Statistics. As businesspeople, we swim in them, we swallow them, we do our damnedest not to drown in them. Mark Twain usually gets credit for the famous saying about the malleability of statistics, though Wikipedia says the author himself attributed the quote to Benjamin Disraeli, a 19th-century British Prime Minister:
Today’s lesson in statistical stretchers (“stretchers” in the Huck Finn sense) comes courtesy of Washington Post reporter Brad Plumer, who argues in a recent entry in his highly recommended Wonkblog that there are many more people working in manufacturing in this country than get reported by the Bureau of Labor Statistics.
Not just a few more: between half a million and 2 million more.
Plumer’s key point is that many companies aren’t counted as manufacturers by the government even though they indisputably participate in the manufacture of goods. Most of these phantom manufacturers fall into the category of factoryless goods producers—companies that design and coordinate the manufacture of goods but don’t have actual factories or assembly plants. They contract the assembly part of the process to other companies, usually overseas.
Apple (IW 500/4), for example, performs virtually all of its manufacturing activities in the United States except the assembly of its iPhones and iPads, which it outsources to China.
To be clear, IndustryWeek unequivocally categorizes Apple as a U.S. manufacturer.
Way More Than Peanuts
Citing a recent study by two Dartmouth College economists, Plumer says in his Wonkblog post that if the government classified all of the factoryless goods producers as manufacturers, it would have added between 431,000 and 1,934,000 workers to the U.S. manufacturing sector in 2007.
The official government count for U.S. manufacturing jobs in 2007 was 14 million. So we’re not talking peanuts; we’re talking about a material statistical disparity.
The study’s authors, Andrew Bernard and Teresa Fort of Dartmouth's Tuck School of Business, caution against reading too much into the foregoing, however. They say they don’t know how the relevant numbers have changed through the years because historical data on the production of factoryless goods is inconsistent. Therefore they can’t say, for example, how much of the manufacturing job loss since 2000 is a result of the shift toward the production of factoryless goods. (They plan to explore this question in the future.)
On the other hand, the study’s authors do assert that the rise in factoryless production definitely canNOT be said account for the entire decline in American manufacturing jobs—from roughly 17 million in the 1990s to 12 million today. That job loss, they say, is clearly a “true decline.” But the overall decline may be mitigated by the fact that some relevant jobs have stayed within the United States. They just aren’t being counted as U.S. manufacturing jobs by the Bureau of Labor Statistics.
Figures often beguile me, but I don’t think it’s any stretch to say that for precision’s sake—for apples-to-apples’ sake—the government needs to revisit how it classifies manufacturers and counts manufacturing jobs.