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.

Suppose we were back in the 1990s, and unemployment was 5.0%.  But now suppose the economy was growing slowly due to slow growth in the working age population and slow growth in productivity.  A “Pop Keynesian” says that we can solve this problem with fiscal stimulus.  What do the smart 1990s Keynesians say in reply?

What do they say today?

Will the press ever learn?  Here’s Reuters:

Data released before the Tokyo market opened showed that Japan’s economy slipped back into recession in the July-September quarter, contracting at a 0.8 percent annualised rate, compared with the median estimate for a 0.2 percent contraction.

Last year I had fun mocking press reports of a 2014 “recession” in Japan, which I believe was the 4th in 6 years.  And now another.  However the more sophisticated press is finally beginning to catch on.  Here’s the FT:

Bright and early on Monday morning, expect to hear some gloomy news — that Japan is in recession, for the fourth time in just five years. Doom will be mongered; foes will declare it a fresh blow to Prime Minister Shinzo Abe’s “three arrows” of economic stimulus. And I will not believe a word of it.

That is not because the cries of recession will be false — even though the unemployment rate is low and falling, the corporate mood is one of cautious optimism and Japan is manifestly not entering a slump. Output is expected to fall by 0.1 per cent in the third quarter after a 0.3 per cent decline the quarter before — and, on the standard definition, two consecutive quarters of contraction equals a recession.

The trouble is this definition of a recession is anachronistic and misleading. It is an idea that no longer means what we think it does — that is, a significant decline in activity. Even worse, the way we use it causes harm, not just in Japan but also in an increasing number of advanced economies. The R-word needs a rethink.

Japan keeps entering “recessions” because its population is falling and so the economy has little potential to grow. The Bank of Japan puts trend growth at 0.5 per cent or less, compared with the US Federal Reserve’s estimate of 2 per cent for America.

Actually those are both too high; trend growth is zero in Japan, and 1.2% in the US. But at least the press is starting to figure it out.  That’s progress.

[Fourth city in 5 nights.  Tired.  I wrote this a few weeks ago to post while traveling. I hope to respond to comments for this post and earlier posts on Sunday.]

So I started rereading the General Theory, just for the heck of it, and noticed something very interesting.  Early in the book Keynes lays out the musical chairs model of the business cycle.

The General Theory begins with an appalling caricature of “classical economics”. Keynes says that the classical economists believed that unemployment was caused by sticky wages.  This is true.  Then Keynes suggests that the classical economists believed all unemployment was “voluntary”.  The idea was (according to Keynes) that the classical economists believed that full employment could be reached at a sufficiently low wage level, and hence if there were unemployment then it must be voluntary—workers refused to take the pay cuts that would restore full employment.

This makes no sense to me, as it seems to confuse collective action and suffering at the individual level.  No single automobile factory worker can regain a job by cutting his or her wages, and thus their unemployment is every bit as involuntary as if the sticky wages were caused by an Act of God. But that’s not how Keynes looked at it.  Then we come to a sentence that reminds me of the style of a certain famous NYT columnist:

Obviously, however, if the classical theory is only applicable to the case of full employment, it is fallacious to apply it to the problems of involuntary unemployment—if there be such a thing (and who will deny it?).

Who will deny it?  Keynes has just told us that the classical economists deny the existence of involuntary unemployment. Then why say, “Who will deny it?”  Keynes wants the readers to nod their heads, and note that there’s clearly lots of involuntary unemployment 1936, it’s just that the classical economists are too dense to see the obvious.

But what is so obvious about involuntary unemployment, as defined by Keynes? We all agree that there were lots of people without jobs. We all agree that lots of them wanted to be working.  We all agree that lots of them were miserable.  I call that “involuntary unemployment.”  But I think they were unemployed because of sticky wages, and that if workers collectively accepted lower wages then we would have had full employment in 1936.  And Keynes tells us that if we hold the belief that I hold, and that many interwar economists also held, then we are not entitled to say we think there is such a thing as involuntary unemployment.

OK, so what is Keynes’s theory of unemployment?  Brace yourself.

Sticky wages!

The only difference is that Keynes also believes that if workers did accept wage cuts, it would not solve the problem.  Prices would also fall in response, and hence real wages would not fall.  I don’t think he’s right, but the key point is that he thinks that assumption somehow magically turns “voluntary unemployment” into “involuntary unemployment.”  I can’t imagine that the unemployed workers even understand that distinction, or care.  Nor do I understand how the reader is supposed to think it’s “obvious” that there is one type of unemployment and not the other.  Does a miserable unemployed worker look different if his plight is “voluntary” at a level that he has no control over?

