It’s complicated

Macro is endlessly complicated, as I was reminded while reading Tyler Cowen’s recent post on his view of macroeconomics.  One could write an entire book evaluating Tyler’s claims, but I’ll limit myself for now to this observation:

5b. Monetary stimulus to be effective needs to be applied very early in the job destruction process of a recession.  It is much harder to put the pieces back together again, so urgency is of the essence.

I sort of agree, but there’s much more here than meets the eye, so let’s play around with it a little bit:

1. What does Tyler mean by monetary stimulus?  Does he mean “concrete steppes”, such as money creation and/or cuts in the Fed funds target?  Or does he mean an easier monetary policy as I would define it—higher expected NGDP growth?

2. In other words, there are two ways to read Tyler’s claim; that it’s hard to boost AD quickly when the economy is already sliding into a deep recession, or that even a quick and vigorous boost to AD will have limited impact on employment, in the short run. Both claims are defensible, but they are very different claims.

3. For instance, when the economy is deep into recession, the Wicksellian equilibrium interest rate is generally falling rapidly.  That means a cut in the fed funds target may not effectively ease monetary policy.  Using monetarist language, V may be falling and so a monetary injection may not boost M*V.

4. If NGDP is sharply boosted in a recession (and I believe it can be) then there’s also the “job-matching” problem.  People who lose jobs typically are not re-employed at the same company.  Thus RGDP may not rise, even if NGDP does.

5. On the other hand, the net change in employment, even during a recession, is quite small compared to the background rate of job creation and job destruction. I seem to recall that roughly 31 million jobs are created each year, and 30 million are destroyed.  (Relying on memory, correct me if that’s wrong.) If so, then the job-matching problem is not likely to prevent a quick end to the recession, and vigorous recovery.  Just do enough to stop the layoffs, and the flow of new jobs that is always occurring in the background will rapidly boost employment.

All of those perspectives barely scratch the surface of the issues raised by Tyler’s claim. Here’s another way of looking at it:

1.  A sound monetary policy will not fix recessions, it will prevent them.  That’s because recessions tend to occur when expected future NGDP (1 and 2 years forward) drops significantly.  And that sort of drop reflects bad monetary policy. The way to stop a recession that has already begun is to prevent it from happening in the first place.

2. And no, I’m not simply making a “If I was headed to Vegas I wouldn’t start from here” snide remark, I am quite serious.  A recession is a sustained drop in output, but it technically begins when output starts dropping.  That means a recession can be prevented from occurring, even after it has started (indeed up to about the 6 months point.)

3. Here’s an exchange rate peg analogy.  Suppose you are pegging exchange rates, and are asked how to react to a 10% fall in the value of the currency.  What is the answer? My answer is, “don’t let it fall 10%.”  But what if it does fall 10%?  Then you try to punish the speculators by quickly pushing it back up again, so that the bears lose a lot of money.  By analogy, the best solution is to have a monetary regime where expected NGDP growth always remains on target.  But if it slips for some reason, then bring it back up immediately.  Then there is the issue of how quickly actual NGDP responds to a recovery in NGDP expectations.  I say “really fast” but can’t prove that.  It would be an interesting hypothesis to test.

So far I’ve been hinting that Tyler’s claim is too simple, that perhaps a quick recovery via monetary stimulus is possible.  But there’s a third way of looking at this issue, which tends to support Tyler’s claim:

1. If markets are efficient, then NGDP expectations are rational.  If I’m right that recessions are caused by sharp declines in expected future NGDP, which then depresses current NGDP, then a quick recovery can only occur if the market forecast was wrong.

2. Now think about Tyler’s claim from the following perspective.  How optimistic should we be that a monetary regime can solve problems that the market currently thinks it will fail to solve?  An EMH fan like me would say, “not very optimistic at all.”  If your policy regime is to “vigorously boost NGDP expectations whenever you go into recession”, then you should never go into a (demand-side) recession in the first place. It’s like if your policy were to immediately bring an exchange rate back to the peg anytime it fell more than 1%, then it should never fall 10%.  If it does, your policy has already failed, or more precisely has already been expected to fail.

3. Most economists think in terms of; “What do we do at the zero bound?”  I think that’s a really bad way of thinking about the problem, like discussing what to do if the bus is flying over the guardrail.  You want a policy that avoids zero rates, by avoiding really low NGDP growth expectations.  If you have a recession and/or a zero interest rate, it’s pretty clear your policy has failed.  So I can’t blame Tyler for being really pessimistic about the prospects for monetary “rescue” at that point, it would require policymakers to be smarter than the market, which they ain’t.

To summarize, we have a recession because the market thinks a monetary rescue is really unlikely to occur.  That doesn’t mean there is any technical barrier to a monetary rescue, but it does suggest that the institutional structure of the monetary policymaking process is not conducive to a quick recovery.  So pessimism is in order, albeit perhaps for a slightly different reason from what most people assume.

Note that this analysis does not apply to the business cycles of the 1920s, 1940s or 1950s, when recoveries from recession were very quick.  I don’t know why we don’t have those sorts of recoveries anymore (although of course I have theories).  If elite macroeconomics were as successful as its practitioners would like us to believe, then we’d have an answer to this question; we’d know why recoveries have become agonizingly slow.

PS.  I strongly disagree with Tyler’s point 5a on sticky wages, but will wait for the clarifying post he promises.

PPS.  Lars Christensen will be doing a speaking tour in 2016.  I strongly recommend that you book him for any topic other than “Why China will never be the largest economy in the world.”

PPPS.  I hope people didn’t miss the excellent piece on the euro, by Matthew Klein.

PPPPS.  Although I don’t agree with every single detail, I strongly support the general thrust of this Paul Krugman post on terrorism.

 

 


Tags:

 
 
 

61 Responses to “It’s complicated”

  1. Gravatar of Major.Freedom Major.Freedom
    17. November 2015 at 18:06

    “A sound monetary policy will not fix recessions, it will prevent them. That’s because recessions tend to occur when expected future NGDP (1 and 2 years forward) drops significantly.”

    That is incorrect. Recessions are caused by previous inflation which causes malinvestment. The inevitable correction cannot be prevented by more inflation. The drop in aggregate spending, which is associated with drastic changes in relative spending, is a healthy market force determined change. Central banks attempting to reverse this with more inflation will only set the economy up for an even deeper recession later on. Perpetual “stable” NGDP growth is an impossibility without eventual hyperinflation, since more and more money would be needed to offset the growing corrective, deflationary forces building up in response.

    Central banks are not omnipotent.

  2. Gravatar of Ray Lopez Ray Lopez
    17. November 2015 at 19:01

    Sumner reaches new heights of lunacy with this post.

    Sumner: “What does Tyler mean by monetary stimulus?” – same thing you mean by “Pop Keynesianism”, anything you want it to mean.

    “A sound monetary policy will not fix recessions, it will prevent them” – riiiight. Abolish the business cycle. Better yet, have Congress pass a law abolishing the business cycle, done! I think the Communists promised the same thing with their system.

    “But if it slips for some reason, then bring it back up immediately. Then there is the issue of how quickly actual NGDP responds to a recovery in NGDP expectations. ” – two points made, back to back. First, what if the economy does not come back up “immediately”? Sumner’s answer in another post is simply to ‘print more money’, which is potentially hyper-inflationary notwithstanding that money is largely neutral. Second, Sumner himself invokes the metaphysical ‘expectations fairy’ yet admits he doesn’t know what the effect will be: “I say “really fast” but can’t prove that. It would be an interesting hypothesis to test.”

    Sumner, you can’t prove anything with your wordy model, and your ‘interesting hypothesis to test’ could be the ruin of the US economy. You should let deep thinking be done by somebody more qualified, like your critic Jason Smith.

    PS–I see MF also picked up on your ludicrous ‘abolish the business cycle’ passage. Time to use that fancy WordPress editing tool and strike out the offending Sumnarian prose , Sumner.

  3. Gravatar of TravisV TravisV
    17. November 2015 at 19:09

    Someone wise once said “Do or do not. There is no try.”

  4. Gravatar of Ray Lopez Ray Lopez
    17. November 2015 at 19:27

    OT- to see how ludicrous Sumner’s conception of monetarism is, note this May 2013 passage where Sumner states a difference of a few months (two months, to get to zero interest rates, and BTW Sumner, as Jason Smith points out, ignores the implications of a Keynesian liquidity trap at 0%) in easing monetary policy is the difference between having a Great Recession and arguably avoiding one. As if anything in nature is so abrupt (it is not). Put another way, if zero interest rates were necessary to jump start the economy, and in fact such zero interest rates occurred in December 2008 rather than September as Sumner wished, then why didn’t business pick up in Dec 2008? You mean to tell me that the two months delay in cutting rates permanently crippled business expectations of recovery? (the Expectations Fairy again). That it resulted in such damage that the economy could not bounce back? Rehiring a fired worker is that hard?

    Sumner (May 2013): Werning is expressing the conventional wisdom about interest rates in late 2008, but in fact the Fed was shamefully slow to ease policy. A Svenssonian “target the forecast” approach would have forced the Fed to cut rates fast enough so that expected NGDP growth was always on target. And yet rates didn’t hit zero until mid-December 2008, by which time it was obvious to everyone that we were in a severe recession. The Fed did not even cut rates by a measely 1/4% in the meeting of mid-September 2008, 2 days after Lehman failed. They weren’t targeting the forecast.

  5. Gravatar of Philo Philo
    17. November 2015 at 20:10

    You write: “People who lose jobs typically are not re-employed at the same company.” On the face of it, this claim looks dubious. I think what you mean is this. When a company has fired or laid off workers because of a recession caused by the expectation of lagging NGDP, and then NGDP expectations are suddenly boosted back to trend, the proportion of its own fired/laid-off workers that a company will rehire is inversely related to the length of time between the firing/laying-off and the recovery of NGDP expectations. Is that it? (And you also assume, not implausibly, that job-churning is bad for RGDP.)

  6. Gravatar of John Handley John Handley
    17. November 2015 at 21:21

    Scott,

    “Most economists think in terms of; “What do we do at the zero bound?” I think that’s a really bad way of thinking about the problem, like discussing what to do if the bus is flying over the guardrail.”

    Except that we are at the zero lower bound and have been for several years. You can’t just pretend it’s not a problem by suggesting that we never should have ended up at the zero lower bound.

    “You want a policy that avoids zero rates”

    Exactly, this is why New Keynesians advocate for Taylor Rules. Models say that, if monetary policy is optimal (maximizes the households utility function), then the zero lower bound will rarely bind and that Taylor Rules are a good approximation of optimal policy.

