New York's congestion pricing disaster

And the larger crisis of Deep Blue governance

Photo by Spencer Platt/Getty Images)

I’m both very frustrated with Kathy Hochul’s efforts to kill New York City congestion pricing at the last minute and also sort of weirdly sympathetic to her.

At some level, she’s just trying to help House Democrats win races on Long Island, a worthy and important goal. The congestion pricing plan is unpopular there, and she is prioritizing policy moderation and winning over idealistic pursuit of a niche interest group goal. But, as someone who really likes congestion pricing, I’m annoyed that transportation reformers’ ox is the one getting gored here. Why did New York Democrats sink so much political capital into a bail reform program and other soft on crime measures that, even in the absolute best case scenario, had minimal upside for most people’s lives?

Congestion pricing was always a political risk. We know from London, Stockholm, and Oslo that voters resist the idea of paying for something that they’re accustomed to getting for free. But we also know from those cities that once congestion pricing is in place, it’s proved reasonably popular and politically durable. In that sense, it’s the opposite of bail reform: a known pain point that actually does make most people’s lives better and that comes to be a lasting achievement.

The other frustrating aspect is that in New York, a lot of time and money was sunk into winning Democrats a friendlier set of state court judges who would give them a freer hand in redistricting. But once that work was done, they didn’t use it to enact an aggressive gerrymander that would have eliminated these Long Island political concerns. Instead, they shored up a few incumbents and went home.

Now, New York is not addressing its major housing problems or transportation problems, and it’s had to U-turn on criminal justice reform thanks to a state leadership that can’t set priorities or get things done. That said, what’s happened has happened and Hochul doesn’t have a time machine, so even though I’m mad, I do see where she’s coming from. And I want to talk about some broader issues here.

Where congestion pricing is uncontroversial

Just outside of DC, in Northern Virginia, they have a lot of congestion pricing. Except they don’t call it that. Instead, they say that various NoVa highways have added “express lanes” that you can access in exchange for a toll that varies according to the time of day. This is a totally uncontroversial policy to the point where I can’t even tell whether people support it.

Because nobody really talks about it. It’s just a good, pragmatic solution to the endless dilemma of public budgeting — everyone driving on the road wishes someone would build a bigger road, but nobody driving on the road wants to pay higher taxes. So instead, they sold bonds to build the extra lanes and are paying off the loan with toll revenue from the new lanes. As an individual, the absolute worst case scenario under this plan is that you never drive on the new toll lanes but receive some minor benefit from the fact that other people use them. More likely, you (like me) occasionally use the toll lane and exchange a bit of money for a faster drive. Many others commute on those roads daily, and I’m sure some fraction of those commuters are regular express lane users. They, by definition, have self-selected into paying more for a faster ride.

Everyone is happy except for anti-sprawl ideologues who are angry that improved road capacity generates “induced demand” and more people commuting from further flung Virginia exurbs.

This is probably true, but from a normal person’s point of view, it’s part of the reason for doing it. Adding transportation infrastructure unlocks more abundant Virginia housing and powers the state’s economic growth. I think it’s a bad idea for states to make additional highways and induced demand their only strategy for improving housing abundance and transportation infrastructure — Virginia should also make it easier to build infill housing in already developed areas, especially near commercial corridors and job centers and near WMATA and VRE stations. But there’s nothing wrong with making the state’s highway infrastructure better, and congestion pricing is a logical way to do it.

The reason this works, though, is because Virginia had highway right-of-ways with enough extra space to build the express lanes. The construction projects involved weren’t trivial and did cost a lot of money, but they were feasible, and the scope of the project was limited by cost and logistical considerations.

Often, a highway has normal lanes running in both directions, and then a single pair of reversible express lanes in the middle, with the express lane available in the peak commuting direction.¹ This is not the case in New York, which does not have a logistically feasible option to double the size of the Manhattan street grid and then make the new lanes toll-only. For congestion pricing to work, they can’t just add new paid options, they need to eliminate existing free options. That’s naturally more controversial.

