How do we tax the digital economy? Generally speaking, the OECD is proposing the shift of taxable profit from low-tax jurisdictions to market jurisdictions. Net worldwide taxable profit would be unchanged. Net worldwide tax revenue would rise. Exactly how this will happen remains a mystery, but for the first time the OECD has provided details of a possible proposal. When we estimated that proposal, using real (but far from perfect) data, the following facts emerged:
The United States is likely to gain revenue.
Amazon, a premier digital multinational business, could be unaffected by the new regime.
Tax havens, as expected, suffer revenue losses while market countries benefit.
The fuzzy and yet-to-be-determined scope of any proposal will have a major effect.
Reallocating Residual Profits
On February 13 OECD economists held a webcast and presented slides outlining their data, their methods, and some preliminary results of the revenue and economic effects of the base erosion and profit-shifting pillar 1 and 2 proposals. Pillar 1 concerns taxation of digital businesses. Pillar 2 concerns minimum tax proposals roughly similar to the global intangible low-taxed income and base erosion and antiabuse tax provisions introduced in the Tax Cuts and Jobs Act.
Pillar 1 has three parts: Amount A would reallocate profits with a newly developed taxing right for market jurisdictions. Amount B would use a simplified, fixed percentage (or percentages) of sales to determine returns for routine marketing and distribution functions. Amount C concerns dispute resolution. The calculations in this article apply only to revenue estimates of amount A.
Each multinational (or business segment of a multinational) that is “in scope” (that is, is large enough and has specific business characteristics) must follow several steps to estimate its amount A reallocation of worldwide profit. First, each in-scope multinational must calculate residual (or non-routine) profit. This is equal to the excess, if any, of worldwide profits before tax (from financial statements, not tax returns) minus X percent of sales. The news in the February 13 presentation is that the OECD created estimates for X equal to 10 percent and X equal to 20 percent. Estimates here present results using the assumption yielding greater non-routine profit, that is, X equal to 10 percent. So in addition to being in scope, a multinational must have profit before tax (PBT) in excess of 10 percent of sales for the proposal to apply.
Second, only a portion of residual profit — presumably corresponding with the share of total profit from intangibles attributable to marketing intangibles — is reallocated. The other bit of news is that the OECD assumed that share to be 20 percent. (This will be disappointing to many taxpayers that commonly used 10 percent in their examples.) There seems to be no reasoned justification for the choice of that specific percentage.
Once the reallocable residual profit is determined, it is reallocated either to or from a jurisdiction based on whether the excess of S% over P% is positive or negative. S% is the share of a multinational’s sales in a jurisdiction as a percentage of worldwide sales of that multinational. P% is the share of a multinational’s profit in a jurisdiction as a percentage of worldwide sales of that multinational. So corresponding with the formula (“Pillar 1 revenue effects”) on slide 29 of the OECD presentation, we can calculate each country’s share of tax revenue gain or loss as the amount:
Tax rate * 20 percent * (PBT - 10 percent of sales) * (S% - P%)
with the tax rate calculated for each country, and PBT, sales, S%, and P% calculated for each multinational in each jurisdiction.
It is possible to obtain reasonably good estimates of the U.S.-versus-foreign reallocation of profit from financial data in annual reports (assuming all business of any given multinational is in scope). Annually reported data include worldwide sales, worldwide PBT, U.S. sales, and U.S. PBT. For the calculations here, we assume the appropriate U.S. tax rate is 21 percent. With this information we can calculate U.S. revenue gain or loss for each country, assuming the OECD’s illustrative proposal applied in a prior year.
Let’s run through the calculations for Apple as an example. In 2017 Apple had $229.2 billion of worldwide sales revenue and $64.1 billion of PBT. The ratio of PBT to sales was 28 percent. Any percentage greater than 10 percent indicates there is non-routine profit to be reallocated. Non-routine profit equals before-tax profit in excess of 10 percent of sales (that is, $64.1 billion - 0.1 * $229.2 billion = $41.2 billion). The market allocation fraction is assumed to be 20 percent, so $8.24 billion (0.2 * $41.2 billion) is the total amount of profit to be reallocated worldwide.
