At the start of 2025, New York City became the first US city to introduce cordon-based congestion pricing. Under the new rule, passenger vehicles pay $9 to enter NYC's central business district (CBD) between 5am–9pm on weekdays and 9am–9pm on weekends. The design of the policy raises questions about how drivers will respond. Will the new rule reduce overall congestion in the city? Or will drivers simply reroute to unpriced roads?

New research from Cody Cook (Yale), Aboudy Kreidieh (Google), Shoshana Vasserman (Stanford), and coauthors estimates the impact of the policy along a range of dimensions, including traffic (measured in terms of driving speeds), air quality, consumer spending, and foot traffic.

What We Learned

• Introducing congestion pricing made driving in the CBD about 11% faster. The gains were largest in the afternoon—historically the most congested time—and have persisted since the policy’s launch.
• Speeds also rose outside the CBD, with gains that scale with each road's exposure to CBD-bound traffic. Road segments throughout the metro area experienced traffic speed gains in proportion to their exposure to the policy, as measured by the share of cars on the segment headed to or from the CBD. The main approaches into the CBD, like the Lincoln and Holland Tunnels (where 80-100% of cars are headed to or from the CBD), sped up by 6.4% on average. Many areas predicted to absorb displaced traffic, including the Bronx and the New Jersey counties facing Manhattan, instead experienced some of the largest speed gains.
• Most of the policy's benefit to drivers goes to those who never pay the toll. To evaluate the policy’s impact on welfare, the authors compare drivers’ time savings to their cost paid. For a driver paying to enter the CBD, the toll is steep relative to the average time savings of 2.3 minutes; they are only better off if their value of travel time is at least $207/hour. However, thanks to spillovers outside the CBD, speeds on many unpriced trips also increase. While the average unpriced trip saves just 9 seconds, unpriced trips outnumber priced trips by over a hundredfold. Factoring in the time savings on these trips, the overall break-even value of time falls to $26/hour.
• Beyond traffic conditions, the policy had little-to-no effect on consumer spending, overall foot traffic, or air quality in the CBD. While estimated fuel efficiency improved, there has been no detectable effect on fine particulate matter (PM 2.5), much of which comes from non-vehicle sources (e.g., cooking gas, heating oil, and outside pollutants carried to NYC by wind).

Policy Implications

• The speed gains were permanent, not temporary. Introducing congestion pricing led to a lasting increase in speeds, both in the CBD and in other parts of the city. This is different from simply widening roads, which past research has shown tends to attract more cars until traffic is just as bad as before. A toll instead makes driving into the zone permanently more expensive, so enough drivers stay away to keep speeds above where they started.
• The gains did not come at the expense of lower-income New Yorkers. A common concern with congestion pricing is that its costs fall hardest on lower-income drivers. In this case, speeds rose in neighborhoods across the income distribution, with larger gains in lower-income tracts.
• Whether other cities would see similar gains is an open question. Whether these speed gains can be replicated elsewhere depends largely on how sensitively each city's roads respond to fewer cars. NYC's roads were operating on steep portions of their “congestion functions”—the engineering relationship between car density and speeds—where even small drops in traffic can produce outsized speed gains. The authors find that NYC's roads are more "elastic" than those in Philadelphia, Baltimore, Boston, or Atlanta; only Chicago's roads appear even more responsive. This suggests that cordon pricing may also be effective in Chicago, but how much drivers cut back in response to a toll may vary by city—for instance, there may be less switching away from cars in cities where public transit options are less readily available.
• The politics remain hard. The fact that most of the policy's gains flow to drivers who never pay the toll, while the costs fall on a visible group that pays every day, underscores a tension central to the political economy of congestion pricing. The policy’s costs are concentrated on drivers on trips to the CBD, while the benefits are spread diffusely across the many unpriced trips throughout the city.

Data and Methods

The authors use data from anonymized trips taken with Google Maps, air quality data from PurpleAir sensors, spending at shops and restaurants from a sample of credit/debit cards, and foot traffic from a sample of GPS devices. They compare changes along each of these dimensions in New York upon introduction of the policy to changes in similar cities where no congestion pricing policy was introduced. These "synthetic controls" stand in for how New York would have changed had the policy never taken effect. The authors compare New York City to these synthetic controls for roughly four months before the policy launched and eighteen months after (September 2024 through June 2026).