Tuesday, November 24, 2015

Early Fisherism

John Taylor has an interesting blog post with a great title, "Staggering Neo-Fisherian Ideas and Staggered Contracts." John goes back to a paper he wrote in 1982 for the Jackson Hole conference, on the issue of that time, how to lower inflation. He presented simulations of a model with staggered wage setting, which I reproduce below.


So as far back as 1982, here is a model in which lower interest rates correspond with lower inflation, both in the short run and the long run.  John's model has money in it, so the mechanics are a pre-announced monetary contraction.

Sargent's famous "Ends of four big inflations"  tells an even more radical story.

On solving the governments' fiscal problems, inflation ends instantly. Sargent and Wallace alas do not have interest rate data, but from the inflation data it's pretty plausible that interest rates fell like a stone when the fiscal reforms are implemented. They have money stock measures -- and the ends of these inflations did not have any monetary tightening at all. Money stock measures all expanded substantially as inflation ended.

I've been having an interesting back and forth with a correspondent about Milton Friedman's views. In  "Do higher interest rates raise or lower inflation?" I quoted Friedman's 1968 address, and said he believed that an interest rate peg is unstable. Not so fast says my correspondent, and passed on a lovely memo written by Milton Friedman -- better still once owned by Anna Schwartz. (Yes I checked that it's ok to post this)




and later



As I read this quote, Friedman emphasizes that lower interest rates come only with lower inflation in the long run, so there is some Fishery theory here. But in the short run, if the Fed lowers money growth, then interest rates will first rise but then decline as inflation declines. So the implied short run relationship goes the other way.

As I read it, then, Friedman says it is possible to target interest rates. But to do so requires particularly active money growth policy to offset the instability that would result from simply announcing a fixed interest rate.

That leads to a very interesting question, how the same interest rate path could be supported by different money growth paths.

23 comments:

  1. Could you publish the whole paper? It would be an interesting read. It comes from a time when (IIRC) inflation was just taking hold as a result of rising deficits.

    My subjective experience in the early 1980s was that high interest rates stopped inflation in house prices in its tracks.

    ReplyDelete
  2. Some basic ideas that are coming out of your blogs:

    1. Permanent Inflation results from permanent changes in monetary supply.
    2. Changes in monetary supply are the net of treasury's borrowing and Central Bank lending.
    3. Short-term changes tend to have the opposite long-run effect. (For example, a shock to VIX or risk will shoot securities down, but increase the returns after the shock. A binge of treasury borrowing will lower cash supply immediately, but raise it in the future when the treasury pays interest on what it borrowed.)
    4. The risk-free rate reflects expectations of net changes to monetary supply. (So long as central banks can control those expectations, they can control risk-free rates. )
    5. Real interest rates result from volatility (ie: volatility raises the value of the implied short-call on profits.)
    6. Central Banks can only reduce economic volatility as much as they can forecast price changes.
    7. If Central Banks can't forecast prices, they can't reduce volatility, which means they can't reduce real interest rates either.

    ReplyDelete
  3. Only Sargent wrote The end of four big inflations

    ReplyDelete
  4. Interest rates were pegged during the big inflation in Germany:

    http://informationtransfereconomics.blogspot.com/2015/10/pegged-interest-rates-hyperinflation.html

    Afterward the inflation ends they shot up, fell back down only to be pegged again in the 1930s.

    ReplyDelete
  5. Yes, Milton Friedman said low interest rates are the result of tight money, and that a central bank cannot tighten its way to higher interest rates, at least not for a long.

    NeoFisherianism raises an interesting point: If inflation can be quelled by pegging interest rates at zero, then should the goal of macroeconomic policy also be "labor shortages"?

    Seems so... not the least because workers vote. How will workers vote in a nation of chronic high unemployment, or a nation of labor shortages? I suspect there will be greater appreciation of free enterprise in the latter scenario.

    Fighting inflation by suffocating the economy is the road to socialism.

    ReplyDelete
  6. We seem to have come full circle, back to where we started. What Friedman says in that letter corresponds very closely to what I always understood as orthodox opinion, before the Neo-Fisherian kerfuffle started.

