Jennifer Blouin and Leslie Robinson, tax accounting professors at Wharton and Dartmouth, respectively, have an important paper on the geographic allocation of multinational enterprises’ (MNEs) reported income. [Link] The paper recently presented at the University of North Carolina’s Tax Symposium [Link] shows that the US Bureau of Economic Analysis (BEA) data, used by many tax researchers, significantly overstates the amount of income reported in tax haven countries. This affects the magnitude of some estimates of the fiscal effects from base erosion and profit shifting (BEPS).
Bottom-line: Some estimates of BEPS are overstated due to BAD (Best Available Data, see also 8/31/16 blog), but tax rate differences still significantly encourage profit shifting to minimize taxes. Better understanding of the data should increase greater transparency of MNEs’ tax and economic activities globally.
Blouin and Robinson provide critical insights into how MNEs report their financial information. MNEs can report the geographic allocation of their net income in three ways: consolidated, equity method, or cost (dividend) method. Consolidated reporting eliminates double counting, but then requires some allocation method (e.g. arms’-length transfer pricing, apportionment) for assignment of income to individual countries. The equity method includes the earnings of lower tier directly-owned (50%+) subsidiaries in the subsidiaries’ countries and also in the country of the direct owner. Since many multi-tiered corporate structures place holding companies or conduits in tax havens, those countries are assigned the income earned in the tax haven plus the earnings of lower-tier subsidiaries in operating countries. The cost method includes the dividends received from lower tier subsidiaries in the income earned in the country of the upper tier holding company.
The BEA Activities of Multinational Enterprises [Link] requires US MNEs to report the earnings of their foreign subsidiaries using the equity method. Before the Blouin/Robinson analysis, concerns about possible double counting were expressed, but not quantified. Their analysis, working directly with the underlying BEA data, revels the magnitudes of the double counting and the distorting geographical allocation of the aggregate statistics. The analysis finds that 41% of US MNEs’ foreign source income in 2016 is from tax havens when the double counting is removed, compared to 62% when not corrected. The adjusted tax haven percentage is much closer to the 35% I calculated from the 2017 US MNE CbCR report, when all large US MNEs were required to report. [Link]
They also re-estimate the regression analysis of a well-regarded statistical analysis of BEPS. The analysis also shows that the elasticity of profit shifting to differential tax rates falls from -2.7 to -1.8, which reduces the estimated fiscal effects, but is still quite high. A recent IMF staff meta-analysis of profit shifting [Link] found a mean elasticity of -1.0, which is also high. An elasticity of -1.0 means a 10-percentage point difference in tax rates results in a 10% shifting of profits from higher-rate to lower-rate countries. A -1.8 elasticity means that a 10-percentage point difference results in an 18% shifting of profits. Even with a 21% US federal corporate tax rate and a zero-rate tax haven, the tax rate differential could result in a 21-38% shifting of profits, absent other tax base protection measures such as Controlled Foreign Corporation, interest limitations, or minimum tax rules.
The Blouin/Robinson paper finds that studies of BEA data by Zucman et. al as well as US tax return data suffer from double counting and geographical misallocation. The G20/OECD BEPS Country-by-Country Reporting (CbCR) data does not have the geographical misallocation but includes some double-counting in the “Stateless Income” reported. Some stateless income is from fiscally transparent entities, such as partnerships, as well as from some of their owners that meet the reporting requirement of $850 million of annual revenue. The OECD BEPS Action 13 public consultation [Link] on possible revisions to the CbCR notes several approaches to greater transparency of “Stateless Income,” which could show the potential double-counting and geographic misallocation, as well as highlight entities that don’t report taxable income to any jurisdiction. The Blouin/Robinson analysis show that 2020 revisions to the CbCR are needed.
More detailed reporting of multi-tiered MNE income is needed to improve tax administrators’ ability to perform high-level risk assessment and for policy analysts to more effectively use financial data (accounting, tax, and CbCR) to analyze BEPS and the effects of BEPS policy changes. Many MNEs have enormously complex corporate structures for tax minimization, regulatory and business reasons. Greater transparency in financial accounting and/or CbCR would highlight MNEs structuring tax conduits through tax havens.
Blouin and Robinson have made a significant contribution to the financial accounting, tax and BEPS research with their focus on the details of the underlying BEA data, used by many analysts. The BEA data has been the Best Available Data (BAD), but now we have a better idea of how BAD it was, and also how it can be improved for future analysis. Their analysis should make us question the magnitudes of many BEPS analyses, but it still shows that BEPS was still a significant problem in 2016 and still requires continued focus and tax policy changes to reduce remaining BEPS.
Twenty plus countries’ CbCR data is being used in the OECD’s economic impact assessment of the digitalization tax proposals and is scheduled to be included in the OECD’s second Corporate Tax Statistics publication later this year. That information will help assess the extent to which BEPS was principally a US MNE phenomenon or more widespread. Additional analysis of the BEA and CbCR data could also assess how pervasive BEPS is by size of MNE, and whether the CbCR reporting threshold of $850 million of annual global revenue should be lowered.
Tom Neubig