The Cochrane Debate

A few days ago I debated the famous John Cochrane on monetary policy.  (No, not the “if the glove don’t fit . . . ” Johnnie Cochran, rather the University of Chicago professor.)  He seemed to go easy on me, perhaps because I was from a small school.  So it was a pleasant debate.  But we certainly found many points where we disagree.  I probably talked too much.  Here is the link.

I can’t bear to watch myself, so I’ll rely on memory.  There seemed to be three areas where Cochrane had trouble accepting my views:

1.  He couldn’t understand how I could claim monetary policy was “tight” last fall.

2.  He was skeptical that the Fed could do much to create inflation, at least if its tools were restricted to things like setting an explicit inflation target, negative rates in ERs, and quantitative easing.  He does believe that a combined fiscal/monetary “helicopter drop” could get the job done.  And I recall that his first choice was having the Fed buy riskier assets.

3.  He also worried about overshooting toward high inflation.

It’s hard to make complex points in a debate format (or at least it’s hard for me), so I’ll try to briefly sketch out what I tried to say.

I am pretty sure that fiscal theorists like Cochrane have the same basic view of commodity regimes as I do.  Thus FDR’s dollar depreciation program was a “foolproof” way out of the liquidity trap, because it directly reduced the value of the dollar against a real good, by essentially changing the definition of the dollar (from 1/20.67 oz of gold to 1/35 oz of gold.)

The question is whether we can do the same under a fiat money regime.  Suppose we devalue the dollar against basket of commodities?  That’s essentially no different from devaluing against gold, and so I presume Cochrane would accept the effectiveness of that idea as well.  We can’t construct a “spot market” for all the goods and services in the CPI, because the values are only know with a lag, and the goods are not easily stored and traded.  But both of those problems vanish if we peg a futures contract linked to the CPI.  That would work much like a gold standard except that it would stabilize the future expected CPI, rather than the current price of just one good.  And again I am pretty sure that Cochrane would agree that that might work.  But even that sort of radical regime change is probably too much for the Fed to do right now.  So does the Fed have any practical options?

My argument is that they could simply commit to a well-defined target path for the price level, or better yet NGDP.   Cochrane seemed to agree that level targeting was better than rates of change.  The usual argument for level targeting is that it stabilizes expectations better.  So if we briefly fall below target, investors will then expect a fast “catch-up period,” which will help to boost AD right now.  But I seem to recall Cochrane was a bit skeptical of whether this sort of Fed promise would be credible.

I really don’t see why a Fed promise for modest inflation wouldn’t be credible.  Has there ever been a situation where a central bank promised to inflate and wasn’t believed?  And don’t say Japan, because the BOJ never made that promise.

I don’t get out much, and I think over the years I developed a certain naivete about the state of modern macro.  I read all these foolproof plans for escaping a liquidity trap by big names like Woodford, Svensson, McCallum, Mishkin, and yes even Bernanke, and just assumed that these published papers were now the start of the art view of the profession.  So I was really shocked when this crisis hit to find that most economists still had the simple Keynesian “pushing on a string” view of monetary policy at zero rates, and that they therefore believed that we now needed to rely on fiscal policy.

[By the way, Cochrane does not have that simple-minded view.  He also opposes fiscal stimulus, and would prefer that any stimulus involve unconventional monetary policy (such as a Fed policy of buying riskier assets.)  So we are actually not that far apart.  But I don’t think they even need to buy risky assets.]

Well, I have learned a lesson; just because the textbooks say that monetary policy can be highly effective in a liquidity trap, doesn’t mean that the economists who teach out of those books believe it.  And exactly the same applies to the question of how to define the stance of monetary policy.  In a previous post I mentioned Joan Robinson’s silly claim that money couldn’t have been expansionary in the German hyperinflation, as interest rates weren’t low.  Laugh all you want, but you can’t imagine how often I have heard people say monetary policy couldn’t have been tight last year, because interest rate did fall to low levels.

I suppose some might defend their views by saying that Robinson didn’t understand the Fisher effect, and we now have a more sophisticated view that focuses on real interest rates.  But that explanation won’t work, for two reasons.  First, a contractionary monetary policy can actually reduce real rates, by dramatically lowering real growth expectations.  But more importantly, the only objective estimate of real interest rates that we have—the TIPS spread—soared much higher late last summer and into the fall.  So if the response to Robinson is that we should look at real rates, then why aren’t economists looking at real rates?  When they do, they usually seem to mention backward-looking real rates, when what we need is (ex ante) forward-looking real rates.

Other economists would point to the big rise on the monetary base, but Friedman and Schwartz showed that that indicator was equally unreliable when people and banks are hoarding base money during a deflation and/or financial crisis.  Cochrane did understand this problem, and I believe he mentioned that the monetary aggregates have also gone up in the last year.  But even that won’t work.  The lesson of the 1980s is that even the monetary aggregates are extremely unreliable indicators of the stance of monetary policy.  So what is left?  I say the only sensible indicator is Svensson’s idea of the forecast of the policy goal variable.  But most economists seem to feel that the right indicator of the stance of monetary policy is something like “whatever my gut instinct tells me after I look at a few indicators like interest rates and the base.”  I find that attitude very discouraging.  And this is something on which I think Cochrane and I can agree.  He puts it much more eloquently in a recent paper on the crisis:

Some economists tell me, “Yes, all our models, data, and analysis and experience for the last 40 years say fiscal stimulus doesn’t work, but don’t you really believe it anyway?” This is an astonishing attitude. How can a scientist “believe” something different than what he or she spends a career writing and teaching? At a minimum policy-makers shouldn’t put much weight on such “beliefs,” since they explicitly don’t represent expert scientific inquiry.

