They have special rights of way

... which come at the expense of taking away existing universal-use lanes, worsening all other traffic congestion.

Which then gets people on the bus, easing the congestion. This is a well documented effect, where the opposite happens, adding more lanes increases traffic. https://en.wikipedia.org/wiki/Induced_demand#Reduced_demand_...

  a well documented effect

Not really. Had the "road diet" actually resulted in a shift from cars to transit, the transit agencies would show a corresponding increase in their census and farebox return. They didn't.

Okay, you're right about the Wikipedia article not showing an increase in those statistics for mass transit. However, this paper does show a substitution effect for busses: http://clinlawell.dyson.cornell.edu/InTech_public_transit_ch...

However, there is an induced demand effect where because there are less cars, more people will drive. So in effect an extra lane appears, empty of the people riding the bus.

Short-term, that is correct.

But long-term the maintenance is significantly higher, as you need to replace the road surface very frequently.

A usual bus stop, or a BRT lane, has to be replaced every 6 months if it’s asphalt, and every 18 to 24 months if it’s cobblestone. That’s a significant expense.

While you can build and use light rail for decades.

What about concrete? US interstates can go decades with high truck traffic if they are concrete.

Pure concrete leads to significant shifts between the plates over the years, leading to the well-known bump every time you pass between plates. This obviously affects the bus’s suspension, and passengers, negatively.

Concrete with asphalt on top of it has the issue of the asphalt having to be replaced frequently.

I’m not sure if US interstates with concrete are different, but these were the only options discussed here in Germany.