Wednesday, April 23, 2014

The Neo-Fisherite Rebellion



Over the last three years, a quiet rebellion seems to have sprung up in macroeconomic circles. So far, it's limited to a few whispers, a couple of papers, and the odd blog post or dinner speech, but it represents a striking break from conventional thinking. And despite my best efforts, I find myself unable to convince myself that it's wrong. The rebellious idea - which I've decided to call "Neo-Fisherite" - is that low interest rates cause deflation, and high interest rates cause inflation.

First, the basic idea. The Fisher Relation says that nominal interest rates are the sum of real interest rates and inflation:

R = r + i

That's not an assumption, that's just a definition (actually it's an approximation, but close enough). What I call the "Neo-Fisherite" assumption is that in the long term, r (the real interest rate) goes back to some equilibrium value, regardless of what the Fed does. So if the Fed holds R (the nominal interest rate) low for long enough, eventually inflation has to fall. This is exactly the opposite of the "monetarist" conclusion that if the Fed holds R very low for long enough, inflation will trend upward.

The opening shot of the Neo-Fisherite rebellion was fired by Minneapolis Fed President Narayana Kocherlakota, in a speech in 2010:
The fed funds rate is roughly the sum of two components: the real, net-of-inflation, return on safe short-term investments and anticipated inflation. Monetary policy does affect the real return on safe investments over short periods of time. But over the long run, money is, as we economists like to say, neutral. This means that no matter what the inflation rate is and no matter what the FOMC does, the real return on safe short-term investments averages about 1-2 percent over the long run. 
Long-run monetary neutrality is an uncontroversial, simple, but nonetheless profound proposition. In particular, it implies that if the FOMC maintains the fed funds rate at its current level of 0-25 basis points for too long, both anticipated and actual inflation have to become negative. Why? It’s simple arithmetic. Let’s say that the real rate of return on safe investments is 1 percent and we need to add an amount of anticipated inflation that will result in a fed funds rate of 0.25 percent. The only way to get that is to add a negative number—in this case, –0.75 percent. 
To sum up, over the long run, a low fed funds rate must lead to consistent—but low—levels of deflation. (emphasis mine)
This sparked a furious backlash in the blogs (see here, here, and here for just a few examples). Especially note the detailed critiques by Andy Harless and Brad DeLong.

Following that controversy, Kocherlakota famously converted to a more conventional monetarist position and began arguing for more QE and higher inflation (raising the interesting, but almost certainly fantastical, possibility that his conversion is a deep-cover operation and he really still thinks "easy" monetary policy will be deflationary!).

But others emerged to pick up the banner of rebellion.

The second outbreak of the rebellion came when Steve Williamson wrote a paper in which QE is deflationary. This touched off possibly the most entertaining and explosive debate in the history of the macroeconomics blogosphere. "Monetarist"-leaning bloggers like Scott Sumner and Nick Rowe teamed up with "Keynesian"-type bloggers like Brad DeLong and Paul Krugman to assail Williamson's idea, while a few mavericks like Tyler Cowen, David Andolfatto, and me came cautiously to Williamson's defense (without believing in his model per se). I have no idea how much of a splash the paper made within academia itself, but that would be interesting to know.

Anyway, unlike Kocherlakota, Williamson has stuck to his guns. Last December, he wrote a blog post suggesting that Paul Volcker whipped inflation in the early 80s not by raising interest rates in the short term, but by lowering them in the long term. In February he wrote that "there's a clear Fisher relation in the data" for both Japan and the U.S., posting this graph:



The graph shows a positive contemporaneous correlation between inflation and interest rates (which monetarists, of course, would claim is evidence that interest rates are lowered in response to deflationary shocks).

In a recent "debate" with Mark Thoma (who took a fairly conventional monetarist view), Williamson again reiterated the Neo-Fisherite idea.

Last December, Williamson's revolt got some heavyweight support when it was joined by John Cochrane:
I've been following with interest the rumblings of economists playing with an amazing idea -- what if we have the sign wrong on monetary policy? Could it be that raising the interest rate raises inflation, and not the other way around? 
Conventional wisdom says no, of course: raising interest rates lowers inflation in the short run and and only raises inflation in a very long run if at all. 
The data don't scream such a negative relation. Both the secular trend and the business cycle pattern show a decent positive association of interest rates with inflation, culminating in our current period of inflation slowly drifting down despite the Fed's $3 trillion dollars worth of QE. 
Cochrane proceeds to write down a reduced-form model in which a long period of low interest rates causes inflation to jump up for a short while but then trend downward in the long run. He concludes cautiously:
Obviously this is not the last word. But, it's interesting how easy it is to get positive inflation out of an interest rate rise in this simple new-Keynesian model with price stickiness. 
So, to sum up, the world is different. Lessons learned in the past do not necessarily apply to the interest on ample excess reserves world to which we are (I hope!) headed. The mechanisms that prescribe a negative response of inflation to interest rate increases are a lot more tenuous than you might have thought. Given the downward drift in inflation, it's an idea that's worth playing with. 
I don't "believe" it yet (I hate that word -- there are models and evidence, not "beliefs" -- but this is the web, and it's easy for the fire-breathing bloggers of the left to jump on this sort of playfulness and write "my God, that moron Cochrane 'believes' monetary policy signs are wrong" -- so one has to clarify this sort of thing.) We need to explore the question in a much wider variety of models.
About a month ago, Cochrane followed up with a simple structural model that includes the Fed and the Treasury, and gets the same results. He also notes how normal monetary policy deals with the Fisher Relation:
Nominal rate = real rate plus expected inflation. Tradition says though that you temporarily steer the wrong way. First lower the nominal rate, then inflation picks up, then deftly raise the nominal rate to match inflation. If you instead raise rates and then just sit there waiting for inflation to catch up all sorts of unstable things happen...But maybe not.
Just a few days ago, Williamson evaluated Lars Svensson's policy advice from a Neo-Fisherite point of view, summing up the core of the rebellion quite well:
For simplicity, think about a world with perfect certainty. In the long run, standard asset pricing gives us the Fisher relation, which is 
R = r + i, 
where R is the short-term nominal interest rate, r is the real interest rate, and i is the inflation rate...Therefore, in the long run, if R is targeted at its lower bound by the central bank, 
i = - r - t 
So, if we think that r is invariant to monetary policy in the long run, then if the central bank pegs the nominal interest rate at its lower bound, and central bank liabilities are taxed, this will make long-run inflation lower.
(Note that the Neo-Fisherite idea is NOT that a fall in interest rates causes the price level to jump up and then drift back down to its original level. If the Neo-Fisherites are right, holding interest rates at low levels for a long time will cause a long-term deflationary trend that will eventually push the price level lower than if interest rates had been kept at the old, higher level.)

So why do I find myself unable to convince myself that the Neo-Fisherite idea is wrong? Well, like Cochrane, I don't really "believe" in any of these specific models (or any others). But two things make me unwilling to discount the Neo-Fisherites. The first is a thought experiment, and the second is the evidence.

The thought experiment is this: Suppose there was no government debt, and the Fed raised nominal interest rates to 20% and held them there forever. What would happen to real rates? Well, they wouldn't rise to 20% forever, because there's just no way that our society can physically, technologically deliver a 20% riskless rate of return to bondholders. Eventually, one of two things would have to happen: either 1) the Fed's control over the nominal interest rate would break down, or 2) inflation would rise (the Neo-Fisherite result). If the Fed can't control the nominal interest rate, then our standard models of monetary policy all break down, and we have to think about the microeconomics of money demand, which is hard to do. But the only alternative would be the Neo-Fisherite result.

As for the evidence, the U.S. case is more interesting to me than the case of Japan. In Japan, population growth is negative and Total Factor Productivity is stagnant, raising the obvious possibility of a Larry Summers-style "secular stagnation", where interest rates look low but are in fact high. But in the United States, productivity and population both continue to grow at a reasonable clip, so while "secular stagnation" seems possible, it would require some more stuff in the basic model that I haven't seen yet. Meanwhile, QE, while it produces slight jumps in inflation expectations whenever a policy change is announced or an asset purchase is made, has overall coincided with a downward drift in inflation. Monetarists will say that this is because of expectations that the Fed's asset purchases won't be permanent, and of course we can't rule that out. We can't really rule anything out in macro, even serious academic macro, much less the "eyeball-it-and-see-what-you-think" approach taken on the blogs (and in much of the private sector).

But what I'm saying is, there seems as of yet no obvious reason for me to write off the Neo-Fisherite idea. And - as Cochrane and Williamson have both shown - it's possible to write down models where the idea works. The structural models so far all rely on rather odd and rigid fiscal policy rules, so the microfoundations are still kind of crazy. But I see no reason why those models are substantially crazier than any of the more mainstream, monetarist-type (or RBC-type) models. So for now, count me on the side of the Neo-Fisherite rebels, just because I think the idea is neat, and potentially very important, and I want to see where it leads.

