The (Not So) Simple Analytics of the New Amount A

As outlined in the OECD blueprint, the original purpose of pillar 1 was to respond to the problem that “in an increasingly digital age, in-scope businesses are able to generate profits through participation in a significant/active and sustained way in the economic life of a jurisdiction, beyond the mere conclusion of sales, with or without the benefit of local physical presence.”¹ The OECD’s proposed solution to that problem was amount A, which would allocate a portion of the global residual in-scope profit (GRIP) of multinational enterprises to market jurisdictions, with the allocation key being a market jurisdiction’s share of MNE in-scope destination-based sales. In contrast to the current system in which MNE global profits are taxed where the profits are earned — that is, source taxation by host countries — or where MNE shareholders or headquarters are located — that is, residence taxation by home countries — amount A would move some portion of MNEs’ global profits from source and residence jurisdictions to market jurisdictions.

How is the new amount A, as defined in the OECD’s July 2022 progress report,² different from the old amount A, as outlined in earlier OECD documents?³ How do the versions differ and how might the differences matter — for example, in terms of creating winners and losers?

To answer those questions, I start with the simple analytics of the old and new amount A formulas and compare them using numerical examples. I then compare old and new amount A in terms of their implications for winners and losers, opportunities for governments and MNEs to game the formulas, and mechanisms for tax base relief. There are major differences: New amount A is less aligned with the original goals of pillar 1; is more difficult to estimate, implement, and administer; offers greater opportunities for manipulation; and is less likely to benefit low-income countries.

The Simple Analytics of the Old Amount A

The Old Formula

The original amount A proposal in the impact assessment and pillar 1 blueprint starts by determining which MNEs are in scope based on their global revenues, profitability, and industry or activities. Financial accounting adjustments are used to create an adjusted measure of global pretax profit for in-scope MNEs. From that measure a percentage of the MNE’s global revenue is deducted as a proxy for routine returns. The remainder is multiplied by a fixed reallocation percentage to determine the total available for reallocation, referred to as “amount A.”⁴

Under the amount A formula, the net impact on the tax revenues of a specific jurisdiction (J) can be estimated as⁵:

Where Ʃ indicates that the variable is calculated at the global, not jurisdictional, level:

• A is the MNE’s global residual in-scope profit (ƩGRIP), determined by subtracting a profitability threshold (equal to a percentage of MNE global revenues) from the MNE’s global pretax profit;

• B is the reallocation percentage, which when multiplied by component A, determines the total MNE tax base available for reallocation;⁶

• C is the tax-base-receiving allocation key, defined as the ratio of the MNE’s global in-scope destination-based sales (GIDS) in a jurisdiction⁷ divided by GIDS for the MNE as a whole (ƩGIDS) — that is, a jurisdiction’s GIDS/ƩGIDS share;

• D is the tax rate to be levied on the tax base received by a market jurisdiction;

• E is the tax-base-relieving allocation key, defined as the ratio of the MNE’s GRIP in a jurisdiction divided by global residual in-scope profit for the MNE as a whole (ƩGRIP) — that is, a jurisdiction’s GRIP/ƩGRIP share;⁸ and

• F is the rate of tax base relief (via credit or exemption) to be provided by a tax-base-relieving jurisdiction to tax-base-receiving jurisdictions.

Conceptually, equation 1a can be broken into two parts that are multiplied together to determine a jurisdiction’s net gain or loss in tax revenue:

• A * B = amount A, the portion of MNE global residual in-scope pretax profit to be reallocated among jurisdictions; and

• (C * D) - (E * F) = a jurisdiction’s positive or negative net share of amount A.

We can reorganize the components in equation 1a as the difference between two amounts:

The first bracket in equation 1b measures a tax-base-receiving jurisdiction’s revenue gain, which equals its tax base gain (A * B * C) multiplied by its tax rate (D). The second bracket measures a tax-base-relieving jurisdiction’s revenue loss, which is its tax base loss (A * B * E) multiplied by its tax rate (F).

Table 1. Numerical Example of Old Amount A

Estimate of Amount A Available for Reallocation Among Jurisdictions
Global revenue (ƩR)		2,000
Adjusted pretax profit (ƩP)		400
Profitability threshold	0.1 * ƩR	200
Residual profit (ƩGRIP)	ƩGRIP = ƩP - 0.1 * ƩR	200
Reallocation percentage (RP)		25%
Amount A	0.25 * ƩGRIP	50
Estimated Change in Jurisdiction J’s Tax Base
Revenue in J (R)		1,000
Adjusted pretax profit (P)		150
Residual profit in J (GRIP)	GRIP = P - 0.1 R	50
Calculation using the C - E gap
Component C	Component C = GIDS/ƩGIDS = R/ƩR	50%
Component E	Component E = GRIP/ƩGRIP	25%
Change in J’s tax base	Amount A * (GIDS/ƩGIDS - GRIP/ƩGRIP)	12.5
Calculation Using Gross vs. Net Amount A
Tax base receipt	Amount A * component C	25
Tax base relief	Amount A * component E	12.5
Change in J’s tax base	Net amount A allocation	12.5

What is implicit in equation 1b is that market jurisdictions are also residence and source jurisdictions. In other words, the typical country is a host to foreign MNEs and a home to its own MNEs, with both home and host MNEs selling both there and abroad, as well as buying from other countries. The reality of the 21st-century global economy is that domestic and foreign MNEs are ubiquitous in most developed and developing countries. Thus, to avoid double taxation, the jurisdictions receiving taxing rights under amount A must subtract their existing taxing rights.

