, and not stopping.