    “by avoiding really low NGDP growth expectations.”

    Ignoring liquidity and changes in time preference, the nominal interest rate on safe assets should be a function of expected NGDP growth (see the consumption Euler equation in your microfounded mathematical macro model of choice), so this basically amounts to “avoid the zero lower bound by avoiding the zero lower bound.”

    “To summarize, we have a recession because the market thinks a monetary rescue is really unlikely to occur. That doesn’t mean there is any technical barrier to a monetary rescue, but it does suggest that the institutional structure of the monetary policymaking process is not conducive to a quick recovery.”

    Except when there is a technical barrier, like the zero lower bound.

    “easier monetary policy as I would define it—higher expected NGDP growth”

    So, basically high nominal interest rates (i = E_t pi_t+1 + f(g_t+1) where i is the nominal interest rate, pi is the rate of inflation, and g is the growth rate of the economy) are consistent with easy monetary policy, in your view. Of course, raising nominal interest rates should cause future NGDP to be substituted into the present, which would mean lower current NGDP, but high expected NGDP growth… It seems you’ve been using the same faulty indicator of monetary policy as the New Keynesians…

    For a bit of an explanation of why nominal interest rates should indicate expected NGDP growth, consider the following linearized consumption Euler equation:

    c_t = c_t+1 – a*(i_t – pi_t+1 – rn)

    where c_t is consumption, i_t is the nominal interest rate, pi_t is the inflation rate, rn is the rate of time preference, and a > 0.

    solving for i_t yields

    i_t = (1/a)*(c_t+1-c_t) + pi_t+1 + rn

    assuming output is equal to consumption, this can be rewritten as

    i_t = (1/a)*g_t+1 + pi_t+1 + rn

    where g_t is the growth rate of real output. Assume a is about unity, so the equation becomes

    i_t = (g_t+1 + pi_t+1) + rn

    or i_t = n_t+1 + rn

    where n_t is the growth rate of NGDP. If rn remains relatively constant, then the nominal interest rate should be a direct indicator of expected NGDP growth, your preferred indicator of the stance of monetary policy. If you raise the nominal interest rate, NGDP growth will certainly be expected to be higher in this model, but this could be because current NGDP is lower (something I would think you’d define as a recession) or because future NGDP is higher.

    Optimal policy in this model is to peg the nominal interest rate at a constant difference from the time preference rate (or perhaps the natural real interest rate). If this ever happens to go below zero, then expectations of NGDP growth will be higher than the central bank’s target. In other words, if the natural rate of interest falls too much, the NGDP growth expectations will rise specifically because interest rates are too high which will cause current NGDP to fall and current output to fall.

    Now, in this model, we’ve run into the same problem that New Keynesian’s have with interest rates, just concerning a different variable. “Does higher expected inflation mean higher future prices or lower current prices?” (Neo-Fisherian argument) becomes “Does higher expected NGDP growth mean higher future NGDP or lower current NGDP?”

    You can rely on a couple different techniques to get yourself out of this problem. One option would be to join John Cochrane in supporting the fiscal theory of the price level, another would be to follow the lead of this (http://www.patrickminford.net/wp/E2009_21.pdf) paper (I tried to explain it here, but I realized it would take far too long and is a bit too off topic, it’s a good read though).

  7. Gravatar of jknarr jknarr
    17. November 2015 at 22:32

    Re #3, I’ve always wondered why we cannot just call V=interest rate. They are the same for all intents.
    https://research.stlouisfed.org/fred2/graph/?g=2Bh9

    Please also consider Tyler’s #9. Debt rearranges AD temporally – adding to the present and taking from the future. Existing debt is anti-AD, in short.

    Surely a huge accumulated load of anti-AD would affect economic performance and interest rates.
    https://research.stlouisfed.org/fred2/graph/?g=2Bhq

  8. Gravatar of Ray Lopez Ray Lopez
    17. November 2015 at 22:43

    OT- For those of you that still believe in the flawed Friedman et al. narrative about the tight Fed being behind the Great Depression, I invite you to read this short paper, found online, and recommended by Tyler Cowen. Note C. Romer is an expert on Depression era history, and she skewers Friedman though she is clearly sympathetic to him. The ONLY part of Friedman’s narrative that fits monetarism is the episode where the UK/US went off the gold standard, and it ‘helped’ them (other nations did not, and it made no real difference, but I digress). From that SINGULAR example, the monetarists try and make a big deal, including B. Eichengreen. It’s pathetic.

    RL

    Title: NBER Macroeconomics Annual 1989, Volume 4, Volume Author/Editor: Olivier Jean Blanchard and Stanley Fischer,
    editors, Publisher: MIT Press, Volume URL: http://www.nber.org/books/blan89-1
    Conference Date: March 10-11, 1989, Chapter Title: Does Monetary Policy Matter? A New Test in the Spirit
    of Friedman and Schwartz, Chapter Author: Christina D. Romer, David H. Romer
    Chapter URL: http://www.nber.org/chapters/c10964

    sample: “The issue of whether monetary or non-monetary forces were primarily responsible for the initial two years or so of the collapse of economic activity that began in 1929 has been sufficiently debated that there is no need for us to argue that the case in favor of a monetary interpretation ***is not clear cut***.”

    PS–I see John Handley is taking on Sumner at his own game, which assumes money is largely non-neutral, complete with fancy math. Let’s see if Sumner blows him off like he does me, via misdirection and changing goalposts.

  9. Gravatar of Ray Lopez Ray Lopez
    17. November 2015 at 22:47

    @myself – since the quoted C. Romer et al. language is ambiguous, I add the followup sentence. She is saying monetary factors were not clearly responsible for the post 1929 collapse. Romer et al. – “As in the other episodes we have discussed, non-monetary forces were strongly contractionary during this period (see Temin 1976, and Romer 1988b)”

  10. Gravatar of Jason Odegaard Jason Odegaard
    18. November 2015 at 03:01

    @Major.Freedom

    “Central banks are not omnipotent.”

    Omnipotence is not needed for a government agency to execute actions that improve the coordination of humans involved in the system influenced by the government. While the course of action may not be perfect – nothing is perfect – it is still better than no action. This certainly applies to monetary authorities.

  11. Gravatar of Benjamin Cole Benjamin Cole
    18. November 2015 at 05:09

    Why slow recoveries from the 1980s recessions onward?

    Isn’t the simple answer that central banks have been fighting inflation? They have used recessions to crank down inflation.

    They have succeeded admirably. But, along the way, central banks became ossified and sanctimonious.

  12. Gravatar of ssumner ssumner
    18. November 2015 at 06:10

    Ray, You said:

    “But if it slips for some reason, then bring it back up immediately. Then there is the issue of how quickly actual NGDP responds to a recovery in NGDP expectations. ” – two points made, back to back. First, what if the economy does not come back up “immediately”?”

    Sorry to have to tell you this, but “it” is not “the economy.”

    Philo, I think the data support my claim about re-employment, but perhaps someone else can chime in.

    John, You said:

    “Except that we are at the zero lower bound and have been for several years. You can’t just pretend it’s not a problem by suggesting that we never should have ended up at the zero lower bound.

    “You want a policy that avoids zero rates”

    Exactly, this is why New Keynesians advocate for Taylor Rules. Models say that, if monetary policy is optimal (maximizes the households utility function), then the zero lower bound will rarely bind and that Taylor Rules are a good approximation of optimal policy.”

    I disagree with almost everything here:

    1. I don’t believe we are at the zero bound. The term “bound” means the limit of zero is stopping rates from being lower. But the Fed doesn’t want lower rates, rates are right on target, ands likely to rise. (Ands of course the Europeans have shown that negative rates are possible, as Yellen has acknowledged.)

    2. The Taylor Rule would have called for even tighter monetary policy in 2008 than actual policy, and hence would have been a disaster. That’s why we need to replace the Taylor Rule with NGDPLT, combined with NGDP futures targeting. Each of those changes would have made it more likely that economy avoided the zero bound.

    You said:

    “Ignoring liquidity and changes in time preference, the nominal interest rate on safe assets should be a function of expected NGDP growth”

    It’s a function of both NGDP growth, and the level of RGDP relative to trend.

    You said:

    “So, basically high nominal interest rates (i = E_t pi_t+1 + f(g_t+1) where i is the nominal interest rate, pi is the rate of inflation, and g is the growth rate of the economy) are consistent with easy monetary policy, in your view. Of course, raising nominal interest rates should cause future NGDP to be substituted into the present, which would mean lower current NGDP, but high expected NGDP growth… It seems you’ve been using the same faulty indicator of monetary policy as the New Keynesians…”

    This is the neoFisherian fallacy, not my view. Please see one of my many recent posts criticizing the neoFisherian view.

    Interest rates are a faulty indicator because (contrary to your claim) they are not reliably linked to expected NGDP growth, at least in the short run. If interest rates did equal expected NGDP growth, then we could peg expected NGDP growth merely by pegging interest rates.

    It does absolutely no good to write down a model where the nominal interest rate is expected NGDP growth, because the model is wrong. Yes, the two variables are strongly correlated, but not identical. And the differences are too significant for interest rates to be used as a policy target.

    The basic problem is that NKs talk like changes in the nominal interest rate are a policy, but they aren’t, they are the effect of various monetary policy choices. You can raise nominal interest rates with a tight money policy, or an easy money policy, as I’ve shown in other blog posts. If you raise interest rates with a tight money policy then expected NGDP growth will fall, if your raise rates with an easy money policy then expected NGDP growth will rise.

    You said:

    “You can rely on a couple different techniques to get yourself out of this problem.”

    I don’t like the fiscal theory of the price level, there’s no empirical support in the US. The best fix is to link the medium of account to another asset price. In the old days they used gold, but the demand for gold is too unstable. Today we could peg it to a basket of goods (CPI futures), or better yet NGDP futures. But by all means banish interest rates from the monetary policy process—ignore them.

    jknarr, Debt does not affect NGDP, unless monetary policy is incompetent. And there are 100 ways policy can be incompetent, and hence 100 ways it could affect AD.

  13. Gravatar of ssumner ssumner
    18. November 2015 at 06:17

    John Handley, BTW, the paper you link to seems like a reasonable approach, based on the abstract. I’ve always believed there was an implied real backing for money. Thus the public assumes that if fiat money is ever abolished, it will be converted into another asset at fair market value (based on the current price level.) That is, of course, what happened to holders of the various European currencies, when they joined the euro, although in this case the euro was just another fiat currency. But as long as you promise to exchange money for some real asset in an extreme case, it solves the Cochrane indeterminacy problem, doesn’t it?