The U.S. percentage of worldwide sales is 36.8 percent. The U.S. percentage of worldwide profit is 26.7 percent. The difference, 10.1 percent, is the portion of reallocable profit to be reallocated to the United States. (The difference can be positive or negative, so profit can be allocated either to or from any jurisdiction.) The amount of profit reallocated to the United States is $832 million (10.1 percent of $8.24 billion). At a 21 percent tax rate, the estimated extra tax that Apple would pay to the IRS if the OECD’s illustrative proposal applied to all the business conducted by Apple would be $175 million.
Table 1 shows the estimated change in the U.S. income tax base due to the amount A reallocation for 25 multinational corporations using this method. (These multinationals were chosen unscientifically for various reasons: They were among the world’s largest or had significant excess profit or to illustrate the result for one member of a specific industry.)
A January 29 report from the OECD explained that financial service industries (with perhaps some exceptions for the consumer-facing component of their business) and natural resource extraction industries would be carved out of the proposal. It also explained that (1) digital service businesses and (2) businesses with “active and sustained engagement in a market jurisdiction, beyond the mere conclusion of sales, without necessarily investing in local infrastructure and operations” would need to calculate amount A reallocations. (OECD, “Statement by the OECD/G20 Inclusive Framework on BEPS on the Two-Pillar Approach to Address the Tax Challenges Arising From the Digitalization of the Economy” (Jan. 29-30, 2020).)
Unfortunately, how these qualitative thresholds will be defined is unclear. So which corporations in Table 1, and which of their business components, would be in scope we don’t try to determine here. We just do quantitative work. But it is easy to see that even with this small sample, large amounts of revenue are at stake depending on how these squishy terms are defined.
In Table 1, Amazon and Caterpillar do not have any amount A reallocations because they have no residual profits. In 2017 Amazon profits before tax are 2 percent of sales. (In 2019 Amazon is still below threshold, with profits before tax equal to 5 percent of sales.) Most or all of Amazon’s businesses would seem to fit comfortably under the OECD’s conception of business in scope of amount A because they provide digital content, an online marketplace, and cloud computing web services.
Foreign Reallocations
Foreign-country-specific data are not usually available from published financial reports. To fill in the blanks — admittedly imprecisely — we have applied available aggregate 2017 data to distribute each corporation’s foreign share of sales and profits. Two sets of data are available. The first is on the foreign activities of U.S. multinationals from the Commerce Department Bureau of Economic Analysis. The second is the country-by-country reporting data published by the IRS Statistics of Income division. Prior articles in these pages have discussed significant weaknesses in these data. (Prior analysis: Tax Notes Federal, Jan. 6, 2020, p. 26; and Tax Notes Federal, Jan. 13, 2020, p. 198.)
The estimated reallocation of foreign profits using each data set is shown in tables 2A and 2B. These estimates show, as expected, a reallocation of profit tax havens like Luxembourg, Netherlands, and the Cayman Islands. (But the picture is not so clear-cut for Singapore, Ireland, and Bermuda.) Significant revenue would be shifted into larger countries among those shown, including Germany, France, Brazil, and China.
