Last January, New York City launched a bold traffic congestion pricing program, the first of its kind in the United States. Drivers of personal cars entering the newly specified “congestion relief zone” – the part of Manhattan south of 60th Street – must pay $9 each day they enter the zone.

Note a key difference between this program and tolled express lanes that are becoming increasingly common across the United States in high-traffic areas: there is no toll-free alternative to driving in the zone. This lack of a gratis option has made the New York program controversial, to say the least. It was challenged in a flurry of lawsuits, many of which are still pending. And less than a month after President Trump was sworn in for a second term, his administration moved to rescind federal approval of the program, which had been granted in the final months of the Biden administration, echoing the sore point that it “leaves drivers without any free highway alternative.”

The New York Metropolitan Transportation Authority, which administers the CRZ and is counting on the revenues to anchor $15 billion in bonds for capital improvements for mass transit, promptly went to court to block the administration’s move and refused to suspend or pause the program. The various challenges to the CRZ are currently progressing through the legal system, but the MTA is not giving ground. “We’ve been sued, as I say, in every courthouse east of the Mississippi,” explained Janno Lieber, the authority’s CEO. “It’s pretty clear that they are all going to lose.”

Actually, the Trump administration tacitly agrees. An internal memo leaked from the Department of Justice cautions that its case against the CRZ is unlikely to convince the courts. So the program is here to stay, at least for a while. And the big questions now: How well is it working? Are there changes that could make it work better?

Devils and Details

The program was initially slated to launch in June 2024. Drivers of personal cars were to be charged $15 electronically when they entered the CRZ during peak hours (5 a.m. to 9 p.m. weekdays and 9 a.m. to 9 p.m. on weekends), and a quarter that amount for entry off-peak. Larger private vehicles would be charged $24 or $36, depending on their size (with off-peak entry rates similarly set at 25 percent the peak rate). Meanwhile, drivers of taxis and app-based, for-hire vehicles (FHVs) like Ubers and Lyfts would be exempt from the direct $15 charge; instead, their passengers would pay surcharges per ride of $1.25 for taxis and $2.50 for FHVs, 24 hours a day, on any trip to, from or within the CRZ.

This fee, as explained by the Traffic Mobility Review Board (the five-member citizen board created by the legislature to recommend a tolling structure), was set by dividing $15 by the estimated average number of trips that hired cars made in Manhattan’s central business district before congestion pricing went into effect. Whatever the rationale, though, reality is a bit messier since taxis already paid a $2.50 per ride ($2.75 for for-hire services) congestion fee for operating below 96th Street in Manhattan. Hence the new effective per-ride congestion fee was originally set at $3.75 for taxis and $5.25 for ride-shares.

All pricing details were set, and all infrastructure was ready to go, but under pressure from a variety of interests – suburbanites who commute by car, Manhattan retailers worried that they would lose shoppers, taxi-owner lobbies, Uber and Lyft, Congressional Democrats wary of alienating suburban voters – New York Governor Kathy Hochul “indefinitely paused” implementation just before the planned launch.

After the November election when the plan was revived, the peak fee was cut to $9 for personal vehicles, $14.40 for small trucks and $21.60 for larger trucks to make the medicine go down a little easier. Meanwhile, 50 cents per trip was clipped off the taxi fee and $1.25 from the FHV fee. The idea now is to raise the fees in 2028 (to $12 for personal cars), and then again in 2031 (to $15 for personal cars).

Congestion Questions

While the design of the plan looks straightforward and intuitive, in our view it has a major shortcoming: it charges different categories of road users different amounts – in some cases, dramatically different amounts – per “unit of congestion” that they are likely to generate.

