September 04, 2003


This was not unexpected, but it is still amazing to see it in print:

The Bureau of Labor Statistics of the U.S. Department of Labor today reported revised productivity data-as measured by output per hour of all persons-for the second quarter of 2003. The seasonally adjusted annual rates of productivity change in the second quarter were: 7.2 percent in the business sector and 6.8 percent in the nonfarm business sector.

Posted by DeLong at September 4, 2003 09:03 AM | TrackBack


How does Ed Prescott interpret this data? Surely the big positive technology shock should be producing big employment gains, unless there's a huge income effect on labor supply. And it's not like labor market regulation has increased in the last 3 years so he can't blame another New Deal for the employment declines.

Posted by: P O'Neill on September 4, 2003 09:31 AM

Interesting that you have been so right about high productivity numbers and moderate growth of demand and GDP resulting in less demand for American labor. But, for much of the economic press this relation is a complete mystery.

GDP grows at 3.1%, productivity grows at 7.2%, and there is Less need for labor. Perhaps if you write again and again in a year or so the relation will sink in.

Demand recovery has been below the post WWII average since the end of the recession, despite the 3 tax cuts and increase in military spending since 2001. Productivity growth since the end of the recession has been above the post WWII average. Well, we can always cut taxes for the rich again. Duh.

Posted by: anne on September 4, 2003 09:34 AM

Here's a productivity question that perhaps Prof. DeLong or other posters can answer: A couple of days ago on NPR an International Labor Organization representative, discussing an ILO study on productivity in different countries, said that the U.S. had higher labor productivity overall than Europe, but that the difference was due to Americans' working longer hours than Europeans and that when you took into account Europeans' shorter workweeks and longer vacations, it was Europe that actually had higher productivity. Huh? I thought productivity was value PER HOUR (OR OTHER UNIT) WORKED, so the measurement shouldn't be affected by the NUMBER of hours worked. Was the ILO rep (who had an accent) saying something infelicitously in English? Or was the ILO study measuring something other than "productivity?" Or what?

Posted by: Richard Riley on September 4, 2003 09:50 AM

"Surely the big positive technology shock should be producing big employment gains."

Anne explained this, high productivity growth and slow demand growth result in no need for additional corporate workers. Also, competitive production capacity is rapidly expanding in China and India. We are growing too slowly to cut into unemployment.

Posted by: jd on September 4, 2003 10:01 AM

Richard Riley,

I have seen reports of the study you mention. I believe the first statistic was dollars of output per worker, in which the US leads. When restated as dollars of output per worker-_hour_, Norway leads.


Posted by: John Casey on September 4, 2003 10:27 AM


Yes, that's right. The theme on productivity coming out of Brussels over the past year and now out of the ILO (Rome, right?) is that productivity leadership depends on the unit of measure for work. They chose a different unit and get a different answer.

Among the responses to the Groshen and Potter post were some that said the results from Groshen and Potter show the notion of an "output gap" to be nonsense. I'm not sure why that would be, but the output gap gets wider, all else equal, as productivity rises. Today, Fed Governor Bernanke made a pretty strong statement in favor of the "output gap" view of the world, saying that the Fed might need to ease further if all output growth is the result of rising productivity, rather than employ gains. (Two-year Treasuries took off like a shot, but that's beside the point.)

Posted by: K Harris on September 4, 2003 10:32 AM

Different denominators, different answers. Fair enough. (And probably a worthwhile international comparison, actually.) Thank you John Casey and K Harris.

Posted by: Richard Riley on September 4, 2003 10:44 AM

"Fed Governor Bernanke made a pretty strong statement in favor of the "output gap" view of the world, saying that the Fed might need to ease further if all output growth is the result of rising productivity, rather than employ gains."

Just so! There is an output gap, and the idea that was is needed is ever more job losses to allow for new production structure is absurd. Remember, there is ample high level production capacity in China and India that American corporations can draw on. There is too little demand. We are simply not growing fast enough to absorb global capacity and high productivity growth and add to employment in America.

Posted by: anne on September 4, 2003 11:00 AM

Can I ask a naive question? I believe I understand Anne's point about how productivity growth and demand growth inter-relate. But we've been seeing reports of this extraordinarily high productivity growth for quite some time now.

The naive question is, is this real? Is it a real thing that's going on in the way people are working now, or is it just a numerical artifact? If it's real, how can it be happening quarter after quarter in what I remember as the absence of really significant capital investment?