Company Name | Profit Before Tax (PBT) | Sales (S) | PBT/S | Non-Routine Profit (PBT - 0.1 * S) | U.S. Share of Sales (S%) | U.S. Share of Profits (P%) | Percentage to U.S. (S% - P%) | Change in U.S. Tax Base |
|---|---|---|---|---|---|---|---|---|
Apple | $64.1 | $229.2 | 28% | $41.2 | 37% | 27% | 10% | $0.8 |
JPMorgan Chase | $35.9 | $100.7 | 36% | $25.8 | 77% | 75% | 2% | $0.1 |
Alphabet | $27.2 | $110.9 | 25% | $16.1 | 47% | 36% | 11% | $0.4 |
$20.6 | $40.7 | 51% | $16.5 | 44% | 34% | 9% | $0.3 | |
Intel | $20.4 | $62.8 | 32% | $14.1 | 20% | 55% | -35% | -$1 |
Microsoft | $29.9 | $96.6 | 31% | $20.2 | 53% | 23% | 30% | $1.2 |
Johnson & Johnson | $17.7 | $76.5 | 23% | $10 | 52% | 28% | 25% | $0.5 |
Micron Technology | $5.2 | $20.3 | 26% | $3.2 | 14% | -1% | 15% | $0.1 |
Walt Disney | $13.8 | $55.1 | 25% | $8.3 | 76% | 91% | -16% | -$0.3 |
PepsiCo | $9.6 | $63.5 | 15% | $3.2 | 62% | 36% | 26% | $0.2 |
Broadcom | $1.8 | $17.6 | 10% | $0.1 | 7% | 115% | -108% | $0 |
Pfizer | $12.3 | $52.5 | 23% | $7.1 | 50% | -56% | 105% | $1.5 |
Boeing | $10.1 | $94 | 11% | $0.7 | 44% | 96% | -51% | -$0.1 |
Visa | $11.7 | $18.4 | 64% | $9.9 | 47% | 72% | -25% | -$0.5 |
Amazon.com | $3.8 | $177.9 | 2% | $0 | 66% | 148% | -82% | $0 |
Procter & Gamble | $13.3 | $65.1 | 20% | $6.8 | 42% | 68% | -26% | -$0.4 |
IBM | $11.4 | $79.1 | 14% | $3.5 | 38% | 5% | 33% | $0.2 |
Amgen | $9.6 | $22.8 | 42% | $7.3 | 73% | 46% | 27% | $0.4 |
Altria Group | $9.8 | $25.6 | 38% | $7.3 | 100% | 100% | 0% | $0 |
American Express | $7.4 | $39 | 19% | $3.5 | 74% | 86% | -12% | -$0.1 |
Coca-Cola | $6.7 | $35.4 | 19% | $3.2 | 42% | -10% | 52% | $0.3 |
Merck | $6.5 | $40.1 | 16% | $2.5 | 43% | 53% | -10% | -$0.1 |
Caterpillar | $4.1 | $45.5 | 9% | $0 | 41% | 6% | 35% | $0 |
McDonald’s | $8.6 | $22.8 | 38% | $6.3 | 35% | 26% | 9% | $0.1 |
Starbucks | $4.3 | $17.7 | 24% | $2.6 | 74% | 79% | -5% | $0 |
Sum of 25 above | $365.8 | $1,609.7 |
| $219.8 |
|
|
| $3.6 |
Source: Author’s calculations using Form 10-K reports. U.S. Tax Base Change = 20% * (PBT - 10% of S) * (S% - P%). Assuming, as is highly unlikely, all segments of all 25 corporations are “in scope.” | ||||||||
While tables 2A and 2B show estimated taxable profit shifted among foreign jurisdictions, Table 3 shows estimated revenue (applying country-specific tax rates) for the four big-tech “GAFA” companies (Google, Apple, Facebook, and Amazon) and a synthetic composite for all 25 companies combined. The estimates show revenue gains for most countries (including the United States) except for most tax havens that suffer revenue losses. Worldwide, there are net revenue gains for governments and tax increases for corporations (except for Amazon and Caterpillar, which do not have residual profits to reallocate under amount A).