Others say that we should have a fiscal stimulus to “give people confidence,” even if we have neither theory nor evidence that it will work. This impressively paternalistic argument was tried once with the TARP. Nobody could say how it would work in any way that made sense, but it was supposed to be important do to something grand to give people “confidence.” You see how that worked out.Public prayer would work better and cost a lot less. Seriously, as social scientists, economists don’t have any special expertise to prescribe what intrinsically meaningless gestures will and will not give “confidence,” so there is no reason for anyone to listen to our opinions on that score.

“Well,” I’m often asked, “we have to do something. Do you have a better idea?” This is an amazingly illogical question. If the patient has a heart attack, and the doctor wants to amputate his leg, it’s perfectly fine to say “I know amputating his leg is not going to do any good,” even if you don’t have a five-step plan to cure heart attacks. As a matter of fact, as above, there are perfectly good answers to this question, but even if there were not, it simply makes no sense to “do something” that you know won’t work.One of the most important things that scholars can do is to explain ignorance. I often say “I don’t know, but I do know with great precision why nobody else knows either.”Ninety percent of good economic policy is, “first, do no harm.”

When I read this I found it very inspiring.  It was great to see another economist make a heartfelt plea to rely on reason, not instinct.  How ironic then that Krugman of all people accused Cochrane of going back to a “Dark Age of macroeconomics”.  (Krugman doesn’t seem to have read the second half of the paper, where Cochrane relaxes the constant velocity assumption.)  If one substitutes “interest rate indicator of monetary policy” or “monetary ineffectiveness” for “fiscal stimulus” in the first paragraph of the quotation, then you pretty much have my view of the whole mess.  The economics mess, and the mess within macroeconomics.

Indeed my only objection to Cochrane is that as rigorous as he is, he is not quite rigorous enough.  I don’t like hearing him say that although markets indicate low inflation, he is worried about the risk of high inflation.  Good University of Chicago economists don’t make predictions, they infer market predictions.

PS.  I’ve tried to be fair in this post, but I am 99% sure I misinterpreted Cochrane on at least one or two points.  I will email him this post and make appropriate corrections in an update.

Update:  After posting this I realized that Krugman’s “Dark Ages” metaphor could be turned on its head.  Remember those medieval monasteries in England and Ireland, perched on a rocky islet in storm-tossed seas?  The ones that kept the tradition of classical learning alive during the long Dark Ages.  Which university played that role during the long dark period in the mid-twentieth century?  Who continued to keep classical economic ideas alive while governments were socializing investment?  Who preserved models that the world could turn to when it woke from its horrible statist nightmare around 1980?  Wasn’t it John Cochrane’s university, perched on the shores of a cold Great Lake?



32 Responses to “The Cochrane Debate”

  1. Gravatar of Bob Murphy Bob Murphy
    19. July 2009 at 09:38

    Nice post, Scott. I tried watching the debate but I can’t stand those video formats; it always looks like the guys are on the space shuttle talking to mission control.

    When you get a chance, can you elaborate on why even M1 and M2 aren’t good indicators of monetary looseness? I have never seen you discuss that before.

  2. Gravatar of Richard A. Richard A.
    19. July 2009 at 12:17

    The large increase in the monetary base was needed to offset the collapse of the M1 money multiplier. Collapse of the M1 money multiplier is due to banks holding onto excess reserves and this coincides with the Fed beginning to pay interest on reserves. I know of no other recession in recent times where the M1 money multiplier acted this way. When banks increase their reserves, it shows up as an increase in the monetary base–but this is contractionary.

  3. Gravatar of Alan Alan
    19. July 2009 at 14:39


    As a long time fan of Cochrane (particularly his finance work; I’m less familiar with his policy work, but I did find his fiscal stimulus note very impressive), I agree with most of the points you’ve raised above in regards to Cochrane’s apparent “Joan Robinson view” and his prediction of higher future inflation, contrary to market expectations.

    My question is not directly related to this post and is based on this paragraph above (while my question is based on price-level path targeting, I think the question can be adapted to (level or growth) NGDP-targeting):

    “My argument is that they could simply commit to a well-defined target path for the price level, or better yet NGDP. Cochrane seemed to agree that level targeting was better than rates of change. The usual argument for level targeting is that it stabilizes expectations better.”

    If I’m not mistaken, I think these arguments depend on an assumption that agents have completely forward-looking expectations. Also, in a model where there are costs to disinflation (which would include a model with backward-looking expectations or less than full credibility of the central bank) the gains from full predictability of the price level may not offset the costs of occasional disinflations following positive price level shocks.
    Finally, the lack of long-term indexed contracts suggests that the benefits of long-run price predictability may not be that large.

