Setting the right toll
We in transport economics generally say roads are underpriced and travel is heavily subsidized by society at large. As a result, we over-consume travel. The solution is raising the price of travel.
One way of raising the price of travel is tolling new roads rather than letting them be free.
Scenario: A privately owned toll road or lane in the context of free roads will set profit maximizing toll, subject to regulation. Because the road would not have been built but for the private investors, this regulation will typically have been negotiated in the franchise agreement to be a light touch, so tolls will be close to profit maximizing. Since this is not a competitive market as taught in Econ 101, profit maximizing tolls are not equal to welfare maximizing, and instead will likely be too high.
Given that drivers have alternative roads, tolling the road too high will thus drive vehicles away from this road and onto other roads, which will be more congested (and polluted) than necessary. Total travel overall will be higher than without this road. However on competing routes, travel will be slightly lower (recognizing induced demand eats up some of the capacity, though consumer surplus should still rise), than if the road did not exist, so this capacity, compared to its absence generally improves the transport welfare (consumers surplus) of travelers. However, this priced capacity compared to it being free, reduces the transport welfare of travelers.
Scenario: All roads tolled (these might be virtual tolls, like mileage based user fees, rather than traditional tollbooth or gantry style tolls) can only be achieved with the cooperation of the road owners, almost universally involving some level of government. While the system might be administered under contract, it will undoubtedly be at worst regulated like a monopolistic utility, with cost plus rate-of-return rates. The costs might include externalities, and not just direct costs, or maybe the externalities will be taxed separately (carbon taxes at refineries, e.g.; mandatory car insurance). These tolls will be close to welfare maximizing after covering costs (though the welfare maximizing toll may not cover costs, depending on the shape of the cost function and whether the road is congested).
Based on their history, individual toll roads are not an especially good business to invest in. They are high risk and the forecasts cannot be trusted. New toll roads have tended to over-estimate demand and under-estimate costs. Eventually they need to pay their creditors, and the investors get wiped out. There are lots of reasons for this, but pity not the toll road builders, they do it to themselves.
Still, the key point is that tolling some of the roads may very well be worse than either tolling none of the roads or all of the roads. It introduces distortions, may result in new underutilized capacity in an environment where other roads remain congested, and discredits the premise of pricing in the mind of politicians and voters. While incrementalism is generally a good strategy, the steps need to be in the right direction (for instance, tolling all of the roads some may be better than tolling some of the roads a lot and others not at all).