LSE - Small Logo
LSE - Small Logo

Helena Vieira

August 19th, 2016

Most multinationals use unaffiliated suppliers in their supply chains


Estimated reading time: 5 minutes

Helena Vieira

August 19th, 2016

Most multinationals use unaffiliated suppliers in their supply chains


Estimated reading time: 5 minutes


The global nature of supply chains has rapidly come to dominate international trade. This column presents new evidence on production fragmentation and intra-firm trade. For US corporations, cross-country shipments of goods between units of the corporation are rare, despite the fact that most US manufacturing parents own foreign affiliates in upstream or downstream industries.

In 1997, the chipmaker Intel invested $330 million to build an assembly and test plant near Rio Segundo, Costa Rica. This plant assembled parts produced mostly by other Intel-owned affiliates. By 2006, Intel employed 3,500 workers and was responsible for almost 5 per cent of total Costa Rican GDP and 20 per cent of its exports (CEFSA 2006).

At about the same time, Apple began contracting with Taiwanese firm Hon Hai Precision Industry Co. – better known as Foxconn – to assemble Apple products out of parts produced almost entirely outside of Apple. In 2013, almost one third of Foxconn’s one million Chinese employees worked solely on producing the iPhone 5s.

Both Intel and Apple created these arrangements to take advantage of cost savings abroad, but they did so in very different ways. Intel created a wholly owned subsidiary, while Apple contracted ‘at arm’s length’ with an unaffiliated party. The cross-country production fragmentation embodied in these relationships – the global supply chain – has rapidly grown to the point that it now dominates international trade (Johnson and Noguera 2012). How are these supply chains structured?

In principle, these global supply chains can exist inside or outside the firm. In recent research using affiliate-level data on the activities of US multinational corporations from the Bureau of Economic Analysis (BEA), we document that the case of Intel is exceptional among the US multinational corporations in manufacturing. Cross-country fragmentation of production within the boundaries of the firm, with the purpose of providing goods to other parts of the firm, is not the norm.

New evidence on intra-firm trade and production fragmentation

  • Fact 1: Trade within US corporations is relatively rare.

The median affiliate ships nothing to – and receives nothing from – its US parent. Only 10 per cent of the foreign affiliates of US multinationals are exclusively dedicated to providing goods to the rest of the corporation, while more than half of them ship less than 10 per cent of their output to other members of the corporation. This is evident from Figure 1, in which we plot the histogram of US manufacturing affiliates by the share of their total sales shipped to affiliated parties. Most affiliates of multinational firms appear to exist with the purpose of serving consumers, or other firms, in their markets of operation. The concentration of intra-firm trade in a few affiliates is even more apparent in Figure 2 – 5 per cent of the foreign affiliates of US multinationals account for three quarters of the total intra-firm trade observed in the data.

Figure 1. Distribution of affiliates by share of total sales shipped to affiliated parties

rappoport fig1

  • Fact 2: Independent of the size of their host markets, those affiliates that do ship goods within the firm tend to be large – such as Intel’s massive investment in tiny Costa Rica – and are often the largest affiliates within the corporation.

In fact, affiliates that trade with their parents employ, on average, twice as many workers as those that do not, while the largest 5 per cent of foreign affiliates accounts for half of the total parent-affiliate trade recorded at the Bureau of Economic Analysis. The skewness of intra-corporation trade towards large affiliates is reminiscent of the skewness of manufacturing exports in large firms (Bernard and Jensen 1995) and of multinational activity in even larger firms (Helpman et al. 2004). The concentration of intra-firm trade in the largest firms is consistent with Antras and Helpman’s (2004) contract theory of the multinational firm, in which only the largest firms choose to keep offshore activities within the firm.

  • Fact 3: Though most parents own affiliates in upstream and downstream industries, the striking fact is that we find no relationship between the parent and the affiliate industries’ input-output relationship and intra-firm trade.

In our research, we document this regularity using the input-output table*. This table shows, for example, that firms in the ‘agricultural chemicals’ industry use goods from the ‘pharmaceutical industry’ as inputs. Thus, it is natural to expect that an affiliate in the pharmaceutical industry owned by a parent in the agricultural chemical industry would ship goods to its parent – this is what the classical theory of the boundaries of the firm would predict. The data show that even though the overwhelming majority of affiliates are in industries tightly linked by an input-output relationship to their parent’s industries – a fact also documented in Alfaro and Charlton (2009) – we find no significant relationship between the flow of goods between members of the multinational corporation and the input-output links between them.

Figure 2. Distribution of affiliate’s trade to affiliated parties

rappoport fig2

Using data on multi-plant firms within the US, Atalay et al. (2014) also find a lack of trade between upstream and downstream plants within the firm. Our finding is even more surprising given that factor price differences – the theoretical motivation for production fragmentation and the intra-firm trade that accompanies it – are much larger across countries than across US cities.

Interpretation of findings and challenges

Our findings raise the obvious question: Why do multinational corporations own facilities in industries with strong input-output relationships, if not for the shipment of goods along the production chain? As Atalay et al. (2014) suggest, it may be that the firm’s boundaries are determined not by the ability to transfer goods within the firm, but by the ability to transfer capabilities. Production fragmentation between two parts of the same corporation working in two industries with strong input-output links may signal the use of a common set of intangible inputs, knowledge, or expertise. These intangibles are a source of comparative advantage for the multinational firm, which entails specialisation in the production of input-output linked goods, even in the absence of physical shipments between within related parties.


Our finding that there is little international production chain fragmentation within the boundaries of the corporation raises a new set of questions to be explored in future research. Multinational corporation configurations are not always as simple as the Apple and Intel models. The ways that the multinational firm uses related parties and third-party suppliers are more complex than previously thought – for instance, to what extent do these third-party suppliers interact with different parts of the same corporation? Are multinational firms replicating their production chains abroad? The challenge in answering these questions is the availability of even more detailed data.

*Note from the LSE Business Review editor: input-output tables describe the sale and purchase relationships between producers and consumers within an economy.



Natalia RamondoNatalia Ramondo is an Assistant Professor of Economics at the University of California at San Diego (UCSD), School of Global Strategy and Policy, and a faculty research fellow at the National Bureau of Economics Research (NBER). Her research focuses on international economics with especial emphasis on the behavior of multinational firms.Before joining UCSD in 2013, she was an Assistant Professor at the University of Texas at Austin and Arizona State University, as well as a research fellow at Princeton University.  She received her Ph.D. in Economics from the University of Chicago in 2006 and her B.S. in Economics from Universidad de Buenos Aires in 1997.

Veronica RappoportVeronica Rappoport is an Assistant Professor at the Managerial Economics and Strategy Group, Department of Management, LSE. Her research interests are in international trade and investment and international finance.



Kim RuhlKim Ruhl is an Assistant Professor of Economics at the New York University Stern School of Business.  His research focuses on international economics, models of firm heterogeneity, and national income accounting. Before joining Stern in 2006, he was an Assistant Professor at the University of Texas at Austin and a Visiting Scholar at the Federal Reserve Banks of Minneapolis and Philadelphia.  He received his Ph.D. in Economics from the University of Minnesota in 2004 and his B.S. in Economics from Bowling Green State University in 1999.

About the author

Helena Vieira

Posted In: LSE Authors

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.