  14. Gravatar of collin collin
    18. November 2015 at 06:26

    I actually agree with Tyler that fiscal and monetary stimulus work best when they work fast. I called my position on fiscal stimulus as an One and Done Monetary Stimulus. (Long term Keynesian like Japan creates an economic system around monetary stimulus).

    Working in an office, I still don’t see the libertarian position on laying people off versus sticky wages. What I have seen is firms tend to lay people off versus larger wage cuts because:

    1) If a firm cuts wages for all workers, then the top workers will simply find another position with company that pays higher. (It might take 18 months.)
    2) Firms not only have low marginal productive workers but low marginal productive departments. (Often the department were created to expand the firms core business and are not long term successful.) So it is easier for firm to cut a marginal department altogether versus cutting into productive core business.
    3) Firms can measure the marginal productivity of a worker so it is easier to identify ones to lay off.
    4) And finally, firms like the subtle message of creating fear with their workers that if you don’t produce then you can be laid off.

  15. Gravatar of Ray Lopez Ray Lopez
    18. November 2015 at 06:46

    @Sumner – what is “it” then? You said: “where expected NGDP growth” and that’s what I reference as ‘the economy’. NGDP is not the economy then?

    You’re quite clear at being obscure.

  16. Gravatar of Ray Lopez Ray Lopez
    18. November 2015 at 06:50

    Never heard of him until now… https://en.wikipedia.org/wiki/Cliff_Asness

    Seems like he’s lived quite the charmed life so far.

  17. Gravatar of Ray Lopez Ray Lopez
    18. November 2015 at 06:51

    @myself -oops, sorry wrong blog…was meant for Marginal Revolution.

  18. Gravatar of Patrick R. Sullivan Patrick R. Sullivan
    18. November 2015 at 07:13

    I usually agree with Scott’s take on things, but not his praise for Krugman’s terrorism column. Which is the equivalent of whistling as you walk past the graveyard at midnight.

    The Muslims who have settled in Europe do NOT have the same values (polls show that a majority of Muslims actually sympathize with the goals of the terrorists, if not their methods). They certainly don’t agree with Western ideas of freedom to speak ill of Islam.

    We have a century’s worth of evidence that crackpot ideas held by small numbers of people (with names like Lenin, Stalin, Mussolini, Hitler) can grow and dominate a broader culture. Russia gave us Tolstoy, Dostoyevski, Rachmaninov…and Communist totalitarianism. Germany; Bach, Goethe, Schiller…and Nazism. Italy; DaVinci, Michaelangelo…Fascism.

    Not to mention that Mohammed’s ideas were once confined to a tent in a middle eastern desert. Krugman is engaging in self-deception.

  19. Gravatar of Patrick R. Sullivan Patrick R. Sullivan
    18. November 2015 at 08:01

    The New York Times has, over the years, been guilty of underestimating ‘terrorism’ of all sorts. Famously, in December 1923 they reported on the ‘Beer Hall Putsch’ as having comic opera touches. and of Hitler’s ‘”coup”.’

    They were back at it in March of 1924, reporting on the trial of Hitler and Ludendorff for ‘high treason’, with; ‘the sequel to their comic opera beerhall revolt of Nov. 8.’

  20. Gravatar of Patrick R. Sullivan Patrick R. Sullivan
    18. November 2015 at 08:10

    Of course there was also the original NY Times report on the Hitler putsch on Nov. 10, 1923, under the headline ‘BAVARIAN OPERA BOUFFE.’ In which we can read about the ‘fantastic project upon which Ludendorff and Hitler imprudently embarked and which has so quickly come to disastrous wreck.’

    Because the Germans were united! Hitler and Ludendorff ‘left entirely out of their reckoning the powerful sentiment of the essential need of unity, which has been bred in the hearts of two generations of Germans since 1870.’

    So, nothing to worry about. Take it from the NY Times.

  21. Gravatar of Patrick R. Sullivan Patrick R. Sullivan
    18. November 2015 at 08:36

    Mark Steyn is a lot closer to the truth (and much funnier) than Paul Krugman and John Oliver;

    http://www.steynonline.com/7300/croquembouche-lie-people-die

    ‘As for bringing a suicide bomb to a pastry war, I like a croquembouche as much as the next fellow. But take it from a notorious Islamophobe: one of the least worst things about the Muslim world are the pastries.’

  22. Gravatar of Patrick R. Sullivan Patrick R. Sullivan
    18. November 2015 at 09:09

    It’s not very hard to find people who are more knowledgeable and sensible than Very Serious NY Times Columnists. Such as Claire Berlinski (she lives in Paris, and is the granddaughter of Jewish refugees);

    https://ricochet.com/claire-berlinski-answers-viktor-orban-part-1-of-4/

    ————–quote————-
    …my perspective is colored by the decade I spent in Turkey. It is impossible to say to someone who lived as long as I did in the Islamic world that Islam is a singular thing. I have too many Muslim friends to think that, but even if I didn’t, I hope I’d have worked that out just on the basis of common sense. But that’s not my key point. The key point is the other thing I saw in Turkey, the thing that marked me most permanently. I saw the rise of an authoritarian regime — a real one — and how such regimes take power, step-by-step.

    Turkey was of course never a liberal paradise. But I lived through a long, twilight period where the lights grew dimmer and dimmer, even as the rest of the world insisted they were blazing strongly. This didn’t happen because Turkey is a Muslim country (and in fact, it is not, technically; it is a secular country with a majority Muslim population). It happened because the safeguards against the authoritarian temptation weren’t strong enough.

    I’ve come to believe that absent those safeguards, it’s not — as I and many others’ hopefully thought in the wake of the fall of the Berlin Wall — mankind’s natural disposition to move forward into broad, sunlit uplands of liberal democracy. I no longer believe there’s a natural trajectory of history at all. But certain forms of government seem to become popular at various periods of history. The form of government for which the early 21st century will be noted by historians is not liberal democracy, but competitive authoritarianism, Viktor-Orban style.
    ———-endquote—————–

  23. Gravatar of TravisV TravisV
    18. November 2015 at 09:57

    David Beckworth has a fantastic new post:

    “How to Trigger a Panic Attack at the Fed”

  24. Gravatar of TravisV TravisV
    18. November 2015 at 10:28

    Does anyone know whether David Tepper is a Market Monetarist or something else?

    http://www.bloomberg.com/gadfly/articles/2015-11-17/david-tepper-s-not-the-only-one-who-should-watch-yuan

    “David Tepper of Appaloosa Management told the Robin Hood Investor’s Conference on Monday that the yuan is massively overvalued, Bloomberg reported, likely echoing the thoughts of many hedge-fund peers if not Donald Trump.”

  25. Gravatar of Wednesday assorted links Wednesday assorted links
    18. November 2015 at 11:22

    […] 5. I like it when “it’s complicated.” […]

  26. Gravatar of John Handley John Handley
    18. November 2015 at 17:00

    Scott,

    “I don’t believe we are at the zero bound. The term “bound” means the limit of zero is stopping rates from being lower. But the Fed doesn’t want lower rates, rates are right on target, ands likely to rise. (Ands of course the Europeans have shown that negative rates are possible, as Yellen has acknowledged.)”

    Maybe we aren’t currently at the zero lower bound at the moment, but that misses the point. What New Keynesian’s want to find out is “what do we do at the zero lower bound” so that we know what to do if the zero lower bound does bind. You are ignoring the zero lower bound in general by saying it’s not a problem now. You’re basically suggesting we don’t research it because it’s not currently a problem (I suppose New Keynesians did the same thing in the ’90s, but at least now they’ve realized that we should understand the zero lower bound). What do you think a central bank should do if the zero lower bound does bind? (and please don’t just answer that the zero lower bound is unimportant because we’re not at it anymore or that the zero lower bound is unimportant because we’d never hit it in an optimal monetary policy regime, neither are effective arguments for what policy should be at the zero lower bound, they just dodge the question.

    Maybe the zero lower bound is the wrong term. What I am talking about is the point at which monetary expansion becomes ineffective. This is the case whenever money demand is indeterminate which, based on the amount of excess reserves, I think is obviously the case.

    “Interest rates are a faulty indicator because (contrary to your claim) they are not reliably linked to expected NGDP growth, at least in the short run. If interest rates did equal expected NGDP growth, then we could peg expected NGDP growth merely by pegging interest rates.”

    So, then, what part of your model is inconsistent with basic utility maximization with perfect asset markets? Are asset markets imperfect, is the time preference rate endogenous?

    “The Taylor Rule would have called for even tighter monetary policy in 2008 than actual policy, and hence would have been a disaster. That’s why we need to replace the Taylor Rule with NGDPLT, combined with NGDP futures targeting. Each of those changes would have made it more likely that economy avoided the zero bound.”

    To what Taylor Rule are you referring? I was speaking generally about New Keynesian models in which Taylor Rules do approximate optimal policy. To be closer to reality, maybe someone needs to derive optimal policy in New Keynesian models with financial frictions, some asymmetrical pricing frictions (http://worthwhile.typepad.com/worthwhile_canadian_initi/2015/01/inflation-targeting-destroyed-its-own-signal.html), and/or some significant nominal wage rigidity. Maybe then we would come up with a monetary policy rule that is close to optimal in the real. Or, of course, we could simply assert that our preferred target is the best and call it a day.

  27. Gravatar of Scott Sumner Scott Sumner
    18. November 2015 at 18:11

    Patrick, Yes, even Hitler started small. But there are a whole lot of people who started small and stayed small. My hunch is that ISIS will be in that group.

    In any case, Krugman was arguing that our (or the French) reaction may cause more damage than the attack itself. Not that far-fetched, given American history over the past 15 years.

    John, Back in 2007 the conventional wisdom was that the zero bound was not a problem. Mishkin’s Money and Banking text (which is the best seller) said monetary policy remains “highly effective at the zero bound. That’s what we were teaching our students. Bernanke said zero rates didn’t stop the Japanese from inflating, they were just being stubborn. Krugman made the same complaint. Svensson talked about “foolproof” plans for reflation in a liquidity trap. I agreed with all of them.

    And then the profession suddenly changed its mind. But why? What new facts did we learn that suggested the 2007 consensus was wrong? No one have ever presented me with any such facts. All I get are supposed “facts” that just aren’t so, like the claim the Japanese tried QE and it failed. If all I’m hearing are false claims being made, on what basis should I shift my view away from the 2007 consensus? In 2007 I agreed with Mishkin’s claim that monetary policy is highly effective at the zero bound, and I still believe that. Mishkin doesn’t seem to believe it any more, but I have no idea why.

