It was the nominal shock: 2 more pieces of evidence
By “it” I mean the sharp fall in real GDP all over the world that began in August 2008. There are two parts to my hypothesis. One part is that the intensification of the financial crisis beginning in September 2008 was mostly caused by the sharp fall in NGDP, not foolish lending. Lenders had a right to expect that the Fed would keep nominal GDP growing, (it had been continuously growing since 1958.) In Part one I repeat part of my recent reply to Hamilton; those who already saw it should skip to part 2:
Part 1
For the sake of argument assume that $1 trillion of the estimated $4 trillion in financial losses was due to poorly thought out mortgage lending, i.e. loans that would have been defaulted on even with stable five percent NGDP growth. Then assume that the other $3 trillion in losses, which became apparent only late last year when estimates of NGDP plunged sharply, represented higher quality residential mortgages, as well as commercial and industrial loans that would have been sound at a five percent NGDP growth rate. If I am right, and most of the financial losses last fall and winter were due to falling NGDP, then a modest uptick in nominal growth should modestly trim those losses. I think we have had a modest uptick in nominal growth in the past few months. Yes, it is very small relative to the huge decline in NGDP, which over the past 12 months has fallen about eight percent below trend. But nevertheless things are looking a bit better.
Take a look at this AP story showing how this modest upswing affected the IMF’s estimates of total losses from the financial crisis.
ISTANBUL, Turkey – Likely losses from the financial crisis in the three years to 2010 have been reduced by $600 billion to $3.4 trillion as the world economy grows faster than previously expected, the International Monetary Fund said Wednesday.
And now imagine not a slight uptick, but a much more expansionary policy in 2008, one that kept NGDP growing at five percent rate in 2009 and 2010, even if a somewhat weak fourth quarter was unavoidable. Then assume the recent uptick represented 20 percent of the total collapse in NGDP. What kind of estimate does that imply for the share of financial losses that are due to a weak economy? I think it is very possible that only about $1 trillion of the $4 trillion in losses is due to bad lending. After all, look how little economic growth it took to trim $0.6 trillion off those estimated losses.
Part 2 The Great Recalculation?
Arnold Kling discusses several schools of thought that see the problem as real, not nominal. One is “real business cycle” theory, the view that recessions are caused by productivity shocks. Another is the “recalculation theory,” which has a more Austrian flavor. In my view the strongest case for that theory is that both housing and finance over-expanded in the period between 2004 and 2006. After the bubble burst employment declined in those industries. Because jobs are highly specialized, these workers had trouble finding jobs in other industries. So frictional unemployment increased. This also led to less total employment in housing and finance, and a much smaller increase in employment in all other industries.
Unfortunately, I was not able to find employment data for residential housing, but I do have employment data for the major industry groups. Government employment is almost exactly unchanged over the past year (so much for fiscal stimulus), so let’s focus on private sector employment:
Sector August 2008 employment August 2009 employment
Total Private 114,497 108, 736
Mining 787 707
Construction 7,177 6,093
Manufacturing 13,387 11, 771
Trade, Transport 26,354 25, 145
Information 2,990 2,821
Finance 8,141 7,706
Professional 17,727 16,600
Education 18,950 19,321
Leisure 13,454 13,156
Other services 5,530 5,416
Employment in construction and financial services fell by 1.5 million, but employment in other areas did not rise slightly, as expected. Instead, more than 4.2 million additional jobs were lost in other sectors. In addition, I think the 1.5 figure is a gross overestimate of jobs lost in housing and banking, as the construction field also includes the building of infrastructure, offices, factories, malls, etc. According to the BEA data, between 2008:2 and 2009:2, residential construction declined by $118 billion and nonresidential construction fell by $93 billion. But I am convinced that with 3rd quarter to 3rd quarter data the two would be about equal. This is because housing has declined continuously, whereas nonresidential construction was flat between 2008:2 and 2008:3, and then fell at an increasing rate. It looks to me like the fall in non-residential construction was a consequence of the sharp break in GDP that began in August 2008, not a cause. So I believe only about half of the lost construction jobs can be blamed on the housing bust. Thus it seems like less than 1 million out of the total of 5.7 million jobs lost over the past year can be attributed to the housing and banking problems. The other job losses could have been prevented with a more expansionary monetary policy that kept NGDP growing at about 5%.
