Reuters reports on the view of Fitch ratings of different types of tolling as an investment risk, in the UK context (but with some applicability more generally. It gives a useful insight into how investors view toll road projects, in separating them between those that offer new routes (and so offer a consistent saving and become the long term preferred route) and those that bypass congestion (and are dependent upon conditions on untolled route):
We would expect toll roads built to relieve congestion to have a weaker credit profile than toll roads built to service a new route. As "peaker facilities," they benefit more from rises and suffer more from falls in traffic volumes. At times of economic growth, when overall traffic volumes in a particular transport corridor rise, growth on the free road is constrained by congestion. Most of the growth is therefore captured by the toll road. When the economy weakens and overall volumes fall, toll road traffic in turn experiences a greater contraction. Toll-free travel may also be more attractive in a downturn.
This volatility, which we call the "corridor leverage effect", has been seen for example in Spain, where a relatively large number of toll roads have competition.
In the managed lane system that has been developed in the U.S., drivers can choose between toll and free roads in the same transport corridor based on real-time estimated travel time information and dynamic pricing. Managed lanes feature higher revenue volatility. We therefore only see managed lanes as viable in heavily congested areas.
I am more bearish than that about managed lanes, as the congestion needs to not only be severe, but consistent over long periods of the day to help make such lanes viable. It is notable that toll lanes have only been seen in the USA and Israel so far, and the former in most cases simply to get more utilisation out of poorly used HOV lanes.