What the records show about the 2024 Penn Station feasibility study

This page covers the October 2024 feasibility study Doubling Trans-Hudson Train Capacity at Penn Station, prepared by WSP/FXC for Amtrak, the MTA, and NJ Transit. It is separate from the Regional Plan Association’s April 2026 Penn Station report and from the Federal Railroad Administration’s (FRA) Service Optimization Study (SOS) with Deutsche Bahn, now underway. The first part sets out what public-records responses show about the study’s modeling. The second part, which absorbs the January 2025 essay “Beyond the Terminal Trap,” makes the case for running trains through Penn Station instead of expanding it as a terminal.

What the records show

Public-records responses later showed one record about Amtrak’s possession of simulations, another about whether simulations were performed, and a third in which FRA sought access to a Penn Station Rail Traffic Controller (RTC) model developed with the three railroads. Separate Service Optimization Study records released by NJ Transit show the railroad challenging RTC outputs proposed for public capacity reporting.

Amtrak’s response on appeal

After a request for the study’s underlying data, Amtrak’s legal counsel responded on appeal about the records it holds.

“[T]he search did not reveal any simulations or modeling data in Amtrak’s possession...”

“[T]he appendix in the earlier draft versions contained a cost estimate for an alternative that was no longer being considered by the project team.”William Herrmann, Executive Vice President, Chief Legal and Human Resources Officer, Amtrak. Amtrak appeal letter (PDF)

NJ Transit’s response

Asked for the same modeling data under New Jersey’s Open Public Records Act (OPRA), NJ Transit said it did not possess the requested simulation records.

“Please be advised that NJ TRANSIT does not possess records responsive to this Request.”OPRA Unit, Office of the General Counsel, NJ Transit. NJ Transit response (PDF)

FRA’s request for a Penn Station RTC model

Records NJ Transit later released under OPRA add a step to the chronology. On November 6, 2025, the FRA asked NJ Transit to release a model it described as a

“Penn Station-specific RTC model developed in coordination with MTA and Amtrak”Lyle Leitelt, Federal Railroad Administration, November 6, 2025. FRA request record (PDF)

NJ Transit’s objections to the federal study’s RTC files

The same production includes March 2026 correspondence about the federal Service Optimization Study. NJ Transit identified technical problems in the study’s RTC files and objected to public-facing trains-per-hour capacity figures, citing

“significant deficiencies in both the modeling assumptions and the resulting analysis”Kerry Donovan, NJ Transit, March 24, 2026. NJ Transit RTC challenge record (PDF)

The issues included shortened train consists, zero-second switch handling, mismatches between track drawings and the RTC case, dwell-parameter violations, and optimistic delay reporting. This record concerns the federal study’s analysis, not the October 2024 WSP/FXC study.

The records do not settle the modeling question. Amtrak framed one response around possession and another around whether simulations were performed, and NJ Transit’s records later refer to a Penn Station RTC model while challenging the separate federal study’s RTC outputs.

What the study team said before publication

In August 2024, before the study was published, project representatives discussed their plans at a public forum. Attendees asked about the data and methodology. Foster Nichols of WSP answered: “we're certainly happy to produce all the documentation and calculations and there'll be a report submitted sometime after Labor Day.”

• 0:00. Layla Law-Gisiko of The City Club asks for the “methodology” and “general assumptions” used in the study.
• 0:58. Asked whether the southern expansion meets the stated goal of 48 trains per hour, an Amtrak official says: “We believe that it will.”
• 1:13. Asked for clarification, the response describes the plans’ status: “We’re still in concept design...”
• 1:53. Foster Nichols of WSP says the team is “happy to produce all the documentation and calculations” and that a report will be submitted “sometime after Labor Day.”
• 2:52. A speaker describes the approach of first testing whether the goals can be met within the existing station footprint before evaluating expansion.

The report was published in October 2024.

The case for through-running

Penn Station is the busiest rail hub in the United States, and it still runs like a dead end: trains pull in, sit, reverse, and block the tracks behind them. Cities that rebuilt their central stations for through-running, where trains pass through instead of ending there, moved more people more reliably for less money. New York can do the same without demolishing neighborhoods for more stub-end tracks. It has to stop operating like it’s 1910.

The core problem

Stub-end tracks force long dwells: passengers unload, crews change, trains reverse, and each must be dispatched back out. Those minutes add up to lost capacity, missed schedules, and crowded platforms. With through-running, trains arrive, open their doors, and continue; shorter dwells mean more trains on the tracks that already exist. Philadelphia, Paris, Tokyo, and London broke their terminal bottlenecks this way, which is why New York’s $16.7 billion plan to expand a stub-end terminal is a costly detour.

At a City Club of New York webinar on January 15, 2025, Philippe Crist of the OECD asked listeners to imagine a city they could drive into at different points but could not drive through, and said rail needs the same through connectivity.