    Sorry if you’ve already addressed these issues (I tried unsuccessfully to use the search link above), so feel free to forward me onto the relevant post. Thanks


  4. Gravatar of JKH JKH
    19. July 2009 at 15:18


    Interesting debate.

    I have a few suggestions for future ones.

    I tend to view a debate as an intellectual gladiatorial contest, so please take these comments in that context.

    If you’re going to use the aggregate demand / financial crisis causality as your opening statement, you need to figure out a way of articulating it and delivering it with a one two punch. I understand it’s a difficult one, but you should work on it. If you didn’t see the video, Cochrane’s body language was not good when hearing this. Moreover, he seems to have an aggressive debating style. You need to hit somebody like that with a full paradigm sledgehammer every 3 or 4 minutes or so until he gets the message he’s not in charge. I think you’re a lot closer to being able to do that than you might realize, because of your deep knowledge of your subject.

    Cochrane’s theme all the way through was that there’s not a whole lot the Fed can do. He must have repeated this close to a dozen times. You need to spell out what you want the Fed to do so he gets past this failure of imagination. I think you did a pretty good job of this, but the message could still be strengthened.

    If I were framing it, I would bifurcate as follows. First, spell out what it is you want the Fed to do. Second, spell out what it is you want the banking system to do, BASED ON how you expect the Fed’s ideal behaviour should affect banking system behaviour. THEN, link these two ideas by spelling out the role of excess reserves.

    I.e. you want the Fed ideally to use NGDP futures and then you want the Fed to do OMO based on NGDP futures. And you want the Fed to charge interest on reserves in order to encourage commercial bank balance sheet expansion and money creation. That’s a specific battle plan for the Fed, as per all your lengthy dialogue on your own blog and with Nick Rowe and Co. You pretty much came through with this plan at the end, but it would be good to present this strongly in the middle of a debate like this. These debates may seem friendly, but the viewer wants to see who has the stronger argument. Quite frankly, I didn’t see much from Cochrane other than objection to your ideas.

    Around the 35 minute mark, you and Cochrane got into a discussion of the role of excess reserves. I think this was the most interesting and the most important part of the entire debate, but the weakest. With all due respect, this part of the explanation of how you see excess reserves playing out needs to be strengthened – both the reasoning behind charging interest on them, and what you expect the banks to do as a result of paying interest. You’ll have noted that one of Cochrane’s objections was the idea that banks wouldn’t take risk based on the reserve penalty rate, to which your response was they should be buying t-bills. This needs to be upfront as your expectation of what the Fed should expect the banks to do if the Fed charges interest on reserves. Finally, I still disagree with your technical explanation of what will happen. The banks will buy bills and create deposits. The Fed won’t necessarily flood the system with currency (i.e. Fed notes), but that’s how you seemed to explain what you expect.

    Well done, Scott, but Cochrane’s own combative inclination warrants legitimately intimidating debating techniques, such as overpoweringly superior arguments which you are quite capable of.

    That’s my take.

  5. Gravatar of Bob Murphy Bob Murphy
    19. July 2009 at 15:43

    Richard A., it’s not clear whom you’re answering there, but if it’s me: Let me just clarify that I understand why Scott thinks the explosion in base, per se, isn’t proof of easy money. But since M1 went way up, to me that proves Bernanke pumped in more base than necessary, in order to counteract the drop in the M1 multiplier. (Look at the level of M1 over the last few years; the Fed clearly increased the stock of money held by the public.)

    So I wanted Scott to explain this; I’ve never heard him say even an expanding M1 can be consistent with tight money.

  6. Gravatar of Winton Bates Winton Bates
    19. July 2009 at 17:26

    As someone who has little knowledge of monetary economics the main point of disagreement between you and John seemed to be about the Fed’s capacity to make a credible commitment to an inflation level target using conventional monetary policy instruments. It seems to me that this issue might be beside the point, given that the Fed has shown willingness to use unconventional instruments. Could it be that the main thing needed to make policy credible is the announcement by the Fed of an explicit price level target for say 10 years hence?

  7. Gravatar of StatsGuy StatsGuy
    19. July 2009 at 18:05

    I thought you did pretty decent, but you are very mild-mannered. Just way too nice.

    Here are some things to ponder over:

    1) A goldman sachs report placed the optimal Taylor Rate at something like -6% in January… So by that standard, yes, the Fed was being massively contractionary after the crisis had started. Unless, of course, Cochrane doesn’t believe in Taylor Rule targets.

    2) You guys had a largely semantic disagreement over whether creative Fed policies were monetary or “fiscal”. Honestly, it didn’t really matter.

    A lot of what cost the Fed credibility in September/October was its failure to live up to its implied promises to buy government debt in vast quantities. They didn’t announce any intention to buy in September, October, November, December, January… most of February. When they did, the effect was nearly instantaneous (and they didn’t even announce a large purchase – merely 300 billion, which is maybe a sixth of the anticipated deficit).

    So even if Cochrane can’t get off the notion that “the only thing the Fed can do is buy and sell government debt”, he should at least admit they WEREN’T EVEN DOING THAT RIGHT.

    Cochrane could argue the purchase of govt. debt was fiscal, since only reason there was ample debt to purchase was because of the huge deficits. But WHO CARES? No matter how you cut it, the Fed was pathetically weak and disingenuous. They wanted everyone to believe they had the guts to use their toolbox, but desperately hoped they would not have to use it.