Because the idea is very important. Everything the Fed does, pretty much, is based on the idea that the longer you hold interest rates at low levels, the "easier"- i.e. the more pro-growth and inflationary - your monetary policy is. The Neo-Fisherite idea doesn't just discount the effectiveness of monetary policy(like RBC models do, or like the MMT people do) - it stands that whole monetary policy universe on its head. If the Neo-Fisherites are right, then not only is the Fed massively confused about what it's doing, but much of the private sector may be reacting in the wrong way to monetary policy shifts.

Anyway, blog debates are fun, but I'm even more interested to see if Neo-Fisherite papers are starting to appear in the academic literature in increasing numbers. If they do, that means the rebellion can no longer be written off. (On this note, see updates below...)


Updates

Specifically, what I'd be interested to see is for someone to find some microfoundations for the Neo-Fisherite result that don't depend on fiscal policy reaction functions.

Blogger Edward Lambert of Angry Bear declares his support for the Neo-Fisherites.

Someone on Twitter points me to a recent paper by Stephanie Schmitt-Grohe and Martin Uribe in which a Neo-Fisherite result holds. The driving force behind the model is a form of animal spirits. That's interesting, because when I was thinking of what kind of microfoundation other than a weird fiscal policy rule might give a Neo-Fisherite result, I thought this kind of behavioral explanation might do the trick...so it's cool to see my intuition confirmed. Here's a Simon Wren-Lewis blog post about that paper.

Ryan Avent considers the Neo-Fisherite idea. He rejects it in the end, but before he does that, he comes up with a plausible-sounding story of how it might be right.

Scott Sumner, unsurprisingly, is no fan of the Neo-Fisherite idea. Neither is David Beckworth.

Tony Yates reveals that there is actually a fairly large literature on this question, most of which concludes that if the Fed holds down interest rates forever, inflation becomes...indeterminate!

David Glasner considers the Neo-Fisherite idea, and finds himself unable to reject it. The behavior of expectations, he notes, is key.

117 comments:

  1. If the Neo-Fisherites are right, then not only is the Fed massively confused about what it's doing, but much of the private sector may be reacting in the wrong way to monetary policy shifts.

    The NFs are committed to RatEx, so if the private sector is reacting in the wrong way their theory has a serious problem.

    ReplyDelete
  2. Overeating inevitably leads to dieting and weight loss. So obviously, if you want to lose weight, you are advised to overeat.

    ReplyDelete
    Replies
    1. It is called cognitive dissonance. You just can't fathom your belief is wrong, so you assume it is correct and the alternatives are " crazy".

      Delete
    2. +1 Max what you're citing is a Non-sequetor (it does not follow) analogy i.e. comparing weight gain in a human organism to the general price level of goods and services.

      Leon Festinger's theory of cognitive dissonance focuses on how humans strive for internal consistency. When inconsistency (dissonance) is experienced, individuals largely become psychologically distressed. His basic hypotheses are listed below:

      "The existence of dissonance, being psychologically uncomfortable, will motivate the person to try to reduce the dissonance and achieve consonance"
      "When dissonance is present, in addition to trying to reduce it, the person will actively avoid situations and information which would likely increase the dissonance" --- Wikepedia

      Delete
    3. Nathanael8:27 PM

      Max has a completely correct analogy.

      Low interest rates lead to lots of borrowing (overeating), which leads to a high household debt load. Unless wages go up to pay off the debt, the indebted people then try to pay down their debt, which causes debt deflation (dieting & weight loss).

      Delete
  3. o. nate3:19 PM

    I would be more inclined to take option (1) in your thought experiment. Unless the rate of inflation was high already, I don't think the Fed could maintain a 20% interest rate for very long. The correlation data seems to be a clear case of confusing cause and effect. By the way, Yichuan Wang had a good reply to Williamson: http://synthenomics.blogspot.com/2013/12/why-dynamic-stories-are-important.html

    ReplyDelete
    Replies
    1. o. nate3:24 PM

      However, rereading your post, I see that you're positing a world with no government debt in your thought experiment. So in such a world it's not even clear how the Fed would be conducting its policy, since the usual way they set rates is through buying and selling government debt. I guess if we imagine a different Fed mechanism, such as paying interest on reserves, then yes, the Fed could set the interest rate arbitrarily to 20% and this would be inflationary, because the Fed would be printing tons of money to pay the IOR. However, I don't think this is the scenario that Williamson et al are envisioning.

      Delete
    2. yes, the trick in that thought experiment is to say "1) can be ruled out because funny things happen if the Fed can't control nominal rates". But funny things are *meant* to happen if you make such weird assumptions!

      Delete
    3. louis1:50 PM

      Agree. If the Fed wants short term interbank loans to be transacted at 20% nominal rates, it needs to make money very hard to come by. Meaning, engage in open market operations to suck up reserves by selling assets (FX reserves, gold?) into the market for cash, or paying 20% overnight interest so all reserves sit at the Fed. Either way, money would be very tight and the effect would surely be highly deflationary. The marginal project would indeed have a 20%+ real rate of return, because all inferior return projects would not be funded in such a tight money environment. It's not strange for a sharp reduction in supply to yield an increase in price and decrease in quantity.
      Eventually you would see the dollar replaced as medium of exchange and real rates go back down again as the economy returned to full employment under a new monetary regime.
      Basically this is the result you see with emerging markets that try to peg unsustainably high exchange rates, isn't it?

      Delete
  4. In your thought experiment is interseting. In order to raise and keep the interest rates at 20% in a world without governemt debt, the fed would have to issue it's own liabilities in order to hit it's interest rate targets. So if the return on real capital is less then the CB bonds, people will try to sell real capital and exchange it for money to then by CB bonds.

    So it seems obvious that this would be deflationary as people would hoard money to exhcange it for bonds and sell real assets.

    But then what about the fisher equation. Above we assumed that R is 20 and r is less than R, so i = R - r, and must be a positive number.

    I think the problem is linking r in the Fisher equation to the return on real assets. If you don't assume they are the same thing, then the real (fisher) return is actually greater than the nominal return. This actually makes sense because in a deflationary environment, your nominal assets are able to buy more and more real goods.

    ReplyDelete
  5. Anonymous3:20 PM

    Heh heh! OK, but on Volcker you have to remember: the game was to increase the Fed rate AND to lower wages for average working people. "Inflation," sub rosa, was always defined as "your wages."

    ReplyDelete
  6. Anonymous3:22 PM

    Your first thought experiment is weird. There's no government debt at all? What are we using for money in this world? Would our monetary system really look anything similar to what we're seeing today in US and Japan? Haven't you broken the connection to anything that's real? You just leapt though a discontinuity, haven't you? And who says that fed, in a no-government-debt, low-inflation world would be able to enforce 20% nominal interest rates? No one would be lending to government and would be able to undercut the banks and the fed, right?

    I mean, you might as well ask what would happen to the monetary system if the moon were made of green cheese.

    ReplyDelete
  7. This is highly strange. No one ever thought the monetary autority control long term interest rates directly. That's not in new keynesian models, not in market monetarist models, not even in IS-LM.

    And everyone knew long term interest rates correlated with inflation.

    There are any reasons to believe monetary autority control long term interest rates directly?

    These confusions is why we should forget interest rates and think money.

    ReplyDelete
    Replies
    1. Anonymous6:37 AM

      Perhaps it reflects excessive deference to the price mechanism.

      Delete
  8. The Fed's liabilities are: currency in circulation, deposits of depository institutions and treasury balance. How do you 'tax' these? and does that not imply the same thing as charging fees for deposits?

    ReplyDelete
  9. In theory, the Fed can continue to offer 20% interest on reserves forever (though I'm not sure it would be legal if it pushed the Fed deeply into technical insolvency), so its control over nominal interest rates would not break down (unless legal constraints forced it to do so). But it's obvious that the dynamics of very high interest on reserves push toward lower prices in the short run. The opportunity cost of holding cash would be very high, and the return to deferring expenditure would be very high, so there would be less expenditure, and prices would go down. The equilibrium requires that prices be expected to rise, but this is not the usual Neo-Fisherite result, because you don't get to that equilibrium except by an initial drop in prices. I can vaguely imagine a perfect RatEx world in which that initial drop doesn't have to happen, but that's clearly not the world we live in, because the slightest deviation from RatEx (and the slightest deviation from perfect credibility of the 20% interest rate target) brings in the initial deflationary dynamic.

    IIRC Williamson essentially assumes away the initial price drop: he uses fiscal policy to tie down the initial price level. But that's cheating. Sure, you could (in an ideal world) do a fiscal policy that would avoid the initial price drop, but in that case you're not really studying the effect of monetary policy in isolation. If you want to study the effect of monetary policy, your baseline fiscal policy has to be some reasonable approximation of what a "passive" fisc would actually do.

    ReplyDelete
    Replies
    1. and glancing at the second Cochrane link, it looks like he uses a different version of the same trick. Basically the Fed' s interest rate policy commits the Treasury to a fiscal policy that will be consistent with the Neo-Fisherite result. Not a good model for how a fisc actually behaves.

      Delete
    2. John Cochrane's reduced-form model includes the short-run effect.