In algebraic terms, there is no difference between equations 1a and 1b. Equation 1a multiplies amount A by a jurisdiction’s net share ((C * D) - (E * F)), which may be positive or negative, to determine its net gain or loss in tax revenues. Equation 1b calculates a jurisdiction’s revenue gain as a tax-base-receiving jurisdiction and subtracts its revenue loss as a tax-base-relieving jurisdiction. Mathematically, either method yields the same result, as illustrated in Table 1. I discuss later how equations 1a and 1b, while identical on the surface, can be quite different in practice, especially if governments can game the amount A formula.

The formula can also be simplified, with the benefit that empirical estimates can be more easily calculated, if we assume that jurisdiction J levies the same corporate rate on its tax base gained and lost under amount A — that is, components D and F are identical and set equal to t, the jurisdictional corporate rate.⁹ With that assumption, we have:

Table 2. Empirical Publications on Amount A

Publication	Country	Industry
OECD, “Tax Challenges Arising From Digitalisation — Economic Impact Assessment” (Oct. 2020)	Global	ADS & CFB
Allison Christians, “Taxation of the Digital Economy: Preliminary Analysis of OECD Pillar 1 Impact Assessment + KPMG Transfer Pricing Study of Amounts B & C (Presentation Slides)” (Mar. 8, 2020)	Illustration	ADS & CFB
Lorraine Eden, “A Leap of Faith: The Economic Impact Assessment of the Pillar One and Pillar Two Blueprints,” 49 Tax Mgmt. Int’l J. 591 (Dec. 2020)	Illustration	ADS & CFB
Eden, “Pillar One Tax Games,” 50 Tax Mgmt. Int’l J. 4 (Dec. 2020)	Illustration	ADS & CFB
Eden, “Winners and Losers: The OECD’s Economic Impact Assessment of Pillar One,” 49 Tax Mgmt. Int’l J. 597 (Dec. 2020)	Global	ADS & CFB
Eden, “Canada and the United States: Winners or Losers From Pillar One Amount A?” 50 Tax Mgmt. Int’l J. 143 (Mar. 2021)	Global	ADS & CFB
Eden, “The Simple Analytics of Pillar One Amount A,” 50 Tax Mgmt. Int’l J. 137 (Mar. 2021)	Illustration	ADS & CFB
Eden, “Taxing the Top 100: U.S. Estimates of Winners and Losers From Pillar One Amount A,” 50 Tax Mgmt. Int’l J. 301 (June 2021)	U.S.	Top 100
Eden, “Winners and Losers: U.S. Country and Industry Estimates of Pillar One Amount A,” 50 Tax Mgmt. Int’l J. 222 (May 2021)	U.S.	Top 100
Michael Devereux and Martin Simmler, “Who Will Pay Amount A?” EconPol Policy Brief 36 (July 2021)	Global	Top 100
Vladimir Sarkov and Alexis Jin, “How Would Amount A Affect U.S. Corporate Income Tax Revenue?” Tax Notes Int’l, Nov. 22, 2021, p. 865	U.S.	Top 100
Kartikeya Singh, “What’s It to (the) U.S.? An Impact Analysis of Pillar 1 for U.S. Multinationals,” Tax Notes Int’l, Apr. 12, 2021, p. 151	U.S.	Top 100
Singh, “Amount A: The G-20 Is Calling the Tune, and U.S. Multinationals Will Pay the Piper,” Tax Notes Int’l, Aug. 2, 2021, p. 597	Global	Top 100
Singh, “Relieving Double Taxation of Amount A: Different Ways to Spread the Pain,” Tax Notes Int’l, Feb. 7, 2022, p. 675	Global	Top 100
Martin A. Sullivan, “Which Companies Could Be Caught in the Pillar 1 Net?” Tax Notes Int’l, Oct. 25, 2021, p. 390	Global	Top 100
Sarkov and Jin, “A Tough Call? Comparing Tax Revenues to Be Raised by Developing Countries From the Amount A and the UN Model Treaty Article 12B Regimes,” South Centre Research Paper 156 (June 2022)	Global	Top 100

Using that result, the gap between components C and E enables us to estimate which tax jurisdictions should, on net, gain or lose from amount A: Winners are those whose C - E gap is positive (GIDS/ƩGIDS - GRIP/ƩGRIP > 0), and losers have a negative gap. The total gain or loss in corporate tax revenues for each jurisdiction can be estimated by multiplying the jurisdiction’s corporate rate times its C - E gap times amount A.

Winners and Losers Under Old Amount A

The impact assessment estimated winners and losers from amount A, assuming two in-scope industry groups (automated digital services (ADS) and consumer-facing businesses (CFB)), a 10 percent profitability threshold, and a 20 percent reallocation percentage. The progress report predicted that the overall impact of amount A would be positive but small (0.2 to 0.5 percent of global corporate revenues, or $5 billion to $12 billion), with the gain created by differences in corporate rates across tax-base-receiving and -relieving jurisdictions. Three bar charts summarized the likely winners and losers from amount A. High-, middle-, and low-income jurisdictions were predicted to receive small percentage gains in tax revenues, with the largest percentage (less than 1.5 percent of corporate revenues) going to low-income jurisdictions. Investment hubs would lose significant tax base but because their corporate rates were very low, their maximum revenue loss was also small (3.9 percent of corporate revenues).

