Our Unhinged Fed

hobby horse, testimony, ioer, Federal Reserve, reserves, excess reserves
Wooden and Corduroy Hobby Horses in a Bucket in a Courtyard at Arlington Court, near Barnstaple, Devon, England, UK Thinkstock http://www.thinkstockphotos.com/image/stock-photo-arlington-court-the-country-estate-barnstaple/531228483/popup

hobby horse, testimony, ioer, Federal Reserve, reserves, excess reservesIf you haven't seen much of me on these pages lately, that's because I've spent most of the last few weeks feeding and grooming my favorite hobby horse: that's right, the Fed's policy of encouraging banks to hoard reserves by paying above-market rates on their Fed reserve balances.

Well, last Thursday morning I rode the old gal to Capitol Hill, where I put her through the paces before the House Financial Services Subcommittee on Monetary Policy and Trade, at its hearing on "Monetary Policy v. Fiscal Policy: Risks to Price Stability and the Economy." Mickey Levy of Berenberg Capital Markets, Eric Leeper of Indiana University at Bloomington, and Jared Bernstein of the Center on Budget and Policy Priorities, also took part.

Below I reproduce my five-minute spoken testimony, in which I attempt to summarize the twenty-two thousand word written testimony I submitted beforehand.  All four written testimonies, including my screed, can be read here.  Those who wish to see the entire show, including my and the other participants' replies to members' questions, will find a video embedded further below.

***

Chairman Barr; Ranking member Moore; distinguished committee members: In October, 2008, the Federal Reserve began paying interest on banks’ reserve balances with it. My testimony today concerns the economic consequences of that step.

The Fed was originally supposed to start paying interest on reserves in 2011, to reduce the implicit tax burden reserve requirements placed on banks. But as the 2008 crisis worsened, the Fed received Congress’s permission to start paying interest on reserves immediately. Its goal then was, not to relieve banks of a required reserve burden, but to get them to hoard reserves it was creating by its emergency lending, so that that lending wouldn’t result in increased bank lending and inflation.

To make interest on reserves serve this role, the Fed set the rate on reserves above comparable market rates, where it has kept it ever since. It thereby ignored the laws' stipulation that the rate was “not to exceed the general level of short-term rates.”

As an anti-stimulus measure, interest on reserves worked very well: so well that within weeks the Fed did an about-face. Now it hoped to stimulate the economy by purposefully creating large quantities of fresh bank reserves.  All told, the subsequent three-rounds of “Quantitative Easing” created another $2 trillion of additional bank reserves. Yet because reserves still paid an above-market rate of interest, banks just kept on accumulating them, as they had done — and as the Fed had wanted them to do — before QE, when it was worried about inflation.

If  “insanity is doing the same thing over and over again but expecting different results,” then I fear it must be said that at least some Fed officials were not quite in their right minds.

Although the QE stimulus was disappointingly small, the Fed’s actions had other, big consequences. By acquiring trillions of dollars worth of Treasury and mortgage-backed securities, and borrowing from banks to pay for them, the Fed dramatically increased its footprint on the U.S. credit system. Before IOR and QE, bank reserves were less than 1% of bank deposits; bank loans, in contrast, were almost 100% of bank deposits. Today bank reserves are 20% of deposits, and loans are just 80% of deposits. Before IOR and QE, the Fed’s assets were  7% of commercial bank assets. Today the figure is 27%.

Commercial banks are expected to invest the public’s deposits productively, subject to certain regulatory guidelines. Central banks aren’t. They’re tasked instead with regulating the scale of commercial bank lending and deposit creation. According to the Fed’s own guidelines, as set forth in a pre-crisis publication, it is supposed to “structure its portfolio and … activities so as to minimize their effect on… credit allocation within the private sector.” The reason for this, the same guidelines state, is that hard-earned experience shows that “in general…market-directed resource allocation fosters long-run economic growth.”

In fact there’s a vast economics literature on what’s known as “financial repression.” The term refers to the harmful consequences of policies — mainly in less-developed countries — that divert savings from commercial banks to central banks, and thus from more to less productive uses. That literature blames such policies for much of the world’s poverty.

The Fed’s current operating system, with its above-market interest rate on reserves and bloated balance sheet, is very financially repressive. That is one reason for the continuing post-crisis “productivity slowdown.” Yet the same system, far from at least improving basic monetary control, has prevented the Fed for 5 years running from meeting the 2% inflation target it set in 2012.

Distinguished committee members, Chairman Barr, a central bank that cannot control inflation, and especially one that cannot make inflation go up, is a central bank that is unable to perform its most fundamental duty.

To close, the Fed’s new operating system, based on above market interest on reserves, has had disastrous consequences. Yet despite these consequences, the Fed’s current plan for “normalization” would keep much of the current arrangement in place. I hope, for the general public’s sake, that Congress will not let that happen.

  • Ray Lopez

    Nice speech George! I'm sure the congressmen did not read any of your written report (maybe one or two of their aides did) but they did hear your well reasoned oral testimony.

    However, I will say it seems that these statistics need clarification and/or they are misleading: "Before IOR and QE, bank reserves were less than 1% of bank deposits; bank loans, in contrast, were almost 100% of bank deposits. Today bank reserves are 20% of deposits, and loans are just 80% of deposits. Before IOR and QE, the Fed’s assets were 7% of commercial bank assets. Today the figure is 27%."

    Isn't it true (I think, it's hard to get stats for a layman like me on this issue) that bank reserves expand and contract according to the economy? Hence during 'boom times' reserves and/or equity (given they are not the same thing, but correlated it seems to me) are always very thin (less than 1%) and during contraction they get fatter (i.e., in 2007 bank multiples were 30:1 it was said, and during the Great Recession they came down to 10:1). The stat I'm looking for (and have not yet found) is this: what multiple of liabilities/assets (I get the two confused for banks) divided by equity (or say reserves, though not the same thing as equity) is present during the business cycle? Does this multiple fluctuate? If so, then the banks behavior above is 'typical'. However, the question for now is why banks are not lending more. And the answer may be either: (1) they get free interest on reserves so why bother, or, (2) nobody is borrowing money anymore, given that the retail sector is devastated (Tysons Corner VA shopping mall is empty I am informed) and people cannot afford to buy houses (the two big customer groups for banks).

