Last week, David Levinson asked whether we are “building a city” in the Minneapolis-St. Paul metropolitan area. He noted that rail investments only make sense in areas where there is sufficient density to support those investments. He further wondered whether Minneapolis and St. Paul could achieve those requisite densities for a majority of residents in core areas in 20 or 30 years.
While we may only be dabbling with building a city at the center of our region, Minneapolis already has densities sufficient to support a respectable transit system (St. Paul does not). This is encouraging because it means that smart transit investments in the city have a high likelihood of success.
But it is also discouraging that much of Minneapolis is transit-ready because it means the city should have better transit than it does. Perhaps the city’s poor transit system is the result of our decision makers’ priorities and experiences. As has been pointed out here and elsewhere, Met Council board members – who exercise control over Metro Transit – rarely “incorporate public transit into their daily lives — if they use it at all” because they are “often unable to include the additional time needed for transit in [their] schedule,” or they “can’t take it for work because [they] have to go to a lot of meetings.”
Maybe we can provide inspiration for smart local transit investments. After all, countless busy lawyers and financiers in New York and politicos in D.C. take public transit every day (as do a few busy people I know in the Minneapolis legal community). If they make it work, we can all make it work.
HIGH QUALITY TRANSIT IS A VIABLE OPTION IN MINNEAPOLIS TODAY
Let’s start with the basics. Roughly 40% of Minneapolis’s residents live in neighborhoods dense enough to support fixed rail transit. The adjacent map shows those Minneapolis neighborhoods with residential density greater than Montreal (11,701 people per square mile) in green, greater than D.C. (9316 people per square mile) in blue, and employment density greater than 9600 per square mile in red. I chose these cities as comparison cities because they both have significant transit systems.
Nearly all these dense Minneapolis neighborhoods are contiguous. The block is about 6.5 miles long (from Holland to Kingfield) and about 3.25 miles wide (from East Isles through Phillips). Because all the neighborhoods are contiguous, Minneapolis could build a mini world class transit system entirely within the already-dense neighborhoods and then gradually lengthen lines as peripheral neighborhoods became dense enough to support transit. This mini transit system might look something like this:
The map above is part of Metro Transit’s existing Hi Frequency bus map. These Hi Frequency buses have reasonably short headways (wait time between vehicles) but can be excruciatingly slow. Yet the Hi Frequency map shows that Metro Transit knows where good transit is needed.
What can we do to improve speed in this dense Minneapolis core? Some might point to the city’s approved streetcar network, which covers much of the same territory as the Hi Frequency system:
But the city and Metro Transit have indicated that our streetcars will operate within personal automobile traffic (the Midtown Corridor streetcar will be an exception if it is built). Therefore, any discussion about how to improve streetcar timing applies to our current bus network and vice versa.
So why not push to maximize our current system, even through pilots, before laying track? We could learn a lot, and improve a lot of peoples’ everyday lives, by experimenting with speedy bus corridors.
EXAMPLE LINE: THE #18 BUS
The #18 bus, which primarily runs along Nicollet Avenue South and Nicollet Mall in Minneapolis, provides a nice case study for reliable yet inefficient service and for opportunities to fix these inefficiencies. Much of the #18’s route is part of Minneapolis’s first planned streetcar line.
According to Metro Transit’s own information about the #18 bus, it should take anywhere from 27 to 47 minutes to get from 38th Street to 4th Street if the bus is running on time (details in footnote [1]). According to Google Maps, it takes 13 minutes to drive from 38th and Nicollet to 4th and Nicollet on city streets.
WHY THE #18 IS SO SLOW
The #18’s timing woes are the result of many factors. Most can be classified into three general categories: (1) necessary delay, (2) delay that occurs due to the transit system’s priorities, and (3) delay that results due to prioritization of other modes of transportation (some causes of delay, such as requiring that riders pay fares on the bus instead of at the stop, do not neatly fall into one of these categories).
First, buses necessarily travel slower than other vehicles because they must take time to slow down as they approach a stop, must stay stationary to pick up passengers, and must take time to get back up to speed. This delay can be mitigated if buses are given their own lane.
Second, Metro Transit has determined that the #18 should stop nearly every block for much of the route. The agency believes this will serve more users and more types of users.
Finally, other factors explicitly prioritize cars over buses. Those factors with the most impact include requiring buses to lose time pulling in and out of traffic at stops and timing lights for cars instead of giving buses signal priority or preemption.
On routes where personal automobile traffic counts are magnitudes higher than bus ridership, these last factors may be logical from an efficiency standpoint, although the traffic counts and ridership numbers are influenced by the relative efficiencies of the respective systems. Commute figures on Nicollet, however, are not magnitudes apart.