    (In general, these debates would be more fun to watch if you threw out words like “pathetic”, “weak”, “disingenuous”, and “desperate” when talking about the Fed. Maybe “sadly incompetent” too. It’d be therapeutic. Try it.)

    3) The Fed’s commercial paper program was pretty decent; at the time, I recall thinking this was pretty snappy (although not big enough, and focusing only on highest grade asset backed paper, which was consistent with the Fed’s view that this was merely a liquidity crisis not a demand crisis). That’s a far cry from “buying and selling government debt”, and yet the Fed managed this under the old regime (Bush Co.) without any new legislative authority.

    4) Cochrane’s best point is that the Fed is scared to commit to printing enough money to break the deflationary cycle because it’s worried about an inflationary rebound. (And this view is precisely what restrained the Fed; it was well represented in committee hearings in which Bernanke repeatedly was challenged on the Fed’s ability to mop up excess liquidity when the crisis ended.) This view was THE ENEMY. In essence, this view was partly responsible for our current crisis because it crippled the Fed.

    It also might be correct, particularly if inflation expectations follow a tipping point model.

    Frankly, there are two ways to fix this – give the Fed tools to better control velocity (but the tools it had, like reserve rate, were being precisely mismanaged), OR give the Fed more legal tools to mop up liquidity after velocity recovers. Brad DeLong has recently echoed this latter idea (give the Fed the ability to sell bonds, and suck reserves out of the system).

    If the Fed had stated an intent to buy 500 billion in govt. debt back in October, while petitioning Congress for Fast Track legislation to authorize it to sell bonds (and declaring an intent to target NGDP, or even to target 2-3% inflation over a 2 to 5 year time horizon)…

    Well, we wouldn’t have 9.5% unemployment.

  8. Gravatar of StatsGuy StatsGuy
    19. July 2009 at 18:17

    Also, let me echo one of JKH’s points:

    When Cochrane articulated (to paraphrase) “when interest rates are coming in at 15% to 16% due to the risk premium – if you can even get a loan – it seems like 0.75% interest on excess reserves is a third order effect”, that came across as a pretty strong point. If you’re going to continue to feature this prominently, you need a rejoinder.

    Here’s a suggestion: “It isn’t just about the 0.75%; it’s about the signal this sends to banks. And that signal is a big flashing neon sign that says the Fed is not serious about getting nominal growth back to target.”

    Elaboration: “If the Fed had put a 3% penalty on excess reserves, there’s a good chance we would not have seen those 15% interest rates. The Fed was obsessed with shoring up bank balance sheets through a hidden transfer to banks, but in the process they caused vastly more damage to bank balance sheets by instigating a universal collapse in asset values and dramatically intensifying the demand/employment components of the recession.”

  9. Gravatar of ssumner ssumner
    19. July 2009 at 18:34

    Bob, In a financial panic with low inflation, FDIC-insured deposits are a refuge for ordinary people. They are the working and middle class’s T-bills. So the real demand for those assets rises. I vaguely recall that Milton Friedman warned about a big resurgence of inflation when the M’s grew fast in 1982. It didn’t happen, and that was another nail in the coffin for monetarism.

    Richard, Yes the multiplier collapsed, but that was because the fed started paying interest on reserves. It still doesn’t explain why real M1 demand rose.

    Alan, Good points, but I disagree in a couple areas. I think we need to start thinking in terms of what would happen if the Fed committed to do something, and kept doing it over and over. I agree that models are a bit unclear as to what happens when policy isn’t credible. But the solution is for the Fed to do what it says it will do. I know that answer isn’t satisfactory, but it’s the best I can do tonight. I feel more strongly about this assertion you made:

    “Finally, the lack of long-term indexed contracts suggests that the benefits of long-run price predictability may not be that large.”

    I’ve read lots of articles arguing that small menu costs can produce huge welfare losses. There is a sort of externality to wage and price stickiness. So I think the consensus is that business cycles could result from unexpected price shocks even if individuals would only have a small gain from hedging inflation risk. Sorry I can’t cite anything specific, but I am pretty confident on this point.

    JKH, Thank you for the very helpful comments. I made a mistake by not deciding ahead of time how to handle the issue of futures targeting, which I see as a foolproof way out of the liquidity trap. I assumed that Cochrane was less worried about monetary policy ineffectiveness than he actually was, and I hesitated which way to go when it came up in the debate. I needed to fixate on one idea, not waver between Fed inflation targets and a NGDP futures targeting regime.

    Winton, That’s about right, but John didn’t even think QE would work, if it involved T-bonds (as it has recently.) The key difference is that he didn’t think an explicit Fed price level target would be believed, and I do.

    Statsguy, I agree with most of what you say. By the way, the Fed already has all the tools it needs to mop up, and I think the TIPS market knows this. But I have no objection to more tools, if it helps restore confidence.
    Yes, I know I am not aggressive enough. That’s why I was relatively unknown at age 53.
    I am more aggressive as a writer than debater.

