Chicago talks about its transit recovery as if it were one thing. It is two things, and they are not close. Between January and May of this year the CTA’s buses carried 83.3 million rides, 86 percent of what they carried in the same months of 2019. The trains carried 47.0 million, 65 percent. The bus system has nearly come back. The rail system is still missing a third of its riders.
Chicago’s buses are back to 86% of what they carried before the pandemic.
The ‘L’ is at 65%.
Rides taken January through May, 2026 against 2019 · CTA’s own daily counts
That gap has quietly rearranged the CTA. In 2019 the bus accounted for 57 percent of the rides across both systems. This year it is 64 percent. Whatever the CTA is planning around, close to two thirds of the rides it carries now happen on a bus.
Seven bus routes are carrying more riders than in 2019
- 94California119%
- 9Ashland115%
- 60Blue Island/26th113%
- 4747th109%
- 93California/Dodge105%
- 157Streeterville/Taylor103%
- 3King Drive101%
Seven routes are at or above their pre-pandemic ridership. Measured the same way over the same months, only three of the L’s stations clear that bar: Cermak-Chinatown, Morgan-Lake and Washington/Wabash, out of the 143 stations that recorded entries in both windows. Six of the seven bus routes are long neighborhood arterials: two California routes, plus Ashland, King Drive, 47th, and Blue Island through Little Village and Pilsen.
The seventh is worth naming rather than burying, because it cuts against the pattern. Route 157 does run through downtown, and it is above 2019. It is also the only winner with a hospital campus at each end, linking Streeterville’s medical district to the Illinois Medical District on the West Side. We cannot see who is on it — the data counts boardings, not occupations — so we will not tell you it is full of nurses. What we can say is that the one downtown-running route on this list is built around two places where the work is done on site.
One footnote on the L comparison, since we have published a different number. Our station-by-station look counted two stations above 2019, not three, because it compared average weekday entries. Washington/Wabash clears its 2019 mark on all-day totals but not on weekdays alone. Both numbers are right; they answer slightly different questions, and the all-day figure is the one that matches how we counted the buses here.
The route that should have been on that list, and isn’t
An earlier version of this article had eight routes on it. The eighth was 53 Pulaski, at 106 percent of 2019, and it does not belong there. On August 17, 2025, the CTA extended route 53 south to Ford City and cut route 53A back to the Pulaski Orange Line station, a change its board approved that June, and added route 53 to the frequent network that runs every ten minutes. Route 53 took over part of route 53A’s street and got more buses to do it with.
The ridership numbers show the handoff cleanly. Through 2025 the two routes rose and fell together every year. Then in the January-to-May window this year, route 53 gained about 434,000 rides and route 53A lost about 338,000. Counted separately, one looks like the recovery story of the year and the other looks like a collapse. Counted together, the Pulaski corridor carried 89 percent of its 2019 riders: a few points better than the bus system as a whole, and nowhere near the comeback that route 53 on its own appears to be having. Neither route is in the lists above, because their 2026 figures are answering a question about CTA planning rather than about Chicagoans.
On Ashland, the corridor held. The mix inside it moved.
Ashland Avenue
Western Avenue
Ashland Avenue carries two CTA routes along the same north-south corridor: the local, route 9, which stops every few blocks, and X9, a limited-stop express that skips most of them. Neither one runs into the Loop. The local is now running above its 2019 ridership while the express sits at about two-thirds, and Western Avenue shows the same split.
Add the two routes together and Ashland is carrying slightly more bus riders than it did in 2019, about 102 percent of the old total. The corridor did not lose its ridership; the mix inside it moved, with the express falling from 27 percent of Ashland boardings to 17. Route 49 and the X49 together are at 88 percent, with the same shift toward the local; fold in the 49B up at the north end of Western and the whole street comes to 85. What the counts cannot tell you is why. They record boardings, not decisions, so nothing here shows an individual choosing the slow bus over the fast one. How often each bus runs is part of the story too, and the CTA changed how it spread service across routes even as it restored the total.
Two different ways to end up at the bottom
- 2Hyde Park Express53%
- 135Clarendon/LaSalle Express57%
- 52Kedzie60%
- 115Pullman/115th62%
- 119Michigan/119th64%
- 156LaSalle66%
- X9Ashland Express66%
- 134Stockton/LaSalle Express66%
- X49Western Express69%
- 111111th/King Drive70%
- 91Austin72%
That is the bottom eleven of the seventy-eight, in order, with nothing left out between them. The express routes that run into the Loop sit in the middle of it. The Hyde Park Express is at 53 percent, the Clarendon and Stockton expresses down LaSalle Street at 57 and 66. All three are weekday-only routes built around office hours, and they are the least recovered part of the bus system. That is the same shape we found underground on the Red Line, where the State Street subway stations that serve the Loop run near half of 2019 while neighborhood stops did better.
