Thursday, April 21, 2022
Scott Dyreng (Duke; Google Scholar), Robert Hills (Penn State; Google Scholar) & Kevin Markle (Michigan State; Google Scholar), Tax Deficits and the Income Shifting of U.S. Multinationals:
Significant controversy has emerged about the scope of the international tax planning of U.S. multinational firms, with estimates of income shifted out of the U.S., profits recognized in tax havens, and revenue loss ballooning over time. Most studies that derive these empirical estimates use macroeconomic data which support inferences drawn at an aggregate level but are not conducive to analyses at more granular levels. In this study, we use microeconomic data from firms’ publicly available financial statements to derive firm-year estimates that we use to evaluate existing aggregate estimates. We find that many estimates based on macroeconomic data are significantly overstated. We also use our firm-year estimates to analyze the distributions of these amounts within the economy. We show that all dimensions of international tax planning are concentrated in three industries and dominated by a small number of very large firms.
We use firm-year data to re-examine prior macroeconomic studies of the tax avoidance behavior of multinational firms. First, we show that the downward trends in foreign effective tax rates highlighted in prior research (e.g., Wright and Zucman ) are driven by a few very large, profitable firms and that, therefore, the inferences drawn from those trends may not be applicable beyond those firms. Indeed, our data suggest that the typical firm recognizes about 10% of its foreign profits in tax havens, significantly less than the 50% inferred by Wright and Zucman  using macro data. We also estimate U.S. revenue loss (i.e., the U.S. tax deficit) due to income shifting out of the U.S. Our estimates suggest revenue loss figures that are about 20% of those in Clausing , and are consistent with those in Blouin and Robinson .
Our findings suggest caution in matching appropriate data to the empirical question being asked. Because macroeconomic data are implicitly weighted by firm size (often foreign income), conclusions drawn from macroeconomic data might be driven by a few large firms and not reflective of the typical multinational firm. To the extent that policies are implemented to solve income shifting problems and affect all firms when the problem resides in only a few firms, significant economic costs could be imposed on smaller, less profitable firms. Thus, researchers and policymakers should exercise caution when extrapolating from the macro level to the micro level.