September 12, 2013    Volume 20, No. 12

Search Back Issues
Guest Editorials
Trading Exchanges
About Us

Lean Machines Sixteen Case
Studies On Lean
From Manufacturing &
Technology News

Free e-mail newsletter

Output Growth Is 50 Percent Lower Than Reported:
U.S. Economic Statistical System Has Not Measured The Real Costs Of Outsourcing And Import Substitution

By Richard A. McCormack, Editor

The federal government's statistical agencies have acknowledged that they have not measured -- and have even mis-measured -- the real impact on the U.S. economy of surging imports and outsourcing of U.S. production and services to low-cost foreign locations. As a result, the U.S. government has substantially over-reported productivity growth, manufacturing output and gross domestic product. Inflation and import data may also be wrong, as well as the assumption that automation is the primary cause for the loss of millions of American jobs.

It is not a minor issue.

The problem stems from the lack of an "input price index" that captures the substantial savings companies achieve when they switch from domestic to cheaper foreign sources of parts, components, products and services. Government statisticians have investigated how to fix the problem and last year proposed a $1 million pilot program to create an input price index. But Congress did not fund the program. The full cost of implementing such an input price index for manufactured goods would be $11 million annually.

By not gathering price data on import substitution, economists are making wrong assumptions about the U.S. economy, according to economic researchers involved in studying the subject. Economists have effectively argued that there is no need for policymakers to take action to boost manufacturing since output measures paint a rosy picture of a sector that is highly competitive, with output remaining at historically high levels, despite the heavy loss of manufacturing jobs.

But researchers involved in economic statistics say that economists who make the argument that the United States is producing more than ever are out of date on the current state of research on a statistical system that does not measure the impact of globalization on the U.S. economy.

"The shift by U.S. manufacturers from domestic to low-cost imported intermediates has likely imparted a significant upward bias to measured growth in the sector's real value added," according to researchers from the Upjohn Institute and MIT. " 'Offshoring bias' occurs because the price decline associated with a shift from a high- to a low-cost supplier is generally not captured in the import or input price deflators. As a result, real import growth is understated, and the growth in real domestic value added and multifactor productivity is overstated."

So long as imports were only a small part of the economy, "it wasn't too critical" not to have an input-price index, says William Alterman of the Bureau of Labor Statistics. "But as imports have gone from 5 percent to 20 percent of GDP, it has become a much bigger issue. It isn't a one-to-one correspondence, but if GDP is overestimated, then there is an assumption that productivity estimates are also overestimated for the same reason. We are mis-measuring the gains from terms of trade as productivity gains. We are getting gains but they are not productivity gains. Clearly, we have missed something."

Substituting imported products for those that were domestically produced "resulted in an overstatement of the annual growth in real value added by 0.2 percentage points to 0.5 percentage points per year from 1997 to 2007," according to the Upjohn/MIT research. This might not sound like a lot, but it is huge, "implying that real value-added growth was upward biased by as much as 50 percent" for the majority of U.S. manufacturing during that period, says the study. "Estimates on the bias from materials offshoring to multifactor productivity ranged from about 0.1 to 0.2 percentage points per year for all of manufacturing and about 0.2 to 0.4 for the computer and electronics industry."

This statistical lapse is the likely reason why productivity rates have increased, but wage gains have not -- a situation that poses an intellectual quagmire for most economists who have been ignorant about the statistical anomalies associated with globalization's impact on the United States.

The United States government does produce an import price index, but this only tracks prices of the same products that are imported every month. It does not tally the savings to companies from import substitution.

BLS is investigating a system that measures the price impact of import substitution and offshoring as it pertains to the manufacturing sector, since there are potential sources of data for a new index. It sent representatives last year to a meeting of the Institute for Supply Management (ISM) to seek information from company purchasing agents on the feasibility of gathering pricing data for their materials, components, parts and other inputs.

ISM is one of the largest and most respected supply management associations in the world with a total membership of nearly 40,000, notes Alterman. The management of ISM "believed establishments would definitely have the price information necessary to produce these price indexes," he says. "Further, they offered to help us set up a more systematic process for eliciting information and guidance from ISM member companies."

BLS then met with a group of companies including Hewlett Packard, Caterpillar, Celanese and Northrop Grumman to seek further information on how often companies purchase items (daily, monthly, quarterly and annually); whether they kept data on the re-pricing of the same item on a regular basis; the types of prices BLS should seek (contract prices or transaction prices); how often the quality and specifications of the purchased items change; whether buyers have data on the total value of purchases in a referenced year; and how often they change suppliers of major inputs and why.

"In general, the focus group participants indicated that their establishments would almost certainly have the data available that BLS would need to construct these indexes and they did not believe cooperation issues would be any different from what the Bureau currently experiences with establishments," says Alterman. "[I]t was also felt that the strong support from the staff at the ISM itself could help in obtaining cooperation from the member companies as well as suggestions for how best to design data collection and other materials."

In the 1970s, BLS proposed the creation of both an output price index (taken from the wholesale price index) and an input price index. The agency created the output price index because it was considered more critical given that there was not enough money to also create an input price index. BLS "was aware of this as something we wanted to produce but it never got off the ground," says Alterman.

BLS is focusing on a input price index for manufacturing because it is also able to draw data from the Economic Census produced by the Census Bureau every five years. The BLS has been working with that data and has found it adequate in providing pricing details for many manufactured goods.

But the problem of import substitution and outsourcing goes well beyond the manufacturing sector. It could even have a more pernicious impact on the U.S. service sector, reducing officially measured output, trade accounts and productivity in IT, software development, call centers, legal, accounting and medical. Savings from outsourcing those functions are also not being picked up in the import data. The result: the American economy looks like it is producing more with fewer Americans, which is not the case.

"For the rest of the economy and for services [the five-year Economic Census] doesn't have the level of detail" necessary to produce an input price index, says Alterman. Creating one for the manufacturing sector would cost between $10 million and $11 million a year, but manufacturing represents only 11 percent of the U.S. economy. "So it raises the issue of how do you do that with services," says Alterman. "Regardless of whether a call center is in the United State or is foreign, what is the output of that call center? Should you measure calls completed per hour, or the number of problems solved or not solved? Regardless of whether it is being moved from domestic to foreign, you still have the question of what is the measure of output. The input measure might be along costs, but you still have to figure out what to measure and that is difficult, and becomes even more problematic for inter-company transfers of services."

Technical papers on these subjects are available for viewing at the Upjohn Institute website, "Measuring the Effects of Globalization," located at

Provide us with a comment on this article.

We'll notify you as issues and free stories like this one appear on this site. Sign up for a content-rich, e-mail newsletter. (You will NEVER receive spam.)

Please consider subscribing to Manufacturing & Technology News. You will have access to all back issues dating to 1998, plus receive the current issue electronically and via regular mail. It is all original reporting on the most important stories facing U.S. industry. No advertising. The cost of a new subscription is $495 per year.

Scan Back Issues Comments | About Us | How To Order

Reproduction Rights 2012 Are Granted To This Story So Long As A Link Is Provided To This Source Of Original Content.