The Missing Profits of Nations

Transcription

1 The Missing Profits of Nations Thomas Tørsløv (U. of Copenhagen) Ludvig Wier (U. of Copenhagen) Gabriel Zucman (UC Berkeley) June 2018

2 Introduction Why are corporate tax rates falling globally? Standard explanation: globalization competition to attract real capital race to the bottom But today s largest multinationals don t move much K across borders (don t even have much tangible K) Instead they shift accounting profits, including to places that collect 0 tax (Google in Bermuda) Tax competition model cannot explain this pattern Need to study profit shifting, why it rose and persists

3 This paper: New data and explanations First contribution is empirical: produce new series of global profit shifting using macro data. Key novelties: New database: profits of local v foreign corp in each ctry Complete map of where profits booked globally Direct estimate of size of profits shifted to havens Forensic analysis of tax haven data show out of which countries profits are shifted Improved macro stats: we provide estimates of GDP, profits, & factor shares corrected for shifting Second contribution is theoretical: provide new explanation for persistence of profit shifting

4 Our results Main empirical results: 40% of multinationals profits shifted to tax havens EU is the main loser; US firms are the main shifters Policy failure explains persistence of shifting: High-tax countries focus enforcement on transactions that shift profits to other high-tax places They ignore tax havens, where bulk of shifting occurs In effect, high-tax countries are stealing from each other while letting tax havens flourish

5 Implications for future of taxes and inequality Tax competition model: corporate tax rate 0 Capital moves race to bottom inevitable Progressive income tax will disappear (impossible to enforce with low corp. tax rate: the rich incorporate) Globalization fuels inequality Our results: corporate tax may rise in the future Capital does not move; paper profits do Policy failures explain this shifting Can be fixed corp tax could even if no coordinat o Domestic policies, more than globalization, are key

6 The size of global profit-shifting

7 Key challenge in the literature: Little data on profits in tax havens No reference estimate of size of global profit shifting Widely-used source (eg, by OECD 2015 for its official estimate): financial accounts micro-data (Orbis) log(π ict ) = α + β(1 τ ct ) + δfirm it + γcountry ct + ɛ ict Extrapolate global shifting from ˆβ Problem: limited reporting in tax havens most shifted profits not visible in financial accounts (i) ˆβ downward biased (ii) biased inferences about size and location of shifted profits

8 Most of Google s profits are invisible in available financial accounts data Bn. Google's profits in Orbis True global profits Sum of observable profits

9 Most of Apple s profits are invisible in available financial accounts data Bn. Apple's profits in Orbis True global profits Sum of observable profits

10 None of Facebook s profits are visible in available financial accounts data Bn. Facebook's profits in Orbis True global profits

11 Most of Nike s profits are invisible in available financial accounts data Bn. Nike's profits in Orbis True global profits Sum of observable profits

12 Only 17% of multinationals profits are visible in financial accounts micro-data Fraction of firms 30% % 10% The missing profits in Orbis Weighted average: 17% Weighted average: 0.17 = % 0% %.4 40% 60%.6 80%.8 100% 1 Share of global of profits found in in Note: This graph shows the imperfect coverage in Orbis. For each multinational firm we take the sum of profits made by all subsidiaries registered in Orbis and divide by the global profits of the same multinal firm. Whenever the share is lower than 1 this means that we only see part of the global profits in Orbis..

13 Our appraoch: we combine and analyze global macro data in a systematic way New national accounts data: Key novelty: exploit new foreign affiliates statistics to decompose profits into local vs. foreign firms Better than Orbis because relies on much more info. (including tax returns & census-like surveys) Improved balance of payments data: Bilateral trade & intra-group payments shows out of which countries profits are shifted Ultimate-owner direct investment statistics shows which multinationals shift profits

14 A new global database on profits (2015) Trillions of current US$ Global gross output (GDP) 75,038 Depreciation 11,940 Net output 63,098 % of net corporate profits Net corporate output 34, % Net corporate profits 11, % Net profits of foreign-controlled corp. 1,703 15% Of which: shifted to tax havens 616 5% Net profits of local corporations 9,812 85% Corporate income taxes paid 2,154 19%

15 Methodology to estimate amount of profits shifted to tax havens

16 How multinationals shift profits offshore Three ways to shift profits to low-tax countries: Manipulation of intra-group export and import prices (= transfer prices) Intra-group interest payments (tax deductible) Strategic location of intangibles We construct a macro indicator of profit-shifting that captures all channels of shifting