BTW, I suppose some job losses in the home furnishing industries are indirectly related to housing. Specifically, the portion of home furnishing that go into new homes, not existing homes. But that industry is not large enough to dramatically change the stylized facts discussed above. So let’s summarize:
1. A recent IMF study came up with estimates that strongly imply a large share of the roughly $3 trillion is financial losses since last September were due to a weakening economy. With more robust NGDP growth, those losses would not have occurred. (Note that even in a flexible price classical model nominal shocks affect debt defaults if debt contracts are not indexed.)
2. Regional housing data suggests that the second wave of the housing market decline, which began in August 2008, spread to many heartland regions that had been spared any speculative excess during the sub-prime crisis. A root cause of the earlier stability is that in markets like Texas housing is a roughly constant cost industry. The sharp fall in house prices since August 2008 in those heartland markets is most plausibly explained by falling nominal incomes (i.e. falling NGDP) at the national level. And that’s monetary.
3. Even if 100% of the post-August financial crisis had been due to reckless lending, and none was an endogenous response to falling NGDP, it appears that only about one million of the 5.7 million jobs lost are due to that crisis. The rest occurred in industries that should have been expanding according to “real” models of the recession.
So in one sentence; it looks like the post-August 2008 crisis was mostly an endogenous response to the weak economy, and even if it wasn’t, it can only explain a small share of the recession.
So how about it Tyler? Can I bump you up from one third to 35% or 37%?
Tags:
2. October 2009 at 06:59
Casey Mulligan writes a lot about trends in nonresidential construction.
2. October 2009 at 07:34
I’ll go up to 36!
2. October 2009 at 07:36
Scott:
I trust you saw Josh Hendrickson’s replies (here and here) as well as Bill Woolsey’s reply to Arnold Kling’s views on money. Somehow, though, Arnold continues to hold the view that money doesn’t matter. Also, here is some evidence consistent with your view that monetary policy was tight late last year.
2. October 2009 at 08:42
Scott, I the have data on residential contruction employment. Send me an email before 5:00 if you want it.
gbush@cppib.ca
2. October 2009 at 10:09
A couple things you might be forgetting is that there was a great deal of exchange rate instability during the recession and a lot of the job losses may have been in trade related industries. Also, financial markets such as commercial paper froze up and the TED spread rose through the roof, and many jobs were shed in industries not directly to finance or housing but dependent on finance. I’m not sure how an increase in NGDP would replace say the commercial paper market.
2. October 2009 at 11:36
What would Arnold Kling say?
Part 1 – losses are accounting fiction and they have nothing in common with true economic loss. Losses are also heavily distorted by subsidies.
Part 2 – most of these jobs would have disappeared anyway, as they are based on unsustainable risk premiums and unsustainable trade deficits. Recalculation will create new jobs primarilly in export sectors and this can happen only slowly. Traditional monetary stimulus won’t work because at first all money will be hoarded by insolvent financial system. In the medium term when velocity returns to higher levels we will get pure inflation with no real effects.
2. October 2009 at 16:30
There are two parts to Kling’s Recalculation Argument, and _both_ must be true to believe his argument that stabilizing NGDP has no effect on the real economy:
1) ALL of the people who lost their jobs SHOULD have lost their jobs. The economy just hasn’t yet “figured out” what to do with them yet.
2) Stabilizing NGDP does not impact the speed at which a “recalculation” occurs.
Both of these arguments are weak.
1) EVEN if certain sectors haven’t “figured out” they should be growing, those sectors should at least know they should not be shrinking. But they all shrunk.
2) If the price signal is the mechanism that transmits information about what sectors should be growing, how can he possibly think that nominal price declines facilitate the clear transmission of that signal? (If he believes that companies budget their headcount in “real” dollars, he needs to get a “real” job.) And this leaves out the problem of debt, which is to say that companies may _want_ to grow, but may be constrained by deflationary prices when prices drop and their cash flow is sucked away to cover debt obligations.
