David Beckworth has a new Mercatus paper that examines the Fed's decision to adopt a "floor" system for interest rates. Beginning in October 2008, the Fed began paying interest on bank reserves. This effectively created a floor on market interest rates, as banks would have no incentive to lend money at rates lower than they could receive on reserves held on deposit at the Fed. Prior to 2008, the Fed controlled short-term interest rates by adjusting the supply of base money, a "corridor system." Now they have two independent policy tools, changes in the money supply (open market operations), and changes in money demand (done via interest on reserves.)
David sees several flaws in this new system:
The Fed's floor system, then, may be a drag on economic growth for two reasons. First, it may weaken aggregate demand growth by setting the target interest rate above the natural interest rate. Second, it may inhibit credit and money creation by removing banks' incentives to rebalance their portfolios away from excess reserves. If so, the critics are right to be worried about the Fed's floor system, because it would constitute a Great Divorce for monetary policy.
I worry that deposit insurance biases banks toward too much lending, so at the moment I'm most worried about the first issue. In monetary history, one recurring theme is central banks misjudging the stance of monetary policy because they focused too much on interest rates and not enough on the money supply. Thus during late 2007 and early 2008, the Fed wrongly assumed that it was "easing" monetary policy, even as the growth in the monetary base came to a halt.
Admittedly, this excessive focus on interest rates can occur even without IOR. But the system of interest on bank reserves makes the mistake even more likely to occur, as the quantity of money becomes even less informative. Monetary policy is seen as being all about changes in interest rates, not changes in the supply and demand for base money. The Fed's monetary policy stance during the fall of 2008 would have almost certainly been less contractionary if Congress had not authorized the Fed to pay interest on reserves.