The End of the Clinton-Era Productivity Growth Boom?

Source: Congressional Budget Office

John Schmitt and Dean Baker fear that the Clinton-era productivity growth boom is at an end. I think it is possible, but unlikely, and way too early to start panicking. Note the word "arguably" in their first sentence and the phrase "is sustained" in their second:

Comment is free: The real economic crisis: [A]rguably an even more fundamental problem facing the US economy: the sharp deceleration in productivity growth since the middle of 2004.... [I]f sustained... the deceleration in US productivity growth since the second half of 2004 is striking by historical standards. Between 1947 and 1973, the golden age of postwar capitalism, productivity growth averaged about 2.8% per year in the United States. At that pace, the output of the average worker was set to double about every 25 years, allowing roughly comparable increases in national living standards. From 1973 through 1995, however, productivity growth took a nosedive, with the average rate dropping to just 1.4%. At this lower rate, average worker output would take about 50 years to double, implying far slower progress in living standards.

From the mid-1990s on, however, official productivity growth again accelerated rapidly, returning to a 2.9% rate reminiscent of the golden age. Quite suddenly, though, in the second half of 2004, productivity growth dropped sharply. From the third quarter of 2004, productivity growth rate, at 1.3% per year, has not even managed to match the 1.4% growth rate of the productivity bust of 1973-1995.

Some productivity optimists argue that the downturn is a blip. But, this is a blip that just turned three years old - fully one-third the length of the nine-year 1996-2004 boom that the optimists champion. Other optimists dismiss recent performance as cyclical - related to the downturn in the US economy.... The productivity numbers are likely even worse than they look. The most important reason is that the official productivity figures don't handle the rapid depreciation of new technology very well...

Comments

[Bruce Wilder]

The division, here, should not be characterized as a division between "optimists" and "pessimists", but between those who attribute productivity to magic and those, who look to some rational analysis and evidence.

Among economists, magic clearly has the upper hand, and that's a very sad circumstance.

Those, who are inclined to rational analysis and evidence are not likely to just mumble something about "technology" and computers. They might notice the contrast between Clinton policy and Bush policy on deficit spending, savings and investment, and the contrasting implications for (inferred) per capita capital stock. That would not be the end of the analysis, of course -- just a beginning, but a beginning founded on some analysis of production and on evidence of something.

[2slugbaits]

"Other optimists dismiss recent performance as cyclical - related to the downturn in the US economy"

Huh? If productivity is defined as output per labor hour, then wouldn't we expect to see rising productivity numbers in a weakening economy? Or are businesses in the habit of letting go of their most productive employees first and retaining marginal workers?

[JO'N]

I'm not an economist, but the one I don't play on TV predicts that productivity slowdowns would more-or-less match slowdowns in wage increases. (After all, if economists claim that CEOs won't work as hard if they're not paid tractor-trailers full of cash and options, then I'd think that the same argument would apply to everyone else.) And, in fact, it can't be an accident that the eras of better productivity growth (1948-1973 and the 1990's) are exactly the eras when average wages were increasing robustly. Surely economists can't view this as an accident, but I can't recall ever hearing them link the two in this way. (Though that's perhaps a failure of my memory rather than a failure of economists.)

[me]

"Or are businesses in the habit of letting go of their most productive employees first and retaining marginal workers?"

In the case of the IBMs and Intels of the world they fire US employees, including the most productive since they ditch entire groups, Say programmers, and hire unproductive Indians. Now those unproductive Indians are costing about half of US programmers and with the rising rupee, productivity is falling. No surprise here.

All of these phony transfers for labor arbitrage and to avoid taxes and not being picked up in the official government statistics, but the numbers are finally coming home to roost.

[Bruce Webb]

Well I'll ask the question I always ask? Where are the labor effects?

I am not an economist but I track the labor series closely as a lead indicator for Social Security solvency and it has simply been wacky in the last couple of years with revisions galore.

If we look at the series as it appears today you do indeed see a sharp dropoff in mid 2003 and a sustained decline since.
http://www.bls.gov/ (click on the productivity dinosaur and scroll to the table).
But that sharp dropoff was not picked up initially, the table looked much different and much more positive right through Q2 2006, when the BLS dropped this on us in their quarterly press release:
http://www.bls.gov/news.release/archives/prod2_09062006.pdf
This for the then current quarter was disturbing enough:
"In total manufacturing, the change in productivity was revised down from a preliminary estimate of 3.0 percent."

For one thing it suggests that paying attention to the preliminary number is a fools game.
But this made me throw up my hands (from the Q2 preliminary)

"In the manufacturing sector, growth in output and productivity were revised down in both 2004 and 2005. Because hourly compensation was revised down more than productivity in 2005, the increase in unit labor costs was also revised downward. In 2004, a large downward revision to productivity eclipsed the downward revision in hourly compensation, and unit labor costs edged up 0.1 percent rather than falling 3.1 percent as reported June 1. In 2003, an upward revision in productivity combined with a downward revision in hourly compensation to yield a much smaller increase in unit labor costs than previously reported. For the entire 1987 to 2005 period for which manufacturing data are available, the average annual rate of productivity growth was unchanged at 3.7 percent per year. The revisions did lead to changes in recent periods; productivity growth from 1995 to 2000 was revised up from 4.1 percent per year to 4.7 percent per year, while the average rate of growth from 2000 to 2005 was revised down from 5.0 percent to 4.1 percent."

Maybe professionals are equanimical in the face of data sets that years after the fact can get adjusted +/- 25%. And what is the rate of growth for Q2 2006 reported today? 0.6%. I understand that this series is volatile from quarter to quarter, I can even understand how you could have a 25% swing between preliminary and final, though it bothers me, but this intratemporal stuff between years is a little much. "Did I say 3.0%? No I meant 2.6%. Oh except I proceeded to revise that down to 0.6%". And all of this from the same agency within an 18 month period.

Why were we not having this same discussion in 2005? Well probably because the numbers showed an entirely different picture then they did in Q2 2006 and now with Q2 2007 in the books. (Interestingly Q1 2007 was revised upwards from 0.2% to 0.7% between preliminary and final in the Q2 release.)

So lets say my faith in this series is not what it was. For my specific purposes the really poor revised numbers for 2005 were bad news indeed. Economists who know stuff tell me there has been a historic link between productivity and real wage increases and that moreover that link broke down some in the early 2000's, that whole 'rising inequality' thing. Given that and a series that currently (because who knows what it will look like after Q3 revisions) looks like this:
2002 4.1
2003 3.7
2004 2.7
2005 1.9
2006 1.0
shouldn't we be seeing strong employment or wage effects or both? Line workers should be getting hammered. Yet Social Security receipts, which would seem to be an excellent proxy for working class wages, came in ahead of even the optimistic model for both 2005 and 2006. We have a key econmic measure shrinking by 75% over a four year period and no one seemed to notice. It was only last year that people were crowing about how recent growth numbers were proof that Bush tax cuts worked. Not that I particularly agreed but at least they had some numbers on their side. Then.

I have more questions than answers here. At a minimum this series as it stands today is not exactly proof of the magic of tax cuts. And Social Security receipts continue to come in in line with projections.

[wood turtle]

I think productivity is a relatively meaningless number.