September 08, 2003

2002-2003 Productivity Growth

Next February the Commerce Department's Bureau of Economic Analysis is going to release its first estimates of production and productivity for the year 2003 relative to 2002. Next February's data release is highly likely to show a 2002-2003 year-over-year labor productivity growth rate of 4.0% or so--less than the 5.4% year-over-year of 2002 relative to 2001, but still faster than the growth rate in any year in the late 1990s boom.

The rapid rates of innovation and efficiency growth of the new economy are alive and well, but buried underneath slower-than-desired demand and a truly lousy labor market.

Posted by DeLong at September 8, 2003 09:33 PM | TrackBack

Comments

professor...

why do you get such a smiley-face for productivity?

it seems to me (a novice in econ) that increasing productivity can only eliminate jobs, given a fixed (adjusted for population inflation) increase in consumption and a growing increase in automation.

when taken to it logical conclusions (extreme automation, human labor elimination) doesnt productivity growth invalidate capitalism completely?

how do economies work when the human-performable-job/human-on-the-planet ratio drops near zero?

this is an issue i have nightmares about...and i have never taken an economics class...

please help

Posted by: sampo on September 8, 2003 10:55 PM

Brad,
I have a dumb question for you about productivity statistics:

If Jane leaves BigCorp to become self-employed, and then sells BigCorp the exact services that she was providing when she was an employee, does measured productivity rise? I worry that it does, because BigCorp's sales have stayed the same, but Jane might no longer be counted in the payroll employment survey.

One of the puzzles of the labor market is that the household survey, which treats the self-employed as working, is showing less job losses than the payroll survey, which has a harder time picking up self-employment (does it even try?). Perhaps the measured productivity growth we're seeing is related to that puzzle.

I feel dumb, because I don't know how the statistics are calculated.

Posted by: Arnold Kling on September 9, 2003 05:17 AM

Arnold, from the BLS itself:

"The primary source of hours and employment data is the BLS Current Employment Statistics (CES) program, which provides monthly survey data on the number of jobs held by wage and salary workers in nonfarm establishments. The CES also provides average weekly paid hours of production and nonsupervisory workers in these establishments. The Office of Productivity and Technology estimates average weekly paid hours of
nonproduction and supervisory workers. Weekly paid hours were adjusted to hours at work using the BLS Hours at Work survey, conducted for this purpose. Data from the National Compensation

Survey are used for recent years. Data from the BLS Current Population Survey (CPS) are used for farm labor, nonfarm proprietors, and nonfarm unpaid family workers. Estimates of
labor input for government enterprises are derived from the CPS, the CES, and the National Income and Product Accounts (NIPA) prepared by the Bureau of Economic Analysis (BEA) of the Department of Commerce."

Have a nice day,

Edward

Posted by: Edward Hugh on September 9, 2003 06:30 AM

Arnold again:

Sorry, I missed this part:

"Hours data for the labor productivity and cost measures include hours for all persons working in the sector-wage and salary workers, the self-employed, and unpaid family workers."

http://www.bls.gov/news.release/prod2.tn.htm

This, I think makes it clear that the self employed are included, whilst they are excluded from the establishment survey. So I'm afraid Jane of BigCorp doesn't help us.

Robert Gordon here:

http://faculty-web.at.northwestern.edu/economics/gordon/WEFTEXT.pdf

does.

The effect seems real enough, and seems related to mismeasurement of intangible human capital in the past, driven forward by current 'learning by doing'. Unlike to Gordon I think just how long this present technological wave has in front of it is anybody's guess.

http://bonoboathome.blogspot.com/2003_08_31_bonoboathome_archive.html#106277785510014371

Maybe it's also interesting to remind everyone just what labour productivity stats actually tell us (same BLS source):

"These productivity measures describe the relationship between real output and the labor time involved in its production. They show the changes from period to period in the amount of goods and services produced per hour. Although these measures relate output to hours at work of all persons engaged in a sector, they do not measure the specific contribution of labor, capital, or any other factor of production. Rather, they reflect the joint effects of many influences, including changes in technology; capital investment; level of output; utilization of capacity, energy, and materials; the organization of production; managerial skill; and the characteristics and effort of the work force."

Posted by: Edward Hugh on September 9, 2003 06:57 AM

Stephen Roach wrote a good article on productivity last week and I pretty much agree with him. Measuring productivity in a service based economy is very tricky: measuring the productivity of a hairdresser, lawyer or a programmer is inherently difficult. I would also like to know how truthful companies are in reporting work hours figures etc. to the Bureau of Labor Statistics. And is the BLS audited or monitored by an independent body ? The BLS has been changing its methodology several times in the past decade.

