Transit UPT Index vs 2019: Mid-Size Agencies Near 88, Mega Urban Stuck Near 74

Unlinked passenger trips — every boarding counted once, transfers counted again — are the Federal Transit Administration's workhorse ridership unit. When boards ask "are we back?", they almost always mean: how does calendar-year UPT compare with 2019, the last full pre-pandemic year in the National Transit Database (NTD).

The answer at the industry level is now stable enough to parse by agency size and mode. Drawing on FTA NTD annual and monthly ridership through CY2024, APTA's December 2024 mode recovery snapshot, and Census ACS commute shares for context, this desk finds a national UPT index near 79 (2019 = 100), a bus–rail gap of about 14 points (86 vs 72 as of December), and a size gradient that puts mid-size and small urban reporters closer to the baseline than mega and large systems that still dominate national totals.

The interactive dashboard above traces the national index path, median recovery by size band, a mode radar, an agency scatter of 2019 UPT against 2024 index, and ACS commute-share deltas for selected metros.

National UPT is near four-fifths of 2019 — not full recovery

FTA's 2024 National Transit Summaries and Trends put public-transit UPT near 7.6–7.7 billion trips for the year — roughly a 10% gain over 2023, and still short of the ~9.9 billion trips that anchored 2019. APTA's Ridership Trends dashboard framed year-end 2024 ridership around 79% of the comparable 2019 period. Those two readings — annual NTD totals and APTA's same-period index — are close enough to treat ~79 as the working national UPT index for this piece.

Service did not shrink as much as ridership. Agencies operated on the order of 93–95% of 2019 vehicle revenue miles while carrying closer to four-fifths of the trips. That wedge is the empty-seat problem in one chart: operators restored schedules faster than passengers returned to seats, especially on peak-oriented rail.

The path is a classic pandemic V with a long right shoulder. The trough in 2020 cut rail harder than bus. The 2022–2024 climb favored bus and demand response; rail closed the year still near 70–72 of its 2019 level.

Size bands matter more than the national average

National averages are pulled by a handful of very large urban reporters. Group full reporters by 2019 total UPT and the recovery story splits:

• Mega urban (≥500M trips in 2019): desk median index near 74. These systemsNew York City Transit foremost — still carry a large share of national UPT, so their incomplete recovery anchors the industry total.
• Large urban (50–500M): median near 78, with a wide band from rail-heavy underperformers (BART-style indices in the mid-50s) to bus-leaning Sun Belt agencies in the mid-80s.
• Mid-size (10–50M): median near 88closer to the baseline, and often bus-dominated.
• Small urban (<10M): median near 92, with a visible right tail of agencies that exceeded 2019 after fare-free experiments or route redesigns.

Put differently: the agencies closest to "back" are not the ones that move the national total. Mid-size and small systems can post 90–100+ indices while mega and large reporters still leave a 20–30 point hole relative to 2019. That is why boardrooms in mid-size cities hear a different recovery story than congressional staffers reading the national NTD headline.

Among the largest ~150 agencies by 2019 UPT, only about 7% have returned to or surpassed their pre-pandemic trip counts. The median change in that universe is on the order of −22%. Recovery is real; full restoration is rare.

Bus recovered faster than rail — not the other way around

A common intuition after every downtown revival story is that "rail is coming back first." The NTD/APTA mode cut says the opposite for the national stack.

As of December 2024 versus December 2019 (APTA Policy Brief, May 2025):

• Demand response: ~93
• Bus (all): ~86
• Light rail / streetcar: ~76
• Heavy rail: ~71
• Commuter rail: ~70
• All rail (aggregate): ~72

Bus leads the all-rail aggregate by roughly 14 index points. Heavy and commuter rail — the modes most tied to central-business-district office peaks — remain the softest. Light rail sits in between. Demand response, which never depended on five-day office towers, is nearest to full recovery.

Agency-level mode splits echo the national pattern. In the curated panel, bus-heavy systems (bus ≥70% of 2019 UPT) cluster higher on the scatter than rail-heavy peers at similar size. CTA, WMATA, MBTA, SEPTA, and BART all show rail indices below their bus indices where both are separable. LA Metro, Houston METRO, and Miami-Dade — more bus-weighted — sit closer to the mid-80s overall.