Commuters who enter the congestion zone in the morning, park and then drive home later in the day in effect pay $4.50 per trip (the $9.00 daily fee spread over two trips). They are typically on the street adding to congestion within the CRZ for perhaps an hour a day. Meanwhile, taxi, Uber and Lyft passengers pay a lot less per trip, especially if one takes into account the fact that approximately half the time, they drive around empty and create congestion without paying anything at all for it. And an Amazon, FreshDirect or FedEx delivery vehicle that enters the tolled zone before 5 a.m. pays just the off-peak rate of $3.60 (25 percent of the $14.40 peak rate for package delivery trucks) but can still spend a full workday (sometimes more than 100 stops) delivering packages, thereby causing vastly more congestion than a single commuter.

These differences between “per unit of congestion” prices for different types of travelers are problematic for two reasons. First, they seem unfair. If the primary purpose of congestion pricing is to reduce travel time and pollution at minimal cost to travelers, why should, say, a nurse from Queens driving to work in Manhattan pay substantially more than a business traveler taking a taxi there from LaGuardia, or orders of magnitude more than the individual beneficiaries of deliveries from a single Amazon truck’s meandering route through the congestion zone?

Second, the structure of the congestion fees is unlikely to be effective at solving the traffic problem within the CRZ. The fees hit private cars the hardest, raising the cost of entering the congestion zone from nothing to $9. But the Traffic Mobility Review Board reports that private autos and motorcycles represent only 35 percent of traffic in the CRZ. Meanwhile, taxis and FHVs represent 52 percent of traffic – and the minor increases in their total prices are unlikely to lead to a substantial reduction in demand. Similarly, delivery vehicles, which also cause a large amount of congestion, are unlikely to be deterred from entering the CRZ because the fees are spread over many, many deliveries.

The possibility that the impact of New York’s congestion pricing plan on traffic within the CRZ will be modest is consistent with the experience of London’s congestion pricing scheme. London now charges the equivalent of about $20 per day for private cars. But it charges FHVs only once per day and exempts taxis from the tolls entirely, thus making the average amount collected per unit of congestion minimal and ineffective.

Not surprisingly, then, London still suffers from heavy traffic congestion. Indeed, a 2024 study by the navigation company TomTom awarded London the dubious honor of being the most congested city in the world (out of 387 for which TomTom has collected traffic data). Of course, it is possible that without congestion pricing, London traffic would be even worse. But it is nevertheless safe to say that the scheme has not lived up to expectations that it would go a long way toward solving London’s congestion woes.

What’s Happened in New York

Now that a few months have passed since the launch of congestion pricing in Manhattan, we can track the initial impact. The results are encouraging, at least on some dimensions. But at the same time, they confirm that the troubling design issues we highlighted above are very much present.

Specifically, congestion pricing’s impact on traffic entering the CRZ is substantial. Crossing times are down 48 percent at the Holland Tunnel connecting New Jersey, and 30 percent at both the Williamsburg Bridge (connecting Brooklyn) and the Queensboro Bridge (connecting Queens). By contrast, the gains within the CRZ seem modest: bus speeds are up just 4 percent since the introduction of congestion pricing – an improvement, but hardly a dramatic one.

These observations are consistent with those reported in an independent, recently released Stanford University working paper, “The Short-Run Effects of Congestion Pricing in New York City.” The authors confirmed that speeds on the bridges and tunnels that constitute entry points to the CRZ are up as hoped, but travel times on trips within the CRZ have decreased by only 8 percent – not a trivial difference, but less than one might hope for a place where it is often still faster to walk than drive at rush hour. Incidentally, those who are really, really curious can slice and dice the data to their hearts’ content on the Congestion Pricing Tracker website, which compares travel times on various routes through the city across the days of the week and hours of the day before and after the January 2025 imposition of congestion pricing.

Path Forward

Fortunately, having identified this problem, it would be relatively straightforward to address it. Rates are currently scheduled to go up in lockstep in 2028 (to $12 for personal cars) and then to $15 in 2031. Our calculations do suggest a better alternative.