The highly aggregated nature of the OECD’s estimates spurred many researchers (myself included) to perform their own estimates of the total and jurisdictional impacts of the amount A proposal; a list of those studies is shown in Table 2. Early estimates, following the OECD impact assessment and the pillar 1 blueprint, assume ADS and CFB are the only industries in scope for amount A. The U.S. government’s April 2021 proposal to replace ADS and CFB with the top 100 MNEs led to some modifications in the 2021 statement and generated new empirical estimates. Some studies have been done at the global level, others from the perspective of a single country (for example, the United States), and still others used hypothetical numbers rather than statistical data.

Two simple examples may suffice to highlight the key results. Winners from amount A were predicted to be tax jurisdictions where foreign MNEs have a major market presence (high GIDS) but declare little profit (low GRIP). Examples include MNEs with large destination-based sales but no foreign affiliate or permanent establishment or with foreign affiliates that earn limited returns (for example, low-risk distributors). In those jurisdictions, the C - E gap is positive and the jurisdiction is, on net, tax-base-receiving under the formula. Predicted losers from amount A were jurisdictions where foreign MNEs have small host-country sales (low GIDS) but earn large profit (high GRIP); examples are tax havens and investment hubs. Those jurisdictions are, on net, tax-base-relieving under the amount A formula and expected to be large losers in terms of forgone tax base.

The Analytics of New Amount A

The amount A progress report outlines the building blocks of the new amount A in the form of seven titles and 10 schedules. Titles 1-4 are relevant for the amount A formula:

• Title 1 (scope) defines in-scope MNEs as those with MNE global revenue > €20 billion and profitability > 10 percent. Financial services and extractives remain out of scope.

• Title 2 (charge to tax) defines the new tax base for a market jurisdiction as the portion of an MNE’s adjusted pretax profit associated with revenues arising in that jurisdiction under amount A’s revenue-sourcing rules, assuming nexus rules are met.

• Title 3 (nexus and revenue sourcing) sets the floor for nexus at MNE revenue ≥ €1 million for countries with GDP ≥ €40 billion, which is reduced to MNE revenue ≥ €250,000 if GDP < €40 billion. Revenue-sourcing rules are provided for different categories of transactions but if reliable indicators are unavailable, macro-level indicators (even as broad as country share of global GDP) can be used to allocate MNE revenues among countries.

• Title 4 (determination and allocation of taxable profit) provides rules for determining and allocating taxable profit to a jurisdiction.

The New Formula for Profit Allocation

As noted, title 4 of the progress report outlines a new profit allocation rule under which 25 percent of an MNE’s global adjusted pretax profit minus 10 percent of its global revenue is allocated to eligible market jurisdictions. Each jurisdiction’s share of amount A is determined first as the share of MNE revenues in that jurisdiction, and second as adjusted by a marketing and distribution safe harbor (MDSH). Separate equations are used to calculate the two steps.

To compare the old and new versions of amount A, the most useful section of the progress report is article 6, which outlines the two formulas used to calculate amount A, and title 7, which provides definitions of key variables.

The first of the two formulas determines the amount of the MNE’s tax base (profit) to be allocated to a jurisdiction, J. The amount J receives in tax base equals the difference between the MNE’s global adjusted pretax profit and global revenues times a 10 percent profitability threshold times a 25 percent reallocation percentage times the ratio of MNE revenues arising in that jurisdiction as a share of the MNE’s global revenues. We can write the first formula as (where Ʃ indicates that the variable is calculated at the global, not jurisdictional, level):

Where:

• Q is the new amount A tax base allocated to J;

• P is an MNE’s pretax adjusted profit in J;

• R is an MNE’s revenues in J;

• 10 percent is the profitability threshold; and

• 25 percent is the reallocation percentage.

Transfer pricing professionals familiar with the old amount A formula will recognize the similarities between that new profit allocation formula and the old formula (equations 1a-c). While the titles and definitions have changed, the basics behind amount A are the same. Equation 2a simply measures J’s share of amount A as a tax-base-receiving jurisdiction as the product of components A through C in equation 1a, where:

• A is the MNE’s global residual in-scope profit (ƩGRIP) = ƩP - ƩR * 10 percent;

• B is the reallocation percentage of 25 percent;

• C is a jurisdiction’s tax-base-receiving allocation key (R/ƩR = GIDS/ƩGIDS);

• A * B = amount A; and

• A * B * C = J’s profit allocation (its tax base gain) under the amount A formula.

A minor difference between equations 1a-c and 2a is that component D, the tax rate to be levied by a jurisdiction on its newfound tax base, is missing from equation 2a. However, the component could easily be included by multiplying 2a by the jurisdictional corporate rate to determine its gain in tax revenues.

One key change between the old and new amount A formulas is the shift from GIDS to revenues. The old formula focused on MNEs’ outward (destination-based) sales activities in the form of exports and in-country sales by MNE affiliates. The new formula no longer has that focus; in fact, the word “sales” has been replaced by the word “revenues” throughout the progress report. Thus, a core driver of the amount A project — that MNEs were making large profits from sales in market jurisdictions without a taxable presence there (for example, through ADS) — has been diluted by shifting from GIDS to revenues.¹⁰ The new amount A formula therefore has moved even further away from the original problem that motivated pillar 1.