Congestion and uncertainty

Every debate about congestion pricing in New York’s “take something away from people” mode features the following two perspectives:

1. Someone arguing that traffic congestion is bad, and congestion pricing is a well-known way to address it and also raise revenue.

2. Someone else insisting that the elasticity of driving to Manhattan with respect to price is essentially zero, that everyone will reduce their consumption of everything that is not “driving to Manhattan” and congestion will remain the same while the government secures more tax revenue.

It’s challenging to prove who’s right about this. I think that if you really push someone espousing the second argument, he will probably admit that the elasticity is not actually zero, that at least one person on any given day will switch to riding the LIRR, that someone else will carpool with a buddy, and that a third person will decide to skip the trip altogether.

But that’s only three people! Three people aren’t going to transform the congestion situation. And the person making the first argument has to admit that it’s difficult to know exactly what the elasticity of driving behavior with respect to price is. How much faster will traffic actually become?

When Jed Kolko wrote his piece for Slow Boring on what kinds of research are helpful to policymakers, one veteran of a prior administration reached out to say that when he was doing economic policy, he was constantly wishing for more papers estimating elasticities. I think that kind of work often strikes academics as a little bit boring, since there’s nothing particularly interesting from a theoretical standpoint about the question of how exactly drivers respond to tolls. Obviously, they drive somewhat less. How much less? Well, it depends on how high the toll is and what the alternate options are. Just running study after study after study to try to nail down the exact parameters more precisely isn’t very interesting and isn’t going to make you famous.

But as a policymaker, it would be really nice to have better estimates. Raising the alcohol tax will generate additional tax revenue and also cause people to drink less. But those two goals are somewhat at odds with each other. If the elasticity is high, the tax will have a large public health impact (people drink less) but not raise that much revenue (people are abstaining rather than paying the tax). If the elasticity is low, you have a really good source of tax revenue but only a marginal public health benefit. As a hot take guy, I think it’s easy to say this is good policy either way, but if I were a state legislator considering a bill, I would want to know what my bill is accomplishing. That would change how I talk about it, who I try to partner with on it, and even how much effort I put into it. I’d also want to know about other parameters. California and Texas are large states with population centers far from state borders, so I doubt alcohol taxes there generate a lot of “leakage” where people cross borders and buy cheap booze out of state. But a state like Pennsylvania should probably worry about this a lot.

With congestion pricing, not only are there fewer elasticity studies than you’d ideally like, but the reality is that each city is different. Behavior in London or Stockholm tells you something about behavior in New York. But these are different places with different geographies. Any charge you levy is going to be something of a guesstimate, the very existence of which will give you much better data about the parameters, and then you’re going to want to adjust. So the question then becomes largely one of trust — do you think the powers that be will impose a charge, see what happens, and then optimize based on better data? Or do you think they’re scumbags?

The problem of social trust

American politics has become more polarized along lines of social trust in recent years, with higher-trust people becoming Democrats and lower-trust people becoming Republicans. This is often misportrayed by conservatives in the discourse as a question of people with lower trust in elite institutions becoming Republicans, because conservatives think that mistrust is warranted. But as Eric Levitz explored recently, we are seeing partisan realignment about interpersonal trust as measured in questions like “generally speaking, would you say that most people can be trusted or that you can’t be too careful in dealing with people?”

And if you don’t trust your neighbors or coworkers, then you are probably not going to trust impersonal institutions like the MTA or the State of New York.

And while I think there’s a set of anti-system leftists who believe that spreading messages of despair and corruption will drive people toward embracing progressive political radicalism, the New York case reflects the reality that low trust drives people to small-c conservatism. The question of congestion pricing is less “would I be better off with a $15 fee and less traffic in Manhattan” and more “when the initial version of this isn’t quite right, do I think the people in charge will respond in a principled and data-driven way to make it better?”

