1946

Over at Econlog I did a post discussing the austerity of 1946.  The Federal deficit swung from over 20% of GDP during fiscal 1945 (mid-1944 to mid-1945) to an outright surplus in fiscal 1947.  Policy doesn’t get much more austere than that! Even worse, the austerity was a reduction in government output, which Keynesians view as the most potent part of the fiscal mix.  I pointed out that employment did fine, with the unemployment rate fluctuating between 3% and 5% during 1946, 1947 and 1948, even as Keynesian economists had predicted a rise in unemployment to 25% or even 35%—i.e. worse than the low point of the Great Depression.  That’s a pretty big miss in your forecast, and made me wonder about the validity of the model they used.

One commenter pointed out that RGDP fell by over 12% between 1945 and 1946, and that lots of women left the labor force after WWII.  So does a shrinking labor force explain the disconnect between unemployment and GDP?  As far as I can tell it does not, which surprised even me.  But the data is patchy, so please offer suggestions as to how I could do better.

Let’s start with hours worked per week, the data that is most supportive of the Keynesian view:

screen-shot-2016-09-24-at-9-32-18-am

Weekly hours worked dropped about 5% between 1945 and 1946. Does that help explain the huge drop in GDP?  Not as much as you’d think. Here’s the civilian labor force:

screen-shot-2016-09-24-at-9-34-38-am

So the labor force grew by close to 9%, indicating that the labor force in terms of numbers of worker hours probably grew.  Indeed if you add in the 3% jump in the unemployment rate, it appears as if the total number of hours worked was little changed between 1945 and 1946 (9% – 5% – 3%).  Which is really weird given that RGDP fell by 22% from the 1945Q1 peak to the 1947Q1 trough–a decline closer to the 36% decline during the Great Contraction, than the 3% fall during the Great Recession.

That’s all accounting, which is interesting, but it doesn’t really tell us what caused the employment miracle.  I’d like to point to NGDP, which did grow very rapidly between 1946 and 1948, but even that doesn’t quite help, as it fell by about 10% between early 1945 and early 1946.

Here’s why I think that the NGDP (musical chairs) model did not work this time. Let’s go back to the hours worked, and think about why they were roughly unchanged.  You had two big factors pushing hard, but in opposite directions. Hours worked were pushed up by 10 million soldiers suddenly entering the workforce.  In the offsetting direction were three factors.  A smaller number of (mostly women) workers leaving the workforce, unemployment rising from 1% to 4%, and average weekly hours falling by about 5%.  All that netted out to roughly zero change in hours worked.

So why did RGDP fall so sharply?  Keep in mind that while those soldiers were fighting WWII, their pay was a part of GDP. They helped make the “G” part of GDP rise to extraordinary levels in the early 1940s.  But when the war ended, that military pay stopped.  Many then got jobs in the civilian economy.  Now they were counted as part of hours worked. (Soldiers aren’t counted as workers.)  That artificially depressed productivity.

It’s also worth noting that real hourly wages fell by nearly 10% between February 1945 and November 1946:

screen-shot-2016-09-24-at-9-30-12-am

This data only applies to manufacturing workers. But keep in mind that the 1940s was the peak period of unionism, so I’d guess service workers did even worse.  So my theory is that the sudden drop in NGDP in 1946 was an artifact of the end of massive military spending, and the strong growth in NGDP during 1946-48, which reflected high inflation, helped to stabilize the labor market.  When the inflation ended in 1949, real wages rose and we had a brief recession.  By 1950, the economy was recovering, even before the Korean War broke out in late June.

Obviously 1946 was an unusual year, and it’s hard to draw any policy lessons.  At Econlog, I pointed out that the high inflation occurred without any “concrete steppes” by the Fed; T-bill yields stayed at 0.38% during 1945-47 and the monetary base was pretty flat.  Some of the inflation represented the removal of price controls, but I suspect some of it was purely (demand-side) monetary—a rise in velocity as fears of a post-war recession faded.

