Finally I’ve been getting a flood of German data.  I’ll rely on data from Mark Sadowski, but would also like to thank Johannes Fritz, libertaer, and Martin.  Here’s Mark:

The data is as follows:

RGDP (annual rate)
2006Q1 – 2,263,994.4 million euro
2008Q1 – 2,429,934.8 million euro
2012Q4 – 2,463,022.8 million euro

The average rate of growth fell from 3.6% to 0.3%

Productivity
2006Q1 – 40.7 euro
2008Q1 – 42.6 euro
2012Q4 – 42.6 euro

The average rate of growth fell from 2.3% to 0.0%

Employment (1000)
2006Q1 – 38,940.0
2008Q1 – 40,250.0
2012Q4 – 41,719.0

The average rate of growth fell from 1.7% to 0.8%.

Hours worked at an annual rate (calculated)
2006Q1 – 1,428.5
2008Q1 – 1,417.2
2012Q4 – 1,385.9

The average rate of increase fell from (-0.4%) to (-0.5%).

Unemployment Rate
2006Q1 – 10.6%
2008Q1 – 8.0%
2012Q4 – 5.4%

So employment rose by 7.1% over a period of nearly 7 years, and hours per worker fell by 3.1%.  That means hours rose by 4%.  Productivity rose by 4.7% and RGDP rose by 8.8% (roughly half from more hours worked, and half from improved hourly productivity.)

Many commenters pointed to various job sharing programs, or subsidies to keep workers employed during the recession.  Libetaer used the metaphor of putting 2 workers in one chair.  The one chair reflects relatively meager growth in NGDP, and the two workers represent job sharing.

It would be a mistake to assume that the German miracle was an illusion due to all this job sharing and slow productivity growth (caused by keeping excess workers around.)  The welfare loss to a society from a 5% RGDP shock is much greater if 5% of workers lose their jobs, as compared to all workers staying employed, but working 5% less hard.  Yes, the miracle seems marginally less impressive, but I’d still prefer Germany’s labor market to France’s.

PS. Mark’s data doesn’t show a significant break in hours worked after 2008, so that factor may be overrated.

PPS.  The low level of hours worked in Germany shows the absurdity of blaming economic problems elsewhere on “lazy workers.”

Most people seem to have misinterpreted my previous post, probably because I “buried the lede” at the end.  Just to be clear, I have no problem explaining why German unemployment fell between 2006 and 2012.  Although NGDP growth was fairly slow, Germany did lots of labor market reforms, which I’ve discussed in previous posts, and this opened up lots of low wage jobs.  That could also explain the relatively weak productivity figures, although as I indicated earlier there are lots of issues with German data, so I’m not sure exactly what happened to productivity.

Mostly I was asking if anyone knew of any accurate data for German employment, hourly productivity, and hours per year between 2006:1 and 2012:4.  No one seemed to know, so I assume the data is very hard to find, if available at all.  And even if I had the data, it wouldn’t really have any causal implication, it would simply be an accounting-like partitioning of growth into its components (labor employed, hours per year, and productivity per hour.

What I found amazing was the fact that total nominal compensation kept growing at roughly 3% per year, even as NGDP growth slowed from 4.8% to 1.3%.  In case anyone doesn’t think that is amazing, consider that the invaluable Mark Sadowski did research on this and found that nothing like that happened in any of the other eurozone members:

I checked to see how Germany compares with the 15 other eurozone members for which there is data (using the most recent quarter for which they have data) and what I found was Germany is the *only* country *not* to have a big drop in the rate of nominal labor compensation growth since 2008Q1. This seems to support David Beckworth’s argument that ECB monetary policy is primarily geared towards Germany’s needs.

I would encourage readers to not jump to conclusions when thinking in terms of the “musical chairs model.”  There’s a temptation to say; “the real problem was X,” when X was simply a manifestation of the musical chairs problem, i.e. the tendency for the unemployment rate to rise when wages are sticky and nominal income growth slows sharply.  Let me be more specific.  Suppose someone said the real explanation was the labor market reforms of around 2003-05.  Here’s how I would respond:

1.  Yes, that’s why Germany could create more jobs with a given amount of NGDP growth than the other eurozone members.  Call that the “miracle of 2006-12”

2.  But no, that doesn’t explain why German unemployment continued falling after 2008, when NGDP slowed sharply.  But the fact that aggregate nominal employee income continue rising at roughly 3% does explain why German unemployment kept falling after 2008.  Call that the “miracle of 2008-12”.

