Garcia-Schmidt and Woodford on Neo-Fisherism

In the 1950s and 1960s, many liberals were anti-anti-communists.  In retrospect that was not a very wise political stance, but it was far better than being a communist. After reading part of Mariana Garcia-Schmidt and Michael Woodford’s new NBER paper on Neo-Fisherism, I’m inclined to call myself an anti-anti-NeoFisherian.

John Cochrane had shown that some New Keynesian models led to NeoFisherian conclusions, i.e. that cutting interest rates leads to lower inflation.  Garcia-Schmidt and Woodford propose an alternative approach, which replaces rational expectations with “reflective expectations”:

We argue that predicting what should happen as a result of a particular policy commitment requires that one model the cognitive process by which one imagines people to arrive at particular expectations taking that information into account. In this paper, we offer a simple example of such an explicit model of reasoning. Under our approach, a perfect foresight equilibrium (or more generally, a rational-expectations equilibrium) can be understood as a limiting case of a more general concept of reflective equilibrium, which limit may be reached under some circumstances if the process of reflection about what forward paths for the economy to expect is carried far enough. Our concept of reflective equilibrium is similar to the “calculation equilibrium” proposed by Evans and Ramey (1992, 1995, 1998): we consider what economic outcomes should be if people optimize on the basis of expectations that they derive from a process of reflection about what they should expect, given both their understanding of how the economy works and (as part of that structural knowledge) their understanding of the central bank’s policy intentions.

Here’s where Garcia-Schmidt and Woodford lose me:

In particular, we show that in our model, a commitment to maintain a low nominal interest rate for a longer period of time “” or to maintain a lower rate, for any fixed length of time “” will typically result (under any given finite degree of reflection) in increased aggregate demand, increasing both output and inflation in the near term, though the exact degree of stimulus that should result depends (considerably) on the assumed degree of reflection. This is true regardless of the length of time for which the interest-rate peg is expected to be maintained, and even in the limit of a perpetual interest-rate peg. Thus consideration of the reflective equilibrium resulting from a finite degree of reflection yields conventional conclusions about the sign of the effects of commitments to lower interest rates in the future, and does so without implying any non-negligible effects of changing the specification of policy only very far in the future.  (emphasis added)

I’m not denying that “the model” produces this result, but how is this any less far-fetched than Neo-Fisherism?  Indeed it seems even worse.  What would you think of a statement by the BOJ that they intended to keep nominal rates at zero forever?  I’d probably have the same reaction as John Cochrane.  Perhaps I’m missing something, these models are way over my head.

Let me propose a solution that is intermediate between the two extremes.  There is no paradox.  Easy money always raises inflation.  But easy money may or may not raise nominal interest rates. Even with complete rational expectations. Consider these two cases:

A.  The BOJ uses pure discretion.  Wages and prices are sticky, and expectations are rational.  The BOJ suddenly reduces the money supply unexpectedly.  Real money balances fall, and as a result short term interest rates rise. Inflation falls. This is all consistent with rational expectations. I don’t know whether it’s consistent with ratex NK models, but if it isn’t then perhaps the models should be revised.

B.  The BOJ commits to a 2% per year depreciation of the yen against the dollar (or 2% more than currently expected). Interest rates rise at all maturities due to the interest parity condition.  They simultaneously do a once and for all currency depreciation large enough to offset the impact of the higher interest rates.  There is no instantaneous change in AD, by assumption, but over time inflation rises by 2% due to PPP.

In both cases the BOJ raised interest rates immediately, and in case A the inflation rate fell and in case B it rose.  One produced a NeoFisherian result and one produced a conventional result.  This is how the world actually works, AFAIK. And we were able to get this result without jettisoning rational expectations.  Also notice that in case A the money supply fell, and in case B it probably rose, and certainly rose in the very long run.

Instead of focusing on ambiguous indicators such as the path of interest rates, we need to come up with a clear, unambiguous real time indicator of the stance of monetary policy.

NGDP futures anyone?