This era shows that you can have a lot of “reallocation” and a lot of austerity, without necessarily seeing a big rise in unemployment.  And if you are going to make excuses for the Keynesian model, you also have to recognize that most Keynesians got it spectacularly wrong at the time.  Keynesians often make of big deal of Milton Friedman’s false prediction that inflation would rise sharply after 1982, but tend to ignore another monetarist (William Barnett, pp. 22-23) who correctly forecast that it would not rise.  OK, then the same standards should apply to the flawed Keynesian predictions of 1946.

Tyler Cowen used to argue that 2009 showed that we weren’t as rich as we thought we were.  I think 1946 and 2013 (another failed Keynesian prediction) show that we aren’t as smart as we thought we were.

Update:  David Henderson has some more observations on this period.

 

The Fed begins to see the light

Each Fed statement, they inch a bit closer to market monetarism.  The newest statement lowered the forecasts for the future level of interest rates (the so-called dot plot) by about 50 basis points.  That’s still too high, but no longer as out of touch as they were a year or two ago.  Kudos to Kocherlakota and Bullard for seeing the light before the rest of the FOMC.  The long-term trend RGDP growth estimate was lowered again, this time from from 2.0% to 1.8%.  That’s still too high, but it’s getting closer to my estimate of 1.2%.  (The actual growth rate over the past decade has been 1.28%, but I believe the trend is still slowing.)

Question:  Has any school of thought been more accurate than market monetarism (over the past 8 years) regarding these issues:

1.  QE is expansionary.

2.  Negative IOR is expansionary.

3.  Forward guidance can be expansionary.

4.  Low rates are here to stay.

5.  Inflation won’t be a problem.

6.  NGDP and RGDP growth will be slower than the Fed expected.

7.  Sweden erred in not following Svensson’s advice.

8.  Trichet screwed up in 2011.

9.  Monetary policy would offset fiscal austerity in 2013

10.  Cross-sectional evidence for fiscal stimulus vanishes when confined to countries with independent monetary policy.

11.  Denmark would not be forced to revalue their currency upwards.

12.  Ending extended unemployment insurance in 2014 would accelerate job growth by about 1/2 million.

13.  Abenomics would increase Japan’s inflation rate.

14.  Switzerland would make their zero bound problem worse by revaluing the franc.

15. “Austerity” would not stop the UK unemployment rate from falling to full employment.

Why the BOJ policy move (mostly) lacked credibility

The BOJ’s recent decision is likely to end up being far more important than anything the Fed does or does not do today.  But as of now it raises more questions than answers:

1.  The BOJ announced it would cap 10-year bond yields at 0%, and also that it would attempt to overshoot its 2% inflation target.

2.  The BOJ did not announce lower IOR or more QE.

Today’s market reaction is hard for me to gauge.  The initial reaction was clearly positive, as stocks rose nearly 2%, and the yen fell by almost 1%.  Later, however, the yen more than regained the ground it lost.  That doesn’t mean the BOJ action had no impact, just that whatever impact it had was at most slightly more than markets anticipated.

The overshoot promise could be viewed as either Krugman’s “promise to be irresponsible”, or as a baby step toward level targeting.  Martin Sandhu gives a third interpretation—a signal that the target really is symmetrical, despite all the talk about a de facto 2% inflation ceiling in many countries.  There’s no reason that all three interpretations could not be a little bit true—after all, central bank policy is made by committees.

I vaguely recall Bernanke suggesting something like a long term interest rate peg—can anyone confirm?  I have mixed feelings about this idea.  On the one hand, I like moving monetary policy away from a QE/negative IOR approach, and toward a price peg approach.  But I view interest rates as almost the worst price to peg, for standard NeoFisherian reasons.  Are low long-term rates easy money, or a sign that money remains tight?  That’s not at all clear.

