Friedman on Keynes (and Krugman too?)

Here’s Milton Friedman (1974, p. 169):

One consequence of my rereading large parts of the General Theory … has been to reinforce my view that absolute liquidity preference plays a key role.  Time and again when Keynes must face up to precisely what it is that prevents a full-employment equilibrium, his final line of defense is absolute liquidity preference.

Hicks made a somewhat similar comment in 1937.  I’m going to talk about a recent post by Paul Krugman, which reminded me of this statement.  Here’s an excerpt:

That’s not the only factor. We also seem to have a slowdown in the rate of growth of potential output, mainly because of demography – the baby boomers are now exiting the work force, all the women are in, etc. – but also perhaps because of slowing productivity. Just to put a number to it, here’s the CBO projection of the rate of potential growth:

.   .   .

CBO thinks that we’re looking at potential growth around 1 percentage point slower than it was during the Great Moderation. To think about how this affects demand, consider the simple “accelerator”, in which producers, other things equal, invest enough to keep the ratio of capital to output constant as the economy grows.

At first I thought Krugman made a bonehead error, confusing aggregate supply with aggregate demand. But Krugman never makes bonehead errors.  His errors are infrequent and subtle and hard to detect.

[Memo to conservatives:  If you think Krugman has made a bonehead error, reread my miscommunication post.]

Later in the post things start to come into focus:

.   .  .  This says that other things equal, a 1 percentage point drop in potential growth would reduce investment spending by 2 percent of GDP.

So between the end of rising leverage and slowing potential growth, we seem to be depressing aggregate demand by 4 percentage points. That’s a lot!

Now, this effect can be offset to some degree by reducing interest rates. But can this be enough?

OK, so this is not the way I’d work through the problem, but I see his point.  He’s holding interest rates constant, and hence productivity shocks affect monetary policy (as I define it) by lowering the Wicksellian equilibrium rate.  I would note that the Fed doesn’t target interest rates except in the very short run, and Krugman is worried about a long run problem.  The Fed targets inflation, or the Taylor Rule.  Indeed the Taylor rule (when it actually works properly) is not all that far from a NGDP target, under which productivity shocks have no impact on AD.  So it’s an odd assumption for Krugman.  Unless . . .

Krugman continues:

Here’s the real short-term interest rate over the Great Moderation era, calculated as the difference between the 3-month T-bill rate and one-year inflation expectations in the Survey of Professional Forecasters:

.  .  .

The average real rate during the GM years was 1.9 percent. Given the factors I’ve described, it seems hard to avoid the conclusion that the average real rate looking forward will have to be negative. If inflation stays relatively low, e.g. 2 percent, this would mean an economy that often, perhaps usually, finds itself in a liquidity trap.

The difference between Krugman and a vulgar Keynesian is that Krugman understands Friedman’s point (even if he doesn’t agree with Friedman’s broader appraisal of how revolutionary the GT actually is.) Krugman understands that if you are going to return to this way of talking about the world, the pre-new Keynesian way of thinking about things, you can’t have the Fed successfully running a Taylor Rule.  And that means the zero bound underlies everything in this sort of analysis.  AS changes become AD problems.   In the GT it is what allowed Keynes to take issues that were traditionally non-macro (say trade) and claim they were actually macro issues. And Krugman also gets to trade at the end:

What might change this scenario? One key point could be trade. Before the 1980s, the US had more or less balanced trade. During the Great Moderation era, it ran an average current account deficit of 3 percent of GDP. Eliminating that deficit somehow would reverse most of my shortfalls. I would say, however, that the most likely way to reduce the deficit would be via a weaker dollar, achieved through low real interest rates, achieved in turn with a higher inflation target.

This is the only place where Krugman loses me.  I can see how a higher inflation target might lower the CA deficit in the short run (although even that depends on the relative strength of the substitution and income effects) but how does this work in the long run?  Does a higher inflation target permanently lower the real exchange rate?  I’m skeptical.

So perhaps there is a subtle and hard to detect problem in the post after all.  Or perhaps I still have not overcome the communication problem.  Hopefully commenters can fill me in; that’s their purpose after all.

PS.  If I was in a mischievous mood I’d say that Keynesians are like that firefighter that started fires to preserve his job.  They insist the Fed should target interest rates, a policy that contributes to liquidity traps by making the Fed “mute” at zero rates, and then insist that the liquidity traps created by their policy require Keynesian solutions.  But after the Kaminska post I’m not in a mischievous mood, so I’ll instead acknowledge that Krugman’s 4% inflation would solve the problem, although I believe 4.5% NGDP targeting, level targeting, will solve the problem with a more politically feasible and economically superior policy.