The New Safe Harbor

Another key change is that the second half of the formula in equations 1a-c, which identifies who provides tax base relief under the formula, is missing from equation 2a. In other words, components E and F have been removed from the new amount A formula and replaced by the MDSH, which is calculated separately. The new safe harbor “adjusts the allocation of amount A for market jurisdictions that already have existing taxing rights over the Group’s residual profits.” It is meant to prevent double taxation by jurisdictions that already have a taxing right because there is a taxable entity (a subsidiary or PE) in that jurisdiction. Double taxation would occur if a market jurisdiction could tax the MNE’s residual profits twice — first under existing source and residence tax rules and second by taxing an amount A allocation to the jurisdiction.

The safe harbor is poorly named because it is not limited to marketing and distribution activities but rather applies to all MNE profit earned by an MNE’s subsidiaries and PEs in a market jurisdiction. The name comes from the pillar 1 blueprint, which proposed (but left for later work) a safe harbor for routine marketing and distribution activities. When in-scope industries were limited to ADS and CFB, the MDSH name was appropriate; now that amount A includes all industries except finance and extraction, it is no longer accurate. Nor is it clear who benefits from the safe harbor — jurisdictions or MNEs?

To calculate the MDSH, the progress report starts with a measure of the MNE’s adjusted pretax profit in a jurisdiction, referred to as elimination profit (EP). Global elimination profit (ƩEP) and global adjusted pretax profit (ƩP in equation 2a) are both financial accounting measures of pretax profit but they are not identical.¹¹ Also, the method for calculating EP differs from that for GRIP. EP does not have a profitability threshold of 10 percent of revenues deducted whereas GRIP does; in other words, GRIP is a residual measure of profit and EP is not.

Unfortunately, the progress report is confusing about how the two formulas interact to determine a jurisdiction’s net tax base gain under amount A. In fact, the MDSH formula on page 17 of the report appears to be incorrect, so my analysis instead follows the wording on page 16. In the text, a jurisdiction’s portion of amount A (as measured by Q in equation 2a) is reduced by the safe harbor as applicable. The jurisdiction’s net amount A ranges between Q and zero; that is, the maximum allocation to a jurisdiction should be Q and the minimum allocation zero, depending on the safe harbor. The second equation in the new amount A formula can therefore be written as:

Where Ʃ indicates that the variable is calculated at the global, not jurisdictional, level:

• M is J’s net tax base gain, which equals its profit allocation minus the safe harbor;

• Q is J’s profit allocation from amount A, as determined by equation 2a;

• The safe harbor is (EP - PEP) * Y where:

  • EP is the MNE’s elimination profit in J;

  • ƩETR is the MNE’s elimination threshold return = 10 percent of the MNE’s global revenue (ƩR) divided by its global DPC (ƩDPC) — that is, ƩETR = 10% * ƩR/ƩDPC;

  • PEP is a deemed amount of profit based on the portion of EP that results in a return on depreciation and payroll (RODP) that is the higher of ƩETR or 40 percent;

  • RODP is the ratio of EP to depreciation and payroll costs (DPC) — that is, RODP = EP/DPC;

  • EP - PEP is the deemed amount of residual profit; and

  • Y is the offset percentage — that is, the portion of residual profit (EP - PEP) eligible for offset under the safe harbor, which could be 100 percent or some multiple.

In other words, a jurisdiction’s net portion of amount A is its gross gain in tax base (Q in equation 2a) minus the safe harbor as determined in equation 2b. If the safe harbor exceeds Q, the jurisdiction’s net gain would be negative; the formula therefore has a built-in floor of zero. It also appears the OECD did not consider the possibility that the MDSH could be negative, which would raise a jurisdiction’s net gain from amount A to more than Q. I therefore assume that a jurisdiction’s gain under the new amount A varies from a floor of zero to a ceiling of Q.

Some simple examples may help. Using equation 2b, if Q = 10 and MDSH = 4, M = 6. Similarly, if Q = 10 and MDSH = -2, M = 10. Lastly, if Q = 10 and MDSH = 16, M = 0. Thus, M varies between a floor of zero and ceiling of Q depending on the safe harbor.

Table 3. Numerical Example Comparing Old and New Amount A Formulas (in billions of dollars)