Those are two very different “miracles,” which require very different explanations.

[Update:  Marcus Nunes has an excellent post that presents a different interpretation, one more favorable to NGDP targeting.  I’m still thinking about the issue.]

PS.  This was one of the most successful labor market reforms in history.  The data suggests it increased the share of German GDP going to the working class.  However by creating low wage jobs it may have also made German wages more unequal.  So naturally the party responsible for one of the most successful labor market reforms in history now wants to scale back the reforms, and even the Christian Democrats are calling for a minimum wage.  Here’s The Economist:

WHEN Gerhard Schröder took to the podium in the Bundestag on March 14th, 2003, Germany was called the “sick man of Europe”. More than 4m Germans (11.6% of the workforce) were on the dole. A widespread assumption was that unemployment could never be defeated, merely “administered,” says Wolfgang Clement, who was the former chancellor’s labour and economics minister. Countering that spirit, Mr Schröder unveiled a package of reforms that he called Agenda 2010. The leader of the opposition, a little-known physicist from east Germany called Angela Merkel, derided it as unambitious. But it soon became clear that the agenda would transform Germany’s labour market.

Ten years later, what is the verdict? Financial crises be damned, Germany stands as an economic beacon, with record employment and the lowest youth unemployment in Europe (see chart). Some countries are studying Agenda 2010 as if it were a manual. Mr Schröder is being feted at conferences all over the world.

Only the Germans themselves, and notably Mr Schröder’s own Social Democrats (SPD) and the Greens, who were his coalition partners, seem unsure whether Agenda 2010 was a blessing or a curse. Now in opposition, but hoping this year to defeat Mrs Merkel, who replaced Mr Schröder as chancellor, the SPD cannot exactly disavow its own reform. But it is signalling to its blue-collar base and its own left wing that it might undo parts of it.

So the one European economy with a healthy economy wants to undo the reforms that made it successful.  Go figure.

This is a fairly important post, but it will take me a while to get to the punch line.  I’ve been trying to collect German data, but have not had a lot of success.  As far as I can tell, between the 1st quarter of 2006 and the 4th quarter of 2012, the German unemployment rate fell from about 12% to 5.4%.  But there seem to be different indices.  In any case, that’s the jobs “miracle.”  So then I began to look for other data.

RGDP rose by 8.8%, while NGDP rose by 16.94%.  At annual rates that’s about 1.3% RGDP growth and about 2.4% nominal.  So why such a big drop in unemployment?  The German population does seem to be gradually falling, but as far as I can tell that’s not the main factor.  German employment seems to have risen by around 7% to 10%, depending on the data source.  Can someone help me?  So that suggests either fewer hours worked per worker, or flat productivity.  But as far as I can tell productivity (RGDP/hour) rose around 4% total (with incomplete data), and hours per year only fell by around 1%.  But the hours number may be wrong, as I can’t tell if it includes part-timers.

So how did unemployment fall so much with slow NGDP growth?  Perhaps a bit of everything.  Low inflation helps; that means more RGDP for each euro of NGDP.  But 8.8% RGDP growth is not very much over 6.75 years.  So you then add in slow growth in worker productivity, or less hours worked per worker, or some combination.

BTW, employment in the US is down almost 1% over the same period where German employment has risen by 7% to 10%.  Our RGDP is up by 6%, but that’s not good for a country with robust population growth.

Now here’s where things get interesting.  NGDP grew very fast from 2006:1 to 2008:1, by 9.8% total, or about 4.8% per year.  Then NGDP started falling, and despite a rebound by 2012:4 was only up another 6.2%, barely 1.3% per year.  Market monetarism says that Germany should have done much worse after 2008:1 than before.  Initially they did do poorly, but over the entire 2008:1 to 2012:4 period they did quite well, with the unemployment rate falling from 8.2% to 5.4%.  So what’s wrong with market monetarism?

The short answer is that I took some shortcuts, instead of staying true to the “musical chairs model.”  In the past I’ve often argued that a fall in NGDP causes unemployment because there is less income to pay workers, and yet hourly wages are sticky.  Some workers end up sitting on the floor.  The logic of that model suggests that the real problem is not unstable NGDP, but rather instability in a component of NGDP, namely total wages and salaries.

