I spoke at two symposiums last week in London (Economic Statistics Centre of Excellence) and Luxembourg (EC's Eurostat on distortions to countries’ national economic statistics as a result of multinational companies’ tax planning involving intellectual property. Current national statistics in a globalized intangible-intensive economy with sophisticated cross-border supply changes may not be reliable for economic policymaking due to the effects of tax base erosion and profit shifting (BEPS). A recent NBER Conference on Research in Income and Wealth explored similar topics.
Two examples below illustrate the problem. Fortunately, recent global and national tax initiatives will reduce the distortions in the future. Distortions in national statistics from tax BEPS aren’t limited to Ireland and the United States, but both counties illustrate the problem.
· 25% GDP growth rate in Ireland: In 2015, Ireland’s Gross Domestic Product grew at an astronomical rate due to several factors, including changing the location of headquarters (so-called tax inversions) and also the transfer into Ireland of IP of foreign MNEs combined with foreign contract manufacturing. Ireland’s Central Statistics Office now also presents an “adjusted” Gross National Income to better measure actual economic activity of Irish residents. Distorted high macroeconomic growth might just tighter macroeconomic policy than appropriate. See the analysis by John Fitzgerald of Trinity College Dublin.
· Profit-shifting out of the high-tax-rate US to lower-tax-rate countries raised the US trade deficit, by lowering the measured value of US exports and raising the measured value of US imports. A recent NBER paper by Guvenen, Mataloni, Raissier and Ruhl estimate the distortion could have been as large as one-half of the US trade deficit in 2012. Would the Trump Administration be pushing tariffs and quotas as strongly if the measure of the trade deficit was not distorted by BEPS?
My recent World Intellectual Property Organization (WIPO) working paper finds that BEPS distorts the measure of Charges for the Use of Intellectual Property (CUIP). Additional empirical analysis finds that CUIP receipts relative to prior year Research and Development expenditures is significantly reduced by high statutory tax rates on Intellectual Property income.
The bad news is the national statistics have been distorted and they would be unlikely to be improved by attempted imputations, for instance through formula apportionment of global income. Similar to double non-taxation, some world GDP does not get counted due to mismatches resulting from MNE tax planning. National statisticians are grappling with how to measure economic ownership, similar to tax transfer pricing rules determining the allocation of cross-border profits by functions performed, risks assumed and assets used.
The good news is that the national statistics distortions from BEPS will be reduced (not eliminated) by recent global tax initiatives and national tax changes. These include:
· the OECD/G20 BEPS Project’s tighter linkage between intangible pricing and value added in the revised OECD Transfer Pricing Guidelines;
· shutting down tax haven “cash boxes;”
· restraint on harmful tax completion by requiring economic nexus for special tax rates, including for Intellectual Property;
· new transparency rules requiring country-by-country reporting (CbCR) and special country tax rulings;
· reduction in the variation across country tax rates, particularly after the 2017 US tax change, see November 20, 2017 blog;
· Ireland’s ending the tax rules enabling the Double Dutch Irish Sandwich next year; and
· increased focus in both tax policy and administration on hybrid structures and transfer mispricing.
National statistical offices (NSOs) are grappling with how best to measure national production and income in an increasingly global and intangible-intensive economy. Many of the same issues are being addressed by tax policymakers and tax administrations. NSOs often worry that asking companies about tax in their surveys would reduce response rates, but tax statistics may be the best available data for national measures of income of multinational corporations.
One of the potential action steps at the Eurostat symposium was increased data sharing across NSOs. I would also recommend more data sharing between tax administrations and NSOs. In the U.S., the Bureau of Economic Affairs benefits from being able to utilize tax statistics under strict confidentiality arrangements. The CbCR by MNEs to tax administrations is to be used for high-level risk assessment, and also “where appropriate, for economic and statistical analysis.” Reducing the distortions in countries’ macroeconomic national statistics is an important reason for policymakers to have the CbCR aggregated data published by their tax administrations and also to give their NSOs limited access to the new CbCR information.
Improving national statistics is important and continued focus and cooperation between tax and national statistics will be critical to addressing future measurement issues, such as cross-border income from digitalization of the global economy.