Productivity - Wage Gap Grows

Labor Department announced today that productivity of U.S. workers rose 6.4% in the second quarter of 2009. Productivity is measured by output per hours of workers. The largest increase since the third quarter of 2003.

Here is the kicker:

The huge increase was due to hours working declined faster than output.

Oh, wait there is more:

Hourly compensation in the nonfarm business sector increased 0.2 percent in the second quarter of 2009, compared to a decrease of 2.4 percent one quarter earlier. When the 1.3 percent rise in consumer prices was taken into account, real hourly compensation fell 1.1 percent in the second quarter of 2009 (seasonally adjusted annual rates).

So, wages are not even competing up with inflation.

We have the productivity-wage gap growing. We have hours worked falling. We have real hour compensation decling.

This all translate to more destruction of the middle class and more income inequality. Our economic growth model is broken, as Dr. Palley argues, and this report is just another example of that.

I am adding these quotes from this news story because, in my opinion, show the total disregard for the survival of the middle class:

"It's good because it helps keep inflation low; labor costs are pretty benign," said Scott Brown, chief economist at Raymond James & Associates in St. Petersburg, Florida.

"On the other hand it means you can do more with fewer people," he said.

What an asinine thing to say. You would think increasing peoples income would help grow the economy. Then there is the obsession with inflation and wage growth:

Unit labor costs, a gauge of inflation and profit pressures closely watched by the Federal Reserve, fell 5.8 percent, the biggest decline since the second quarter of 2000. Analysts had expected unit labor costs to fall 2.4 percent in the second quarter. Unit labor costs dropped by a revised 2.7 percent in the January-March quarter.

It seems to me that an increase in labor costs particularly in the form of real wages would help the economy. This neo-liberal economic growth model is not sustainable and it will destroy the middle class.

Update: RO here. I was writing a post at the same time so to keep topics consistent, I am attaching it to this post.

The Bureau of Labor Statistics of the U.S. Department of Labor today reported preliminary productivity data--as measured by output per hour of all persons--for the second quarter of 2009. The seasonally adjusted annual rates of productivity change in the second quarter were:

  • 6.3 percent in the business sector
  • 6.4 percent in the nonfarm business sector

Productivity gains in both sectors were the largest since the third quarterof 2003, and were due to hours worked declining faster than output.

In manufacturing, the preliminary productivity changes in the second quarter were:

  • 5.3 percent in manufacturing
  • 3.9 percent in durable goods manufacturing
  • 2.0 percent in nondurable goods manufacturing

Manufacturing productivity grew 5.3 percent in the second quarter of 2009, as output fell 9.9 percent and hours worked decreased 14.4 percent (seasonally adjusted annual rates). This was the largest quarterly gain in manufacturing productivity since the first quarter of 2005, when output per hour increased at a 7.3 percent annual rate.

These are some astounding blow out productivity numbers.  While the report attributes the record breaking increases to a drop hours worked, there might be more to the story in offshore outsourcing and the drop in trade imports.   Stay tuned for a more detailed analysis.  In the interim, see Table A below:


Table A. Productivity and costs:  Preliminary second-quarter 2009 measures
(Seasonally adjusted annual rates)
                                                 Hourly    hourly    Unit
                    Produc-                      compen-   compen-   labor
Sector              tivity    Output   Hours     sation    sation    costs
                       Percent change from preceding quarter
Business            6.3       -1.8      -7.5        0.1     -1.2     -5.8
Nonfarm business    6.4       -1.7      -7.6        0.2     -1.1     -5.8
Manufacturing       5.3       -9.9     -14.4        5.8      4.4      0.5
  Durable           3.9      -16.5     -19.6        8.7      7.3      4.7
  Nondurable        2.0       -3.4      -5.3        2.0      0.7      0.0
                     Percent change from same quarter a year ago
Business            1.9       -5.4      -7.1        1.1      2.1     -0.7
Nonfarm business    1.8       -5.6      -7.3        1.3      2.2     -0.6
Manufacturing      -1.3      -15.0     -13.9        6.0      7.0      7.4
  Durable          -5.0      -21.1     -16.9        7.1      8.1     12.7
  Nondurable       -0.2       -8.9      -8.6        4.9      5.9      5.1


Subject Meta: 

Forum Categories: 


I hope you don't mind me tacking on some more "information at a glance" in the Productivity report. It just made sense vs. having another title that writes about the same thing.