Why the $16.7 billion expansion misses the point

The railroad partners say more terminal tracks will cure congestion. They will not: more gates at a cul-de-sac do not make it a street.

• Capacity per dollar. When New York State announced the southern expansion in January 2020, it was projected to handle 40 percent more trains.
• Operational waste. Stub-end operation ties up platforms and interlockings with turnbacks and midday yard runs. Through-running keeps equipment moving and frees the core.
• Hidden costs. Bigger terminals need more storage, staff, and layovers. Through-running cuts midday storage needs and puts crews to work in service instead of waiting.

The southern expansion would also raze Block 780, including a church, homes, and small businesses, to extend the same operating pattern. New York can add capacity without erasing a neighborhood.

Proof from elsewhere

• Philadelphia. The Center City Commuter Connection linked two rival terminal systems with a 1.7-mile tunnel, opened in 1984. It provided the same coverage with fewer trains and crew hours, revived the Market East district, and got more value from existing lines while helping catalyze growth.
• Tokyo. The world’s most complex multi-operator network works because its operators adopted through-running step by step, with multi-voltage trains, shared protocols, and negotiated labor rules. If Tokyo can harmonize that, Penn Station’s three operators can too.
• Paris. The RER threaded suburban lines through new central tunnels and ran frequent, metro-like service across agencies and power systems, giving riders direct trips, drawing massive ridership, and tying the region together as one network.
• London. Thameslink started small by reopening an old tunnel, proved the demand, and grew in phases. The core was built for 24 trains per hour; the working timetable for October 6, 2026 schedules 20 in the peak hour.

The barrier is governance

Penn Station’s geometry is workable. The obstacle is that Amtrak, NJ Transit, and the MTA each plan around their own yards, rules, schedules, and territories.

“In Berlin there are 41 different operators of transit in the region ... but as a user you’d never know it ... there’s one ticket, there’s one timetable ... there’s one network.”Paul Lewis, Deutsche Bahn E.C.O., City Club of New York webinar, January 15, 2025

New York needs the same rider-first approach: a single Penn Station through-running authority with the mandate to run the core as one railroad: unify the timetable (one slot plan, sensible line pairings, shared dispatch), resolve labor rules (cross-territory crews, cross-training, clear operating protocols), and own capital planning so approach tracks, platforms, and yards are funded and sequenced for the integrated plan rather than three separate wish lists. Working groups have not delivered through-running; an authority can.

The usual objections

• The columns. Some columns need work for targeted platform widenings, which can be phased.
• Train storage. Through-running reduces storage demand at the core. Outlying yards in Secaucus, Queens, and the Bronx can clean and turn trains where space is cheaper, leaving Penn Station for movement.
• Lopsided reverse-peak demand. Pair lines with similar volumes, short-turn some services, and add frequency where demand exists, as Tokyo and Paris do every day.

A phased plan

1. Operational changes, 12 to 24 months. Standardize the platform interface where feasible and tighten dwell targets; pilot cross-running between NJ Transit and the LIRR or Metro-North on limited patterns with dual-power or multi-system equipment; set up integrated scheduling and dispatch software and publish an integrated slot diagram; and move to shared fare media so the system works as one network for riders.
2. Targeted capital, 2 to 5 years. Rebuild selected interlockings, upgrade signals to cut headways, and add a few high-value connections near Sunnyside and Secaucus; widen platforms where passenger flow constrains dwell; and buy or retrofit rolling stock for cross-system operation.
3. Scale and optimize, 5 to 10 years. Expand through patterns across the peak and add mid-distance runs linking New Jersey with Queens and Long Island; continue decentralizing yards, adding crew flexibility, and refining the slot plan; and grow in phases, as Thameslink did, so riders see each gain.

Why it matters beyond trains

Through-running is a mobility upgrade, a climate strategy, and an equity policy at once. Riders get time back through shorter dwells, fewer transfers, and better reliability. One-seat rides put more jobs within reach, including reverse-commute and cross-region trips to places like Newark and Queens. More rail capacity at all hours means fewer car trips and lower emissions. And stations with frequent, direct service attract homes and jobs without razing blocks to store trains.

The question to answer

At a House hearing on December 9, 2021, Representative Seth Moulton asked NJ Transit’s chief executive what would happen to Penn Station’s capacity if NJ Transit trains ran through to Long Island and the reverse. Moulton then said: “We looked at Boston, and it increased capacity at South Station by about eight times.” The question for Penn Station is not “how many more stub-end tracks can we bolt on,” but “how do we turn Penn from a terminus into a network?”

Choose integration

Across four world capitals and one Philadelphia tunnel, the pattern holds: through-running works when leaders choose it and align institutions to deliver it. New York’s “too hard, too expensive” line is a policy choice, not a technical fate. Create the authority, start the pilot, fix the chokepoints that matter, and build support by showing results. Penn Station should serve the region it anchors, and through-running is how.