Country | Sales (S) | Profit Before Tax - Adjusted (PBT) | PBT/S | Country Sales as Percentage of All Foreign Sales (S%) | Country Profits as Percentage of All Foreign Profit (P%) | S% - P% (reallocation of non-U.S. excess profit) |
|---|---|---|---|---|---|---|
All non-U.S. | $6,221.4 | $736.2 | 11.8% | 100% | 100% | 0% |
United Kingdom | $643.5 | $70.5 | 11% | 10.3% | 9.6% | 0.8% |
Canada | $587.5 | $35.7 | 6.1% | 9.4% | 4.9% | 4.6% |
Singapore | $437.3 | $39.2 | 9% | 7% | 5.3% | 1.7% |
China | $375.6 | $28.2 | 7.5% | 6% | 3.8% | 2.2% |
Ireland | $369 | $102.6 | 27.8% | 5.9% | 13.9% | -8% |
Switzerland | $351.4 | $44.9 | 12.8% | 5.6% | 6.1% | -0.4% |
Germany | $339 | $12.3 | 3.6% | 5.4% | 1.7% | 3.8% |
Netherlands | $289.8 | $99.8 | 34.4% | 4.7% | 13.6% | -8.9% |
Mexico | $249.4 | $21.7 | 8.7% | 4% | 2.9% | 1.1% |
Japan | $235.3 | $25.3 | 10.8% | 3.8% | 3.4% | 0.3% |
France | $200.2 | $6.6 | 3.3% | 3.2% | 0.9% | 2.3% |
Brazil | $178.1 | $3.1 | 1.8% | 2.9% | 0.4% | 2.4% |
Belgium | $161 | $9.2 | 5.7% | 2.6% | 1.2% | 1.3% |
Australia | $153.4 | $10 | 6.5% | 2.5% | 1.4% | 1.1% |
Hong Kong | $140.3 | $12 | 8.5% | 2.3% | 1.6% | 0.6% |
Italy | $114.7 | $6.3 | 5.5% | 1.8% | 0.8% | 1% |
India | $94.1 | $10.5 | 11.2% | 1.5% | 1.4% | 0.1% |
Spain | $83.6 | $5.5 | 6.6% | 1.3% | 0.8% | 0.6% |
Luxembourg | $82.4 | $29.9 | 36.2% | 1.3% | 4.1% | -2.7% |
South Korea | $70.9 | $6.4 | 9% | 1.1% | 0.9% | 0.3% |
Cayman Islands | $62 | $31.7 | 51.1% | 1% | 4.3% | -3.3% |
Thailand | $56.2 | $5.9 | 10.4% | 0.9% | 0.8% | 0.1% |
Bermuda | $53.6 | $9.9 | 18.5% | 0.9% | 1.3% | -0.5% |
Malaysia | $47 | $3.8 | 8.1% | 0.8% | 0.5% | 0.2% |
Total of 24 above | $5,375.2 | $630.7 | 11.7% | 86.4% | 85.7% | 0.7% |
All other | $846.2 | $105.5 | 12.5% | 13.6% | 14.3% | -0.7% |
Source: Author’s calculations using aggregate Bureau of Economic Analysis data for 2017. PBT is computed by subtracting 80 percent of “income from equity investments” (suspected of being double counted) from net income. (Prior coverage: Tax Notes Federal, Jan. 6, 2020, p. 26.) | ||||||
Country | Sales (S) | Profit Before Tax - Adjusted (PBT) | PBT/S | Country Sales as Percentage of All Foreign Sales (S%) | Country Profits as Percentage of All Foreign Profit (P%) | S% - P% (reallocation of non-U.S. excess profit) |
|---|---|---|---|---|---|---|
All foreign | $6,827 | $638 | 9% | 100% | 100% | 0% |
United Kingdom | $673 | $18 | 3% | 9.9% | 2.8% | 7% |
Canada | $516 | $32 | 6% | 7.6% | 5% | 2.6% |
Singapore | $530 | $55 | 10% | 7.8% | 8.6% | -0.8% |
China | $393 | $27 | 7% | 5.8% | 4.2% | 1.6% |
Ireland | $520 | $29 | 6% | 7.6% | 4.6% | 3% |
Switzerland | $438 | $49 | 11% | 6.4% | 7.7% | -1.3% |
Germany | $338 | $7 | 2% | 5% | 1.1% | 3.9% |
Netherlands | $339 | $40 | 12% | 5% | 6.3% | -1.3% |
Mexico | $220 | $16 | 7% | 3.2% | 2.4% | 0.8% |
Japan | $236 | $25 | 11% | 3.5% | 3.9% | -0.4% |
France | $205 | $5 | 2% | 3% | 0.8% | 2.2% |
Brazil | $177 | $6 | 3% | 2.6% | 0.9% | 1.7% |