In other words, what is this number really telling us?

Posted by: Altoid on September 4, 2003 11:22 AM

I the late 90s we were awed by seemingly impossible numbers of economic expansion. It turns out, there was a lot of corporate book cooking. Will we find more errors sometime in the future?

Posted by: bakho on September 4, 2003 11:38 AM

No, because that was during the Evul Librul Klinton MalAdministration. Now we've been blessed with our God-Sent Dear Leader, who has restored Honor and Responsibility and Dignity to the White House.

Therefore, we can absolutely trust any and all numbers coming from the White House or businesses (current numbers, that is - old numbers are obsolete, and only commie mutant traitors would even remember them).

In fact, to not trust these pronouncements is a sure sign of IslamofascistSympism, ObjectivelyProSaddamiteism, and DoublPlusUnGoodAmericanHateism.

As a matter of fact, bakho (or should I call you 'Saddam'?), the mere raising of such 'questions' is a sign that You Hate America.

Posted by: Barry on September 4, 2003 12:01 PM

My employer improved productivity dramatically by eliminating most of the R&D department. No immediate effect on revenue, major effect on staffing. Viola! I wonder how much of these numbers reflect short-term decision making that will ultimately be unsustainable.


Posted by: msw on September 4, 2003 12:07 PM

I keep thinking--in my relativel naive uneducated (economically, that is) fashion that productivity has been growing as a result of a recovery from the overcapacity of the bubble. That is to say, isn't it possible that not only was there overinvestment in capital resources, there was an overinvestment in labor? As unnecessary jobs are shed and more efficient labor allocation occurs, wouldn't that necessarily mean productivity would go up?

That guess is probably quite naive and wrong in fity different ways, but the general idea is something that I intuitively feel is correct...that is, that the abnormally high productivity gains we're seeing are an artifact of the combination of the bubble, *and* of the lagging effects of the technological changes that supposedly was the rationale for the bubble. Maybe we're getting those benefits sort of "all at once" as all the much needed restructuring has finally occured. The gains will go away, but since the actual economy is growing, those gains will result in increased demand down the road and the unemployment situation will work itself out. I mean, productivity can't grow at this rate forever, can it? Not unless something truly radical has happened technologically.

Posted by: Keith M Ellis on September 4, 2003 12:10 PM

"If productivity gains are real, how can it be happening quarter after quarter in what I remember as the absence of really significant capital investment?"

The reason is that there was always significant capital investment being made. Bard DeLong has shown on a number of occasions that investment in technology has been strong from 2000 on. Prices for technology products have been falling and purchases of the goods rising, so it may look as though there is little investment because total spending has not increased rapidly.

Deceptive. We have added amply to our technical equipment, but with falling prices we have barely changed our budget. We are investing and adding to productivity, but not paying ever higher prices for equipment. By shopping about, we actually get all sorts of added prices breaks.

The productivity gains are real, the continued investment is real, the ability to produce more goods with the same or fewer workers is real. Demand is too low.

Suppose we spent the kind of money needed to build our infrastructure. That would surely boost demand, but there is little chance now. Demand really is too low. There is an output gap.

The "need for re-forming labor market" folks are wrong, and I believe thay have been wrong about Japan.

Posted by: anne on September 4, 2003 12:32 PM

No problem. Who needs employment when there are always more tax cuts?

Bush Firm on Tax Cuts as Remedy for Economic Languor

President Bush's reiteration that temporary tax cuts be made permanent today made it clear that he is not daunted by predicted huge budget deficits....

Posted by: jd on September 4, 2003 12:53 PM

"People are more likely to find work if businesses and their workers can be certain that the lower tax rates of the last years will stay in place," Mr. Bush told an audience at the Greater Kansas City Chamber of Commerce. "Today, you don't have that confidence they'll stay in place. And there's a good reason: because under the laws that were passed, tax relief is set to expire."

Posted by: jd on September 4, 2003 12:55 PM

Pardon someone not well-educated in economics, but isn't high productivity just the other side of the coin to high unemployment? At least in this instance?