    • George Selgin

      You can look up bank loans and leases on FRED and compare them to commercial bank deposits and you will see that they are almost the same year after year before 2008.

  • International Macro

    Bloody good summary. Hope it has the desired impact.

    • George Selgin

      Thanks I.M.

  • Kevin Dowd

    George says that “By acquiring trillions of dollars worth of Treasury and mortgage-backed securities, and borrowing from banks to pay for them, the Fed dramatically increased its footprint on the U.S. credit system."

    Unfortunately, the phrase "and borrowing from banks to pay for [the asset purchases]" involves a serious error.

    As a general rule, the Fed does not purchase new securities by borrowing from the banks. Instead, it purchases them by issuing new base money with a stroke of the keyboard. This is standard textbook stuff.

    George’s passage equates commercial banks and central banks. Were the Fed to operate like a commercial bank, then its operations would have the same significance as those of, say, BAC. But the Fed is much more significant precisely because it and it alone can create new base money out of nothing.

    Now I grant that it is possible that the Fed could buy more securities on the open market by borrowing from the banks (e.g., via repo operations). Such operations would not increase the size of the Fed’s balance sheet, however.

    The enormous ballooning of the Fed's balance sheet proves that the Fed has been financing most if not all of its asset purchases by expanding base money, not by borrowing.

    Had George simply written, “By acquiring trillions of dollars worth of Treasury and mortgage-backed securities, the Fed dramatically increased its footprint on the U.S. credit system" then he would have been correct. But had the Fed merely borrowed from banks to pay for these purchases – which it didn’t – then I put it to George that it would haven’t have been able to make those purchases in the first place.

    The Fed could only pursue QE because its unique powers as a central bank enable it to create base money out of nothing!

    • George Selgin

      In the final analysis the Fed's command over assets depends more on the real demand for its liabilities than on the nominal quantity of such that it chooses to create. This fact is also to be found in the (better) textbooks!

      • Kevin Dowd

        Your answer is a head scratcher, George. I don't understand what you mean when you say that the Fed's "command over assets" (whatever that might mean) depends MORE on the real demand [etc] and LESS on the nominal supply [etc], if I have paraphrased your statement accurately.

        But in any case, I understand your passage well enough to know that it does not address the points I made.

    • Ray Lopez

      Dr. Dowd, an immovable object in monetarism, runs into the unstoppable force of Dr. Selgin, but who, Dr. Who, will win? I say, by George, the smarter man wins!

      Dowd: "the enormous ballooning of the Fed's balance sheet proves that the Fed has been financing most if not all of its asset purchases by expanding base money, not by borrowing". Wrong wrong Wronskian (apologies to any math readers, btw he was also an economist: https://en.wikipedia.org/wiki/Józef_Maria_Hoene-Wroński)! In fact, the Fed ballooned their balance sheets from $1T to $4T by buying (junk and otherwise) commercial paper, at least initially. The "keyboard" purchases Dowd mentions happened later. And I would say the "keyboard" fiction–widely repeated by monetarists–is linguistic nonsense that even Chomsky would disapprove of. What is the difference between Dowd's "By acquiring trillions of dollars worth of Treasury and mortgage-backed securities, the Fed dramatically increased its footprint on the U.S. credit system" then he would have been correct" and what Selgin wrote? A whole lot of nothing! If a central bank gives base money in exchange for junk paper (or for a promise of junk paper) to a delinquent bank, saving it (think BAC, C) is it any different than 'borrowing' money from a deadbeat bank? No. It's the same thing. Both 'debtors' are in no position to pay and the central bank is bailing them out, exactly as what happened with BAC, C and a host of other financial institutions in 2008.

      Reader request: BTW, if you're reading this far, I would like to see an article on how the money supply is expanded–I do understand it's from the interest rates–in today's economy where the Fed cannot print money and just hand it to the Treasury to spend. An article on how Bitcoin (a fixed money supply) would also work in an economy that's expanding would also be welcome, as well as seigneurage (if somebody owns all of bitcoin mining, how does that person 'get rich' the most. Do they print all of bitcoin at once, or slowly over time? I think money neutrality is irrelevant in this discussion).

      • Kevin Dowd

        "Wrong wrong Wronskian (apologies to any math readers, btw he was also an economist: https://en.wikipedia.org/wi…ózef_Maria_Hoene-Wroński)! In fact, the Fed ballooned their balance sheets from $1T to $4T by buying (junk and otherwise) commercial paper, at least initially."

        Wow! What I actually wrote was "The enormous ballooning of the Fed's balance sheet proves that the Fed has been financing most if not all of its asset purchases by expanding base money, not by borrowing." Nothing in your post contradicts that claim.

        Your "keyboard fiction" is an accurate description of the Fed's operating procedure: the Fed pays digitally, rather than by physically printing cash.

        As for your not being able to see the difference between my position and George's, well, that is up to you.

        If you cut out the noise in your postings, the truth might become clearer. Over and out.

        • George Selgin

          There is in fact no substantive difference between Kevin's understanding and mine. A fiat-money issuing central bank acquires assets by exchanging its liabilities for them, where those liabilities are, in essence, perpetuities. (John Exter cleverly labeled them "IOU nothings," but there's no practical difference between "nothing" and "something, never.) In other words, "buying" and "borrowing" amount to two ways of saying the same thing. Kevin has tried here to turn a semantic quibble into a substantive disagreement; I wish he would just recognize it for what it is. Paying with a liability or security = "borrowing."

