Expansionary OMOs are still expansionary

Tyler Cowen links to a John Cochrane post, which discusses a WSJ piece by Andy Kessler:

The usual thinking is that bank reserves are “special.” They are connected to GDP in a way that Treasuries are not.  In the conventional monetary view, MV = PY.  Bank reserves, through a multiplier, control M. The bank or credit channel view says that bank reserves control lending and lending affects PY. The red M&Ms, though superficially identical, have more calories.

In Andy’s view (my interpretation), that is turned around now. Now, Treasuries supply more “liquidity” needs than bank reserves, and (more importantly) the supply of treasuries is more connected to nominal GDP than is the supply of bank reserves.

Part of this inversion of roles is supply. In place of the usual $50 billion, we have $3 trillion or so bank reserves. Bank reserves can only be used by banks, so they don’t do much good for the rest of us. Now, they just sit as bank assets in place of mortgages or treasuries and don’t make a difference to anything. More treasuries, according to Andy, we can do something with.

More deeply, constraints only go one way. Normally, the banking system is up against a constraint. Reserves pay less interest than other assets, so banks use as little as possible. Now, they are awash in liquidity. You can’t push on a string, as the saying goes.

It makes me a bit dizzy seeing John Cochrane using the Keynesian “pushing on a string” metaphor that Milton Friedman discredited decades ago.  More seriously, there are several fallacies in this argument.  Let’s assume for the moment that reserves and T-bills are both non-interest-bearing; why would monetary policy have any impact?  The answer is simple; markets don’t expect interest rates to be at zero forever.

OK, but why does that make base money special now, when T-bills also pay no interest?  Because base money is the medium of account, whereas T-bills are not.  In the future a change in the supply of T-bills will not (significantly) change the price level, it will merely change the nominal price of T-bills.  In contrast, a change in the supply of base money can never affect the nominal price of base money, which is always one.  Instead, the price level and NGDP adjust to changes in the market for base money.  None of this has anything to do with the lending channel.

Some might argue that this will only work if the central bank has credibility, if investors believe that QE will raise the future level of the base, and hence future NGDP.  But we know for a fact that investors do believe this, as markets react violently to even tiny hints of changes in QE.  And they do so in a way that clearly indicates that QE is expansionary.  That’s why we know for certain that Milton Friedman and the MMs are correct, and John Cochrane and Andy Kessler are wrong.  As a fellow Chicagoan, I strongly urge Cochrane to dump the Keynesian pushing on a string metaphor, and go back to the Chicagoan monetarist tradition.



29 Responses to “Expansionary OMOs are still expansionary”

  1. Gravatar of jknarr jknarr
    24. May 2013 at 11:09

    Scott, isn’t he right in that reserves are only a medium of account for the banking system? How would greater bank-only reserves affect nominal prices, except through lending?

    Expectations don’t seem sufficient — in the early 1930s, the Fed also expanded reserves dramatically — absorbing gold in the process — and it similarly failed to boost NGDP.

    Bank-only reserves can be demanded/translated into public currency, but only with the intervening creation of private demand deposits. Zero rates imply very low private demand for borrowing, and so reserves languish, no?

    In other words, there is no pushing on a string, but there is pushing on excess reserves — if the Fed truly wished to boost NGDP, it should create private demand deposits by buying non-bank-held assets.

  2. Gravatar of Scott Sumner Scott Sumner
    24. May 2013 at 11:50

    jknarr, That’s not really accurate. When the Fed tried QE in 1932 there was a heavy gold outflow, which reduced the expansionary impact. It’s much hard to have a credible inflationary policy when you have a currency pegged to gold. As soon as the gold peg was dropped, prices rose sharply. As we’ve seen in the US and Japan, a central bank can adopt an expansionary policy anytime they wish. The problem has been getting them to wish to do so.

  3. Gravatar of Scott Sumner Scott Sumner
    24. May 2013 at 11:52

    The key is to focus on market responses to monetary policy initiatives. The central banks need to let markets guide their policy. Markets currently are indicating that more QE boosts expected NGDP growth.

  4. Gravatar of jknarr jknarr
    24. May 2013 at 12:46

    I think that the 1930s vector was largely the US Treasury — the Fed was buying a few Treasurys in 1932 to make up for the (small, it appears to me) gold outflows and the ending of the discount bill program. Massive reserve formation only took off in 1933-34 when gold holdings accelerated.