Belgium | $161 | $6 | 4% | 2.4% | 0.9% | 1.4% |
Australia | $143 | $15 | 10% | 2.1% | 2.3% | -0.2% |
Hong Kong | $166 | $12 | 7% | 2.4% | 1.9% | 0.5% |
Italy | $109 | $7 | 7% | 1.6% | 1.1% | 0.5% |
India | $102 | $12 | 12% | 1.5% | 1.8% | -0.4% |
Spain | $85 | $4 | 4% | 1.2% | 0.6% | 0.7% |
Luxembourg | $124 | $25 | 20% | 1.8% | 3.9% | -2.1% |
South Korea | $73 | $5 | 7% | 1.1% | 0.8% | 0.2% |
Cayman Islands | $87 | $59 | 67% | 1.3% | 9.2% | -7.9% |
Thailand | $57 | $6 | 10% | 0.8% | 0.9% | -0.1% |
Bermuda | $78 | $32 | 42% | 1.1% | 5.1% | -3.9% |
Malaysia | $48 | $5 | 10% | 0.7% | 0.7% | 0% |
Total of 24 above | $5,822 | $496 | 9% | 85.3% | 77.6% | 7.6% |
All other | $1,005 | $143 | 14% | 14.7% | 22.4% | -7.6% |
Source: Author’s calculations using aggregate Statistics of Income country-by-country data for 2017. | ||||||
Country | 25 Companies Aggregated | Alphabet | Amazon.com | Apple | |
|---|---|---|---|---|---|
United States | $784 | $75 | $0 | $64 | $175 |
United Kingdom | $163 | $13 | $0 | $14 | $40 |
Canada | $57 | $4 | $0 | $5 | $14 |
Singapore | -$22 | -$2 | $0 | -$2 | -$5 |
China | $41 | $3 | $0 | $4 | $11 |
Ireland | $59 | $4 | $0 | $5 | $15 |
Switzerland | -$31 | -$3 | $0 | -$3 | -$7 |
Germany | $198 | $16 | $0 | $17 | $48 |
Netherlands | -$35 | -$3 | $0 | -$3 | -$8 |
Mexico | $16 | $1 | $0 | $2 | $4 |
Japan | -$50 | -$4 | $0 | -$4 | -$11 |
France | $133 | $10 | $0 | $12 | $32 |
Brazil | $83 | $6 | $0 | $7 | $20 |
Belgium | $50 | $4 | $0 | $4 | $12 |
Australia | -$21 | -$2 | $0 | -$2 | -$5 |
Hong Kong | $4 | $0 | $0 | $0 | $1 |
Italy | $28 | $2 | $0 | $3 | $7 |
India | -$48 | -$4 | $0 | -$4 | -$11 |
Spain | $23 | $2 | $0 | $2 | $6 |
Luxembourg | -$5 | $0 | $0 | $0 | -$1 |
South Korea | $4 | $0 | $0 | $0 | $1 |
Cayman Islands | -$10 | -$1 | $0 | -$1 | -$2 |
Thailand | -$9 | -$1 | $0 | -$1 | -$2 |
Bermuda | -$15 | -$1 | $0 | -$1 | -$4 |
Malaysia | -$5 | -$1 | $0 | $0 | -$1 |
Total of above 24 | $613 | $44 | $0 | $55 | $155 |
All other non-U.S. | -$382 | -$32 | $0 | -$33 | -$90 |
United States | $784 | $75 | $0 | $64 | $175 |
All non-U.S. | $231 | $11 | $0 | $22 | $65 |
Total tax change | $1,015 | $86 | $0 | $86 | $239 |
Source: Author’s calculations using Bureau of Economic Analysis data, including tax rates computed from BEA data, and calculations from Table 1. | |||||
Table 4 shows company-by-company estimates of U.S. and non-U.S. companies for the two different data sets. The U.S. revenue changes reasonably indicate what might have occurred in 2017 if the OECD’s illustrative proposal was in effect and the U.S. tax rate was 21 percent. The foreign revenue changes are speculative because they are based on aggregate data, not taking into account company-by-company differences in the composition of foreign profits and sales. Estimates of additional U.S. tax liabilities are greatest for companies with large residual profits and with significant excess of U.S. sales percentage (S%) over U.S. profit percentage (P%) — like Apple, Microsoft, and Pfizer.