  10. Gravatar of david glasner david glasner
    19. July 2009 at 19:27

    Bob Murphy:

    Perhaps you can explain to us why you would assume that if prices are expected to fall generally and asset prices are tanking, people would not find holding money a more attractive form in which to hold their wealth than most other available assets? On the other hand, if you concede that in those circumstances the demand for money (aka liquidity preference) increases, how can looking at M1 or M2 by themselves tell you anything about whether monetary policy is tight or easy? More generally, perhaps you can explain to us why the amount of deposits created by the banking system in voluntary non-coercive exchanges (which is what M1 and M2 are measuring) ever tell you anything about whether monetary policy is tight or easy. The amount of deposits created by the banking system (at a given price level and stock of base money) is (at least at a first and maybe a second or third approximation) exactly the amount that the public wants to hold. If money is in some sense a hot potato which must be held by someone, that is certainly not the case for deposits, which can be extinguished at any moment by the public and the banks. It is only base money, the output of the monetary authority not the banks, which has to be held by someone.

    Scott, I haven’t yet watched your debate with John Cochrane. But I will bring up again one of the few points on which we don’t seem to agree, which is that conventional monetary and macro theory has no theory that provides for a determinate price level or rate of inflation. Thus, under conventional monetary theory, the Fed has no monetary instrument by which to control the rate of inflation. All it can do is posture and hope that the public will take it seriously. This was Fischer Black’s take on monetary theory for which he was seriously abused by Milton Friedman back in the 1970s when Black was at Chicago. But his view is now pretty much the orthodoxy now endorsed by Woodford et al. and other Chicagoans. According to this view, the public derives its inflation expectations from the fed funds rate targeted by the Fed in pursuit of an implicit or explicit inflation target. But an exogenous change in the public’s inflation expectations cannot, as long as those expectations are maintained, be counteracted by the monetary authority. So monetary policy is all smoke and mirrors. That’s what accounts for the policy skepticism of Cochrane. Now as a practical matter, I agree with you that the Fed should now be focused exclusively on increasing inflationary expectations and that this objective (even if we assume that is now an objective of the Fed) is being undercut by all the hand-wringing of people like Cochrane about the need to start fighting inflation as soon as it rears its ugly head. If so, fiscal policy (aka stimulus) may be performing a useful function now by creating large deficits which support expectations of future inflation.

    The two main points of my ruminations are first the importance of supporting inflationary expectations (which we agree on) and second the lack of a monetary theory that tells us how to do that in a fiat money world, because the price level and the rate of inflation are both indeterminate in that world (at least according to the conventional monetary theory that we now have). In this environment targeting the forecast certainly is worth a shot, but if you really take Fischer Black’s theory seriously (and I do), then it is not clear that even targeting the forecast is a workable policy.

  11. Gravatar of JKH JKH
    20. July 2009 at 03:01


    This is what you’re up against:

    Note the following, which has been my interpretation for some time of why the Fed has been paying interest on reserves:

    “”Interest on reserves is an important part of the exit strategy,” Fed Vice Chairman Donald Kohn said at a conference at Princeton University May 23.”

    You’ve correctly pointed out that the Fed could have modified their reserve compensation profile by charging interest at the margin on a sub-piece specifically targeted at banking system deposit multiplication.

    But I think the above bolsters my argument as to what their thinking has actually been:

    First, the “core” excess reserve position (basically all of it under the existing compensation structure) has not been targeted toward bank deposit multiplication; instead, it has served the separate purpose of acting as a specific funding source for extraordinary Fed asset expansion.

    Second, the payment of interest on reserves is a structural feature that the Fed has instituted in a more or less permanent way, in order to prepare the market for the fact that rates can go up in the future, notwithstanding the state of its balance sheet expansion at such point, and that this feature and this argument are anticipatory defenses in response to the more hawkish political and economic views on Fed monetary exposure at this time.

    It will interesting to hear what Bernanke has to say in his testimony over the next 2 days about payment of interest on reserves. I think he’s reached the tipping point where he’s going to be forced to talk about it in some detail in the Q&A.

  12. Gravatar of bill woolsey bill woolsey
    20. July 2009 at 03:35


    I have been telling you that the Fed would need to purchase risky and longer term assets.

    I have no problem with the Fed buying the safe and short term stuff first, but when the signal is expand and it has bought that up, then, it will have to go to the riskier and longer term stuff.

    Still, if the public is accumulating safe and low risk assets, creating an excess demand for money, then having the Fed buy low risk and short term assets and creating money will do no good. By reducing the quantity of the short and low risk stuff on the market, it exacerbates the shortage of it, which shifts over to money, exacerbating the shortage of money. Of course, the quantity of money is rising, but you having done anything to correct the underlying shortage.

    Of course, buying up the short and low risk stuff does not harm either. It is just a wash. But as long as we avoid this foolish notion–Base money is high by historical standards, inflation is about to take off–then so what? Buy the lowest risk and shortest stuff until it is gone, and then go riskeir and longer term.

    But to say, it is all about whether the increase in base money is permanent or not is a red herring. It isn’t permanent!!!!!

  13. Gravatar of ssumner ssumner
    20. July 2009 at 05:31

    Statsguy, I should have you debate for me next time, I agree that I didn’t have a good rejoiner for the 15% interest rate comment.
    I don’t believe interest rates are where we should focus. They are a symptom of the problem. But most people look at things through the lens of interest rates, so that’s why I face such an uphill battle.

