Two Futures for NYC’s Aging Highways:

Rebuild them to current, wider federal highway standards, or replace them all with street-level transit, trees, parks, and neighborhood connections. This model prices both over a 30 year period to visualize what we currently plan to spend and what we could save if we were to make different choices. Challenge the assumptions.

Because Robert Moses shouldn’t get to choose for you, too.

The Off-Ramp Model — What if NYC removed its highways?
$0.0B
30-Year NPV Savings · Removal vs Rebuild
Removal wins

Removing the Brooklyn-Queens Expressway (BQE), the East Side Highway (FDR), the Cross Bronx Expressway (XBX), and the West Side Highway (WSH) and replacing them all with transit at street level costs less than rebuilding the highways over a 30-year period.

How to read this

Every number is a 30-year net present value, 2029–2058, bond-financed on identical terms for both futures. Gray bars are the cost of rebuilding each highway in place to current standards. Colored bars are the cost of demolishing it and building transit on the land — construction, vehicles, and operations, minus fares, property tax, and land sales. The gap is public money saved. Every assumption below is adjustable; the formulas are not.

Rebuild vs Removal · 30-yr NPV by corridor

Rebuild Removal + transit

Corridor detail

CorridorModeRebuild NPVRemoval NPV Savings% savedStationsFleetCapacity

Not counted anywhere above: property near a removed highway appreciates. At a conservative 5% uplift on adjacent value (Embarcadero and Cheonggyecheon precedents run 5–15%) and a 1.5% effective tax rate, the four corridors would throw off roughly $3.2B in additional 30-year tax increment NPV ($0.5B BQE · $1.3B FDR · $0.4B XBX · $1.0B WSH at model defaults). We exclude it from every number on this page.

Assumptions 🔒 Formulas locked — inputs only

Corridors · Ridership & Mode
Set all:

Daily boardings each corridor must carry, and the transit technology replacing the highway.

Subway is priced as new grade-separated structure — it cannot street-run on a cleared at-grade corridor, so most segments carry tunnel or new-structure costs.

Financing & Design

Assumptions by transit mode The actual numbers — challenge them

AssumptionLight RailElectric BRTSubway
Guideway rate card · $M per route-mile by segment type
Segment typeLight RailElectric BRTSubway
Station rate card · $M per station by station type
Station typeLight RailElectric BRTSubway

Where the transit money goes 2026$, before escalation & financing

CorridorGuidewayStationsSystems Contingency*Vehicles= Transit capex$/route-mi DemolitionAnnual O&MLand proceeds

* Contingency is charged on both sides. The removal scenario carries 50% on every dollar it spends — guideway, stations, systems, and demolition alike, per FTA practice for pre-engineering estimates. Highway rebuild carries 35%, matching Flyvbjerg’s +34% average overrun for bridges and tunnels — the right benchmark for rebuilding urban viaducts and trenches under live traffic. Both are adjustable above; the removal side carries the heavier burden by default.

People vs. pavement · per lane, per hour

One highway lane moves about 2,300 people per hour (1,800 vehicles at NYC's 1.3 average occupancy). The same width of right-of-way, given to transit:

How this works

This sandbox runs the same arithmetic as the full Off-Ramp NYC financial model: per-segment highway rebuild costs (NYC DOT, NYSDOT, and FHWA unit rates, escalated to spend year, rebuilt to current standards with the widening premium above), versus demolition plus transit built on the freed right-of-way — guideway, stations, systems, and vehicles, with 50% contingency on transit and demolition, 35% on highway rebuild (Flyvbjerg fixed-links benchmark). Both futures are bond-financed on identical 30-year terms, and each bond tranche is carried through its full 30-year term — including payments that fall beyond the model's final year (2058). Transit revenue counts fares (ramping over three years), incremental property tax, and a one-time sale of a share of freed land. Each corridor carries its own ridership target and mode. All figures are 30-year net present values, 2029–2058.

Conservatism still runs against removal where it matters: the fare is blended down for Fair Fares, and land sale revenue is set to zero by default — freed land is carried as public benefit, not monetized (the slider below tests revenue scenarios; currently 0% of genuinely sellable land is sold). Each segment first carries a sellable fraction (0% where the highway sits over a mapped street, along parkland, or on a waterfront; 50% for filled trenches, per the Rochester Inner Loop precedent; 40% for highway-owned elevated rights-of-way). Interchange and cloverleaf footprints — fee-owned land the linear math misses — are counted separately, net of the ramp movements that must survive. The conclusion survives anyway.

This is a financial exercise, not an engineering plan. Numbers may be refined as the model evolves.