Two pieces of context re: bus ridership declines

The news from the CTA last week was that a longtime trend of convergence between bus and rail ridership is, if anything, picking up speed:

It’s hard to overstate how big a transformation of Chicago’s transportation landscape this is. As recently as the 1990s, there were well over two bus riders for every rail rider; the city’s transit was as much about a crosstown streetcar-era bus grid as it was about a downtown-focused heavy rail network.

But rail ridership has nearly doubled in the last 20 years, and bus ridership – which stabilized in the early 1990s after a long period of precipitous decline – has begun to fall again in the last few years.

The reasons for that aren’t entirely clear. Ridership began a notable dip right after the 2008 economic crash, which makes sense – fewer employed people means fewer commutes. But the more recent decline, which seems to have started in 2012, is harder to explain. Precisely because it’s hard to explain, I suspect, many of the theories offered are quite vague: many of them are of the “people just don’t like buses” variety.

"Revenue miles" is just a measure of service. If there's one bus route that's 10 miles long, and it's served by one bus that travels from the beginning to the end and then back again, that's 20 revenue miles.
“Revenue miles” is just a measure of service. If there’s one bus route that’s 10 miles long, and it’s served by one bus that travels from the beginning to the end and then back again, that’s 20 revenue miles. Source: National Transit Database

But I think any discussion of bus ridership in Chicago needs to include this chart, and take two things away from it.

1. First of all, declining bus ridership is not actually a “long-term” trend, though it’s often framed that way. (Or, to be more specific: decline is typical of the last 50 years, but not the last 10 or 20.) In fact, as recently as the mid-2000s, ridership was growing. And other than the deep recession years of 2009-2010, 2013-2014 represents the first multiyear ridership decline since the mid-1990s. This isn’t meant to wave the problem away: it actually makes it worse, since it suggests that far from experiencing a long, slow decline driven by structural factors, something specific has changed recently that’s made buses less attractive.

2. Secondly, service matters. I think it is probably not a coincidence that ridership growth in the 2000s came at a time when the CTA was adding service: reducing wait times between buses, expanding their hours, and introducing express routes. (Between 2002 and 2006, the CTA created ten “X” routes, which mostly followed existing bus lines, but stopped every half mile instead of every eighth. Almost all of them were discontinued in 2010 because of a budget shortfall.)

I think it is also probably not a coincidence that the CTA has had a difficult time recouping its bus ridership losses from the recession, given that its dramatic recession-era service cuts have mostly remained in place.

To be clear, this isn’t at all a slam on the CTA, which can’t raise significant revenue without raising fares. Moreover, it has launched some bus service improvements in the last few years – the Jeffery Jump, the “Loop Link” busway, and the proposed Ashland BRT – that make a good template for expansions into the rest of the system.

In constant 2013 dollars. Source: National Transit Database
In constant 2013 dollars. Source: National Transit Database

But the money for more broad-based, bread-and-butter improvements to service just hasn’t been there. And unlike rail customers – who are generally downtown commuters, and have to think about paying downtown parking rates if they want to ditch public transit – bus riders who already have a car face no such massive disincentive to using it. To be competitive, buses need to run frequently and reliably, and make decent time along their routes. They are absolutely capable of doing that, given relatively modest investments in operations funds, technology, and space. But we’re not making nearly enough of those investments.

37 thoughts on “Two pieces of context re: bus ridership declines

  1. The previous rounds of service cuts weighed heavily on bus vs. rail. But this convergence may show what I feared might happen after those cuts, namely the CTA is becoming a de facto commuter rail agency, with the primary focus on taking people to and from work in the Loop less so than as general transportation. I’m not sure what your on the ground perspective is of this. What is the breakdown on peak vs. off peak ridership these days?

    1. I’m not sure. The Red Line on the North Side seems to almost always be pretty well-used, but other lines, not so much. You’re right that peak/off-peak would be really, really good numbers to have, though.

  2. Since 2009, there are daily failures of the bus network because the CTA is stretching every last efficienciency it can out of the operating budget. The proposed state budget cuts to operations funding are scary, and based on the false idea that the CTA can find further efficiencies on its own.
    There are efficiencies to be found for transit in the region, but they depend on changing state and local policy around the prioritazation of transit on our streets and highways. No such policy changes, such as Lakeshore Drive bus lanes, are forthcoming from the state.

    1. Obviously Uner and Lyft, began in 2012, has by far the greatest impact on buses. Stunning this article completely missed it. Ride-sharing will grow in popularity, buses will continue to decline. No investment in bus system should be made.

      1. No, that’s definitely not the case. I don’t know offhand how many rides Uber and Lyft give, but I have seen maps of their service, and the places that they are strongest also happen to be the neighborhoods where bus ridership has fallen less sharply than in the city overall.