      And yes, so far the structural models all involve Treasury policy rules that are not really credible...but no less credible than models of money demand, eh?

      Delete
    3. Andy, I do not think SW needs to assume away the initial price drop. In his model, an increase in the IROR reduces currency and increases reserves. The drop in currency lowers current prices in the centralized market relative to future prices, which is equivalent to higher inflation. As far as output is concerned, because money is neutral in the centralized market, the increase in the IROR has no effect. What is non-neutral in a liquidity trap is an increase in liquid interest-bearing assets (e.g. Treasuries), which facilitates transactions in the decentralized market thereby increasing output. In turn, this requires a larger fiscal deficit until the yield of Treasuries rises above the IROR, whatever it may be. My problem is that it is difficult to swallow that a drop in cash-transactions would be neutral under current conditions.

      Delete
    4. I think Noah's Fisherite dilemma is probably right. If there were no treasury debt instruments, but the Fed offered 20% interest on reserves forever (with say a 21% penalty rate for discount borrowing), and there were no other changes and - for some reason - very low inflation initially, then there would be little correlation between the policy rate and commercial rates. The cost of borrowing additional reserves would be over 20%, but no bank would ever need to borrow additional funds, since the Fed would be pumping free money into reserve accounts at a whopping 20% per year, resulting in a nearly 250% increase in aggregate reserve holdings in five years and a 600% increase in 10 years, just for doing nothing. If nominal GDP growth remained below 20%, then competitive pressures would push nominal commercial rates down to near zero to reflect what would be effectively a zero marginal cost for additional funds. Indeed no bank, even one growing at an astronomical rate, would ever choose to borrow additional funds. They would simply convert other assets into cash as needed and park the cash in their reserve account.

      But another way of putting this is that IOR is just an alternative type of government bond. If the Fed gives the banks 20% interest on reserves, that is not much different than offering them three month t-bills at 4.7%.

      But back to Noah's scenario, the actual long-run outcome would probably be a continued correlation between the policy rate and commercial rates, with massive inflation. Starting with low inflation and 20% IOR, competitive pressures would result in banks beginning to offer their depositors huge interest returns on time deposits, and even ordinary demand deposits, but very low interest rates on loans. If there is only 2% inflation, banks can still make money by passing most of their reserve account income onto their customers, and by charging those customers very small amounts for loans. But the gigantic influx of interest income for depositors would then cause massive inflation, and the economy would restabilize with high interest rates on deposits and loans, with inflation approaching 20%.

      The policy upshot to consider here is that the central bank's policy interest rate is actually a key cognitive input for inflation expectations. If that is the case, it is no wonder that central banks keep failing to hit their inflation targets in a ZIRP regime.

      Delete
    5. Noah, models of money demand are clumsy attempts to rationalize the behavior that we observe. I don't see an analogy with models of fiscal policy that fly in the face of observation.

      Dan, I don't see how any loans would get made. Any money a bank has to loan out, it can make a higher return by keeping it as reserves. So why would it make any loans at all? But it's true that 20% IOR would commit the Fed to create a lot of money, since it wouldn't have enough assets to sell to offset the credits to reserve accounts, so it would technically become insolvent, and maybe expectations of the value of money would decline. So maybe Noah's thought experiment works, but again this is a trick (a different trick than the Neo-Fisherites use). The thought experiment only (possibly) works for absurdly high interest rates. The same logic does not apply in reverse if interest rates are low because low interest rates do not constitute a commitment to contract the money stock in the future.

      Delete
    6. So suppose that the inflation rate is zero, the nominal interest rate is zero, and government liabilities are constant over time. If the nominal interest rate on reserves rose to 5% then consumers would want to reduce currency holdings, as currency does not offer a return. The drop in currency would drive prices down relative to the future. The process should continue until the real rates of return on currency and reserves become equal again (equilibrium). Since the nominal return on currency cannot change, the only thing left to adjust is the current price level relative to the future price level. Specifically, current prices must fall 5% below prices a year ahead, which implies a 5% inflation, so that the real rate on reserves will fall to zero, the same as for currency. This is simple arbitrage!.

      Delete
    7. Anonymous5:17 PM

      Tsss... Noah - I would have thought you have more humour... But still another try with more explanations for the finance professors without drive to click links...

      4th paragraph from the bottom:
      "Montarists will say that this is because of expectations that the Fed's asset purchases won't be permanent, and of course we can't rule that out"
      ...Montanist takes a "n", not a "r", see
      http://en.wikipedia.org/wiki/Montanist

      For the link-disliking finance professors:
      The Montanists were a sect claiming personal holy visions - something that, in my opinion, combined with your typo - makes a good opportunity for a small joke, because I see the Monetarist view as sometimes possed by the same spirit, especially when defending against heretical ideas (like the one you talk about...)

      And if this comment irks you again - and you delete it - PLEASE FIX THE TYPO!

      5. try - typo still not fixed - long live copy&paste...

      it could be so easy to get rid of that comment - fix the typo...

      Delete
    8. For Cthulhu's sake, I thought you were a spambot. Just email me next time!! :-)

      Delete
    9. Peter, I think the difference of opinion regards why credit is limited. Most people think that the problem is low demand. People will not borrow more unless the real interest rate falls further, so lowering the real interest rate (e.g. by increasing inflation) is the answer. My understanding is that Williamson thinks this is a credit supply problem. Privately-issued assets (e.g. mortgage loans) have become less liquid, so banks will not buy them until they have acquired enough liquid assets (reserves and Treasuries). Therefore, to expand credit we must first issue enough liquid assets (i.e. Treasuries) to satisfy the demand by banks. Once this has been accomplished, banks will expand their lending to the private sector. When will this have been accomplished? When banks refuse to hold interest-bearing assets unless they get compensated, i.e., when the interest rate on Treasuries rises above the IROR.

      Delete
    10. CA: "The drop in currency would drive prices down relative to the future." Yes, that's my point. Unless the expected future price level goes up, the immediate effect of the rise in interest rates is that prices go down. That is deflation. Yes, it will be followed by inflation, or at least, it will be expected to be followed by inflation, but before that inflation materializes at lot of other things could happen. The critical question is what determines the expected future price level. The Neo-Fisherites have various tricks (mostly involving fiscal policy) to cause the expected future price level to rise, so that you don't get the initial deflation in response to a rise in interest rates. But I see no plausible reason that we should actually expect the expected future price level to rise. And if the expected future price level doesn't rise, then the immediate effect of a rise in the nominal interest rate (if that rise is exogenous and expected to be permanent) is a fall in the price level.

      Getting back to the relevant application, the question would be whether there is a reason to think that ZIRP caused the expected future price level to rise. I don't see any such reason. And we have not observed a bout of rising prices in response to ZIRP, to set us up for deflation in the case where the expected future price level is unchanged. I conclude that our experience is not consistent with the Neo-Fisherite view.

      Delete
    11. Andy, OK, we are in agreement. The initial drop in prices is not referred to as deflation because in the model the jump is instantaneous, but I agree that in the real world things do not play out as in the model. I think in SW "world" the increase in future prices is driven by the fact that reserves must be used for purchases at some point in the future. A reduction in nominal spending now that increases reserves should, everything else being equal, result in higher nominal spending in the future. However, people have already chosen optimal time paths for real quantities based on their rate of time preference and the real interest rate, so any shuffling in nominal spending across time alters nominal prices across time so as to live quantities unchanged. Again, whether this seems a reasonable representation of what happens in the real world is questionable.

      Delete
  10. Noah, think about it, do you really think the market monetarist explanation for the correlation of low interest rates and low inflation isn't better than the neo-fisherite rebelion?

    Sometimes I think you don't give enough credit to MM's because their theory would imply finance has no place in macro.

    ReplyDelete
  11. Anonymous3:40 PM

    One data point on acceptance of these models: this is actually an implication not just of more recent models, but of standard New Keynesian models with rational expectations: a standard reference is Hagedorn "Optimal Disinflation in New Keynesian Models" (2011 J. Monetary Economics). From the abstract:

    "Central bankers’ conventional wisdom suggests that nominal interest rates should be raised to attain a lower inflation target. In contrast, I show that the standard New Keynesian monetary model with rational expectations and full credibility predicts that nominal interest rates should be decreased to attain this goal. Real interest rates, however, are virtually unchanged. These results also hold in recent vintages of New Keynesian models with sticky wages, price and wage indexation and habit formation in consumption."