    Reserves appear inextricably tied to gold on the Fed’s balance sheet of time — the real NGDP action appeared to be linked to the Fed’s (relatively new) purchasing of Treasury bonds.

    If credibility is the biggest factor, the BoJ should go down to the Tokyo fish market and buy tuna at huge premiums! This would create demand deposits, and ensure that the easing will never be unwound — old fish and all. 😉

  5. Gravatar of Bill Ellis Bill Ellis
    24. May 2013 at 13:05

    Red meat alert.
    Krugman is talking Monetary policy. Says that Ball’s case for a 4% inflation target sounds good. No mention of NGDPLT… 🙁

    The basic point is that a higher baseline for inflation would make liquidity traps, in which conventional monetary policy is up against the zero lower bound, less likely and less costly when they happen. Ball estimates that if we had come into this crisis with an underlying inflation rate of 4 percent, average unemployment over the past three years would have been two percentage points lower. That’s huge “” it amounts to millions of jobs and trillions of dollars of extra output.


  6. Gravatar of Edward Lambert Edward Lambert
    24. May 2013 at 13:07

    We need an important clarification… QE is expansionary for capital, not labor. Thus, it is relatively contractionary for labor.
    We see clearly now that there is a capital sector of liquidity, and a labor sector of liquidity in the economy. Monetary policy has no direct effect upon the labor sector of liquidity, but it does upon the capital sector of liquidity.
    What is expansionary for capital may be contractionary for labor.
    Case in point… Record highs for stock market, while Food stamps reach new highs.

  7. Gravatar of W. Peden W. Peden
    24. May 2013 at 13:13

    Bill Ellis,

    NGDP targeting… Perhaps mean RGDP (i.e. GDP) targeting? I’ve never heard of “NGDP targeting”, but I have heard about interest rates (a monetary policy tool rather than a target) and inflation targeting (when the central bank tightens money if inflation is above a certain level).

    Why would Krugman mention such an idea? I’ve never heard of it and I’m SURE that Krugman hasn’t either.

    A 4% inflation target (with fiscal stimulus if monetary policy shoots its load i.e. interest rates go low and conventional monetary policy loses effectiveness) makes sense to me.

  8. Gravatar of W. Peden W. Peden
    24. May 2013 at 13:15

    Edward Lambert,

    How cruel of you to tease us with exciting claims and then be so coy about arguments for them!

  9. Gravatar of Fed Up Fed Up
    24. May 2013 at 13:31

    MV = PY

    You say M equal monetary base = currency plus central bank reserves, right?

  10. Gravatar of J J
    24. May 2013 at 13:39

    Given that the Fed won’t hit its 2% inflation target, I’m not sure a 4% inflation target would do much good. The Fed is not bound by any of its rules or guidelines or economic constraints. It is limited by its fear of getting yelled at for letting inflation get even close to 2%. That is the only explanation that makes sense anymore, assuming the Body Snatchers haven’t paid the FOMC a visit.

  11. Gravatar of JimP JimP
    24. May 2013 at 13:50


    from the ever fun A E-P


    The critics are right to say it may fail. But are they suggesting that the previous status quo was tenable? If Mr Bass happens to read this blog, I would like to know what he would do if he were prime minister of Japan.

    As for the countless readers demanding an apology from me for backing QE, Abenomics, and all the sins of monetarism: I defy you all.

  12. Gravatar of Fed Up Fed Up
    24. May 2013 at 13:58

    I think I better ask my question more completely.

    MV = PY

    M = money = money supply = medium of account (MOA) = monetary base = currency plus central bank reserves, right?

  13. Gravatar of jknarr jknarr
    24. May 2013 at 14:02

    Have I missed something? — AEP: “What they should have done is to conduct old-fashioned open-market operations, à la Friedman, Fisher, Hawtrey, Cassel, or Keynes himself, buying long bonds from non-banks to force up the M3 money supply. That works, as Ben Bernanke discovered when he finally alighted upon the policy by accident late in the Fed’s QE efforts.”

    I agree with the diagnosis, but has the Fed bought from non-banks? How would we know? The balance sheet expansion has almost entirely been funded by excess reserves, which means they have only been buying from banks, and so unsurprisingly, NGDP is still mired at 3.4% after five years.