Posted by: rea on September 4, 2003 01:49 PM

I have some rather scattered thoughts on this report. First output in manufacturing declined by 2.4 percent and hours worked by 5.9 percent. Overall productivity rose in this sector by 3.7 percent. Not the highest Iíve seen but respectable since manufacturing productivity tends to shrink when its output shrinks. Furthermore productivity in manufacturing has in my experience usually been higher than the aggregates in the Business sector. This is not the case here. Second the, the Business, and Nonfarm Business Sector, productivity are measured differently from manufacturing. For manufacturing the numerator is the manufacturing index prepared by The Federal Reserve, while the Business/ Nonfarm Business section uses Real Gross Domestic Product from the BEA. This makes the two measures not directly comparable, but it does raise the question why one is so eye popping and the other is so lackluster. Output in the Business sector rose 4.2 percent and hours dropped by 2.7 percent while productivity increased 7.2 percent.

One has to ask where these GDP numbers increased and by how much. The biggest increase was a 25.5 percent increase in federal government expenditures, and this reflected a 45 percent increase in national defense. So we would have to say that a large portion of this productivity increase was due to the war in Iraq. (Or a remarkable increase in governmental productivity). But then why would that increase productivity. Are the armed forces that much more productive. Well, that is not the case, quite. You see we have all these weapons that are very expensive and can be blown up with relatively few people. Think cruise missile, then think refrigerator. Hopefully lots of dollars entered into your mind for the cruise missile. Of course these are also produced by workers but they do so in an almost handicraft form of production. So part of this rise in productivity reflects a whole lot of government expenditures, for a small amount of very expensive hardware.

There was a also large increase in durable goods production in the GDP accounts (25 percent) while the manufacturing index from the Fed shows a 3.3 percent decline in durable goods. These might be measured by different organizations, and there might be some counting differences in methodology, but the difference should not be that great, and it should not move in totally different directions. I believe one, or the other, has to be wrong and will be revised. I canít really say which. My experience has been that the Fedís manufacturing index is revised less than the quarterly GDP accounts and we are still in an early stage of the GDP accounts. Still a revision on the order of 28 percent? Something very weird is going on here.

Third there is a troubling question of why? On the one hand manufacturing so-so productivity increase, on the other hand business 7.2 percent increase. In the normal course of things one would expect to see an increase in investment in productivity enhancing "stuff." Like machines and software. Yet for the most part this has not been occurring (although these did increase in the second quarter). If services, rather than manufacturing, are to take a lead in productivity increases then one might expect to see bigger jets, more fiber optic cable, more software, and computers sold. This really hasnít occurred. So whatís up doc? I suspect a statistical aberration.

Posted by: Lawsrence on September 4, 2003 02:20 PM

The truisms at work here: Fewer workers work harder. And scared employees work harder.

Posted by: nick sweeney on September 4, 2003 03:50 PM

If I understand Lawrence correctly, the non-farm number comes from dividing GDP by hours worked. Thus a huge GDP jump caused by military spending in the numerator. And if the time of soldiers is not factored into the denominator, we get a real anomaly, a numerical artifact.

What happens if "hours worked" gets seriously under-reported because much is under the table or mandated to be off the clock? And how does one figure the productivity of someone who doesn't put in billable hours, like, say, a clerk?

I see where productivity goes way up if a company installs or upgrades its order-processing system to cut turnaround by 2/3, or a system that goes direct from design to manufacture. On the other hand, I know people who fly a lot. I drive. I go to stores. Whatever increases real throughput and productivity, there are vast sectors of the American economy that don't have it. So what's going on, really?

Posted by: Altoid on September 4, 2003 07:34 PM

"The naive question is, is this real? Is it a real thing that's going on in the way people are working now, or is it just a numerical artifact?"

Not so naive. Not so real - at least as regards investment.

From "Smoke and Mirrors" in the FT:

"...the largest increase in business investment in three years, revealed in the second-quarter US gross domestic product figures, is also not all it seeems. Stephen Roach of Morgan Stanley notes nearly half the increase is due to a 54 per cent annualised surge in computers and peripheral equipment. Actual spending on computers in current dollars rose by $6.3bn in the second quarter - about a sixth of the reported real increase of $36.4bn. The difference is due to statistical adjustments to take account of product improvements and to rebase the figures to 1996 exchange rates. The adjustments may aid comparisons in real terms, but Mr Roach concludes: "This smacks more of statistical amplification of a small change in an industry that is still labouring." Much the same comment could be applied to the earnings figures."

"Who is so deafe or so blinde as is hee
That wilfully will neither heare nor see?" - John Heywood, Proverbes.