PPS.  The “Is it AS or AD puzzle?” brought back a memory.  Years ago I did a seminar at Bentley and someone asked “shouldn’t your supply curve actually be a demand curve?”  I looked and said, “OK, I guess you are right, but it doesn’t change anything important.” That raised a few eyebrows.  After the seminar my colleague said; “interesting paper, but he’ll never get that published.”

It’s in the JMCB; “Privatizing the Mint.”


Tags:

 
 
 

25 Responses to “Friedman on Keynes (and Krugman too?)”

  1. Gravatar of Kevin Donoghue Kevin Donoghue
    10. December 2013 at 09:56

    “Does a higher inflation target permanently lower the real exchange rate?”

    By itself, maybe not. But if Americans develop a reputation for being quite relaxed about the FX losses incurred by foreign holders of dollar assets, then I’d say yes it could result in a weaker real exchange rate for quite a long time. It might be worth looking at the period 1960-1990 to see how “benign neglect” of the $ influenced the behaviour of foreign fund managers.

  2. Gravatar of Bob Bob
    10. December 2013 at 10:09

    When we get to trade and currency values, we have to remember we are playing an iterative game. The Fed can easily devalue against a commodity, but you’ll only succeed in devaluing against another currency if the other central bank is OK with it.

    If Yelen wanted to make the yuan be worth 40 cents, would she really be able to do so? I have little doubt that a small central bank can do whatever it want to its exchange rate, but it sure seems that Fed policy changes could have major consequences in trading partners, so their central banks would change their policies too.

  3. Gravatar of Kevin Donoghue Kevin Donoghue
    10. December 2013 at 10:16

    As a follow-up to my previous comment, maybe younger readers haven’t encountered the idea of benign neglect applied to US exchange-rate policy, so links might be useful.

    Here’s Samuel Brittan in 2004 discussing past experience.

    I haven’t read this Barry Eichengreen paper, but he’s usually worthwhile.

    Bob,
    I think Scott would like nothing better than to see central banks competing to see who can inflate fastest! (Well for a while anyway.)

  4. Gravatar of Morgan Warstler Morgan Warstler
    10. December 2013 at 10:19

    Scott,

    What if year over year, there is less real demand for credit?

    Govt. needs to borrow less.

    Businesses need to borrow less

    Individuals need to borrow less.

    NGDPLT helps keeps things going along…

    But say over the last 5 and for the next 25, every year, we all simply have less desire to borrow money.

    It’s not a shock, it’s a new equilibrium. That’s my fiat assumption for this question:

    Ok sure, NGDPLT still is the best stabilizer thru a very tough patch of land…

    But IF there was a true change in long term credit demand.

    If everyone is trying to pay off their current loans.

    If everyone is trying to borrow less money in future.

    Do the folks who want to loan money want a HIGHER LEVEL TARGET?

    Do the folks that are trying to not borrow money want a LOWER TARGET?

  5. Gravatar of Mark A. Sadowski Mark A. Sadowski
    10. December 2013 at 10:21

    “This is the only place where Krugman loses me. I can see how a higher inflation target might lower the CA deficit in the short run (although even that depends on the relative strength of the substitution and income effects) but how does this work in the long run? Does a higher inflation target permanently lower the real exchange rate? I’m skeptical.”

    I am too, although Krugman’s area of specialty is international trade so we should keep cognizant of the fact that he probably knows what he is talking about. But here’s what I am thinking.

    I’ve made what I believe is a fairly complete study of trade during zero lower bound incidents from the Great Depression, the Lost Decade to our own era. (I even have one of what a fellow commenter refers to as my “seminars in a comment cage” on this subject. But I’ll spare you until it is more relevant.) One very consistent fact that keeps popping up is that when countries become earnest about doing expansionary monetary policy in zero lower bound episodes both imports and exports increase significantly and net exports decline modestly.

  6. Gravatar of Nick Rowe Nick Rowe
    10. December 2013 at 10:54

    He loses me there too.

    The problem is: he’s reasoning from a CA change. Then he totally loses me when he discusses the possible causes of the CA change.

  7. Gravatar of Doug M Doug M
    10. December 2013 at 11:09

    A country that runs a trade deficit must have a countervailing investment flow.

    If the Fed were to adopt a lower real interest rate and a higher inflation target, that countries financial assets would become less attractive.

    Lower real real interest rates and higher inflation targets will lower the real exchange rate.