Although I am disappointed by the specific steps taken today, these actions do make me more optimistic in one sense.  The BOJ has shown that it’s still willing to experiment, and that it still wants to raise inflation.  Here’s an analogy.  When the Fed first engaged in forward guidance, they did so in a very ineffective manner—low rates for X number of years.  This was criticized as being rather ambiguous—in much the same way the BOJ’s 10-year bond yield cap is ambiguous.  So the next step in forward guidance was to make the interest rate commitment conditional on the economy, a major improvement.  Perhaps the BOJ’s next step will be to switch to price level target, in order to make the size of the inflation overshoot more concrete.  Or maybe instead of capping 10-year bond yields, they’ll peg something more unambiguous, such as the yen against a basket of currencies.  If that’s too controversial (and it probably is) then peg the yen against a CPI futures contract.

The danger is that this specific move won’t work, and the backlash will prevent the BOJ of moving further down the road in the future.  Maybe that’s why the yen reversed course a few hours later.

PS.  It just occurred to me that the 10-year bond yield cap could be viewed as a sort of commitment to enact a policy expected to lead the yen to appreciate by at least 1.68%/year against the dollar (for standard interest parity reasons).  That’s a non-NeoFisherian way of explaining why I’m skeptical.  In the long run, this bond yield cap means that Japanese inflation is likely to be 1.68% lower than US inflation.  They really needed to do the opposite of what the Swiss did early last year—they should have sharply depreciated the yen, and simultaneously raised interest rates.

PPS.  Kudos to Paul Krugman.  Ideas that start out seeming very “ivory tower”, such as promising to be irresponsible, can end up being enacted, at least in part.  Unfortunately, Krugman first proposed that idea under the (quite reasonable) assumption that liquidity traps would be temporary.  The markets don’t seem to believe that any longer, at least with respect to Japan.

Update:  Kgaard added this comment:

Scott — My understanding is that at the post-announcement press conference Kuroda said he would not be upping the amount of monthly bond purchases, and this is what turned the yen. Seems to me that what you wrote a couple days ago is entirely relevant here: If they SAY they want 2% CPI but then take actions consistent with 0% CPI, then what they really are targeting is 0% CPI until further notice, and investors will respond accordingly. Hence stronger JPY …

Update #2:  from commenter Mikio:

Kuroda did not say he will not expand purchases. On the contrary. He repeated again that the BOJ is ready to expand both the purchases as well as cut the IOR. But obviously they did not think they need to act now.

I think the jury is out there about this move. It’s non-progress if you look at the yen, it’s marginally positive if you look at stock market.

Update#3:  HL added the following:

Kgaard and Mikio are both right / wrong

After the longest delay for the statement release, markets learned at 01:18 pm Tokyo time, the following:

No change in policy balance rate at -0.1%
Monetary base expansion until inflation stable above 2%
JGB purchases in line with the current pace
MB/NGDP rate to hit 100% in 1 year (currently 80%)
Yield curve control introduced
Average maturity target scrapped
10 year JGB yield target around 0%
Forward guidance enhanced
Inflation overshooting commitment
Purchases to fluctuate to achieve curve control
Continue easing until inflation stably above 2%
No mention of timeframe
Comprehensive review on policy
NIRP helpful for decline in funding rates
NIRP didn’t seem to change banks’ willingness to lend
NIRP’s impact on yield curve, however, a bit problematic

Then during the preso (03:30~04:42), Bloomberg headlines
03:39 pm “No change in commitment to achieve 2% ASAP”
03:44 pm “Cutting minus rates further is still an option”
03:55 pm “Don’t think BOJ is coming close to limits”
03:55 pm “We just strengthened our framework”
04:12 pm “Expect inflation to hit 2% during FY 2017”
04:22 pm “amount of bond buying changes with the economy”
04:29 pm “BOJ won’t have JPY 80 trillion for JGB buying as fixed”
04:40 pm “New framework isn’t tapering”

So immediately after the statement release, markets had reasons to believe that some guidance on quantity part could be provided by Kuroda (continuing the purchase, etc). Then during the preso, Kuroda made that guidance more ambiguous. But the fact is: there is nothing in these comments to suggest that the BOJ will start tapering soon.