          And as I said before, it is ultimately the real demand for a central bank's liabilities that determines both their value and the value of the central bank's assets. That "head scratch inducing" claim is the essence of the quantity theory, among many other fundamental propositions of monetary economics. By paying IOER, the Fed has dramatically boosted banks' real demand for reserve balances; hence it has gotten those banks to lend more to it, in real terms, than they would otherwise have done. The Fed's monopoly of currency is the ultimate source of demand for its liabilities, of course; but IOER has allowed it to boost that demand more than ever.

          In all such discussions a modicum of good will is called for in assessing others' positions. The summary dismissal of another person's viewpoint, and suggestions that that person lacks basic ("textbook") knowledge of the subject under discussion, when that person can hardly be said to be a newbie in the field, is as imprudent as it is uncharitable.

          • Kevin Dowd

            George:

            I have not "tried" to turn a semantic quibble into a substantive disagreement.

            Your claim that "Paying with a liability or security = "borrowing."" is not correct as a general statement.

            If I pay for something by writing a check, then that per se is not borrowing, and if the central bank pays for asset purchases by issuing more base money then I would claim that is not borrowing in any substantive sense, although one might regard it as such from the perspective of central bank balance sheet accounting. But to conflate the latter with "borrowing" in a normal sense serves to hide the very distinction I was trying to make, namely the distinction between the central bank financing asset purchases by borrowing and by issuing new base money.

            When referring to textbook knowledge I was NOT suggesting that you do not know your textbook theory. I sought instead to emphasize that I was not making some new subtle point that no one had heard before, but was reiterating old basics. For anyone to interpret what I wrote in any other way is itself less than charitable.

            But to repeat, I think we have talked this out and will have to agree to disagree.

          • George Selgin

            "If I pay for something by writing a check, then that per se is not borrowing." Yes, it is, until the check clears. A check is merely a written order to the bank to pay a sum to the person to whom the check is made out; it is, in other words, a promise to pay.

            Here is a typical statement of the point–one of many one can find; this one (the first that popped up on Google books) happens to be from Richard Ely's old principles text (p. 218). I reproduce it because I fear my claim will otherwise be treated as idiosyncratic, when it fact it is perfectly orthodox:

            "I. Instruments of Credit. — Among the instruments of
            credit the simplest and most extensively used is the (1)
            check. A check is an order upon a bank by an individual
            or company requiring the payment of a certain sum of money to the order of a person named or to the holder of the check. In this form of credit the element of time plays a very small part. If money were paid instead of a check, the person receiving it would be likely to deposit it in a bank. Receiving a check, he carries it to the bank. The element of credit here prominent is the trust or confidence involved, the confidence that the check will be honored by the bank upon which it is drawn."

            And while it is your prerogative to withdraw from the discussion at any time, it is not your prerogative to determine when I should so so.

          • Ray Lopez

            OK, apologies accepted by both of you. I will now try and synthesize these distinctions, and school all of you along the way, and during writing this post maybe even learn something new (I'm a science guy, not an economist, though I'm now retired–stopped after age 40, I have enough money, thanks, hope you're doing well too). If my style rubs you the wrong way, well this is the internet, it's been that way for decades. I trolled from a Stanford university account almost thirty years ago. Trigger warning, snowflake reader: if you want to read a journal article, this comments section is not the place!

            First off, this apparently is the sticking point: what is borrowing? Let's start with fundamentals: bank assets = bank liabilities plus bank capital. Bank capital is never used, it belongs to the owners (I think it does not even count as reserves for Fed purposes, could be wrong though). Assets include reserves (paper that sits at Fed member banks), securities (Fed paper and commercial paper) and loans (to bank customers). Liabilities include transaction deposits ('checking accounts'), non-transaction deposits (not sure what that is, except by exclusion, maybe a term deposit like a 1 year CD) and–this is key–borrowings.

            What is/are borrowing(s)? Let's look at what these two Titans of Monetarism argue:

            George: “By acquiring trillions of dollars worth of Treasury and mortgage-backed securities, and borrowing from banks to pay for them, the Fed dramatically increased its footprint on the U.S. credit system."

            Kevin: Unfortunately, the phrase "and borrowing from banks to pay for [the asset purchases]" involves a serious error. As a general rule, the Fed does not purchase new securities by borrowing from the banks. Instead, it purchases them by issuing new base money with a stroke of the keyboard. This is standard textbook stuff.

            George: In other words, "buying" and "borrowing" amount to two ways of saying the same thing. {mild ad hominem deleted} Paying with a liability or security = "borrowing."

            Kevin: Your claim that "Paying with a liability or security = "borrowing."" is not correct as a general statement. If I pay for something by writing a check, then that per se is not borrowing, and if the central bank pays for asset purchases by issuing more base money then I would claim that is not borrowing in any substantive sense, although one might regard it as such from the perspective of central bank balance sheet accounting. But to conflate the latter with "borrowing" in a normal sense serves to hide the very distinction I was trying to make, namely the distinction between the central bank financing asset purchases by borrowing and by issuing new base money.

            George (surrebuttal): "If I pay for something by writing a check, then that per se is not borrowing." Yes, it is, until the check clears. A check is merely a written order to the bank to pay a sum to the person to whom the check is made out; it is, in other words, a promise to pay, the drawee of which implicitly grants credit to the drawer for the time it takes to settle.

            OK, how to square these circles? By understanding that Fed Reserve "liabilities" are Fed member bank "assets". The Fed 'buys' member bank assets by 'paying' with 'base money'. This base money (Fed asset = member bank liability) is then used to create more member bank assets (loans to customers) that multiplies throughout the economy due to fractional reserve banking (the famous geometric series formulae 1/(1-r), at the limit). Thus the Fed 'borrows' paper from member banks to "pay" them money, so George is 'right'. But, upon reflection, Kevin has the more 'orthodox' definition (not understanding, I think Kevin and George are equals). Why is Kevin's 'keyboard' definition correct (and BTW Dr. Dowd, if you're reading this, I'm not talking about electronic money created with the push of a button, please)? Because of fractional reserves, the Fed can ALWAYS create MORE MONEY! To understand this, simply think about how the world would work under a 100% reserves (narrow banking) world. To save space, I won't even describe this, just think about it. Thus today, the Fed can always create more money 'at the push of the button' (Kevin's point) by simply 'borrowing' (George's point) member bank assets, said assets were created using the geometric series expansion of previous Fed base money. Since the Fed can always "borrow" loans created by previous Fed base money to create more base money, using 1/(1-x), the process can expand infinitely in a fractional reserve system. Now THAT's textbook 101, derived from first principles from a guy who took Econ 101 in university 'before both of you were born'!

