Question for Market Monetarists and MMTers: What Happens if IOR Goes to Zero?

January 5th, 2012

For the non-cognoscenti: “IOR” is interest on reserves. Banks keep money in their accounts at the Fed. In October, 2008 the Fed started paying .25% interest on those accounts.

The Fed’s also engaged in “quantitative easing,” a.k.a. open-market purchases on steroids, creating new money and using it to buy $1.6 trillion dollars worth of bonds from banks. The money is deposited in banks’ reserve accounts.

The result: banks have $1.6 trillion dollars in excess reserves (in excess of what they’re required to hold) sitting in their accounts at the Fed.

This is the heart of the “pushing on a string” argument — giving the banks more reserves (making their holdings more “liquid”) doesn’t (necessarily) increase real-economy transaction volumes (on consumption or investment), either directly through spending by the banks or via bank loans to people and businesses who will spend it. This $1.6 trillion in new money issued by the Fed is effectively stuffed in an electronic mattress.

So I’m curious what would happen if the Fed no longer paid IOR.

I asked Scott Sumner this a while back:

if tomorrow the Fed dropped IOR to zero or even negative, what would happen to:

o Excess reserves
o Inflation

He gave a somewhat less than satisfactory answer:

The IOR question is a good one, and at the risk of being annoying I’m going to slightly dodge the question. I do think it would be expansionary, but it’s hard to know how much, because it’s almost inconceivable to me that it would be done by itself, without any other policy changes. It could be slightly expansionary, or if accompanied by other moves, wildly expansionary.

Less than satisfactory (for me) because he often engages in these kind of simplified thought experiments. Change Variable X, ceteris paribus: what would happen?

I’m basically asking for a free education here (hoping others would appreciate such an education as well), but I’m also hoping to spur a discussion on a tightly focused question that has not been cogently discussed, as far as I can find. (I certainly could have missed it. Pointers welcome.)

Cross-posted at Angry Bear.

  1. January 5th, 2012 at 12:05 | #1

    I can help as a Market Monetarist. The reason Scott Sumner does not do ceteris paribus is because his starting point for analysis is: what is the Fed’s target? If they have a specific target, then no matter what the shock, they willl hit the target. Abolishing IOR without any offsetting actions would be vastly expansionary. Before IOR, excess reserves were around 50 billion. Now they are around 1.5 trillion. There is no reason to think they would not decrease back to the pre-IOR levels, at least over time, pushing around around 1.4 trillion dollars of high powered money into the economy and triggering significantly higher inflation. That’s why Scott says ceteris would not be paribus. The Fed would have to offset such an expansionary move with OMOs.

    Here is an article I wrote awhile back:

  2. Becky Hargrove
    January 5th, 2012 at 12:50 | #2

    I wasn’t too crazy about IOR at first. But then, it is a mechanism that can be easily changed as needed which is good. At the present it seems worthwhile in that Europe is in such an iffy position (Like Minsky indicated, the real purpose of the Fed: standing at the ready for emergencies). Plus, banks don’t really have a good reason to increase lending by much now when markets continue to be structured so tightly (too few real consumer options for lower income). So I have actually come around to the position that I’m glad for the present setup. If that changes I’ll let you know!

  3. January 5th, 2012 at 13:22 | #3

    @James Oswald
    “The reason Scott Sumner does not do ceteris paribus is because his starting point for analysis is: what is the Fed’s target?”

    Understood, but he does do it as a way of explaining or exploring an issue.

    “There is no reason to think they would not decrease back to the pre-IOR levels, at least over time, pushing around around 1.4 trillion dollars of high powered money into the economy and triggering significantly higher inflation. ”

    ‘kay that’s two questions out of three (excess reserves and inflation). What about NGDP? And would anyone venture a guess as to what “significantly higher inflation” means, numerically?

    Note: I am not suggesting this should be done. This is just an attempt to get clarity on a single issue.

  4. beowulf
    January 5th, 2012 at 14:11 | #4

    Let us recall Central Banking 101, as boiled down by Jamie Galbraith in his foreword to Warren Mosler’s book.
    “Modern money is a spreadsheet! It works by computer! When government spends or lends, it does so by adding numbers to private bank accounts. When it taxes, it marks those same accounts down. When it borrows, it shifts funds from a demand deposit (called a reserve account) to savings (called a securities account). And that for practical purposes is all there is.”

    When the Fed bought those $1.6T in Tsys, it shifted funds from securities accounts to reserve accounts. Those funds will stay there until drained, either by Tsy taxing it away or by Tsy and/or the Fed buying the reserves back (by issuing debt).

    Tsy could issue new debt (though typically that’s just to pre-drain new federal spending from reserves) or the Fed could sell back that $1.6T in Tsys it holds. There’s also a little known third option. The Fed can issue banks its own new debt— a Fed certificate of deposit.

