> We study the impacts of New York City’s Central Business District (CBD) Tolling Program, the first cordon-based congestion pricing scheme in the United States. Using a generalized synthetic controls design, we find that the policy increased speeds on CBD roads by 11%, with little-to-no effect on air quality, transactions at shops and restaurants, or overall foot traffic in the CBD. Speeds also increased on roads outside the CBD that are commonly traversed by drivers headed to/from the CBD. These spillovers lead to faster trips throughout the metro area, including for many unpriced trips. We develop a simple model to bound the driver welfare effects, and estimate gains of at least $14.3 million/week (before any revenue recycling). These gains are largely driven by diffuse time savings for the many unpriced trips outside the CBD, highlighting the importance of accounting for network-wide spillovers. Finally, we show how characteristics of local travel patterns and road networks can inform the potential impacts of introducing cordon-based congestion pricing in other cities.
> To understand why congestion pricing’s total time savings mostly accrued to those traveling outside Manhattan, consider that most drivers heading into the island traverse roadways outside the congestion relief zone as they approach it from Long Island, New Jersey or wherever they began their journey. By shrinking the number of peak-time cars flowing into the toll zone, congestion pricing reduces traffic on outlying roadways, where remaining drivers — including those who never had Manhattan on their itinerary — can now go faster.
[…]
> Better yet, the coauthors wrote that they found “no evidence of offsetting slowdowns on different road types … suggesting that the policy reduced overall traffic volumes rather than simply displaced congestion.” That conclusion refutes critics’ predictions[1] that congestion pricing would shift drivers toward suburban roadways that would slow to a crawl.
Paper:
> We study the impacts of New York City’s Central Business District (CBD) Tolling Program, the first cordon-based congestion pricing scheme in the United States. Using a generalized synthetic controls design, we find that the policy increased speeds on CBD roads by 11%, with little-to-no effect on air quality, transactions at shops and restaurants, or overall foot traffic in the CBD. Speeds also increased on roads outside the CBD that are commonly traversed by drivers headed to/from the CBD. These spillovers lead to faster trips throughout the metro area, including for many unpriced trips. We develop a simple model to bound the driver welfare effects, and estimate gains of at least $14.3 million/week (before any revenue recycling). These gains are largely driven by diffuse time savings for the many unpriced trips outside the CBD, highlighting the importance of accounting for network-wide spillovers. Finally, we show how characteristics of local travel patterns and road networks can inform the potential impacts of introducing cordon-based congestion pricing in other cities.
* https://www.nber.org/papers/w33584
Seems fairly 'obvious' in hindsight:
> To understand why congestion pricing’s total time savings mostly accrued to those traveling outside Manhattan, consider that most drivers heading into the island traverse roadways outside the congestion relief zone as they approach it from Long Island, New Jersey or wherever they began their journey. By shrinking the number of peak-time cars flowing into the toll zone, congestion pricing reduces traffic on outlying roadways, where remaining drivers — including those who never had Manhattan on their itinerary — can now go faster.
[…]
> Better yet, the coauthors wrote that they found “no evidence of offsetting slowdowns on different road types … suggesting that the policy reduced overall traffic volumes rather than simply displaced congestion.” That conclusion refutes critics’ predictions[1] that congestion pricing would shift drivers toward suburban roadways that would slow to a crawl.
https://archive.is/https://www.bloomberg.com/news/articles/2...
https://archive.is/dhUrE