Brad DeLong and Ryan Avent share new research from Robert Gordon and Robert Krenn on the fiscal multiplier in 1939-1941 (that may be gated).
The conclusion? 1.80, until late 1941 when it hit capacity constraints and was reduced to 0.88.
I'm going to try my best to skim this this weekend and maybe comment further, but what I want to stress is how unsurprising everything we know about multipliers is. When you look in demand constrained times it's high, when you look at it in normal or supply constrained times (Barro) it's low. It would shock me not to see that. The estimate is counter-cyclical (there was a VoxEU article recently on this - I'll try to track it down). It would shock me not to see that. It is high before capacity in manufacturing is hit, and then drops once you hit that capacity constraint (you can only build so many tanks with the existing facilities before hitting supply constraints and crowding out kicks in).
Is this a good paper? I don't know - seems like a surprisingly short time frame to estimate a multiplier. I'm interested to see how they do it.
But I haven't really been surprised by much of anything I see like this. Rapid recovery in 1921 and 1946? Entirely understandable and explainable. It would be surprising not to see that.
This doesn't mean unbalanced capital structure or investors scared of political rhetoric don't matter for anything. It's plausible they do. But until I see better empirical tracking of those concepts and until one of my priors gets upset, I'm going to continue to be a Keynesian that thinks Hayek had some great insights and not a Hayekian that thinks Keynes had some great insights. It is the single best framework we have. Period.