	Variable	Formula	1	2	3	4	5	6
Calculation of Amount A
1	Global revenue	ƩR	$200	$200	$200	$200	$200	$200
2	Global in-scope profit	ƩP	$40	$40	$40	$40	$40	$40
3	GRIP	ƩGRIP = ƩP - 0.1 * ƩR	$20	$20	$20	$20	$20	$20
4	Amount A	0.25 * ƩGRIP	$5	$5	$5	$5	$5	$5
Calculation of J’s Portion of Amount A Before the Safe Harbor
5	GIDS in J = revenue in J	R	$100	$100	$100	$100	$100	$100
6	Component C in J	GIDS/ƩGIDS = R/ƩR	0.5	0.5	0.5	0.5	0.5	0.5
7	Gross amount A to J	Q = amount A * R/ƩR	$2.5	$2.5	$2.5	$2.5	$2.5	$2.5
Calculation of the Safe Harbor in J
8	EP in J	EP	$20	$20	$20	$20	$20	$20
9	GRIP in J	GRIP = EP - 0.1 * R	$10	$10	$10	$10	$10	$10
10	Component E in J	GRIP/ƩGRIP	0.5	0.5	0.5	0.5	0.5	0.5
11	DPC in J	DPC	$10	$20	$30	$60	$30	$60
12	Global DPC	ƩDPC	$40	$40	$120	$120	$120	$120
13	J’s share of global DPC	DPC/ƩDPC	0.25	0.5	0.25	0.5	0.25	0.5
14	Return on DPC in J	RODP = EP/DPC	2	1	0.67	0.33	0.67	0.33
15	ETR in J	ETR = 0.1 * R/DPC	1	0.5	0.33	0.17	0.33	0.17
16	ƩETR	ƩETR = 0.1 * ƩR/ƩDPC	0.5	0.5	0.17	0.17	0.17	0.17
17	Max ETR in J	Max ETR = max (ƩETR, 0.4 or 0.6)	0.5	0.5	0.4	0.4	0.6	0.6
18	PEP = portion of EP where RODP = max ETR	PEP = max ETR * DPC	$5	$10	$12	$24	$18	$36
19	MDSH = (EP - PEP) * Y	(EP - PEP) * Y	$15	$10	$8	-$4	$2	-$16
Calculation of Net Gain/Loss in Amount A to J Under New Formula
20	Amount A before MDSH	Q	$2.5	$2.5	$2.5	$2.5	$2.5	$2.5
21	Amount A after MSDH	Q - MDSH	-$12.5	-$7.5	-$5.5	$6.5	$0.5	$18.5
22	Q - MDSH ≥ 0 (floor)	Min (Q - MDSH, 0)	$0	$0	$0	$6.5	$0.5	$18.5
23	Q - MDSH ≤ Q (ceiling)	Max (Q - MDSH, Q)	-$12.5	-$7.5	-$5.5	$2.5	$0.5	$2.5
24	0 ≤ Q - MDSH ≤ Q (floor and ceiling)	Max (Min (Q - MDSH,0), Q)	$0	$0	$0	$2.5	$0.5	$2.5
Calculation of Net Gain/Loss in Amount A to J Under Old Formula
25	Tax base receiving by J	Amount A * component C	$2.5	$2.5	$2.5	$2.5	$2.5	$2.5
26	Tax base relieving by J	Amount A * component E	$2.5	$2.5	$2.5	$2.5	$2.5	$2.5
27	Net amount A	Amount A * (C - E)	$0	$0	$0	$0	$0	$0
Note: The table assumes that: (1) R and GIDS are both revenue measures; (2) ƩP and EP are both in-scope profit measures; (3) max ETR (row 17, columns 5 and 6 have a floor of 0.6 instead of 0.4); and (4) Y is 100% (row 19).

A Numerical Comparison

While equation 2a is a reworded version of equation 1a, equation 2b is new and confusing. The progress report’s formulas might be easier to understand with numerical examples.

Table 3 is designed to illustrate the similarities and differences between the old and new versions of amount A.¹² It is organized into five sections:

• the global amount A (the pie for reallocation);

• jurisdiction J’s gross profit allocation as a tax-base-receiving market jurisdiction;

• J’s safe harbor under the new amount A;

• J’s net profit allocation after tax base relief is provided through the safe harbor; and

• J’s net profit allocation after providing tax base relief through the old amount A formula.

Columns 1-4 represent the four possible combinations of (1) two MNEs differing (one high, one low) in DPC intensity (proxied by ƩETR) and (2) two MNE affiliates differing (one high, one low) in DPC intensity (proxied by ETR). Examples of affiliates with low DPC intensity are those that generate profits by exploiting intangible assets or sheltering multinational group profits; those with high DPC intensity have profits more closely tied to labor and property, plant, and equipment costs (for example, manufacturing).

Columns 1-2 assume the MNE has a low DPC intensity (ƩETR > 0.4), and columns 3-6 assume the MNE has a high DPC intensity (ƩETR < 0.4). Columns 1, 3, and 5 assume the MNE affiliate has a low DPC/ƩDPC ratio (0.25), and columns 2, 4, and 6 assume the affiliate’s DPC/ƩDPC ratio is high (0.5).

Column 1 therefore illustrates the low-low DPC case in which the MNE has low DPC intensity and its affiliate’s DPC intensity is lower than the group average (for example, the affiliate is a regional headquarters or marketing hub in a pharmaceutical MNE). Columns 4 and 6, on the other hand, illustrate the high-high DPC case in which the MNE has high DPC intensity and the affiliate’s DPC intensity is above the group average (for example, the affiliate is a captive manufacturing plant in an automotive MNE).

To illustrate the impact of the arbitrary 40 percent floor in the calculation of the maximum ETR in equation 2b, columns 5 and 6 assume a 60 percent floor instead of a 40 percent floor. Thus, columns 3 and 5 are identical, as are columns 4 and 6, except for their different floors.

Given the confusing terminology on pages 16-17 of the progress report, I also explore the effects of different floors and ceilings on the safe harbor and the net amount A received by jurisdiction J. Rows 22, 23, and 24 impose different constraints on net amount A. Rows 22 and 24 assume that the net amount A allocation (Q - MDSH) cannot be negative (the floor is zero). Rows 23 and 24 assume that the net amount A allocation cannot be higher than the gross allocation (the maximum net amount A is Q, which implies the MDSH ≥ 0).