USDJPY had its strongest period between the statement release and the start of the preso Q&A. Then it was on a gradual decline even before Kuroda started making comments related to the quantity dimension. Eventually it settled around 101.70 before declining sharply at the start of New York session…

 

How did we end up here?

I’ve finally had a chance to read Paul Romer’s critique of macroeconomics, and it’s every bit as interesting as you might expect.  I’m going to focus on a single issue, which in my view lies at the heart of what’s gone wrong in recent decades—identification.  This issue has been the main focus of my blogging over the past 7 1/2 years, so it’s very dear to my heart. By late 2008, it was clear to me that not only did economists not know how to identify monetary shocks, but also that they were very far off course, and didn’t even understand that fact. Indeed this misunderstanding actually became highly destructive to progress in both economic science and economic policymaking.  One of the two the worst contractionary monetary shocks of my lifetime is generally regarded as “easy money”.  So how did we get here?

1. The earliest monetary shocks were seen as involving the price of money.  Coin debasement was a common example.  No one knew the money supply, and banks did not yet exist. This policy tool was used by FDR in 1933, but today has fallen out of fashion in big economies. Small countries like Singapore still use the price of money (exchange rates) as a policy instrument, but they do not drive the research agenda.  I’m trying to bring it back with NGDP futures targeting.

2.  Although the monetarist approach to identification (i.e. the money supply) dates back at least to Hume, it really came into its own with fiat money, especially during hyperinflationary fiat regimes.  Milton Friedman preferred the M2 money supply as a monetary indicator, at least during part of his career.

3.  Then for some strange reason the profession shifted from the money supply to interest rates, as an indicator of monetary shocks.  But why?

Perhaps you are thinking that you know the answer.  Maybe it had something to do with the early 1980s, when velocity was unstable and monetarism was “discredited”.  If that is indeed what you are thinking, then it merely illustrates that you are even more confused than you know.  Yes, velocity is unstable.  And yes, that means Friedman’s 4% money growth rule might not be a good idea.  But that has absolutely no bearing on the argument for replacing the money supply with interest rates, as an indicator of the stance of monetary policy.

The relationship between i and NGDP is just as unstable as the relationship between M2 and NGDP, probably more unstable.  At least with M2, we generally can assume that an increase means an easier monetary policy, and a reduction means a tighter policy.  We don’t even know that much about the relationship between interest rates and NGDP. Right now, markets expect about a 1% fed funds rate in 2019. Suppose you had a crystal ball that told you that the fed funds rate in 2019 would be 3%, not 1%.  There’s a classic “monetary shock”. The stance of monetary policy changed unexpectedly.  But which way—is that easier than expected policy, or tighter?  I have no idea, and you don’t know either.  Even worse, my best guess would be “easier” but the official model says “tighter.”

Paul Romer says we know that monetary shocks are really important.  I agree.  And he says the Volcker disinflation proves that.  I agree, and could cite many other examples, probably even more than Romer could cite.  So I’m completely on board with his general critique of those who claim we don’t know whether monetary shocks are important.  But Romer then claims that the real interest rate is a useful measure of the stance of monetary policy, and it isn’t—not even close.

Am I denying that if the Fed suddenly raised the real interest rate by 200 basis points, money would be tighter on that particular day or week?  No, I agree that that statement is true.  But it’s also true that if the Fed suddenly raised the nominal interest rate by 200 basis points, money would be tighter on that particular day or week.  Or if the Fed suddenly cut M2 by 10%, money would be tighter on that particular day or week.

So why don’t we use M2 to measure the stance of monetary policy?  Because over longer periods of time, movements in M2 do not reliably signal easier or tighter monetary policy.  But that’s also true of movements in nominal interest rates. If you have a highly contractionary policy, then inflation and nominal rates will fall in the long run.  Hence low rates don’t mean easy money.  And this argument also applies to real interest rates.  If the Fed adopts a tight money policy that drives the economy into a depression, then real interest rates will decline, even as policy is effectively contractionary.  This actually happened in 1929-33 and 2008-09.