    Fed CDs shift funds out of reserve accounts and into the term deposit facility (a de facto securities account). If I had my druthers, Tsy would stop borrowing and just create US Notes whenever it spends. Let the Fed worry about draining excess reserves (and controlling short term rates) by its issuance of its own CDs of varied maturities.
    Since the US Govt is constrained by a debt limit while the Federal Reserve, an agency of the US Government is not, its just common sense to shift debt from the constrained whole to the unconstrained part. :o)

  5. January 5th, 2012 at 17:05 | #5

    I was just hazarding a guess as to why he did not do ceteris paribus reasoning in this case. He has done it in the past.

    I should not have said “inflation” necessarily. Theoretically, the higher NGDP could come as either higher real growth or as higher inflation, but it would increase very quickly. I would say that inflation would be the likely outcome, since there isn’t that much excess capacity.

  6. beowulf
    January 5th, 2012 at 22:06 | #6

    @James Oswald

    Sorry James, I had meant to post that as a general response. My intent was to harass that Steve Roth fellow, not you.

  7. January 6th, 2012 at 07:12 | #7

    @James Oswald

    First, again, lemme repeat: this is not something I’m suggesting. Just using it as a means to understand.

    Here’s what I don’t understand, and what your article doesn’t really explain:

    “pushing around around 1.4 trillion dollars of high powered money into the economy ”

    1. Reserves *are* high-powered money. Increasing that quantity (massively) has had little effect, apparently.

    2. How do banks “push” it into the economy?

    Suppose the banks wanted to hold physical currency instead of reserves, because IOR is zero or negative. (At zero you’d think they’d prefer reserves; less costs to hold, and actually much more liquid.) They’d buy up vaults full of cash and sit on it, *draining* currency from circulation.

    I can’t see how banks sitting on currency instead of reserves would increase anyone’s incentives to go buy things. (Ditto the next step up the liquidity ladder — reserves instead of bonds.)


  8. January 6th, 2012 at 07:16 | #8

    Well-harassed, thank you. 😉

    I get all that, but it doesn’t answer my question. Give the Rube Goldberg system we’ve got, what happens if IOR goes to zero or negative?

    I’m starting to think that the Thoma post linked at the Angry Bear version of this post gives the answer: not much in terms of the real economy, but many financial institutions and arrangements go schitzo.

  9. Becky Hargrove
    January 6th, 2012 at 09:00 | #9

    Given this mornings’ discussions about rehypothecation (Marginal Revolution) all the liquidity and IOR in the world may not be enough for this leaky bucket. I don’t think that Bernanke is sleeping well right now.

  10. beowulf
    January 6th, 2012 at 11:35 | #10

    Yeah, I’d agree with Thoma. Especially since the FDIC “asset tax”– which they’ve levied instead of a “deposits tax” since last April claws back about almost half of the 0.25% IOR anyway.
    “The new assessment implicitly increases the cost of the federal funds by adding that assessment rate onto the fed funds rate.”

    So my fix for the Rube Goldberg system is an even bigger Rube Goldberg system. Drop IOR to zero and peg short term rates by FDIC board adjusting the bank asset tax rate. As the FDIC marks it up the cost of money goes up, it would ripple through as banks charge higher interest rates on loans. Even if Tsy simply “printed money” (or minted coins) to spend, it’d be automatically sterilized when the spending flowed into excess reserves (which ARE part of the FDIC tax base). Except instead of a being a Tsy revenue-negative reserve drain, its a revenue-positive reserve drain– higher rates mean more money in FDIC deposit insurance fund (which naturally Congress could order rebated back to Tsy like Fed earnings).

    The FDIC could screw up the bond market (in a good way) by removing T-Bonds from tax base, higher tax rate charged on reserves would mean lower T-bond rates (since even a 0% T-bond is a better return than say, a -1.0% excess reserves). With “net interest” a Tsy profit center instead of an expense, those Pete Peterson deficit scare charts would all have to be revised in a dramatically less scary manner.

  11. beowulf
    January 6th, 2012 at 11:54 | #11

    Come to think of it, I don’t think Treasuries are in the FDIC tax base now, so maybe the FDIC already screwed up the bond market (that would explain why T-bill rates are so much lower than IOR or actual Fed Fund rates).

  12. January 6th, 2012 at 20:24 | #12

    Basically the answer goes back to Hume:
    “If the coin be locked up in chests, it is the same thing with regard to prices, as if it were annihilated”

    Money, like some bizarre quantum particle, only affects things when it is spent, and a pile of money sitting in a bank vault cannot be spent. The reason why more money is normally associated with more inflation is that under normal circumstances, the more money people have the more they spend. Reserves are never spent and thus have no impact on prices or spending. If the central bank increases reserves, under normal circumstances, the banks will lend the reserves out and they will be spent. Currently however, the banks’ best investment is simply to “lend” the money to the Fed and collect their risk free .25%. But the Fed does not spend the money, nor does the Fed make loans to individuals or businesses.

  13. January 6th, 2012 at 20:29 | #13

    Ok, for the rest of your response:
    2. The banks push the money into the economy by lending it to businesses and individuals who then use the money to buy real goods and services. It’s the claims on real productive resources that lower unemployment and get the economy moving.

    On physical currency, you could convert your bank account to cash. Everyone you know could convert their bank accounts to cash, but a bank cannot, simply because there isn’t that much cash in the entire economy. There is only around 400 billion of currency in the world, and the Fed does not pay interest on currency. Typically, because of inflation, currency yields a negative return, and so people only hold as much as they have to to do transactions. If the Fed credibly committed to a higher inflation or NGDP target, no one would try to horde cash.