17 Conceptual framework Macro indicator of profit-shifting π Country s corporate output Y = F (K, AL) = rk + wl 2 types of corp: f (foreign) vs. l (local) Capital share α = rk/y Net interest paid = p% of rk Pre-tax profits/wage ratio: π = (1 p) α/(1 α) We analyze π for f vs. l firms in each country

18 Global patterns in corporate profitability Recorded profitability varies systematically across countries: Tax havens have abnormally high profits/wage ratios In tax havens: foreign firms are much more profitable than local firms (π f >> π l ) In non-haven countries: foreign firms are less profitable than local firms (π f < π l ) Clear evidence in macro statistics of shifting from high- to low-tax places

19 Firms in tax havens are abnormally profitable 300% Pre-tax corporate profits (% of compensation of employees ) 250% 200% 150% 100% 50% 0% Luxembourg Ireland Puerto Rico Singapore Hong Kong Netherlands Belgium Germany Italy United Kingdom Japan Spain Australia United States Switzerland Canada Average among non-havens: 36% France

20 High profits in tax havens stem from high profits of foreign firms 300% 250% Pre-tax corporate profits (% of compensation of employees) 200% All firms Local firms 150% 100% Non-haven average: 36% 50% 0% Luxembourg Ireland Puerto Rico Singapore Hong Kong Netherlands Belgium

21 In non-havens, foreign firms are always less profitable than local firms 60% Pre-tax corporate profits (% of compensation of employees) 50% Foreign firms Local firms 40% Average: 36% 30% 20% 10% 0% Germany Italy United Kingdom Spain Japan Australia United States France

22 Profits are offshore, losses onshore 1675% 800% Pre-tax corporate profits (% of compensation of employees) 600% 400% 200% 0% Puerto Rico Ireland Luxembourg Switzerland Singapore Hong Kong Netherlands Belgium United States Australia United Kingdom Spain Japan Foreign firms Local firms France Germany Italy

23 Our method to estimate the amount of profits shifted to tax havens Set π f Advantages: in havens equal to profitability local firms π l Simple and transparent Controls for country-level determinants of profitability in tax havens (e.g., anti-labor policies) Easy to track over time & space ( debt/gdp): could be monitored by policymakers to implement sanctions Potential concern: High capital intensity of foreign firms in tax havens?

24 Do machines move to low-tax places?

25 Testing the hypothesis that machines move to low-tax places Maybe tax havens attract highly capital-intensive industries from abroad: With Cobb-Douglas production, this does not affect π With CES production and σ > 1, high K/L high π Test using data on affiliates of US multinationals: US data more detailed than data of other countries (importantly: info on K) Large sample of US multinationals surveyed annually, universe every 5 years back to 1966

26 Tax haven affiliates of U.S. multinationals are abnormally profitable 800% 700% 600% 500% 400% 300% 200% Taxable corporate profits of affiliates of U.S. multinationals (% of compensation of employees) 100% Average among non-haven affiliates: 49% 0% Ireland Bermuda & Car. Luxembourg Switzerland Singapore Netherlands China Japan Mexico Belgium Canada India United Kingdom Italy Germany France Brazil

27 Paper profits move to tax havens. Machines don t. 800% 700% 600% Decomposing the taxable profits / compensation ratio (Majority-owned affiliates of U.S. multinationals) Haven / non-haven ratio 500% 400% 300% 200% 100% 0% K/wL r α/(1-α) 1-p π

28 Tax haven affiliates of US multinationals have been increasingly profitable 350% 300% Pre-tax profits of affiliates of U.S. multinationals (% of compensation of employees) 250% 200% Tax haven affiliates 150% 100% 50% Non-haven affiliates 0%

29 Globalization has been paper profits not machines moving to low-tax places 500% The profitability π of the affiliates of US multinationals (ratio of Haven affiliates / Non-haven affiliates) 400% π = (K/wL).r.(1-p) Operating surplus / physical capital (r) 300% 200% Net interest received (1-p) Physical capital / wage (K/wL) 100% 0%

30 Who Loses? Allocating the Shifted Profits

31 Tax havens run large trade and interest surpluses, all paid back to foreign parents Current account balance (% of national income) United States Australia Canada France Japan Spain Italy Belgium Germany Netherlands Singapore Puerto Rico Ireland Luxembourg Net trade surplus Net intra-group interest received Net intra-group profits received -80% -60% -40% -20% 0% 20% 40% 60% 80% 100%

32 How we allocate the shifted profits We follow destination of tax havens service exports and intra-group interest receipts Services: focus on royalties, management fees, ICT, fin. services most conducive of shifting Advantage of using tax haven data: capture services better than importers data ( 30% gap) The excess profitability (π f π l ) in havens match the amount of excess high-risk transaction with them Distribute excess profits prop. to these transactions E.U. countries are the main losers