Posted by: Nescio on September 9, 2003 07:04 AM

Sampo,

Until the perfessor responds, my guess is that he gets a smiley-face for productivity because it is the source, other than outright shortages of labor, of pay increases for workers. It is also the source, other than new finds of natural resources, of rising standards of living. These benefits are not apparent now because of other circumstances in the economy. But perhaps the good perfessor will speak for himself.

Posted by: K Harris on September 9, 2003 08:46 AM

If a person works 45-50 hours a week, but gets paid for 40 doesn't she look more productive than she actually is? I assume production hours are measured by payroll hours. And lots of folks are doing just that in order to keep their jobs.

Posted by: Michael L on September 9, 2003 10:10 AM

I would like to see your answer to sampo's question, professor. I am also an economics novice. But, I see things like he does. So puhleeeeeeeeze enlighten this worrying soul.

Posted by: eyesopen on September 9, 2003 10:30 AM

I would like to see your answer to sampo's question, professor. I am also an economics novice. But, I see things like he does. So puhleeeeeeeeze enlighten this worrying soul.

Posted by: eyesopen on September 9, 2003 10:33 AM

I would like to see your answer to sampo's question, professor. I am also an economics novice. But, I see things like he does. So puhleeeeeeeeze enlighten this worrying soul.

Posted by: eyesopen on September 9, 2003 10:34 AM

I would like to see your answer to sampo's question, professor. I am also an economics novice. But, I see things like he does. So puhleeeeeeeeze enlighten this worrying soul.

Posted by: eyesopen on September 9, 2003 10:36 AM

I would like to see your answer to sampo's question, professor. I am also an economics novice. But, I see things like he does. So puhleeeeeeeeze enlighten this worrying soul.

Posted by: eyesopen on September 9, 2003 10:37 AM

The inflation rate used in Q1 was 2.4%. The rate for Q2 was .8%. Amazing! This is like somebody slammed on the brakes. This change in inflation rate is a big part of 2Q GDP and productivity calculations.

Greenspan is probably shaking in his boots. This could be the big deflation boogieman he uses to roil the bond market. Or maybe King George said I need some better numbers.

So which is it? Looming deflation or booming GDP? Its pretty hard for the economy to be both at the same time.

Posted by: David on September 9, 2003 10:41 AM

"it seems to me (a novice in econ) that increasing productivity can only eliminate jobs, given a fixed (adjusted for population inflation) increase in consumption and a growing increase in automation."

Why a fixed increase in consumption? As productivity rises and it becomes possible to produce more goods, people will consume more goods. This is known as Say's Law and requires only that people prefer more to less. It is not true at all times, but it surely is in the long run.

"when taken to it logical conclusions (extreme automation, human labor elimination) doesnt productivity growth invalidate capitalism completely?"

If automation goes so far that everyone is completely satiated by the consumption of goods produced without requiring any human labor, why then just relax, lie back, and enjoy yourself. You would have nothing to complain about - remember, you're satiated.

Posted by: Daniel Lam on September 9, 2003 10:58 AM

Sorry about the many entries. It said that it was unable to contact the site, so I kept trying to post that. Weird.

Posted by: eyesopen on September 9, 2003 11:32 AM

Hate to brake it to you - there is one sector where productivity in the US seeminlgy is among the worst. Look at the US healthcare system; if they actually perform some work for the large sums of money they receive there, why don't they stop infant mortality from being that high!?

http://blogofpandora.blogspot.com/2003_09_01_blogofpandora_archive.html#106313163627799020

Posted by: Mats on September 9, 2003 11:54 AM

1Q inflation was 2.4%, 2Q inflation is .8% Did you notice the price of anything dropping? I didnt.

But, when inflation drops, then GDP is adjusted with a smaller inflation adjustment. A large part of the 3.1% 2Q growth is this change in inflation.

If inflation had really dropped to .8% that would mean an accelerating decrease in inflation. Something that Greenspan has mentioned in his talks about deflation. Something that Greenspan has been trying mightily to not let happen.

And naturally if you get more GDP by applying a smaller inflation factor, you also get more productivity. So are the GDP and Productivity numbers real? This game is only good for one Quarter, so we will know soon.

Posted by: David on September 9, 2003 02:46 PM

"As productivity rises and it becomes possible to produce more goods, people will consume more goods."

What? I'm sorry, but did this actually make sense to someone? An increase in supply CAUSES demand??? What?