    David, I’m not sure I can agree with this view:

    “According to this view, the public derives its inflation expectations from the fed funds rate targeted by the Fed in pursuit of an implicit or explicit inflation target.”

    Or maybe I do agree, but only in a very different sense. I believe that the fed funds rate matters only to the expect that it changes the public’s expectations of the future path of the monetary base. Thus a fed funds rate cut is expansionary if and only if it leads to expectations of faster MB growth over time. The problem with the fed funds target is pretty obvious today, it can’t be lowered below zero. Nick Rowe uses the analogy of balancing a pole in the palm of your hand. If you want the top of the pole to move left, you move your hand to the right. But what if you bump up against a wall? How do you get the poll off the wall? Woodford says promise to keep interest rates at zero for a really long time. FDR said devalue the dollar immediately. Svensson says the same. I agree with Svensson and FDR.

    I do understand that with sticky prices every short term change in the MB can be described as an equivalent change in the fed funds target, (as long as the interest target isn’t zero) but I just don’t think that approach is very helpful.

    I think Cochrane vastly overrates the indeterminacy problem. He agrees that it doesn’t apply to commodity regimes. But inflation targeting is pretty similar to a commodity regime. Even if the target isn’t precise, the public has a general idea of what the central bank is up to, and thus I don’t think they have much trouble dealing with “exogenous” changes in expectations. Rather the problem is that expectations change when the public correctly notices central bank incompetence. I don’t ever recall seeing an exogenous change in expectations, indeed I think the law of large numbers makes that impossible. Expectations change for a reason, and the reason is always a lack of confidence that the Fed will continue to single-mindedly pursue NGDP targeting.

    You said:

    “If so, fiscal policy (aka stimulus) may be performing a useful function now by creating large deficits which support expectations of future inflation.”

    The problem here is timing. Nobody expects the debt to be monetized in the near future, the worry is all about the distant future. So this makes the expected inflation gradient (or “yield curve”) steeper. But what we really need is much higher inflation over the next 12 months, not 10 to 20 years into the futures. Wages are flexible over long periods of time, so higher inflation in the distant future won’t boost real growth at all.

    I also worry that if fiscal deficits cause long term inflation expectations to rise, the Fed may be more reluctant to push QE right now. Indeed this may have happened in June, and may explain why the MB fell sharply in late June. The Fed was getting a lot of criticism because the implied 5 to 10 year TIPS spread rose well above 2%.

    Regarding your last point, I think there is a well-developed economic theory for targeting the price level in fiat money world. You peg the price of CPI futures by having the Fed buy or sell unlimited numbers of these futures at a fixed price. That deals with every objection raised by Black. The market determines the MB so there is no indeterminacy, and again it is exactly analogous to a gold standard, except you replace gold with CPI futures.

    JKH, I actually agree with what you say here, but it doesn’t change my criticism of the Fed. Let’s say the Fed does need an interest rates on reserve program as a mop up strategy (because they want to keep a lot of liquidity in the banking system even after they begin to tighten.) Set the rate at a negative level now, and tell the markets that it will be raised sharply when things get better. Is that right, or is there another problem I am missing? (Also acknowledging the possible higher cutoff point for the penalty rate I mentioned in our last conversation.)

    Bill, you said:

    “I have been telling you that the Fed would need to purchase risky and longer term assets.

    I have no problem with the Fed buying the safe and short term stuff first, but when the signal is expand and it has bought that up, then, it will have to go to the riskier and longer term stuff.”

    In one sense I totally agree, I just don’t think it would be likely. If you put the penalty rate on ERs, then all the extra QE will go into RRs and cash. How much is needed? I can’t believe the Fed would have to go past T-bills and T-notes. There are literally trillions of those sorts of bonds in circulation. We’d have hyperinflation long before we got anywhere near the riskier stuff.
    But I think you and I agree on what’s most important. Start with T-bills and go to the riskier stuff if necessary. So I don’t disagree with your policy agenda, I just don’t think we’d need to go as far as you do.

    On the permanent vs temporary question, I mean the part needed to get a permanently higher price level is permanent. I don’t mean everything they do in a liquidity trap is permanent.

  14. Gravatar of Bob Murphy Bob Murphy
    20. July 2009 at 07:45

    Scott (and others who agree with him): So if I understand your position, “tight money” goes hand in hand with price deflation, while “easy money” goes hand in hand with price inflation, right? (Or if you prefer, tight means low price inflation, while easy means high price inflation.)

    So if the demand for money goes up 50%, an increase in supply of only 40% is tight. Is that correct?

    (I’m not trying to trap anybody, I just think the above is a much simpler definition, if that’s what you are really saying.)

  15. Gravatar of Bob Murphy Bob Murphy
    20. July 2009 at 07:47

    Part of my point, is that if the above really is what you mean by tight vs. loose, then Scott could end the argument with a one-liner: “If you think the Fed was loose in late 2008, why did the CPI fall?”

    Right? If someone tries to argue with that, you just say, “We aren’t arguing about the facts, now we’re just quibbling over definitions.”

