You're forgetting: or c) did nothing or d) helped, but not enough.

Fair enough, but those also don't seem like realistic options, given history.

I think Keynesianism is useful because, while it may not be effective, it doesn't really hurt anything either.

What? How about crowding out? How about artificially fabricating demand, leading to longer slumps? How can you say that?

Also, this way we get a bunch of infrastructure upgrades that we needed to do anyway.

We'd already be getting those. What usually happens is that places that need upgrades get ignored in favor of more poltiically-expedient ones.