Park the Bus: How the 2026 World Cup Stress-Tests North American Transit
Causal analysis of ridership, fare equity, and transit strain across six 2026 World Cup host cities. Synthetic control, panel DiD, event studies, and an interactive D3/Leaflet dashboard. Living analysis updated weekly.
What this is
The 2026 FIFA World Cup (June 11 to July 19) sends 48 nations to 16 North American cities. This project asks what that does to the transit systems. How much more do people ride? Which corridors hit capacity? What does it actually cost to reach the stadium, in dollars and as a fraction of local wages? And how do the answers compare to Qatar 2022, Russia 2018, and the other tournaments that came before?
This is a living analysis. The dashboard and report update weekly as data arrives. Event-study results (MTA hourly, CDMX Metro daily) appear first; synthetic control and panel DiD results follow once the FTA National Transit Database releases June 2026 monthly figures in August.
The Interactive Dashboard
The dashboard below (also open full-screen) shows:
- Map: six host cities sized and coloured by peak Transit Strain Index
- City panels: confirmed match calendar with teams and per-match strain ratings; ridership chart (actual vs. baseline with World Cup window shaded)
- Fare comparator: stadium roundtrip cost by city; toggle between $ absolute, % of local minimum wage, and % of international fan daily budget
- Past World Cups: how Qatar 2022, Russia 2018, and Brazil 2014 compare on transit policy, cost, and mode share
- What-if calculator: adjust fill rate, transit mode share, and service uplift to see corridor load in real time; full formula shown
- Model vs. Reality: predictions tracked against actual data as it publishes
The Six Cities at a Glance
| City | Stadium | Capacity | Transit mode share | Peak strain | Roundtrip fare |
|---|---|---|---|---|---|
| New York / NJ | MetLife Stadium | 82,500 | ~45% | 9.5/10 (Final, Jul 19) | $25 standard / $150 VIP |
| Mexico City | Estadio Azteca | 87,523 | ~60% | 7.2/10 | $0.55 (USD) |
| Vancouver | BC Place | 54,500 | ~80% | 7.9/10 | Free (w/ ticket) |
| Seattle | Lumen Field | 72,000 | ~38% | 7.2/10 | $7.00 |
| Atlanta | Mercedes-Benz | 71,000 | ~28% | 9.0/10 (Semi-final, Jul 15) | $5.00 |
| Toronto | BMO Field | 45,736 | ~72% | 6.4/10 | ~$4.84 (CAD) |
The Transit Strain Index = projected incremental riders divided by scheduled corridor capacity in the 3-hour match window. Above 7 means the corridor is being substantially outpaced by demand. The system copes through crush loading, longer egress times, or walk-away crowding.
The NJ Transit Controversy
MetLife Stadium has no direct rail connection to New York City. Fans take NJ Transit from Penn Station to Secaucus Junction, then a shuttle. NJ Transit announced a special World Cup package at roughly $150 per person versus $25 for standard service. That is 8.8 hours at NYC’s minimum wage of $17/hour.
This is not a story about NJT being greedy. It is a story about what happens when a multi-billion-dollar international event is hosted at a stadium served by a single-corridor commuter rail line with no platform redundancy. Price is how you allocate a scarce physical resource when you cannot expand it quickly.
Vancouver chose a different answer: free transit for all ticket holders, funded from a $21.6M CAD World Cup operating budget. The expected result: 80% of BC Place attendees arriving by SkyTrain.
For context, Qatar 2022 was free for all ticket holders. The Moscow Metro in 2018 cost about $0.60 USD per roundtrip. North America 2026 spans a wider range than any prior World Cup, from free to $150.
Methods (brief)
Synthetic control: per host agency, donor-pool weights fitted on 2015-2019 and 2022-May 2026 monthly UPT (NTD). Treatment = June/July 2026. Inference via in-space placebos. COVID years 2020-21 excluded from fitting window.
Panel DiD: log(UPT) on Host x WC_Window, agency and month-year fixed effects, clustered standard errors. Event-time plot tests pre-trends.
Event study (daily): match-day regression on MTA hourly and CDMX Metro daily data with day-of-week and week fixed effects. This is the headline evidence for within-month variation.
Strain Index: planning-tool composite: incremental demand from stadium capacity times mode share times fill rate, against corridor capacity from serving lines times vehicles/hr times vehicle capacity times 3-hour window. Calibrated to actual ridership as tournament data arrives.
Data Pipeline
pipeline/
├── 01_fetch/ FTA NTD · MTA Subway Hourly · CDMX Metro Afluencia · GTFS
├── 02_clean/ typed parquet panels (agency x month, daily NYC, daily CDMX)
├── 03_analysis/strain index + event studies -- results_*.json
└── 04_export/ writes the JSON the dashboard readsRe-run with python run_pipeline.py. Dashboard JSON updates weekly via GitHub Actions.
Status
| Component | Status | Notes |
|---|---|---|
| Dashboard | Live | Match calendar (teams from FIFA schedule), fare comparator, what-if, historical WC |
| City profiles | Live | Six cities, surge plans, confirmed match schedule |
| Event study (MTA/CDMX) | Pending | MTA data lags ~4-6 weeks; expected late July |
| Synthetic control | Pending | Requires NTD June release ~August 2026 |
| Panel DiD | Pending | Same as SC |
| Model accuracy tracker | Live | Updates as actuals arrive |
| Full written report | In progress | Skeleton live; results fill in as data arrives |