It might seem I’m nitpicking, but Keynes’s entire critique of the “classical model” of unemployment falls apart once you allow for the fact that sticky wages can cause involuntary unemployment.  The interwar economists had perfectly fine theories of the business cycle, and Keynes doesn’t lay a glove on them.  So much for chapter 2.

In chapter 3 he argues that demand shocks cause changes in employment, and says that the “classical” economists did not realize this.  The classical economists (supposedly) assumed Say’s Law held true.  And yet during the interwar years the standard model of business cycles was essentially a demand shock model, although economists used different terminology in those days. Fisher did early Phillips Curve models.  Pigou, Hawtrey, Cassel, and the others certainly understood that demand shocks (monetary shocks) affected employment.

I must say that Chapter 3 is beautifully written, especially section 3.  But that must have infuriated the interwar economists all the more.  I can’t even imagine what it would be like to be unfairly attacked by a brilliant economist with a devastating wit. Oh wait . . .

In book II things get much better.  Recall that I often argue that inflation is a vague and undefined concept, which makes real GDP also hard to pin down.  On the other hand I do grudgingly concede that the CPI provides a ballpark estimate of inflation.  I certainly don’t think the actual rate is 8%, as some claim.  Inflation estimates are accurate enough so that we know China has grown faster than Zimbabwe, despite the higher NGDP growth in the latter country.  On the other hand I’ve also suggested that macroeconomists should discard inflation and real GDP, and focus on more easily measured concepts, such as NGDP, wages, and employment.  Now let’s look at Keynes:

But the proper place for such things as net real output and the general level of prices lies within the field of historical and statistical description, and their purpose should be to satisfy historical or social curiosity, a purpose for which perfect precision “” such as our causal analysis requires, whether or not our knowledge of the actual values of the relevant quantities is complete or exact “” is neither usual nor necessary. To say that net output to-day is greater, but the price-level lower, than ten years ago or one year ago, is a proposition of a similar character to the statement that Queen Victoria was a better queen but not a happier woman than Queen Elizabeth “” a proposition not without meaning and not without interest, but unsuitable as material for the differential calculus. Our precision will be a mock precision if we try to use such partly vague and non-quantitative concepts as the basis of a quantitative analysis.

.  .  .

In dealing with the theory of employment I propose, therefore, to make use of only two fundamental units of quantity, namely, quantities of money-value and quantities of employment. . . .

We shall call the unit in which the quantity of employment is measured the labour-unit; and the money-wage of the labour-unit we shall call the wage-unit.

.  .  .

It is my belief that much unnecessary perplexity can be avoided if we limit ourselves strictly to the two units, money and labour, when we are dealing with the behaviour of the economic system as a whole; reserving the use of units of particular outputs and equipments to the occasions when we are analysing the output of individual firms or industries in isolation; and the use of vague concepts, such as the quantity of output as a whole, the quantity of capital equipment as a whole and the general level of prices, to the occasions when we are attempting some historical comparison which is within certain (perhaps fairly wide) limits avowedly unprecise and approximate.

Right on!!  That’s exactly my view. (I did a similar post on this a year ago, with a much longer section of quotations, in case you want more context. Here I’ll explore different implications.)

So Keynes basically has a model with three components:

1.  Aggregate demand shocks (NGDP shocks)

2.  Sticky wage-units  (sticky nominal wages)

3.  Labour unit fluctuations  (employment fluctuations)

Wait, that’s my musical chairs model!

This is all explained by the end of chapter 4, on page 45 of a nearly 400 page book. If I were writing the General Theory, I would have merely added that NGDP is determined by monetary policymakers, and then called it a day after 50 pages.

I probably won’t read any more, because I know what’s coming next.  There will be hundreds of pages explaining what causes demand shortfalls, and monetary policy plays only a modest role in the model.  Why did Keynes and I reach such different conclusions?  Because I am living in a fiat money world where central banks really do determine NGDP.  Indeed if they didn’t then the entire concept of inflation targeting would be nonsensical.  The 2% inflation rate since 1990 would be an amazing coincidence, almost a miracle.

In contrast, Keynes lived in a gold standard world or at least in a world where countries were expected to soon return to a gold peg.  In that sort of world monetary policy was much more constrained.  If you go to the extreme and assume M was fixed, then explaining NGDP is all about explaining velocity.  And what explains velocity?  Things like interest rates and uncertainty.  Here are some things that would lower V when M is fixed:

1.  The public decides to save more–lowering interest rates, which lowers V.

2.  Less animal spirits among businessmen, which leads to less investment, less demand for credit, lower interest rates and lower V.

3.  Fiscal austerity, which lowers government borrowing, lowering interest rates and V.

4.  A big inflow of foreign saving from surplus countries like Germany and China, which depressed interest rates and lowered V.

