THE DESIRE NAMED STREETCAR: in “Why You Should Think Twice Before Building a Rail Transit System,” Aaron Renn of the Manhattan Institute notes:

Metro has a large unfunded maintenance liability. This doesn’t surprise us because we expect American transit systems to have a backlog.

The difference is that unlike NYC, Chicago, Boston, etc., which have systems a century old, the Washington Metro system is actually new.

The oldest part of the Metro opened in 1976. That means Metro is 40 years old – max. Much of it is actually newer than that.

Forty years after opening, Metro already faces a maintenance crisis.

This should give other regions pause when it comes to building a rail transit system. My colleague Alex Armlovich points out that NYC has more or less been on a 40 year refresh cycle, with two rounds of major system investment since the subways opened. This doesn’t seem out of line as a capital life heuristic to me.

So cities need to keep in mind that if they build a rail system, they not only have to pay to build it, they pretty much have to pay to rebuild it every 40 years. This is a challenge because as we see it’s easier to muster the will to build something new than to maintain something you already have.

Given the huge permanent capital outlays implied by rail transit, you only want to build it where there’s sufficient value to justify it. Washington unquestionably achieves this. It simply hasn’t been able to capture the value into a maintenance revenue stream, plus Metro (like many systems) has been badly mismanaged.

Sadly, given the enormous potential for graft, union construction handouts, and featherbedding, most city planners view the prospect of huge capital outlays on a regular basis being spent on a rail transit system’s infrastructure and rolling stock as a hugely desirable feature, not a bug. Which is the very same reason they prefer light rail over simply buying more busses as needed, despite the fact that the busses can run anywhere, and their routes can be easily changed as circumstances demand.