My flawless post showing a strong job market in 2007
Arnold Kling recently criticized my post arguing that the US job market was relatively strong in 2007 and early 2008. I cited a low unemployment rate, and a reasonably good labor force participation rate (LFPR.) Here’s what Kling had to say in response:
I am going to react to three things: Nick Rowe talks about the fact that housing transaction volume is higher when prices are rising; Scott Sumner’s latest attempted swindle; . . .
Scott tries to twist the labor market data to raise doubts that labor demand was weak prior to 2008. I am sorry, but I am using the same labor market indicators that I used years before the crisis, and I think that the decline in labor demand prior to 2008 is pretty evident. See this post.
I eagerly read the rest of Kling’s post looking for all the examples of how I attempted to “swindle” the reader and how I had “twisted” the data, and didn’t find a single problem with my post. I’ll take that as implying there are no flaws. Then I checked the post Kling linked to:
So here is, if you will, daughter of LUCY. You take total nonfarm payroll employment and divided it by the population over 16. (This is basically just an employment to population ratio.) Divided each monthly number by the peak value, reached in December of 1999, then multiply by 100. The resulting index behaves as follows:
This differs from my LFPR in several respects. The LFPR uses population aged 16 to 65 in the denominator, whereas modified LUCY uses all population over 16. Thus an increase in the number of women living to be 93 does not affect my indicator at all, whereas using Kling’s ratio it indicates (assuming those women don’t work) a loss of jobs. I also believe the LFPR includes the total number working in the numerator, whereas LUCY uses nonfarm payroll employment. I don’t object to Kling using this indicator, and never criticized his post. I’m still a little puzzled as to why he criticized my post. Perhaps in his next post he’ll tell us exactly how I twisted the data and swindled my readers.
Update: Not so flawless after all. There are several versions of the LFPR, some stop at age 65 and some go all the way up to infinity. The BLS has no age limit. Commenter Brent explains where the difference between Kling and I probably lies:
Actually, the labor force participation rate published by BLS uses the civilian noninstitutional population aged 16 and over as the denominator. So a 93 year-old woman will be included in the denominator unless she resides in an institution, such as a prison, mental hospital, or nursing home.
The differences between your measure and Kling’s are in the numerators. Your numerator is the number of persons classified as either employed (including the self-employed) or unemployed (that is, actively searching for work). Kling’s numerator is the number of wage and salary jobs“”that is, people with second jobs are counted twice and the self-employed are not counted at all.
The statistics suggest that second jobs grew pretty rapidly during the 1990s and declined during 2000s. It isn’t clear to me whether these changes were driven more by shifts in labor demand or in labor supply. At any rate, for the purpose of discussing missing jobs, I think people are usually thinking of primary jobs rather than second jobs, so I prefer your measure.
PS. Just to be clear, I was not arguing that the job market in early 2008 was as strong as in 2000, just that it didn’t seem nearly as weak as the Jim Tankersley article I linked to had suggested. I think it is quite possible I am wrong, but I’m still waiting for a commenter to successfully refute my post.
Tags: Unemployment
23. January 2011 at 09:32
The job market in Orange County began turning seriously south in 2007:
http://economy.ocregister.com/2009/03/05/oc-lost-64600-jobs-over-the-last-year/5317/
The unemployment rate was going up each quarter in 2007 in OC, jobs were being lost in finance, construction and other sectors, In 2006 jobs were being created at a fast clip, and unemployment was 3.2 %
Here’s a flash graphic that shows the spread of unemployment from the housing bust centers through time and then spreading across the country.
http://boingboing.net/2009/11/23/animated-map-of-us-u.html
23. January 2011 at 09:37
If you aggregated over the whole world, you could even more successfully obscure the malinvestment / recalculation bust of 2007, which is geographically structured as structured in terms of capital good and labor specialization and relative price.
23. January 2011 at 09:44
Who´s twisting? Kling has to do several transformations on the data to get his preffered result!!! Certainly the labor market was not as strong in early 2008. At this time NGDP was a little below trend and employment in non residential (CRE) peaked in march 08. As late as june 08 unemployment was (only) 5.5%. At this point residential construction employment had already dropped 20% from it´s peak in april 06!
23. January 2011 at 09:46
Employment in Orange County, CA tops out in December, 2006:
http://www.calmis.ca.gov/file/lfhist/oran$hlf.xls
23. January 2011 at 09:49
Right, the great recalculation was just getting under way — the first pepples before the avalanche ….