            So both Kevin and George are in violent academic disagreement. And fighting so hard because so much money (not their own) is at stake: the entire money supply of the USA.

            Closing thought: society –including those pols on Capital Hill–implicitly assume money is neutral, short term and long, so none of this matters outside of hyperinflation, which the developed countries don't have (nobody has; only Venezuela). Analogy follows. I dealt professionally with patents. One sound argument I heard my entire life was this (the unsound ones, like 'patents are monopoly, and all monopoly is bad' are not even worth discussing): if patents are needed, why do most Nobel prize in physics winners and even ordinary inventors (see by way of example, https://en.wikipedia.org/wiki/Newman_Darby 'windsurfer' inventor) not have any patents? Why did Newton, Maxwell, Gibbs, Arrhenius, Einstein, many others, invent (or discover, as if there's no invention, ha!) math formulae that changed the world? Because 'inventors invent', without incentive. That is society's assumption, and to date most of the 'real' inventions have been produced under that assumption. The UK hovercraft inventor (https://en.wikipedia.org/wiki/Christopher_Cockerell) found out the hard way what this means (he acknowledged, as have many others, that this assumption by society is wrong but does not prevent society from reaping the benefits of hapless inventors throughout history who have been exploited by society). Numerous other examples in this vein; one paper on my hard drive says inventors only get 5% of their fair share of any worthy invention, society gets the other 95%). To construct a counterfactual where patents matter would require an alternate universe that we don't have. To date, weak and bad patent laws have gotten us this far, so why tinker with inventing a new patent law? No need. Another analogy is by economist Peter Lindert in his book showing the rise of the welfare state in the 20th century did not cut down on GDP growth rates, so what's the bother? What does Cato say? You'd have to construct an alternate universe where less government would mean even greater growth, and that's speculative hypothesis, not proven theory.

            Over and out!

  • Fed Up

    I want to start here. Before about 2008, interest on required reserves and interest on excess reserves were both zero, right?

    After about 2008, interest on required reserves and interest on excess reserves were the same, right?

    • George Selgin

      After 2007, the two rates were different initially, but they were soon made identical.

      • Fed Up

        2006: IORR = 0 ; IOER = 0

        2008: what were IORR and IOER?

        2009: IORR = .25 ; IOER = .25

        1) I thought IOER was about being able to keep the fed funds rate above 0.

        2) You talk about lending (central bank) reserves / hoarding (central bank) reserves. I believe you need to distinguish between lending (central bank) reserves in the fed funds market and "lending" new demand deposits by commercial banks that could possibly create a reserve requirement that needs to be met. I think the best way to do that is to assume there is one central bank, one commercial bank, and the private sector. That eliminates payment clearing and the fed funds market.

        Start with excess (central bank) reserves at zero. Now have the one central bank buy a treasury bond from the private sector. That private entity wants to hold demand deposits at the *one* commercial bank. Now there are excess (central bank) reserves.
        What are the range of possibilities now?

        • George Selgin
        • George Selgin

          On your second point, I am I think quite clear on the distinction you make: my 5 minute spoken testimony summarizes a 22,000 word document in about 680 words, so of course it cannot go into such details. (Nor would the audience have welcomed it had it done so.) Please see the written version for all of the nuts and bolts.

          • Fed Up

            "Nor would the audience have welcomed it had it done so."

            From above, "Its goal then was, not to relieve banks of a required reserve burden, but to get them to hoard reserves it was creating by its emergency lending, so that that lending wouldn’t result in increased bank lending and inflation."

            My guess is if you had quizzed the audience later on about that statement, you would get different interpretations of that (probably many). Some of them would be wrong.

            I am going to try to work on my scenarios. I hope to post more. Time to eat.

        • Fed Up

          If I am following this correctly, you mean IOER is too high. That means the commercial banks do not "lend against (central bank) reserves". Lending against (central bank) reserves means creating a new demand deposit that creates a "new" reserve requirement.

          With banks I am going to assume they only purchase bonds. With all those assumptions, what bonds would the banks buy if IOER was lower? New bonds? Existing bonds? Treasuries, mortgages, personal loans?

  • Philon

    (I wrote this before seeing Keven Dowd's comment.)

    I was surprised to see you write: "By acquiring trillions of dollars worth of Treasury and mortgage-backed securities, and borrowing from banks to pay for them, the Fed [etc., etc.] . . . ." I thought the Fed paid for its QE purchases by creating new money rather than by borrowing from banks. Perhaps your point was that much of the new money thus created by the Fed was given to banks as the Fed purchased Treasury and mortgage-backed securities (since banks owned a lot of these securities), and the banks *in turn* put or left these funds on deposit with the Fed in the form of excess reserves. A bank deposits excess reserves with Fed = the Fed borrows from the bank–is that the idea?

    But a bank probably decides whether to deposit funds with the Fed as excess reserves quite independently of where the funds came from–not just when they came from the Fed's purchase of securities held by the bank. Thus there is little connection between the magnitude of the Fed's QE purchases of securities from banks (and still less from non-banks) and the magnitude of excess reserves built up by banks with the Fed. (Indeed, since the Fed was paying an above-market rate of interest, a bank should sell its entire portfolio of safe short-term securities, even if the Fed did not want to buy all of them, and deposit the proceeds with the Fed as excess reserves; the build-up of excess reserves should have greatly exceeded the Fed's QE.) So the sentence I partly quoted above still strikes me as misleading.

    • George Selgin

      Yes, Philon, what you propose to have been "the idea" behind my statement is indeed what I meant by it.

