International Tax Planning of Dutch Shell Companies (SPEs) By - PowerPoint PPT Presentation

International Tax Planning of Dutch Shell Companies (SPEs) By Arjan Lejour, * Jan Möhlmann, Maarten van ‘t Riet, Thijs Benschop * Presenter, also Tilburg university CPB Netherlands Bureau for Economic Policy Analysis

The Netherlands: foreign investment hub! NLD USA 4084 5332 USA NLD 12412 3725 LUX LUX 14942 5094 CHN UK 3634 HKG CHN 4419 2535 1385 1415 1439 Inward FDI position (bln US, 2016) Outward FDI position Global: 30 000 billion US$ SPEs are key: 80% of the FDI positions is on their balance sheets The Dutch SPEs and Tax Planning CPB Netherlands Bureau for Economic Policy Analysis 17 November 2018 2

Data DNB • Financial information 4500 (balance sheet and 4000 Total foreign assets (x billion euro) annual results) by SPE 3500 between 2004 - 2016, 3000 with geographical 2500 information Sample 2000 Population • Largests SPEs fully 1500 covered; smallers ones 1000 sampled 500 0 • About 1000-1500 SPEs by year in sample • This sample covers 10% of all SPEs and 2/ 3rd of the assets The Dutch SPEs and Tax Planning CPB Netherlands Bureau for Economic Policy Analysis 17 November 2018 3

Incoming and outgoing flows, sample, bln euro 200 160 180 140 160 120 140 Other income Other costs 120 100 Royalties Royalties 100 80 Payments from securities Payments on securities 80 Interest from other loans 60 Interest on other loans 60 Dividend from affiliates Dividend to affiliates 40 40 20 20 0 0 2004 2009 2013 2016 2004 2009 2013 2016 Incoming flows Outgoing flows The Dutch SPEs and Tax Planning CPB Netherlands Bureau for Economic Policy Analysis 17 November 2018 4

Country shares by incoming flow, 2016 SGP GER BRA IRE IRE LUX USA SGP USA ESP CHE UK Dividend Interest Royalty Dividend flows come often not from the host countries of investment • • Part of interest flows can be explained by high CIT rates in origin country, for an other part treaty shopping could be relevant • Royalty flows are only relevant for a few countries The Dutch SPEs and Tax Planning CPB Netherlands Bureau for Economic Policy Analysis 17 November 2018 5

Outgoing shares by destination and ubo, 2016 50 50 100 40 40 80 30 30 60 20 20 40 10 10 20 0 0 0 CHE UK USA LUX JPN BMU USA USA CAY UK IRE CHE LUX Dividend Interest Royalty Large flows are owned by US MNEs: to Switzerland, Luxembourg for dividend to Cayman Islands, Ireland and Luxembourg for interest to Bermuda for royalty The Dutch SPEs and Tax Planning CPB Netherlands Bureau for Economic Policy Analysis 17 November 2018 6

Bilateral flows have to be constructed B 1 A 1 A i B j C B m A n Multiple flows by SPE: proportional measure Multiple flow types : proportional measure Unequal incoming and outgoing flows: minimum value The Dutch SPEs and Tax Planning CPB Netherlands Bureau for Economic Policy Analysis 17 November 2018 7

Direct tax planning gain (without US) 30 25 20 Origin Interest Dest. Interest 15 10 Origin Royalty Dest. Royalty 5 0 2009 2013 2016 A B Large gains for royalty flows, modest gains for interest flows tax rates are weighted by the size of the flows The Dutch SPEs and Tax Planning CPB Netherlands Bureau for Economic Policy Analysis 17 November 2018 8

Tax planning gain: treaty shopping B 40 35 30 dividend direct dividend via NLD 25 20 interest direct C interest via NLD 15 10 royalties direct royalties via NLD 5 0 A 2009 2013 2016 Treaty shopping reduces effective tax rates on royalty flows We do not find lower rates on indirect dividend flows The Dutch SPEs and Tax Planning CPB Netherlands Bureau for Economic Policy Analysis 17 November 2018 9

Regressions = α + α + α + α + β + β + β + + ε Flow G G G T T T D ijt 0 1 it 2 jt 3 ij 1 it 2 jt 3 ijt t ijt Flow NL is bilateral dividend, interest or royalty via the Netherlands Many zero’s: Poisson, Pseudo ML Gravity and tax variables included Regressions by year: no panel About 8000 observations per year Gravity variables have expected sign (and significant) Tax haven dummies have also a positive effect on the flows The Dutch SPEs and Tax Planning CPB Netherlands Bureau for Economic Policy Analysis 17 November 2018 10

Regressions results coefficient bilateral tax rates Flows NL interest dividend royalty 2016 2009-2016 2016 2009-2016 2016 2009-2016 difference 4.533*** 7.392*** 10.15*** 5.998*** 20.19** 24.67*** direct -3.140 4.133*** 2.215 -4.065*** 5.696 25.23*** indirect -9.823*** -12.17*** -14.09*** -7.206*** -24.26* -23.40*** Expect positive sign for direct bilateral tax rate Expect negative sign for indirect bilateral tax rate Expect positive sign for difference A 10%-point higher direct tax rate, increases royalty flow by 250% The Dutch SPEs and Tax Planning CPB Netherlands Bureau for Economic Policy Analysis 17 November 2018 11

Conclusions • Dutch SPEs important for international diversion of financial flows – About 200 bln euro per year › Dividend flow form the larger part • The Netherlands is not the only conduit country in the chain – Luxembourg, Singapore, Ireland, … . – Traditional tax havens are also involved • A large share of these flows is ultimately owned by US MNEs • Tax factors play a major role: A higher tax rate on the direct route increases the flow via Dutch SPEs – Tax planning gain is in particular large for royalty flows The Dutch SPEs and Tax Planning CPB Netherlands Bureau for Economic Policy Analysis 17 November 2018 12

Assets and liabilities (sample), bln euro Category 2004 2009 2013 2016 Assets Affiliates 604 1263 1732 1798 Debt (including bonds) 495 837 931 969 Securities 5 86 90 96 Other assets 0 3 9 2 Total 1104 2189 2762 2865 Liabilities Equity by major shareholders 543 1185 1607 1716 Debt 280 579 636 681 Securities (including bonds) 298 476 449 513 Total 1121 2240 2692 2911 The Dutch SPEs and Tax Planning CPB Netherlands Bureau for Economic Policy Analysis 17 November 2018 13