“At this point residential construction employment had already dropped 20% from it´s peak in april 06!”
23. January 2011 at 10:03
Pebbles do not necessarily turn into an avalanche, unless some nut blows up a few sticks of dynamite. Yes, that´s exactly what the Fed did by letting nominal expenditure tank after june 08!
23. January 2011 at 10:06
During the time the so called recalculation was just “getting on the way”, non residential construction and export sectors were INCREASING employment…
23. January 2011 at 10:27
How did that work out.
“During the time the so called recalculation was just “getting on the way”, non residential construction and export sectors were INCREASING employment…”
23. January 2011 at 11:22
Arguing with Austrians:
Look, here we have the aggregate, national income…
“No, no, you are committing an aggregation flaw! You must look at sectoral breakdown!”
Okay, okay. Look, here we have spending by sector: healthcare, housing, etc. And employment by sector: manufacturing, real estate, etc.
“No, no, you are committing an aggregation flaw! You must look at geographical breakdown!”
…
23. January 2011 at 11:23
Unemployment rates by county since Jan. 2007 — THE MOVIE:
http://www.latoyaegwuekwe.com/geographyofarecession.html
A boom- bust malinvestment / recalculation cycle is structured in terms of: production sectors, geography, labor specialization, length of production, relative prices, financial asset prices, etc.
23. January 2011 at 11:29
(lest anyone think my previous comment was just for snark, I should point out that while it is entirely coherent for such flaws to be hidden in aggregates, an increasingly non-aggregable economy must have increasingly crazy and coincidental excess demand curves for the substitution effects that Austrians love to dominate. It’s the intuitive outcome of arguing, as ABC does, that the capital market interacts so damn crazily with the real interest rate that it explodes at the touch of a feather. Or a Fed-er.)
23. January 2011 at 11:43
Scott, this is what was happening in the tech market in Q3, 2008:
http://gigaom.com/2008/10/08/sequoia-rings-the-alarm-bell-silicon-valley-in-trouble/
This came after a long run up in what many folks considered to be a bubble.
Essentially, EVERY technology start up was INSTRUCTED to fire people, to stop growing, focus on core revenue.
And that’s tech, that’s the most resilient real thing going – it’s not non-health care / public employee / bankster.
It fought off the gripping hands of fear longer than anything else, and came back quicker.
—-
The issue, AS ALWAYS, is that you like to look at the peak “right before” the fall and then say that’s “strong.”
It isn’t, the peak right before the fall is not sustainable.
It’s just like your trend on NGDP, if we trend from “sustainable” say the slow climb right after the tech crash 2001-2003 and before the housing boom – something WE KNOW FOR SURE is sustainable, its puts us on trend now.
No, if there was a crash, there was a boom, and the peak before the crash is not a “strong job market.”
23. January 2011 at 11:57
sorry david, your impression does not have enough randomly capitalized words and phrases.
23. January 2011 at 12:27
I just skimmed through some of Scott’s recent earlier posts and, lo and behold, here is Bob Murphy doing exactly what I was attempting to parody in my 11:22 comment: when you find the aggregates don’t support your theory, pick the aggregate apart into subaggregates. When that still doesn’t work, repeat on the subaggregates.
I mean, look at what he wrote in (3): he has a narrative, and the facts change, so he goes digging for new facts. Arrrrgh. Here’s a bet: he stopped digging for new facts the moment he managed to finagle something that fit his narrative.
23. January 2011 at 12:28
Folks, let me tell you about employment. Some areas and industries are always going through “less good” times. (Let’s talk about the city of Detroit’s long term trends, shall we? While people were beginning to point to California have some very specific California problems rather before the turn, so examples from California are underwhelming.) That’s why aggregates are, well, aggregates. They start to turn when the number of areas and industries having “less good” times start increasing.
As for “malinvestment”, how one can give a definition of that which is not dependent on the level of economic activity beats me. What’s a fine idea in downturn Houston may very much be “malinvestment” in Port-au-Prince. If firms go bust when economic activity drops significantly, what does that prove except the level of economic activity has dropped significantly?
23. January 2011 at 12:50
Scott wrote:
The LFPR uses population aged 16 to 65 in the denominator, whereas modified LUCY uses all population over 16.
Actually, the labor force participation rate published by BLS uses the civilian noninstitutional population aged 16 and over as the denominator. So a 93 year-old woman will be included in the denominator unless she resides in an institution, such as a prison, mental hospital, or nursing home.