    It matters a lot whether you think of the money growth rate as the central bank instrument, so nominal interest rates are endogenous (which is how Friedman viewed the world), or whether you think of the nominal interest rate as the central bank's instrument, so money growth is endogenous (which is how central bankers view the world).

    I note that John Taylor makes exactly this point too: "In the simulations, the money supply provides an anchor and the interest rate is determined in the markets. The Fed is effectively saying that it will set the money supply path and the market will then set the interest rate according to the path."

    ReplyDelete
    Replies
    1. I agree that Friedman essentially has the conventional view -- negative short term relation, positive long term relation. And it's still pretty clear he does not approve of an interest rate target! Both 1970 Friedman and 1982 Taylor are thinking in terms of money targets, whereas more recent theory reflects the fact of pure interest rate targets. These are just interesting early views.

      Delete
    2. OK. But I would like to distinguish between targets and instruments (very short term targets?). I would say the Fed (like most central banks) uses an interest rate instrument to hit an inflation target.

      Let's rephrase the question this way: suppose the Fed uses an interest rate instrument. If the Fed wanted to increase the money growth rate, should it raise or lower the nominal interest rate instrument? The conventional view says "lower". So people and banks will borrow more, so the money stock expands endogenously, raising the inflation rate. But eventually the Fed must raise the nominal rate to match the higher money growth rate and inflation rate, or else you get hyperinflation.

      Delete
    3. By the way: good find on that Friedman letter, and thanks for posting it.

      Delete
  7. I think the entire argument only holds in a semi closed economy like Japan. In the US, capital flows combined with the trade deficit do more to swing monetary conditions than minor action of the Fed. I think interest rates had much more traction in the pre Plaza Accord days when the US dollar was still in relatively short supply in terms of world trade flows, and higher effective interest rates.

    One of the greatest difficulties is looking back to the gold standard transition period and aftermath, and characterizing anything as a new system came into use.

    I don't think of this period as reflective of anything before or after- roughly 1968-1985.

    ReplyDelete
  8. Add on: How does Sargent (or others) explain Japan's chronic minor defaltion in the face of large perennial national deficits?

    And, here in the United States, we just went through a period of large federal deficits yet falling inflation.

    Indeed, if we are to believe the Fed, the outlook for inflation is worsening even as budget deficits shrink!

    Something else must cause inflation other than national deficit. Not that I think minor inflation is much to worry about.

    The real purpose of macroeconomics should be robust economic growth.

    ReplyDelete
  9. What is the relationship between Friedman and Neo-Fisherism, if there's any?

    What Friedman was clearly saying is that if you want, say, lower inflation tomorrow, you lower the money growth rate (that is, raise the nominal interest rate) today, and after inflation is down interest rates will go down as a result, reflecting lower inflation. The opposite happens if - at full employment - the money growth rate rises: the nominal interest rate goes down, inflation increases, and interest rates increase as a result, reflecting higher inflation.

    (Incidentally, as rational exceptions were introduced in this mechanism, this has created the illusion that, even in a liquidity trap, it only takes for central banks to pump more money in and higher inflation will follow, forgetting that if people want to stay liquid they don't spend the additional money and inflation won't move... And similar ineffectiveness - as far as expected inflation goes - one gets in an economy with large unemployment, where more money would most likely have an impact on output prior than on prices, so that targeting higher inflation in order to stimulate spending would simply get the causation wrong...)

    Under Neo-Fisherism, instead - if I understand it right - central banks wanting to target a higher rate of inflation should announce a higher nominal interest rate, since this - based on Fisher's equation - would cause higher expected inflation. It seems to me that this is exactly the opposite causation nexus than in Friedman.

    I gather that Friedman was a Fisherian but not a Neo-Fisherian. Am I right?
    Thank you for this post and for this Friedman's document.

    ReplyDelete
    Replies
    1. The biggest difference between Neo-Fisherism and Friedmanism is that Friedmanism is a policy prescription (k% money growth rate to stabilize the inflation rate) while Neo-Fisherism is an observed relationship (inflation rate tends to rise with the nominal interest rate - all else being equal).

      Finally, Friedman and Neo-Fisherism are both are silent on the debt supply.