    The standard criticism of this result, made in response to Kocherlakota and many others, is not that it is inconsistent with standard models, but that it is an artifact of the way stability analysis is performed in rational expectations models: expectations are allowed to jump discontinuously in response to a policy change, and so in these models the Fisher relation (which should have an E, for expectations, in front of the right side) holds instantaneously, without large residual expectation errors. How reasonable this might be depends on the context. For day-to-day macroeconomic policy, announced changes may not be large or well-understood enough for prices to move instantaneously to equilibrium (explicit price stickiness of course blocks this, but in standard New Keynesian models asset markets do not exhibit stickiness). In such cases, learning models may be preferable. In the case of ending hyper-inflations and other large discrete actions like a currency reform, such an immediate jump may be more plausible. Distinguishing when different cases are relevant will require a more nuanced understanding of asset markets. Note though that, clearly, this property of standard New Keynesian models does not mean that positive unexpected shocks to the interest rate are inflationary in the model: only changes in a policy rule result in such 'perverse' outcomes. However, in the real world, with imperfect Fed communication, a long sequence of negative shocks to the interest rate may be hard to distinguish from a shift in the "true" inflation target: witness the blogosphere debate on whether the Fed "actually" has a 2% target or in fact is undershooting on purpose. So the clear distinction in the model between these two scenarios may not be so clear cut in actuality.

    ReplyDelete
  12. Bit of a flaw in the premise, nominal rates don't really represent the stance of monetary policy.

    Otherwise, it's just a long-term effect balanced against a short-term effect. Fisher and Friedman explained long ago that a regime of tight money would gradually lead to low nominal interest rates as people gradually accepted that long-term inflation was going to be lower and repriced the i component accordingly. The opposite can also happen, a long-term strategy of loose money can lead to higher rates.

    If the Fed wants to inflate, all it has to do is tell markets it's planning to inflate, either through a higher inflation target or NGDPLT. QE should have been used only to make such promises credible. They've confused the hammer with the nail.

    ReplyDelete
    Replies
    1. The Fed still has a 2% inflation target, and they keep missing it low.

      I sometimes wonder how many people remember why inflation targets are generally not negative.

      Delete
    2. It's funny, half the people are saying "THE FED KEEPS HAMMERING! THEY'RE BUILDING INFLATION!!" and the other half are saying "NO INFLATION HAS BEEN BUILT YET, THEY NEED TO HAMMER HARDER!"

      Meanwhile, only a few people seem to notice the Fed isn't using nails.

      Delete
    3. shisaku -- Can you imagine someone in 1989 telling you Japan's nominal GDP would be lower in 2013? Who could have believed that?

      Delete
    4. Although I see now that Nipponese nominal GDP peaks around 1994, then passes that peak in 2009. Still rather shocking for Mr. 1994.

      Delete
    5. "QE should have been used only to make such promises credible. "

      Notice that whenever QE was used inflation actually *fell*. It happened in Japan, it happened in Sweden. So if anything QE makes it credible that inflation is heading lower.

      Delete
    6. Heh, I guess the QE signal to markets has become "we forgot how to inflate."

      Delete
    7. PeterP,
      "Notice that whenever QE was used inflation actually *fell*. It happened in Japan, it happened in Sweden."

      Japan’s original ryōteki kin’yū kanwa (QE) was officially announced in March 2001, although there wasn’t a noticeable increase in the monetary base until December 2001. In March 2006 the BOJ announced the completion of QE and the monetary base was reduced by about 24.4% between January and November 2006.

      How did the Japanese economy do? Here’s a graph of Japans’ annual CPI inflation and harmonized unemployment rate:

      http://thefaintofheart.files.wordpress.com/2013/06/sadowski2b_5.png

      Note that aside from the consumption tax induced increase in 1997 Japan’s inflation rate dropped nearly every year from 1991 through 2002 and then *increased* until there was consecutive years of consumer price inflation in 2006-07 for the first time in nearly a decade. Unemployment increased nearly every year through 2002 and then dropped in 2003 for the first time since 1990, and then continued to drop every year through 2007. And it’s worth noting that even the Nikkei 225 gave its thumbs up during 2003-07, with the index rising from less than 7900 in April 2003 to over 18,000 in June 2007, which is still by far the greatest stock market rally in Japan since the beginning of the Lost Decade(s).

      Delete
    8. PeterP,
      "Notice that whenever QE was used inflation actually *fell*. It happened in Japan, it happened in Sweden."

      The Riksbank increased its monetary base by more than the Fed early on in the crisis. By October 2008 it was already 121% larger than in August compared to 25% for the Fed. By December 2008 it was 225% larger compared to 101% for the Fed. However, unlike the Fed which moved to a zero interest rate policy by December 2008, the Riksbank did not lower its repo rate to 0.25% until August 2009, and only kept it at this rate through June 2010.

      Although the Riksbank was slower to move to ZIRP, during this time it maintained a (-0.25%) deposit rate, the first central bank to institute a negative interest rate, and the monetary base was maintained in the range of 270% to 350% larger than it had been in August 2008. In contrast the Fed was paying (+0.25%) on reserves and the monetary base was only 100% to 140% larger than it had been in August 2008 during this period.

      Here’s a graph of NGDP for Sweden, Denmark, the US and the Euro Area since 2007. It is indexed to 100 in 2007Q3.

      http://research.stlouisfed.org/fred2/graph/?graph_id=153008&category_id=0

      Note also that Sweden opened up a very wide lead in NGDP with respect to the other currency areas between 2009Q4 and 2010Q4. NGDP grew by 13.4%, 6.9%, 6.9% and 11.2% in 2010Q1 through 2010Q4 respectively. Thus NGDP soared by 9.6% year on year between 2009Q4 and 2010Q4.

      And yes, the last time that the Reichsbank attained its year on year target inflation rate of 2.0% for any sustained period was from November 2009 through April 2010, during the time it did QE.

      Delete
  13. Anonymous4:17 PM

    No, this tale is fine, Pooh sir. they are merely shining their tale to keep from being bored by yours, that is a nasty tale.A very nasty tale, old boy!!!!!

    ReplyDelete
  14. I'm definitely a novice here, but why doesn't the recent European experience of increasing rates with low inflation counter this idea? Krugman has been writing a lot about Sweden's experience with that lately, and it seems to me to rule this out. Shouldn't inflation have increased for them? Or do we expect the lag to be multiple years?

    ReplyDelete
  15. Noah,
    You sense the same thing that I do. I commented on Williamson's post pretty much your thoughts. I am in agreement with the Fisher effect.
    I simply want to thank you for having sufficient wisdom and constitution to write this post.

    ReplyDelete
  16. Rob Rawlings5:14 PM

    In a world where prices and interest rates adjusted instantly to money supply changes then the ONLY way a CB could successfully target 20% interest rates would be to expand the money supply so that inflation ran at (20-r)%.

    Its only because prices are sticky in the short term that reducing the money supply causes rates to rise (money becomes more scarce and therefore more valuable compared to other goods so people will need a bigger incentive to part with it?). As prices fall to match the new money supply so will interest rates fall too , so if the CB is serious about 20% rates (and wants to avoid a deflationary spiral) it had better reverse policy and increase the money supply to achieve its goal

    ReplyDelete
  17. High interest rates cause inflation? Erdogan told us so: http://www.bloombergview.com/articles/2014-01-17/interest-rate-obsession-may-doom-turkey-s-erdogan

    ReplyDelete
  18. The thought experiment is this: Suppose there was no government debt, and the Fed raised nominal interest rates to 20% and held them there forever. What would happen to real rates? Well, they wouldn't rise to 20% forever, because there's just no way that our society can physically, technologically deliver a 20% riskless rate of return to bondholders. Eventually, one of two things would have to happen: either 1) the Fed's control over the nominal interest rate would break down, or 2) inflation would rise (the Neo-Fisherite result). If the Fed can't control the nominal interest rate, then our standard models of monetary policy all break down, and we have to think about the microeconomics of money demand, which is hard to do. But the only alternative would be the Neo-Fisherite result.

    Which rate are we talking about, the fed funds rate? The 30-year Treasury rate?

    The Fed could easily produce the latter simply by credibly promising a 20% long-term inflation target. The former would cause a massive deflationary spiral, until eventually people would simply stop using the currency because there wasn't enough of it around to be useful for exchange (this happened in TGD, notably; people starting issuing their own scrip).

    So that's how you square that circle: the Fed has absolute control over its currency, but zero ability to force people to actually use it.

    ReplyDelete
  19. Anonymous5:36 PM

    Cochrane, Williamson et al. got causality all wrong !!! Milton Friedman explained all this in his 1968 paper 'The role of monetary policy' and I assume that none of these guys ever bothered to read it. In that paper, he also does a very similar thought experiment in which he explains what happens if the CB pegs interest rates forever at very low levels (below the Wicksellian rate), which would force the monetary authorities to engage in larger and larger market operations... Furthermore, he explains the following: "...the monetary authority could assure low nominal interest - but to do so it would have to start out in what seems like the opposite direction, by engaging in deflationary monetary policy. Similarly, it could assure high nominal interest rates by engaging in in an inflationary policy and accepting temporary movement in interest rates in the opposite direction." Milton Friedman explains everything in that paper and I think that Cochrane et al. are seriously confused about the entire issue. They also do not seem to understand the income and liquidity effect

    ReplyDelete
  20. Here is one very big bit of empirical evidence against the "Neo-Fisherite" theory:

    For the last 20 years the Bank of Canada has been targeting 2% inflation. And the average inflation rate over that same 20 years has been almost exactly 2%.