    You asked:

    “What do you think a central bank should do if the zero lower bound does bind?”

    My first choice is to peg the price of NGDP futures contracts.

    My second choice is buy enough Treasury bonds so that the central bank’s internal forecast of NGDP growth is on target. I would hope they have a sensible enough target so that they don’t have to buy a lot of bonds. But hey, there is no law against mass stupidity, and if the FOMC is so stupid that they set a target so low that it leads to zero interest rates EVEN WHEN THE MARKET EXPECTS THE TARGET TO BE HIT, then I can’t stop them. In that case they’d have to but a whole lotta bonds.

    Are you claiming the Fed might buy all the T-bonds (and MBSs) in existence, and still fail to hit their target?

    You said:

    “So, then, what part of your model is inconsistent with basic utility maximization with perfect asset markets? Are asset markets imperfect, is the time preference rate endogenous?”

    I don’t understand what perfect asset markets have to do with interest rates equaling expected nominal GDP growth. Are you confusing a particular model that features perfect asset markets, with all models that feature perfect asset markets? For instance, nominal interest rates are basically the same in all 50 states, but NGDP growth varies from state to state. There is no presumption in basic macro theory that nominal interest rates must equal expected NGDP growth. The level of RGDP relative to trend is also important.

    You said:

    “To what Taylor Rule are you referring? I was speaking generally about New Keynesian models in which Taylor Rules do approximate optimal policy”

    You keep talking about what NK models imply, but why does that matter? I agree that NK models imply that some sort of Taylor rule is optimal, but these models are flawed. I doubt even Woodford favors strict adherence to a particular Taylor Rule. If they are giving us clearly inaccurate policy advice, then we need to stop paying attention to the model, and get a new model. My suggestion would be to shift over to the Lars Svensson approach, target the forecast. You can’t argue that Svensson doesn’t understand NK models.

  28. Gravatar of Bob Murphy Bob Murphy
    18. November 2015 at 18:39

    Scott, just being honest here, when you and Nick Rowe talk about “concrete steppes” I have no idea what you’re talking about. I would have to go look up your earlier posts to see the analogy first laid out, to then know what the macro analog is.

    In contrast, I totally get the (alleged) Chuck Norris Effect.

  29. Gravatar of Ray Lopez Ray Lopez
    18. November 2015 at 19:34

    @Bob Murphy – concrete steppes is (partly) the post below, which combines two older Sumner posts and also referenced Nick Rowe, who came up with the term. It’s as clear as mud. Excepts below [MY COMMENTS IN CAPS]. BTW I have no idea what Chuck Norris effect is for this blog. Maybe: ‘Chuck Norris’ monetary policy is so strong he doesn’t need monetary policy, he just expects the economy to improve and it obeys’ (about as sound as Sumner’s logic).

    http://www.themoneyillusion.com/?p=13472

    Nick Rowe often talks about “The People of the Concrete Steppes;” commenters who don’t like explanations based on central banks steering expectations, and demand to know what “concrete steps” the central bank should take.

    In my view (and I think [SUMNER IS NOT SURE OF ROWE’S VIEWS? INTERESTING] Nick’s as well) money matters [NO IT DOES NOT, BERNANKE ET AL FAVAR 2003 PAPER FINDS EXTREMELY WEAK FED POLICY SHOCK EFFECTS. I KNOW YOU WILL ARGUE FED POLICY IS NOT NGDPLT BUT THAT’S ANOTHER MATTER. TO DATE THE FED DOES NOT MATTER.]; an increase in the money supply will raise nominal aggregates in the long run via something like the “hot potato effect.” [WHAT IS HOT POTATO EFFECT? IS THIS MONEY VELOCITY?]

    [AFTER ADOPTING THE DEFINITIONS OF TWO CRITICS OF MONETARISM, WHICH IS LUDICROUS IF YOU THINK ABOUT IT, AND ACKNOWLEDGED LATER IN THE POST AS INSANE BY SUMNER HIMSELF, AND SHOWING THE CRITIC’S DEFINITIONS WERE MET, SUMNER STATES:]

    Fine, so let’s call higher real rates a tight money policy in the eyes of the people of the concrete steppes. Did monetary policy become far tighter in the second half of 2008? [FIGURE 2 SHOWS REAL INTEREST RATES ROSE SHARPLY AFTER THE LEHMAN FAILURE OF OCT 2008. AMAZINGLY, THIS LEHMAN EVENT, EXOGENOUS TO THE FED, IS USED AS EVIDENCE THE *fed* TIGHTENED! ASTONISHINGLY BACKWARDS LOGIC!]

    [SUMNER THEN BLAMES THE FED FOR BELIEVING IN HIGH INFLATION IN THE SEP. 18, 2008 FOMC MEETING, BUT SEE THE ACTUAL MINUTES AND THE PICTURE IS NOT SO CLEAR, GOOGLE FOR THESE MINUTES (“Members agreed that keeping the federal funds rate unchanged at this meeting was appropriate. The current low [!!! DO YOU SEE THIS SUMNER? FED THINKS THE *real* FED RATE IS *low*. YOU DISAGREE? PROVE THEM WRONG WITH CONTEMPORANEOUS DATA TO SEPT. 2008.] real federal funds rate appeared necessary to provide adequate counterweight to the restraining effects of tight credit conditions and of continued declines in the housing market on spending and output…. [THE FED BELIEVED IT HAD EASED AND IT FURTHER HAD TO ACKNOWLEDGE THE INFLATION HAWKS LIKE LACKER OF RICHMOND FED WITH THIS FOLLOWING STATEMENT, WHICH SUMNER SEIZES UPON BUT CLEARLY IS NOT THE MAIN THRUST OF THE MINUTES] …However, the possibility that core inflation would not moderate as anticipated was still a significant concern. With substantial downside risks to growth and persisting upside risks to inflation, members judged that leaving the federal funds rate unchanged at this time suitably balanced the risks to the outlook. Some members emphasized that if intensifying financial strains led to a significant worsening of the growth outlook, a policy response could be required; however, such a response was not called for at this meeting. Indeed, it was noted that, with elevated inflation still a concern and growth expected to pick up next year if financial strains diminish, the Committee should also remain prepared to reverse the policy easing put in place over the past year in a timely fashion.”[AH, NO DOUBT SUMNER WILL INVOKE THE ‘EXPECTATIONS FAIRY’ AND USE THIS LAST SENTENCE AS EVIDENCE PEOPLE THOUGHT THE FED WOULD TIGHTEN LATER, NOT THAT IT MATTERS SINCE MONEY IS LARGELY NEUTRAL]

  30. Gravatar of John Handley John Handley
    18. November 2015 at 19:50

    Scott,

    “What new facts did we learn that suggested the 2007 consensus was wrong?”

    There was this thing that the US and the UK tried called QE and inflation barely budged. Surely there’s a reason that a 400% increase in the monetary base results in lower than average inflation. I call that ‘monetary policy ineffectiveness at the zero lower bound’ and I think I’ve explained my reasoning for that to you quite well in the past. If you’d like me to go over it again, I’m happy to.

    “My first choice is to peg the price of NGDP futures contracts. My second choice is buy enough Treasury bonds so that the central bank’s internal forecast of NGDP growth is on target.”

    So, do two things that most models where the zero lower bound is endogenous imply are improbable (in other words, there’s no way for the central bank to force them as an equilibrium, so some other actor – e.g. the household choosing one of the rational expectations equilibria subjectively or the fiscal authority forcing one of them – must require the desired result to occur).

    “I don’t understand what perfect asset markets have to do with interest rates equaling expected nominal GDP growth. Are you confusing a particular model that features perfect asset markets, with all models that feature perfect asset markets? For instance, nominal interest rates are basically the same in all 50 states, but NGDP growth varies from state to state. There is no presumption in basic macro theory that nominal interest rates must equal expected NGDP growth. The level of RGDP relative to trend is also important.”

    I am referring to the basic consumption Euler equation in every standard NK and RBC model which is derived from a combination of 1. utility maximization 2. perfect asset markets and 3. a constant exogenous time preference rate. I have no clue where you get the “level of RGDP relative to trend is also important” idea from, maybe it has to do with the altered version of the Euler equation in NK models that has the output gap instead of just consumption. Look at the rearranging of a standard consumption Euler equation that I did in my first comment for the reason that, if the assumptions of the vast majority of macro models are approximately correct, nominal interest rates should be roughly indicative of expected NGDP growth.

    “You keep talking about what NK models imply, but why does that matter? I agree that NK models imply that some sort of Taylor rule is optimal, but these models are flawed.”

    I agree. This is why I said: “to be closer to reality, maybe someone needs to derive optimal policy in New Keynesian models with financial frictions, some asymmetrical pricing frictions (http://worthwhile.typepad.com/worthwhile_canadian_initi/2015/01/inflation-targeting-destroyed-its-own-signal.html), and/or some significant nominal wage rigidity. Maybe then we would come up with a monetary policy rule that is close to optimal in the real. Or, of course, we could simply assert that our preferred target is the best and call it a day.” Or, more simply, I’m not inclined to believe your model is not flawed unless it can quantitatively match the data and I’m willing to add more frictions than just sticky prices to NK models in order to more closely approximate the real world (and determine a more ideal monetary policy). I find your assertion that an NGDPLT would be better than an inflation targeting regime highly dubious; the model that you provide makes NGDPLT optimal by assumption either explicitly because you suggest “employment fluctuations are driven by variations in the NGDP/Wage ratio” or because you only list nominal wage rigidity as a nominal friction which necessarily means that NGDPLT is better than an inflation target.

    “I doubt even Woodford favors strict adherence to a particular Taylor Rule.”

    Probably. I’m not sure I agree with Woodford.

    “If they are giving us clearly inaccurate policy advice, then we need to stop paying attention to the model, and get a new model.”

    Define “clearly inaccurate” please. Do you mean “clearly inaccurate when put through the lens of my subjective view of economic theory?” Maybe we do need a new model. Just realize that a model that is the same as NK except that it has a money demand function and has some nominal wage rigidity is still an NK model. In this way, you’re just a New Keynesian who doesn’t realize that he is. You’re probably a bit more reasonable than the average NK economist because you don’t think interest rates are a good measure of the stance of monetary policy, but, then again, you keep trying to shoe-horn NGDP into places that it doesn’t need to be.

  31. Gravatar of Derivs Derivs
    19. November 2015 at 01:14

    #3 seems kind of odd. Currency devaluation has been the equivalent of NGDP targeting for many countries this past year. What would a commodity producing countries NGDP have looked like this year had all those countries currencies not be devaluated so much?

