JP Koning commented on my last post suggesting that floating interest rates on reserves would lead to "incredible" volatility in short term interest rates.
First, I must admit that the institutional framework I described is consistent with the Fed doing interest rate smoothing as much as it desires. I don't favor such a policy, but it would be possible.
If, instead, we consider a k percent rule for reserves (including k=0,) then a floating interest rate on reserves would tend to increase the volatility of short term interest rates relative to what would occur with a fixed interest rate on reserves, including keeping it at zero.
In my view, the added variability of short interest rates would be desirable to the degree it reflects changes in the supply or demand for credit. This is equivalent to variation in the short run Wicksellian natural interest rate. In that situation, a fixed interest rate on reserves, much less a policy of adjusting the quantity of reserves to smooth short run interest rates, is disequilibrating. It is keeping the market rate from tracking the natural interest rate and creating undesirable short run fluctuations in aggregate nominal expenditures.
However, to the degree that shifts in the demand to hold money (or reserves) is creating short run fluctuations in interest rates, interest rate smoothing is exactly what should occur. The quantity of reserves would be adjusted to the demand to hold them without creating short run distortions in interest rates or, more importantly, in expenditures on output. If changes in the quantity of reserves are not permitted due to a quantity rule, then keeping the interest rate on reserves fixed would at the very least dampen the undesirable changes in other interest rates and in spending on output. That would be the least bad option.
Since if favor allowing the monetary authority (or central bank) to make changes in the quantity of reserves more or less continually, I see no particular value in a market process that would lessen the harm given a fixed quantity of reserves. (I favor index futures convertibility as a constraint.)
Anyway, my view on what would happen with a fixed quantity of reserves is that borrowing and expenditure plans would be slightly postponed or perhaps hastened based upon changes in short term interest rates. Suppose that there is an increase in the demand for bank credit and banks demand more reserves. This raises the interest rate banks must pay for reserves. If the interest rate the central bank pays on reserves balances is fixed, even at zero, this will result in some banks reducing the amount of reserves they desire to hold.
If there are no reserve requirements, this effect is given its maximum possible effect. It is this short run liquidity effect that is emphasized in money and banking theory--at least since Keynes.
If the interest rate on reserves rises in this situation, then the liquidity effect disappears. The interest rate must rise until the quantity of bank credit demanded is back to its initial level, or else new deposits are attracted, for example by selling negotiable certificates of deposit. Short run credit markets must clear.
Considering the effects on the demand for output, if some want to borrow from banks to purchase more output then others must be deterred from borrowing from banks so that they purchase less output. Or else, others might be induced to lend more to banks, thus saving and spending less on output. This added saving would reduce the demand for output, offsetting the increased demand for output by borrowers.
Certainly, it is possible that this would all occur by responding to changes in short term interest rates. However, we can easily imagine spikes in interest rates being avoided by rationing. Borrowers would be told that funds are not immediately available and that because money is tight, you must wait a few days before making the planned purchase. Low interest rates might well result in lots of calls to potential borrowers stating that funds are available now.
But, what if the problem is that the banks have no additional credit demand but simply decide that holding more reserves is a better policy. Those traders in the money market office have made one mistake due to "too clever" manipulations too many. What should happen is that the monetary authority create additional reserves. But if it fails to do so, then efforts to obtain more reserves by selling securities or else temporarily restraining lending will tend to raise interest rates. If the interest rate on reserves is fixed, then this raises the opportunity cost of holding reserves. Some bank or other releases reserves. While this increase in interest rates is disequilibrating, if the interest rate on reserves increases as well, then there will be no release in reserves! There is no tendency for higher interest rates to result in a lower demand for reserves. Only as expenditure on output (or really, lower output or lower prices) result in lower demand for reserves, will there be a return to monetary equilibrium. The liquidity effect is gone and we are back to a pre-Keynesian world where nominal income must adjust so that the demand to hold money (or reserves) matches the given quantity.
But, of course, the quantity is not given and the monetary authority should expand the quantity of reserves in this situation so that there is no impact on short term interest rates or expenditures on output.
So, what is the effect on volatility of interest rates? I think that they will fluctuate however much the monetary authority wants them to fluctuate. Hopefully, they will fluctuate with changes in the short run natural interest rate only. But nothing in the regime prevents the a central bank from changing the quantity of money to intentionally create monetary disequilibirum to smooth interest rates.