Progress Report, Joint Census/BEA Research Project

The Business Dynamics of Multinational Corporations and Global Value Chains
Accomplishments in Year 1
Link BEA and Census Data
This project has developed the data infrastructure to identify multinationals (U.S. parents and
U.S. affiliates of foreign parents) and services traders (exporters and importers) in the Census
Bureau’s Business Register. This is accomplished by linking the following surveys to the Business
Register:
 BE-10/BE-11 (Surveys of U.S. Direct Investment Abroad) for the years 2007-2013;
 BE-12/BE-15 (Surveys of Foreign Direct Investment in the United States) for the years
2007-2012; and
 BE-120/125 (Surveys of Transactions in Selected Services and Intellectual Property with
Foreign Persons) for the years 2007-2012.
An entity in one of the three BEA surveys mentioned above contains a name, Employer
Identification Number (EIN), and/or address. Therefore, an entity in a BEA survey is linked to a
firm in the Business Register using a combination of EIN, name, and/or address matching. In
addition to implementing computer matching algorithms, we also undertake clerical matching. We
perform clerical matching on approximately 500 U.S. multinational parent enterprises (with at
least 100 employees) that appear in almost all years between 2007 and 2012 and for which we did
not obtain matches using EINs. Table 1 describes the final results of the matching.
BEA Survey
BE-10/11
BE-12/15
BE-120/125

Average Weighted Match Rate
99%
94%
N/A

Average Unweighted Match Rate
96%
61%
94%

Recommendations for new data collections related to factory-less goods producers in the 2017
Economic Census
This project has developed the data infrastructure to assess the responses to the special
inquiry about the purchase of contract manufacturing services in the 2012 Economic Census (EC),
an establishment level survey, and the Company Organization Survey (COS), an enterprise level
survey integrated into business register processing, in order to explore the feasibility of identifying
factory-less goods producers (FGPs). The OMB’s Economic Classification Policy Committee
(ECPC) recommends measuring FGP activity at the establishment level. However, cognitive
research by the Census Bureau suggests that company headquarters rather than individual
establishments may be the best respondents to answer the special inquiry questions intended to
measure FGPs.
No single existing survey contains an ideal set of questions to identify FGPs. Therefore,
we have linked the following confidential micro datasets for 2012 sourced from the Census Bureau
and the Bureau of Economic Analysis:
 Economic Census (EC)
 Company Organization Survey (COS)
 Linked Firm Trade Transactions Database (LFTTD)

 Longitudinal Business Database (LBD)
 Annual and Benchmark Surveys of US Direct Investment Abroad (BE-10/11)
 Foreign Direct Investment in the United States (BE-12/15)
Preliminary analysis utilizing the above sources of data is currently underway and will enable
the most detailed analysis of “factory-less” manufacturing activities to date and pave a path
towards offering a feasible way of identifying firms that engage in these activities.
Use linked data sets to construct a supply and use table for 2012 following initial specifications
The project has started creating supply-use tables for 2012 by estimating the components
of extended (by firm type) supply-use tables for the “Semiconductor and other electronic
component manufacturing,” NAICS 3344 for 2012. Firms are categorized by type of ownership,
firm class size, and whether the firm exports. Preliminary results show that most of the gross
output in this industry is concentrated in a few permutations of these firm types. By type of
ownership, value added as a share of gross output is largest for U.S. MNEs and lowest for
foreign MNEs. Also, trade as a share of gross output and labor productivity is larger for MNEs
than non-MNEs. This finding is consistent with results from Fetzer and Strassner (2015) on
manufacturing. Type of ownership and exporter status seem to better capture heterogeneity than
firm class size in this industry. This analysis will help inform the initial specifications used for
the supply-use tables for all industries.
References
Fetzer, James and Strassner, Erich H., Identifying Heterogeneity in the Production Components
of Globally Engaged Business Enterprises in the United States (June 10, 2015), BEA Working
Paper WP2015-13.