In the 2000s, U.S. manufacturing suffered its worst performance in American history. Not only did America lose 5.7 million manufacturing jobs, but the decline as a share of total jobs (33 percent) exceeded the rate of loss in the Great Depression.
Despite this unprecedented negative performance, most economists, pundits and elected officials remain remarkably blasé, if not downright giddy about what has transpired. In a HuffPost blog Jay Pelosky writes that "America's manufacturing competitiveness is its best in decades." American Enterprise's Mark Perry writes "manufacturing is doing 'amazingly well." Wow, I'd hate to see what doing poorly looks like.
Many, like Paul Krugman cite some very modest manufacturing employment growth, claiming manufacturing is "one of the bright spots of a generally disappointing recovery." Now the story is that America's poised for a manufacturing renaissance. In fact, the U.S. lost 2.2 million manufacturing jobs during the Great Recession, and since then a little over 300,000, or 13.6 percent, have returned. At the rate of growth in manufacturing jobs in 2011, it would take until 2022 to return to where the economy was in terms of manufacturing jobs before the Great Recession.
How can so many economists and pundits look at a loss of one-third of manufacturing jobs and say things are great? The answer is easy: it's all due to superior productivity growth they claim. For them, companies are getting much more productive, enabling them to produce more with fewer workers. Far from a cause for concern, the dramatic loss in manufacturing jobs should be seen as an indication of just how great we are doing. The steeper the losses, the better. While they may pose a problem for workers, the thinking goes, the job losses benefit consumers and the American economy. All that might be needed are better programs to help laid-off production workers. Certainly no need for a determined national manufacturing competitiveness strategy.
This consensus view knows know political affiliation. Greg Mankiw, Chairman of the Council of Economic Advisers under President George W Bush states, "The long-term trends that we have recently seen in manufacturing mirror what we saw in agriculture a couple of generations ago." Obama CEA Chairman Austan Goolsbee agrees, writing "Employment in the [manufacturing] sector ... get smaller and smaller almost as proof of how productive it has become. It is exactly the same process that agriculture went through." The Congressional Budget Office agrees, writing "Strong growth in productivity and a slower rate of growth in the demand for manufactured goods have necessarily entailed a decline in manufacturing's share of total employment."
So why is elite consensus opinion so firmly behind the "this is a repeat of the agriculture story"? There are two main reasons. First, most economists and pundits do not extend their analysis beyond one macro-level number (change in real manufacturing value-added as a share of GDP) which at first glance appears stable. But this number masks real decline in many industries. In 2010, 13 of the 19 U.S. manufacturing sectors (employing 55 percent of manufacturing workers) were producing less than in 2000. At the end of 2010, the U.S. motor vehicles industry was producing 45 percent less than it had in 2000. The primary metals industry, 36 percent less. Textiles, 47 percent less.
Second and more fundamentally, U.S. government statistics significantly overstate the change in U.S. manufacturing output, and by definition productivity, in part because of massive overestimation of output growth in the computer and electronics sector and because of problems with how manufacturing imports are measured.
Measurement of the computers and electronics industry (NAICS 334) is a particular problem. Because of Moore's law computers get more powerful every year. But when a company makes a computer that is twice as fast than the one it did two years, ago, the government counts it as if they produced two computers. This is why according the government statistics, from 2000 to 2010, the computer and electronics sector increased its real U.S. output over 5.17 times. Compare this with electrical equipment, which saw a decline of 12 percent.
It is hard to believe that the U.S. computer and electronics sector is producing 5.17 times more in the United States than it was a decade ago, given the fact that its employment declined by 43 percent and according the U.S. Census Bureau's the number of units of consumer electronic products shipped from U.S. factories actually fell by 70 percent.
ITIF corrects for these biases in a new report and we find that U.S. manufacturing output actually fell 11 percent over the last decade while GDP increased 17 percent, something that has not happened before, at least since WWII.
Moreover, manufacturing productivity grew just 32 percent, not the reported healthy 72 percent. If productivity growth was actually 72 percent, one would expect that U.S. manufacturers would have added plenty of machines and factories over the last decade to be more productive, as they have done every decade since WWII. In fact, total U.S. manufacturing capital stock increased just 2 percent, compared to historic rates of growth of between 20 and 50 percent per decade.
Thus, while superior productivity increases played some role in declining manufacturing employment, the overriding factor was output decline, highlighted by a striking result: if from 2000 to 2010 manufacturing output had grown at the same rate as that of the rest of the business sector, the United States would have 3.8 million more manufacturing jobs today and at least another four to six million jobs from the multiplier effect.
As such the conventional wisdom that U.S. manufacturing job loss is simply a result of productivity-driven restructuring (akin to how U.S. agriculture lost jobs but is still healthy) is fundamentally flawed. U.S. manufacturing lost jobs because manufacturing lost output, and it lost output because its ability to compete in global markets - some manipulated by egregious foreign mercantilist policies, others supported by better national competiveness policies, including much lower corporate tax rates - declined significantly.
Contrast that to other nations, like Austria, China, Finland, Germany, Japan, Korea, the Netherlands, Sweden, and Taiwan where manufacturing is either stable or growing. The loss of so many U.S. manufacturing jobs is not due principally to higher productivity growth or to some inexorable shift to a post-industrial economy. It is due to a failure of U.S. policies (for example, underinvestment in manufacturing technology support policies and a corporate tax rate that is increasingly uncompetitive) and the expansion of other nations' mercantilist policies.
So much of the debate and rhetoric around U.S. manufacturing is erroneous, precisely because the core data on manufacturing output and productivity are so flawed. It is time for this debate to be informed by accurate data and thorough analysis. We need to arrive at a point where anytime someone asserts that the loss of manufacturing jobs is due principally to superior productivity growth, the statement is challenged as inaccurate. Likewise, with statements like, "since we've lost so many manufacturing jobs it proves that it's no longer an important industry."
This issue is too important for policy to be informed by poor data and shallow analysis. It is time for economists, journalists, policy makers and others to get the story right.
Follow Robert D. Atkinson, Ph.D. on Twitter: www.twitter.com/Robatkinsonitif