Lori Stuntz (IRS) presents Using a Gravity Model to Predict Cross-Border Tax Avoidance (with Michael Udell (IRS, RAAS)) at Georgetown today as part of its Tax Law and Public Finance Workshop hosted by Brian Galle and Day Manoli:
International trade economists use gravity models to explain cross-border flows of goods and services across features of the two countries that either encourage (mass, or size) or discourage (distance) these flows. We repurpose this concept to measure the attractiveness for cross-border tax avoidance.
Our approach recasts the gravity model mass concept as a composite measure that includes the withholding tax rate (WHT) on payments between countries, ownership requirements associated with each WHT, and capital gains taxes imposed by the destination country. The distance term becomes measures of tax administration transparency across the border (with less transparency increasing the attractiveness for crossborder tax avoidance) and an index that measures regulator quality in each country. Unlike the classical gravity equation, which is measured as the attraction across two masses, this repurposed gravity equation for tax avoidance is not limited to an attraction across a single border. Instead, we develop a framework to measure the attraction across multiple borders as a sequence. An important feature of our model is that the order of the countries in a sequence matters.
We create a database of treaty dividend withholding tax rates and associated ownership percentages for qualified corporate dividends across 230 countries using treaty information from the International Bureau of Fiscal Documentation (“IBFD”). For each possible country pair, we record up to 4 different dividend withholding rates and required company ownership percentages, for a total of 59,018 possible bi-lateral cross-border dividend withholding rates. We combine these treaty WHT rates with country level data on domestic tax rates capital gains (also obtained from IBFD), World Bank measures of regulator quality and political stability for each country, and indicators for country participation in various exchange of information programs with the United States. These data allow us to identify countries that are possible “destinations” (low or no income tax and low transparency) or “conduits” (low or no withholding tax and politically stable).
To generate sequences, we begin with all possible dividend withholding rates and associated ownership requirements for dividends between each of the 230 countries in our database (Country A – Country B) and we join in all of the dividend treaty WHTs for each of the 59,018 country B’s with the rest of the world to create over 15 million potential 3-country sequences. We calculate the gravity equation for tax avoidance for each sequence with weights estimated using a measure of foreign financial investment flows.
We develop a framework to chain gravity indexes and evaluate if it is advantageous to add an additional country to the sequence. The gravity model for cross-border tax avoidance is flexible and can be generalized so that sequences originate from any of the 230 countries in our dataset.
Editor's Note: If you would like to receive a daily email with links to tax posts on TaxProf Blog, email me here.