  8. Gravatar of Kenneth Duda Kenneth Duda
    10. December 2013 at 11:20

    Scott,

    > This is the only place where Krugman loses me. I can see how a higher
    > inflation target might lower the CA deficit in the short run (although
    > even that depends on the relative strength of the substitution and
    > income effects) but how does this work in the long run?

    FWIW, my interpretation is that all Krugman is saying is that a weaker dollar lowers the CA deficit, and a higher inflation target weakens the dollar in the short run (presumably through people exiting dollar-denominated assets to seek higher returns in other currencies). i.e. there is nothing controversial here.

    May I say how much I appreciate your analysis. You are not just a great economist — you are also a great person, with the candor and intellectual honesty you bring to every discussion. I also feel that understanding your lay-press writings on NGDPLT was eye-opening for me, and I do hope that NGDPLT will one day become Federal Reserve policy. It would make the country and the world a better place.

    Sincerely,
    -Ken

    Kenneth Duda
    Menlo Park, CA
    kjd@duda.org

  9. Gravatar of Kevin Donoghue Kevin Donoghue
    10. December 2013 at 11:30

    Nick Rowe: [Krugman is] reasoning from a CA change.

    Surely not; he writes:

    “…the most likely way to reduce the [current account] deficit would be via a weaker dollar, achieved through low real interest rates, achieved in turn with a higher inflation target.”

    He’s reasoning from a higher inflation target. That makes US assets less attractive to the likes of yours truly, we shift our portfolios and the desired result follows.

    Maybe he’s wrong but I can’t see that he’s at all unclear.

  10. Gravatar of Steve Steve
    10. December 2013 at 12:08

    “Memo to conservatives: If you think Krugman has made a bonehead error, reread my miscommunication post”

    I’ve always been under the impression that Krugman’s errors are deliberate.

    PS
    If anyone wants to watch John Taylor debate Alan Greenspan, tune into Kudlow tonight.

  11. Gravatar of Dan W. Dan W.
    10. December 2013 at 12:34

    Morgan,

    Nice points to which I will add that if you look closely you will see there is an asterisk just above the T in NGDLT. This asterisks takes you to a statement that states: “assuming all else remains constant”

    As an engineering student I solved many academic problems using that asterisk. This clause and the assumption of the frictionless track are educational necessities, otherwise the answer to every problem would require much greater examination before one could even begin answer it.

    NGDP is one of many possibly helpful economic indicators. If one had to choose one indicator by which to centrally manage an economy I suppose it is not the worst of options to use. But it is terribly presumptuous to declare that it is the best and it is bad science to make this claim while not making full disclosure of all inaccuracy, ambiguity and abstraction contained in the measurement.

    Of course the very first problem facing NGDPLT is this: What is the T and who decides? The second problem that will arise is this: To what extremes will Monetary Policy go to obtain L? NGDPLT is a fantastic academic exercise but its implementation, as with any Monetary Policy, must occur in the real world, where most everything is not constant and where friction most certainly exists.

  12. Gravatar of Vivian Darkbloom Vivian Darkbloom
    10. December 2013 at 12:34

    “Surely not; he writes:

    “…the most likely way to reduce the [current account] deficit would be via a weaker dollar, achieved through low real interest rates, achieved in turn with a higher inflation target.”

    He’s reasoning from a higher inflation target.”

    I think that’s right. It’s like the House that Jack Built read backwards (the current account deficit that Krugman tore down).

  13. Gravatar of Joseph Joseph
    10. December 2013 at 13:05

    “…weaker dollar, achieved through low real interest rates, achieved in turn with a higher inflation target”

    Isn’t he still basically assuming near zero nominal rates. Or else he’s completely neglecting the fisher effect which would be bone headed mistake

  14. Gravatar of Ashok Rao Ashok Rao
    10. December 2013 at 13:05

    Why do you conclude a higher inflation target would not reduce long run interest rates? It is an appealing proposition, but if expected inflation increased, demand for cash would fall increasing demand for other assets. There would be a less than one-to-one increase in nominal rates.

    This would be a “hydraulic” explanation that Krugman dislikes. But from my (limited) understanding of the literature, much of the welfare costs from higher inflation comes from uncertainty in investment due to inflation variability which outweighs the increased demand for capital goods. This is likely not to be significant between 2 and 4%.

    Evidence seems to suggest that there is a relationship between the two, adjusting for the correct variables: http://mpra.ub.uni-muenchen.de/36443/1/MPRA_paper_36443.pdf

  15. Gravatar of Kevin Donoghue Kevin Donoghue
    10. December 2013 at 13:58

    Joseph, yes, he’s assuming low nominal interest rates; the Fed refuses to compensate Euro-dwellers (like me) with higher nominal rates, so the $ falls to bargain-basement levels. That’s how the current a/c improvement is obtained.