Local exceptions exist. APTA notes that MBTA commuter rail has, at times, led heavy/light rail recovery in Boston. Those exceptions do not overturn the national ordering: for the United States as a whole, bus recovered faster than rail.

Hybrid work shows up in ACS commute shares

UPT counts every trip — errands, school, weekends, tourists. ACS means of transportation to work isolates the commute. Comparing 2019 and 2023 ACS public-transportation commute shares for selected metros shows where the office-peak hole is largest:

• San Francisco–Oakland: roughly −5.8 percentage points
• New York–Newark: about −4.8 points (still the highest absolute transit-commute share)
• Washington and Chicago: around −3.3 points each
• Houston and Phoenix: small absolute drops on already thin baselines
• Tucson: a rare flat-to-up tick alongside fare-free bus ridership above 2019

Commute-share declines do not map one-for-one onto UPT indices — non-work trips returned faster in many networks — but they help explain why rail-heavy mega systems lag. When Tuesday–Thursday office occupancy stays below 2019, peak rail loads stay soft even if weekend and midday bus ridership looks healthier.

What the curated agency panel shows

The dashboard's 24-agency panel is a desk sample spanning size bands, not a census of every NTD reporter. Rounded UPT totals are aligned to published annual agency figures and urban-area summaries.

Illustrative anchors:

• NYCT near 79large enough to set the national tone, with bus ahead of subway.
• CTA near 66, with rail weaker than busa Midwest rail-commute story.
• LA Metro near 83 on a bus-weighted network.
• BART near 55almost pure heavy rail into soft office peaks.
• MARTA near 57; Valley Metro near 60large systems still far from baseline.
• Houston METRO and Miami-Dade in the mid-80sSun Belt bus networks closer to 2019.
• GRTC (Richmond), Sun Tran (Tucson), DASH (Alexandria), and CDTA (Albany) at or above 100small and mid systems that crossed the line, often with fare policy or bus-only networks.

Filter the scatter to rail-heavy and the cloud drops; filter to bus-heavy and it lifts. Filter to small and the recovered dots (green) appear. The national 79 is an average of very different recoveries.

Caveats and how to read the index

Several measurement cautions apply:

1. Unlinked trips double-count transfers. A bus-to-rail journey is two UPT. Networks that redesign for more transfers can raise UPT without raising linked journeys. Linked-trip series (where available) are a useful cross-check but are not the NTD headline.
2. Agency boundaries and purchased transportation can shift year to year. Mergers, new modes, and reporting changes move UPT even when passenger behavior is flat.
3. Fare-free periods (Richmond, Tucson, and others) boost measured trips; treat those indices as policy-affected, not pure demand recovery.
4. Size bands are desk constructs based on 2019 UPT thresholds. FTA does not publish an official "mega / large / mid / small" recovery product; medians here are derived from public agency totals.
5. 2024 annuals and APTA monthly indices are close but not identical windows. This piece uses ~79 for the national story and APTA's December mode cut for bus-vs-rail.
6. ACS commute shares lag (2023 vs 2019 here) and cover workers onlynot students, tourists, or off-peak riders.

None of those caveats erase the central pattern: large urban reporters remain furthest from 2019, bus leads rail, and national recovery is still a fifth short of the pre-pandemic trip count.

What boards and funders should take from the size split

If capital and operating decisions are framed only against the national 79, mega and large systems look "almost there" when many are actually in the mid-60s to mid-70s, while mid-size systems near 88 can look average when they are already closer to replacement levels. Mode mix compounds the error: a rail-heavy large agency at 70 and a bus-heavy mid-size agency at 90 are not in the same recovery regime.

For service planning, the empty-seat wedge (VRM restored faster than UPT) argues for peak restructuring on CBD rail before blanket headway restoration. For equity and essential-worker service, the bus lead is a reminder that midday and coverage routes carried more of the pandemic floor — and more of the rebound.

The UPT index versus 2019 is not a verdict on transit's usefulness. It is a size-and-mode map of how far demand has come back. On that map, mid-size America is nearer home than the largest urban reporters — and the bus, not the train, has been the faster vehicle.