PS.  I would appreciate any help you can offer with the second quoted paragraph.  I should add that Garcia-Schmidt and Woodford are not recommending that central banks make commitments to hold rates low regardless of macro conditions, and later point out that this sort of open-ended policy could lead to wildly unstable results.

HT:  Thomas Powers


Tags:

 
 
 

32 Responses to “Garcia-Schmidt and Woodford on Neo-Fisherism”

  1. Gravatar of luis enrique luis enrique
    7. October 2015 at 05:21

    Scott

    you seem to have seized upon the “even true in a limiting case” as if it’s their main case

  2. Gravatar of marcus nunes marcus nunes
    7. October 2015 at 05:25

    Having taken money out of the equation, “models” are becoming more complicated than reality! And thus of very little use.

  3. Gravatar of Benjamin Cole Benjamin Cole
    7. October 2015 at 05:30

    Well, let me answer this way. Over the wires today (10/7):

    “[Japan] stocks then leapt in mid-afternoon as lawmaker Kozo Yamamoto of the ruling Liberal Democratic Party — one of the architects of Prime Minister Shinzo Abe’s “Abenomics” economic strategy — told reporters in Tokyo that monetary easing is at the heart of the reflationary policies, brokers said.
         Tsutomu Yamada, market analyst at kabu.com Securities Co., said expectations are building for the BOJ to roll out further stimulus at its next meeting. The central bank, while leaving its policy unchanged at its two-day board meeting through Wednesday, is expected to revise its price forecast at the next policy meeting on Oct. 30.”

    Imagine a prominent legislator in the U.S. wholeheartedly advocating monetary easing and more inflation. Or anybody besides me, for that matter.

    But you know what prices are dead in Japan? Demand is weak, that’s why.

    The neoFisherians must think people on the street pay attention to the Fed. They do not. Most people do not know what is the Fed, or who is Yellen.

    In the business world, actors know what they can charge for a service, product or labor. If demand is good and strong, you might raise prices.

    Egads. Do you suppose sellers of real estate look at interest rates and conclude house prices will be dead for years on end and so no sense in raising your selling price? Really? How do you explain the property recovery from 2008?

    And how did Volcker do what he did?

    NeoFisherians must look at helium balloons and conclude anti-gravity waves are at work.

  4. Gravatar of Nick Rowe Nick Rowe
    7. October 2015 at 06:57

    Scott; On the second paragraph. Start in a stationary equilibrium, where nothing is expected to change. Now suppose the central bank lowers the interest rate for T periods, and people know this. Each individual firm thinks:

    1. For the same inflation rate, real interest rates will be lower, so demand for my goods will be higher, so I will raise my price by (say) 1% (when the Calvo fairy lets me) and keep it there.

    2. Hang on! Every other firm will do the same thing, and raise their prices by 1% too, so real interest rates will be even lower, so demand will be even higher, so I should raise my price by (say) 3%.

    3. Hang on! Every other firm will…..

    etc. But after n of these reflections, people stop thinking. So inflation rises, but by a finite amount.

    There is no money in this model, at least not explicitly.

  5. Gravatar of Ray Lopez Ray Lopez
    7. October 2015 at 07:01

    Shades of George Soros’ “reflexivity” in Garcia-Schmidt and Woodford’s “reflective expectations”. Metaphysics.

    The monetarist crowd, like the Ptolemic astronomers of the Middle Ages, will do anything to preserve the fiction that money is non-neutral to a significant degree. This latest attempt is like a epicycle to explain retrograde motion where the earth is the center of the universe.

    Occam’s Razor would simply dictate we do away with the Quantity Theory of Money as a predictive equation, and accept money is neutral. Sumner should show courage and be the first significant practitioner to fall on his sword.