    Note also that (even worse) QE2 and QE3 purchases were funded by reserve creation at US branches of foreign banks — not at large or small domestic banks in the US. Note that foreign banks have only trivial lending operations in the US, also — I’m not sure how excess reserves in Europe are supposed to generate NGDP in the US.

  14. Gravatar of ChargerCarl ChargerCarl
    24. May 2013 at 14:58

    From what I’ve read of Cochrane he doesn’t seem to think demand shocks exist at all. It’s all supply 100% of the time for him.

  15. Gravatar of DOB DOB
    24. May 2013 at 15:28

    “OK, but why does that make base money special now, when T-bills also pay no interest? Because base money is the medium of account, whereas T-bills are not.”

    This makes no sense. What if the Fed tomorrow printed red dollars instead of greens and said that greens are the medium of account but reds aren’t, but they’ll exchange one for another at 1:1 on demand? If they converted half of the money supply from green to red, do you really believe this would have any macro significance? Of course not, people would treat redbacks and greenbacks as perfect substitute.

    T-bills and cash are perfect substitutes at the zero bound. Reserves created today can be destroyed tomorrow. The Fed is NOT implying or even hinting that increase it makes in the base today is going to stay for some period of time.

    QE’s effect is entirely due to pressure on longer term interest rates, and not at all due to increase in MBase. If the Fed repo’d out all the Treasuries it purchased through QE (“sterilizing”) and the base went back to $1tn, it would have zero macro impact. In fact it would be good from a housekeeping standpoint.

    Quantity only matters insofar as it impacts interest rates. Interest rates always matter. Therefore there’s no reason to ever think about Quantity.

  16. Gravatar of 123 123
    24. May 2013 at 15:46

    Kessler and Cochrane have an old fashioned broad money monetarist claim. It may well be the case that if you do the proper Divisia style calculation, when the Fed purchases the treasuries, the broad aggregates contract. But the effect is small ( it is basically a tax on the repo business ), and the expected future policy channel is much stronger.
    That said, it would be better if the Fed focused on buying a broad diversified private sector asset portfolio.

  17. Gravatar of Daniel J Daniel J
    24. May 2013 at 18:43

    I don’t know what they put in his coffee but Avent just wrote a tour de force. http://www.economist.com/blogs/freeexchange/2013/05/busts#comments

  18. Gravatar of flow5 flow5
    24. May 2013 at 18:52

    Big differences between credit creators & credit transmitters. Lending by the non-banks (NBs) is non-inflationary. And money flowing thru the NBs never leaves the CB system. The NBs, pre-Great/Recession, represented 82% of the pooling & lending markets (Z.1 release: MMMFs, CP, GSEs…).

    In order to eliminate Reg Q ceilings, the BOG turned 38,000 financial intermediaries into 38,000 CBs. The resulting “free for all” caused the 2nd S&L crisis.

    The IOeR policy (by stopping the flow of savings thru the NBs, as well as sopping up safe-assets), destroyed the NBs.

    The economy can’t “recover” until the CBs are put out of the savings business (prevented from out bidding the NBs for savings), & the remuneration rate is the principal problem.

  19. Gravatar of ssumner ssumner
    25. May 2013 at 05:59

    jknarr, Yes, but after 1933 NGDP did grow rapidly.

    Bill Ellis, I prefer NGDP targeting, but I certainly agree with much of the logic in that post.

    Edward, ???????

    Fed up. Yes.

    Jknarr, The fact that almost all the new money is ERs does not imply the bonds have been directly purchased from banks

    JimP, He’s always excellent.

    DOB, You said;

    “The Fed is NOT implying or even hinting that increase it makes in the base today is going to stay for some period of time.”

    Just the opposite–as we know from the Evans Rule. But I agree that they aren’t doing enough of that.

    And as for monetary policy working through interest rates, I presume you have not been following recent events in Japan, where interest rates are virtually unchanged from 6 months ago.

    123, I see nothing “market monetarist” in his post. And I oppose the buying of private assets–there’s no need to when there is plenty of Treasury debt to buy–enough to create hyperinflation.

    Daniel, He’s always excellent.

  20. Gravatar of 123 123
    25. May 2013 at 06:31

    Let me put it this way. Due to the all the private sector repo infrastructure that exists, every treasury note is a virtual bank, with a treasury as an asset, and a repo liability on the other side of the balance sheet. So if the Fed buys the treasury, in effect it is buying a virtual very liquid private sector asset that should be counted in M3.
    Suppose Bank of England did QE by absorbing all those Scottish private sector one pound notes, such as Clydesdale bank one pound notes, and as a result the turnover in the Glasgow pubs goes down a bit. Bernanke is doing the same – thats the point Cochrane and Kessler are making.

