Income shifting and U.S. international trade in goods statistics

Highlights

• Path analysis of how tax-motivated income shifting affects trade in goods statistics.

• Transfer pricing does not appear to affect intrafirm trade in goods statistics.

• Location decisions do significantly affect intrafirm trade in goods statistics.

Abstract

Intrafirm trade represents greater than one-third of total U.S. international trade in goods. Since these are not arm’s-length transactions, trade policymakers have voiced concerns that income shifting may distort international trade in goods statistics through the manipulation of transfer prices. Using country-level data on intrafirm exports and imports, we estimate a path analysis that simultaneously tests how and to what extent tax-motivated transfer pricing and real investment decisions affect intrafirm trade in goods statistics. Contrary to speculation, we do not find an economically significant relation between transfer pricing and intrafirm trade in goods statistics. In contrast, we find that tax-motivated location decisions create a 21 (20) percent or $819.7 ($927.1) million difference in mean intrafirm exports (imports) between the U.S. and a low- and high-tax country. This study provides trade policymakers with relevant information about the extent to which real investment decisions and accounting manipulations affect intrafirm trade in goods statistics and contributes to the international trade and income shifting literatures.

Introduction

This study investigates how and to what extent tax-motivated income shifting affects U.S. international trade in goods statistics. Trade in goods statistics measure the value of exports and imports of physical goods flowing between the U.S. and foreign countries. As the U.S. and world economies have become more interconnected through international trade, trade in goods statistics have become one of the most closely watched economic indicators in the U.S. (Bureau of Economic Analysis, 2014). For instance, U.S. trade policymakers, such as the President and the Office of the United States Trade Representative, use trade in goods statistics to assess the effectiveness of trade agreements, whether trade partners are engaging in unfair trade practices, and the effects of trade relationships on employment and wage growth in the U.S. (Executive Order 13786, 2017). Likewise, international trade economists use trade in goods statistics to examine the determinants of international trade, as well as to understand the relation between international trade and U.S. economic and employment growth (e.g., Burtless, 1995, Singh, 2010). And, the business press is replete with headlines trumpeting the latest release of trade in goods statistics and the effects of international trade on the U.S. economy (e.g., Swanson, 2019, Harrison, 2020).

Understanding how income shifting affects U.S. international trade in goods statistics is of interest for at least three reasons. First, intrafirm trade makes up a significant portion of total international trade in goods. Specifically, when a U.S. multinational enterprise (MNE) engages in an intrafirm sale (purchase) of goods with its foreign affiliates, the transactions are reported in the U.S. international trade in goods statistics as exports (imports) between the U.S. and a foreign country. Over the last 30 years, the value of these intrafirm exports (imports) represents between 25 and 39 (33–43) percent of total U.S. exports (imports) of goods (see Fig. 1).

Second, because intrafirm trade in goods is not an arm’s-length transaction, trade policymakers have voiced their concern that the manipulation of transfer prices for tax purposes may distort aggregate trade statistics. For example, in an Organization for Economic Co-operation Development (OECD) policy paper, Bonturi and Fukasaku (1993, 146) state, “The use of transfer pricing in intrafirm trade may introduce an element of uncertainty into the value of a fairly large part of international trade…” Moreover, in another OECD policy paper Lanz and Miroudot (2011, 25) state, “From a trade policy perspective, the most crucial question is then how transfer pricing affects the measurement of trade: Does the pricing of intra-firm trade distort the actual trade patterns between countries? And to what extent is intra-firm trade over- or underestimated?”

Third, U.S. tax policy may indirectly induce intrafirm trade in goods via tax-motivated location decisions. In particular, prior to the Tax Cuts and Jobs Act of 2017 (TCJA), the U.S. had one of the highest corporate tax rates in the world. Thus, U.S. MNEs could lower their tax burden by shifting investment from the U.S. to low-tax countries (Grubert and Mutti, 1991, Hines and Rice, 1994). Moreover, since foreign direct investment (FDI) is a prerequisite for intrafirm trade, it is likely that once an MNE establishes an affiliate in a foreign country, there will be an increase in intrafirm trade with that country (Grubert and Mutti, 1991, Clausing, 2000).

To examine how tax-motivated income shifting affects U.S. international trade in goods statistics, we obtain intrafirm exports and imports of goods data at the country level from the Bureau of Economic Analysis (BEA) for the years 1983–2015. Following prior research in accounting (Collins et al., 1998, Klassen and Laplante, 2012), we compare the average effective tax rate (ETR) of foreign country i with the U.S. tax rate as our proxy for the incentive to engage in income shifting.