The differences between your measure and Kling’s are in the numerators. Your numerator is the number of persons classified as either employed (including the self-employed) or unemployed (that is, actively searching for work). Kling’s numerator is the number of wage and salary jobs“”that is, people with second jobs are counted twice and the self-employed are not counted at all.
The statistics suggest that second jobs grew pretty rapidly during the 1990s and declined during 2000s. It isn’t clear to me whether these changes were driven more by shifts in labor demand or in labor supply. At any rate, for the purpose of discussing missing jobs, I think people are usually thinking of primary jobs rather than second jobs, so I prefer your measure.
23. January 2011 at 13:58
I thought it was a joke. A follow-up to “Scott Sumner fights dirty – he looks at the data” (or similar words).
23. January 2011 at 16:28
David.
This isn’t that hard.
Success came in providing a sound causal explanation of the origin of species when Darwin threw over-board Plato fixed and interacting aggregates — and when he followed his theory and looking for the appropriate causal diaaggregation within appropriately selected aggregates to causally explain both the existence of those aggregates — and their change and replacement through time.
Menger & Jevons did the same thing in economics — inventing the economic way of thinking and solving the aggregation paradox of the valuation of water and diamonds, etc.
Theory tells us where to look for the causal force that changes the aggregates — and it tells us which aggregates are significant for thinking about the process.
The economic way of thinking tells us that the causal mechanism works at the margin — and anyone who is not incompetent knows that every choice involves a time element — choices take place at the margin of more or less time.
Creationist make arguments like David all the time because they are ignorant of the Darwinian way of thinking, a theoretical lens which tells you which aggregates matter and how change at the margin leads to macro changes in larger collectives.
David doesn’t know which margins to look at and which aggregates matter for process analysis because he has never thought to use the economic way of thinking when the choice involves more or lass time — which is the central coordination problem of production and labor economics.
23. January 2011 at 16:33
Lorenzo you are ignorant of Darwinian biology, you are equally puzzled by concepts like “adaptation” or “fitness”, etc.
The economic way of thinking applied to time decisions in production good choices is a theoretical lens that gives you such a definition and gives you eyes to see such things.
Ignorance of theory leaves one blind to processes evident to those who know the science — as any Darwinian biologist will tell you.
Lorenzo writes:
“As for “malinvestment”, how one can give a definition of that which is not dependent on the level of economic activity beats me.”
23. January 2011 at 16:35
Make that:
_If_, Lorenzo, you are ignorant of Darwinian biology, you are equally puzzled by concepts like “adaptation” or “fitness”, etc.
Lorenzo might write:
“As for “adaptaton”, how one can give a definition of that which is not dependent on the overall size of a species population beats me.”
23. January 2011 at 16:45
Lots of things involve time, Greg. Asserting that time-structure is the source of macroeconomic fluctuations is quite another.
Screaming about time-preference is as convincing as Keynesians screaming about liquidity-preference; as I’m sure you are aware, like Keynesian models of liquidity, most of the simple renditions of the Austrian intuition about time preference run into assorted empirical problems. Tyler Cowen seems to like to emphasize comovement. I won’t rehash the usual litany here.
23. January 2011 at 17:01
Greg, You said:
“The job market in Orange County began turning seriously south in 2007:”
Yes, that’s exactly my point. My critics say housing construction wasn’t shedding jobs–but it clearly was. The fact that the national unemployment rate didn’t go up shows that jobs lost in housing construction do not have a serious effect on the national unemployment rate.
Regarding your second comment; you focus on local sectoral issues, I focus on aggregate issues. Macro is about aggregates, indeed that’s the definition of macro. You are right that OC was sliding into recession in 2007. You are wrong about the US sliding into recession in 2007.
Marcus, Do you have a source for residential construction employment (separate from other construction?)
Greg said:
“How did that work out.”
Pretty well until the Fed let NGDP fall sharply below trend.
Greg, The unemployment by county supports my argument. It shows jobs were being lost locally, but not nationally.
Morgan, Gee, what else happened in 2008:3?
Lorenzo and david, Exactly.
Thanks Brent, I’ll do an update.
Nick, Yes, I assumed he was partly joking. Mine was also supposed to be funny–but that may have been unclear, as humor isn’t my strength.
23. January 2011 at 17:21
The economic way of causal explanation involves causation on margins — which changes and eplains countable aggregates.
You can attempt to create a fake science doing something else by stipulative fiat — but it will be the scientific equivalent of explaining the origin of species without causation at the margin, and with only natural kind aggregates interacting magically as aggregates.