  16. Gravatar of david glasner david glasner
    20. July 2009 at 08:16

    Scott, As you know, over twenty years ago I actually described an indirect convertibility regime in my book on free banking. But if you recall in my book I also tried to make the price level determinate (again following Earl Thompson) by specifying a real demand for the monetary base. That tax component of the demand for the monetary base at least provides an upper bound for the price level for any given quantity of the monetary base. But without the anchor of a fixed real tax obligation, the price level for any given nominal quantity of money remains undetermined. So my argument in the previous post was from the perspective of conventional monetary theory which posits a demand for fiat money but no real demand for the monetary base to discharge tax liabilities, i.e., the Woodford/Black model. I was asking what constrains inflation/price level expectations in that world. Nothing as far as I can tell. But you may be right that a regime in which the central bank commits to buying and selling futures contracts at a fixed price in unlimited quantities may be enough of a regime change to tie down expectations if the commitment is credible. The commitment is analogous to pegging a countries exchange rate in terms of another. Note, however, that a sufficiently strong exogenous change in expectations may overwhelm an exchange rate peg and could also overwhelm a futures peg.

    About fiscal policy, I totally agree that we are getting very little bang out of those bucks, but we have to take what we can get where we can get it.

  17. Gravatar of Current Current
    20. July 2009 at 09:24


    The point that Scott and David Glasner are making is quite simple.

    Demand for money is demand to hold money. Money is held against uncertainty. So, if the general economic environment becomes less certain then it is likely that the demand to hold money will rise. That rise will not be associated by a corresponding immediate rise (or fall) in prices. Mises writes about this in Theory of Money and Credit somewhere near page 270 though I’m not sure quite where.

    (David Glasner poses the problem of comparing money against return on bonds, this idea is not really necessary.)

    In the current banking environment in the US many banks have far more reserves than the minimum required by law. Consequently they can expand the monetary base when demand for money increases and contract is likewise. There is a complication though, the interest on reserves. That means that banks will not necessarily do what expand the base.

    So, it is makes sense to say that “So if the demand for money goes up 50%, an increase in supply of only 40% is tight.” Whether that’s the situation we have now in the US is a different question. I’m not really sure.

    I definitely think that the interest on reserves means that it could be the situation and that’s why that policy is unwise. Though that doesn’t mean I agree with the rest of Scott’s arguments.

    FWIW I disagree with quite a lot of stuff in this thread but I haven’t got time to post on it. I will do in the future.

  18. Gravatar of JKH JKH
    21. July 2009 at 01:58


    Nothing missing on last.

    So here’s Bernanke in today’s WSJ, talking about “the exit strategy”.

    A lot of detail on reserves, and interest on reserves, as I anticipated.

    At one point, he positions paying interest on reserves ahead of draining reserves as the priority tightening mechanism:

    Based on my assessment of your views, I suspect you think he should be writing more about an entrance strategy than an exit one.

  19. Gravatar of bill woolsey bill woolsey
    21. July 2009 at 02:17


    Yes, “tight money” would be if the demand for money rises 50% and the quantity of money rises 40% (assuming they were equal to begin with.) At least at first pass. Supply and demand are equally important.

    Tight money isn’t necessarily associated with a falling CPI, but with falling aggregate expenditures–more or less, falling GDP.

    Further, it is a drop in nominal GDP from its expected growth path. (The expected growth of nominal GDP is embedded in the growth rate of money demand.) Tight money
    leads to disinflation rather than deflation.

    And, Sumner argues that the relevant measure of money is the montary base. Tight money is when the quantity of base money is less than the demand to hold it. However, as a practical matter, tight money is when the quantity of base money is at a level so that expected nominal GDP falls below its targetted growth path.

  20. Gravatar of ssumner ssumner
    21. July 2009 at 05:04

    Bob, I have consistently defined tight money as a policy expected to produce below target growth in prices or NGDP or whatever the target is. The key word is “expected.” By that standard, money was clearly tight by October last year. By the way, most economists believe money was tight in the early 1930s, and yet the monetary base went up around 30%, but the demand for money went up even more. So your thought experiment is pretty close to what actually happened in the early 1930s, and the standard view is that money was tight during that period. That doesn’t mean I am right, but it doesn’t mean that my definition is not at all weird, but rather very conventional.

    David, I apologize for the following, as I am probably missing something in your argument.

    I do not carry cash to pay taxes. If tomorrow the government said that from now on all income taxes must be paid with T-bills, then I would still carry exactly as much cash as before. Then every April 15 I would take my cash, buy a T-bill and give it to the government.

    I believe that cash and wallets play roughly the same role. Both make shopping easier and more convenient. Yet we don’t motivate the demand for wallets by insisting that the government accept wallets in payment for taxes. So why cash?

    Cash is extremely useful, so much so that even during hyperinflation people continue to use cash. There are countries that have averaged double-digit inflation for decades, without people giving up on cash. They could have gone to the store with interest-bearing bonds in their pockets, but cash is so convenient that they continue to use cash.

    I think if one really insists on having a rigorous theory of cash demand the most plausible are as follows:

    Eventually cash will be replaced by electronic money. When that occurs the cash will be redeemed by the government. That redemption value anchors the future price level.

    Or, the public believes the Fed will adjust the supply and demand for money to roughly anchor the CPI over time. This makes for a quasi-commodity regime.