We present the details in a working paper, but they can be summarized succinctly. Instead of raising the rate for private vehicles, we think it should be kept at $9. Meanwhile, the surcharges on taxis and FHVs should be raised to the level that roughly equalizes the CRZ fees they pay per unit of congestion with the fees that private vehicles are paying. Note that occupancy data show that for every paid taxi and FHV trip, the vehicles spend a comparable amount of time cruising empty and continuing to create congestion. So if the surcharge is to be collected from paying passengers, it needs to cover both that trip itself and the corresponding empty trip.

Setting congestion fees on vehicle size (small or large), typical vehicle usage (passenger or truck) or paid trips (taxis, etc.) will never exactly mesh with the goal of setting fees on a per unit of congestion basis. But the flat fee is an especially bad fit for delivery vehicles that may make one stop or one hundred in the congestion zone. A better approximation would be to shift to truck fees per mile, adjusted for vehicle size. And happily, that approach is practical since most fleet delivery vehicles (Amazon, FedEx, etc.) are already tracked by their companies in real time, so new infrastructure would not be needed to record their routes in the congestion zone. Our plan would set the fee at $2.25 per mile, adjusted for vehicle size.

Putting together these numbers, our revisions would generate slightly higher revenues than the current long-term plan. Yet it might be more palatable politically because (a) it is more equitable, since it charges sources of traffic congestion in proportion to the amount of congestion they generate, and (b) it shifts some of the burden from personal to business vehicles, instead of overcharging the former while undercharging the latter. The table below summarizes our estimates.

* * *

The launch of congestion pricing in New York City was a monumental achievement, capping decades of hard work by economists, policymakers and staff at various levels of government. It is already showing signs of what’s possible. We hope it survives the current legal challenges, continues to improve traffic in New York and eventually serves as a model for other congested metros around the world. At the same time, whenever a revolutionary economic mechanism is implemented, some shortcomings usually become evident – real-world progress is often iterative.

Congestion pricing in New York is no exception: the early experience points to some clear areas for improvement. Fortunately, these shortcomings can be addressed before the next version of the pricing schedule is due to be finalized for 2028. We are optimistic that with the appropriate future adjustments, it will go down in history as a spectacular example of economic policy innovation.

Loose Ends

One possible concern with the higher surcharge levels on taxis and FHVs – both as a matter of politics and equity – is the negative impact on drivers’ incomes. For example, an August 2023 letter from 25 New York politicians (including leading mayoral candidate Zohran Mamdani) calls for exempting taxi and FHV drivers from any direct congestion fees. We agree – and in fact that is what the current scheme does. However, congestion fees on the passengers, both of taxicabs and FHVs, need to be substantial, commensurate with the negative externalities that they create on New York traffic, and will need to go up in the future if the societal value of congestion pricing is to be maximized.

It’s important to emphasize that the impact on ride-share drivers would be limited. First, only about a third of Uber and Lyft trips in New York City start or end (or both) in the Congestion Relief Zone, and of those, only about three-quarters take place during peak hours. Thus, only a quarter of all FHV trips would be affected by the surcharge. Moreover, with trips taking less time (because of reduced traffic congestion), each affected trip would be faster and cheaper (net of the congestion fee), further moderating the negative effect of congestion surcharges on consumer demand and drivers’ earnings.

The demand is more concentrated for taxis in Manhattan. But any potential reduction in demand would again be partially offset by the fact that with reduced congestion, trips would be faster. In fact, early evidence suggests that the demand for taxis in the CRZ has risen since the introduction of the program in January.

Another important piece of the puzzle is the impact on delivery vehicles. Crucially, many deliveries can be made during off-peak hours. In fact, shifting deliveries away from peak times has long been a goal of policymakers in New York City, who have tried various policies to encourage this shift. It is thus critical to keep congestion tolls on delivery vehicles at or near zero during off-peak hours to maximize their incentives to undertake this shift. Remember, too, that some of the additional revenues from increased surcharges during peak hours can go toward expanding the existing programs that help minimize the frictions associated with the shift in delivery times.