To an extent, Hochul is just a victim of larger social trends that are out of her control. But I have, from the beginning, thought that the specific design of this initiative is poorly suited to deal with the social trust problem. The most successful congestion charge programs, in Stockholm and Singapore, are implemented by governments with a strong reputation for public sector performance. They were also accompanied by significant new visible infrastructure investments, that in Stockholm’s case included new road investments as well as new transit. In New York, by contrast, the basic idea has been to use the congestion tax revenue to plug a hole in the MTA’s budget. If the MTA were an agency with a strong reputation for transparency and delivering projects with a high ratio of benefits to costs, people might feel excited about championing all the great new things the MTA is going to build.

The actually existing MTA is just the opposite.

Last January, for example, they announced plans to spend $3 billion on excessively heavy, excessively slow rolling stock for Metro North and LIRR commuter rail. Advocates put a fair amount of effort over the years into getting the federal government to allow American transit agencies to buy lighter, cheaper, faster European trains, and we won what we were asking for. And then the most important transit agency in the country just didn’t do it. The article about that I linked to appeared in the New York Post, a right-wing newspaper that doesn’t really favor better mass transit and that’s spearheaded a sort of braindead opposition to congestion pricing. They just like to point out the MTA’s flaws to bring down trust in the state and advance rightwing politics. But liberals sticking their heads in the sand and failing to reform flawed agencies isn’t working. People do not believe the MTA will make smart decisions or put money to good use, and those suspicions are warranted. Note in particular that as you adjust a congestion fee and learn more about elasticity, you will face certain decisions about maximizing revenue versus maximizing congestion impact. Which is the MTA going to prioritize?

Blue state blues

The proximate problem for Hochul is that spiking the congestion fee does not address the MTA budget hole.

Whatever the political downsides to congestion pricing, it’s not like there’s some other tax that everyone’s going to love. Meanwhile, as a legal matter, it seems like MTA bondholders may be able to insist that the law creating the congestion tax be implemented. For all that I’ve said here both to sympathize with Hochul’s political calculations and for skeptics of the congestion pricing plan, I think the only way out is through at this point. It would have been better to reform the MTA to make it a more credible agency, to pitch this as a congestion measure rather than a revenue measure, and to try to avoid the toxic cars-versus-transit culture war space in which this policy has landed. It also would have been much better for New York Democrats to be bolder on transportation and redistricting and more cautious on crime.

But we are where we are.

A larger issue here is that I continue to think most Democrats are excessively blasé about the crisis in blue state governance quality. Precisely because New York City and California feature such a high cost of living, they are populated almost exclusively by extreme outliers in terms of the value placed on the unique lifestyle amenities provided by those places. But when New Yorkers tell me they could never move to the suburbs of Raleigh, they’re not saying they would miss the high quality public services provided by their state government. New York State has more people than the Netherlands, a GDP per capita nearly as high as Qatar, and the same combined top state and federal income tax rate as Denmark. What are residents getting for that?

The upshot is that while I don’t disagree with the folks dumping on Hochul, I think everyone reducing this to a question of her decision-making or the pressure from Hakeem Jeffries or Chuck Schumer are missing the forest for the trees.

New York has significant governance challenges, and I don’t think anybody really believes that the people leading the state are trying to address those challenges. Can the answer really be that New York’s main problem is that taxes are too low and the big question is which tax should be raised?

When I think about the need for moderate partisan factionalism on the Democratic side, I think there are two wings to that which are distinct but complementary. On the one hand, Democrats need to be able to compete for statewide elections in places like Ohio, Iowa, Florida, Texas, and North Carolina in order to build national governing majorities. But they also need to be able to show that states like Massachusetts, Maryland, New York, and California are effective models of governance that can deliver good things. That means a style of blue state moderate politics that is oriented toward actually fixing the problems of the blue states, rather than just a politics of merely saying “no” to the wildest leftist schemes. Call it “abundance” or call it something else, but there needs to be a real push for growth and reform that starts with the premise that something is badly wrong with the model.

1. It would be twice as expensive (or more) to add express lanes in both directions, and you would raise way less than twice the revenue, so they don’t do it.