5.  Wage cuts would transfer income to capitalists, which would boost saving, reduce interest rates, and reduce V.

But Keynes did not build the General Theory around the Equation of Exchange—that would not have been revolutionary.  Instead he created the Keynesian cross and hinted at IS-LM.  Instead of pointing out that shocks to V affect NGDP and hence (because of sticky wages) also employment, he made it seem like thrift, and bearish expectations, and austerity, and beggar-thy neighbor trade policies directly caused unemployment.  He appealed to the prejudices of what he calls the “uninstructed.”  These are, after all, the common sense views of most people, even politically conservative people who are not well versed in economics.  Here’s Keynes:

That it [classical theory] reached conclusions quite different from what the ordinary uninstructed person would expect, added, I suppose, to its intellectual prestige.

The General Theory screams out that the uninstructed people were correct, and then whispers that they were correct for the wrong reason, and not at all when monetary offset applies.  The problem here is that he was too successful.  We now have several generations of reporters and politicians who think this these bearish factors are always contractionary, even when not at the zero bound.

Of course Keynes did talk about the possibility of monetary stimulus, and was pessimistic that it would be sufficient in a deep depression, for standard liquidity trap reasons (actually its more complicated, but the liquidity trap is still a necessary condition for complete monetary policy ineffectiveness.)  And even if there was not a complete liquidity trap, under a gold standard the ability to print money was limited.

To summarize, Keynes started with the same basic business cycle model as I use—the musical chairs model.  Combine deficient nominal spending with sticky wages and you end up with high unemployment.  But he did not become a market monetarist, for two reasons:

1.  He did not think that wage cuts would help, they would merely lead to further cuts in AD.

I think that’s wrong, but I also think it’s a pretty minor dispute.  As a practical matter I don’t think wage cuts are a very effective solution to a collapse in NGDP. The big difference is the second distinction:

2.  He did not think the monetary authority could provide adequate levels of AD (NGDP) during a depression.

There’s an interesting similarity between these two points.  Keynes was very clear that excessively high real wages were a root cause on unemployment, but didn’t think nominal wage cuts would help.  Keynes was very clear that inadequate spending in money terms was the root cause of labor markets being out of equilibrium, but didn’t think printing more money would (necessarily) solve the problem.  So in both cases he correctly diagnosed the problem, but drew back from the obvious solution.

However I do think we need to cut him some slack here.  Things looked very different in a gold standard world, and most other economists were just as confused as he was.  A few such as Fisher, Hawtrey and Cassel saw the true nature of the (monetary) problem more clearly than Keynes did, but they were the exception.

I’m a bit skeptical about the utility of the Keynesian model, even in a world of constrained monetary policy.  But here I need to admit that an awfully lot of very smart people see things differently than I do, so it’s at least plausible that the Keynesian model has merit as a sort of backdoor way of explaining movements in V when M is fixed, and hence as a theory of demand side business cycles under certain types of monetary regimes.

But I don’t think I’ll read anymore.  I can’t bear to work though 100s of pages of convoluted (implicit) explanations for why certain shocks might impact V.  If only Keynes had stopped at page 45 and said:

“So the central bank should create a NGDP futures market and target the futures price along a track rising at 4%/year.  That makes the world safe for classical economics.

The end.”

I define monetary policy in terms of NGDP growth.  So tight money occurs when NGDP (or expectations of NGDP) are falling.  I also claim that tight money causes most recessions.  Some commenters argue this is a tautology.  That complaint is actually really stupid.  Just think about Zimbabwe, where RGDP plunged as NGDP rose by zillions of a percent.  But it’s also an unintentional compliment, as we’ll see.

I frequently point out that the unemployment rate is strongly correlated with the ratio of wages to NGDP.  Some people argue that this is a tautology.  That’s actually a somewhat more reasonable complaint, although it’s not accurate.  I say reasonable, because a slightly different version of the “musical chairs model” (MCM) would be a tautology. This version:

1/employment (in hours) = W/Aggregate nominal labor income.

I replaced unemployment with 1/employment (in hours), and I replaced NGDP with total labor compensation, which is well over half of NGDP, and highly correlated with NGDP.

So what’s going on here?  Is this a model or a tautology?  And if it is a model, how do we falsify it?

Here it would be helpful to think about how a real business cycle (RBC) economist, AKA “new classical” economist, would criticize the MCM, if one of them actually took the time to examine my pathetic little model.

They’d clearly accept the fact that NGDP and RGDP growth are highly correlated in the US, and that unemployment is highly correlated with W/NGDP.  That’s not the issue. The issue is the policy counterfactual.  We observe that NGDP is often more volatile than W, especially during periods like 2008-09, where NGDP suddenly fell by 3%, or roughly 8% below trend, and wage growth only slowed very slightly.  So W/NGDP rose sharply, as did unemployment.