  6. Gravatar of Patrick R. Sullivan Patrick R. Sullivan
    7. October 2015 at 07:19

    From page 471 of ‘The Courage to Act’;

    ‘Prior to the crisis, the Fed affected the federal funds rate by changing the supply of bank reserves. In particular, to raise the funds rate, we would sell some of our securities. The payments we received would reduce bank reserves and leave banks with greater need to borrow on the interbank market. More borrowing by banks would in turn push up the federal funds rate…. But our securities purchases under QE1 had flooded the banking system with reserves, to the point that most banks now had little reason to borrow from each other. With virtually no demand for short-term loans between banks, the funds rate had fallen close to zero. In this situation, moderate reductions in the supply of reserves would be unlikely to affect banks’ borrowing needs and thus influence the federal funds rate. In short, the Fed’s traditional method for affecting short-term interest rates in its pursuit of maximum employment and price stability would no longer work.’

    There are other details of how the Fed manipulated the fed funds rate, which are interesting insights into Bernanke’s mind. Clearly, he’s committed to monetary policy conducted through interest rates.

  7. Gravatar of ssumner ssumner
    7. October 2015 at 07:34

    Luis, My real problem is with moving away from rational expectations. There’s no need to. See Nick’s comment.

    Marcus, Yes, and if you insist on taking money out of the model, replace it with NGDP futures prices, not interest rates.

    Ben, That’s good news from Japan.

    Nick, That’s very helpful. Bennett McCallum once suggested that if we take out rational expectations, then our policy implications will no longer be reliable. This seems like one of those cases.

    Ray, You said:

    “Occam’s Razor would simply dictate we do away with the Quantity Theory of Money as a predictive equation, and accept money is neutral.”

    Kindergardeners say the darndest things!! You do realize that money neutrality IS THE QUANTITY THEORY OF MONEY

    Patrick, Yes, and I think that was a mistake.

  8. Gravatar of Patrick R. Sullivan Patrick R. Sullivan
    7. October 2015 at 09:08

    The discussion on NGDP targeting in Bernanke’s book starts on page 517. On page 522, this must have Milton Friedman spinning in his grave;

    ‘…the fundamental reason that interest rates were low was that a weak economy can’t generate healthy returns on savings and investments. True, our policies pushed rates down even further, but we were doing that to promote economic recovery.’

  9. Gravatar of Nick Rowe Nick Rowe
    7. October 2015 at 10:05

    Scott: ” Bennett McCallum once suggested that if we take out rational expectations, then our policy implications will no longer be reliable. This seems like one of those cases.”

    The Neo-Fisherite equilibrium is extremely sensitive to *exact* rational expectations. If the Fed pegs a nominal rate of interest and lets MB be demand-determined at that rate of interest, then tiny departures from pure RE can lead to massive deviations from equilibrium. We should choose a policy rule that is much more robust, so it doesn’t matter much if expectations aren’t perfectly rational.

  10. Gravatar of ssumner ssumner
    7. October 2015 at 10:59

    Patrick, He seems to think the weak economy just happened.

    Nick, Completely agree about Neo-Fisherism. But I was making another point. The Garcia-Schmidt/Woodford model produces some really implausible results, precisely because they move away from rational expectations.

    We need a model with rational expectations, and without NeoFisherian conclusions.

  11. Gravatar of W. Peden W. Peden
    7. October 2015 at 11:13

    Scott,

    http://www.iea.org.uk/minutes/shadow-monetary-policy-committee-votes-five-four-to-raise-bank-rate-in-october

    Peter Warburton: “A failure to raise UK interest rates materially in 2015 and 2016 opens the possibility that a new downturn may arrive while policy interest rates remain
    extraordinarily low, offering little scope for policy to be eased at that time.”

    A conclusion in search of an argument?

  12. Gravatar of Njnnja Njnnja
    7. October 2015 at 12:04

    The paper spends a lot of effort dealing with the problem that there should be very little difference between a permanent change in central bank policy versus a very long term change in central bank policy. And so in your case A, if the reduction is credibly believed to be permanent, then of course you decrease inflation. But what if the central bank only promises that it is temporary, say, for 2 years. Then both the deflation from the decrease and the inflation from the increase occur at the same time, and there is no change. But what about 10 years? Or 30 years? Or 500 years? How do you allow for ratex, but not get silly results like a “temporary” decrease in the money supply for 500 years is not deflationary?