I'm going to try to pick this apart in more detail, but the first thing that popped out at me is the correlation to the 2003 time frame. At that time, manufacturing as well as services was being offshore outsourced in mass. I do not know if this has anything to do with today's numbers, but the data point is suspect.

I don't mind at all.


What does an economist cost in India? We might think about that next time there is a vacancy at the Fed.

Frank T.

Frank T.

I would prefer China

as the location of origin for the next Fed chair...firstly then China could manage their Treasuries and exchange rate issues much more effectively and then it seems the skill level is more advanced in China (sic, i.e. China is literally capturing U.S. industries whereas India seems hell bent on simply trading people vs. developing their own markets).

When they really ramp up and start offshore outsourcing attorneys...that is when we would see more than lip service to global wage arbitrage...oh yeah, they would be on it like flies on .....

Bad for economy

I think that as this wage gap continues to grow it is going to further damage our economy, and this is going to show up in our currency. The dollar has already weakened a lot over the past decade, and I think it will continue to do so because our government keeps promoting the same policies, even though they have proven they do not work and ultimately lead to even more government intervention.

key is to prove it

to connect up GDP with the middle class, with your average Jack and show how squeezing average Jill will cause the GDP to go tumbling down the hill.

On productivity, most economists will claim it's just technological advances...

It's like they have brain death on productivity.

I think you just need to make a sufficiently compelling case

which in fact will be MUCH harder than "proof" (if economics wants to be a science it has to abandon that term. Science doesn't do proofs, only disproofs). Because "compelling" doesn't mean logically compelling, it means emotionally compelling to those whose emotions are hard-wired to money. IOW, the people the system selects for.

Exhibit A: bankers are doing their damndest to get back to double digit returns and the practices that got us here, which logic tells us is beyond stupid.

Exhibit B: The practices that got us here were widely touted by everybody from the 401K holder to the Washington Consensus neoliberals to the Chair of the Federal Reserve to Joe House-flipper to be a New Economy where the old rules no longer apply. Not by any logically compelling "proof", despite far more compelling disproof, but because they were making money hand over fist.

It is easy to make a logically compelling argument for wealth inequality for being a destructive force in a consumer driven economy. It is not so easy making one that is emotionally compelling to those in power.


I throw in plenty of rage, dismay and sarcasm in every post...but hey, I'm a math head. Since when doesn't science prove a hypothesis?

Einstein got to the heart of the matter

when he said "A hundred experiments will not prove me right; a single experiment can prove me wrong".

Science is all about pragmatism: it tests hypotheses, and if the hypothesis fails, modify or try again. If it doesn't, go with it until it does fail. A theory may pass with flying colors for centuries, until technology improves enough to test it to failure. Classical mechanics is a prime example: it passed innumerable tests, but now we know it is fundamentally incorrect. It is still useful, for many things far more useful than quantum mechanics, and it is still taught at the graduate level (and it would be nice if more people understood it) -- but we know it fails at a very basic level. Math does proofs, science doesn't. Science does what works, math ... well, sometimes the connection to the real world can get a bit tenuous.

This isn't about anger or cynicism: I'm a physical scientist, not a pure mathematician, I'm more interested in understanding (even if poorly) how things work than in beautiful elegant proofs of what may be completely unrealistic. No doubt a mathematician might pose it otherwise.


let's not

a = \frac {dv}{dt}

an argument.


proof as in with mathematical models, the theory as well as the evidence, the statistics.

So, let's switch to evidence.

"Evidence" much better

Scientific theories always have to be able to stand the test every time they're used. That's how it should be.

finally dawned on me

I think one truly can see these results in the equations. Look at the reads section (in EP menu), Gomory is a very good example of use of trade theory equations and showing that it's right there in the math, which says under certain conditions, certain variables trade is not a "win-win" as portrayed.

These are the types of things I am looking at. Take George Borjas, he shows repeatedly, with mathematics that increases in immigration will cause wage repression at certain levels.