All of the traditional indicators are unreliable.  The smarter New Keynesians will say that money is tight when the interest rate is above its Wicksellian equilibrium rate. But how do we know when that is the case?  After all, the Wicksellian equilibrium rate cannot be directly observed.  You need to look at outcomes; Wicksell said interest rates were above equilibrium when prices were falling, and vice versa. But that means we can only identify easy and tight money by looking at outcomes; are prices rising or falling?

Today we would substitute above or below 2% inflation, or 4% NGDP growth, but the basic idea is the same.  Money is tight when it’s too tight to hit your target, and vice versa. Ben Bernanke got this right in 2003, and then lost track of this concept when he joined the Fed:

Do contemporary monetary policymakers provide the nominal stability recommended by Friedman? The answer to this question is not entirely straightforward. As I discussed earlier, for reasons of financial innovation and institutional change, the rate of money growth does not seem to be an adequate measure of the stance of monetary policy, and hence a stable monetary background for the economy cannot necessarily be identified with stable money growth. Nor are there other instruments of monetary policy whose behavior can be used unambiguously to judge this issue, as I have already noted. In particular, the fact that the Federal Reserve and other central banks actively manipulate their instrument interest rates is not necessarily inconsistent with their providing a stable monetary background, as that manipulation might be necessary to offset shocks that would otherwise endanger nominal stability.

Ultimately, it appears, one can check to see if an economy has a stable monetary background only by looking at macroeconomic indicators such as nominal GDP growth and inflation. On this criterion it appears that modern central bankers have taken Milton Friedman’s advice to heart.

Others will object that New Keynesians understand that it’s the level of interest rates relative to the Wicksellian equilibrium rate that matters.  For instance, a recent paper by Vasco Curdia shows that money was actually quite contractionary, during and after the Great Recession.

screen-shot-2016-09-19-at-4-42-20-pm

Yes, but that paper was written in 2015.  Back in late 2008 and throughout 2009, market monetarists were just about the only people claiming that monetary policy was highly contractionary—and that was the period when we most needed clear thinking.  Others were lulled by meaningless indicators like low nominal and real interest rates, as well as a ballooning monetary base.

How did we end up using interest rates as an indicator of the stance of monetary policy?  Romer provides one possible clue in his paper:

By rejecting any reliance on central authority, the members of a research field can coordinate their independent efforts only by maintaining an unwavering commitment to the pursuit of truth, established imperfectly, via the rough consensus that emerges from many independent assessments of publicly disclosed facts and logic; assessments that are made by people who honor clearly stated disagreement, who accept their own fallibility, and relish the chance to subvert any claim of authority, not to mention any claim of infallibility.

I fear that economists have deferred too much to the “central authority” of central banks.  When I talk to macroeconomists, they seem to think it’s natural to use interest rates in their monetary models because the central banks actually target short-term interest rates.  But that’s a lousy reason.

Another problem may be that some economists are infected by a popular prejudice—that low interest rates are a “good thing” for the economy.  We visualize that we would be more likely to buy a new house if interest rates fell, and extrapolate from that to the claim that low rates would boost GDP.  That’s obviously an example of the fallacy of composition.  Yes, I’d be more inclined to borrow money if interest rates fell, ceteris paribus.  But some other guy is less inclined to lend me the money if interest rates fell, ceteris paribus.  Of course ceteris is not paribus if interest rates fall, and it all depends on whether they fall because of an increase in the money supply (expansionary) or more bearish expectations from the public (contractionary.)

Elsewhere I call this “reasoning from a price change”, and even Nobel laureates do it:

Real interest rates have turned negative in many countries, as inflation remains quiescent and economies overseas struggle.

Yet, these negative rates haven’t done much to inspire investment, and Nobel laureate economist Robert Shiller is perplexed as to why.