      The Fed could purchase anything it wished in anticipation of banks' accumulating excess reserves and thereby funding the purchases by increasing the real demand for Fed balances by a like amount.

      The banks could not collectively dispose of loans already on their books before QE.
      At most some banks could sell loans to other banks. But even doing that much would not be cost effective in most cases.

      • Philon

        "At most some banks could sell loans to other banks." I believe banks can also sell loans to non-banks.

  • Jerry O'Driscoll

    I must query George about his statement that "a central bank that cannot control inflation, and especially one that cannot make inflation go up, is a central bank that is unable to perform it most fundamental duty." Have you joined the Fed Up crowd calling for higher inflation? Tell me it isn't so.

    • Kevin Dowd

      I agree with you Jerry.

      So is George suggesting that the central bank cannot make inflation go up or that it cannot control inflation at all? If so, I would suggest that that is a radical departure from sound monetary orthodoxy.

      Lets go back to first principles: if I was running a central bank and wanted to increase inflation, I know just how to do it. I would issue base money like crazy and make no effort to sterilize it. Works every time, right? But if I can make inflation go up, then I also have the means of controlling it.

      And if he is suggesting (as he seems to be) that the Fed is thereby failing in its most fundamental duty, then that raises the question of what George sees the Fed's most fundamental duty to be.

      I thought its most fundamental duty had something to do with promoting monetary and financial stability, or controlling inflation, or maybe dual mandate.

      • George Selgin

        No fellows: I'm not in favor of more inflation: I'm bemoaning the Fed's loss of monetary control. Jerry, you have yourself expressed this concern, no?

      • George Selgin

        Kevin, what you propose as a sure-fire way to raise the inflation rate is just what the Fed attempted between 2008 and 2015, to the tune of several trillion dollars of new base money creation. Yet, since the Fed first announced an explicit inflation target of 2% in January 2012, it has failed to reach that target every quarter since. If that is not a valid test of your suggested solution, and one that establishes its inadequacy, what test would qualify?

        • Kevin Dowd

          "what you propose as a sure-fire way to raise the inflation rate is just what the Fed attempted between 2008 and 2015, to the tune of several trillion dollars of new base money creation. ..If that is not a valid test of your suggested solution, and one that establishes its inadequacy, what test would qualify?"

          What I actually proposed (and I believe I was quite clear) was that if the Fed wanted to increase inflation then it could do so by (a) issuing a lot of base money and (b) making no effort to sterilize the new base money. I grant that it did (a) but it did not do (b): Exhibit 1 is IOER.

          George, I know its your favorite hobby horse and you have spent a lot of time grooming the old gal, but – and I hate to have to say this my friend – maybe her next outing should be to the glue factory.

          • George Selgin

            Kevin, I just don't get it. My whole point is that IOER made it impossible for the Fed to achieve its inflation target.You responded by saying that I was wrong, that reaching the target was only a matter of making sufficiently large open market purchases. I pointed out how the QE episode contradicted that assertion. Now you reply in turn that I overlook the role of IOER! For goodness sake, the whole point is that above-market IOER undermines ordinary monetary control! Yet having thus thrown my own argument back at me you proceed to conclude that I should abandon it as hopeless! The whole thing leaves me dumbfounded.

          • Kevin Dowd

            George, you are conflating different issues here.

            "My whole point is that IOER made it impossible for the Fed to achieve its inflation target."

            I did not challenge that argument. I think we can agree that IOER prevented the Fed from achieving its inflation target.

            "You responded by saying that I was wrong, that reaching the target was only a matter of making sufficiently large open market purchases."

            I said that you were wrong on other specific points. I never asserted that achieving the target was only a matter of making sufficiently large open market purchases. Let me quote what I wrote:

            "What I actually proposed … was that if the Fed wanted to increase inflation then it could do so by (a) issuing a lot of base money and (b) making no effort to sterilize the new base money. I grant that it did (a) but it did not do (b): Exhibit 1 is IOER."

            Returning now to quotes from your response just above:

            "I pointed out how the QE episode contradicted that assertion"

            The assertion that you contradicted was not one that I made.

            "Now you reply in turn that I overlook the role of IOER! For goodness sake, the whole point is that above-market IOER undermines ordinary monetary control!"

            Let me repeat: I do not challenge the argument that IOER undermines ordinary monetary control.

            I think we have talked this one out, George.

        • Lawrence White

          When Kevin writes "issue base money like crazy and make no effort to sterilize it" he means "make no effort to offset its effect on money held by the public, e.g. M2." When George uses "sterilize" he means "offset its effect on M0." You guys are talking past each other by using the same term differently. There isn't any substantive disagreement here. The Fed's QE programs expanded M0, but didn't expand M2 because of IOER.

          • George Selgin

            No, Larry: it is more than that. In response to my complaint that _thanks to IOER_ the Fed could no longer achieve its inflation target, and that it had therefore lost its ability to perform its most basic duty (meaning the duty of monetary control), Kevin wrote: "So is George suggesting that the central bank cannot make inflation go up or that it cannot control inflation at all? If so, I would suggest that that is a radical departure from sound monetary orthodoxy."

            Naturally I took him to mean that my "suggestion," which clearly took IOER as its starting point, was incorrect. Does his argument not read that way? Does he not seem to suggest that I am mistaken in holding that IOER caused the Fed to lose the ability to control inflation? Read him again before deciding! In any event, I could hardly imagine that, given the context, by assuming away "sterilization" he meant as well to assume away IOER–that is, he meant to allow for the very qualification upon which my own initial claim rested!

            That's not us talking past each other: that is Kevin arguing in a circle! If he understood IOER to be "sterilizing" all along, he ought to have agreed with my original claim, instead of suggesting that it was unsound, and wasting all of our time.

          • Kevin Dowd

            Thanks Larry

            "When Kevin writes "issue base money like crazy and make no effort to sterilize it" he means "make no effort to offset its effect on money held by the public, e.g. M2.""

            This is correct.