  16. Gravatar of ssumner ssumner
    10. December 2013 at 14:06

    Kevin, Keep in mind that from an inflation perspective, the 1960-90 period has a sharp break at 1981, when inflation started falling fast. The CA deficit continued to get worse in the lower inflation period, if I’m not mistaken.

    My model of exchange rates is complicated. Monetary policy determines nominal rates, saving investment imbalances determine real rates (including government saving as in China) and then monetary policy also determines real rates for a few years, before money neutrality kicks in. I think that’s broadly right, although I’m willing to entertain the idea that money is not 100% superneutral in the long run vis-a-vis real exchange rates.

    Bob, In a nominal sense yes (via hyperinflation), but it’s real rates that matter, and you are right to be skeptical on that score.

    Morgan, That’s where the real-nominal distinction kicks in. With a 2% inflation target that might happen. One option is to raise inflation high enough where people want to borrow. There is always some inflation rate that makes borrowing a good deal. You get lousy real returns, but it’s better than a depression or Keynesian socialism. But I don’t see it getting quite that bad.

    Mark, Excellent empirical work. When are you going to turn all your research into a book?

    Nick, Glad I’m not the only one confused. Oh wait, didn’t I say the same thing in response to your comment on Kaminska? It’s all your fault!!

    Actually on second thought the later comment by Kevin is right. I’m still confused, but I don’t think it is reasoning from a CA deficit.

    Doug, Short run yes, but in long run the Fisher effect kicks in. Perhaps it’s the tax problem, inflation makes assets less attractive due to the high inflation tax on capital.

    Ken, Thanks, that’s very kind. Yes, it might be the short run he was focused on. I do think others read it as long run, but perhaps it was a sort of throwaway line by Krugman and he wasn’t really getting into that distinction.

    Steve, No, I don’t think so. If I was looking for a flaw it might be oversimplifying the arguments of his opponents, or focusing only on data that supports his views. But I think he believes his analysis, he’s sincere in that respect. And he does occasionally mention data that makes him change mind.

    Dan, My claim is that NGDP is the best indicator of monetary policy, not “the economy” in a real sense.

    Joseph, Yes, if he assumes zero rates then a higher inflation rate would be a lower real rate. Then the question is whether the zero rate is a long run phenomenon. That might be it.

    Ashok, Maybe, but if my read of the literature is right that’s not an appealing solution. The lack of superneutrality occurs because we accumulate less capital, and become slightly poorer.

    Oddly no comments on the heart of the post. I guess people like controversy.

  17. Gravatar of Doug M Doug M
    10. December 2013 at 14:53

    “Doug, Short run yes, but in long run the Fisher effect kicks in. Perhaps it’s the tax problem, inflation makes assets less attractive due to the high inflation tax on capital.”

    The Fisher effect… the Fed cannot move the real interest rate and hence monetary policy has no affect on the economy… You can’t be a monetarist and believe that that is true. At lest, you cannot believe that the Fisher affect holds over some time frames.

    Anyway, if you buy Krugman’s argument that productivity is dropping… a drop in expected productivity would suggest a drying up in US investment opportunities, US assets appearing to be less attractive and a drop in real rates independent of the Fed.

    Taxes aren’t it, we are talking about overseas investors buying US assets. They are not paying taxes — at least not US taxes.

  18. Gravatar of Andy Harless Andy Harless
    10. December 2013 at 15:21

    Krugman’s 4% inflation would solve the problem, although I believe 4.5% NGDP targeting, level targeting, will solve the problem with a more politically feasible and economically superior policy

    More politically feasible, yes, but I don’t think it entirely solves the problem. Suppose the potential growth rate is 2.5% and the natural interest rate is (as Krugman implicitly suggests) just below -2%. In that case there is no equilibrium with 4.5% NGDP growth. So you get maybe unstable monetary policy or maybe a zig-zag growth pattern, where you alternate between recessions and catch-up periods. (At least level targeting assures that you will get the catch-up periods, but this doesn’t seem like a complete solution.)

    Now admittedly Krugman’s reasoning would imply that the potential growth rate is less than 2.5%, so technically, if all his analysis is right, there would be an equilibrium with 4.5% NGDP growth, but that seems to be cutting it awfully close. I certainly wouldn’t trust the natural interest rate to be constant at just below -2%; I’d expect it to rise and fall, and I’d expect there to be times at which an equilibrium cannot be reached. I would advocate for a combination of your solution with Krugman’s solution — an NGDP level target, but higher than 4.5%. Ideally maybe 6% (say, implicitly, 4% inflation plus 2% potential growth) but for political feasibility maybe a nice, round 5% would be acceptable, would give us an equilibrium most of the time.