Fortunately I was able to find German data for aggregate worker compensation (wrong link, try this.):

2006:1   286,190 mil. euros   (50.5% of NGDP)

2008:1  303,710 mil. euros  (48.8% of NGDP)

2012:4  347,820 mil. euros  (52.5% of NGDP)

To say these number blew my mind would be an understatement.  I immediately saw the annual rates were similar, and my calculator showed 3.0% growth during the huge German boom of 2006:1 – 2008:1, and then an almost identical 2.9% during the global recession and recovery, when their NGDP growth rate plunged from 4.8% to 1.3%.  No wonder the Germans love the ECB’s tight money policy!!

The logic of the musical chairs model is that you should stabilize the path of aggregate nominal wages and salaries, because hourly wages are sticky.  When I started blogging I assumed wage targeting would be politically impossible, and knew that NGDP targeting was already a well-regarded concept.  So I latched on to NGDP targeting.  But in retrospect I wish I’d latched on to aggregate employee income.  Call it “income targeting.”

Recessions are not caused by less spending; they are caused by less income going to workers.  Usually the two go hand-in-hand, but the German miracle tells us that when they diverge, it is employer employee income that matters most.

PS.  Now that market monetarism riding high, I figured it was time for a vicious internecine struggle for the soul of market monetarism.  Consider this the first shot.  🙂

PPS.  Just to anticipate one objection from my fellow MMs, I do realize that German NGDP was above trend in 2008:1, and that one could argue that it is currently near the trend line.  But in that case the unemployment rate should have fallen during 2006-08, and then risen.  But it fell fairly steadily.

PPPS.  Another advantage of total compensation over NGDP is that it is easier to measure, not susceptible to the sorts of definitional problems discussed by James Hamilton here.  (I.e., is R&D part of GDP?)

John Taylor recently criticized central bank policies in the major economies, and re-iterated his support for a clear rule for the policy instrument.  John Cochrane responded:

John Taylor has a lovely little blog post, encapsulating so much in a few sentences. An excerpt with comments (emphasis mine)

…there is a crucial issue which explains much of the enormous difference of opinion between critics and supporters of the Fed’s current policy. Critics such as me and Allan Meltzer … argue that monetary policy should focus on a clear strategy for the instruments of policy. A goal for inflation or other measures of macro performance is not enough if it is simply part of a whatever-it-takes approach to the instruments. Such an approach results in highly discretionary and unpredictable changes in policy instruments with unintended adverse consequences, as we have been seeing in recent years.

Supporters such as Adam Posen… are just fine with the Fed using, even year after year, a whatever-it-takes approach to the instruments of policy as long as there is an overall goal. With such a goal in mind, so their argument goes, the central bank can and should always intervene in any market, by any amount, over any time frame, with any instrument or program (old or new), and with little concern for unintended consequences in the long run or collateral damage in the short run (say on certain groups of people or markets) as long as it furthers that goal.

Critics are very concerned about those unintended consequences and collateral damage; they are also concerned about an independent government agency wielding such a great deal of power as it carries out a year-after-year whatever-it-takes approach. Supporters are much less concerned.

I have always had this problem with nominal GDP targets, inflation targets, and so forth. Ok, the Fed adopts your target. Now what? If nominal GDP doesn’t do what the Fed wants it to do, what should the Fed do about it? Talk more? (Monetary policy is starting to look more and more like foreign policy here).

.  .  .

Taylor, of course, would like the Fed limited to the instrument of short-term rates, and to follow the Taylor “rule” for setting them. But the principle is larger than that instance.

Like Cochrane and Taylor, I’m not happy with current policy, but I don’t believe they have the right answer.  Let’s start with the fact that the Taylor rule is a sort of flexible inflation target, so it’s not clear to me why Cochrane would oppose inflation targeting and praise Taylor’s defense of the Taylor Rule.  Any instrument-based monetary regime must have a nominal anchor, or else the price level becomes indeterminate.  Of course there are some instruments (the price of gold, for instance) that do anchor the price level, but I doubt Cochrane favors a return to the gold standard.

If we’ve learned anything in the past few years it’s that the Taylor Rule (and indeed any regime based on an interest rate instrument) is a lousy system.  It freezes up just when you need it most, at the zero bound.  I like to compare it to a car that has steering that works fine, except that it locks up when driving on twisty mountain roads.