2. October 2009 at 16:38
123 writes:
“Recalculation will create new jobs primarilly in export sectors and this can happen only slowly.”
Presumably, Kling also believes that depreciating the currency does not affect the growth rate of the exporting sector…
The Recalculation argument is not just weak, it is patently dangerous. Kling should seriously think about this. As a respected economist, if people listen to him or use his arguments as an excuse to do nothing (believing that nothing can be done and the economy will heal automagically), then he should take personal responsibility for some percentage of the people whose lives are destroyed as a result. This debate is about more than intellectual stimulation.
2. October 2009 at 17:26
what statsguy said, about kling.
the recalculation idea is interesting in a way, but i think it is fundamentally flawed.
the ‘sectoral structure’ of the economy is interesting, but people will substitute consumption and work in one sector for another if they have money and are spending money. there is no reason to believe that sectoral shifts should be sudden or more difficult than people are willing to pay for. they should cause scarcity (within sectors that are enlarging) and inflation within those sectors. etc, etc.
the metaphor i come to is that money supply is a bit like temperature — it allows molecules to move around — and when money is removed molecules are stranded — and it might appear that removing money ‘revealed’ some underlying structure, what you perceive is actually is a quite different state than existed before. (its main difference is that it is much less free. people are stuck. kling’s narrative doesn’t say anything about the necessity to keep people from getting stuck en masse — that is a fine state to recalculate in!)
another problem i have with the ‘recalculation’ story is that it fails as metaphor. more on that another day.
but there are a couple things i don’t understand.
one is what the fed could have done, given that its peer institutions in the banking sector were hoarding cash. i understand what the fed could do now, but i don’t understand what the fed could have done late last year. ssumner, you probably covered this in earlier posts and i missed them. could you point to some?
another piece of the story which i don’t hear much of, certainly from kling and the like, is the process of deleveraging. the fact that people and institutions were highly leveraged and, more importantly, sloppily financed (by historical standards) made them much more susceptible to a downturn. individuals and companies were (no data, sorry) in my impression more susceptible to a drop in NGDP because they carried more debt than equity. in oct-mar i could hear the sound of a hundred million stop loss triggers going off.
2. October 2009 at 19:39
The causal elements in “recalculation” are more numerous and complex
than you and Arnold have imained to conceive.
Accelerating profits in short period consumption sector can juice employment in he late stages of the artificial boom — e.g. people spending 2nd mortgage money on restaurants and landscaping.
After the boom, machines become comparatively more cheap than labor as compared to before the artificial boom. Substitution of machines for labor effects many sectors.
And this is just a tiny hint of the “recalculations” that are blocked from he imagination of the economists thougth as a pure artifiact of the economist’s premises going into his math models. “Recaculation” relations across time and between production streams and labor streams are rules out by the stipulations of the formalism.
2. October 2009 at 19:46
If you mean to include the work of Hayek, this isn’t true. See e.g. Hayek, Profits, Interest, and Investment, 1939
Scott wrote,
“The rest occurred in industries that should have been expanding according to “real” models of the recession.”
3. October 2009 at 08:18
Thanks TGGF, Nonresidential definitely turned down quite a bit after residential. This supports the point that there were two separate shocks.
Thanks Tyler. You are the most open-minded blogger, and also highly intelligent (a rare combination.) So your response is my benchmark. I hope you weren’t just being polite.
David, Thanks for the VAR tests–I just posted a reply. I generally agree with Bill, Josh, and you; I think Kling overlooks a lot of evidence that the most clearly identified monetary shocks have a powerful short run impact on prices and output. In addition, theory suggests that if monetary policy affects long run inflation, it should also immediately impact current asset prices.
Gregor, I didn’t get to this in time. Perhaps when you get a chance you could post in these comments a few basic numbers. Roughly how much did it decline between mid-2006 and mid-2008 (say July or August) and then how much since? Even rough estimates would be helpful.
rrm364. That is a good point, but it fits my analysis perfectly. I argued that the dollar appreciated strongly during the tight money period of July to November 2008. This reduced jobs in export industries, and reflects the effect of tight money. Prior to mid-2008 we had easier money and a weaker dollar, and this helped maintain employment in export industries. Of course much of the job loss was also due to weak overseas economies, but that reflected the fact that money was too tight at a global level.