The table is constructed such that amount A is $5 billion (row 4) and J’s gross portion is $2.5 billion (rows 7 and 20) in all six cases. Also, I have set the gap between components C and E in the old formula equal to zero — so the net amount A under the old formula is zero (row 27) — to focus attention on the effects of the new amount A.

Some of the key findings include:

• While the net amount A received by each jurisdiction under old amount A is zero (row 27), the amount varies widely under new amount A (rows 21-24), depending on the MDSH and the floor and ceiling restrictions.

• Without a cap on amount A or floor on the safe harbor, the net amount A under the new formula varies from -$12.5 billion to $18.5 billion (row 21).

• Comparing columns 3 and 4 with columns 5 and 6 shows that raising the fixed percentage used in the maximum ETR calculation from 40 percent to 60 percent (row 17) raises PEP (row 18) and lowers the MDSH (row 19).

• Row 19 also illustrates that the MDSH is typically positive but can turn negative for high-high DPC entities (columns 4 and 6). The safe harbor is negative (-$4 billion) in column 4 and even more negative (-$16 billion) in column 6 when the floor for calculating the maximum ETR is set at 60 percent instead of 40 percent (row 17).

• Because the MDSH can be negative, a ceiling is necessary to prevent the net amount A to a jurisdiction from being larger than its gross amount A. Row 21 shows the net amount A without a ceiling or floor; row 23 has a ceiling but no floor. The ceiling caps the net amount A at Q ($2.5 billion); without the ceiling, the net amount A can be seven times higher (column 6, row 21).

• The MDSH can also be so high that the net amount A turns negative (-$12.5 billion in columns 1-3, rows 21 and 23). Setting the floor on net amount A at zero removes that effect (columns 1-3, rows 22 and 24).

• Capping net amount A with both a floor of zero and a ceiling of Q significantly narrows the range (row 24), which now varies from a low of zero (columns 1-3) to $0.5 billion (column 5) to a high of $2.5 billion (columns 4 and 6).

In sum, the widely varying results in Table 3 reflect the complications introduced by the safe harbor, which suggests that administering the new formula will be much more difficult and provide many opportunities for gaming amount A. The variation also points to how far the new formula has moved from proxying a redistribution of MNEs’ residual profits to market jurisdictions based on their GIDS and GRIP shares.

A Comparative Analysis

The tax-base-relieving process can significantly affect the winners and losers from amount A, as well as its administrative and dispute settlement costs, because the process determines who pays for amount A. This section explores differences between the old and new formulas for providing tax base relief and how that relief would be provided.

The Formulas

Under the old amount A formula, tax base relief was provided through component E, a jurisdiction’s GRIP share (GRIP/ƩGRIP). Component E was designed to ensure that market jurisdictions that already had taxing rights over MNE profits would receive only a profit allocation exceeding their existing taxing rights.

Conceptually, tax base relief under equations 1a-c can be viewed as being provided in two steps. First, each jurisdiction subtracts its existing MNE tax base from what it expects to receive in new tax base under the amount A formula. Second, jurisdictions with negative C - E gaps must also give up some of their tax bases to other market jurisdictions; their overall tax bases decline if they participate in the amount A reallocation game.

The subtraction in the first step prevents a jurisdiction from taxing its own tax base twice — once as a residence or source jurisdiction and second as a market jurisdiction — thus providing a safe harbor (from double taxation) to MNEs in the market jurisdiction. Market jurisdictions where MNEs have subsidiaries or PEs must subtract the full amount of their tax bases in those jurisdictions from the jurisdictions’ allocation under amount A. For example, if an MNE has GIDS of €100 million and GRIP of €60 million, the jurisdiction receives €40 million in new tax base under amount A, not €100 million.

That is so even if a jurisdiction chooses not to tax an MNE’s tax base in its jurisdiction and thus receives no tax revenue on that base. That could be the case if the jurisdiction is a tax haven, exempts the tax base, or provides tax preferences that offset the owed taxes. A market jurisdiction that already has taxing rights over an MNE’s base — regardless of whether it exercises those taxing rights — cannot levy an additional amount A tax on that base. Its share of the MNE’s existing corporate base (component E in the formula) must be subtracted from its share of amount A (component C) to determine its net amount A.

Also implicit in the old amount A formula is that jurisdictions with a negative C - E gap — that is, GIDS/ƩGIDS < GRIP/ƩGRIP — must give up their own tax bases equal to that gap. For example, a jurisdiction with an expected gain in tax base (A * B * C) of €100 million and an expected loss in tax base (A * B * E) of €150 million not only receives no amount A but also must give up €50 million of its own tax base to other jurisdictions. The jurisdiction providing tax base relief under amount A in effect transfers to other jurisdictions its taxing rights to that portion of the MNE tax base in perpetuity without necessarily knowing which jurisdictions are receiving that base or what will be levied on the base.

That point has not been well understood in analyses of amount A. Each tax-base-relieving jurisdiction must in effect commit — and a new multilateral convention would be created to enforce that commitment — to give up its taxing rights over some portion of its MNE tax base and no longer levy taxes on that forgone base. It is hard to imagine that a government would willingly give up in perpetuity its sovereign right to tax MNE profits under existing residence and source principles, even if does not currently tax that base. Not only would a government be limiting its own choices, but it would also be limiting the choices, and tying the hands, of future governments — unless it were to terminate its participation in the multilateral convention. National sovereignty is an important value for nation states.¹³ That a country would commit to giving up its sovereign taxing rights for the foreseeable future (as long as it remains a party to the convention) is not a well-understood cost of the amount A proposal.