Better Data
With access to confidential, company-specific data sources, the OECD and government tax collectors will have better data than used here. But as the OECD repeatedly emphasizes, existing data sources are incomplete for the task at hand. Given year-to-year fluctuations in profitability and the allocation of profits, the difficulty of defining businesses and business segments that are in scope, and difficulties in determining the location of sales, it is hard to determine anything more than general trends and orders of magnitude when estimating the revenue effects of amount A.
| Assuming Distribution of Foreign Profit Same as Aggregate 2017 BEA Data | Assuming Distribution of Foreign Profit Same as Aggregate 2017 IRS-SOI Country-by-Country Data | ||||
|---|---|---|---|---|---|---|
United States | All Non-U.S. | Worldwide Total | United States | All Non-U.S. | Worldwide Total | |
Apple | $175 | $169 | $343 | $175 | $65 | $239 |
JPMorgan Chase | $17 | $54 | $71 | $17 | $28 | $44 |
Alphabet | $75 | $43 | $118 | $75 | $11 | $86 |
$64 | $59 | $123 | $64 | $22 | $86 | |
Intel | -$205 | $262 | $57 | -$205 | $187 | -$19 |
Microsoft | $255 | -$67 | $188 | $255 | -$84 | $171 |
Johnson & Johnson | $104 | -$17 | $87 | $104 | -$28 | $75 |
Micron Technology | $19 | $17 | $36 | $19 | $6 | $26 |
Walt Disney | -$54 | $58 | $4 | -$54 | $43 | -$11 |
PepsiCo | $36 | -$10 | $25 | $36 | -$12 | $23 |
Broadcom | -$3 | $3 | $0 | -$3 | $2 | -$1 |
Pfizer | $312 | -$171 | $141 | $312 | -$154 | $158 |
Boeing | -$15 | $15 | -$1 | -$15 | $11 | -$4 |
Visa | -$104 | $128 | $23 | -$104 | $92 | -$13 |
Amazon.com | $0 | $0 | $0 | $0 | $0 | $0 |
Procter & Gamble | -$74 | $93 | $19 | -$74 | $66 | -$8 |
IBM | $48 | -$8 | $40 | $48 | -$13 | $34 |
Amgen | $82 | -$34 | $49 | $82 | -$34 | $48 |
Altria Group | $1 | $0 | $0 | $1 | $0 | $0 |
American Express | -$18 | $22 | $4 | -$18 | $16 | -$2 |
Coca-Cola | $70 | -$26 | $43 | $70 | -$28 | $42 |
Merck | -$11 | $23 | $12 | -$11 | $15 | $4 |
Caterpillar | $0 | $0 | $0 | $0 | $0 | $0 |
McDonald’s | $24 | $29 | $53 | $24 | $12 | $36 |
Starbucks | -$5 | $11 | $6 | -$5 | $7 | $2 |
Source: Author’s calculations using Bureau of Economic Analysis data (including tax rates computed from BEA data), SOI country-by-country data (including tax rates computed from SOI data of firms with positive income), and calculations from Table 1. | ||||||