[AnthonyMouse]

> It's widely accepted that less congestion will encourage people to travel by car more, which will soon eat up the gains

This is not only not widely accepted but extremely misleading. The nature of "induced demand" is really that congestion suppresses demand. Which behaves in an entirely different way than the name implies.

If demand was being induced then all attempts to relieve congestion would be permanently impossible, which is nonsense. There are empirically places with less traffic congestion. Whereas if demand is being suppressed then an apparent capacity shortfall in the amount of 0.5X may really be a shortfall of X. Adding 0.5X then doesn't work, because congestion will remain until you add X, with the suppressed demand filling back in to the extent that you satisfy it.

In order to relieve the congestion you have to add enough capacity to satisfy the demand that would exist in the absence of congestion, which is more than the amount that exists now, but not an infinite amount that unconditionally consumes all possible capacity or efficiency increases.

[inigyou]

There's a related paradox - I don't remember the name - that the speed of traffic is determined by the speed of public transit. Cities with less car congestion probably have faster public transit.

[AnthonyMouse]

That doesn't seem like a paradox at all. If you have more convenient public transit then more people use it. If more people use it then there is less traffic.

This is also why things like bus lanes are ridiculous. The theory is supposed to be that making the bus faster than cars will get more people to use it. But if that actually created a benefit that exceeds its cost, i.e. got more people to take the bus than the wasted capacity of not allowing cars to use that lane when a bus isn't, then it would reduce the traffic congestion that it inherently requires. It's a thing that can't even be part of the solution to the problem because any means of actually solving the problem would eliminate its mechanism of operation.

[inigyou]

It's a paradox because making car traffic faster doesn't make car traffic faster, but making metro trains faster makes car traffic faster.

[AnthonyMouse]

Except that they both make car traffic faster, with the result that cities with slower trains on average have more car traffic.