      Consider the action matrix of the central bank:

      1A. Buy government bonds - Central bank sets price they will pay, market sets quantity it will sell (As currently practiced by U. S. fed)

      1B. Buy government bonds - Central bank sets quantity it will buy, market sets price it will sell at (U. S. Fed cannot do this, market may set price higher than what the bonds are worth)

      2A. Sell government bonds - Central bank sets price it will sell at, market sets quantity it will buy. (As currently practiced by U. S. Fed)

      2B. Sell government bonds - Central bank sets quantity it will sell at, market sets price it will buy at (U. S. Fed cannot do this, market may set price at a significant discount to what the bonds are worth)

      3B. Lend money - Central bank sets price (interest rate) it will lend at, market sets quantity of loans it will take on (as currently practiced by U. S. Fed)

      3A. Lend money - Central bank sets quantity of loans it will give, market sets price of loans it will take on (possibility of negative interest loans may preclude U. S. Fed from doing this)

      4B. Call loans - Central bank sets price of called loans, market sets quantity of loans that it will pay back at behest of central bank (This is an odd one, usually the U. S. Fed lends for such short time frames that it's unlikely to happen though it is conceivable).

      4A. Call loans - Central bank sets quantity of loans that it will call, market sets price for called loans (Again this is an unusual Fed action and unlikely given the potential for the market to set the call price of direct Fed loans at $0)

      And so the U. S. Fed currently practices 1A, 2A, 3B, and potentially 4B.

      From a Friedmanite view - 1A and 3B may increase the money supply.
      From a Neo-Fisherian view - 2A and 3B may increase the nominal interest rate.

      Both would agree that action 3B may raise the inflation rate - Friedman from increased money supply, Neo-Fisherian from increase nominal interest rate.

      Actions 1A and 2A leave room for disagreement between the two camps. Friedmanites would say action 1A will raise the inflation rate (increase money supply), Neo-Fisherites would say that action 1A will lower the inflation rate (lower interest rates) - score one for the Neo-Fisherites.

      Friedmanites would say action 2A will lower the inflation rate (decrease the money supply), Neo-Fisherites would say action 2A will raise the inflation rate (higher nominal interest rate). It remains to be seen whether a Fed unwind would result in higher or lower inflation.

      Finally, both Friedmanism and Neo-Fisherism leave out productivity gains, fiscal position, and a lot of other things so take what I am saying with a tablespoon and two grains of salt.

      Delete
  10. Biagio,

    Neo-Fisherism doesn't make the leap to causality or policy recommendations. It simply provides an observed relationship - increases in the nominal interest rate offered by the central bank tend to raise the inflation rate absent other changes (productivity change, public / private demand for credit, etc.).

    In fact, what should be apparent from a U. S. standpoint was that the central bank during the 1970's was responding to the increasing demand for credit in the U. S. (public and private) The central bank was presumably hoping that by raising the nominal interest rate they could cool that demand for credit.

    The rise in the inflation rate during the same period was a result of the increased debt load (interest and principle) with no corresponding increase in productivity.

    And so the notion that raising the nominal interest rate will always and everywhere raise the inflation rate is overly simplistic.



    ReplyDelete
  11. Frank,

    Thanks for your clarifications. I have two observations on two of your points:

    "Neo-Fisherism doesn't make the leap to causality or policy recommendations. It simply provides an observed relationship"

    But then it doesn't say anything new besides what's already in Fisher equation, after all.

    "Friedmanites would say action 1A will raise the inflation rate (increase money supply), Neo-Fisherites would say that action 1A will lower the inflation rate (lower interest rates) - score one for the Neo-Fisherites."

    Do Neo-Fisherites really score on this one? It seems to me that if you keep pumping money in an economy with large resource unemployment, you may see large increases in resource absorption before you see any significant rise in inflation. Moreover, the way money is pumped in the economy matters: QE, for instance, is a very indirect way of stimulating demand, so you may see slow output responses and even slower price responses following large injections of money. Not to mention those cases where liquidity preference is so high that velocity of circulation drops and no rate of money growth can see to get any output or price effect before confidence is restored; in fact both output and prices might even continue to weaken, keeping nominal interest rates at their lowest: nothing so much Neo-Fisherian, I should note.