But two very different histories are sitting side by side on that list, and the same dataset separates them. Run the routes back to 2010 and the Hyde Park Express was not a declining route: it was carrying 36 percent more riders in early 2019 than in early 2010, growing right up to the pandemic before losing half its riders in a single year. Some of the South Side routes further down the list got there the opposite way. The 111th Street bus was already down to 45 percent of its 2010 ridership by 2019, and the Michigan/119th bus to 62 percent, both of them shedding riders across the decade before the pandemic arrived. One group fell off a cliff in 2020. The other had been walking downhill since before the decade started, and the pandemic simply found it further along.
The schedule explains less than you would expect
The obvious objection is that ridership follows the schedule, and a route with fewer buses will carry fewer people. That objection has an answer here. The CTA restored its scheduled bus service to pre-pandemic levels at the end of 2024, the last of it taking effect that December, as local reporting covered at the time. The 2026 numbers here are measured against a system running a full schedule, not a shrunken one.
At the route level the caution still applies, and the agency’s own figures show it has teeth. The CTA restored total service hours while changing how they were spread across routes, and in the same announcement it reported that routes which had received extra service were up 14 to 15 percent year over year that October, against 10 percent for routes that had not. An individual route’s number here reflects both who wanted to ride it and how often a bus came. Moving service between routes largely washes out of a systemwide total, which is why the 86-against-65 gap is the finding here and the route list is the texture around it.
How we counted
- Source: the CTA’s own daily ridership counts, published by the city: bus totals by route (jyb9-n7fm) and station entries for the L (5neh-572f), both current through May 31, 2026.
- The window: every figure compares January 1 through May 31 of one year against the same months of another, so the partial 2026 year is never set against a full one.
- Routes shown: routes carrying at least 300,000 rides in the 2019 window, which leaves 80 comparable routes. Route names come from the CTA’s bus-route file.
- Route figures are per service day. Some routes run weekdays only and some run every day, and a few have changed which days they run since 2019, so we divide each route’s rides by the number of dates it actually reported service. Comparing raw totals instead flags exactly the same routes as being at or above 2019, so the choice of method does not change who is on the list. Route 93 is the only one of the 80 that moves materially between the two methods: it reads 124 percent on raw totals and 105 percent per service day, because it added Sunday service it did not run in 2019 (Monday through Saturday then, every day now). Systemwide figures are unaffected, since both windows cover the same 151 calendar days and every one of those days is present in the data for both years.
- Did cancelled routes inflate the survivors? Only two routes that ran in the 2019 window were gone by 2026, and together they carried 58,686 rides, six hundredths of one percent of the bus system. Two routes appear in 2026 that were not in 2019, the larger being the X4 Cottage Grove Express.
- We also checked for routes that were rebuilt rather than cancelled. A route that gets extended or truncated keeps its number, so it survives that first test while its ridership figure quietly stops meaning what it used to. We compared every route’s 2025 window against its 2026 window and flagged any that moved sharply against the systemwide trend. Five did. Two are the Pulaski pair described above, and both are excluded from the lists here. Of the other three, the 93 gained the Sunday service already accounted for, and the X49 and 115 ran on the same number of days in both years, so neither drop is a change in which days the route operates. This dataset carries daily ride counts and not scheduled service hours, so it cannot tell us whether the bus came less often within those days — on any single route, that remains a live possibility.
- The L’s two odd stations. State/Lake carried 1.5 million entries in the 2019 window and has since been shut for reconstruction; dropping it from both years lifts the L recovery from 65 to 67 percent and leaves the gap with the bus intact. Damen-Lake on the Green Line opened after 2019, so it appears in the 2026 data with no 2019 figure to be measured against, and is not part of the 143-station comparison either way.
- What the two systems count: bus figures are boardings; L figures are station entries, which count a rider once per trip even if they change trains. The two are not interchangeable measures, so we compare each system against its own 2019 rather than against each other in absolute terms.
- What this does not show: why any individual route moved. Ridership reflects demand and service together, and route restructuring can move riders between routes that share a corridor.
Computed by KCM Desk from CTA records through May 31, 2026; published July 19, 2026. We will run these windows again when the full year lands. If you spot an error, corrections come first.