During the 2000s, as U.S. manufacturing was transformed by devastating job losses, prominent economists and presidential advisers offered comforting words.
The paring of the manufacturing workforce – it shrank by a third over the decade – actually represented good news, they said.
It meant that U.S. workers and factories had become more efficient and, as a result, manufacturing companies needed fewer people.
What happened to manufacturing? In two words, higher productivity, Robert Reich, former labor secretary in the Clinton administration, wrote in 2009.
The decline in U.S. manufacturing employment is explained by rapid growth in manufacturing productivity over the past 50 years, said Glenn Hubbard, former chairman of the U.S. Council of Economic Advisers under President George W. Bush.
But a handful of economists have begun to challenge that explanation, chipping away at the long-offered assurances that the state of U.S. manufacturing is not as bad as the jobs numbers make it look.
Instead, they say, its significantly worse.
The job losses, in their view, were caused more by the declining global competitiveness of U.S. manufacturers than more-efficient factories. The apparent productivity gains reflected in the official U.S. statistics have been miscalculated and misrepresented, they say, a view that has been at least partly validated by new research.
I bought into this idea for a long time that it was superior labor productivity that caused most manufacturing job losses, said Rob Atkinson, of the Information Technology and Innovation Foundation, a nonpartisan think tank. Then I began to dig into the numbers.
The arguments, which get a full airing in a report issued last week by the foundation, are being mounted as economists and politicians on the presidential campaigns debate what, if anything, to do to help the nations manufacturers.
Among the options are tax incentives, trade assistance and education credits. But the underlying question is whether U.S. manufacturing is as robust as some readings of the productivity figures suggest, or faltering amid global competition.
These numbers have been tossed about to say, Look how productive U.S. factories have been, said Susan Houseman, senior economist at the W.E. Upjohn Institute, coauthor of a paper with three Federal Reserve economists raising questions about the accuracy of the productivity numbers.
The reality is a lot more complex, and not as flattering.
At first glance …
As calculated by federal statisticians, the productivity growth of U.S. factories has seemed quite impressive. Between 1991 and 2011, productivity more than doubled, meaning that a single worker today produces what two did 20 years ago, according to Bureau of Labor Statistics figures.
Many economists have concluded that the loss of manufacturing work should be considered a success story. Just as farming became more efficient over the previous century and fewer Americans found jobs on farms, U.S. manufacturing is simply becoming more efficient, as economists such as N. Gregory Mankiw, chair of the CEA under Bush, and Austan Goolsbee, recently the CEA chair under Obama, have argued.
But as Houseman, Atkinson and others have pointed out, relying on that single number for manufacturing productivity is misleading and masks a far more complex reality.
For starters, as Houseman and her Federal Reserve colleagues have shown, the productivity gains that have been reported may be overstated because the statistics that the U.S. collects do not adequately reflect the changes that have come with globalization.
Calculating labor productivity depends on determining the value of U.S. manufacturing output and dividing it by the number of manufacturing worker hours.
But in a time when factories increasingly have turned to imported parts and outsourcing, it can be difficult to determine what is U.S. manufacturing output and what should be properly counted as output from a foreign factory.
Critically, Houseman and others have shown that the price-savings that U.S. factories have realized from outsourcing have incorrectly shown up as gains in U.S. output and productivity.
This bias may have accounted for as much as half of the growth of U.S. manufacturing output from 1997 to 2007, excluding computers and electronics manufacturing, Houseman and her co-authors have estimated.
The federal statistical agencies, which have helped fund Housemans work, agree the bias exists, though they say there might be other problems that are offsetting.
Figuring out where the productivity gains actually happened can be difficult, said Brent Moulton, associate director for National Economic Accounts at the Bureau of Economic Analysis, which generates the output figures used in calculating commonly used measures of labor productivity.
On the agencys website, a new message says the bias could account for overestimating the nations GDP and productivity by 0.1 or 0.2 percentage points.
A deeper issue
But there may be another, broader problem with the manufacturing output and productivity figures, too: Those numbers lump all manufacturing together when there are actually two very different trends afoot.
Between 2000 and 2010, computer and electronic products manufacturing output rose at a remarkable rate of almost 18 percent per year.
Over the same period, output in the rest of U.S. manufacturing remained roughly flat, according to BEA figures tallied by Houseman. Thats a dismal showing for a decade.
It is only when computer and electronic products are included that overall manufacturing output registers the impressive increases. Though it represents 15 percent or less of manufacturing output, the sectors strong growth makes the rest of U.S. manufacturing seem much more robust than it really is.
Moreover, as the critics point out, there are reasons to question what the remarkable growth in computer and electronics production really means.
For one thing, much of the nations production of computers and electronics has moved offshore. The number of consumer electronics shipped from U.S. factories dipped about 70 percent between 2000 and 2010, according to Censuss Current Industrial Report.
Moreover, at least some of the productivity gains shown in U.S. computer manufacturing reflect the increasing power and decreasing prices that come with innovation.
When a computer chip doubles in efficiency, that can turn up in a doubling of output and productivity in computer manufacturing. But that is not what is ordinarily thought of as manufacturing efficiency.
It is innovation that makes it look like theyre manufacturing a whole lot more in the U.S. than they really are, Atkinson said.
Why the job losses?
Finally, the critics note that even if one accepts the accuracy of the U.S. productivity numbers, high productivity shouldnt necessarily mean more job losses.
During the 90s, for example, when reported U.S. productivity was also high, the job losses in manufacturing were very slight, compared to the drama of the 2000s, when employment in the sector dropped from 17.2 million to 11.6 million, according to BLS figures.
When pressed, economists acknowledge that it amounts to a matter of emphasis.
Atkinson and other critics of the productivity story concede that indeed, some of the job loss was caused by increasing productivity. Many factories are now filled with automated machines that are far more efficient, after all, so one worker can account for far more output.
But he thinks the dominating factor was that U.S. manufacturers lost their edge amid increasing world competition. Based on that view, he favors changes to government policy that he believes could return millions of manufacturing jobs to the U.S.
Reich, now at the University of California-Berkeley, acknowledges this.
I dont think theres any question that outsourcing has played an important role. The question is: What role?
Regardless of what caused the job loss, Reich said, its difficult to see a huge number of jobs coming back in manufacturing.