Peter Swan and Michael Belzer don’t think much of toll-road privatization. If you hike road tolls to maximize profits, they say, then you end up with a large number of trucks taking second-best routes – something which is inefficient, unpleasant, and downright lethal in terms of increased crashes.
I understand that tolling roads has significant negative externalities in terms of the costs associated with the "free" roads that trucks use as an alternative. My issue with the paper is different: it’s with its emphasis on privatization. Privatization is a decision which is made long after the decision to toll a road, and in any event all toll-road operators have their prices heavily regulated by the local government. "If governments allow private toll road operators to maximize profits," write Swan and Belzer, and I more or less stopped taking them seriously at that point, because as far as I know there isn’t a privatized road in the world where the government has allowed the operator to maximize profits without any kind of price cap.
The decision to privatize a toll road is separate from the decision as to how high to allow tolls to rise. Once the government has decided what tolls are acceptable, it then makes the decision as to whether it’s going to continue to run the road itself, or outsource that service to a private company.
Brand-new roads, on the other hand, are a different beast entirely. Any truck traffic on a new road will take traffic away from old roads – which means that a new road is likely to have positive externalities, or at least much less in the way of negative externalities, no matter how high its tolls.
I am broadly sympathetic, however, with the conclusion that road tolling should be thought out strategically, at the level of the Department of Transportation, rather than as an ad hoc tactical method of raising money. So long as there is a "free" alternative, tolling has the effect of simply shunting costs elsewhere, and therefore on a net basis raises much less money than the optics might suggest.
Incidentally, this is in no way an argument against congestion charging in cities – in that case, there is no "free" alternative.