1. Empirical comparisons of profit shifting under worldwide and territorial KEVIN MARKLE, UNIVERSITY OF WATERLOO ITPF/AEI CONFERENCE, WASHINGTON, D.C. MARCH 31, 2014

2. Estimating profit shifting Profit shifted to a jurisdiction = Profit reported – Expected profit (i.e., no shifting) 3/31/2014MARKLE 2

3. Estimating profit shifting Profit shifted to a jurisdiction = Profit reported – Expected profit (i.e., no shifting) ObservableNot observable, have to estimate 3/31/2014MARKLE 3

4. Estimating profit shifting Profit shifted to a jurisdiction = Profit reported – Expected profit (i.e., no shifting) 2 methods commonly used to estimate “expected” profit: 1.Use information in 10Ks to determine average associations between domestic sales and foreign profits and between foreign sales and domestic profits. 2.Use financial statements of reporting entities in the corporate group to calculate an expected profit based on inputs: fixed assets, labor, productivity. ObservableNot observable, have to estimate 3/31/2014MARKLE 4

5. Estimating profit shifting Profit shifted to a jurisdiction = Profit reported – Expected profit (i.e., no shifting) 2 methods commonly used to estimate “expected” profit: 1.Use information in 10Ks to determine average associations between domestic sales and foreign profits and between foreign sales and domestic profits. 2.Use financial statements of reporting entities in the corporate group to calculate an expected profit based on inputs: fixed assets, labor, productivity. To test for tax-motivated income shifting:  Add a tax incentive variable to the model.  To the extent that reported profit varies with tax incentive, profit shifting is inferred. ObservableNot observable, have to estimate 3/31/2014MARKLE 5

6. Comparing worldwide to territorial Method 1: Using the information in 10Ks  Sample is US MNCs only 3/31/2014MARKLE 6 Method 2: Using the information of subs  Sample is subs of MNCs from many countries

7. Comparing worldwide to territorial Method 1: Using the information in 10Ks  Sample is US MNCs only  Key assumption: the need to repatriate foreign earnings in the short term is what prevents worldwide with deferral from being de facto territorial. 3/31/2014MARKLE 7 Method 2: Using the information of subs  Sample is subs of MNCs from many countries  Key assumption: the sub of a German MNC is expected to behave the same as the sub of a U.S. MNC except for the international regime of the MNC.

8. Comparing worldwide to territorial Method 1: Using the information in 10Ks  Sample is US MNCs only  Key assumption: the need to repatriate foreign earnings in the short term is what prevents worldwide with deferral from being de facto territorial.  Use “financially unconstrained” firms as proxy for territorial.  Advantage: allows estimation of income shifted out of the U.S. 3/31/2014MARKLE 8 Method 2: Using the information of subs  Sample is subs of MNCs from many countries  Key assumption: the sub of a German MNC is expected to behave the same as the sub of a U.S. MNC except for the international regime of the MNC.  Compare the shifting of subs of worldwide MNCs to that of subs of territorial MNCs.  Advantage: does not need to proxy for “territorial”

9. Empirical results Method 1: Using the information in 10Ks  Dyreng and Markle paper  Sample: US MNCs, 1998-2011  Results  “Territorial” MNCs shift 2% - 19% more income out of the U.S.  Caveats  Assumes that all else is held constant (i.e., base erosion measures are same for two groups)  Firm-level measures of financial constraint are not as precise as we would like 3/31/2014MARKLE 9

10. Empirical results Method 1: Using the information in 10Ks  Dyreng and Markle paper  Sample: US MNCs, 1998-2011  Results  “Territorial” MNCs shift 2% - 19% more income out of the U.S.  Caveats  Assumes that all else is held constant (i.e., base erosion measures are same for two groups)  Firm-level measures of financial constraint are not as precise as we would like 3/31/2014MARKLE 10 Method 2: Using the information of subs  Markle paper  Sample: MNCs from many countries, 2004- 2008  Results  Territorial MNCs shift a higher percentage of income, on average, than worldwide MNCs  The difference is not statistically significant in a subsample of MNCs with foreign growth opportunities  Caveats  The worldwide/territorial classification of a sub looks through any intermediate structures