            What George means I am not sure, but I would hope that we could all agree with your final sentence that “The Fed's QE programs expanded M0, but didn't expand M2 because of IOER.”

            My concern is that George makes a number of mistakes beyond making this general argument.

            He makes three in his original posting:

            1. “By acquiring trillions of dollars worth of Treasury and mortgage-backed securities, and borrowing from banks to pay for them, the Fed dramatically increased its footprint on the U.S. credit system."

            I believe the phrase “and borrowing from banks to pay for them” is an error that undermines an important argument that George is making, and one that I would otherwise agree with.

            2. and 3. In his posting, George writes that “a central bank that cannot control inflation, and especially one that cannot make inflation go up, is a central bank that is unable to perform it[s] most fundamental duty.”

            I believe this involves at least two errors, namely, (a) that it asserts (in an unqualified way) that the Fed cannot control inflation or make it go up, and (b) it suggests a notion of the Fed’s “fundamental duty” that many (including me) would argue is erroneous.

            Then he makes further mistakes in his comments in this discussion section.

            4. “Paying with a liability or security = "borrowing."”

            5. I say that If I pay for something by writing a check, then that per se is not borrowing, but George says ”Yes, it is, until the check clears.”

            But paying with a liability is only borrowing when one’s net balance goes negative. If I have funds in my account and write a check against them and do not overdraw, then I am not borrowing. I am merely running down the amount I have in my deposit account.

            George insists that our disagreements are semantic and not substantial. I do not believe this to be true as regards the first three points, but he MIGHT be right on points 4 and 5 (in which case we presumably have quite different notions of the term “borrowing” and I would assert that mine is the more natural) and he is right that we do largely agree in the sense that had he made small changes in his blog posting addressing the points I made above (points 1-3 again) then I would have been the first to agree with his analysis. And in so far as George’s main argument relates to the inadequacy of IOER as a policy then YES, I certainly agree.

  • Hu McCulloch

    Zero interest on required reserves (IORR) indeed acts like an implicit tax on bank deposits, equal to the Fed Funds Rate. Banks already pay income taxes on their profits on this business, so I see no reason to have this extra tax, which just induces them to go through contortions (such as retail Sweep Accounts) to evade reserve requirements.

    But interest on excess reserves (IOER) is quite different. Instead of removing a tax, it puts the Fed in the business of financial intermediation, borrowing (at the IOER rate) from some parties in order to lend to others (the Treasury, mortgage borrowers, and bailout clients). The Fed should leave financial intermediation to the financial sector, and National Debt maturity policy to the Treasury.

    • George Selgin

      Thanks for this explanation, Hu. It accurately expresses my own understanding.

    • George Selgin

      Hu, I'd be very interested to have your comments on my written testimony, linked above, which is an extended essay on the damage done by above-market IOER. I am, as you can see, quite in the minority in thinking it has mattered much, but I understand that you may be among the few in my camp and so would very much welcome your thoughts, here or at my Cato email address.

  • Mattyoung

    If the IOER were too high then why would the borrower put up with the charges? Haven't heard any Congress critters complain, mainly because they get the interest subsidy via the seigniorage stream. In fact, the loan undercut the market, committing the same crime.

    • George Selgin

      In this case "the borrower" is the Fed, which is able to pay above-market IOER by (1) using some of its seignorage earnings on non-interest-paying currency to cross-subsidize banks' excess reserve holdings and by (2) taking-on considerable duration risk. In both cases the cost is actually borne by third parties: proximately by the U.S. Treasury, and ultimately by U.S. citizens generally.

      And there have been complaints from some Congressman, including Mr. Barr, to whose subcommittee my remarks were directed.

  • George Selgin

    There was no substantial hoarding of excess reserves between the 1930's and October 2008. As for other causes of the drop in lending, some of them also matter; but it is wrong to treat this as a. either/or proposition. For example, the lower loan demand, the lower the IOER rate required to tempt banks to hoard reserves. Finally, above market IOER is not merely a reduced tax: it is a subsidy; and subsidies are no less harmful to efficient resource allocation than non-neutral taxes can be. That the banks don't mind being subsidized is itself not relevant. What's good for them ( or some of them) may still harm prospective borrowers; and that's just what the Fed does when it encourages them to accumulate excess reserves using above-market IOER.

  • George Selgin

    No, Richard: it's not my logic but your empirical claim that's flawed. Hoarding of excess reserves was not common at all prior to the advent of positive IOER. Here is a FRED chart showing total and excess reserves from 1984 (the earliest date available for excess reserves). The only excess reserve spike is that for Sept. 11, 2001: https://fred.stlouisfed.org/graph/fredgraph.png?g=exdk

    One must go back to the 30s to find another episode of reserve hoarding.

    • RM Salsman

      So the excess reserve hoarding that occurred in the 1930s and in 1999 and in 2001 – when there was no IOER to speak of – simply doesn't count? Selective empiricism is no cure for faulty logic. Also, excess reserves data go much farther back than 1984; for example: http://research.stlouisfed.org/fred2/series/M1486BUSM156NNBR

      • George Selgin

        I myself happily admitted that there was a spike in 2001, and that before that one had to go back to the Great Depression, so I don't believe I was being selective: I merely wished to point out that high excess reserves were exceptional, and I stand by that claim.

        FRED's excess reserve series didn't go back before '84, but I am happy to be informed of the alternative series (it is not strictly one for total excess reserves). That series itself ENDS in '44, and so doesn't quite suffice for the purpose either. It does show, however, that I overlooked the accumulation of excess reserves during WWII, and I readily accept the correction.

        Still I do not believe you will find any postwar accumulation until the 2001 spike, and then nothing until 10-2008. To remind readers of the issue, your claim was that banks "hoarded many times in the past in the absence of receiving interest on reserves"; while my counterclaim was that "Hoarding of excess reserves was not common at all prior to the advent of positive IOER." I do not believe that many persons would consider the almost complete lack of reserve hoarding in the postwar era, until October 2008, supportive of your claim rather than my counterclaim.

