Martin Feldstein tries to answer the question: “Why has the Federal Reserve’s printing of so much money not caused higher inflation?” and comes up with a seemingly obvious answer - because the Fed pays interest on excess reserves. Like many others, Feldstein sees the payment of interest on excess reserves (IOER) as a “fundamental” change in Fed policy. The reality however is that the payment of IOER is a necessary prerequisite for any regime in which the Fed wishes to sustain a positive Fed Funds rate in the presence of excess reserves. IOER is a red herring and there is simply no way in which the Fed can generate inflation by tinkering with it.

The easiest way to understand this is to look at all the possible configurations of the Fed Funds rate and IOER. Ignoring the ludicrous scenario in which IOER is greater than the Fed Funds rate, there are three other configurations:

  1. Fed Funds and IOER are both equal to 0%.
  2. Fed Funds is above 0% and IOER is equal to 0%.
  3. Fed Funds is equal to 0% and IOER is below 0%.

In the first configuration, payment of interest on reserves clearly does not matter. If the Fed Funds rate itself is at zero, then clearly banks have no incentive to try and get rid of excess reserves.

The second configuration is often invoked as a scenario that could generate inflation. But if the Fed Funds rate is above 0% and IOER is 0%, then there can be no excess reserves in the system. If the central bank wants to sustain a positive Fed Funds rate, it must either pay interest on reserves or mop up all excess reserves. If there are any excess reserves, the Fed Funds market rate immediately falls to 0%. And we’re back to configuration 1 where both the Fed Funds and IOER are equal to 0%. To put it differently, if IOER is equal to 0% and the Fed Funds rate is above 0%, there cannot be any excess reserves in the system.

The third configuration is more interesting. Even if we have hit the zero-bound, why can’t the Fed enforce a negative IOER to force banks to try and get rid of their excess reserves and trigger the monetarist ‘hot potato’ effect? If the central bank charges a small negative rate on reserves, the effects will be negligible. Banks will pass on this cost to deposit-holders in the form of negative deposit rates or extra fees. In the absence of any alternative liquid and nominally safe investment options, most depositors will pay this safety premium.

But what if the Fed charges a significantly negative interest rate on reserves? For example, what if it costs 5% to hold excess reserves? In a world where all money is electronic, this may just work. But in a world where bank depositors possess the option to take their cash out in the form of bank notes, highly negative interest rates on reserves are impossible to enforce. In other words, if IOER is -5%, then you and I can earn a higher interest rate of 0% by simply taking our money out of the bank and holding currency instead.

To summarise, there is no avalanche of inflation coming our way no matter what the Fed pays out as interest on excess reserves.



I'm not sure I understand why we should ignore "the ludicrous scenario in which IOER is greater than the Fed Funds rate" given that this is our current situation. IOER is at 0.25%,, while Fed Funds are trading at 0.09 ± 0.06%, Am I missing something important here?

Ashwin Parameswaran

earwulf - apologies. I should have said scenario in which IOER is significantly above the Fed Funds rate.

JP Koning

But if the Fed reduced the IOR from 0.25% to 0%, would you agree that there would be *some* increase in inflation. We could argue about the degree, ie. avalanche vs trickle, but I just don't see how a change in the IOR could ever be neutral.

Ashwin Parameswaran

Absolutely there will be a small increase in the price level. As I said if we make IOER say -1%, we will probably get an increased price level with not much of a move into physical cash by depositors. But this is a small and one-time dose of inflation. There is no "escape velocity" in this approach.


The gap between IOER and FFR stems from this: "Note that some nonbank institutions, for example government-sponsored enterprises (GSEs), still hold deposits at the Fed that do not earn interest. These institutions may be willing to engage in transactions below the rate for interest on reserves, explaining why the federal funds rate may still drop below that rate at times."


I find one flaw here: "if the Fed Funds rate is above 0% and IOER is 0%, then there can be no excess reserves in the system." Suppose we have 2 banks, each with $100 required reserves. There are $250 total reserves in the system ($50 excess reserves). Bank A has $50, bank B has $200. It's not true that bank B will simply lend to bank A at 0%. Even if there are more banks in the system, the overnight (Fed funds) rate will be small, but not zero unless ALL banks have excess reserves.


Hi Ashwin, I think the critical aspect of negative nominal rates does not occur at the short end. There is no real prospect of -5% on bank deposits! A -0.25% on excess reserves, compounded over 25 years leads to a much more significantly negative rate at the long end. So in the NNR regime, the curve is inverted and anchored at a very slight negative rate at zero duration but at perhaps several hundred basis points negative at the long end. This makes sense because any NNR regime is taxing reserves and deleting the receipts from that tax, so shrinking the monetary base and the price level but at quite a slow pace. So NNR are a way of accommodating contraction.

john a

at this juncture, how about imposing a noticeable negative (-1 or -2%) ioer simultaneous with a more aggressive tapering , then gradual liquidation of the feds QE book? of course the banks will cry bloody murder as they now have the backup of the fed to protect their yield curve leveraged trade, and the admin will protest too as it imposes discipline on their spending, but this is the way to solve this great imbalance. eventually the ioer can revert to zero or positive as the feds bond book declines.