Vehicle counts on Nicollet Avenue between 25th and 26th Streets are 9106 per day. Assuming 78% of vehicles are single occupancy (Minnesota’s average) and assuming 2.5 people in the remaining 22% of vehicles (this number is a guess), total car ridership at that point on Nicollet is about 11,500 per day.
Daily bus ridership on the full #18 route was 10,949 in 2010. It appears reasonable to assume that well more than half the #18’s rides traverse part of the route between 38th and 4th – anecdotally the bus is most crowded between Lake and 7th Street in Downtown, and the boarding map on page 93 of this Metro Transit technical memorandum shows the route’s heaviest boarding between downtown and Lake Street.
These numbers do not provide an apples-to-apples comparison because they contrast route-wide boarding data with traffic counts at a specific place. However, the figures provide enough information to show that the number of people traveling by car along Nicollet do not vastly outnumber bus patrons along the route. Notably, car passenger and bus ridership numbers are similar magnitudes even though the #18 is anywhere from 2 to 4 times slower than a private car on Nicollet.

$512 of ungathered city revenue on Nicollet Mall. This photo taken after I made my way through the intersection of stand-still vehicles.
Car travel from one point to another on Nicollet is more than twice as fast as the bus in part because those factors that explicitly prioritize cars over buses – placing buses in car traffic, making buses pull over to stops, and timing traffic lights for car speeds – are all in place for much of the route. Yet even on Nicollet Mall in downtown, where the city prohibits private automobile use, signal priority remains non-existent. In addition, private automobiles on cross streets are constantly blocking the intersection when buses have a green light. Although this “blocking the box” is a violation under state law that earns box blockers a $128 ticket, I have never seen someone ticketed for this offense in downtown.
These street design elements, traffic signal timing, and lack of enforcement reveal city, regional, and cultural priorities in favor of the automobile.
RESHUFFLING PRIORITIES COULD DECREASE TRAVEL TIME BY 40%
Metro Transit takes a lot of flak for inefficiencies and deficiencies in the system. Some of it is deserved, and the agency is gradually addressing some rider concerns. However, much of the responsibility for our bus system’s inefficiencies lies with municipal governments.
Cities, not Metro Transit, are usually responsible for traffic signals. Although there has been a lot of attention paid to green line signal priority and route speed on streets.mn and elsewhere, few discuss that signal priority and preemption could just as easily be used for our existing bus systems.
In addition, Twin Cities municipal and county governments are generally responsible for street design. While citizens increasingly hold cities and counties accountable for lack of dedicated bike lanes, we rarely hear outcry about lack of dedicated bus lanes.
So what would happen if the City of Minneapolis prioritized the #18 bus within the Minneapolis density zone by giving it dedicated lanes and full signal preemption?
Signal prioritization decreased the Green Line’s travel time by 28% and, according to Mike Hicks and Metro Transit’s earlier travel time estimate, preemption could have decreased travel time by about 40% (I am using initial end to end travel times of 67 minutes, current scheduled travel times of 48 minutes, and initial estimated travel time of 40 minutes).
Applying these travel time gains to the hypothetical 38th Street to 4th Street #18 trip, trip time could decrease from 27-47 minutes to 19-36 minutes (28% decrease) or 16-32 minutes (40% decrease) (calculations in footnote [2]).
I am not an engineer and cannot calculate how bus stop frequency vs train station frequency would affect efficiency gains, nor can I calculate how much time would be gained from giving the #18 dedicated lanes where it does not need to pull in and out of personal automobile traffic. Nevertheless, it appears safe to assume that signal prioritization or preemption would dramatically decrease travel time on the #18.
(Update 4/6/15: This section was amended to correct increases in travel times with decreases in travel times.)
A PROPOSED PLAN
Here is a quick and easy way for Minneapolis (and perhaps eventually St. Paul and surrounding
on-the-grid suburbs) to become a more functional urban environment: take all segments of all Hi Frequency routes that run through neighborhoods that achieve a transit-supporting density and give those routes signal preemption and dedicated lanes (where feasible), even it it means removing a travel lane and/or parking lane for cars on the route.
This would reward the 162,000 Minneapolis residents who live in urban densities with urban facilities and might jolt other Minneapolis residents into realizing the benefits of density in their back yard. It would also make life instantly better for those dependent on transit and would make it possible for people with busy schedules to feel that they can take transit at least some of the time.
While explicitly rewarding density with transit investments may seem like a high political hurdle, it is not without precedent: Peter Albert, at SFMTA, gave a talk at the Pro Walk Pro Bike conference on doing just this in the Bay Area.
WHY WE NEED EFFICIENT TRANSIT
Public transit has the potential to be a positive and defining feature of a city and region. However, it has to be pleasant and efficient. Yes, trains are wonderful and I am glad that the Twin Cities is investing in rail. However, we cannot wait the decades it will take to build out our rail network to better our transportation system, and we should not settle for continued slow transit.