    The Bank of Canada has said it has been doing the exact opposite to what Neo-Fisherites would recommend: whenever the BoC fears that inflation will rise above 2% it raises the nominal interest rate, and whenever it fears that inflation will fall below 2% it cuts the nominal interest rate.

    If the BoC had been turning the steering wheel the wrong way this last 20 years, there is no way it could have kept the car anywhere near the centre of the road. Unless it was incredibly lucky. Or was lying to us all along.

    ReplyDelete
    Replies
    1. But in order to model this formally, you really do need to think of the nominal interest rate as a signal that reveals information about the central banker's type. Nominal interest rates are a really stupid way to signal, but that's how central bankers do it. Here's my model: http://worthwhile.typepad.com/worthwhile_canadian_initi/2013/12/getting-the-right-sign-on-the-nominal-interest-rate-signal.html

      Delete
    2. Rob Rawlings6:29 PM

      Doesn't the BoC adopt that policy because it wants R = r + 2% in the long run?

      Seems like if you are targeting inflation you will hit it by using a short-term interest rate target which you will move in the same direction as any change in inflation. If you target an interest rate then you will (probably) hit it via a short-term inflation target which you will move in the opposite direction to any change in the interest rate.

      If you're targeting inflation you live in a non neo-fisherite world.
      If you're target interest rates then you live in a neo-fisherite world.



      Delete
    3. The BoC "hit" its target because an economy with unemployment and a moderate output gap staggers along with a near zero ~2% inflation. Did the BoC hit its target in 2008? No. The economy stopped complying.

      Delete
    4. W. Peden12:57 PM

      PeterP,

      Last time I checked, the Bank of Canada DID hit its inflation target in 2008-

      http://research.stlouisfed.org/fred2/graph/?g=ysZ

      It missed it in 2009, but inflation was well above target in 2011, and an average of the two years gives a figure of about 1.6% i.e. within 0.5% of the target. So if that's what an economy looks like when it "stops complying", I don't see what the problem is...

      Delete
    5. This comment has been removed by the author.

      Delete
    6. If a nation is composed of 100% devout Neo-Fisherites, is it possible that the power of expectations (that they would have) could overcome the usual mechanism here, and that increasing interest rates might cause an increase in the inflation rate?

      Delete
  21. Anonymous6:31 PM

    Where does this explicit assumption of real risk-free rate of 1-2% come from? Let that rate be negative and Kocherlakota's example falls apart.

    ReplyDelete
  22. Anonymous6:37 PM

    "Kocherlakota famously converted to a more conventional monetarist position and began arguing for more QE and higher inflation (raising the interesting, but almost certainly fantastical, possibility that his conversion is a deep-cover operation and he really still thinks "easy" monetary policy will be deflationary!)"

    Noah, you sure have a gift for making Fed stuff sound exciting. Keep it up.

    As for the Neo-Fisherian thinking, I am not overly convinced by the logic (and I am still a little annoyed at Williamson's moon visitor Volcker story). That said, monetary policy frameworks do not tend to have long life spans at center stage and given some surprising patterns in inflation and interest rates in recent years I suspect there will be some tweaks to what is conventional wisdom about how the Fed affects the economy. It is very cool to see such debates happening in a lot of forums too, so thanks.

    ReplyDelete
  23. From Robert Waldman, in a comment on my post on Kocherlakota:

    "The assumption that the economy will end up in a rational expecations equilibrium does not imply that a low nominal interest rate leads to an equilibrium with deflation. It might lead to an equilibrium in which dollars are worthless.

    I'd say the experiment has been performed. From 1918 through (most of) 1923 the Reichsbank kept the discount rate low (3.5% IIRC) and met demand for money at that rate.

    The result was not deflation. By October 1923 the Reichsmark was no longer used as a medium of exchange."

    ReplyDelete
    Replies
    1. The hyper inflation of Germany had a complicated story. It does not correspond.
      http://en.wikipedia.org/wiki/German_hyperinflation_of_1922%E2%80%9323

      Delete
    2. The Reichsbank pegged the discount rate at 5.0% from January 1914 through June 1922, a period of 8.5 years:

      https://research.stlouisfed.org/fred2/graph/?graph_id=174465

      If you look at the Index of Internal Prices on Table XIV (Pages 156-59):

      http://mises.org/books/hyperinflation.pdf

      You'll note that inflation averaged 35.2% in 1915, 7.7% in 1916, 17.0% in 1917, 21.2% in 1918, 91.8% in 1919, 257.7% in 1920 and 28.6% in 1921. In June 1922, the last month that the discount rate was pegged at 5.0%, year on year inflation reached 413.1%.

      Here's the discount rate and the year on year rate of inflation by month after that point:

      Month--interest--Inflation
      1922m7-----5.1------603.5
      1922m8-----6.1------900.0
      1922m9-----7.3----1,286.5
      1922m11---9.2----3,274.3
      1922m12--10.0----4,126.4
      1923m1---10.9----7,488.6
      1923m2---12.0--13,522.0
      1923m4---13.6----8,095.0
      1923m5---18.0--11,447.0
      1923m8---29.6--53,142.8
      1923m9---62.0--75,446.8
      1923m10-90.0-122,291,313.4

      It seems to me we've already run this experiment with predictable results (several times).

      Delete
    3. Mark, from the data you present there, could you say that by the time 5% was abandoned, it was too late? The Reichsbank never got out ahead of the hyper-inflation past 1922m7, and they were always playing catch up from that point forward? Or you tell me: why did rates start rising after 1922m7? A failed attempt to control inflation or something else?

      Delete
    4. Tom, the increases the Reichsbank made in the discount rate after June 1922 were not doing anything meaningful when inflation was running at a rate over two magnitudes greater. It must have been purely symbolic.

      Delete
    5. Correction:

      "The Reichsbank pegged the discount rate at 5.0% from January 1914 through June 1922, a period of 8.5 years:..."

      Should read:

      "The Reichsbank pegged the discount rate at 5.0% from January 1915 through June 1922, a period of 7.5 years:..."

      Delete
  24. The one thing this demonstrates to me is that we should be shutting down economics departments all over the country and starting over.

    ReplyDelete
    Replies
    1. Anonymous7:56 AM

      This talk has being going on in England. Particularly at Manchester University where action is being demanded. Human Resources at the Bank of England and Government are also not happy.

      http://rogerfarmerblog.blogspot.co.uk/2014/04/teaching-economics.html#comment-form

      In fact the links to this discussion - including the Report - are particularly good. Perhaps Noah will bring that discussion over and offer his views on this blog.

      Delete
  25. Michael Pettis wrote a very interesting post over at his blog a few years back that backed up this idea and was very persuasive. He used China as an example, saying the increase in interest rates led to lower savings rates and more spending and vice versa. I found it fairly compelling. I forgot the name of the post but you can find it over at his site. Highly reccomend it as a companion piece to this.

    ReplyDelete
    Replies
    1. louis1:56 PM

      Pettis's analysis was specific to China's "finacially repressed" banking system where consumer choice of savings vehicles is limited by law.

      Delete
    2. Yes, but I'm not sure how that disproves my point.

      The whole thesis of the article is that the Chinese chose to save less when rates were higher, because the rate of return on savings was higher, which decreased the savings rate. Spending and inflation rose as the need to save a larger portion of income diminished.

      It's not a very difficult idea and I'm not sure what consumer choice of savings vehicles have to do with any of that. Were interest bearing bank deposits outlawed recently? I mean, it's not like they have to choose between the mattress or spending it immediately. We're talking about China, not Somalia.

      Delete
  26. Impact of rates on spending is agonizingly low and the private sector is massively long govt debt => at debt levels high enough the lost interest income will swamp the former effect.

    Numbers bear it out:
    http://mikenormaneconomics.blogspot.com/2013/11/lowering-interest-rates-is-deflationary.html?

    ReplyDelete
    Replies
    1. The graph shows that interest rates cause inflation, but not in the way that you're suggesting. This is evident when one focuses on the period from 1976-1986:

      https://research.stlouisfed.org/fred2/graph/?graph_id=174464

      Note that from 1976 through 1980 when inflation was accelerating, real short term rates were consistently below the inflation rate. But from 1981 through 1986, during the period of disinflation, real short term rates were always at least 3 percentage points above inflation.

      Delete
    2. Correction. That last part should read:

      "Note that from 1976 through 1980 when inflation was accelerating, short term rates were consistently below the inflation rate. But from 1981 through 1986, during the period of disinflation, short term rates were always at least 3 percentage points above inflation."

      Delete
  27. Williamson has a marketing problem with his idea. If he'd just come up with a nice abbreviation his theory could compete with Nominal GDP Targeting.

    I suggest NIRLT (Nominal Interest Rate Level Targeting). The fed announces that from now on it will pin the ten year treasury to an arbitrary level picked out of thin air. If rates on the ten year drop below the level the Fed is targeting, they sell treasuries to raise rates (or do the buy them? That damn QE thing again). Oh well, we'll worry about that detail later.