      2. LOL! Ain’t nobody taking Uber to work at rush hour, at least not in numbers that would even show up in these stats. Do you honestly think working people will regularly prefer spending $20 on an Uber over a $2.25 bus ride?

  3. One counterpoint: there’s a medium-term trend of L ridership increase in conjunction with bus stagnation. So something must have induced more Loop transit commutes. Is the Loop gaining jobs?

    1. Yes. The Loop is gaining jobs, and the Red/Brown/Blue line corridors, in particular, have been gaining Loop workers. On top of that, the larger downtown/north lakefront/dense northwest corridor area has been gaining retail, restaurant, and cultural destinations at an even faster clip. I would guess a large percentage of people get to those places by train, because parking is scarce and expensive, and frequently they involve drinking.

  4. I would have to think that Divvy has taken up more bus trips than rail trips as well. Not sure if that would account for much in the grand scheme of things, however.

    1. Yeah, even if every Divvy ride displaced a bus ride – which is definitely not the case – it wouldn’t come close to accounting for the drop over the last two years.

  5. What portion of the decline in bus revenue miles is attributable to fewer buses versus changing ridership patterns?

    1. That’s really hard to say. What *doesn’t* seem to be the case is that people are literally switching from buses to trains – the areas of the city where L ridership is increasing the fastest are *not* the areas where bus ridership is falling the fastest. What’s confusing, again, is that there are any number of reasons why you’d expect a long-term decline in bus ridership, at least compared to rail – the geography of jobs is changing, population is down away from major L lines, etc. But this most recent decline really only began in 2012, which makes it hard to use decades-long trends to explain it.

      1. So recently someone was trying out a privately run bus from Lakeview to the Loop. This is essentially the route of the old Addison Limited of decades ago but withnicer seats and wifi. I would believe they were not cannibalising CTA local bus service so much as the sardined out Red/Brown/Purple Lines.

      2. I guess the thing I’m struggling with is why Revenue Miles jumped, then declined so drastically, as ridership stayed relatively constant. As I think about it, rider mix shouldn’t really matter, because it seems like Revenue Miles is based on routing, not on ridership, strictly speaking. It seems to me that A) the number of buses went up, then down, B) the route mix shifted toward, then drastically away from, routes with faster average speeds, C) overall average route speeds went way up then way down, or D) some combination of these. I get that it’s hard to disentangle, but the change is so drastic…

      3. I’m not quite following – where does the speed part come in? Vehicle revenue miles shouldn’t be affected by average speeds – it just basically multiplies the number of scheduled runs by route lengths. If a bus finishes its route extra quick, that shouldn’t change the VRM unless they decide to send it for an extra run, which I don’t think really happens.

        I think one potential explanation is that ridership is relatively sticky. If you take the bus, and your route’s frequency drops from every 5 minutes to every 10 minutes, you’re probably not going to stop riding immediately. But obviously the bus has become notably less convenient, and so over time, you’ll be more likely to choose some other form of transportation. Meaning, I think, that service cuts could have ridership effects for several years in the future.

      4. I think I was trying to look at this in the context of shifting route mix, but the answer you proposed sounds a lot simpler: “CTA ran the buses more hours/runs, then cut way back after 08, but ridership didn’t really change because it takes people a while to change their transpo habits in either direction.”

        FWIW, my thinking with the speed thing (which is not based on any actual knowledge of how bus scheduling works beyond getting yelled at for trying to board the 6 while it was waiting to start its run on Wacker) was that, if we are comparing two routes, each served by one bus, one of which produces fewer daily revenue miles, then that route must have a lower average speed. Either the bus idles longer waiting for the next run to start, or its in traffic a lot and thus can’t do as many runs. … I was probably overcomplicating things…

  6. Sounds like an all of the above issue. SInce no one reason jumps out likely several reasons are converging to drive the numbers. Divvy/Service cuts/demographic changes/increased auto use on some routes, whatever.

    And whatever the reason, the problem is that the bus network is hugely important in maintaining a healthy grid based transit network. And that’s why BRT is now so important in maintaining that grid based network by in filling the missing rapid transit gaps. Pretty much every 4 lane arterial should become BRT.

  7. I will just point out that the #66 Chicago bus saw an 9% drop in ridership despite the CTA decrowding program adding rush hour buses to this route. Reliability may be more important than increased service, seeing as these drops occurred on routes that added service & routes that lost service. We can look to Saturday & Sunday service to get an idea of how off peak has been impacted, but from a cursory look it seems to be rising & falling similarly to weekday ridership.