    So the “reflective” model presented here is about allowing in some level of agent “expectation” without running into the paradox that you get if you assume perfect foresight.

  13. Gravatar of Patrick R. Sullivan Patrick R. Sullivan
    7. October 2015 at 12:37

    ‘Patrick, He seems to think the weak economy just happened.’

    He even compares himself to Bill Murray in ‘Groundhog Day’. Over and over the same stuff just kept happening.

  14. Gravatar of Ray Lopez Ray Lopez
    7. October 2015 at 16:48

    Sumner: “Kindergardeners say the darndest things!! You do realize that money neutrality IS THE QUANTITY THEORY OF MONEY”

    But you’re not reading me correctly (Don Geddis made the same mistake, so you’re in good company, then again I consider him an idiot). I said (***emphasis added***): “Occam’s Razor would simply dictate we do away with the Quantity Theory of Money ***as a predictive equation***, and accept money is neutral.”

    The QTM is an accounting identity, but only looking backwards when all is said and done. Looking forward, as you are attempting to do (as did Friedman) it’s worthless. Marcus Nunes has a nice series of graphics showing how different regimes had different velocities; Friedman was in an expansionary period where it seemed his theory held as a universal truth, until velocity no longer was stable.

    Just do it: fall on your sword. As Bernanke’s book title says, “Courage”. (When will you ever read his FAVAR 2003 paper showing the Fed has very little power over the economy?)

  15. Gravatar of marcus nunes marcus nunes
    7. October 2015 at 19:15

    Scott:
    In the Dec 1982 FOMC discussion of NGDPT, alternate FOMC member Morris anticipated your idea of an NGDP futures market:
    MORRIS. “I think we need a proxy-an independent intermediate target- for nominal GNP, or the closest thing we can come to as a proxy for nominal GNP, because that’s what the name of the game is supposed to be.”

  16. Gravatar of Dustin Dustin
    7. October 2015 at 19:25

    Given a permanent ZIRP, I would expect AD and inflation to rise. What’s the problem with that? I realize the challenge in using interest rates as a measure of policy, but a forever peg implies some sort of sustained monetary injection. Conversely, a forever ZIRP could not be achieved with a permanently static or decreasing monetary base.

    Both the examples you provided highlight the challenge in using interest rates as a policy guide, but they are one-time shock scenarios that would not result in a permanent interest rate.

  17. Gravatar of ssumner ssumner
    7. October 2015 at 20:13

    Thanks W. Peden, I’ll do a post.

    Njnnja, Under ratex a temporary increase in the money supply lasting 500 years can be inflationary for a while.

    Ray, You said:

    “The QTM is an accounting identity”

    I know I can count on you.

    Marcus, Yes, that’s pretty interesting. He’s close.

    Dustin, You said:

    “Conversely, a forever ZIRP could not be achieved with a permanently static or decreasing monetary base.”

    Actually that’s the only way I can see it, in the very long run.

  18. Gravatar of Benjamin Cole Benjamin Cole
    7. October 2015 at 22:12

    Ray Lopez: I retract my earlier statement. I no longer think what you do not say is not unmeaningless.

  19. Gravatar of Njnnja Njnnja
    8. October 2015 at 04:59

    Njnnja, Under ratex a temporary increase in the money supply lasting 500 years can be inflationary for a while.

    That’s precisely the problem that they are trying to deal with in this paper. Using their model for agent decision making, a rational-expectations equilibrium (“perfect foresight”) *does not* give you the inflationary effect for a 500 year (“temporary”) increase in the monetary supply that one would expect.

    But if instead of requiring agents to have perfect foresight, you stop their deliberations after n forward steps, this paper shows that you *can* get the expected result of temporary inflation. Hence, it is a valuable addition to the literature (if you feel that raising rates now would lead to lower long term inflation, not higher).

    By allowing agents to think forward (their cognitive model), they don’t ignore the Lucas critique, but that doesn’t mean that you have to assume the spherical cows of perfect foresight.