    With interest rates always positive, inflation will always be positive, and growth will always be positive. It's perfect. I don't have time to write out a model, but just trust that this will work. Now go forth and spread the gospel!

    ReplyDelete
  28. Surely this just gets it the wrong way around. If you have a fed tracking interest rates to control inflation then there is going to a correlation. However, this doesn't mean the fed rate is "pushing" inflation in that direction. In fact you should be able to test this and show there is a lag between the fed rate and the movement of inflation after the fact. I seem to recall this being done somewhere it must be classic text book.

    ReplyDelete
  29. There is another tradition that argues that higher interest rates can be inflationary, some associated with the functional finance version of Post Keynesianism. They argue on cost-push inflation grounds that higher interest rates raise costs for business borrowers who will pass those on as higher prices, although I have never seen a study that showed that this real effect offsets other effects operating in the opposite direction.

    Barkley Rosser

    ReplyDelete
  30. I think neo-fisherites are playing games with terms here. Yes, lowering the interest rate would cause a long-run deflationary trend. It would also cause a long-run decrease in the growth rate of money supply. That is, in the neo-fisherite worldview, QE causes interest rates to rise, not fall. You can't argue that low interest rates cause deflation and then argue QE causes deflation--that's a contradiction!

    ReplyDelete
  31. Does David Glasner count as a "Neo-Fisherite", Noah? He wrote this working paper on the Fisher effect a little over three years ago.

    http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1749062

    ReplyDelete
  32. Anonymous1:16 AM

    Maybe low inflation is here because it's impossible to get wage rises because of smashed labor power? Mass unemployment so people have to live off friends/family? Cuts in gov spending? High personal debt levels meaning less spare cash for spending? Loads of part-time employment? If the QE money was stuffed through everyone's letter-boxes it might have worked, but central bankers are obsessed with undershooting their inflation targets for reputational reasons. Unemployed poors be damned.

    ReplyDelete
  33. A rebellion no less! How lovely!
    Now, what's the logic here, or rather the explanation? The empirical relation is present, between i and p, but Larry Klein demonstrated causation can run either way, back in the 1990s. Any so called model would have to allow causation to switch.
    So, we are reduced to the same episodic thinking that performed so well for some, back in '08. Forget that, or rather extend. Well, higher interest rates do allow investors to undertake more risky investments, but lower disposable capital for projects, not a problem if the marginal efficiency of capital is high. But, that's not the case today, and was also not the case in 2007. So therefore, higher interest rates, but _really_ shouldn't we be saying "financial returns" can only be obtained through non-prudential lending, whereupon a few intermediaries can make money, but in the absence of exponentially increasing velocity, the majority of investors must eventually experience the thrills of a fresh Minsky moment.
    Meanwhile, back in the real world, Keynes saw low interest rates as _a_ remedy for deflationary depression, a relief on restricted profits, but hardly an end-all.

    So, unless the rebels can present a coherent narrative based on current, and anticipated TPP business conditions, they must stay in the clubhouse, plotting.

    ReplyDelete
  34. Anonymous2:45 AM

    Lol, this was a bit of a 180 wasn't it? One day, QE and zero interest rates were sure to cause hyperinflation. SW then realized this didn't quite pan out, so he decides it's the QE and low interest rates that are causing the low inflation.

    Just priceless logic...

    Fit for a JME editor: http://www.journals.elsevier.com/journal-of-monetary-economics/editorial-board/.

    One can only fathom some of the nonsense arguments he must have penned as a referee/editorial decisions he's made. At a LAC near U there are are probably teaching professors who lost tenure decisions because they wrote down models where monetary policy had the standard impact on the economy. You can see where his self confidence comes from -- few people tells a Senior Associate Editor at JME editor he knows nothing of his field. There's no way he can ever admit a mistake in this context -- if he confesses to being wrong about basic Macro 101 stuff, he'd have to admit to himself that he's more or less a fraud, and that all those Macro models he's spent years inflicting on students were in essence a waste of time.

    ReplyDelete
    Replies
    1. Anonymous2:57 AM

      Everyone is making fun of him on : http://www.econjobrumors.com/topic/deflation-in-sweden-svenssons-advice-is-the-problem-williamson

      from where your post have been copied

      Delete
  35. Anonymous8:23 AM

    Simply put you live in the age of IT where real interest rate in the USA is likely to be negative because a lot of capital has now become obsolete. That implies nominal inflation below zero which is impossible which implies the prices cannot fall to clear the markets which implies that the economy is in a liquidity trap which implies that printing money will not raise inflation which probably implies the fisher effect.
    The fisher effect is probably simply a strong indication that the US economy is in a liquidity trap and not evidence that carrying an umbrella will create rain.

    ReplyDelete
  36. Real rates (paid by the banks) are near zero, inflation is the driver.

    When you factor the lending activities, the real rates are negative because you can not get a loan at any rate! This stifles economic growth and incomes. So, even a small inflation is a big deal! Or, we enter into deflation. I would like to add a caveat: inflation is variable depending on location and income levels; the inflation rate measured by govt and feds is at best a figment of imagination.

    ReplyDelete
  37. Anonymous8:27 AM

    Why can't there be a phase transition like in thermodynamics? I.e. for a big enough monetary base relative to the economy size monetary policy looses track. It worked in the past and it works for countries with relatively low monetary base, but not anymore for EU, US, Japan.

    ReplyDelete
  38. ReturnFreeRisk11:52 AM

    I am a novice. But I will give it a shot.

    First, here is my understanding of monetary policy - rates, money supply, QE, devaluation it really does not matter. They all work roughly in a similar manner when it comes to inflation.

    1) The central bank eases (through any of above measures) and it has a short run effect of raising inflation expectations trhough depreciating the currency, raising local currency denominated commodity prices and can be seen in the break even inflation measures of linkers.
    2) This also has the effect of raising prices of all real assets - stocks, real estate, commodities, gold etc. However, this is a one time price change NOT as the central bank hopes a persistent change in ongoing changes in prices (inflation). The central bank can not manufacture persistent inflation outside of these areas without external help. Hyperinflation is usually in war torn countries (reduction in potential output OR massive fiscal spending). To see this, consider the price of stocks or other real assets. The long term value is determined by ability of the economy to produce profits, rents on land etc etc. That is dependent upon productivity growth etc. The central bank can not change the long term value. All it does is to borrow returns from the future into the present and front loads it. It can not raise the return stream on any real asset permanently. It is the same for the prices of goods. Relative changes do not matter.
    3) The longer term effects of lower rates, QE, deval etc come into play - lowering interest income tranfered from government to private sector (on govt bonds) and the lowered purchasing power of income (higher commodity prices and import prices). Lower income and a higher tax EQUAL depressed economy. This is the channel that brings about lower inflation rates.
    4) In the extreme, the central bank's ability to lower real rates (or raise expected inflation is limited). Real rates in US/UK under QE went down to -1% to -2% ranges for longer term securities. This is via people bidding up prices of linkers in the initial phase of the monetary easing. Linkers are real assets like stocks. But they can not just keep going lower in yield. The people who are buying linkers at negative rates get made whole by actual inflation. But if inflation actually happens, nominal yields go up, which drags up real yields toward zero. And if inflation does not appear (as has been the case in Japan, US etc), then the investors in linkers will sell that non-performing asset - raising real yields. Counterintuitive but that is exactly what happened. Absence of actual inflation raised real yields.
    5) I contend the process works in reverse as well. Tightening policy will first raise currency, lower commodity prices in local currency and lower expected inflation, hit real asset prices. But in the longer run, higher interest income and lower commodity prices are good for growth (story of 2013 in the US). This should be positive for wages etc that form the backbone for inflation in the long term. However, again, there is a limit to the process on the up side as well. Limited by output gaps (just as on the downside); potential growth of the economy, productivity, fiscal policy etc etc.

    So I contend that YES, lower rates in the long run lower inflation and vice versa but there are limits to what will happen on both sides. The limits to lower inflation have been commented on by Krugman and others (although he has his own story to fit the data). There are surely limits on the upside as well. Raising rates to 20% will not raise inflation to 20% on its own - it is absurd - but then so is the opposite that cutting rates to -20% will somehow raise inflation.

    Now shoot me down.

    ReplyDelete
  39. The thought experiment fails unless conditions are already nearly in place at about the 19.5 per cent level.

    Nick Rowe does a nice job of explaining this. It's all about the path.

    And you have to be meticulous about the path if that's the desired objective.

    Otherwise you get Volcker.

    ReplyDelete
    Replies
    1. Anonymous1:49 PM

      What is wrong with Volcker? He took us off the gold standard in 1971 and broke inflation in 81-82.

      Delete
    2. Nothing wrong in that sense

      Different point

      Delete
    3. Anonymous5:15 PM

      Tsss... Noah - I would have thought you have more humour... But still another try with more explanations for the finance professors without drive to click links...