How is tax base relief to be provided under the new amount A formula? First, instead of subtracting component E (each jurisdiction’s GRIP/ƩGRIP share) from component C (each jurisdiction’s GIDS/ƩGIDS share) as in equations 1a-c, the new formula calculates a jurisdiction’s share of global revenues and subtracts a safe harbor based on the MNE’s labor and capital costs in that jurisdiction, as outlined in equation 2b. Clearly, there is no linear relationship between a jurisdiction’s residual profit and its share of MNE DPC or the return to those costs, as illustrated in Table 3. For 21st-century MNEs, residual profit is much more closely tied to their intangible assets, not to their property, plant, and equipment or labor costs. Thus, the effects of changing from the old formula to the new will be different, as well as difficult to predict.

Second, the new formula for calculating the net amount A to a jurisdiction has a floor of zero and a ceiling of Q, the original profit allocation, which means the process for providing tax base relief is now completely divorced from equations 2a and 2b. No tax base relief is provided under the new amount A formula; everything is shifted to a separate tax-base-relieving mechanism. Identifying who wins is separate from who pays, and it should encourage more opportunistic behavior.

The Tax-Base-Relieving Mechanisms

The pillar 1 blueprint outlines a four-step process for identifying which jurisdictions provide tax base relief, which is inconsistent with the tax-base-relieving mechanism outlined in equations 1a-c. It proposes four steps for determining who pays: three metrics (activities, profitability, and market connection priority) and a backstop, pro rata allocation that strips jurisdictions with highly profitable MNEs of their tax bases according to their profitability.¹⁴

The pillar 1 progress report briefly outlines the tax-base-relieving mechanism in title 5 (elimination of double taxation for amount A). Title 5 determines which jurisdictions must provide tax base relief and how much — that is, who pays for amount A. It lists potential tax-base-relieving jurisdictions (or specified jurisdictions) based on the smallest number of jurisdictions for which the MNE’s global EP is 95 percent of global EP, plus those jurisdictions with EP ≥ €50 million. The group is then separated into four tiers based its profitability (RODP) in that jurisdiction relative to the MNE’s global profitability. Using a waterfall approach, tax relief is provided by jurisdictions with highest RODP and cascades through the tiers until full relief has been provided. The four tiers are: RODP > 1,500 percent, RODP > 150 percent, RODP > 40 percent, and RODP > 10 percent. The method for identifying which MNEs are entitled to tax relief is separately determined; both tax exemption and credit are mentioned.¹⁵

The progress report’s proposal to allocate tax base relief using a waterfall approach based on a jurisdiction’s RODP makes even less sense than the four-step formula proposed in the pillar 1 blueprint. Both the old and new proposed tax-base-relieving methods are clearly inferior to the original process for who pays outlined in equations 1a-c. The differences across MNEs and entities within a multinational group in terms of their factor intensities (which include not only labor and capital but also intangible assets) and the returns to those factors, as illustrated in Table 3, suggest that the incidence of who pays amount A will be both highly uncertain and uneven in any formula based on DPC.

The progress report is mostly silent on that matter, other than a short discussion of the proposed waterfall method in title 5. What is clear here is that by separating the question of tax base relief – that is, who pays — from tax base receipt — that is, who wins — jurisdictions are encouraged to focus on what they can expect to get from amount A and pay less attention to what that will cost them later in forgone tax base. Separating who wins from who loses encourages governments to view amount A as a free lunch.

Pillar 1 Tax Games

Essentially, amount A carves out a portion of MNE global pretax profit and applies revenue-based global formulary apportionment as an allocation key to redistribute that tax base.¹⁶ Formulary apportionment would be layered on top of existing international corporate income tax rules, creating more complexity and uncertainty and increasing administrative and governance costs.

An important caveat regarding the predicted winners and losers from amount A is that the OECD’s pillar 1 reports and estimates all fail to account for national sovereignty and self-interest behaviors, as well as the difficulties of administering amount A in today’s global economy. MNEs and governments are assumed to abide by the rules of the game — that is, they ignore their own self-interest and national sovereignty and do not behave strategically with respect to the formula. Most public finance scholars view national sovereignty and administrative feasibility as important constraints on designing tax policy.¹⁷ For nation states, participating in amount A means “the renunciation of self-interest — in this context, some measure of sovereign fiscal authority” and “the risk that others are not as (immediately) like-minded despite being members of a coalition of ostensibly shared committed interests.”¹⁸

There are many opportunities for governments and MNEs to game the original amount A formula.¹⁹ As equation 1a demonstrates, to raise its net share of amount A, a jurisdiction should raise component C (its GIDS/ƩGIDS share) and lower component E (its GRIP/ƩGRIP share), both of which would affect its net tax base, and raise component D (the corporate rate on the tax base gained) and/or lower component F (the corporate rate on the tax base lost), both of which would affect its net revenues.²⁰

Large jurisdictions with many MNEs in scope for amount A have the ability and desire to affect components A and B; small jurisdictions are less able to affect the rules of the game. For example, pharmaceuticals are in scope for amount A but financial services are not. Both industries earn above-normal returns and are highly regulated, yet one is in and the other is out, which affects the distribution of winners and losers. As a second example, the U.S. government’s top 100 MNEs proposal brought large business-to-business companies in scope for amount A, reducing its impact on ADS and CFB but generating complexities for determining GIDS at the transactional level.²¹

The old amount A transfers MNE pretax residual profits into market (tax-base-receiving) and out of residence and source (tax-base-relieving) jurisdictions. In equation 1a, a jurisdiction focuses on its net share, which should be positive and as large as possible. In equation 1b, a jurisdiction maximizes its tax revenue gain and then minimizes the tax relief it provides to other jurisdictions. Thus, pillar 1 tax games can be played for each of the components, both in size and timing (especially by large jurisdictions).