    Would Neo-Fisherites score on this only because they observe nominal interest rates and inflation going down at the same time? And would they then be right in drawing from this that raising nominal interest rates would drive inflation up?!


    ReplyDelete
  12. Biagio,

    "But then it doesn't say anything new besides what's already in Fisher equation, after all."

    Correct.

    "Do Neo-Fisherites really score on this one? It seems to me that if you keep pumping money in an economy with large resource unemployment, you may see large increases in resource absorption before you see any significant rise in inflation."

    I only mention that the Neo-Fisherites score on this one given recent events. Obviously, the U. S. government has undergone a belt tightening, private credit demand is restrained, and so isolating cause and effect is murky at best - is it low interest rates, reduced private credit formation, or fiscal consolidation that has kept inflation at bay?

    Which brings me to this - neither Neo-Fisherites nor Friedmanites address the supply and demand for credit. The central bank does not directly "pump" money into an economy. They loan money to the private banking sector and / or they buy government bonds from private banks so that those private banks can lend money into the real economy.

    And so before trying to gauge the economic effect of central bank actions, the first thing you need to look at is what effect those central actions have on the supply and demand for credit.

    That is why actions 1A and 2A create conflict between Neo-Fisherites and Friedmanites.

    Friedmanites might say that the increase in monetary aggregate will be immediately lent back into the economy by the private banking system regardless of risk, profitability, or any sane accounting measure.

    Neo-Fisherites might say that private banks are for profit enterprises that have considerations other than the amount of money available to lend - what is my profit spread, what is my exposure to particular parts of the economy (overexposed to mortgages for instance), what are my shareholder expectations for future profitability, etc - meaning funds available for lending may not get lent.

    ReplyDelete
  13. Biagio,

    "And would they then be right in drawing from this that raising nominal interest rates would drive inflation up?!"

    Again, Neo-Fisherism isn't about drawing a causal link between interest rates and inflation nor is it a policy recommendation that tells the central bank to raise it's interest rate target to raise the inflation rate - note, no where in the Fisher equation is an interest rate target even addressed.

    If you are looking for a causal link to inflation then you must consider the risk free rate of return that is available in an economy (not the interest rate).

    ReplyDelete
  14. Thanks Franks.
    Fully agree on your first comment.
    Regarding your last one, I actually thought that Neo-Fisherians do reverse the causal relationship between inflation and the nominal interest rate, and that they do derive prescriptive propositions. I thought they criticize the Fed for keeping interest rates low, thus causing low inflation, and that the remedy would be of course doing the opposite. But of course I could be wrong, and I will double check.
    In any case, if they don't draw a causal link and don't derive policy recommendations, what's so new in what they say that Fisher hadn't said already?
    I'd be glad if also Prof. Cochrane would comment on this and correct me.

    ReplyDelete
    Replies
    1. Biagio,

      "I thought they criticize the Fed for keeping interest rates low, thus causing low inflation, and that the remedy would be of course doing the opposite."

      I don't think so. The only "policy group" that I know of that recommends higher inflation or a higher inflation target as a policy measure is the Keynesian aggregate demand bunch.

      Neo-Fisherites would be within their rights to criticize the central bank in saying - "Hey central bank, you might have causality reversed". If you want higher inflation, you may want to increase (not decrease) your interest rate target. And they would have some historical evidence backing that up.

      That doesn't mean that Neo-Fisherites recommend a higher inflation rate, a higher inflation rate target, a higher interest rate, or a higher interest rate target as a matter of policy.

      And it doesn't mean that the direction of causality runs from higher interest rates to higher inflation.

      All that the Neo-Fisherites would say is look at prior history and judge for yourself - prior incidents of higher inflation were often preceded by higher nominal interest rates. Does that establish causality - nope. Would that knowledge be helpful to central bank policy makers - I would think so.

      The difference between Irving Fisher and the Neo-Fisherites is that Fisher lacked the data to back up his formula. The Neo-Fisherites have the data.