So, this site has not delved into any heavy theory or use of equations and as you know I just added mimeTeX so we could....

So I think this is where the misunderstanding lies and from your comments I believe you can handle reading Gomory (he has a descriptive text but the second half of the book is completely mathematics).

Gomory looks interesting

I don't know Borjas or quant details. But even without going into gory math there are some things that look obviously wrong by inspection, or more accurately approximations not intended to be universally true. Here's a well known example:

Pr[T_A  < 1,T_B < 1] = \phi_2(\phi^{-1}(F_A(1)), \phi^{-1}(F_B(1)), \gamma)

The fixed \gamma is a clearly only a first order cut at deorthogonalizing the system. Fine when used as an approximation, but for a better model you need a functional form, and how are you going to get that, especially as it is certain to depend on external variables? And if you're going to be predictive, you must have a better model.

When this formula is used, are the F's assumed or measured? Measured will still have issues (time behavior), assumed can be really problematic.

There's no acknowledgment of distributions changing on different timescales. I haven't seen much quant math, but most of what I have seen is in the frequency domain. What little I've seen in the time domain appears to be Markovian (various Monte Carlo exercises), which we know cannot be generally true. There also seems to be an assumption that the system is frictionless.

Now none of these things are showstoppers if you're using the formula for analysis (as opposed to prediction). But it seems more like people didn't care about the details, weren't interested in taking the time to look too closely, because it appeared to work and they had to Make Money Now. Besides, if they looked too closely they might not have liked what they found. So to me the problem isn't in the math, it's in the money managers

you might put that below

the post on Copulas. We're having our own little chat here but I hope to let others read this stuff who do not "speak math" else, it's kind of like two people speaking a foreign language at a party.

that gamma as I recall is CDS values at the end of market..
which also would destroy the validity of a Copula.

Yeah, first time I looked over any of this stuff but I am like gee wiz, this crap isn't even valid for mathematical modeling.

I went through this here and if you search on Copula should find more...
but I was arguing solo because ya know who the hell can read that?

The F's in Li's gaussian copula

Are distribution functions. They are basically conditional probability marginal distribution of the time to default for security A and then security B. The original paper on the above formula and then these are something from advanced probability and statistics...

the "F" is usually denoted to describe probability distribution.

If mortgage A defaults, what's the chance of mortgage B defaulting and believe this or not, but gamma is based on CDSes which is a pure violation, they are not 1:1 and not correlated to actual default of underlying asset, their payout of default since they are traded..i.e. underlying asset goes bust, but one could payout 1 million CDSes on just that one asset and their evaluation is decoupled from the underlying asset value as well...

how GS made out like a bandit frankly. (unless of course AIG was allowed to not pay out 100% on their CDSes).

Actual paper here.

I cannot believe they are betting trillions in the great casino where only the "big guys" get a seat at the table on these things.

with the model itself

I believe it's static, i.e. a normal bivariate, gaussian distribution and a lot of this was so they could number crunch (right o, hey Wall Street, heard of super computers?)

hence, in a gaussian distribution, time invariant, the crash and burn effect would be in the tail and minimized in probability...(way down there or up there in the bell curve).

(say like a bunch of liar loans and mythical people and say overinflated prices and people signing up for $500k loans on $9/hr "salary, etc.) would still be way down in the tail and it's value minimized, regardless of it being more closely to the "norm" in a Gaussian distribution.

Anywho, I still say the thing that freaks me out more is the correlation coefficient being based on me, that's the nutso thing most of all.

I mean you're sitting out here in the peanut gallery, commenting on how our government is busy offshore outsourcing your career field and treating even the most advanced skills like disposable diapers..
you start reading about this stuff...

you go out and be a mathematical tourist and venture into the land of structured finance and find SHIT on the first pass read!

Ya know what I thought....geeks revenge on wall street.

I think I followed all of that.

Modern finance is based on a standard normal distribution. Anything that deviates from that is written off as an anomaly or noise. - Financial Information for the Rest of Us.

The Black Swan - Taleb

There should be a little image somewhere to the book, The Black Swan.

but I think it's pretty good in explaining some of these problems in English...

But the problem is he kind of condemns all things shaped like a Bell Curve, but trying to explain these types of mathematics in English is no trivial task...