“If I can borrow at a negative interest rate, I ought to be able to do something with that,” he tells U.K. magazine MoneyWeek. “The government should be borrowing, it would seem, heavily and investing in anything that yields a positive return.”

But, “that isn’t happening anywhere,” Shiller notes. “No country has that. . . . Even the corporate sector, you might think, would be investing at a very high pitch. They’re not, so something is amiss.”

And what is that?

“I don’t have a complete story of why it is. It’s a puzzle of our time,” he maintains.

Actually there is no puzzle.  Shiller seems unaware that it’s normal for the economy to be weak during periods of low interest rates, and strong during periods of high interest rates.  He seems to assume the opposite.  In fact, interest rates are usually low precisely during those periods when the investment schedule has shifted to the left.  Shiller’s mistake would be like someone being puzzled that oil consumption was low during 2009 “despite” low oil prices.

I know what commenters will say—I’m a pigmy throwing stones at Great Men. They are right.  Guilty as charged.  Look, I’ve made the mistake of reasoning from a price change numerous times—it’s easy to do.  But that won’t stop me from criticizing the ideas of people much more famous than I am.  In Paul Romer I’ve found a kindred spirit.

PS.  Since I’m nearly 6’4″, perhaps I should be PC and add, “Not that there’s anything wrong with being a pigmy”.

PPS.  This link has videos to the recent Mercatus/Cato conference on monetary policy rules.

Paul Romer on the identification crisis

LK Beland directed me to a paper by Paul Romer, which I’ve only had time to skim.  But the abstract is great:

For more than three decades, macroeconomics has gone backwards. The treatment of identification now is no more credible than in the early 1970s but escapes challenge because it is so much more opaque. Macroeconomic theorists dismiss mere facts by feigning an obtuse ignorance about such simple assertions as “tight monetary policy can cause a recession.” Their models attribute fluctuations in aggregate variables to imaginary causal forces that are not influenced by the action that any person takes. A parallel with string theory from physics hints at a general failure mode of science that is triggered when respect for highly regarded leaders evolves into a deference to authority that displaces objective fact from its position as the ultimate determinant of scientific truth.

He focuses on RBC theory, but the problems go far deeper.  New Keynesian models also fail at identification.  Nick Rowe has an excellent recent post on this problem:

Suppose we model monetary policy as M(t) = bX(t) + e(t), where M is the money supply, X is some vector of macroeconomic variables, and e is some random shock. Or, if you prefer, as i(t) = bX(t) + e(t), where i is a nominal interest rate. We estimate (somehow) that monetary policy reaction function, and call our estimate of e(t) the “monetary shock”.

Let us suppose, heroically, that our estimate of the monetary policy reaction function is correct. The econometrician, by sheer luck, gets it exactly right. Whatever that means. And then we use that estimate of monetary shocks to see what percentage of macroeconomic fluctuations (somehow defined) was caused by those “monetary shocks”, and what percentage was caused by other shocks. And suppose, again heroically, we get it right.

This is nonsense. We are making exactly the same mistake that the people were making in my Gold Standard examples above. If the central bank had been following the monetary policy reaction function exactly (or if the econometrician had a complete data set and correct model of the central bank’s behaviour so the estimated reaction function fitted exactly) then by definition there would have been no “monetary shocks”. And so “monetary shocks” would explain 0% of anything, because there weren’t any. 100% of macroeconomic fluctuations were caused by other, non-monetary shocks. Any deterministic monetary policy will have zero “monetary shocks”, by that definition, and any organisation’s behaviour is deterministic, if we understand it fully enough. That is not a useful definition of “monetary shocks”.

Monetary shocks are not the e(t). Monetary shocks mean the central bank chose the wrong monetary policy reaction function. It’s the choice of parameter b, and the choice of which variables belong in the vector X.

People get sick of me always talking about “the stance of monetary policy”.  But the misidentification of easy and tight money is THE central problem in macroeconomics.  Everything else is a footnote.

On a related topic, check out my new post at Econlog—I’m interested in feedback on my graph.