33 The EU loses 20% of its corporate tax revenue, the US 15% 20% 18% 16% 14% Tax revenue lost due to profit shifting (% of corporate tax revenue collected) 12% 10% Global average: 10% 8% 6% 4% 2% 0% EU US Developing countries Rest of OECD

34 Which multinationals shift profits offshore? We track to which countries the profits booked in tax havens ultimately accrue: Allocate shifted profits prop. to direct investment equity income paid (dividends + retained earnings) Using new ultimate beneficial owner direct investment statistics Shows where the big shifters are headquartered U.S. multinationals are the biggest users of tax havens

35 Who shifts most? The US. Who loses most? EU & developing ctries 50% Allocating the profits shifted to tax havens % of total profits shifted to tax havens 40% 30% 20% 10% Where the shifted profits come from To whom the shifted profits accrue 0% EU US Developing countries Rest of OECD

36 Explaining the rise of profit-shifting

37 Beggar-thy-neighbor pays off Incentives of havens can explain the rise of shifting: With source taxation & no coordinat o or sanction, havens can earn revenue by attracting artificial bases Key result: revenue-max. rate 0 < τ <5%: havens with τ τ generate very large tax revenue Can explain the rise of the supply of tax avoidance schemes (e.g., tax rulings: Apple Ireland) Some countries have won and others lost from financial globalization ( textbook free-trade model)

38 Many havens collect a lot of tax revenue... 8% Corporate income tax revenue (% of national income) 7% 6% 5% 4% 3% 2% 1% 0% Malta Luxembourg Hong Kong Cyprus Japan Australia Ireland Switzerland Canada Belgium Korea Singapore Mexico United Kingdom Netherlands France Puerto Rico Spain Germany United States Italy Average among non-havens: 3.5% Poland

39 ... By applying very low rates to the huge artificial tax base they attract Corporate tax revenue collected & tax rate on shifted profits 100% 80% Revenue collected on shifted profits, % of total revenue Tax rate on shifted profits 60% 40% 20% 0% Malta Puerto Rico Ireland Luxembourg Cyprus Hong Kong Singapore Netherlands Switzerland

40 As profit shifting skyrocketed % Pre-tax corporate profits (% of compensation of employees) 200% 150% Ireland 100% 50% United States 0%

41 ...Tax revenue rose in many havens, while they or stagnated in high-tax countries Corporate income tax revenue (% net national income) 5% Ireland 4% United States 3% 2% 1% 0%

42 The lower the rate, the higher the revenue Corporate income tax revenue vs. tax rate in Ireland 5% 50 4% 40 3% 30 2% Tax revenue (left) (% of national income) Nominal tax rate (right) 20 1% 10 0%

43 Explaining the persistence of profit-shifting

44 The policy failure of high-tax countries Why have high-tax countries failed to protect their tax base? Incentives of tax havens can explain avoidance schemes, but not why high-tax countries have let their base shrink Our explanation: failure of tax enforcement In current current international tax system, tax authorities have perverse incentives They try to relocate base booked in other high-tax countries, not base shifted to havens

45 The incentive problem of tax authorities e1 re-located to France is worth the same to France whether it comes from Germany or Bermuda But much easier to relocate e1 booked in Germany: Feasible: information exists (Orbis) More likely to succeed: no push-back from firms Quick: cooperation via dispute settlement agreements Crowds out enforcement on havens: hard (no data), costly (legal defense by firms), lengthy (lack of cooperation) Analysis of transfer price corrections shows most enforcement is against other high-tax countries

46 Most transfer price enforcement is against other high-tax countries # of times country is among top 3 targets Countries most often targeted in transfer price disputes 0 United States Germany Japan Netherlands Switzerland France Korea UK Australia Austria Norway China Singapore Cayman Denmark Canada Czech Taiwan Finland Poland British virgin Hong Kong Panama Sweden

47 Can more cooperation and better information solve the problem? Facilitating dispute settlement can backfire: Ongoing initiative to cooperation among OECD countries Problem: crowds out enforcement on non-oecd havens, where bulk of shifting takes place Better information can help, but not enough: Even with perfect info, firms will always fight more to protect profits they book in low-tax places Internalizing this, tax authorities will keep going after high-tax places

48 Even when tax havens cooperate, tax authorities do not target them % of total Counterpart in Mutual Agreement Procedures in the EU 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% EU high tax country EU tax haven

49 Multinationals outspend tax authorities 1000' employees Governments vs. corporate transfer pricing specialists Wage bill 15 Billion Government Private Source is LinkedIn, but the government count is corroborated by the EY Transfer Pricing Tax Authority Survey (2014). The wage bill is estimated by applying the average salary of an EY Transfer Pricing Specialist (Source: Glassdoor).