We use path analysis to simultaneously investigate two potential links between our proxy for the incentive to engage in income shifting and intrafirm trade in goods statistics. First, we examine a direct link between tax-motivated transfer pricing and intrafirm trade in goods statistics (i.e., the direct transfer pricing effect). If MNEs significantly underprice (overprice) intrafirm exports (imports) of goods with foreign affiliates in low-tax countries, the direct transfer pricing effect will be positive (negative). Second, we examine an indirect link between the tax incentive to invest real business activities in low-tax countries and intrafirm trade in goods statistics (i.e., the indirect location effect). If MNEs increase FDI in low-tax countries, which in turn increases intrafirm trade with affiliates in these countries, the indirect location effect will be negative for both intrafirm exports and imports of goods.

We estimate our path analysis using a structural equation model (SEM) that controls for country-level characteristics and trade costs examined in the trade literature. We find that the direct transfer pricing effect for intrafirm exports and imports is not statistically nor economically significant. This finding is consistent with the notion that MNEs have limited flexibility in manipulating the transfer prices of physical goods (Grubert, 2003, Swenson, 2001). In contrast, we find that the indirect location effect for both exports and imports is statistically and economically significant. Specifically, we find that on average a one percentage point decrease in the foreign ETR results in a 1.08 (1.01) percentage point increase in U.S. intrafirm exports (imports). This translates into a difference of approximately 21 (20) percentage points in intrafirm exports (imports) of goods between the U.S. and a low- and high-ETR country or roughly $819.7 ($927.1) million using the mean intrafirm export (import) value.¹

We perform several additional analyses to further examine and corroborate our findings. We find qualitatively similar results when we analyze only OECD countries, which account for nearly 80 percent of all U.S. intrafirm trade. In addition, the main findings hold when tax haven countries are excluded, suggesting the results are not driven by only a subset of low-tax countries. The findings also hold after including country fixed effects in the model, are robust to the treatment of outliers, and are largely consistent over our entire sample period. Further, the results are not sensitive to using the difference between the statutory tax rate of the foreign affiliates (rather than the ETR) and the U.S. tax rate as the incentive to manipulate the transfer prices of goods.

Our paper addresses an important issue of relevance to trade policymakers. Since U.S. MNEs have a tax incentive to strategically choose transfer prices that underprice (overprice) intrafirm exports (imports) to foreign affiliates in low-tax countries, trade policymakers have speculated that such accounting manipulations may distort actual trade patterns between countries and explain an important portion of the U.S. trade in goods deficit. However, we do not find any evidence that the manipulation of transfer prices significantly biases trade in goods data. Rather, our findings suggest that U.S. tax policy indirectly induces intrafirm trade in goods via tax-motivated location decisions. This insight suggests that rather than debating whether and how to measure intrafirm trade in goods (Eden, 2001), trade policymakers should consider how U.S. tax policy indirectly affects the flow of goods. This insight is also consistent with remarks by the Semiconductor Industry Association (SIA)—a U.S. trade and lobbying group representing the semiconductor industry—in 2017 on potential factors that may cause trade deficits. Specifically, in response to the Omnibus Report on Significant Trade Deficits requested by President Donald Trump (Executive Order 13786, 2017), SIA stated “America’s corporate tax code is outdated and uncompetitive, creating disincentives for making domestic investments…”²

This insight also extends prior work by Clausing (2006) who suggests that transfer pricing decisions are associated with intrafirm trade balances (exports less imports). Because the direct transfer pricing and indirect location effects are motivated by similar local tax incentives, we use a path analysis that simultaneously estimates and isolates the effects of the two channels while controlling for important country-level characteristics and trade costs. In contrast, Clausing (2006) uses a single-equation approach that does not adequately distinguish between the two channels nor control for other relevant country-level characteristics and trade costs. In an additional analysis, we find that transfer pricing decisions do not have a statistically nor economically significant effect on intrafirm trade balances once these concerns are addressed.

Our study will also be of interest to researchers that seek to better understand how U.S. MNEs shift income to low-tax countries. We provide further evidence that the manipulation of transfer prices on goods is not economically significant (Swenson, 2001), which is consistent with Grubert’s (2003) assertion that U.S. MNEs predominantly manipulate the transfer prices on intangible rather than tangible assets. We also contribute to the growing literature on the non-tax consequences of income shifting.³ While the prevalence and income tax effects of income shifting have been widely studied and debated, the non-tax consequences of income shifting are far less clear. We provide evidence of how income shifting via tax-motivated location decisions affects international trade in goods statistics, which play a vital role in the U.S. government’s monitoring, analysis, and projections of macroeconomic activity.

Lastly, our study contributes to the developing literature on the determinants of intrafirm trade. While there is an extensive literature in international economics that examines the determinants of international trade in general, studies of the determinants of intrafirm trade are limited (Lanz and Miroudot, 2011). We provide evidence that tax-motivated location decisions are an economically significant determinant of the amount of intrafirm trade in goods.