        More fundamentally, I never claimed that IOER was the _only possible_ cause of excess reserve accumulation. I have merely maintained, with plenty of evidence that can be found in my written testimony, that IOER is the most likely cause of the recent excess reserve build-up. I understand that you think I'm being illogical in claiming that an increase in the return on reserves from zero to something above market rates on most other liquid short-term assets (also documented at length in my testimony) can have inspired banks to hold substantial excess reserves. I am not sure why you claim that, as it seems to me quite consistent with the ordinary way portfolios respond to changes in relative returns. In any event, it was what Bernanke and Co.intended to have happen when they introduced the policy, as I also document. Perhaps they, too, were guilty of flawed logic. But if so, their plan seemed to work as intended nonetheless.

        • Basho

          "I have merely maintained, with plenty of evidence that can be found in my written testimony, that IOER is the most likely cause of the recent excess reserve build-up."

          Apologies for being so late to this discussion.

          I don't think IOER is a cause for excess reserve build-up at all. Taking the banking system as a whole, aren't excess reserves entirely under the Fed's control? Each bank can seek to get rid of its excess reserves by lending or purchasing securities but, as per your first comment, to the extent it succeeds they will end up with some other bank or banks within the system upon settlement.

          Take your September 2001 example. Per H.4.1 that spike in excess reserves you point to in the chart above stemmed from the Fed adding roughly $60 billion to the system in the days following 9/11.

          I'm not suggesting IOER isn't vitally important. As you say, it necessarily discourages banks from lending or purchasing securities even if how much does so depends on a multitude of factors.

          • Basho

            George,

            Looking through recent posts I see I made the same argument a couple of months ago in response to your "IOER And Banks Demand for Reserves, Yet Again". (http://disq.us/p/1jw23lb)

            My apologies.

            Back then you suggested I should counter your arguments, not simply repeat Whalen's assertions. I didn't because it seemed to me I had broadly acknowledged your points and getting into any sort of repetition didn't at the time seem a good idea!

            Anyway, having raised the issue again I'd like to do so now.

            Back then, you said that "although the Fed's security purchases largely determine the total outstanding quantity of bank reserves (and currency), those purchases don't determine the outstanding quantity of excess reserves . . . ".

            No argument; as I said "Strictly speaking that's true." In your reply you added that "because individual banks' continued efforts to get rid of unwanted excess reserves necessarily contribute to growth in system deposits, these do in effect serve, ultimately, to reduce excess reserves."

            Again, I agree. My only quibble is with the relative importance of this effect. From Sept 2008 to now, system deposits grew by a little over $4.5 trillion while required reserves increased by roughly $100 billion. It was, admittedly, an extraordinary period. Still, if we assume demand deposit growth is about 20% of future growth in system deposits, and reserve requirements stay the same, excess reserves would only fall at a rate of 2% of the growth in system deposits.

            One small final comment. In the June post you suggested the "point is perhaps best illustrated by looking at statistics from before 2008. Back then, banks hardly held any excess reserves; yet ongoing Fed security purchases (and sales) caused the total quantity of reserves to vary considerably, at least by pre-2008 standards:".

            Didn't banks held minimal excess reserves pre-2008 because that was Fed policy? I thought a relatively tight rein on excess reserves was essential for them to hit their target rate through OMOs. It was only the flood of excess reserves in the wake of various Fed initiatives during the crisis that necessitated the introduction of IOER.

          • George Selgin

            "From Sept 2008 to now, system deposits grew by a little over $4.5 trillion while required reserves increased by roughly $100 billion. It was, admittedly, an extraordinary period." You fail to understand my point, Basho, which is precisely that the usual deposit multiplier stopped working after October 2008, thanks to the tendency of banks to accumulate excess reserves because such reserves were then yielding more than short-term loans. Without IOER, banks would have lent, and created corresponding deposits, much more aggressively. It was therefore only owing to IOER, and not to any inherent impossibility of banks' ridding themselves of excess reserves, or to QE per se, that excess reserves accumulated.

            A forthcoming post by Hu McCulloch will make the same point.

          • Basho

            I confess you've got me puzzled, George.

            Had banks (for whatever reason) lent much more aggressively, so that system deposits over that period expanded by say another $4.5 trillion, wouldn't the net result have been excess reserves dropping to about $2 trillion from the current $2.1 trillion (excluding, of course, any Fed instigated changes)? Put another way, based on present settings, required reserves for those additional deposits would probably only be about $100 billion, yes?

            From your posts, you agree that "the Fed's security purchases largely determine the total outstanding quantity of bank reserves (and currency) . . . ". The only way the banking system as a whole can reduce excess reserves is, as you've said, by increasing required reserves through lending and/or purchasing securities. Trouble is, as per the hypothetical above, isn't there an order of magnitude difference between cause and effect? To radically reduce excess reserves the required bank credit expansion would, it seems to me, be almost unimaginably large.

            In any event I'll leave you in peace now. Thanks for the heads up; I look forward to Hu McCulloch's post.

          • George Selgin

            Yes, Basho, the expansion would have to be huge. But no, that doesn't contradict the point. Every hyperinflation shows that, under normal circumstances, confronted with a vast increase in reserves, banks expand deposits enough to maintain desired reserve ratios. Indeed, in most hyperinflations reserve ratios decline. This is a point I have already made in a number of places. (I wish I had $1 trillion for each occasion!)

            Of course the Fed would never have dared to expand B as much as it did, had IOER not killed the usual multiplier, for then it would have generated the massive expansion to which you refer. But none of this contradicts that point that, absent IOER, banks can and generally do dispose of excess reserves, maintaining only relatively slim reserve cushions, and that they can do so no matter how many reserves the central banks tosses at them.

          • Basho

            Fascinating. Thanks. In past episodes, has extreme bank deposit expansion generally been through lending or security purchases?

