A few years back, the government of Ontario twinned the highway 115 which leads from the highway 401 to Peterborough. Now, at least a decade later, the 115 remains a smoothly flowing highway which is rarely subject to traffic jams (as far as I know, anyway). The reason is that there are only so many people who want to go to Peterborough and vice-versa. This number grows with time, but relatively slowly so the highway remains relatively clear.
Now consider the highway 407 built across the top of Toronto roughly around the same time, 6 full new lanes of traffic in each direction. Only ever increasing tolls keep the 407 from becoming as busy and packed as the 401 itself.
The difference is that in Toronto there was a huge latent demand of people who would drive more if only the roads weren't so busy. People who would move further out into suburbia if the commute was going to stay the same length. People who would visit friends across town more often if it didn't take as long. People who would play in a hockey league across town if the drive was shorter and on and on. Because of all this latent demand, building new roads in a big city doesn't lead to any less congestion - as the last few decades have shown clearly. And this latent demand is so enormous that no amount of new road building could ever really bring an end to congestion on city roads.
Not to say that new roads don't have a benefit. If people can live a little further from work or can play in a better hockey league or whatever, those are real benefits, I'm just saying that new roads won't reduce traffic congestion.
The result is that in a place like Peterborough, the level of traffic congestion is a function of population and road building and mechanical 'plannable' factors, but in a place like Toronto, the level of traffic congestion is a function of what people are willing to put up with. The roads just fill and fill until those who are least tolerant of the congestion either move closer to work, move to a different city or take transit.
And this last point is an important one. Given that the upper limit on traffic in a big city is what people are willing to put up with, the upper limit on traffic in a city could be seen as a function of the effectiveness of its transit system. The better the alternative to driving, the less congestion on the roads people will put up with - so a better transit system might actually provide the biggest benefits to those who drive - on roads blessedly free from all the people who took transit instead.
Of course, this effect would be pretty weak when we are talking about buses and streetcars because, as congestion increases, these options slow down just like car traffic does so they aren't much of an alternative.
Anyway, like I said, I'm not an urban planner, and I don't know if this idea is considered nonsense or conventional wisdom or what among those who actually have some expertise in this area, but it makes intuitive sense to me, I'd be interested to hear what others think/have heard.
Controlling for things like total population size, do we find that roads are less congested in cities with more effective alternate transit options? Within cities, are the roads less busy in corridors for which there are good transit options than in corridors for which transit is less effective?