  20. Gravatar of jonathan jonathan
    8. October 2015 at 05:53

    I actually like GSW’s approach quite a bit.

    Here’s a simple statement of Neofisherian arguments: Suppose we’re stuck in a ZLB, low inflation equilibrium in a NK-style model. There is also an equilibrium with full employment and high inflation. This equilibrium features higher nominal interest rates (though lower real rates). Then to reach this equilibrium, we set nominal interest rates higher (which involves lower real rates).

    Now this requires inflation expectations to spontaneously jump after the announcement of the higher path of nominal rates. Why do inflation expectations jump? Because that is what is required to bring about the rational expectations equilibrium consistent with those rates that ultimately converges to the steady state.

    Many people find this implausible. Why? In the back of our heads, we have an intuitive story about how people form expectations. And this story says that raising interest rates should make people expect less inflation, not more. And those expectations should lead them to expect still less inflation, etc. We never get to higher inflation expectations.

    GSW try to formalize this notion. Here’s one way to put it. What is a rational expectations equilibrium? Well, suppose agents hold beliefs about future variables X given by Xe, and suppose they optimally pick X based on those beliefs. This defines a mapping from Xe to X:

    F:X -> Xe

    A rational expectations equilibrium is then a fixed point of this mapping: X* = F(X*).

    GSW are saying that we should consider the process of learning using this mapping. Suppose people have beliefs X0. Then they would choose X1 = F(X0). But then they assume everyone does this, and so they update to X2 = F(X1), etc. This produces a sequence Xn. Then we either take the limit of this sequence, or just go to a finite degree of reflection.

    (Actually, GSW do a continuous version of this updating: they define X(n) by integrating forward the differential equation dX/dn = F(X) – X)

  21. Gravatar of jonathan jonathan
    8. October 2015 at 05:56

    Correction (duh):
    F: Xe -> X

  22. Gravatar of TravisV TravisV
    8. October 2015 at 07:31

    Dear Commenters,

    Could someone please write a blog post about these contradictory passages in Bernanke’s new book (Chapter 19):

    “Our goal was to bring down longer-term interest rates, such as the rates on thirty-year mortgages and corporate bonds. If we could do that, we might stimulate spending””on housing and business capital investment, for example…..Similarly, when we bought longer-term Treasury securities, such as a note maturing in ten years, the yields on those securities tended to decline.”

    …….

    “A new era of monetary policy activism had arrived, and our announcement had powerful effects. Between the day before the meeting and the end of the year, the Dow would rise more than 3,000 points””more than 40 percent””to 10,428. Longer-term interest rates fell on our announcement, with the yield on ten-year Treasury securities dropping from about 3 percent to about 2.5 percent in one day, a very large move. Over the summer, longer-term yields would reverse and rise to above 4 percent. We would see that increase as a sign of success. Higher yields suggested that investors were expecting both more growth and higher inflation, consistent with our goal of economic revival. Indeed, after four quarters of contraction, revised data would show that the economy would grow at a 1.3 percent rate in the third quarter and a 3.9 percent rate in the fourth.”

    …….

    FOOTNOTE: “I tried, without success, to get the media and markets to use the term “credit easing,” a phrase suggested by Dave Skidmore, rather than “quantitative easing.” Quantitative easing was the term applied to (unsuccessful) Japanese programs earlier in the decade, which differed from our securities purchases in many respects. In particular, the Japanese QE programs were aimed at increasing the money supply, while the Fed focused on purchasing longer-term Treasury and mortgage-backed securities as a means of reducing longer-term interest rates.”

  23. Gravatar of Ray Lopez Ray Lopez
    8. October 2015 at 07:47

    I see Sumner and his pals are heavy on the insults but they don’t want to say anything substantive. Perhaps they are afraid of making a mistake, knowing full well I would capitalize on it. So they result to ambiguous phrases like “I no longer think what you do not say is not unmeaningless” and other such nonsense like “I can count on you”.

    Face it, you have nothing. But like Friedman you play a good game of poker with a weak hand.