      4th paragraph from the bottom:
      "Montarists will say that this is because of expectations that the Fed's asset purchases won't be permanent, and of course we can't rule that out"
      ...Montanist takes a "n", not a "r", see
      http://en.wikipedia.org/wiki/Montanist

      For the link-disliking finance professors:
      The Montanists were a sect claiming personal holy visions - something that, in my opinion, combined with your typo - makes a good opportunity for a small joke, because I see the Monetarist view as sometimes possed by the same spirit, especially when defending against heretical ideas (like the one you talk about...)

      And if this comment irks you again - and you delete it - PLEASE FIX THE TYPO!

      6. try - typo still not fixed - long live copy&paste...

      it could be so easy to get rid of that comment - fix the typo...

      Delete
  40. Also, QE and IOR are not directly relevant to this question.

    ReplyDelete
  41. Thinking practically as opposed to through a model, it seems likely if you raise the short term rate to 20% or 200%, there is no bond market. There is a very small money supply. To the extent the economy survives, people move to equity finance, barter, so the answer is 1) Fed loses control of the 'nominal interest rate' in the sense of how real activity is financed.

    And I'm not sure what this story says about moving from 4% to 0%. You still need a story for how that creates deflation. The model has to explain a real story.

    ReplyDelete
  42. I'm confused. Doesn't Noah's thought experiment (as others have already pointed out, with no government debt---hence the Fed isn't pushing on rates via Treasuries) simplify to just printing money until inflation rises and nominal rates hit 20%?

    But this is an example of the money supply expanding until the real economy reaches a point of inflation---which drives nominal rates higher. Why this would have any implications for the Fisher-ite prediction at the ZLB---where the interest rate supposedly is to follow the money supply, not vice versa, and especially where inflation is to exist in some other way than consequent on price levels---is not at all clear to me.

    It seems like this follows as an implication from the behavior of the private supply of money at non-ZLB rates---when rates fall, fewer dollars stay deposited---and assumes that this will behave the same way when the natural rate of interest is below zero. But I just can't understand how depositors would behave that way at the ZLB: they're not satisfied with their return in deposits, but they're not likely to get a greater return in the real economy (and nor, evidently, are they willing to consume instead---the economy still being in a depression and all). So the supply of deposits keeps bumping up against a ZLB demand for it, which keeps paying very low rate, and I understand why they think they should be entitled to a rate consistent with a higher inflationary figure, but it seems obvious they're not going to get it.

    So is what happening that they're really receiving a negative real interest, plus inflation sufficient to get a small positive interest rate, but convincing themselves that they're receiving a positive interest rate, plus a negative inflation rate (or minus deflation)? It makes sense to me that people might think so---but it equally makes sense to me that they're mistaken, and that no one else ought to think so.

    ReplyDelete
  43. I guess both things can be true - inflation impacts interest rates and interest rates impact inflation. I suspect that people form a great deal of their expectations based on their experiences. A low inflation rate in the past makes them guess a low inflation rate in the future. Expectations of low interest rates lowers the "natural" i.e. inflation neutral level of interest rates. Lagged expectations of the average citizens and rational expectations of financial actors (and economic shocks) cause a wobbling of these economic variables around their natural rate. So depending on your timing and intent, you can shift the general level of rates in your economy or take an equilibrium rate as given.

    ReplyDelete
    Replies
    1. Why? Look at long-term interest rates in industrialized countries and we're already there. Besides, most people don't really make decisions with the whole "civilization is doomed" mindset in the background. They just don't care because in the long run, we're all dead anyway.

      Delete
  44. Noah
    Someone asked me if you are a Troll. "He seems to know a great deal about a wide variety of subjects, and manages to get a big crowd really frothed up".

    On a somewhat different matter are you still officially a Megan McArdle fan?

    http://www.bloombergview.com/articles/2014-04-22/piketty-s-tax-hikes-won-t-help-the-middle-class

    I think she is recommending yoga for the unwashed masses, as a path to alleviating their own lot.

    ReplyDelete
  45. Anonymous5:12 PM

    PS: ~582 people type like you...
    https://www.google.at/#q=Montarists

    ReplyDelete
  46. Anonymous5:14 PM

    Tsss... Noah - I would have thought you have more humour... But still another try with more explanations for the finance professors without drive to click links...

    4th paragraph from the bottom:
    "Montarists will say that this is because of expectations that the Fed's asset purchases won't be permanent, and of course we can't rule that out"
    ...Montanist takes a "n", not a "r", see
    http://en.wikipedia.org/wiki/Montanist

    For the link-disliking finance professors:
    The Montanists were a sect claiming personal holy visions - something that, in my opinion, combined with your typo - makes a good opportunity for a small joke, because I see the Monetarist view as sometimes possed by the same spirit, especially when defending against heretical ideas (like the one you talk about...)

    And if this comment irks you again - and you delete it - PLEASE FIX THE TYPO!

    5. try - typo still not fixed - long live copy&paste...

    it could be so easy to get rid of that comment - fix the typo...

    ReplyDelete
  47. Anonymous5:16 PM

    Noah, I'd like to see a bit more discussion of the Cambridge K in the neo-Fisherite rebellion.

    As far as I can tell, (nominal GDP / monetary base) is equivalent to three-month interest rates. NGDP/MB = t-bill rate This implies that (MB constant), increasing rates boosts NGDP.

    But central bank's only power is really the monetary base -- so how they raise rates without (relatively) shrinking the MB is nonsensical. Rates only rise when nominal GDP rises faster than the monetary base.

    ReplyDelete
  48. The inflation of the 70s was "whipped" by destroying the bargaining power of labor to increase wages. High unemployment caused downward pressure on wages. In addition Carter energy policy controlled commodity inflation in the oil and energy sector.

    Inflation is necessary to allow relative prices to reset. When wages are deflating or rising at a low rate, there will not be enough demand to drive price inflation. Wage stagnation has been a period of low inflation. Inflation did not surge in the 1980s as the Fed cut rates because workers lacked bargaining power.

    ReplyDelete
  49. Anonymous5:30 PM

    This could have been much easier for both of us - thank you...
    "Monetarists will say that this is because of expectations that the Fed's asset purchases won't be permanent, and of course we can't rule that out."

    ReplyDelete
    Replies
    1. Thanks for being persistent. I owe you one!

      Delete
    2. Anonymous1:40 AM

      Chapeau!

      Delete
    3. I usually only get persistent when I see a typo in a post title (as David Beckworth and Mike Sax can attest)... only because I feel embarrassed for the author.

      Delete
  50. I posted some ideas on the "steady-state" nature of the Fisher effect.
    http://angrybearblog.com/2014/04/dangerous-logic-fisher-effect.html

    ReplyDelete
  51. Bill Ellis11:28 PM

    "Eventually, one of two things would have to happen:"
    How about a third thing ? People would stop borrowing.

    ReplyDelete
  52. In terms of the thought experiment, it's clunky to assume that no government debt exists. Without government debt, as in your though experiment, the FED controls interest rates by offering 20% return on reserves. This, however, increases the money supply. With an increasing money supply, it isn't too difficult to see the Neo-fisherite result holding.

    I think the more interesting, and more difficult, thought experiment is to assume that the FED controls interest rates by buying and selling government debt. In this scenario, the government can only increase the interest rate to 20% by selling a large quantity of bonds. In return for selling these bonds they get cash back, thus reducing the money supply. As I see it, it is much harder for the Neo-Fisherite result to hold once interest rates and the money supply move in opposite directions (especially if wages are inflexible downward).

    ReplyDelete
  53. Anonymous10:30 AM

    Suppose we had "full employment" and the resulting greater demand, then what? Low interest rates lead to deflation and high rates to inflation or the other way around?

    ReplyDelete
  54. 20% interest payment on central bank accounts (let's assume anyone can open one to avoid all the irrelevant distractions intermediaries introduce) is sustainable without inflation forever in at least one scenario: if most (enough) people just reinvest their interest payments to see their balance go up in perpetuity, and only spend an amount that is commensurate with the real stuff available in the economy.

    If the bulk of the accounts with non trivial balances is held by people with more money than ideas on how to spend it -- this is more and more common! -- and whose social status depends on the balance rather than actual spending, this may even be possible and sustainable in real life.

    ReplyDelete
  55. I think you have to relate this to tinkerbell theory, that if everyone has the same model and expectations and act in accordance with it, their expectations will be fulfilled. A serious problem with this model is not everyone believes in it. Even if we accept the Feds adoption of this model means a majority now believe in it, those who don't can prevent self fulfillment.

    ReplyDelete
  56. But why can't real rates be higher than nominal rates?

    Aren't real rates higher under deflation by definition?

    ReplyDelete
  57. You are all so funny. Treating toy models so seriously. A good modeller, it seems, can make a reasonable case for true = false. But you all keep believing them.

    A free market is powerful. Not something weak that needs tweaking of the numeraire. Keep base money per capita constant and you will have a robust currency and a fair, productive economy.

    ReplyDelete
  58. Anonymous7:59 PM

    Is it possible that we really only have a grasp of the impact monetary policy when the monetary and fiscal authorities are working together? Clearly, printing money and spending is inflationary. Conversely, reducing the money supply and retiring government debt (say via taxation) would be deflationary. But when monetary policy is truly independent, maybe we don't understand its impact on inflation. QE is an asset swap - inflationary, deflationary, or irrelevant?