  14. beowulf
    January 7th, 2012 at 03:14 | #14

    “If the coin be locked up in chests, it is the same thing with regard to prices, as if it were annihilated”

    Exactly, that’s a great point. Compulsory savings has the same disinflationary effect as compulsory taxation and yet is far more popular with the public (a point Keynes made in “How to Pay for the War”). So fund Social Security out of TGA (Social Security is kept off-budget, good, lets keep it that way) and FICA taxes could then be allocated to TSP or 401(k) accounts.
    Here’s where the magic happens, Tsy could adjust FICA tax rate counter-cyclically with zero impact on the federal budget. Since none of the money was going to TGA anyway, it’d be a revenue-neutral stimulus plan.

    Congress won’t ever give Tsy discretion over tax rates, they take the whole “power of the purse” thing seriously. But allocating the percentage of private compensation going to either wages or savings, that’s a different kettle of fish. That I can see them doing (since, in truth the Secretary would be powerless to deny the government or any citizen so much as a dollar of their own money). Maybe that’s what Keynes meant by socialization of investment. :o)

  15. January 7th, 2012 at 07:35 | #15

    James, Beo:

    Thanks. I totally get all that. Except:

    “The banks push the money into the economy by lending it to businesses and individuals who then use the money to buy real goods and services.”

    I don’t see how — in the current situation, at least — lowering IOR would cause more lending.

    Which means the effect would be as described in the Thoma-linked article reffed and discussed over at the Angry Bear thread — serious disruption of the financial system because it breaks various financial entities’ business models, causing negative second- and third-order effects on the real economy.

    But the assertion that it would cause first-order effects via “pushing” money via lending doesn’t seem to have legs.

  16. January 7th, 2012 at 07:37 | #16

    Which perhaps explains why the Fed instituted it when they did, as rates went to zero. (Another question I’ve been wondering about.) They do at least seem to understand the mechanics of the financial system, and knew that they needed to preclude that problem.

  17. flow5
    January 11th, 2012 at 20:23 | #17

    IOeRs’ are the functional equivalent of required reserves. They are a credit control device. IOeRs’ are used to offset the FED’s liquidity funding programs (expansions), on the asset side of the FED’s balance sheet. I.e., quantitative easing, or the FED’s POMOs, were sterilized by adjusting the remuneration rate which restricts bank lending & investment. For the sterilization process to work, IOeRs by definition, must be contractive. I.e., the bank’s cash assets were transformed into earning assets, If the BOG raised the remuneration rate on excess reserves (vis a’ vis other competitive financial instruments, yields, & returns ), the VOLUME of IBDDs will increase and during this process IOeRs will also absorb savings inducing dis-intermediation.

  18. flow5
    January 17th, 2012 at 06:48 | #18

    The pertinent question is what would happen if inflation increased beyond tolerable limits & the remuneration rate was raised. The fact is: the higher the remuneration rate on excess reserves, the more destructive the outcome would be to the money market (higher levels of dis-intermediation), i.e., money would flow OUT of the economy into the commercial banking system (i.e., the intermediaries would shrink in size but the CB system would be unimpaired).

  19. January 17th, 2012 at 07:40 | #19

    @flow5 “The pertinent question is what would happen if inflation increased beyond tolerable limits & the remuneration rate was raised. ”

    That’s definitely a pertinent question. But even with just the two variables (and the unstated implicit variables associated), the question is far more complicated than the rather simplistic one I ask here.

    See the video here:

  20. flow5
    January 20th, 2012 at 11:43 | #20

    Right, Keynes’s liqudity preference curve has always been a false doctrine. And IS/LM is further corrupted in that savings don’t equal investment. Savings held within the commercial banking system are impounded (lost to investment or to any type of expenditure). I.e., banks don’t loan out existing deposits saved or otherwise, they always create new money in the lending & investing process.

  21. BD
    February 2nd, 2012 at 08:03 | #21

    From an MMT perspective, if the Federal Reserve stopped paying IOR, banks would stop receiving income from reserves. Other riskless asses such as T-bills, currently yield 6 basis points right now. As banks try to move from reserves to t-bills, which they can’t do, the T-bill rate drops to zero. Private banks do not increase lending, but they do experience a loss of income.

    Since, the Federal Reserve is no longer paying income on reserves, its net income, which is remitted to the treasury, increases. The decision by the federal reserve decreases the size of the deficit, reduces private bank income and lowers riskless, short-term interest rates by 6 basis points, as of 2 February 2012. The impact on net income for the Federal Reserve is about $4 billion annually on $1.6 trillion in reserves (times 0.25%). The federal government’s borrowing costs also decrease slightly, perhaps by $1 or $2 billion per year. Big Deal. The money multiplier still doesn’t exist, even if the Federal Reserves abandons its IOR policy. If the money multiplier is real, why don’t Canadian banks expand the money supply to infinity, since they don’t have any reserve requirements. Come on, people.

Comments are closed.