  24. Gravatar of marcus nunes marcus nunes
    8. October 2015 at 08:35

    TravisV Such a post was written 2 years ago!
    https://thefaintofheart.wordpress.com/2013/06/02/somethings-fishy/

  25. Gravatar of marcus nunes marcus nunes
    8. October 2015 at 09:24

    Travis, or this one from 41/2 years ago:
    https://thefaintofheart.wordpress.com/2011/04/25/monetary-policy-powerless/

  26. Gravatar of TravisV TravisV
    8. October 2015 at 09:41

    Thanks, Marcus.

    It would be nice if Bernanke wrote something new about the liquidity effect, inflation effect and growth effect and how they tend to drive interest rates in opposite directions.

  27. Gravatar of Min Min
    9. October 2015 at 11:31

    “In the 1950s and 1960s, many liberals were anti-anti-communists. In retrospect that was not a very wise political stance”

    Really! Civil liberties are for the birds.

  28. Gravatar of Patrick R. Sullivan Patrick R. Sullivan
    10. October 2015 at 09:13

    Min, Communist countries have no civil liberties. So, to be ‘anti-anti-communist/ is to be opposed to civil liberties.

  29. Gravatar of Min Min
    10. October 2015 at 11:09

    Patrick, therein lies the irony of anti-communism. To be anti-anti-communist was not to be pro-communist. It was to be against the policies of the anti-communists, which were inimical to civil liberties. “We have become that which we feared,” as the saying goes.

  30. Gravatar of Patrick R. Sullivan Patrick R. Sullivan
    11. October 2015 at 13:26

    Min, you need to count better.

    Communist = anti-civil liberties
    anti-Communist = anti-anti-civil liberties = pro-civil liberties

  31. Gravatar of tesc tesc
    11. October 2015 at 14:08

    Travis,

    ‘Our goal was to bring down longer-term interest rates, such as the rates on thirty-year mortgages and corporate bonds. If we could do that, we might stimulate spending””on housing and business capital investment, for example…..Similarly, when we bought longer-term Treasury securities, such as a note maturing in ten years, the yields on those securities tended to decline.’

    Correct, Bernanke is noticing that the market from which the FED buys assets, will have a reduction in the interest rates. When buying from the short term market, the interest rate will drop in short term bonds””typical monetary policy tool””because of increase supply of investment funds, the new cash the FED is introducing. But, if the expansionary policy has effect on expectations, the interest rates on other markets, like 10 year bonds, will increase because of higher demand for investment funds.

    Now, what about the within the same market? The new supply of liquid funds drives the interest rate down in a short time period, because of the new supply of investment funds, the FED’s new cash. However, within time, investors will respond by increasing investment demand because the injection of new cash has created credibility that the economy will have an increase in NGDP and therefore, an increase in the nominal value of assets.

  32. Gravatar of Scott Sumner Scott Sumner
    12. October 2015 at 18:20

    Njnnja, Instead of dropping the rational expectations assumption, they should keep ratex and make their model more market monetarist, so that there is no paradox.

    Jonathan, You aren’t addressing the argument in this post.

    The problem with dropping ratex to “disprove” NeoFisherism, is that NeoFisherism is true under certain conditions. And those conditions do not necessarily imply an implausible level of foresight, as I showed in this post. They’d be better off keeping ratex, and figuring out when NeoFisherism holds and when it doesn’t.

    Min, You may be too young to remember what it was like. Don’t rely on Hollywood films. Anti-communists were sneered at, regardless of their views on civil liberties. The anti-anti-communists should have been anti-communists who favored civil liberties.

    You said:

    “It was to be against the policies of the anti-communists,”
    which were inimical to civil liberties.”

    Not even close. You might want to go back and read things that liberals said about Mao and Castro and Ho and Che back in the 1960s. It’s pretty embarrassing.

    By the way, it’s obviously symmetrical. Liberals abused civil liberties going after fascist sympathizers. That’s regretable, but it shouldn’t stop someone from being an anti-fascist.

Leave a Reply