  19. Gravatar of Geoff Geoff
    10. December 2013 at 15:59

    Those who don’t understand the subtlety of capital structure in a division of labor, tend to always blame a lack of currency, or lack of currency in circulation, for why there are economic ills in society such as less than full employment.

    Money as a tool of economic calculation is a concept that goes over the heads of Monetarists and Keynesians alike. Everything is in aggregates to them. Not enough supply, not enough demand, too much supply, too much demand, etc.

    Sigh…it’s the new religion.

  20. Gravatar of ssumner ssumner
    10. December 2013 at 17:12

    Doug, I don’t follow, monetarists believe the Fed can affect short term real rates but not long term real rates. Or maybe I misread you–lots of commenters seem to be making comments too terse to understand.

    Andy, Two points, one of which I’m sure you understand:

    1. The split between RGDP and inflation seems to matter to Krugman, but in my view has almost no bearing on the probability of a liquidity trap. Less inflation means more growth, as you noted. Both affect rates.

    2. I am pretty sure the minus 2% refers to the current economy, which is depressed. But in my level targeting proposal the expected NGDP growth rate would rise well above 4.5% during depressed periods. Having said that I might be wrong about he 4.5%, and elsewhere I’ve argued we should let conservatives decide:

    Inflation or socialism?

    Do conservatives want a bit more inflation, or a much bigger Fed balance sheet, perhaps with private sector assets. Neither is appetizing to conservatives, but one or the other is inevitable if the equilibrium real rate gets low enough. I’d prefer a bit more inflation in that case, and it seems you would as well. But we can cross that bridge when we come to it.

  21. Gravatar of ssumner ssumner
    10. December 2013 at 17:13

    Andy, Also, thanks for you comment backing me over at DeLong.

  22. Gravatar of Pacemaker Pacemaker
    11. December 2013 at 02:22

    I think I can see how a higher inflation target might lead to a permanently lower real exchange rate. Because of the zero bound, a higher inflation target allows the Fed to more reliably hit the (lower) Wicksellian rate instead of letting the real interest rate hover above it. Under real interest rate parity, lower real interest rates on average would then imply a permanently lower real exchange rate .

  23. Gravatar of ssumner ssumner
    11. December 2013 at 05:29

    Pacemaker, Yes, that’s possible. Then the question is whether a long run zero bound is a plausible assumption. Maybe, but the yield curve suggests otherwise.

  24. Gravatar of Erik Erik
    13. December 2013 at 02:54

    Krugman says the US have had unpresidented austerity! And yet the US economy is growing… I really like the way you describe Krugman, but I think it is somewhat surprising that he does not reconsider his views at least somewhat.

    http://krugman.blogs.nytimes.com/2013/12/12/unprecedented-austerity/

  25. Gravatar of Vivian Darkbloom Vivian Darkbloom
    13. December 2013 at 07:01

    Fiscal stimulus proponents are really having a rough time putting together a coherent story as to why the economy continues to improve after the rate of prior extraordinary spending growth declines or even after a small decrease in actual real spending (less than one-half of one percent per his graph) which was due to the sequester, ending of extraordinary stimulus measures, as well as the reduction in stabilizer spending. Notably, Krugman hasn’t even tried. I always thought that counter-cyclical spending was *supposed to go down* in an improving economy and this has frequently been the promise in Krugman’s earlier posts.

    For example:

    “So if you’re a serious Keynesian, you’re for maintaining and even increasing spending when the economy is depressed, even though revenue has plunged; but you’re for fiscal restraint when the economy is booming, even though revenue has increased.”

    http://krugman.blogs.nytimes.com/2011/05/02/hard-kerynesianism/

    I find it hard to believe that even a “serious Keynesian” as here defined would interpret this to mean going from deep deficits to equally big surpluses from one year to the next. Surely, some transition is necessary even in that model. Our economy is not “booming” by nearly anyone’s definition; however, I also don’t think that “maintaining spending” means maintaining the same rate of extraordinary spending growth even though that would be convenient if one’s true goal is to expand government rather than the economy as such.

    Apropos that graph, I like how he has presented the change in the *rate* of growth in spending and labelled it merely “growth in spending”. Many readers were predictably led to believe by that that actual real spending had declined throughout the entire period when the line goes down (but not below zero).

    I wish, once and for all, economists would agree on a common definition of “austerity” and stick to it.

Leave a Reply