  32. Gravatar of Jose Romeu Robazzi Jose Romeu Robazzi
    19. November 2015 at 04:25

    I still don’t get why people believe that the monetary authority cannot devalue a fiat currency, if it wishes so. The fact that the money demand function is unknown, and probably non-linear, should not prevent people from believing in marginalism…

  33. Gravatar of Ray Lopez Ray Lopez
    19. November 2015 at 05:20

    Readers note: John Handley savages Scott Sumner, check and mate. Never mind both of them are wrong in different ways, it’s fun to watch.

  34. Gravatar of TravisV TravisV
    19. November 2015 at 06:17

    Wooo!!

    http://www.zerohedge.com/news/2015-11-18/did-goldman-sachs-just-find-smoking-gun-todays-fomc-minutes

    “Did Goldman Sachs Just Find The Smoking Gun In Today’s FOMC Minutes?”

  35. Gravatar of TravisV TravisV
    19. November 2015 at 06:18

    Randall Forsyth in Barron’s: “Fed Setting Long-Range Goal of 0% Interest Rate”

    http://www.barrons.com/articles/fed-setting-long-range-goal-of-0-interest-rate-1447936520

  36. Gravatar of TravisV TravisV
    19. November 2015 at 06:58

    Bob Murphy,

    Re: “concrete steppes,” I highly recommend this post:

    http://www.themoneyillusion.com/?p=27741

    “Men with two feet on the ground”

  37. Gravatar of ssumner ssumner
    19. November 2015 at 08:45

    Ray, “It” is NGDP expectations, whereas “the economy” is actual RGDP. Couldn’t be more different.

    Bob, Concrete steps are things like changes in the fed funds target, or changes in the monetary base. Actions taken by the Fed that are clearly observable.

    John, You said:

    “There was this thing that the US and the UK tried called QE and inflation barely budged. Surely there’s a reason that a 400% increase in the monetary base results in lower than average inflation. I call that ‘monetary policy ineffectiveness at the zero lower bound’ and I think I’ve explained my reasoning for that to you quite well in the past. If you’d like me to go over it again, I’m happy to.”

    That doesn’t answer my question, because that’s not a “new fact”. We did QE in the Great Depression, even under Hoover. Japan had done QE well before 2007. All this was widely known to macroeconomists. How does the recent US experience change what we already knew in 2007? Friedman and Schwartz showed that the monetary base is not a useful indicator of the stance of monetary policy. And yet I see people point to QE as evidence that Fed policy has been expansionary. It makes me want to pull my hair out.

    You said:

    “So, do two things that most models where the zero lower bound is endogenous imply are improbable”

    The models are very incomplete. If we took them literally, then Japan could buy up all of planet Earth without creating a tiny bit of inflation, and live off the work of others. Literally no economist believes that is true. Thus the debate is an empirical one—how much would they have to buy? Would there be enough Treasuries, or would they have to buy riskier assets? That’s why I said the real issue is the target, not the QE. Is the target set so low that the central bank will need a big balance sheet to hit the target? Is the inflation target 0%, 2%, or 4%. Those are the real issues. The amount of QE you need is endogenous, once you pick a target. At higher expected inflation (or NGDP growth) people want to hold less base money.

    You said:

    “I am referring to the basic consumption Euler equation in every standard NK and RBC model which is derived from a combination of 1. utility maximization 2. perfect asset markets and 3. a constant exogenous time preference rate. I have no clue where you get the “level of RGDP relative to trend is also important” idea from, maybe it has to do with the altered version of the Euler equation in NK models that has the output gap instead of just consumption. Look at the rearranging of a standard consumption Euler equation that I did in my first comment for the reason that, if the assumptions of the vast majority of macro models are approximately correct, nominal interest rates should be roughly indicative of expected NGDP growth.”

    You and I look at things very differently. You look at models and assume they describe the actual world. I look at the actual world and see that models don’t describe the world accurately. As I pointed out, NGDP growth does not equal nominal interest rates. Are they positively correlated? Yes, I’ve said that 100 times in my blog. Indeed I’ve argued interest rates are better correlated with NGDP growth than inflation. But the correlation is far from perfect. And the correlation is nowhere near close enough to use for policy purposes.

    The level of RGDP relative to trend obviously impacts interest rates. When output is low, investment falls as a share of GDP, and interest rates are lower than otherwise. After 1933, RGDP growth average 8%, and rates were at zero due to the low LEVEL of output. We had 4% NGDP growth after 2008 and zero rates. You say that doesn’t show up in the models? OK, then why doesn’t someone fix the model? It’s obviously true, so why isn’t it incorporated in the model?

    The models you refer to are so simple, and leave out so much that is important, that they are virtually useless for policy evaluation purposes.

    You said:

    “Or, of course, we could simply assert that our preferred target is the best and call it a day.” Or, more simply, I’m not inclined to believe your model is not flawed unless it can quantitatively match the data and I’m willing to add more frictions than just sticky prices to NK models in order to more closely approximate the real world (and determine a more ideal monetary policy). I find your assertion that an NGDPLT would be better than an inflation targeting regime highly dubious; the model that you provide makes NGDPLT optimal by assumption either explicitly because you suggest “employment fluctuations are driven by variations in the NGDP/Wage ratio” or because you only list nominal wage rigidity as a nominal friction which necessarily means that NGDPLT is better than an inflation target.”

    You don’t seem to understand my role; I’m a blogger, not a model builder. I don’t even buy into the methodology of post-1970 macroeconomics, I’m a dinosaur. But I’m a dinosaur that has been pushing NGDP since the 1980s, and have seen in the past 5 years one big name economist after another switching from inflation targeting to NGDP targeting as a preferred policy option. I’ve seen Christy Romer endorse NGDPLT in the New York Times, and in her endorsement link to my paper advocating NGDPLT, saying it provided “the logic” for NGDP targeting. So while you may find my arguments to be unpersuasive (and I can understand why) lots of people find them persuasive. As long as I keep making progress, I see no reason to become something I am not.

    I have encouraged people to take my approach and turn it into a formal mathematical model. I know enough econ to know that it wouldn’t be hard for a younger economist who was just out of grad school. But no one has taken me up on the offer . . . yet.

    You said:

    “Define “clearly inaccurate” please. Do you mean “clearly inaccurate when put through the lens of my subjective view of economic theory?” ”

    Yup, and let me give a few reasons. I’ve been making these arguments for many years. At first I was laughed at. Now I see serious economists coming around to my view. I said money was clearly too tight in late 2008. In Bernanke’s memoir he concedes the Fed blew it by not cutting rates after Lehman failed. I’ve been describing the stance of monetary policy as being “tight” in the 2008-14 period, and now a formal research paper by Vasco Curdia makes the same point. But when I made this argument in late 2008 and early 2009 I was laughed at.

    I’ve been saying that the Fed has been overestimating the “normal” level of interest rates, and that they won’t be able to raise them as much as they assume. And the new Fed minutes (released yesterday) show that the Fed is now reaching that conclusion as well.

    In late 2012 I said the looming fiscal cliff would not slow growth. In calendar 2013 the budget deficit plunged by $500 billion. 350 Keynesian economists wrote a letter complaining that the austerity” might lead to recession. I said it would not slow growth. In early 2013 Krugman said 2013 would be a “test” of “market monetarism.” In fact, growth from 2012:4 to 2013:4 was higher than the previous year. The Keynesian model failed Krugman’s test. I said that eliminating the extended unemployment benefits in 2014 would speed up employment growth. In early 2014 Krugman (who disagreed) said 2014 would be a “test” of that hypothesis. Growth in employment in 2014 was higher than in 2010, 2011, 2012, 2013, or 2015. Again, Krugman’s Keynesian model fails.

    I said markets would react to news of QE as if it was expansionary. They did. I said the Japanese could depreciate the yen, even at the zero bound. Krugman pointed out that the NK model implied they could not depreciate the yen at the zero bound. The Abe government showed it could be done.

    I was the first to publish a paper mentioning the option of negative rates on reserves (in early 2009). Some claimed the policy was impractical and could never be done. Some claimed it would be contractionary. It now has been done in many countries, and the markets seem to react to news of negative IOR as if it is expansionary (i.e. the exchange rate falls on the news)

    These are just a few reasons to come to mind, if I dug through my 1000s of blog posts I’d find dozens of others.

    That’s how I work, I don’t rely on a single abstract model to figure out how the world works, I look at lots of simple (partial equilibrium) models, and try to put the pieces together. In other words, I use the same methodology as Milton Friedman. I’m a dinosaur.

    And I’m a blogger, don’t expect formal models.

  38. Gravatar of ssumner ssumner
    19. November 2015 at 08:48

    Derivs, I suppose NGDP would have been lower.

    Jose, They think cash and T-bonds are perfect substitutes.

    Thanks Travis.

  39. Gravatar of Jose Romeu Robazzi Jose Romeu Robazzi
    19. November 2015 at 09:42

    @Prof. Sumner
    Just ask the treasury to start selling only perpetual bonds. And do NGDP LT. Actually, if I were responsible for debt management at the treasury, that is what I would do.

  40. Gravatar of Don Geddis Don Geddis
    19. November 2015 at 09:58

    @Ray Lopez: “concrete steppes … referenced Nick Rowe, who came up with the term. It’s as clear as mud. … BTW I have no idea what Chuck Norris effect is for this blog.

    Rather than posting a long comment where you explicitly state your own ignorance, have you considered doing some research to try to learn, first?

    Nick Rowe has a lot of (great!) posts, but here is the concrete steppes post … where he also explains the Chuck Norris analogy.

    Your comments might be more productive, if you would attempt to learn and understand, prior to stating your strong opinions.

  41. Gravatar of TravisV TravisV
    19. November 2015 at 10:37

    Prof. Sumner,

    A couple issues with this blog lately:

    (1) Before I enter my comment, I have to fill out “Name (required)” and “Mail (will not be published) (required)” every single time. I didn’t have to do that before. My name and email address were always remembered.

    (2) There are now long delays in the “Recent Comments” section on the right side of this blog.

  42. Gravatar of Derivs Derivs
    19. November 2015 at 11:51

    “Derivs, I suppose NGDP would have been lower.”

    To your credit, it does appear from that example that the market does have an inherent desire to help maintain a stable NGDP.

  43. Gravatar of TravisV TravisV
    19. November 2015 at 13:15

    Prof. Sumner,

    I’m curious why the 2008 impact of Hurricane Ike on monetary policy is not discussed more by market monetarists.

    Please see your discussion of that here: http://www.themoneyillusion.com/?p=13472

    I’m surprised I haven’t seen that point made anywhere else.

