A Formulary Approach for Attributing Measured Production to Foreign Affiliates of U.S. Parents

Abstract

Consistent with international guidelines, the U.S. Bureau of Economic Analysis measures and attributes output to foreign affiliates of U.S. parents based on data determined under separate accounting according to generally accepted accounting principles. If a multinational company is structured in a way that attributes accounting measures to an affiliate based partially or solely on economic activity resulting from shared inputs that are not employed by the affiliate, measured output may be attributed to an affiliate with relatively few or no local inputs and relatively little or no economic activity. In this paper, we use formulary apportionment, which is also consistent with international guidelines, as a substitute for separate accounting to attribute measured valueadded to majority-owned foreign affiliates (MOFAs) of U.S. parents. Formulary apportionment attributes output based on factors that reflect economic activity. We find overall reattribution from MOFAs to U.S. parents is relatively small—less than 5 percent of value-added attributed to all MOFAs and U.S. parents under separate accounting. However, reattributions across global regions and industry sectors are relatively large. In addition, preliminary work to apply formulary apportionment to imports and exports between U.S. parents and their foreign affiliates yields a relatively large decrease in imports—approximately 3 percent of published total private service imports—but no meaningful change in exports. The overall effect on GDP is only a small increase—approximately 0.1 percent. We conclude that formulary apportionment yields a picture of measured output by industry sector and country that is more congruent with economic activity and more consistent with expectations than related measures generated under separate accounting. JEL Codes: D24, E01, F23

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 10 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.