On the Ethan Allen Express (Amtrak) I took this week, the boarding steps to the cars had to be manually deployed by train staff, along with a little step stool. When I got on, there were two people doing this, so only two train cars were boardable.
I think non-Americans underestimate our ability to not automate things that can clearly be automated through some combination of of inertia, union power, and sheer incompetence.
It’s because Amtrak basically doesn’t matter. What’s amazing about this story is the ability of these make-work policies to survive in one of the most demanding urban transport systems. NYC baby.
One reason is that it's already so expensive to operate a regular train, that the expense of having one employee (or even two) isn't as significant compared to a individual transportation. Paying the taxi driver is a significant part of the cost of a taxi trip. Paying the train operator isn't a significant part of the cost a train ride.
Edit: The article claims the opposite, and maybe that's true in NYC? I did find a breakdown of costs in Germany, for a municipal light rail service: operating the train is 1860 EUR per journey overall, paying the people operating the train (one operator, possibly one conductor) is 350 EUR of that. That ratio is smaller than I would've guessed, but it's not a majority.
It depends on whether you calculate it as including the amortized fixed costs (e.g. the cost of building the tunnels) or the incremental cost (what does it cost to have one additional passenger). If it's the first one then the cost is way higher, but then you'd have to do the same thing in the other case and include the cost of building roads etc.
However, fixed costs are better funded by general taxes than by usage fees because otherwise you pay a huge fixed cost to build something with a low incremental usage cost and then under-utilize it because recovering the sunk cost through fares causes high fares which deters uses whose value exceeds the incremental cost.
Meanwhile human labor is a significant proportion of the incremental cost, when you have humans doing things per-trip that could reasonably be automated.
There are already many autonomous trains operating all over the world. They have centralised control centers to monitor them, and then maintenance crews that can travel to work on any malfunctions or breakdowns.
This is already happening in Paris, London, Copenhagen, Singapore, Tokyo, and many more places. They all still have staff that move around the network to work on things not related to driving the train though.
So, I think you're right in pointing out that they still need many people constantly monitoring and working on the trains. But they don't need a driver per train any more, and they especially don't need two drivers per train.
No. There is no variance in the daily use of a given vehicle. It's not "today A-line, but tomorrow this train goes to Schenectady." An entire fleet is bought for use only on a set of particular routes.
I think non-Americans underestimate our ability to not automate things that can clearly be automated through some combination of of inertia, union power, and sheer incompetence.