  44. Gravatar of Patrick R. Sullivan Patrick R. Sullivan
    19. November 2015 at 17:46

    TravisV, I’m also having trouble commenting here. I’ve spent the better part of the day trying to figure out why. Some issues I’ve resolved, but I can’t get a comment to post if I give my gmail address, but can if I use another one.

  45. Gravatar of John Handley John Handley
    19. November 2015 at 17:50

    Scott,

    “That doesn’t answer my question, because that’s not a ‘new fact’.”

    So, to be clear, everyone in 2007 new that QE would be useless but believed nevertheless that monetary policy at the zero lower bound was highly effective? Maybe the consensus had something to do with forward guidance of some such policy (i.e. not monetary base expansion).

    “I have encouraged people to take my approach and turn it into a formal mathematical model. I know enough econ to know that it wouldn’t be hard for a younger economist who was just out of grad school. But no one has taken me up on the offer . . . yet.”

    Here you’re just being plain unscientific. You seem to want someone to make a model in which NGDPLT is optimal. This is completely backwards from a scientific perspective. You don’t start with the conclusion in science, you come to the conclusion by performing experiments (which, in the case of theoretical macroeconomics means using a set of assumptions about the way people behave to derive a model, determining its implications, and trying to match it with the available data). If all you want is a model that roughly matches the assumptions about people’s behavior that you seem to have (i.e. nominal wage rigidity, interest elastic money demand), then even I could write a model for you (honestly, all you need is some basic calculus and a rough understanding of DSGE to write down a model). Regardless, I’m sure there are un-gated papers that feature nominal wage rigidity and money demand in their models.

    “The level of RGDP relative to trend obviously impacts interest rates. When output is low, investment falls as a share of GDP, and interest rates are lower than otherwise. After 1933, RGDP growth average 8%, and rates were at zero due to the low LEVEL of output. We had 4% NGDP growth after 2008 and zero rates. You say that doesn’t show up in the models? OK, then why doesn’t someone fix the model? It’s obviously true, so why isn’t it incorporated in the model?”

    This is not a result of the fundamentals of interest rates, just the policy decisions that the Fed happens to make.

    “In late 2012 I said the looming fiscal cliff would not slow growth. In calendar 2013 the budget deficit plunged by $500 billion. 350 Keynesian economists wrote a letter complaining that the austerity” might lead to recession. I said it would not slow growth. In early 2013 Krugman said 2013 would be a “test” of “market monetarism.” In fact, growth from 2012:4 to 2013:4 was higher than the previous year. The Keynesian model failed Krugman’s test.”

    What you don’t realize is that the US had been contracting fiscal policy for a couple years by 2013. Look at the level of federal spending since 2008 and you’ll notice that 2013 was just a continuation of the same austerity. Read Simon Wren Lewis and Jason Smith for a much better explanation of this.

    “I said that eliminating the extended unemployment benefits in 2014 would speed up employment growth. In early 2014 Krugman (who disagreed) said 2014 would be a “test” of that hypothesis. Growth in employment in 2014 was higher than in 2010, 2011, 2012, 2013, or 2015. Again, Krugman’s Keynesian model fails.”

    The idea that UI reduces unemployment is certainly not something that comes out of any model I can think of. If you pay people not to work (whether you are using something like the DMP model or the plain neoclassical labor market), then less people will work. UI does increase utility, though. Perhaps you’re confusing Krugman’s opinions with the implications of New Keynesian models.

    “At higher expected inflation (or NGDP growth) people want to hold less base money.”

    But, given rational expectations and a non-MIUF money demand model, you can’t increase expected inflation by increasing the monetary base at the zero lower bound. Maybe better monetary models than cash-in-advance imply a different result. Perhaps someone should ask David Andolfatto or Stephen Williamson.

    “Friedman and Schwartz showed that the monetary base is not a useful indicator of the stance of monetary policy. And yet I see people point to QE as evidence that Fed policy has been expansionary.”

    So the only thing the Fed actually has control over is not a good indicator of the stance of monetary policy, you say? So the fact that increases in the monetary base seem to only be irrelevant at the zero lower bound is irrelevant?

  46. Gravatar of Ray Lopez Ray Lopez
    19. November 2015 at 19:31

    Sumner: “I have encouraged people to take my approach and turn it into a formal mathematical model. I know enough econ to know that it wouldn’t be hard for a younger economist who was just out of grad school. But no one has taken me up on the offer . . . yet. “- why is that? If it’s ‘not hard’ and you’re ‘almost famous’, then surely some young aspiring PhD would gladly turn your mushy views into a hard math model? If it was so easy…but it’s not easy to turn 0/0 logic into something tractible in math, L’Hopital’s Rule notwithstanding.

    Sumner: “I said money was clearly too tight in late 2008. In Bernanke’s memoir he concedes the Fed blew it by not cutting rates after Lehman failed.” – but this is bogus hindsight. The Sep 2008 FOMC minutes, which I posted a link to the other day, shows the Fed thought at the time they were being expansive with their low rates. Perhaps, as you claim, a NGDP futures market would have helped dispell that view, but it’s a view they had at the time. Monday morning ‘coulda, woulda, shoulda’ by Bernanke is just an old man’s regret, nothing more substantive.

    Sumner: “These are just a few reasons to come to mind, if I dug through my 1000s of blog posts I’d find dozens of others” – I would say that despite your hide-bound ideology, which you use to promote your NGDPLT since that’s your baby, is also your greatest strength, since apparently you don’t believe in much moderation of your blog, unlike most others (Krugman, Econlog, Brad DeLong historically, to name a few). You and Tyler Cowen have liberal blog moderation policies.

    @Don Gheddis – thanks, I’ve seen this blog post but I could not find it on Google when I posted before. As for strong, wrong, opinions, look in the mirror.

  47. Gravatar of Don Geddis Don Geddis
    19. November 2015 at 20:29

    @John Handley: “You don’t start with the conclusion in science, you come to the conclusion by performing experiments

    Fair enough.

    which, in the case of theoretical macroeconomics means using a set of assumptions about the way people behave to derive a model, determining its implications, and trying to match it with the available data

    No, that’s not the only way to do “macro science”. And perhaps not even the best way.

    An alternative approach is to start with the data, notice regularities, form hypotheses about them, and only at the end formalize the theory as a mathematical model, and only for the purpose of “double checking” the necessary assumptions.

    Nothing in science requires you to start the process with microfoundations about individual economic behavior. In fact, you may observe a regularity in the aggregate data, whose microfoundations you don’t understand. But that doesn’t stop you from using that observation as evidence to form further conclusions.

  48. Gravatar of Don Geddis Don Geddis
    19. November 2015 at 20:34

    @Ray Lopez: “As for strong, wrong, opinions, look in the mirror.

    Certainly, I may be wrong. But what I like about you, is that your ignorance is self-admitted: “concrete steppes … It’s as clear as mud. … I have no idea what Chuck Norris effect is

    It’s a rare talent to not allow the shame of ignorance to in any way impact the arrogance of pontificating. My hat is off to you, sir.

  49. Gravatar of John Handley John Handley
    19. November 2015 at 21:26

    Don,

    “An alternative approach is to start with the data, notice regularities, form hypotheses about them, and only at the end formalize the theory as a mathematical model, and only for the purpose of “double checking” the necessary assumptions.”

    The only problem I have with this is that, by making the model to fit the data, you might come up with the wrong reasons for the “regularities” in the data. Further, I was talking specifically about theoretical macro. You could certainly start with the data and estimate a VAR, for example, but I still think that you should observe the micro data, add the observed microfoundations to your model, and then derive the result and see if it lines up with the aggregate reality.

    “In fact, you may observe a regularity in the aggregate data, whose microfoundations you don’t understand. But that doesn’t stop you from using that observation as evidence to form further conclusions.”

    I’m worried about people observing a regularity and concluding it happens for the wrong reasons. For example, Sumner sees that NGDP falls during recessions and concludes that the recession happened because of the fall in NGDP (maybe I’ve made a straw man, feel free to disagree with my summary of your views, Scott). There really is no way to prove that the fall in NGDP caused the recession without some underlying theory. If you start with the data, you may impose your poorly thought out theory to explain it.

  50. Gravatar of Jose Romeu Robazzi Jose Romeu Robazzi
    20. November 2015 at 02:52

    @John Handley: “You don’t start with the conclusion in science, you come to the conclusion by performing experiments”

    Keynesians start with the only solution they accept, deficit spending, and work their way backwards.

    Also, you said: “Sumner sees that NGDP falls during recessions and concludes that the recession happened because of the fall in NGDP (maybe I’ve made a straw man, feel free to disagree with my summary of your views, Scott). There really is no way to prove that the fall in NGDP caused the recession without some underlying theory.”

    The underlying theory is that NGDP is a nominal quantity, therefore it is linked to the quantity of base money. Therefore, if you print enough money, you probably get NGDP up, the relationship may be ustable and non-linear, but it is there. Maybe you get just more inflation by doing that, but you can get NGDP up. And no, money and T-bonds are not perfect substitutes. You can’t by ice cream with a fraction of a perpetual T-bond. Under a simple marginalist model, one can always devalue a fiat currency. You may believe this model is wrong, but you can’t claim there is no underlying model.

  51. Gravatar of R.McGeddon R.McGeddon
    20. November 2015 at 03:47

    John Handley:

    “There was this thing that the US and the UK tried called QE and inflation barely budged. Surely there’s a reason that a 400% increase in the monetary base results in lower than average inflation. I call that ‘monetary policy ineffectiveness at the zero lower bound’”

    The purpose of QE wasn’t to cause inflation, it was to prevent deflation – which it did successfully.

    In the 1930s – the last time interest rates hit a ‘lower bound’ – inflation went down to -10%. Raising the inflation rateby abot 10% is hardly ineffective.

  52. Gravatar of Patrick R. Sullivan Patrick R. Sullivan
    20. November 2015 at 07:22

    ‘Krugman was arguing that our (or the French) reaction may cause more damage than the attack itself. Not that far-fetched, given American history over the past 15 years.’

    Actually, Krugman is asserting that;

    ‘France is not going to be conquered by ISIS, now or ever. Destroy Western civilization? Not a chance.’

    I think American (and world) history going back a little further than the last fifteen years, says something different. Judging by what the French are doing right now, so do they.

  53. Gravatar of ssumner ssumner
    20. November 2015 at 13:02

    Travis and Patrick, We are working on it.

    John, You said:

    “So, to be clear, everyone in 2007 new that QE would be useless but believed nevertheless that monetary policy at the zero lower bound was highly effective?”

