...Archive for November 2013

Why (and when) interest-on-reserves matters…

Paul Krugman writes:

Incidentally, small nerdy note. Some people argue that the concept of the monetary base has lost its relevance now that the Fed pays (trivial) interest on reserves. I disagree. Reserves and currency are fungible: banks can turn one into the other at will. But the total of reserves and currency is fixed by the Fed — nobody else can create either. That, as I see it, makes them a relevant aggregate — and anyone who believes that all those reserves are sitting idle because of that 25 basis point reward is (a) silly (b) ignorant of Japan’s experience, where the BOJ sharply increased the monetary base without paying interest on reserves, and what happened looked exactly like our own later experience.

Nerdy indeed! Some might even describe it as dork-ish.

But I think that Krugman is mischaracterizing the view he is arguing with. I’m not sure he would even argue with the view properly characterized. Of course he might!

Perhaps there are people in the world who think that paying 25 basis points of interest on reserves means that base money doesn’t matter, but I have not met any of them. I certainly agree with Krugman that those 25 basis points have a pretty negligible macroeconomic effect now.

The view of people who think that interest-on-reserves permanently diminishes the macroeconomic meaning of base money is contingent on a conjecture that, henceforth, the Fed will always pay interest-on-reserves at a rate comparable to the rate paid by short-term US Treasury securities. If that conjecture is false, then the quantity base money will someday matter again.

In either case, the quantity of base money that exists right now is largely meaningless , and would be meaningless if the Fed were not paying any interest on reserves, because Treasury rates are near zero. This is Krugman’s liquidity trap, an effect of the negative unnatural rate of interest.

The quantity of base money is meaningless right now even if I am wrong about the Fed henceforth always paying interest on reserves at the short-term Treasury rate. Because if I am wrong, the only way that the Fed can create a spread between the interest rate paid on base money (whether zero or something higher) and the Treasury rate is to dramatically reduce the quantity of base money, so that some convenience yield on holding scarce base money offsets the opportunity cost implied by a T-bill / base money spread. Current quantities of base money simply can’t coexist with with a spread between the interest paid on reserves and the interest paid on Treasury bills. (Unless some unexpected thing dramatically increases the convenience yield of holding base money!).

So, there is almost no direct informational value to the current quantity of base money. Perhaps there is indirect information that matters. It could well be that the rate of change of the quantity of base money contains information about the likelihood of future interest rate changes, so it is not irrational of market participants to respond to rumors of tapering or moar QE. Perhaps there are institutional quirks related to the fact market participants can only hold interest-paying base money indirectly via banks, while the stock of risk-free securities depleted by monetary expansion can be held (and hypothecated) by non-bank actors. (If so, “monetary expansion” might be contractionary!)

But the first-order effect of monetary policy is gone. Changes in the base used to engender straightforward imbalances between a direct opportunity cost and the convenience yield of holding money. A reduction of interest rates / expansion of the monetary base would lead to an increase in the direct cost of holding money rather than Treasuries, and put the economy in disequilibrium until NGDP or (too frequently) asset prices adjusted to increase the convenience yield attached to the monetary base. A contraction did the reverse. While we are stuck at zero we can argue over expectations or collateral chains, but the old, blunt, simple channel no longer functions.

And it will never function, as long as the Fed always pays interest comparable to Treasuries on base money. There is nothing special about zero, or 25 bps. What makes a liquidity trap is that the rate of interest paid on money is greater than or comparable to the rate of interest paid on Treasury bonds. So long as that is true, whatever the level of interest rates of interest, macroeconomic outcomes will be much less sensitive to changes in the quantity of money than in once-ordinary times. That is not to say, full-stop, that monetary policy is impotent. Those squishy expectations and institutional quirks may matter. But post-2008, we live in a world where insufficiently expansionary monetary policy has meant tripling the monetary base. Pre-2008, tiny changes in the quantity of base were sufficient to halt an expansion or risk an inflation. The relative impotence of changes in the monetary base is not a function of the zero-lower-bound. It is a function of the spread between base money and risk-free debt, a spread which may well be gone forever.