The problems of trade cycle theory and monetary economics existed when marginalist scientists tackled those problems — the the invention of the word “macro” didn’t change those problem — and it didn’t solve them either.
It led to 80 years of intellectual fashion churn — one thing dumped after another. It didn’t lead to respectable science or success in running the economy — or in persuading even the profession of economists that macroeconomists were doing real science. You certinly haven’t convinced real scientists in other fields.
Stipulations of what is “science” in this area by definitional fiat lacks the sort of persuasive impact the explanatory success and causal explanatory power have.
Mainstream macroeconimcs lack both of the latter — which perhaps explains the dependence on persuassion by definitional fiat (begging the question).
23. January 2011 at 17:25
david wrote:
“I just skimmed through some of Scott’s recent earlier posts and, lo and behold, here is Bob Murphy doing exactly what I was attempting to parody in my 11:22 comment: when you find the aggregates don’t support your theory, pick the aggregate apart into subaggregates. When that still doesn’t work, repeat on the subaggregates.”
This is the traditional “vulgar” Austrian approach. It is complemented by the “prove that my counterargument is totally irrelevant” approach. And then they wonder why they are the laughing stock of the econ blogdom.
They sort of remind of religious zealots forecasting the end of days except that they don’t come handing out pamphlets on Sunday mornings.
23. January 2011 at 18:28
At some point not so long ago, my understanding is that 1/4 of those unemployed came from the construction industry — this doesn’t count those unemployed in microeconomically connected ways, e.g. finance, real estate, local vendors, forestry, construction materials, or those affected by the lost income of the unemployed from construction, etc.
I’m at a lost to understand what sort of game you are playing here, Scott.
It’s an interconnected economy.
The housing crash was even interconnected with the supply of shadow money.
And the appeal to “and then we use magic” to re-coordinate these massive coordination failures my a magically injecting money just where it is needed to re-coordinate everything in a sustainable manner over time — well, “and then a miracle occurs” isn’t science.
Enough money could be dumped in exactly those places were it is needed, unendingly across time, to continue the impossible to sustain production and consumption plans of 2006.
And using the distraction of “look over here, there’s a nice waving red flag over here” (and don’t look over there), doesn’t change anything.
23. January 2011 at 18:29
Make that:
“Enough money could not be dumped in exactly those places were it is needed, unendingly across time, to continue the impossible to sustain production and consumption plans of 2006.”
23. January 2011 at 18:36
Morgan, Gee, what else happened in 2008:3?
Scott, my point was, when Tech ran up the flag, that means everything else had long before it.
Your said, “my post arguing that the US job market was relatively strong in 2007 and early 2008”
And I’m saying there’s no way you can say early 2008 was “relatively strong” and by October it was DOOMSDAY.
If you data says it, you have shitty data, the wrong data, you are misreading the data.
It means late 2007 early 2008, things were headed South, and had been.
23. January 2011 at 18:39
Greg Ransome wrote:
“And the appeal to “and then we use magic” to re-coordinate these massive coordination failures my a magically injecting money just where it is needed to re-coordinate everything in a sustainable manner over time “” well, “and then a miracle occurs” isn’t science.”
Actually it’s an appeal to use the greatest marginal benefit relative to marginal cost to coordinate everything. It’s an appeal to use Adam Smith’s invisible hand coupled with a steady growth in the level of nominal expenditures.
23. January 2011 at 18:41
My appologies, I meant “Ransom” of course.
23. January 2011 at 19:19
Apropos, from a terribly long time ago:
In fact Greg commented on the entry, saying:
So there we have it: Greg does, indeed, believe that people leave $500 bills on the sidewalk. His reading of Hayek is consistent – if you posit a market that detonates under Fed-induced nominal shocks, because the patterns of coordination at 1% higher rates are so ridiculously different, then the market must also melt down under unexpected real shocks. Investors are crazy and stupid, etc.
Us over here, we’ll be noting that the market does not, in fact, regularly melt down over real shocks.
24. January 2011 at 07:28
“I am using the same labor market indicators that I used years before the crisis, and I think that the decline in labor demand prior to 2008 is pretty evident.”
The problem with Kling’s logic is that he imposes his preferred indicators on the discussion. If he is looking someplace other than where you are looking, well then the choice of indicators is the place to begin the discussion. Your decision to look at indicators which do not tell the same story Kling sees in his preferred indicators is not, on the face of it, “twisting”.