    ReplyDelete
  59. This is just so not worthy of you Noah. Yes, sure, if you create a hypothetical world where you assume the natural rate of return is fixed and assume that you can hold the interest rate either well above or well below the natural rate of return indefinitely, than in that strange and wonderful world, you can counterintuitively force the inflation rate to do unexpected things.

    However, in a real world, if the fed can hold the interest rate high above the natural rate of return indefinitely, what that means is that the natural rate of return will rise. If the fed can guarantee a 20% return indefinitely without going bankrupt, everyone will be happy to loan to the fed and accept that 20% return.

    It is not reasonable to assume that the fed can hold the interest rate artificially high indefinitely, and therefore the model is invalid. It's a perfectly good model, but it has nothing to do with how the world actually works.

    If the model is self-contradictory for a thought-experiment that holds interest rates high, it is reasonable to believe the model is probably self-contradictory when you hold interest rates artificially low.


    If you really want to pursue this model, you need to show, in detail, how inflation increases when rates are held artificially high. You need to show why people would prefer to spend excess money on goods and services instead of giving money back to the fed to increase their wealth. The best that you're going to be able to do is: since people spent so long giving money to the fed, infrastructure has been neglected and we deflated the economy by so much that there is no wealth left on which to obtain a return on investment. The rate of return becomes meaningless.

    ReplyDelete
  60. Anonymous11:05 PM

    If the Fisher relation was meant to apply to price-taking lending and borrowing individuals, then why should the Fisher relation apply to the Federal Funds rate? If it doesn't apply then there's no deflation implication arising from the low Federal Funds rate.

    ReplyDelete
  61. Unanimous11:25 PM

    With regard to the micro foundations, perhaps balance sheet effects are part of the story. That is, if interest rates are lowered initially more borrowing occurs causing more money to be spent and inflation to rise, but after a time as the debt mounts less new borrowing occurs because borrowers know that in the long term they need to pay it off, and inflation then falls in the long term.

    ReplyDelete
  62. Thanks for sharing this site, it is very informative for the business personals.Keep on continuing with this.I also provide this service.plz visit my site Ways to make money from home Our compensation plan is one of the reasons that WakeUpNow has had such tremendous growth. We are truly setting people free to live the lives they want.

    ReplyDelete
  63. Nathanael8:25 PM

    "Eventually, one of two things would have to happen: either 1) the Fed's control over the nominal interest rate would break down,"

    This is what happens. You can track this historically. In societies where the "government money" was lent out at very high rates, banks and other organizations started printing their own money *and lending it out cheaper*. Interest rates proceeded to fall.

    This happened even when it was illegal for them to do it.

    Currently the Fed keeps control of interest rates largely through laws restricting the ability of banks to print unbacked money. (Institutional structure matters!) So the banks have to borrow from the Fed. If the Fed interest rates were outrageously high, illegal banks would spring up ignoring the reserve requirements and ignoring the capital requirements and they'd start lending at lower rates.

    Perhaps the key missing element here is an understanding of risk effects. Balance sheet effects are critical in creating a deflationary trap (debt deflation) -- so while low interest rates will be inflationary initially, the debt pileup will be deflationary later.

    If you manage to avoid the debt pileup by one means or another (redistributionary taxes, perhaps) then you won't see the deflation.

    Very few people think in terms of the "natural rate of interest". Some rich rentiers and pension funds do -- they demand their 4% returns and typically adjust their risk level to match.

    But the average 99%er, or the rich rentier who's been burned once, will not be incentivized to borrow by lower nominal rates. "Preservation of capital, not return on capital".

    The microeconomics of supply and demand for money matters.

    It's complicated because there's fundamentally a conflict between demand for store-of-value and demand for medium-of-exchange; a decent microeonomic model has to track the two separate demand curves which are competing for what is basically the same supply of money. (Some bonds are store-of-value but aren't medium-of-exchange, but frankly this is uncommon since you can borrow against most of them at low rates.)

    Furthermore, on the supply side, you have to track creation of money and keep track of *what is considered money* -- "money market funds" were money until one broke the buck, and then they weren't.

    People's reaction to Fed rates, and *bankers'* reaction to Fed rates, may vary depending on conditions, and it is this which determines whether we see deflation or inflation...

    "Tony Yates reveals that there is actually a fairly large literature on this question, most of which concludes that if the Fed holds down interest rates forever, inflation becomes...indeterminate!"

    Glad to see the literature supports me.

    ReplyDelete
  64. Noah, 20% is a crazy level to think about. It surely would be a negative if that happened. There is (and has been for decades) a global carry trade in finance. The Repo, shadow banking and margin loan markets would implode. This would be a systemic shock that would have to lead to a recession.

    But what happens if there is a much smaller rise in % rates? A level that is not punitive, but higher than zero. Assume that the ST % rates set by the Fed in 2013 was 2% versus 0%. Would that outcome have resulted in a higher or lower rate of inflation/economic activity?

    One thing that would have happened is that the cost of servicing the government debt would have been higher (1/3 of all debt is short term). So, all things being equal, the government would have had a larger deficit and the receivers of this higher interest (includes money market funds) would likely spend some of that income. More government spending, bigger deficits and higher consumer spending could potentially add to inflation.

    If there is validity to the neo-fisherite thinking it probably would be a bell-graph outcome. On the left side would be zero % and very low inflation and at 10%+ the inflation rate would also be at zero. The top of the bell would be something a few % from zero.

    If I were running the Fed I would have raised the short end of the curve above zero a few years ago. In my world, the federal Funds rate would be 2% today. I believe that this would have resulted in a slightly higher level of inflation than we have today, and a slight increase in GDP from what was recorded.

    In 2009, when everything was crashing, a zero interest rate was probably appropriate. But five years later the zero rate is more of a drag than a boost.

    Call me a quasi neo-fisherite.

    ReplyDelete
  65. John W9:03 AM

    Apparently the entire point of monetary policy has been forgotten by some.

    If the Fed got interest rates up to 20% people would stop borrowing all but the most essential loans because it would be extremely difficult to achieve a profit after paying a 20% ( + risk premium) interest rate. This would create a vicious cycle of unemployment leading to shortfall of demand leading to unemployment. Basically we are talking about the great depression in the US but even worse this time (because the Fed is being even more austere). This is the whole concept behind the unemployment part of the Feds mandate. In rough cut terms the Fed controls the interest rates to make it so that the amount of investment produces demand for labor appropriate to the supply for labor. To put it another way, if there are excess workers in the market the Fed makes mortgages cheaper so home builders break ground and suck up that excess labor.

    As a historical note, 20% interest rates weren't actually that unusual in pre-renaissence Europe (and other places but Europe is probably best known here) and what we see as the result of usury is very little speculative capitalism, wealth held in land (which can be tended by feudal labor or rented) and speculative inter-city trade being done on a cash basis by the few merchants who managed to hoard the supply of silver. Resorting to the loan market and it's 20% rates was frequently a move of desperation that was very likely to lose you the security for the loan. What we certainly didn't see was this system leading to 20% annual inflation, prices changed less from 0 to 1500 then from 1850 to 2000.

    ReplyDelete
  66. Mitch9:49 AM

    It's depressing to see that economists can't dispense with such a weak argument.

    Let's forget about the Fed. Suppose *Congress*, by fiat, simply decided at what rate loans were allowed to be made. (And we'll also suppose that the ban cannot be evaded.) Let's suppose also that they set the nominal rate R quite high. No one is allowed to lend money except at a high nominal rate of interest.

    Now the relation R = r + i is a tautology, so nothing Congress can do will change it. Now what happens? What happens to economic growth? I am sure we all agree that it declines. But the loans that do get made (and there's no reason to assume there won't be some) will carry a high real rate of interest.

    We leave as an exercise to the reader to imagine what happens if Congress sets a low interest rate - either with or without the added benefit of a lender of last resort who lend at that rate.

    There are therefore two obvious fallacies in the NeoFisherite position:

    a) There's no policy-independent "natural" rate of economic growth. Why should there be?

    b) Even if there is, there need be nothing connecting it to the real rate of return on loans, especially if the government is intervening.

    ReplyDelete
  67. There's a long track record that makes it very clear that, generally, tighter monetary policy leads to lower inflation and looser to higher. You can absolutely dismiss any argument that the opposite is true. The proposed low rates = low inflation model based on the Fisher equation detailed here and the very different Williamson model mentioned here are both junk.

    On the other hand there are some things to think about that might work in the opposite direction in some cases. I think the idea most worth considering is that prolonged loose policy can push capital into a hunt for capital gains and elevated rents on existing assets, thus crowding out investment in productivity enhancement, lowering growth and thus inflation.

    Also it's very important to look at central bank action and not just at interest rates. The fact that low interest rates correlate with low inflation is merely a function of interest rate markets and would be true even in a Hayekian world without central banks.

    ReplyDelete