So, one cannot say the entire modern finance system is based on Gaussian distributions, but one can say, a whole buttload of CDOs sure as hell were and the model sure as hell looks mathematically as well as empirically invalid.

I read the book and thought "what an asshole!" frankly, someone I just don't think would be too fun at parties...
kind of goes on big rants against the machine and to me, turns into some of the noise itself...
but I do not think there is a better book (assholes can have genius level insight) to explain conceptually some of these problems.

I remember I had a modern portfolio theory

class. I had a teacher (PHD candidate) who was doing his thesis on using an equation that tried to predict the movement of atoms and apply it to predict the movement of interest rates. We spend half a semester on this equation.

Now that I thought about this I will check to see if he ever finished his research. Something tells me it didn't work. - Financial Information for the Rest of Us.

quantum mechanics?

How about chaos theory instead? ;) Let me guess, anyone who disagreed with the fundamental premise got a F in the class.

Anyone bother to point out if he could make that sound even remotely sensible, he would get a salary of $250k base hired by Wall street...whereas if he just stuck with Physics he would be lucky to get a $40k Post Doc salary?

I need the class otherwise I would've dropped it

He clearly was using the class to get free help or input from us (like we had any idea). The equation took up three lengths of the chalk board. It was amazing piece of work. After he did that there were very few people left in the class.

I think he was bucking for that Wall Street gig probably w/AIG. Who knows he maybe one of the quants that created this mess. - Financial Information for the Rest of Us.

"Movement of atoms"

Atoms in a vacuum or things like dust motes, liquid molecules, etc? The latter is Brownian Motion (the equation the Langevin equation) and used as a classic example of statistical motion for all sorts of things. Might have relevance to the market in limited cases, but Brownian Motion is Markovian, and the market is not.


That is it. That is the equation that I tried to mentally block out for that past 16 years. What a flashback?

And look what I found:

Modelling of Short Term Interest Rate Based on Fractional Relaxation Equation (PDF File) - Financial Information for the Rest of Us.

Lumped comments

Will likely have more to say after reading the paper. But I don't have a problem with simplifications or even with using a CDS based correlation coefficient. The latter may be useful for model studies of sensitivity to that coefficient. I do have a problem with assuming it is useful in the real world, or as a method of pricing.

Economics is also a strange field in that there is no good way to test the models. Would another have done better? Would one predicting something quite different have been any worse? Is it just the presence of a plausible one that creates a boom?

Re: Black Swan. I agree that was hard to read. 150 pp of real material, 200 pp of I'm So Smart. Desperately in need of a strong editor, there.

look at the properties

of the gaussian copula per the correlation coefficient, that coefficient has to be 1:1, linear and the CDSes by their nature are not, they are not a 1:1 relationship at all to market values of current defaults. i.e. it's a violation of the properties which would make the Copula valid.

Artifacts of Despair

I think it is nothing less than shameful that the mainstream media is having one green-shoot-gasm after another over items that are being driven by the despair of the American people:

"Lower" unemployment - Driven wholly by hundreds of thousands of people having exhausted UI benefits and dropping out of the workforce.

"Better than expected" earnings - Driven mostly by layoffs and cost cuts, effectively shrinking the company!

Productivity - Again, driven by layoffs.

We are through the looking glass folks. A stock market that is a Generational Ponzi Scheme right in front of our face - borrowed TARP funds laundered into the market through the back door. Then the not-real market goes up based on not-real good news. All the while the media crowing over stats that can't exist without the suffering of the people. We are though the looking glass.

word of caution from Dean Baker

productivity #'s erratic around turn arounds.

But I also don't see any graphs and of the ones I have seen....productivity numbers didn't look that crazed during recession bottom calls and this report does.

Might be a good data point to examine with some graphs and regression analysis.

A good example of a mainstream economist ignoring data

And yet these numbers are important to those with alternative values to profit- for they tell a tale of fewer workers doing more work for less pay- EXACTLY the thing that causes a greater gap between the rich and the poor (which apparently, isn't important to Dean Baker because like most economists, he's tenured and could care less if people are starving).

Maximum jobs, not maximum profits.

Maximum jobs, not maximum profits.