Sadly, PhDs often do not grant the holder the ability to see past their own silly notions. An argument that boils down to “the stuff I look at doesn’t show the story that he is telling, so he must be wrong” would not make it through a debate in the undergraduate lounge.
24. January 2011 at 08:11
Greg, You said;
“At some point not so long ago, my understanding is that 1/4 of those unemployed came from the construction industry “” this doesn’t count those unemployed in microeconomically connected ways, e.g. finance, real estate, local vendors, forestry, construction materials, or those affected by the lost income of the unemployed from construction, etc.”
Once again, construction unemployment has no bearing on my argument, check out my newest post. I am concerned with residential construction unemployment.
24. January 2011 at 08:43
Greg,
In response to “It led to 80 years of intellectual fashion churn “” one thing dumped after another. It didn’t lead to respectable science or success in running the economy “” or in persuading even the profession of economists that macroeconomists were doing real science. You certinly haven’t convinced real scientists in other fields.”,
I point out that theoretical physics, especially as it applies to cosmology, has seemed more faddish to me. And macroeconomics is far better grounded in some ways, as there isn’t speculation that ranges from multiple universes, to multiple extra-dimensions.
Macroeconomics is more challenging, because it isn’t an experimental science and things are still messy at the micro level compared to experimental physics. Add to this, that it’s still perhaps very unclear how the micro at the levels of individuals accounts for macro behavior, especially vis-a-vis cognitive biases.
On the other hand, I can show you simple deterministic models of behavior with correlations with experimental results that approach 100%. Take the matching law, for example, when reinforcement is delivered through direct brain stimulation.
I end by saying that it seems to me that those who call macroeconomics a pseudoscience usually know almost nothing about the field. I’ve heard at least a few physicists comment on evolutionary biology and/or economics in very arrogant, ignorant ways.
Besides, how well can physicists model fluid dynamics? How much progress has been made there? As much as has been made in macroeconomics?
24. January 2011 at 08:45
Oh, and we know that nonlinear dynamical relationships are by far more common than the linear ones most physicists deal with. Perturbation theory only takes one so far.
24. January 2011 at 08:48
Morgan,
The unemployment rate was still very low in early 2008 and Scott makes a very compelling case that falling NGDP can account for the rapid deterioration that really picked up during August of that year.
As far as I can tell, the data seems very consistent with his claims and has continued to be consistent with his model since/
24. January 2011 at 09:16
Kharris, yes, it doesn’t seem logical
24. January 2011 at 11:39
Mike, the argument:
The data says things were good in the Economy in early 2008, such that IF only X,Y,Z did not happen, the Tech market would have been fine….
Fails the smell test. It means your data is wrong, or you are mis-interpreting. We were hearing bad news, rounds not getting closed, lower rounds, the fall-2007. And Tech held out longer than anybody, it recovered in real terms faster.
By mid-2008, there was nothing happening out there, and by the time the crash came – it wasn’t like anyone was surprised.
In order for your narrative to fit, it has to be a surprise,we had to wake up and say, “what happened?” and that didn’t happen.
Expectations, according to you folk, matter – and expectations were headed down.
24. January 2011 at 22:25
Mike writes,
“how well can physicists model fluid dynamics?”
I note that when such phenomena is mathematically intractable, physicists have the advantage that they don’t claim the phenomena doesn’t exit.
Things are different in economic. The marginalist mathematics of hetereogenous production good choices through time is mathematically intractable.
Hence “K” and the insistence by modern economists that hetereogenous production process coordination requiring changing choices of more or less time does not exist.
25. January 2011 at 09:05
It can exist, the contention is that it is macroeconomically insignificant.
25. January 2011 at 13:51
Greg, I doubt that any economist ever claimed that re-calculation doesn’t exist.
30. January 2011 at 01:07
Greg: On Darwinian adaptation, have you ever heard of an “extinction event“? I find budget constraints a really useful way to think about natural selection.
An extinction event is when budget constraints shrink massively and suddenly such that whole species disappear. After the event is over, budget constraints massively loosen. A whole lots of new “experiments in living” (known as ‘species’) evolve. There is a surge in the number of species. This stabilises until one gets a fairly even pattern of extinctions and additions. Until the next extinction event.
But how one can have a notion of ‘adapted’ or ‘fitness’ without some reference to the size of the available niche/resources beats me.
After all, the best definition of a living thing is ‘has revealed preferences’. So, of course budget constraints help in thinking about biology.