    1. Yeah – I definitely don’t want this to be taken as saying that service cuts are the only, or even necessarily the main, issue in ridership patterns. On the 66 in particular, I would think that speed and reliability are bigger issues than frequency or span – the 66 comes so often that waiting 5 minutes versus 4, or whatever, is unlikely to cause anyone not to take it. It’s more likely that its incredibly slowness is a problem. Of course, that’s also a service issue: bus lanes, stop consolidation, off-board fare payment, and transit signal priority could make a major dent in travel time.

  8. Could it be that the bus decline and rail increase is mostly a result of changing demographics and geographic population shifts (i.e. losses among all cohorts) from neighborhoods and corridors that are far better served by bus and/or not served by rail? Specifically on the South Side? We already know many of these bus-centric city neighborhoods and corridors are shrinking.

    Further, could transit ridership be a leading indicator of population numbers? Keep in mind the numbers since 2011-2013 are estimates and not a full count. Those numbers have shown citywide stabilization/very modest growth, that the “bleeding has stopped,” and have made some rest a little easier with cause for hope for a turnaround. But consider the average weekday ridership trend for the No. 79 bus, one of the South Side’s workhorse buses that cuts east-west through the heart of the South Side and is consistently among the most ridden weekday bus routes in the city:

    2014: 25791
    2013: 27518
    2012: 31356
    2011: 31897
    2010: 32073
    2009: 33663
    2008: 34453
    2007: 33380
    2006: 32424
    2005: 33316

    I know this route had its rush service decreased from after 2012 (, but those counts were in response to an already declining post-recession ridership.

    The variable of shrinking neighborhoods was not mentioned by the CTA spokesperson in their interview to the Tribune, I think, because the topic is gloom-and-doom and because those discussions quickly lead to a race/class component. That is always the elephant in the room. But DKH touched on the flip side of it above: that North Side neighborhoods along the Red and Brown lines and near Northwest neighborhoods along the Blue line have experienced growth in the number of “trip generators”/destinations, that downtown has experienced a population and jobs increase during the economic recovery, etc., all of which boost rail ridership. But the fact that South Side communities–and possibly West, though the West Side is better served by rail–are hollowing out and hemorrhaging people and jobs in a way that is either a.) still not yet fully reflected in population data or b.) offset by gains in other areas. And my gut is it’s A.

    I know this isn’t a groundbreaking reveal, but it seems important to be included in the conversation. Transportation costs (gas) have mostly gone up over the 2005-2014 period until the oil plummet began last year. There hasn’t been an explosion in bike-share or a new big rail line extension in this area. In a healthy neighborhood corridor–and frankly, a healthy city–ridership would have been a little more resilient through the recession and post-recession period.

    In short, because bus access is far more expansive across the city than the rail system, it feels like the CTA and City are trying to explain away a bus ridership decline by saying it’s partially offset by rail ridership and that 2014 was a brutal winter when, in fact, it’s actually a really, really bad indicator for the city as a whole.

    1. For me, the problem with this is the 2000s. We know that Chicago lost 200,000 people then – 8% of the population – largely from bus-served neighborhoods. Why, then, did bus ridership increase several percentage points over that period, even including the huge drop after the recession in 2009-10? If you leave that out – if you just cover 2000-2008, during which time the city surely lost well over 100,000 people – bus ridership was up almost 10%. What has changed since then? Why does falling population equal rising bus ridership in the 2000s, but falling ridership in 2013?

      1. I think there is a significant time lag in these effects. New buildings take several years to complete and become occupied.
        The census also is a significant lagging indicator with a healthy margin of error. It could be that by the time the population loss had shown up in the census, the population trend had reversed and was already having an impact on ridership statistics.

      2. Costs of car ownership when up dramatically during that decade. Gas prices alone were 3x in 2010 what they were in 2000.

  9. I wrote about declining bus ridership last year and it was interesting to hear from riders how amenities played into this:

    CTA bus tracker was rolled out before train tracker, and now there’s a train tracker at nearly every station (except Wilson), which can help inform your choice of bus v. rail more easily than did before.

    Also, a big complaint I have heard from bus riders is the lack of heat lamps at the bus stations. I know it’s always been cold in Chicago but I heard from riders that they got particular fed up waiting for buses during last winter’s polar vortex. There’s heat lamps at outdoor CTA stations and it’s warmer to wait in a subway station.

    It will be interesting to see how other alternatives to public transit — Divvy, Uber, Lyft, car share — have affected mass transit ridership. Also, a question that’s never really been answered is if the transition to Ventra affected the counting of rides since multiple readers were reported as problematic. (CTA says no.)

    1. Yeah. Actually, do you know anything about plans to roll out bus station heaters? The stop where I wait for the 2 to Hyde Park – at State and Lake – had little heat lamps for a week or two this winter, and then they were taken down. I thought maybe it was a test pilot or something?

