It’s Time to Rethink How We Measure Globalization

A new paper from Teresa Fort finds that current data collection methods don’t capture the full range of U.S. manufacturing firms’ domestic and global operations.

When the world was a simpler place, most manufacturers did their work in-house and in their home country. This included product design and innovation; making inputs and assembling them; and sales, marketing, and distribution. In the early 1900s, for example, Ford made almost all of its 15 million Model Ts at its Highland Park plant, in Michigan. 

Tuck professor Teresa Fort teaches the core MBA microeconomics course and a seminar on firm responses to globalization. Her research is in international trade and industrial organization, with a focus on firms’ global sourcing and production decisions.

But manufacturing activities that were once solidly within the bounds of one firm and one country have, with improvements in information and communication technology, been fragmented across multiple firms and many countries. Ford today has 30 manufacturing plants, 20 of which are in foreign countries, and it also outsources many tasks to arm’s-length partners domestically and internationally. Some “manufacturing” firms, in fact, outsource nearly all of their physical production to other firms in other countries. Nike has 640 manufacturing locations in 38 countries, all of which involve outsourced relationships with contract manufacturers; similarly, most Apple products are made overseas by non-Apple manufacturers. These firms are known as “factoryless goods producers.”

As Tuck professor Teresa Fort shows in a new paper, these modern forms of manufacturing are difficult to measure with the current data-gathering methods used by statistical agencies. For instance, to be classified as a manufacturer in the North American Industry Classification System, which is used by U.S. statistical agencies, a firm must perform “mechanical, physical, or chemical transformation of materials or components into new products.” Meanwhile, a firm that contracts for manufacturing services will generally be classified in non-manufacturing sectors, such as wholesale trade. This system leaves many firms you think of as manufacturers outside of the manufacturing category. One result has been a misunderstanding of the implications of globalization on U.S. jobs and manufacturing skill. 

In “The Changing Firm and Country Boundaries of US Manufacturers in Global Value Chains,” published recently in The Journal of Economic Perspectives, Fort, a faculty research fellow at the National Bureau of Economic Research, uses novel data sets to identify two organizational forms missing from many analyses on global value chains. The first is U.S. firms that perform physical transformation tasks within the firm boundary using exclusively foreign manufacturing plants. The second is factoryless goods producers that outsource all their physical transformation tasks to arm’s-length contract manufacturers. 

Contrary to the fear that U.S. multinationals have offshored most of their jobs, I find that the vast majority of U.S. firms that own foreign manufacturing plants in 2007 also maintain domestic production.
— Teresa Fort, Associate Professor of Business Administration

By identifying these new types of organizations, Fort has grounds to push back on common perceptions about globalization. “Contrary to the fear that U.S. multinationals have offshored most of their jobs, I find that the vast majority of U.S. firms that own foreign manufacturing plants in 2007 also maintain domestic production,” she writes. Moreover, “contrary to the fear that participation in global value chains entails a loss of technological skills, I find that firms with global in-house manufacturing plants and factoryless goods producers both employ relatively high shares of U.S. ‘knowledge’ workers.” Indeed, these firms are disproportionate contributors to R&D, patenting, and trademarks, Fort reports. 

Fort also argues that the current data gathering system isn’t picking up numbers that lead to reliable estimates of gross domestic product (GDP), value-added, and productivity. For example, when Apple sells directly to foreign customers from their foreign suppliers, the profits are counted in U.S. gross national product, but the value-added by U.S. designers and software engineers may be excluded from GDP. If Apple’s 2011 iPad sales were included in value-added figures for that year, the U.S. computer industry’s value added would have been $6 billion higher, “roughly offsetting the decline in domestic computer manufacturing that year,” Fort explains. 

As Fort makes clear, we need a better way to identify and measure modern manufacturers that elude the outdated statistical tools currently used. Fort suggests expanding what statistical agencies collect, including sales, inputs, imports, and exports; expanding data collection across sectors; flagging imports produced by contract manufacturers; and collecting data on firms’ technology use to facilitate studies on cross-country teams. 

“Such expansions of data collection are crucial for a complete picture of global production sharing and accurate assessments of U.S. supply-chain risk,” Fort concludes.