There’s an interesting argument going on between Yonah Freemark and Ryan Avent about road tolls. Freemark makes the usual argument, though with unusual eloquence, that implementing tolls is regressive and that the benefits of congestion pricing come at the expense of the poor. Avent responds with four points:

“One is that failing to price the congestion externality creates costs that will be eliminated by tolling; society saves money on net when congestion is priced.

Second, it may seem fair and logical to pay for transportation improvements through general revenues, but it is actually important to consider the negative effects of taxation. When you tax something you get less of it, so society is much better off taxing things it doesn’t want, like congestion, than things it does want, like income.

Third, it’s really, really easy to make tolling equitable; you just refund some of the revenue to lower income workers. It’s progressive, efficient, and probably the only way you get a plan like this passed in the first place.

And finally, it costs about $10,000 a year, on average to own and operate a car. The really struggling families out there aren’t driving, in other words; they’re taking the bus.”

All of Avent’s arguments are indisputably correct, but they’re answering the wrong questions. It’s an argument that only an economist could love.

On the first two points, Avent is using a strictly utilitarian welfare analysis. He measures society being better off by saying that there will be more total utility floating around, independent of distribution. It is quite possible that congestion pricing will fix the externality by reducing the utility of the bottom third of society and increasing the utility of the top third by the same amount plus 10%. This, for Avent, still makes “society” better off. But is Avent willing to make the claim that if somehow all utility could be redistributed from the bottom 99.5% to the top 0.5%, that would equally good for society? Is he willing to defend the strict utilitarian position that he is taking?

When Freemark challenges tolling on the grounds of equity, he does it while completely acknowledging the net social good; he writes that “It’s true, tolling highways would save “money, time, lives, and emissions.”” But that’s not good enough for Freemark, nor should it be; crude classical utilitarianism of this sort is an awfully unattractive moral stance. Distribution should matter. Thus not only is Avent talking right past Freemark—as they are employing different ethical premises—Freemark’s got the better side of the argument.

Avent does something similar with the average cost of car ownership. It’s that tricky word, average. First, I’m not sure where Avent gets the $10,000 number to begin with. The Bureau of Transportation Statistics pegs it as $8,121  in 2007. More importantly, though, when we’re talking about how much it costs low-income families to drive, the average is a distinctly not useful number. My hunch is that low-income families will pay less than the average to own and operate a car. That said, as Barbara Ehrenreich is always quick to remind us, it’s expensive being poor, and that affects automobile costs; low-income drivers pay more interest on their car loans and pay more for car insurance. But either way, the average cost of car ownership is just not a relevant number if you care about anything other than the aggregate impact.

Avent is much closer with point 3, that you could easily build in progressivity with a refund. That’s quite true. But New York’s proposed congestion pricing didn’t have a refund. London’s doesn’t have one either (though it does spend all the surplus on transportation infrastructure and has put that money to good use). So this is also a bit pie-in-the-sky. So I think that Freemark can be forgiven for maintaining skepticism; the evidence thus far is on his side.

I’d also really like to hear Avent respond to what I think is Freemark’s central contention, namely that congestion pricing will successfully shift commuters’ behavior, but that it will do so primarily by shifting the behavior of lower-income drivers. In other words, the cars that get off the road will be the lower-income cars. Doesn’t it seem very likely that low-income drivers are more elastic to the price of driving? This seems intuitive both because the value of the dollars is relatively higher for the lower-income driver but also because the value of the downtown job might be lower. You could compensate for this somewhat with a rebate, but unless you get rebated more for paying more tolls, I think that you’re still likely to see a lot of the decline in driving come from the lower end of the socioeconomic spectrum. Avent might be OK with that—after all, the environment doesn’t care whether the CO₂ came from a Lexus or an Oldsmobile. But I don’t think it’s a claim that can be ignored by those who care about transportation for reasons that go beyond the environment.

The answer, it seems, is that actually these positions are less in conflict than they seem. I imagine that Freemark would be a lot more excited about seeing congestion pricing if the new transportation bill radically reshifted funding towards transit. I also think that Freemark would admit that it depends on the city. The Socialist mayor of Paris has had remarkable success in reducing car-use while explicitly rejecting congestion pricing on equity grounds due to the specific geography of class in Paris, where the poor live on the periphery. That’s very different than here in D.C., where much of the periphery is middle- or high-income. Avent’s suggestion of a refund is spot on, though I would want to be sure that it is proportional to how often you pay the toll, as otherwise you still will have the effect of pushing the poor off the road. If the money raised goes directly into improving bus service—which can be done extremely quickly—that’s different than if the money goes into the general fund or even if it’s used to fund a new rail line.

Personally, I think that congestion pricing makes sense in a lot of different cities and I was mostly sad that New York didn’t pass it. But I really share Freemark’s concerns and think that if implemented in such a way that its costs remain regressive and there isn’t immediate simultaneous improvement in the transit infrastructure, it is an attack on the mobility of low-income families. That’s not something I’m comfortable with. More tolls, but only if more better tolls.