123, I don’t agree with either point. I think most of the recalculation guys would agree that the recent huge accounting losses in banking do represent economic losses from bad loans. They are always talking about capital being mis-allocated.
I don’t see how a need to increase our exports would cause lower employment. Shouldn’t that lead to more employment in export industries. I also don’t see why lower risk premiums would make people want to work shorter hours.
Statsguy, Both points are excellent.
Statsguy#2, I also think this is a dangerous view, although I don’t doubt that Kling is well-intentioned. But I would hope that those favoring various “real” views of the problem would consider NGDP targeting as a safety valve in case they are wrong. Many real business cycle economists were able to live with the inflation targeting approach, even as they disagreed with the new Keynesian model that underlay this policy. NGDP targeting is just a stable M*V trajectory, something right-wingers should be willing to live with.
q, I like the molecule analogy. Regarding what they could have done, since January 1st I have been publishing articles advocating a negative rate on excess reserves, to discourage hoarding. Lots of people thought I was crazy, until the Swedish Riksbank adopted the idea.
And yes, I have also emphasized that the fall in NGDP made the “real” crisis worse, partly because people were “over-leveraged” as you say. In effect, people became highly leveraged because they thought modern central banks would no longer let NGDP fall. They were wrong.
Greg, I think the main reason machines are substituted for labor is that wages are sticky and the rental cost of machines is pretty flexible. But the root cause of sticky wage unemployment is falling NGDP. If you keep NGDP growing then wages will not get out of equilibrium. If NGDP keeps growing and we still observe the sort of problems we saw last fall, then I will abandon my model and become an Austrian.
Greg#2, I was tying to address the points made in the Garrison slide show that everyone told me to look at. If I got it wrong, then he needs to explain the Austrian view in a more understandable fashion.
3. October 2009 at 08:22
@ greg ransom :
i have no idea what you mean by saying that a ‘boom’ is artificial. you seem to be saying that people bought stuff with ‘money’ that they ‘borrowed’. however, the fact is that the real economy had more than enough productive capability to allow this to happen. so what went wrong? what went wrong was accounting.
the ‘debt’, ie second mortgages, that you are talking about is just an expression of the belief that the economy is going to expand nominally. if the economy expands nominally, a large and increasing amount of aggregate debt is safe; if the economy contracts, it is not safe and may cause insolvencies with debt intermediaries as we’ve seen. the danger of a high debt level is not with the high level of the debt, it is that the effect of negative nominal growth rates is much greater when debt levels are high than when debt levels are low.
3. October 2009 at 08:24
@ssumner, looks like i cross-posted with you. your exposition is clearer and more concise than mine.
3. October 2009 at 11:31
Scott, you wrote:
“I don’t agree with either point. I think most of the recalculation guys would agree that the recent huge accounting losses in banking do represent economic losses from bad loans. They are always talking about capital being mis-allocated.”
Sorry for my unclear language. I wanted to say that recalculation guys think that accounting losses have only slight connection to much larger economic losses. Recent reports about reduction of financial system losses represent two things – changes in accounting standards that allow to hide some economic losses, and fiscal subsidies (some on-balance-sheet, some off-balance-sheet).
“I don’t see how a need to increase our exports would cause lower employment. Shouldn’t that lead to more employment in export industries. I also don’t see why lower risk premiums would make people want to work shorter hours.”
Standard IMF advice for current account deficit countries is to reduce real wages to stimulate exports. Without sharp exchange rate adjustments real wages adjust slowly leading to a long period of lower employment. By the way Kling supports changes in the tax code in order to reduce real wages. Risk premiums were unsustainably low and they allowed to have a huge current account deficits. Now that risk premiums are higher we need exports.
3. October 2009 at 15:53
Greg, I don’t think that it will help to bring up the Ricardo effect.
Something I think that recent discussions on this blog have shown is that most of the posters aren’t really willing to spend the time to learn the Austrian Business Cycle Theory. If they won’t learn that then they’re not going to get around to the Ricardo effect.
We have to get people to look at the general problems of economics a bit differently first.
And I haven’t read Profits, Interest and Investment.