            The Fed really are walking a rather fine line, aren't they? "Success" on the inflation front could quickly present them with the most exquisite dilemma. If it got high enough to produce meaningfully negative real rates, the doomsday machine you describe would presumably get underway with a vengeance. At that point, does the Fed have any palatable choices? Certainly, both reverse QE and drastic hikes in IOER would be high wire stuff.

            If there are articles or research on the dynamics of past hyperinflationary episodes that you'd recommend (including yours of course), I'd be most grateful for any links.

          • George Selgin

            "In past episodes, has extreme bank deposit expansion generally been through lending or security purchases?" Both. It is difficult to generalize, because all episodes differ. For a brief discussion of the Weimar case, see my recent testimony, p. 19: https://financialservices.house.gov/uploadedfiles/hhrg-115-ba19-wstate-gselgin-20170720.pdf

          • Fed Up

            "Of course the Fed would never have dared to expand B as much as it did, had IOER not killed the usual multiplier, for then it would have generated the massive expansion to which you refer."

            I don't think it was IOER. I think there was very little loan demand.

            What happens if there are excess (central bank) reserves and no entity wants to issue a *new* bond to the commercial bank (borrow to spend)?

            Will the commercial bank buy an *existing* bond?

            Also and pre-2008, did the commercial banks buy *existing* bonds from the fed (the fed was the seller) to keep the fed funds rate on target?

          • George Selgin

            "I don't think it was IOER. I think there was very little loan demand."

            A false dichotomy, Fed Up. It is the high level of IOER _relative to_ other rates (where those are themselves low largely owing to low demand) that matters. Of course if other rates had been higher then the same IOER rates might not have tempted banks to hoard so many excess reserves. But had IOER still been zero, even with low loan demand lending of various kinds would have continued to dominate reserve accumulation.

          • Fed Up

            "But had IOER still been zero, even with low loan demand lending of various kinds would have continued to dominate reserve accumulation."

            I am not sure about that. Exactly what kind of lending are you talking about here? Low to no loan demand means very little buying of newly issued bonds. Are you talking about buying existing bonds (if so, who is the seller)? Are you talking about lending in the fed funds market? Other?

            Back in around 2011, 2-year treasuries were yielding less than .25%. I don't see a commercial bank buying those existing bonds even if IOER was zero.

      • George Selgin

        (To supplement my remarks below.) "From 1959 to just before the financial crisis, the level of reserves in
        the banking system was stable, growing at an annual average of 3.0 percent over that period. This was about the same as the growth rate of deposits. Moreover, excess reserves as a percent of total reserves in the banking system were nearly constant, rarely exceeding 5.0 percent. Only in times of extreme uncertainty and economic distress did excess reserves rise significantly as a percent of total reserves; the largest such increase occurred in September 2001."

        https://www.clevelandfed.org/newsroom-and-events/publications/economic-commentary/2015-economic-commentaries/ec-201502-excess-reserves-oceans-of-cash.aspx

  • Joesph Constable

    What can banks do with their reserves that they are not doing because they received interest on them? Without interest, how would these reserves be employed to help. Banks cannot and do not lend out their reserves.

    • George Selgin

      Joseph, banks can (and, until October 2008, actually did) exchange most excess reserves that came their way for other assets–especially short-term securities. That included "lending reserves" on the fed funds market. But more generally, any time that a bank acquires assets, including in making ordinary loans, it tends to part with a like amount of reserves, other things equal, as the funds it supplies to the borrower or security seller are employed (spent) for whatever uses those parties have in mind. Borrowers' receive credit balances. Checks are drawn, and mostly deposited at other banks. Those other banks "return" them (nowadays electronically) for settlement. The Fed debits the lending bank's reserve account, and credits the accounts of the check-receiving banks. Reserves are thus "lent away."

      The frequently repeated claim that "Banks cannot and do not lend out their reserves" takes a superficial view of matters, instead of tracing out the long-run consequences of a bank's decision to grant credit. All bank lending depends on access to reserves for the sake of settling net clearing debits that tend to accompany the lending, other things equal. Sometimes banks acquire the necessary funds after the lending decision is made. But they must either possess or secure funds to settle with or fail.

      I have addressed this point at greater length many times elsewhere, including in my written testimony linked here. So if you find this brief account unsatisfactory, I hope you'll have a look there.

      • Fed Up

        "Banks cannot and do not lend out their reserves."

        Banks cannot and do not lend out their (central bank) reserves to bank customers. I would say true.

        Banks cannot and do not lend out their (central bank) reserves at all. I would say false.

        "Joseph, banks can (and, until October 2008, actually did) exchange most excess reserves that came their way for other assets–especially short-term securities."

        Not sure if exchange is the right word, but was most of that done with the fed to keep the fed funds rate on target?

        "That included "lending reserves" on the fed funds market." I think those should be separated.

        "But more generally, any time that a bank acquires assets, including in making ordinary loans, it tends to part with a like amount of reserves, other things equal, as the funds it supplies to the borrower or security seller are employed (spent) for whatever uses those parties have in mind."

        I believe the loan process and payment process should be separated from each other. It helps to see what is going on.

        "The Fed debits the lending bank's reserve account, and credits the accounts of the check-receiving banks. Reserves are thus "lent away.""

        I would not call that "lent away". I would call that the movement of (central bank) reserves from one commercial bank to another commercial bank.

  • Ray Lopez

    Yes, RM Salsman is right. An answer to my own question in another thread here about whether reserves at banks vary with the business cycle:
    –as of 2009 bank accounting standards (Tier 2 and Basel) have 'incurred loss' provisions that don't allow extra capital to be 'set aside' unless an actual loss has actually incurred;
    –SunTrust bank ran afoul of the SEC in 1998, when it set aside 'excessive' reserves said the SEC, which made SunTrusts stock price seem more stable than the SEC thought; the chilling consequence of this was that, says Daniel Friedman in a 2009 paper , "A Crisis of Politics, not Economics", that banks did not set aside enough reserves during boom times.

    Perhaps the present 'hoarding' of cash is a natural reaction against the late 90s and early 00s, as implied by RM Salsman?