My counterfactual is that had NGDP kept growing at 5% in 2008-09, then RGDP would have also kept growing (although it would have slowed slightly for supply-side reasons) and I claim that wages would have continued growing at about 4%.  An RBCer would not agree.  In their view the counterfactual result would be high inflation and high nominal wage growth, indeed wages soaring at perhaps 10%/year, or something like that. And because wages would have soared by 10%, the stable 5% NGDP growth would lead to 5% fewer hours worked, and the unemployment rate would soar from 5% to 10%. RBCers don’t believe than nominal shocks have real effects.  The Great Recession was caused by real factors, in their view.

The math fits, but how plausible is that counterfactual?  And keep in mind, BTW, this is the ONLY possible counterfactual to my claim that stable NGDP growth would have maintained high employment in 2008-09, (or at least that stable growth in nominal aggregate labor income would have worked, if you want to be picky.)  Even if you are not a RBCer, but blame it on “reallocation”, this HAS TO BE your theory.  For the RBCers to be right:

1.  Wages must be highly flexible, and that flexibility is cleverly hidden by Fed policies that, de facto, keep equilibrium nominal wage growth fairly stable.

2.  It must be true that with a 5% NGDP growth counterfactual, wage growth would soar much higher in a period of fast rising unemployment like 2008-09, all because, well because fast rising wages are mathematically required to produce the big drop in hours worked that the RBCers insist is the “equilibrium” outcome of some mysterious hard to identify technology shock that causes workers to want to take long vacations. Or something like that.

Can you tell that I don’t find the RBCers counterfactual to be particularly plausible?

Other possible flaws in the musical chairs model?  There really aren’t any.  When we get to the policy implications, you can certainly question my claim that:

NGDP = stance of monetary policy,

or that the Fed is capable of stabilizing NGDP growth at 5%, or ask whether fiscal policy is needed too.  But those policy issues are entirely separate; here I’m just trying to figure out what causes business cycles.

And basically it’s fluctuations in W/NGDP.  Which are mostly caused by NGDP shocks. Period, end of story.  It’s a really ugly model, and kinda stupid.  I wish it weren’t true. But it’s also incredibly robust, just unbelievably robust.  Unlike New Keynesians who utilize the Phillips Curve, I can go to sleep at night with serene confidence that I won’t wake up to Bob Murphy or Arnold Kling, or anyone else having some sort of empirical evidence that refutes the model.  At best someone might find an episode where it doesn’t fit it quite as well as usual.

The more interesting question is whether the rest of the profession should take this seriously as a model.  I’m too close to give an objective answer.  Take a look at these graphs, in a blog post by “Robert“:

Now take a look at the bottom graph.  See that spike at zero?  The RBC model predicts no spike there at all, because there in no money illusion allowed in the RBC model.  So immediately we know that the RBC model is wrong.  But all models are wrong, even mine.  The real question is how wrong.  Maybe they are wrong in assuming no wage stickiness, but in fact there’s only a small amount of wage stickiness, and the RBC model is almost true.

Fair enough.  But then they need to convince the rest of the profession on empirical grounds.  And go back and read my points #1 and #2 above.  Good luck convincing the rest of the profession that those counterfactuals are even remotely plausible.

If we can dismiss the RBC critique of the musical chairs model (which no one has made, AFAIK, but it’s the critique that would be made, by anyone who believed the RBC model) then what’s left for me to defend?  Mostly monetary policy.

So let’s say we agree that NGDP shocks cause recessions.  OK, that’s not very controversial; indeed it’s pretty much Keynes’s General Theory (I’ll do a post soon explaining why the General Theory is basically the musical chairs model.)  So I can’t really claim to have invented anything here.  But somehow we’ve drifted away from the General Theory, which focused on NGDP and average hourly wages and employment, to NK models that focus on interest rates and inflation and output.  And that’s obscured Keynes’s musical chairs model.  Instead of using NGDP shocks to explain employment fluctuations, we use interest rates and government output and technology shocks and price shocks to explain output.  It all becomes very confusing, when in fact demand-side recessions are very simple—not enough NGDP to pay the workers.

Maybe if we had an NGDP futures market we could start thinking of NGDP as a policy choice for the central bank, a tool, a lever, and then the musical chairs model would be right in front of us, staring us in the face.  Unavoidable.

No matter how hard you try to focus on something else like price stickiness or technology or investment shocks or whatever, W/NGDP is right there, right in front of you. It’s the sine qua non of demand side macroeconomics.  If the causation doesn’t go from NGDP, to W/NGDP, to employment, then the RBCers are right.  And they ain’t right.

PS.  Robert relied on this paper by Mary Daly and Bart Hobijn.