      In my experience, the on-shelter bus trackers are also a huge improvement. I know I appreciate them – even though I have a phone that can check the time – but I also see other people stepping out to check them very frequently when I’m at a shelter that has one. I suspect that’s also partly a weather thing – when it’s really cold, you don’t want to expose your hands to do stuff on your phone, even for just a minute.

      I’m sure that there’s some displacement of transit rides with Divvy/Uber/etc., but so far, neither the numbers nor the geographic patterns of ridership match up with where the CTA is seeing the biggest declines in bus ridership. I think the answer has to be somewhere else for now.

    2. I know I didn’t have to pay for a bus ride the first 6 months of the Ventra launch (granted I only take the bus a couple times a month).

      I agree with others, its an ‘all of the above’ thing, with the biggest being the service cuts. After they were implemented, and I found myself frequently waiting 30 minutes for the Damen bus on a Saturday, I started either biking (if the weather was nice) or using my wife’s car for any errands/appointments/etc on the weekend.

      I assume a similar calculation has been made by everyone else in the city. In the denser, wealthier parts of the city, people transition to divvy/uber/cars, elsewhere I assume there’s more car sharing and trip consolidation than was previously needed, because no one wants to wait 20-30 minutes for a bus (even with bus tracker).

  10. I think John C. is right on.

    Bus routes, while a nice benefit to an area, are not a growth inducing amenity in the same way permanent infrastructure is. Considering that the US Congress has pretty much refused to fund any major commuter rail expansions in 30-40 years, this is why BRT is so valuable. Because it requires some of the permanent infrastructure associated with rail lines but at a much cheaper cost (and, in all fairness, with less benefit), it is the only real tool that the CTA has in increase bus ridership.
    Personally, i’m very interested to see what happens in Humboldt Park. This area is rapidly gentrifying and becoming more dense. Unlike similar areas, it does not have good access to the EL. If the bus routes serving this area see a notable increase in ridership, then i think you can say with good certainty that land use is driving these numbers.

  11. Up here at Pulaski and Irving Park, during rush hour we regularly have 35min headways on busses (though they are not SCHEDULED to have 35min headways, they consistently drive to a different schedule than the one CTA supplies). This makes it literally faster to walk from Wilson and Pulaski to the Blue Line than to wait for a bus — for those who can walk that far at a moderate pace.

    Add to that the fact that busses get trapped in traffic and you can see why taking more than one bus in a morning’s commute is a recipe for 1-2hr trip times, which is ridiculous and on par with suburban car commutes into the Loop.

  12. I know this has basically nothing to do with the data you’re presenting, but I think it’s relevant to CTA moving forward: was waiting for the #6 a few weeks ago on a freezing night. Saw on TransLoc that the bus wasn’t coming for another 15 minutes. Ordered an Uber-X and split it with the other people waiting for the bus. Ended up being $3 per person instead of the $2 bus fare.

    1. I think the impact of better taxi service on transit ridership is complicated, actually. On the one hand, you have instances like yours, where a ride on Uber or whatever is clearly displacing a ride on the CTA. On the other hand, sometimes transit is not going to serve a particular trip well – either because there’s a screwup, or it’s late at night, or you’re making an unusual trip – and having options like Uber makes it easier to accommodate those instances without just buying yourself a car. In other words, I think good taxi service is part of a transportation ecosystem that lowers the cost of not owning a car, and that will tend to increase the number of people who rely on transit for regular, routine trips.

  13. How delightfully complex. My personal take is that the likely explanations are Service>Demographics>Weather. Untangling the data here could be a thesis, but I think would be doable.

    As you lay out here, changed service is the obvious candidate for most likely culprit, and if you assume year to year noise caused by other shifting variables, and delayed adjustments, I suspect service reductions would explain most of the decrease. A line by line analysis might provide some insight – which lines have sustained the greatest reduction and the greatest gain? The answer to that would go a long way to pointing to the likely cause. On aggregate it’s hard to understand how closely tied service miles and ridership are, but it’s probably fairly easy on an individual line basis.

    I see Demographics as the second most likely culprit. Graphing the population change, income change, and employment change of the tracts that a bus passes through, and comparing those trends to ridership numbers for would go a long way to untangling the influence of demographics.

    Weather intuitively strikes me as a major impact on bus ridership, as bus riders suffer the most of transit users in adverse conditions. We have had a string of brutal winters. Graphing changes in the ridership for May-Sep over the years in question would go a long way to untangling the influence of weather.

    Explanations that cite central job growth, changing behavior due to technological disruption, or preference change all strike me as meme-driven, data-weak analyses. Shifting job locations are a slow trend and not isolated to the last few years, technological disruption of city services is an overblown storyline at this point, and preference change is almost always a flawed explanation that is bias driven.

    The change could be explained by small shifts in 100 variables, but I suspect it’s explained by 1 or 2 major changes that are being obscured by statistical noise.

    A fun and important question.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s