A much better system would be to use NGDP futures as the policy instrument, i.e. make dollars convertible into NGDP futures at a fixed price, which rises by roughly 5% per year.  But that’s not going to happen anytime soon.  So how can I defend recent policies like QE, as being better than nothing?  One answer is that stock markets in the US and Japan love QE.  Conservatives are supposed to believe that markets are efficient, so the stock market response must be telling us something.  Now I suppose one could argue that stocks rose because of much higher expected inflation.  But in fact (PCE) inflation has fallen to 1% in the US, and market indicators of inflation continue to show that market monetarists have been consistently right about inflation (over the past 5 years), and the rest of the conservative movement in America has been consistently wrong.  Surely that counts for something?  My guess is that the markets see more NGDP leading to more RGDP in the short run, and that’s why stock investors love QE.

To summarize; Taylor and Cochrane are right that there is a lot to criticize in current monetary policy.  But the problem is not that central banks are targeting things like NGDP, rather the problem is that central banks are focusing on policy instruments, and are not targeting things like NGDP.  We have monetary policy as a series of “gestures.”  “We’ll do this and that, and then see what happens.”  A policy of NGDP targeting, level targeting, would provide transparency, as markets would know how much NGDP would rise over the next 10 years.  The monetary base and interest rates should be set at a level where the market expects NGDP growth to be on target.  Indeed John Cochrane has endorsed a slightly different market-based approach in previous posts.

The market’s are telling us that the “whatever-it-takes approach” is the right one, and this is why the US and Japan are now doing better than the eurozone.   The markets don’t care how big the monetary base is; they care about NGDP.  We need to focus on goals (hopefully NGDP level targeting) and let the policy instrument be the servant.

PS.  Re-reading Cochrane I’m not sure he does oppose inflation targeting, perhaps he simply opposes inflation targeting regimes that lack a clear instrument rule.  If so, I exaggerated our differences.

HT:  Michael

PPS.  Garett Jones makes an argument that seems related to Cochrane’s post:

So in many policymaking situations, more action means more uncertainty about the effects of the action. And if your goal is to be close to your target, then more uncertainty is a genuine cost of taking more action.

Hence, Brainard’s Conservatism Principle. Alan Blinder, formerly vice chair of the Fed, summed the Brainard Conservatism Principle this way when he applied it to monetary policy:

Estimate how much you need to tighten or loosen monetary policy to “get it right.” Then do less.

Actually you should do exactly as much as the markets think necessary to “get it right.”  (It doesn’t matter what you think.)  Period, end of story.  The mistake people make is thinking about monetary policy is terms of more or less “instrument action.”   There are two big problems with this worldview:

1.  There are many policy instruments; the base, interest rates, exchange rates, etc.  More of one might easily mean less of the other.  I favor NGDP futures targeting, and hence would never change my preferred instrument at all, relative to the trend line.

2.  The second problem is that many people (implicitly) assume that if all this QE hasn’t raised NGDP growth up to 5% or 6%, we’d need even more QE to get there.  On the contrary, if a robust NGDPLT regime were adopted, at say 5%, then the demand for base money would be much less, and hence the Fed could get by with even less QE.

Rather than thinking in terms of “effort,” think of monetary policy as steering an ocean liner.  It doesn’t take more effort to set the wheel at NE as compared to NNE.  So you always want to set the wheel in such a fashion as to equate the forecast of the port you will reach, with your desired destination.  Monetary policy is no different.

HT:  Saturos

Here’s Karl Smith:

Volker’s clout on the FOMC grew the entire time and when he became Fed Chairman in 1979, he dropped the hammer.

.  .  .

The Fed drove the funds rate up from 5% in 1976 to 19% in 1981. I know, I know, Scott Sumner is scowling. Here is the growth rate of NGDP over the same period.

.  .  .

Trending upwards, towards 15%, through the 60s and 70s. Pausing in 1976 and then in two big slams going all the way to a low of 3% in 1982.

I don’t think many people doubt that the policies of the Volcker Fed drove interest rates up and inflation rates down during the early 1980s, but since the general topic of government debt is so contentious its worth reviewing.

I certainly “doubt that.”  The increase in the fed funds rate from 5% in 1976 to 19% in 1981 was caused by easy money policies of Burns, Miller and Volcker, as I explain in this post.  NGDP growth peaked at 19.2% in 1980:4 and 1981:1.

Then in the spring of 1981 Volcker got serious about inflation.  The Fed adopted a tight money policy, and NGDP growth, inflation, and nominal interest rates all began declining, a process that would continue for decades.