    Of course not, I never said QE was useless. The markets respond to QE announcements as if it is (mildly) effective. It was well understood that QE is only effective to the extent it is viewed as being permanent, or at least more effective that way. Notice that QE in Japan began working in 2013, precisely when they raised their inflation target to 2%.

    My point was different, that these earlier examples showed that at the zero bound it is possible to have large increases in the base without much change in the price level. We all knew that already in 2007. So what have we learned since 2007? I’d say nothing—there is no reason to change the conventional wisdom, circa 2007.

    You said:

    “Here you’re just being plain unscientific. You seem to want someone to make a model in which NGDPLT is optimal. This is completely backwards from a scientific perspective. You don’t start with the conclusion in science, you come to the conclusion by performing experiments (which, in the case of theoretical macroeconomics means using a set of assumptions about the way people behave to derive a model, determining its implications, and trying to match it with the available data).”

    I have a very different view. We don’t write down models to discover how the world works, but rather to illustrate our policy intuitions. The models leave out far too much to establish the optimality of any particular policy with any degree of confidence. Instead we reach our policy conclusions through other means, and then use models to illustrate the idea to others. BTW, I don’t actually think NGDPLT is optimal, just better than IT. I’m sure there are further iterations that would be even better than NGDPLT, but you’d never get there by exploring the world with equations.

    As you know lots of mainstream economists (like Woodford) have recently switched over to advocating NGDPLT—they did not do so because their model told them it was optimal.

    You said:

    “This is not a result of the fundamentals of interest rates, just the policy decisions that the Fed happens to make.”

    Of all the things you’ve said, this one I disagree with most strongly. I assure you that the Fed did nothing to cause rates to be close to zero in the 1930s (a period I’ve studied in depth), unless you mean tight money. And in recent years the policy rate has been above the Wicksellian equilibrium rate, so your comment is 180 degrees wrong. The Fed has been holding rates ABOVE equilibrium, with policies like IOR. Check out Vasco Curdia’s recent paper on how policy has been tight since 2008. He’s very “scientific.” (A bogus concept in my view, but that’s another debate.)

    Interest rates over any extended period of time are determined by the economy, not the Fed. The Fed can nudge them a bit higher or lower with the liquidity effect, but their discretion is limited by the need to hit their inflation target. When the ECB tried to raise rates in 2011 they failed spectacularly. Are rates in the eurozone held down by the ECB? Obviously not. So why claim the Fed was holding down rates?

    You said:

    “What you don’t realize is that the US had been contracting fiscal policy for a couple years by 2013. Look at the level of federal spending since 2008 and you’ll notice that 2013 was just a continuation of the same austerity. Read Simon Wren Lewis and Jason Smith for a much better explanation of this.”

    Yes, I notice that Keynesians love to play this shell game. Taxes matter, unless they don’t matter. Then only spending matters. The deficit matters, unless it doesn’t matter, then only spending matters. Austerity produced by tax increases is a huge problem, until it isn’t a problem, and then only government spending matters. Of course most government spending is negative taxes. Is there a model there? No. Are the Keynesians being intellectually honest? No. They change the goal posts whenever things don’t go the way they predicted.

    And by the way, how much of the 2009 “stimulus” that supposedly saved the economy was actually an increase in “G”, i.e, government output? Very little. There simply weren’t many shovel ready projects ready. So do taxes and transfers matter? Keynesians can’t seem to make up their minds.

    If we had gone into recession in 2013, would Wren-Lewis and Krugman and Smith have said “Oops, our mistake, there was no austerity after all, I guess the recession was not caused by the austerity that Keynesians predicted would cause a recession.” I’d be very interested in your answer to this question. Not what you’d say, but what you think most Keynesians would have said.

    And ask yourself this, how much confidence can we have in an ideology that requires the government to manage fiscal policy to stabilize the economy, if the so-called experts can’t even figure out whether austerity occurred or didn’t occur (and it did occur in 2013, BTW) until it is far too late to do any good? The budget deficit fell by $500 billion in calendar 2013, far more than in any other year.

    Regarding what comes out of NK models, that’s been pretty much abandoned by Keynesian mainstream economists, who have moved sharply to the left since 2008, and not just on fiscal policy, but on UI, on the minimum wage, on the taxation of capital income, and a host of other issues. They even see a permanent tradeoff between inflation and unemployment—that’s also not in the NK model. So I’m talking about real world Keynesians like Krugman and Summers, not the model.

    You said:

    “So the only thing the Fed actually has control over is not a good indicator of the stance of monetary policy, you say?”

    I say???? Do you deny that 99% of economists agree with me? The monetary base had been growing at about 5% in the 6 years up to July 2007, then completely stopped growing for 9 months. In an accounting sense (and I emphasize accounting, not causal) that pushed us into recession. How many economists commented on the Fed’s “tight money” policy in late 2007 and early 2008? My guess is approximately zero. Since 1982, economists even stopped using M1 and M2 as policy indicators, except a few lonely monetarists. The bulk of economists NEVER considered the base to be an indicator of easy or tight money, until QE. So no, it’s not just me, it was 99% of the profession in 2007 who thought the base was not an indicator of easy or tight money. The base rose sharply in the early 1930s—can you find a single textbook that says Fed policy was expansionary in the early 1930s? The profession is acting incredibly foolish in suddenly pointing to the base, and doesn’t seem to even realize it. They are using metrics for the stance of monetary policy with zero justification.

    I agree that the base is what the central bank controls most directly, but they have effective control over other variables, such exchange rates and TIPS spreads, if they choose to exercise such control. Most economists look at interest rates, but I think that’s a mistake.

    One other point, Take a look at the path of NGDP in Europe and the US after 2008. Very similar at first, and then a divergence after 2011, a huge divergence. Fiscal policy was quite similar in both places (indeed a bit more austerity in the US), it was all monetary. The monetary policy was less contractionary in the US after 2010, and NGDP growth was far higher. Unemployment went from being about the same in 2010, to being twice as high in the eurozone. So it does make a difference, even at the zero bound. And again, the Japanese GDP deflator has been rising at almost 2% a year since 2013, after falling steadily for almost 20 years. How’d that happen at a time the fiscal policy in Japan was contractionary? It was QE at the zero bound–combined with a more sensible inflation target. QE is not a cure-all, it needs to be combined with a sensible target. But it’s also not nothing.

  54. Gravatar of ssumner ssumner
    20. November 2015 at 13:03

    Ray sees a commenter who actually knows what he’s talking about (John Handley), and tries to imitate him. Cute.

  55. Gravatar of John Handley John Handley
    20. November 2015 at 18:46

    R.McGeddon,

    “The purpose of QE wasn’t to cause inflation, it was to prevent deflation – which it did successfully.

    In the 1930s – the last time interest rates hit a ‘lower bound’ – inflation went down to -10%. Raising the inflation [rate by about] 10% is hardly ineffective.”

    Oh, so your counterfactual is a whole bunch of deflation. The only problem I see there is that the ECB didn’t do QE until recently and there wasn’t a massive deflation in the EU.

    Jose,

    “Keynesians start with the only solution they accept, deficit spending, and work their way backwards.”

    I suppose I’d call myself a Keynesian, and I don’t do that.

    “Under a simple marginalist model, one can always devalue a fiat currency.”

    No. Read my comments on this (http://www.themoneyillusion.com/?p=31151) post and read this (http://ramblingsofanamateureconomist.blogspot.com/2015/11/monetary-policy-effectiveness-in.html) for why. Do you have a different model? If so, lay it out specifically please. Or, at least, link me a paper that comes to the same conclusion as you.

    “The underlying theory…”

    This exactly what I don’t want to exist. The theory should come from observations about micro data. Actually, one of Scott’s assumptions does pretty well in this regard. I can’t site the specific paper, but I’ve read several times about the apparent length between wage changes which is even more significant than the observed length between price changes. It’s where we get to the monetary part of the model that I’m wary of trusting either of you.

    Scott,

    “Yes, I notice that Keynesians love to play this shell game. Taxes matter, unless they don’t matter. Then only spending matters. The deficit matters, unless it doesn’t matter, then only spending matters. Austerity produced by tax increases is a huge problem, until it isn’t a problem, and then only government spending matters. Of course most government spending is negative taxes. Is there a model there? No. Are the Keynesians being intellectually honest? No. They change the goal posts whenever things don’t go the way they predicted. Regarding what comes out of NK models, that’s been pretty much abandoned by Keynesian mainstream economists, who have moved sharply to the left since 2008, and not just on fiscal policy, but on UI, on the minimum wage, on the taxation of capital income, and a host of other issues. They even see a permanent tradeoff between inflation and unemployment—that’s also not in the NK model. So I’m talking about real world Keynesians like Krugman and Summers, not the model.”

    Unfortunately, I increasingly feel that I have no friends in the political arena. Krugman, Delong, Thoma, etc. are certainly way too far left for my taste. I certainly get the idea from their writings that they’ve switched to “see[ing] a permanent tradeoff between inflation and unemployment” and I’m more than a bit annoyed that they don’t pay attention to any economic theory that I can discern on “UI, on the minimum wage, on the taxation of capital income, and a host of other issues.” I wasn’t involved in the econoblogosphere back in 2008 (considering I was only 8 years old), so I can’t really say if Krugman et al have moved to the left since then, but it certainly seems that way based on the difference between the implications of mainstream macro models and their opinions. To be perfectly honest, I’m more inclined to agree with you on a lot of issues than with them (e.g., I think the minimum wage should be abolished and capital taxes should be as low as possible). There really are the two exceptions of the zero lower bound and NGDPLT (which I’m not wholly opposed to).

    “Fiscal policy was quite similar in both places (indeed a bit more austerity in the US)”

    I’d have to disagree. The size of the deficit may have decreased at about the same rate, but this doesn’t account for the contractionary effects of austerity (given a multiplier of perhaps 0.5, which is implied by normal specifications of neoclassical models – see the first couple of pages of http://www.columbia.edu/~mw2230/G_ASSA.pdf). For a rough measure of the actual change in ‘fiscal stance,’ I’d probably multiply the deficit by one minus the IMF’s measure of the output gap.

    Also, generally, I’m open to the idea that monetary policy can be ‘tight’ even at low interest rates. This is why right now I’d like central banks to follow something like a McCallum rule but I’m really not too strongly attached to this either, it may turn out that money demand, when interest rates are not equal to IOR, is not a simple positive function of output and a negative function of interest rates in which case a McCallum rule would be less than ideal.

