Chapter 01 · The Air

How Nitrogen Dioxide Changed Across NYC After Congestion Pricing Began.

Sixteen months of satellite measurements suggest pollution declined more over Manhattan's congestion zone than over the rest of New York City. The placebo comparison helps separate background atmospheric drift from changes more plausibly associated with the toll.

The seasonal cycle, animated

Sixteen Months of Monthly NO₂, Frame by Frame.

Monthly NO₂ (μmol/m²)

≈82 ≈100 ≈125 ≈150 ≈172

Lighter cream = less pollution. Darker green = more pollution.

Pollution rises seasonally during winter and falls during summer. The spatial pattern shifts after January 2025, when the congestion zone began.

Before, after, and change

The Congestion Zone's Footprint in the Air.

The left panels show average NO₂ before and after congestion pricing. The difference map on the right isolates where pollution changed most strongly — concentrated around lower Manhattan and adjacent corridors, rather than appearing evenly across the city.

Before the zone

Jan 2023 – Apr 2024

Average NO₂ over NYC, January 2023 to April 2024

After the zone

Jan 2025 – Apr 2026

Average NO₂ over NYC, January 2025 to April 2026

What changed

After minus before

Change in NO₂ between the two windows

Before / After (μmol/m²)

80 124 167

Change (μmol/m²)

−12 0 +12

Pixels are TROPOMI satellite cells (~5×3 km). The cordon polygon is outlined in cream. Same calendar months in each window remove seasonal cycles from the comparison.

Where pollution declined most

Change in Monthly NO₂ by Neighborhood.

Each shape is a NYC neighborhood (TLC taxi zone). Greens indicate decreases; terracotta indicates increases. The congestion zone is outlined in cream.

NO₂ change by NYC neighborhood, after the congestion zone began A diverging-color map of percent change in nitrogen dioxide. Most NYC neighborhoods are some shade of green, indicating decreases. The congestion zone is outlined in cream.

Percent change in monthly NO₂ between the pre and post windows. decrease increase congestion zone

n=263 zones · 95% CI on the citywide gap [−2.26, −1.46] pp · TROPOMI ~5×3 km native footprint

Why this counts as a finding

Background Atmospheric Drift, Ruled Out.

NO₂ shifts year to year for many reasons unrelated to policy. The analysis tests this directly: the same comparison run a year earlier, before the congestion zone existed, shows little comparable Manhattan signal. The post-2025 pattern reflects more than ordinary atmospheric variability.

Placebo change · 2024 vs. 2023, no policy in effect

−12 μmol/m² 0 +12

cleaner more polluted congestion zone

The placebo year. Roughly flat over lower Manhattan — without the policy, the cordon area doesn't show the kind of anomaly that appears in the real comparison. The methods chapter explains the placebo design step by step.

Sorting neighborhoods three ways

Three Ways of Grouping NYC.

To ask whether different kinds of neighborhoods saw different effects, the analysis sorts NYC's 263 zones into thirds along three traits — income, car ownership, and transit reliance. The maps below show where each group sits geographically; the chart that follows compares how the congestion zone's air benefit reached each one.

Neighborhood income

Lower Mid Higher

How car-free

Most cars Mid Most car-free

How transit-reliant

Least transit Mid Most transit

NYC neighborhoods sorted into thirds for income, car ownership, and transit reliance. Click any map to enlarge. The congestion zone is outlined in terracotta.

Did all neighborhoods benefit equally?

The Pattern Does Not Map Cleanly Onto Neighborhood Wealth.

The congestion zone improved most relative to higher-income, less transit-oriented neighborhoods elsewhere in the city. Lower-income, car-free, transit-dependent neighborhoods often experienced pollution declines similar to the zone itself, producing smaller relative gaps.

This does not suggest the benefits were regressive. Many lower-income transit-reliant neighborhoods saw measurable improvement. The gains were uneven, and some neighborhoods bordering major truck corridors appear to have benefited less than the Manhattan core.

Cordon-vs-rest gap in NO₂ change, by demographic tertile Each row compares the congestion zone to a group of similar neighborhoods elsewhere in NYC. Bars show how much more the zone's air improved than that group's, with 95% bootstrap confidence intervals. −5.0 pp −4.0 pp −3.0 pp −2.0 pp −1.0 pp +0.0 pp Citywide · −1.9 pp After background drift · −4.0 pp Neighborhood income How car-free the neighborhood is How transit-reliant the neighborhood is Lower income −2.3 pp Mid-range −2.8 pp Higher income −0.9 pp Most cars no congestion-zone neighborhoods in this group Mid-range −1.4 pp Most car-free −4.3 pp Least transit −1.1 pp Mid-range −1.8 pp Most transit −3.3 pp How much more the congestion zone's air improved (percentage points)

Each row compares a group of neighborhoods to the congestion zone. Bars show how much more the zone's air improved than that group's, with 95% bootstrap confidence intervals. Two reference lines mark the citywide average and the placebo-corrected zone effect.

Continue · Chapter 02 Methods How the satellite analysis, placebo comparison, and neighborhood stratification work.