Minneapolis is a great place to pilot some better bus lines in anticipation of streetcars:
- The core city has the density to support a serious transit system and already has established Hi Frequency routes that serve that dense area.
- Minneapolis Mayor Hodges wishes to focus on her equity platform. More efficient transit would enable those who cannot afford a car or who cannot drive due to disability to get to work nearly as efficiently as their counterparts who can or choose to pay roughly $9000 per year in personal automobile expenses. This will lead to greater job access for a greater percentage of Minneapolis’s population.
- Increasing bus speed will likely increase transit use (induced demand applies to transit as it does to roads). If the city can move people from cars to transit, it will further the goals in the city’s Climate Action Plan.
- Improving our transit system will help the city attract people who do not want to have to own a car and will, in turn, increase the city’s economic competitiveness.
Improving transit in Minneapolis’s core would not solve all our problems. However, improving transit in the city’s urban core would make many people’s lives better and would incentivize further urbanization of the city. It would take some coordination between the city and Metro Transit, but it would be worth the effort.
FOOTNOTES
[1]
Currently, according to Metro Transit’s own schedule, it takes 27 minutes to take the #18 bus from the commercial node at 38th St. South to the Library Central Branch on 4th St. downtown. This accounts for no delays (the #18’s on-time percentage is about 90, with “on-time” defined as “departing between one minute early and five minutes late.”).
Metro Transit’s #18 schedule also does not account for time spent waiting for the bus (wait time ranges from 5 to 15 minutes from 6:00am to 7:00pm). Some (mainly non transit users) argue that this wait time should not count toward the trip time because they believe riders should acquaint themselves with the schedules. However, I count this time because life (and meetings, for the busy among us) is not always predictable – an important side conversation can delay exit from a home or office and force the rider to miss his or her bus. Meanwhile the same conversation delays a car driver exactly the duration of the side conversation.
Trip + wait time + 5 minute “on-time” grace period means that taking the #18 from 38th St. to 4th St. should take somewhere between 27 minutes and 47 minutes. I did not count time spent walking to the station or bus stop toward the trip time on the theory that both transit users and car drivers must walk to their vehicle (although instances where transit is closer than a parking facility are few).
[2]
A travel time decrease from 67 to 40 minutes is a 40% decrease in trip time. A 40% decrease from 27 minutess would result in a 16 or 17 minute trip. In addition, if buses were 40% faster MetroTransit could run 40% more runs if it ran the same number of buses and drivers.
Increased trips per hour could reduce maximum wait time from 15 to 10.7 minutes (4 buses an hour at 15 minute headways would increase to 5.6 buses an hour). As a side note, headways would start to look really nice at rush hour: Currently between 7:00 and 8:00 am, there are 8 #18s that travel past Franklin. This averages out to 7.5 minute headways. If there were 40% more runs in this same time period, headways would drop to 5.4 minutes.
Using the same logic used in footnote [1] to calculate maximum trip time (trip + maximum wait time + 5 minute “on-time” grace period), the reduced headways and increased trip time would result in a new maximum trip time of trip time of 32 minutes. This is down from existing 47 minute maximum trip time without signal priority.
I followed the same logic to calculate trip times for a trip time reduction from 67 to 48 minutes (28% faster). This would knock 7.7 minutes off our hypothetical trip (to 19.3 minutes). Maximum wait time would drop to 11.8 minutes (rush hour to 5.9 minutes).
Love it. Minne transit system. Build it where we want density
Love it as well. I’d like to add that inline Orange Line BRT stations at 38th Street and maybe Franklin Ave should be built as 35W is rebuilt. Apparently we’ll spend hundreds of millions of dollars adding lanes and ramps for cars within our densest neighborhoods, but we won’t make common sense transit investments? What do you expect from MnDOT, I guess…
Having taken the 18 bus quite regularly for years between Richfield and downtown, I don’t know that I’ve ever once encountered congestion while riding it. Tedious signals, certainly; stops every block, sure; slow bus passengers paying cash, definitely. But never have I actually seen a lot of cars getting in its way. My point (if my experiences are at all representative) is that on that particular line, completely separate right-of-way may not help all that much — and would do big harm by taking away space from parking, bike lanes, or green space. The more compelling change would be signal prioritization, removing bus stops, and off-board fare collection. (All things that would be accomplished at least somewhat by Nicollet BRT, if it weren’t delayed indefinitely for streetcar study.)
Apparently this can be quantified. According to Metro Transit, the 18 spends 3% of its time stuck in traffic congestion — less than 1/10 the time it spends getting passengers on and off the bus.
Surprisingly, this is pretty consistent throughout the studied lines for the aBRT study. Even Lake Street, which I expected to be fairly congested, causes minimal delay to the bus. However, the Lake St line spends 42% of its time on boarding and disembarking. That’s insane!