      Delete
  15. Frank,

    you say Neo-Fisherites could be right in telling the central bank it might have the interest rate-inflation causality reversed, so that if it wants higher inflation, it may want to increase (not decrease) its interest rate target. However, you also say that this doesn't mean that the direction of causality runs from higher interest rates to higher inflation. I see some tension between the two propositions... In any case, how does that happen? What's inside that black box that lead from higher interest rates to hgiehr inflation?

    Nick Rowe and David Andolfatto had a recent interesting exchange on this issue on Andolfatto's blog, from where it emerged that for causality to go from interest rates to inflation, fiscal policy has to play an active role? If that is the case - I myself commented - it would far superior to have a fully-fedged helicopter money operation or overt monetary financing of public deficits.

    Ruling out fiscal policy, how do you think Neo-Fisherites explain it? Do they just observe a statistical association, suggest to exploit it, but really don't have an idea of what is behind it?

    ReplyDelete
  16. Biagio,

    "However, you also say that this doesn't mean that the direction of causality runs from higher interest rates to higher inflation. I see some tension between the two propositions."

    It means that I personally don't agree with the conclusion. Yes, there is historical data pointing to that conclusion, but as I have said before it is overly simplistic.

    "Ruling out fiscal policy, how do you think Neo-Fisherites explain it? Do they just observe a statistical association, suggest to exploit it, but really don't have an idea of what is behind it?"

    Okay if I am ruling out fiscal policy, presumably the government is in a perpetual state of income / expenditure balance - no deficits, no surpluses. Note that this does not describe the state of the U. S. government during the 1970's. How would a Neo-Fisherite explain rising interest rates coupled with rising prices?

    The simplest model I can think of is that the producers of goods borrow money from the consumers of goods at a floating interest rate set by the central bank (no equity financing, no tax deduction for interest paid). Any rise in the interest rate set by the central bank increases the financing costs for producers and increases the income of consumers. Assuming no other changes (worker productivity, technological advancement, tax deferred savings plans, etc.), that should raise prices charged by producers and prices paid by consumers.

    In this model, the central bank makes no loans, prints no money, and just sits there setting an interest rate. Money is created by consumers and lent at the central bank prescribed interest rate to producers.

    ReplyDelete
  17. Frank,

    Thank you for your patience and effort.

    I am glad to hear that you personally don't agree with the Neo-Fisherian conclusions. I therefore doubly admire your effort to come up with a logical explanation of their conclusions.
    Otherwise, my next question to you would have been: how realistic do you think your assumptions are, starting from the simple observation that you must exclude from the model all agents that stand to lose from a higher cost of money, except producers who, as you assume, can simply pass through the higher cost on to consumers. The latter, of course, would need to have access to money free of charge and enough to be able both to lend to producers and to finance their own consumption.

    Yet rather than questioning your assumptions, I appreciate from your example how incredibly heroic the assumptions must be to support the Neo-Fisherian conclusions.

    Thanks again

    ReplyDelete
    Replies
    1. I don't know that I would call consumers lending and producers lending "heroic assumptions", but you are welcome in any case.

      I will go ahead and try to answer your next question:

      "How realistic do you think your assumptions are, starting from the simple observation that you must exclude from the model all agents that stand to lose from a higher cost of money, except producers who, as you assume, can simply pass through the higher cost on to consumers"

      I presume you to mean consumers that borrow. Yes, the model excludes that and a lot of other things. What effect would higher interest rates have on consumers if they too were borrowers?

      It would depend on the debt composition - if consumers own $1 trillion in producer debt and there is only $100 worth of outstanding consumer debt - the interest paid on the producer debt will far and away exceed that owed by consumers.

      U. S. debt composition (government, consumer, banking / finance, and corporate) is tracked by the Federal Reserve here:

      http://www.federalreserve.gov/releases/z1/

      I don't know of anyone that has produced research into how debt composition affects macroeconomic variables, though I think there would be a significant affect.

      Delete

Comments are welcome. Keep it short, polite, and on topic.

Thanks to a few abusers I am now moderating comments. I welcome thoughtful disagreement. I will block comments with insulting or abusive language. I'm also blocking totally inane comments. Try to make some sense. I am much more likely to allow critical comments if you have the honesty and courage to use your real name.