3. October 2009 at 21:04
Scott writes:
“I think the main reason machines are substituted for labor is that wages are sticky and the rental cost of machines is pretty flexible. But the root cause of sticky wage unemployment is falling NGDP.”
Right. The secondary deflation get a lot of its power via various “recalculations”.
And note well, in “Profits, Interest and Investment” where Hayek discusses machines being substituted for labor in the post-boom bust, he assumes sticky wages, wide spread idle resources, and above natural rate unemployment. Add in a bit of secondary deflation and your story is Hayek’s story.
3. October 2009 at 21:11
“q” writes:
“i have no idea what you mean by saying that a ‘boom’ is artificial”
Read some of the essays by Roger Garrison available at his web site.
Google: “Roger Garrison” and “artificial boom”
3. October 2009 at 21:20
Note well that the central explanatory problem Hayek was focused on was the post-1925 unemployment problem in Britain, where everyone recognized both VERY sticky union wages and significant deflation. In his back and forth with Keynes, it was this British situation which was at the front of Hayek’s mind, as a non-American economist from the Continent living & teaching in Britain among the British.
Scott writes:
“I think the main reason machines are substituted for labor is that wages are sticky and the rental cost of machines is pretty flexible. But the root cause of sticky wage unemployment is falling NGDP.”
4. October 2009 at 02:45
Just to add to what Greg has said….
What we should do is compare the benefits that can be had from “capital saving” technology and “labour saving” technology. If in a bust the prices of capital fall, but the prices of labour don’t fall to the same degree then labour saving technology becomes more attractive in comparison to capital saving technology than it was before.
4. October 2009 at 05:46
Thanks q,
123, If we want to lower real wages to boost exports, then the easiest way is with an expansionary monetary policy. Exchange rates move faster than nominal wages. So I entirely agree, but see this as a problem reflecting excessively tight money, not a structural problem.
I think Singapore lowers payroll taxes to boost employment during recessions. If that is what Kling is advocating, then I agree.
Greg, If the secondary deflation is part of what you call the recalculation story, then I would obviously change my tune. I assumed that ‘recalculation’ referred to adjustments that weren’t associated with nationwide nominal shocks.
In retrospect I think the interwar unemployment problem in Britain was structural, not cyclical. It was the same set of factors that caused the natural rate of unemployment to be so high in the eurozone over the past three decades. I agree real wages were a problem, as was unemployment comp. Deflation was a problem at times, but wages were the deeper problem. So I probably would agree with much of what Hayek says on this issue.
One reason I did such a quick pivot in this cycle (from seeing the problem as real in early 2008, to seeing it as nominal in late 2008) is that I am aware that others failed to pivot when necessary, and lost a lot of credibility (fairly or unfairly.)
Current, I agree, and again blame sticky wages.
4. October 2009 at 12:07
Do we have to use this term of Klings “recalculation”? I’m not sure I really understand what he means by it.
4. October 2009 at 12:26
Scott: “I agree, and again blame sticky wages.”
Greg is bringing up here, what is one of the most thorny issues in Austrian economics. It would take some time to explain exactly why it’s so thorny.
To give you a flavour…. Suppose that you agree with me for the case of recessions. Surely therefore it is sensible to take the opposite view about booms?
The amount invested in new capital equipment during recessions is surely less than in booms. But, because of the effect we mention above it is likely that in recessions it will be spent more on labour saving technology. In booms the opposite may be true, since wages are often sticky upwards too. So, although in a boom more may be spent overall on capital purchases, the slant may be more towards capital saving technologies.
This implies that it may not be correct to say that MV=PQ and MV=PT will move roughly in synchrony.
4. October 2009 at 18:47
Scott writes:
“If the secondary deflation is part of what you call the recalculation story, then I would obviously change my tune. I assumed that ‘recalculation’ referred to adjustments that weren’t associated with nationwide nominal shocks.”
The point is: secondary deflation CAUSES “recalculation” problems, problems which produce discoordination effects, e.g. unemployed people with specific skills and unemployed production goods with particular jobs to do. As Steven Horwitz points out in his excellent _Microfoundations and Macroeconomics_ this is the exact flip side of the “recalculation” problem produced by inflation, unjustified credit expansion, etc. i.e. “nationwide” nominal shocks that enter and move through the economy in ways that do NOT fit with the old set of relative prices.