As a frequent rider of the 21/53 along Lake, I can verify with my anecdotal experience that the vast majority of the time the bus is picking up and dropping off people, but I’m guessing they’re including the time it takes to get back into traffic after the stop in this time. I have no idea why most of the bus stops are not on curb-outs. Pulling in and out of traffic is terrible on the 21 because it stops at *every* *single* *stop* pretty much all day long. The 53 is a bit better because of the limited stops, but it often gets stuck behind a slow-moving 21 and, if the traffic is bad, it can’t pass the 21 for a while. If lake street has to stay a 4-lane nightmare, they can at least allow the buses to block the side lanes for stops.
I’ve honestly had plenty of days when the 21 is supposed to be on a ~6min frequency and I can walk over a mile down lake street before a bus shows up going the same way. Then 2-3 tend to show up, and the empty one speeds by, leaving the full bus to cram more passengers on.
Oh, and because it goes through poorer areas, way more people tend to pay cash on the 21 than most other routes I’ve been on, further slowing down the boarding process. If ever there was a case for off-board fare payment, this is it. But I’m sure it’ll be last in line for actual improvements because in fantasy land there’s a midtown streetcar at some indeterminate point far in the future, so we can’t possibly spend money on an existing problem now. Also they have a terrible record of improving transit in areas with a lot of poverty.
“Also they have a terrible record of improving transit in areas with a lot of poverty.”
Sad but true — and completely demented, from a broader public-policy standpoint. Frequent and reliable transit serving poorer areas would expand what one might call the “jobshed” for area residents, allowing them access to a broader range of steady employment options that may assist them in improving their economic situations.
These broader policy issues really do link to one another, but it seems as though much of the country has lost sight of that fact in recent decades. Interestingly, this atomization of policy questions could be said to mirror suburbia’s atomization of varying land uses, social groups, and even individual neighbors…
In ancient times (the 19th century) employers actually built mass transit to bring their poorly paid workers to the factory. They built the housing for the poor — their workers — too.
I think their definition of stuck in traffic must be different than mine. When the bus is stuck behind multiple cars moving slowly and misses a light (or 2), I call that congestion. When cars don’t let the bus re-enter the stream because they’re selfish (moving fast and not yielding) or stuck in traffic, that’s congestion (not dwell time). When cars don’t get to go through an intersection because the box is blocked, that’s congestion. The sheer number of cars on the road requiring lights with relatively long cycles itself is a result of mode share and area congestion. My experience on the 6 is that these all happen very regularly yet the Hennepin study doesn’t give a single percent to traffic. Though I’d probably agree that the 18 sees a lot less of this than the 4, 6, 21, 5, or other routes on fairly heavy arterials.
I’m 100% with you that we should evaluate as many options as possible that speed buses up and are also politically achievable. Stop spacings, far-side stops paired with signal preemption, bulb-outs that negate pull-in/out penalties all make sense on arterials where dedicated lanes really isn’t an option (short-term). But long-term, it sure seems like at least stretches of some major streets should turn lanes over to buses. Particularly areas approaching major intersections and freeways – they seem to be big bottlenecks given light timing/etc.
Unfortunately because almost every tiny little give-away to improve bus service involves a minor subtraction of space/preference for personal cars, there is zero interest in doing it. Imagine the uproar in the star tribune comments section if they dared to suggest giving any preference to buses or making people *gasp* wait behind them while they stop?! It would be matched only by the fervor about how many stop signs people on bicycles run when there’s an article about bike lanes being added somewhere.
Snark aside, I really really wish there was some political will to actually do this. Minneapolis is doing a good job adding density to its core, but it’s going to choke itself relatively soon if it doesn’t do something to actually address transit needs in the city proper. But hey, we have to worry about what the far-flung suburban counties might think since they have equal/greater say in planning our transit.
And far-side bus stops. I have no idea why this change wasn’t made when Nicollet was completely redone south of Lake over the past few years.
CORRECTION: A coworked just noted that the post states signal priority would increase trip time. It would decrease trip time.
And agreed, signal priority would make a much greater difference than dedicated lanes.
Great post, and as a regular rider of the #5 bus, I can say that I would wholeheartedly agree with your recommendations for improving that service.
One quibble, though – you omit Ventura Village from your density map, but I have a feeling that is just an oversight. Ventura Village is part of the broader Phillips community (I noticed that your source doesn’t break out the Phillips neighborhoods and just has all 4 together), and is absolutely above your density threshold, so for future iterations, I think you should make sure to include that with the green shading.
Got it — you are right. I had some of the same issues sorting out the density of the Powderhorn neighborhood and its constituent parts.
Pingback: Sunday Summary – November 23, 2014 | streets.mn