On the potential for a contractionary impact of IOR, they write:
This claim, made even by some good economists, is puzzling. Before December, the Fed paid banks one-quarter of one percent on their reserves. If the Fed had not paid interest, the return to reserves would have been zero. Accordingly, the only potential loans that would have been affected by the Fed’s payment of interest are those with risk-adjusted short-term returns between precisely zero and one-quarter percent—surely a tiny fraction of the total. In fact, over the last four years bank lending has increased at about a 5 percent annual pace (including around a 7 percent annual rate the past two years), with only residential mortgage lending lagging in the aftermath of the housing bust.
There is more here. I cannot say I am convinced. I would focus not on the interstices of lending, but rather on expansionary pressures from the banking system. Without IOR, what do Fed models predict would have been the price level impact of those trillions of new reserves following 2008? (Note that at some margin banks can just convert those new reserves into dividends, without any additional lending, if they are so satiated with trillions of unwanted liquidity. I’m not saying it would happen that way, but think of that as a limiting case.) No, I’m not advocating hyperinflation, but less sterilization of those new reserves would have maintained aggregate demand at a higher level post-2008, boosting investment, output, and employment through a quite traditional channel, as advocated say by the market monetarists.
And note Bernanke and Kohn’s own argument, elsewhere in the post, that soaking up reserves by selling off the Fed’s portfolio may be impractical, disruptive, or too costly. Let’s say that’s true. By limiting its use of IOR, the Fed in essence would have made a more credible commitment that any increase in bank reserves is here to stay, rather than just sitting in a holding tank of sorts. No?
Bernanke and Kohn also rebut the common charge that interest on reserves is a subsidy to banks. They may be right, but they are arguing too hard for “it is not a subsidy at the current margin of further reserve extension,” and not sufficiently rebutting the possibility of an infra-marginal subsidy, initiated right after IOR.
So this is a very smart and well-argued post, as one would rationally expect, but I don’t quite think it lays one’s doubts to rest either.