Posted by: Pooh on September 5, 2003 12:48 AM

I really dislike this arithmetic--which is all that it deserves to be called--that suggests that employment is shrinking because productivity growth is accelerating and growth (for some reason that is beyond me) is a fixed number. An intelligent group of economically literate people like this should detect a bad smell when it is implicitly argued that productivity is in any way a bad thing.

I also really dislike the idea of the output gap being applied to a rapidly restructuring economy like that of the United States. (See yesterday's superb NY Fed citation below on this blog.) This is not a one sector economy producing homogeneous output with homogeneous labor; it is a complex, multisector economy producing many traded goods, and it is experiencing a productivity boom that is very unevenly distributed across sectors. Sectoral adjustments are a good thing, not a bad thing, when market forces are shaping the dynamic. None of this has anything at all to do with macro policy, any more than it did in the mid-1990s. Forget the output gap. It's a useless idea.

There. I feel better now.

Posted by: Jim Harris on September 5, 2003 07:46 AM

Stephen Roach had some interesting comments on productivity this morning. As a frequent reader of Brad's blog I would love to get his and others' reaction to it. Here's a quote from the end:

"I fully realize all this is heresy. But I fear that in our rush to rationalize the wisdom of our ways, we have lost sight of the pitfalls of the Information Age. The white-collar services sector continues to haunt me in this regard. In days of yore, service companies were quintessential variable-cost producers. Their main assets were people, whose headcounts and compensation could be adjusted in close conformity with the ups and downs of the business cycle. But then service companies discovered IT. They spent massively in building an entirely new infrastructure ó not just hardware, but also software and legions of IT-support workers. In going down this road with reckless abandon, the service sector has been transformed from variable- to increasingly fixed-cost producers. Little wonder the mild recession of 2001 gave rise to the worst earnings carnage in modern history. Thatís what lurks on the downside of an IT-enabled economy.

Economics is a big puzzle. There are always a few pieces missing as we try to peer into the future on the basis of what we know about the past. History tells us that productivity is the Holy Grail of economic prosperity. But that was a history that was written for a very different economy ó largely a blue-collar, manufacturing economy. When it comes to services productivity, I am afraid that I continue to believe that weíre guilty of exaggerating the miracles."

Full text:

Posted by: Peter Levinson on September 5, 2003 09:01 AM

I'm with Jim. I'm enough of a simpleton to be very happy about such substantial productivity increases. This kind of productivity growth is one of the few unmitigated Good Things that economists can agree about.

Doesn't anybody remember the misery of the 1970's anymore, with its raging inflation and anemic productivity growth? Productivity growth does not and can never be the reason for job losses (in the aggregate). In fact, it's the opposite--sustained productivity growth is the golden key that opens up new economic opportunities and can raise the standard of living for all wage-earners.

Posted by: Daniel Calto on September 5, 2003 01:19 PM

J. Harris and D. Calto: It really *was* a naive question because I'm not an economist. And I'm far less interested in the moral question (is it a Good Thing) than I am in the questions 1) whether it's really going on, and 2) if it *is* really going on, what will be the real-world consequences, especially short- and medium-term, for real people who need to earn a living and who might not find it so easy to retrain?

This last question might make it obvious that I do not believe that "the market" is an ultimate good, but rather more like an outer bound-- it represents what some people would want to do if they didn't have any restraints to bind them, therefore what they will try to do, and it represents a way to price the value of some things in the short term. It needs to be recognized, but not worshipped.

Economic phenomena happen in the real world. If enough people get frustrated because they can't hold jobs and don't feel they can retrain, or don't feel that retraining is worthwhile (what would it feel like to lose a $17-an-hour machinist's job-- "only" $35K if multiplied out, much less than most academics make-- and retrain as a Walmart greeter?), there have historically been very ugly political consequences. People need to think about that. That's my point.

Posted by: Altoid on September 5, 2003 09:53 PM

An addendum primarily in reply to Daniel Casto. I don't disagree in the abstract that ". . . sustained productivity growth is the golden key that opens up new economic opportunities and can raise the standard of living for all wage-earners." But I have to wonder whether it works the same way in the current global situation where, after about 500 years, we are not bringing new areas into production for the first time but rather are exporting functions integral to our economy into different currency zones.

Absent that, historical experience in this country is that productivity increases multiply jobs and opportunities. But, speaking as a skeptic about the claim that "it's different this time," I still have to ask whether we've ever been in this situation before. At this late hour in the East I can't readily think we have.

Posted by: Altoid on September 5, 2003 10:14 PM
Post a comment