As Scott already has stressed, this is about whether the liquidity trap makes the Fed impotent.
My view is this: The Fed cannot very well control ngdp during a credit collapse (parts of 2008-2009) but it can control ngdp in a so-called “liquidity trap.”
I view my theory as consistent with this graph, rather obviously consistent.
Scott and I think David believe something more like: “The Fed can always control ngdp.”
That means they have to think Bernanke made a huge error and indeed they do think that. (I think Bernanke was slow to react along the longer-run ngdp forecast dimension, but I am more sympathetic to B. than they are, as I don’t think currency is such a useful substitute for collapsing credit, contra Fama 1980.)
Here is the key question: what do the liquidity trappers think?
That this ngdp path is a coincidence? That fiscal policy has been keeping it on an even keel in more recent times? (Implausible for 2010-2012, given other claims they make about fiscal policy, plus implausible more generally.) That some other process — which? — drives the ngdp path?
The problem with the debate, so far, is that we don’t yet know which clear alternative theory of 2010-2012 ngdp determination the liquidity trappers are proposing.
Sufi, by the way, is complaining that Beckworth (and others) are passing over Romer and other empirical pieces on stimulus, but David is pretty clearly covering “the multiplier contingent upon a reaction function of the monetary authority, with the monetary authority moving last and having control over ngdp.” I have discussed these matters with David and he is not behind the academic discussion here but rather very often ahead of it. He has a clearer theory of ngdp determination than do the liquidity trappers, even if I do not agree with his theory in every regard.
I would very gladly have a more transparent debate on ngdp path determination, noting in advance that I personally do not think it would go very well for the liquidity trap hypothesis.
Addendum: This paper surveys some relevant issues from the theory side, although it is not my preferred approach.