  21. Gravatar of Joe Eagar Joe Eagar
    25. May 2013 at 07:24

    I don’t see why acknowledging that base money and Treasuries compete as mediums of account must lead to a “pushing on a string” model. That might be true if a country had a significant current account surplus and no other country was willing to run a corresponding deficit, leading to a depressed level of global demand (as happened during the Great Recession, and may be happening in the eurozone today).

    But the U.S. has a current account *deficit*. So long as our trading partners are willing to tolerate a reduction in their surpluses (and I think most of them are, aside from Germany), monetary policy can turn U.S. fiscal contraction into real growth by accelerating the adjustment in the real exchange rate.

  22. Gravatar of Fed Up Fed Up
    25. May 2013 at 08:16

    I said: “I think I better ask my question more completely.

    MV = PY

    M = money = money supply = medium of account (MOA) = monetary base = currency plus central bank reserves, right?”

    ssumner said: “Fed up. Yes.”

    Duly noted and recorded. zyxzyxooxmi

    I 110% disagree about how M is defined.

  23. Gravatar of ssumner ssumner
    26. May 2013 at 06:43

    123, Yes, that’s the point they are making, but it’s clearly wrong, as we see from market responses to QE.

    Joe, Some liquidity trap believers claim the Fed cannot depreciate the dollar.

  24. Gravatar of ssumner ssumner
    26. May 2013 at 06:44

    Fed up. Surely you don’t disagree that I define it that way. So I presume you think that this is not a useful definition? All definitions are arbitrary.

  25. Gravatar of 123 123
    26. May 2013 at 07:57

    Scott, their argument is right, but not very important, as the expected future policy channel is so powerful. They must have friends in the Glasgow pub business (or repo business) if they are focusing on this.

  26. Gravatar of Fed Up Fed Up
    26. May 2013 at 09:39

    “Fed up. Surely you don’t disagree that I define it that way. So I presume you think that this is not a useful definition? All definitions are arbitrary.”

    I don’t find M = money = money supply = medium of account (MOA) = monetary base = currency plus central bank reserves to be a useful definition of M in MV = PY.

    Let’s start with this:

    Can we agree that central bank reserves do not directly circulate in the real economy?

  27. Gravatar of Jeff Jeff
    26. May 2013 at 13:02

    Cochrane is talking the Gary Gorton theory of the financial crisis: Safe assets (like Tbills) are used as collateral for repos (a particular kind of loan), with a haircut. The money loaned can be used to buy other safe assets, which again serve as collateral. And so on, and so on. It’s the money multiplier story we tell people in Macro 101 about how a $1 increase in reserves ends up boosting deposits by $10 when the reserve requirement is 10 percent.

    Gorton thinks the transmission mechanism for the crisis was an increase in haircuts, which had the same effect on the repo market as an increase in reserve requirements would have in the deposits world.

    What Gorton misses with this story is repos are not money. Nobody uses repos as a medium of account or a medium of exchange. Gorton’s story explains credit contraction, but it does not say anything much about money.

    Policymakers should be forced to write on a blackboard 100 times every day “Credit is not money. Credit is not money. Credit is…”

    It is the same mistake people make when they think of monetary policy as setting an interest rate.

  28. Gravatar of dtoh dtoh
    29. May 2013 at 03:46

    Cochrane has been expounding the “pushing on a string methaphor” for ages.

    If you accepted the financial asset price mechanism, you wouldn’t need to go through the mental gymnastics of explaining how the base is the medium of account.

    It’s much simpler. Even at the ZLB, OMP raise inflation expectations = higher real prices of financial assets which results in a marginal increase in the exchange of financial assets for real goods and services = higher NGDP.

    Scott, this is trivial and irrefutable. If you relied on the asset price transmission mechanism to explain MM, you would only have to write half as many blog posts to explain why otherwise smart people like Cochrane believe dumb things.

  29. Gravatar of QE discussion fun | TVHE QE discussion fun | TVHE
    30. May 2013 at 12:01

    […] this year.  So I missed this post by Cochrane (including a bunch of great comments) and a reply by Sumner.  […]

Leave a Reply