I haven't yet read Gernot Wagner's new book But Will The Planet Notice: How Smart Economics Can Save the World, but I heard him interviewed by Leonard Lopate, and the approach he's promoting seems very sensible. If there is a major economic incentive encouraging people to do negative things (drive, accept plastic bags, buy food sweetened with high fructose corn syrup), is it easier to fight that behavior directly, or to change the incentive?
We can think of this as an application of the Tao, or of Ueshiba's notion of enveloping your adversaries. If you prefer, you can think of it as leverage: if someone is using a lever to magnify their force on an object, is it better to push back on the object, or on the other person's foot? You can also think of it as getting past the superficial story to the real story underneath. The best kind of compromise is when it doesn't matter whether anyone gets what they said they want, but everyone is getting what they really want.
One example of incentives working in transit is Hasselt, Belgium, which is usually held up as a paragon of free public transit, but where it seems that the key was actually converting the government-sponsored inner ring road from an incentive to drive to a "Groene Boulevard" where buses and bikes have priority. However, in order to pass both free transit and the Groene Boulevard, their promoters had to convince the citizens of Hasselt that it was "their" town, their mobility plan, and their bus system. Getting to the levers is not easy, and neither is controlling them once you do.