The whole point, Scott, is a “macro” that pretends that relative prices don’t exit isn’t thinking about the real world. And surprise, because — by assumption — the real world doesn’t exist, then as if by magic changes in money and credit have no effect on the relations between different production streams and different workers and different savings and consumption plans, i.e. the whole structure of production across time, interrelated with changing savings, investment and consumption plans — plans which are often systematically distorted and unsustainable due to money, credit, bandwagon, bubble, and “animal spirit” effects.
And instead of working via relative prices, production streams, and the real world — everything happens in an “aggregate” one good, homogeneous labor, unstructured fantasy world where everything happens “magically” at the “nationwide” level via magical “shocks”.
The plea of the “recalculation” argument is to embrace real causal processes across the structure of relative prices, production, labor, investment, consumption, savings, etc. and to think again about appeals to magic.
Same argument as the Darwinian biologists against the creationist …
Yes money can deflate and deflation has economic consequences — So show us what these are through the network of relative prices and across the structure of production. Don’t give us magic.
5. October 2009 at 06:04
Current; You said;
“This implies that it may not be correct to say that MV=PQ and MV=PT will move roughly in synchrony.”
I am not sure what this means. Is “T” all transactions? In that case I agree that the two equations are completely unrelated.
Greg, You said;
“The whole point, Scott, is a “macro” that pretends that relative prices don’t exit isn’t thinking about the real world.”
I certainly can’t be accused of that sort of macro. I believe that secondary deflations cause real wages to rise. That causes unemployment to rise. And the unemployment is concentrated in certain industries. So I think I agree with Austrians who argue that all sorts of misallocations occur due to relative price changes. I also believe that the relative price distortions caused by secondary deflation could be avoided with a policy of NGDP targeting.
5. October 2009 at 11:41
Point taken.
Studies continue to provide lots of support for the idea that rising relative real wages is a key part most business cycles over the last 100 years. In addition to the new UCLA study on the Depression, Vedder and Gallaway have a good summary of the data in their classic book _Out of Work_. The ongoing rise of the minimum wage can’t be helping in the current environment.
Scott writes:
“I certainly can’t be accused of that sort of macro. I believe that secondary deflations cause real wages to rise. That causes unemployment to rise. And the unemployment is concentrated in certain industries. So I think I agree with Austrians who argue that all sorts of misallocations occur due to relative price changes. I also believe that the relative price distortions caused by secondary deflation could be avoided with a policy of NGDP targeting.”
5. October 2009 at 12:46
Scott, you wrote:
“If we want to lower real wages to boost exports, then the easiest way is with an expansionary monetary policy. Exchange rates move faster than nominal wages. So I entirely agree, but see this as a problem reflecting excessively tight money, not a structural problem.”
Congratulations, you have reached the point where I don’t know the answer to the question “What would Kling say?”.
5. October 2009 at 13:11
Yes, by “T” I mean all transactions. I agree with you that the two equations are not closely related (I’m not sure if they’re entirely unrelated). This causes problems for some monetary disequilibrium theories. I suppose yours isn’t one of those.
6. October 2009 at 16:46
I think we all agree.
Greg, Yes, the minimum wage is a factor in the especially rapid rise in youth unemployment (even compared to 1982 when total unemployment was even higher, but youth unemployment was lower.)
123, Yeah. I have no idea how Kling would respond to this one, as the exchange rate has both nominal and real aspects. Monetary policy can move real exchange rates in the short run, as it isn’t a sticky price.
Current, I was thinking of the transactions in the various financial markets, which are much bigger than NGDP. Include them and you have a very different equation. If you just add intermediate goods and used goods, then the change is much smaller.
8. October 2009 at 04:24
Yes, this is where things get very tricky. I don’t understand this well myself. But, the problem is that Austrian capital theories involve more than GDP components. The involve capital components such as circulating capital like intermediate goods. However, there’s a level of vagueness between when circulating capital starts and capital equipment begins.
I think it’s an area where a new piece of theory is needed.