    Either approach will work

    Current, I agree, and would add that it basically is the current situation in the US.

    JKH, Yes, and to say it is premature to talk of an exit strategy is an understatement. The Fed has become mindbogglingly conservative. Those conservatives who predicted high inflation are in a for a big surprise. The core inflation rate would actually be negative if measured properly. I think it is very possible that zero rates could last for a long time. I hope I am wrong, but I see no sign of the 2% inflation that would trigger a rate increase.

  21. Gravatar of Jon Jon
    21. July 2009 at 21:19

    Or, the public believes the Fed will adjust the supply and demand for money to roughly anchor the CPI over time. This makes for a quasi-commodity regime.

    As we’ve discussed in other forms and places; a CPI target is NOT a quasi-commodity regime.

    At one point, he positions paying interest on reserves ahead of draining reserves as the priority tightening mechanism:

    I found his account incredibly flaccid. He claims that the Fed would–as a fallback–take a few years to work off its purchases of debt. If indeed this is the case, then he is really talking about flattening the yield curve. That is, to hold the ER back the Fed will have to pay an overnight rate in excess of the yield available on those reserves lent for the duration of time it will take the Fed to drain the funds. Worse yet, such an approach is highly unstable. If an inflation perception sets in, the interest payments required will be substantial. Yet the Fed’s holdings will still be low-yield, low-inflation-era notes. They won’t have the cash flow to support the payments without ‘printing’ the money, which will then feedback on the inflation expectations.

    He then pays lip service to the one policy that could work: namely sterilization by means of the SFP. He neglects to explain why SFP are currently 200B despite a supposed program of QE at present.

  22. Gravatar of ssumner ssumner
    22. July 2009 at 05:35

    Jon, You should identify the speaker in each quote. Otherwise it’s hard to go back and find the context. I didn’t know who “he” was. Remember, I stop back here after addressing newer post questions.

  23. Gravatar of Jon Jon
    22. July 2009 at 14:30

    I’m talking about Bernanke’s essay.

  24. Gravatar of ssumner ssumner
    23. July 2009 at 05:34

    Jon, What is the SFB?

    Regarding inflation, it could hurt the Fed the way you suggest, but it also bails out the broader federal government. But I still think inflation won’t just suddenly spring up, it will only happen if the Fed wants it to happen.

  25. Gravatar of Ben Ben
    24. July 2009 at 06:41

    this was a fascinating debate. I wish you could have the same discussion with Frederic S. Mishkin. I would be very curious to know whether or not he bends on your views about targeting nominal demand rather than inflation, charging bank excess reserves, the ultimate power of monetary and fiscal policy to get out of a possible liquidity trap ?
    Could you organise such a debate ? That would be great to confront your views with a former Fed governor !

  26. Gravatar of Greg Ransom Greg Ransom
    24. July 2009 at 21:04

    Thanks for doing this. Great discussion.

  27. Gravatar of Scott Wentland Scott Wentland
    24. July 2009 at 22:05

    I second Greg Ransom’s comment. Thanks for the great discussion and your ongoing commentary.

  28. Gravatar of ssumner ssumner
    25. July 2009 at 06:06

    Ben, Thanks. I doubt it will happen, but then I never thought the Cochrane debate would happen either. It is possible that I will do more debates.

    I took a class from Mishkin in the late 1970s, when we were both very young. I also use his textbook in my money class. I would love to debate him. The areas where we would agree and disagree would differ from the Cochrane debate. We both think monetary policy is effective at zero rates, but he thinks the Fed has done a pretty good job (of avoiding an even worse recession.) For those interested, his May 2009 AER piece gives his views. I’d also like to debate Taylor.

    Thanks Greg and Scott.

  29. Gravatar of This Should be Good | This Should be Good |
    26. July 2009 at 21:14

    […] Scott has a blog post that sums up the […]

  30. Gravatar of Is John Cochrane Sumnerian Now? «  Modeled Behavior Is John Cochrane Sumnerian Now? «  Modeled Behavior
    21. June 2010 at 18:51

    […] It had been a while since I listened to the Sumner/Cochrane debate, so I decided to read Sumner’s take again. It turns out, this paper is actually an academic manifestation of the […]

  31. Gravatar of Morgan Warstler Morgan Warstler
    4. January 2011 at 09:14

    Scott, I lean towards you lie because you refuse to maintain consistency in your application of logic, and you do it on purpose.

    Your position on TARP has been that it has been paid back, it is the incredible shrinking bailout.

    It is not.

    Sans Sumner’s printing of the money, sans mark-to-market, with IOR, eventually ALL BAILOUT FUNDS get repaid, it’s just a matter of how far the horizon is…

    As such, the only VALID criticism of TARP was moral hazard (and the need to crush the losers) because then whether they repay or not, is meaningless – so of course that is what you keep talking about.

    Every criticism of mine has been based on your not being willing to rip off the band-aid, liquidate, bring the full force of moral hazard.

    We don’t need economists, we need referees. You refuse to follow the rules to their brutal final end.

    And you don’t admit. You lie.

  32. Gravatar of ssumner ssumner
    5. January 2011 at 16:50

    Morgan, You reply to a 2009 post to call me a liar? If I admit it will you promise not to tell anyone else?

Leave a Reply