Posted by: Dave Johnson on September 9, 2003 03:16 PM

"If automation goes so far that everyone is completely satiated by the consumption of goods produced without requiring any human labor, why then just relax, lie back, and enjoy yourself. You would have nothing to complain about - remember, you're satiated."

Is this from the same logic that says supply creates demand? If automation takes your job, you certainly are NOT "satiated," you're starving to death!

Posted by: Dave Johnson on September 9, 2003 03:21 PM

the claim that supply creates its own demand is known as say's law, after the early 19th century swiss political economist jean-baptiste say. the point is that to produce a final comsuption good one must first extract raw materials and pay workers to do so, the produce the good through several layers of production, including the production of capital goods, and pay the workers to do so, so that buy the time the consumer good arrives at market there will be a monetary demand to pay for it. this claim depends on the long period equilibrium assumption of classical political economy and a static view of the production hierarchy. it was widely ridiculed at the time and i don't think it is particularly true. for one thing it would depend on the wage rate as a share of the total surplus product as well as the available extension of credit. as applied to modern industrial/post-industrial productivity, it is one of those long run things for when we are all dead.

Posted by: john c. halasz on September 9, 2003 03:45 PM

"What? I'm sorry, but did this actually make sense to someone? An increase in supply CAUSES demand??? What?"

Let's say there's just you and me. Your land is suitable only for raising cows and mine only for growing potatoes. Trading with each other, we obtain both meat and potatoes which we cook to obtain a balanced diet of burger and fries. Most of our labor is expended in farming, with some effort expended on the cooking.

Now one fine day you, being the smarter of the two, invent a magical machine which automatically produces both beef and potatoes with minimal manual effort. Naturally, with this machine giving you all the meat and potatoes you want, you don't need my potatoes anymore: my job supplying potatoes to you has been destroyed by automation. So now where do I get my meat from? Do I starve to death?

Luckily, no. Now that you don't need to engage in strenuous farming to obtain your dinner ingredients anymore, your main daily grudge is the cooking. Seeing poor me on the street without no job and no food, while you have a machine that churns out free food, you hit upon the perfect solution: you hire me to do all your cooking and give me food in return.

In this scenario you capture all the gains to automation while I am the relative loser, but note that I don't become totally useless and starve to death. An increase in supply has indeed created its own demand: higher productivity in beef and potato production led to new demand for cooking services.

Moral of the story: automation can cause displacement by forcing a potato farmer to become a cook, but it does not cause permanent mass unemployment.

Posted by: Daniel Lam on September 9, 2003 04:26 PM

your answers are terribly unsatisfying

theres a scandal in productivity numbers....i wish i had the knowledge to piece it all together.

Posted by: sampo on September 9, 2003 04:55 PM

"it seems to me (a novice in econ) that increasing productivity can only eliminate jobs, given a fixed (adjusted for population inflation) increase in consumption and a growing increase in automation."

Alright, maybe history wil be more convincing than theory. Look at the graph of "Growth in Real World GDP per Capita, 1000-Present" prepared by our kind host at

http://www.j-bradford-delong.net/TotW/berk_fac_lunch/lunch_Berkeley.html

Say it is the end of the 18th century, and you have a crystal ball showing you the path of world productivity growth over the next 200 years. Assuming a "fixed increase in consumption" per capita of roughly the same magnitude as the history of world consumption growth since 1000, you would conclude that there would hardly be anyone left on the planet with a job by 2000.

Posted by: Daniel Lam on September 9, 2003 06:06 PM

daniel lam:
the trouble with your example is the difference
in the ratio of exchange between beef and potatoes
and that between the potato and beef monopolist and the provider of cooking services. you can not ignore the distributional effects of production changes/productivity increases. especially since distributional outcomes will effect the future course of growth in both output and productivity in an economy. of course, this is a complex and systematic question and is not nearly so simple and linear as relatively equal income distribution
equals higher aggregate demand equals greater capital intensities equals higher productivity rates. but distributional issues are a paramount consideration, not just with respect to market position in exchange relations, but with respect to the non-market (sic) issues of selection and development of both business organisation and production technology. by the way, since, by your briskly and brusquely reductive style of argument, i can tell you are an economist, may i point out that your argument betrays telltale rhetorical signs of special pleading: the attribution of production gains to the "smarter",
aligned with the moral sentiments of pity and generosity conveniently identified with enlightened self-interest. "smart" is by no means synonymous with "right", neither cognitively, nor morally. at the top of the all time productivity hit parade is surely the former north carolina truck driver who came up with the shipping container; the idea was so simple that it is highly improbable that no one thought of it before, except that apparently no one did.

Posted by: john c. halasz on September 9, 2003 07:38 PM
Post a comment