    “I say???? Do you deny that 99% of economists agree with me? The monetary base had been growing at about 5% in the 6 years up to July 2007, then completely stopped growing for 9 months. In an accounting sense (and I emphasize accounting, not causal) that pushed us into recession. How many economists commented on the Fed’s “tight money” policy in late 2007 and early 2008? My guess is approximately zero. Since 1982, economists even stopped using M1 and M2 as policy indicators, except a few lonely monetarists. The bulk of economists NEVER considered the base to be an indicator of easy or tight money, until QE. So no, it’s not just me, it was 99% of the profession in 2007 who thought the base was not an indicator of easy or tight money. The base rose sharply in the early 1930s—can you find a single textbook that says Fed policy was expansionary in the early 1930s? The profession is acting incredibly foolish in suddenly pointing to the base, and doesn’t seem to even realize it. They are using metrics for the stance of monetary policy with zero justification.”

    The problem is that, at least to me, it appears that what you (and I suppose everyone else) is saying implies that the stance of monetary policy is defined as something that central banks can’t completely control. Before you suggest that central banks can always control NGDP, keep note that, by ‘completely control,’ I mean peg to whatever value is intended. The Fed certainly couldn’t declare that NGDP will be $500 Trillion next year and actually have that happen.

    “a commenter who actually knows what he’s talking about (John Handley)”

    Thanks!

  56. Gravatar of Jose Romeu Robazzi Jose Romeu Robazzi
    21. November 2015 at 05:04

    @ John handley
    Thanks for the comments. Certainly intelectually honest keynesians don’t start with deficit spending, but to be very frank, in the end, most keynesians tweak data in order to get just that, more deficit spending as the only sensible economic policy.

    On models: a bunch of equations do not make a model (per se), and one can have a model withouht a single (explicit) equation. The reason I believe in the marginalist model is because I see it happening all the time. No matter how “low” the price of something is, if someone big wants to sell it, it will go lower. If I had to lay down a model for the money demand, I would say it is upward rising with base money (no price level effect) until it reaches an inflection point (“saturation point”) where it either stay constant or starts falling, so that more money printing will increase the price level. I don’t know where that point is, if at 6 trillion dollars in the base, or 10 trillion, or 4.5 trillion. But i can believe it exists, sure, if it didn’t, we would had never seen hyperinflation in the past.

    But if you don’t like this model, forget about it. Pick your own model from your blog post. With your own set of equations, I believe that if the monetary authority commits to open ended QE, it will at some point change household expectations for inflation, and there you go.

  57. Gravatar of John Handley John Handley
    21. November 2015 at 10:32

    Jose,

    “we would had never seen hyperinflation in the past”

    No. Hyperinflation is a possibility in every mainstream monetary model (except for New Keynesian models, where it is arbitrarily banned as an equilibrium. Read http://www.patrickminford.net/wp/E2009_21.pdf for an explanation).

    “marginalist model”

    This term is so vague that it implies no meaning whatsoever. You’ve essentially just explained an intuition that you have to me. You assume that, at some point, more money at the zero lower bound means a higher price level. For me to take this claim seriously, it needs to be a result of your model, not something that you just presume to be correct.

    “I believe that if the monetary authority commits to open ended QE, it will at some point change household expectations for inflation, and there you go.”

    That is one equilibrium. The problem is that this equilibrium is not necessarily going to unfold if QE happens. The household can rationally expect any path of the price level, so the Japanese result of 20 years of slightly negative inflation is completely possible, and so is a whole bunch of inflation from QE. It seems, though, that households across the world have chosen the first equilibrium when presented with the scenario in the real world.

    “Pick your own model”

    I can’t shake the feeling that you think my model is non-standard in some way. I assure that, though an oversimplification, Cash-in-advance is a go to model of money demand in the literature. There are also models where money is held because it reduces transaction costs, because it makes the agents happy, because it’s more liquid than other assets, etc. Most of those models either make it impossible for the zero lower bound to happen for a finite money supply (MIUF) or make monetary injections irrelevant at the zero lower bound because money and other assets are perfect substitutes.

    Scott,

    I recently read a paper from 2002 by David Andolfatto called “Monetary Implications of the Hayashi-Prescott Hypothesis for Japan” (link: http://econwpa.repec.org/eps/mac/papers/0307/0307008.pdf). It basically explains my position on liquidity traps. And, as a bonus, it’s not Krugman’s 1998 paper.

  58. Gravatar of Don Geddis Don Geddis
    21. November 2015 at 15:41

    @John Handley: “The Fed certainly couldn’t declare that NGDP will be $500 Trillion next year and actually have that happen.

    That’s an interesting claim. My intuition is the opposite. The nominal price level appears to be just arbitrary numbers, and, as such, could be set pretty much anywhere. (E.g., consider government actions such as overnight peso devaluations.)

    Can you explain in more detail why you don’t believe the Fed could achieve a $500T NGDP next year? (Presumably, we’re assuming that it’s a target that is claimed to have some benefit for the economy; otherwise you would be talking about political roadblocks, not an economic barrier.)

  59. Gravatar of John Handley John Handley
    21. November 2015 at 16:49

    Don,

    In order for NGDP to be $500T next year, the entire path of expected NGDP would have to increase relative to current GDP. This would necessitate a massive increase in nominal interest rates that the Fed cannot (currently) do without a massive monetary tightening unless it somehow induces banks to care about the reserves that they are holding once again. This could probably be done with sufficiently negative IOR, but then there would probably just be a flight to cash. The Fed has a finite ability to make cash unappealing to agents and this is why liquidity traps can be so dangerous.

    A simply practical problem is that there is literally no way the Fed would consciously decide to raise the Fed Funds rate to the level that would be required for this to happen. The FOMC would be committed to low interest rates right now regardless of whether or not the higher interest rates would take the form of a higher NGDP path. In order to even think about this kind of policy, the Fed would need to stop thinking about bond yields, cut IOR as low as possible without causing a flight to cash (hoping that this would be low enough for reserves to be sufficiently less attractive than bonds), and start increasing the monetary base as fast as possible. Ideally, this would provide the necessary fall in demand for real balances and the necessary increase in the supply of nominal money for the price level (and hence NGDP) to increase to the desired level.

    Alternatively, the Treasury could join with the Fed and say that it would sell as many bonds as were demanded at the target interest rate while the Fed doesn’t bother with IOR and just starts increasing the money supply. In fact, this would have a higher chance of being effective than the first option because it doesn’t rely on their being pretty high costs of carrying physical cash. The only problem here is that the Treasury needs to be involved, so the Fed couldn’t do this without the support of the government, something that is unlikely given the way the GOP (and the public) thinks about debt and deficits. It’s unfortunate when the public and the government are both economically illiterate.

    If we weren’t in a liquidity trap, then I guess it would be possible for the Fed to cause NGDP to be as high as it wanted as long as it didn’t try to do so by setting interest rates at a low level. The problem with interest rates is that they reflect spending being substituted between the future and the present. It is possible to substitute spending into the present from the future by lowering interest rates, which could be considered loosening monetary policy, but it is also possible to cause the long run level of spending to fall by lowering interest rates, which could be considered tightening monetary policy. This is why it’s better to measure the stance of monetary policy with something that is directly tied to current spending; i.e. some money aggregate. A higher level of the chosen money aggregate should always be consistent with higher spending, regardless of the interest rate. This is not to say that we should target the growth of a money aggregate, just that we should not be too obsessed with the level of interest rates. Perhaps we should even commit to setting them according to a rule that the Fed could deviate from if it wanted to stimulate demand (actual interest rate rate implied by policy rule). The idea is that the rule would pin down expectations of future monetary policy in such a way that we would never force ourselves into a low-spending-growth-equilibrium.

    Or, we could let the Treasury control interest rates and leave the Fed with just the monetary base.

  60. Gravatar of ssumner ssumner
    22. November 2015 at 10:25

    John, I must say I am impressed that you are only 15 years old, you are more knowledgeable than I was when I graduated from college. Maybe you’re the next Evan Soltas. Sorry if I lumped you in with the Keynesians who annoy me.

    I don’t agree with you on the fiscal austerity being worse in Europe, although I’ll have to admit that I’m relying on the work of others, like Mark Sadowski. I have not researched the issue myself, but Sadowski said the US has done a bit more austerity in recent years, using the cyclically adjusted budget deficit.

    I am current working on a paper discussing what is meant by the term “stance” of monetary policy. It is a very ambiguous concept, which really needs to be pinned down. I understand your comments about controllability, although I would insist that NGDP futures prices are almost completely controllable, if the Fed chose to do so, just as much as interest rates or exchange rates, and more than M2. But I understand that we don’t currently have a highly liquid futures market in NGDP. TIPS spreads might be the lesser of evils at this point.

    You said:

    “No. Hyperinflation is a possibility in every mainstream monetary model (except for New Keynesian models, where it is arbitrarily banned as an equilibrium.”

    This is an interesting point. My concept of monetary theory came from the analysis of high inflation episodes like Brazil and Argentina, where inflation averaged about 75%/year from the 1950s to 1990. I view hyperinflation as the core topic in monetary economics, and everything else as secondary. Thus you start by modeling inflation, and other nominal variables like NGDP. That’s the core of monetary economics. Then when that’s done you add sticky wages, and now nominal shocks can have real effects. It seems like NKs go right from monetary policy (now defined as i, not M) to output, and then tack on an assumption that inflation occurs as the output gap goes away. I think my approach is better able to explain the actual inflation rates in Argentina and Brazil, why it averaged 75% and not 65% or 85%.

  61. Gravatar of ssumner ssumner
    22. November 2015 at 10:48

    John, Haven’t read the entire paper yet, but here’s a quick reaction. They say:

    “My findings can be summarized as follows. A productivity slowdown in the early 1990s combined with tightening monetary policy (which I argue was the case in Japan at this time) is consistent with a contraction in capital expenditure, a decline in then nominal (and real) interest rate, declining inflation, and a steady decline in the money multiplier. It is shown that these shocks may very well drive the nominal interest rate to zero, which I assume happened sometime in late 1995. In the event of a zero nominal interest rate, the model predicts that sales of government securities by the BoJ can have no effect on anything real or even on the price level; the economy at this stage appears to be in a ‘liquidity trap.’”

    My view is that slower productivity (and population) growth explains most of what’s happened in Japan, in real terms. I agree with the above quotation, except the final sentence. I believe that the Abe government has shown that QE “works” by sharply reducing the foreign exchange value of the yen, and boosting inflation and NGDP growth above zero. And this occurred even though working age population growth slowed sharply right as they took power.

    Although I don’t entirely agree about QE, I do agree that QE has had far far less impact than naive monetarist models would predict.

Leave a Reply