50 Conclusion

51 Main findings Main results: 40% of multinationals profits shifted to tax havens E.U. is the main loser; U.S. the main shifter High losses for the EU can be explained by failure of enforcement due to perverse incentives Tax competition model not enough to explain in τ K Policies are key to understand rise & persistence of shifting & in turn decline in corp tax rate

52 There is a policy solution to profit shifting Apportionment of profits proportionally to where sales are made Removes incentives to shift profits and more real activity Works reasonably well for US States Can be done unilaterally Would increase corp tax revenue by 20% in Europe and 15% in U.S.

53 Improving international statistics Our analysis highlights a number of statistical gaps: Foreign affiliates statistics: need to be compiled by more countries & broader (e.g., K, interest, tax) FDI asymmetries: need exchange of micro-data between national statistical authorities Missing national accounts of the Caribbean Corporate registry in the U.S. and havens Decline of the corporate labor share throughout the world is under-estimated

54 Supplementary slides

55 Previous macro approaches A nascent literature takes a macro perspective: UNCTAD (2015) global estimate based on FDI data Clausing (2009), Zucman (2014), Guvenen et al. (2017) for U.S. Pro: does not suffer from Orbis limitations Problems: Hard to infer amount of taxes avoided Hard to infer which countries lose/gain revenues Need to open the black-box of tax havens

56 Foreign affiliates statistics New data: foreign affiliates statistics (FATS) Main national accounts aggregates for affiliates of multinationals (inward and outward) Compiled for a long time in the US Introduced recently in a number of other countries, including EU havens When not available: use direct investment income statistics (BoP) and counterpart country FATS

57 Imputation of profits in foreign firms when no FATS exist Compute profits in foreign firms using direct investment income flows 10% vs. 50% ownership threshold; pre-tax vs. post-tax impute taxes Assume profits / wage same as for US affiliates Imputation when no direct investment income data exist: Estimate direct investment income paid such that world DI income balances to 0 Two reasons why global DI income > 0: missing US profits in Ireland etc.; missing BoP we impute both

58 Anomalies in the world balance of payments Billions of current U.S. dollars The world current account discrepancies Trade balance (missing service imports from tax havens; missing goods imports in China) Direct investment income balance (missing income of US affiliates; Caribbean) -200 Portfolio & other income balance (offshore household wealth)

59 The unrecorded profits of U.S. affiliates in tax havens 300% 250% Pre-tax corporate profits (% of compensation of employees ) 200% 150% 100% 50% 0% Luxembourg Ireland Missing profits of other multinationals Missing profits of U.S. multinationals As reported in the national accounts Average among non-havens: 36% Puerto Rico Singapore Hong Kong Netherlands Belgium Germany Italy United Kingdom Japan Spain Australia United States Switzerland Canada France

60 Service imports from tax havens are under-estimated by importers (B2C sales) Bn. The missing service exports of the six EU tax havens Exports to EU22 recorded by exporter Imports recorded by EU Luxembourg Ireland Belgium Netherlands Malta Cyprus

61 At least 30% of the services exported by EU havens go unreported by the importer 60% Missing service exports, % of total service exports 50% 40% 30% 20% 10% 0% -10% EU22 EU6 Luxembourg Ireland Belgium Netherlands Malta Cyprus Note: Service exports include exports to all EU22 countries (EU26 minus Luxembourg, Ireland, Belgium, Netherlands, Malta, Cyprus).

62 Tax haven firms are abnormally profitable within each sector 800% Taxable corporate profits (% of compensation of employees) 700% 600% Tax havens Non havens 500% 400% 300% 200% 100% 0% Chemical Computers Wholesale trade Information Finance Professional services Other

63 Profits are offshore, losses are onshore 350% 300% 250% Pre-tax profits (% of compensation of employees) 200% Tax haven affiliates of U.S. multinationals 150% 100% 50% Non-haven affiliates of U.S. multinationals 0% % U.S. affiliates of foreign multinationals

64 As settlement is facilitated, high-tax to high-tax disputes are growing Number of cases 7,000 Number of mutual agreement procedures in the OECD 6,000 Inventory of mutual agreement procedures 5,000 4,000 3,000 2,000 New mutual agreement procedures 1,