The ability to game the components of the formulas appears to be even stronger with new amount A and its arbitrary fixed percentages and mandatory floors and ceilings. The safe harbor formula and the proposed waterfall method for providing tax base relief — both of which are tied to RODP — also create more complexity, which encourages opacity and opportunism.

Another way to game amount A is to separate and sequence the formula’s components, as suggested by old equation 1b and as implemented by the two equations for new amount A. An opportunistic jurisdiction would focus first on enlarging its gross tax base gain from amount A and then on minimizing the tax base relief it provides to other jurisdictions under the formula. The two-step process of gains first, losses later could be exploited in several ways. For example, jurisdictions that expect to receive amount A allocations under the first step could be early signatories of the proposed multilateral convention to implement amount A, while those that expect to lose tax base are likely to delay signing.

Tax Base Relief Under Territorial Systems

Regardless of whether the tax-base-relieving mechanism is that in chapter 7 of the pillar 1 blueprint or implied by the amount A formula in the progress report, each faces the same problem: what to do about residence jurisdictions that no longer tax their MNEs’ foreign-source income. It is here that amount A, which is designed to be layered on top of existing international tax principles and rules, creates more problems.

Most countries haves territorial tax systems where the foreign source income of their MNEs is not taxed at home. Residence jurisdictions levy taxes on so-called offshore passive income under controlled foreign corporation rules, and countries that do not have CFC rules were encouraged to adopt them during the first round of base erosion and profit-shifting reforms. A few countries also levy a minimum corporate tax on foreign-source income (for example, the U.S. global intangible low-taxed income regime and base erosion and antiabuse tax), but most of it is exempt from home-country taxation.

Under the amount A mechanism as discussed in the impact assessment, pillar 1 statement, and progress report, relief is to be provided either by exemption or credit. Residence countries exempt their MNEs’ foreign-source income from home-country tax, so they already provide tax base relief to source (host) countries. Even though a residence jurisdiction currently exempts foreign-source income from tax, it is hard to imagine that a sovereign jurisdiction would be willing to permanently forgo its right to levy a residence-based tax on the foreign-source income its MNEs earn on their offshore activities.

Moreover, because residence countries cannot give the same exemption twice, the jurisdictions that must provide tax base relief through the exemption method under amount A must be source jurisdictions — that is, those that host local affiliates of foreign MNEs. That is another point not well understood in the debate over amount A. It is hard to imagine that a sovereign government would be willing to forgo into perpetuity the right to tax the profits earned by local affiliates of foreign MNEs in the host jurisdiction. Under amount A, that tax base would be transferred out of the source country and taxed by one or more market jurisdictions at rates set by that jurisdiction.

Suppose the host government does not provide tax base relief under amount A. That means the local subsidiaries of foreign-owned MNEs are taxed twice, once by the source country under existing source-tax rules and again by the market country under amount A. Double taxation of foreign MNEs would have the unintended (or maybe intended?) consequence of implicitly favoring (by providing an implicit subsidy to) the host country’s own companies — and those companies are likely to encourage that discrimination. Thus, amount A risks becoming a new form of protectionism wielded by source countries, one that is likely to encourage tit-for-tat retaliation by other countries.²²

Conclusions

In this article, the original or “old” amount A concept has been used as a reference point to assess the viability of the “new” amount A as proposed in the July 2022 OECD progress report. While both methods are flawed, the new amount A is more flawed that its predecessor. The new amount A increases the risk of double taxation and leaves room for tax administrations and in-scope MNEs to game the calculation. My conclusions are summarized below.

New amount A is less tied to the original purpose of pillar 1. Pillar 1 was originally meant to compensate market jurisdictions for their inability to levy corporate tax on MNE profits on destination-based sales of ADS.²³ The amount A formula in the impact assessment and pillar 1 blueprint compared a jurisdiction’s shares of GIDS and GRIP to shift MNE profit from residence and source jurisdictions to market jurisdictions. In the new amount A, the portion of MNE residual profit that is taxable in a jurisdiction is determined by an allocation formula based on a jurisdiction’s share of MNE revenues (not destination-based sales), which is to be reduced by a safe harbor tied to RODP costs (not residual profits).

New amount A is harder to understand and estimate. New amount A is harder to model and understand in terms of its simple analytics. My comparison confirms the greater complexity caused by the safe harbor formula. Not only is the equation hard to understand, using DPC as allocation keys for the safe harbor and in the proposed tax-base-relieving mechanism ignores the core role played by intangible assets in value creation within MNEs. RODP is neither an appropriate nor reliable measure of MNE profitability and will likely create more complexities than old amount A.

New amount A offers more opportunities for pillar 1 tax games. Complexity almost always provides opportunities for actors to find loopholes and to arbitrage or manipulate the results. Countries have both motive and opportunity and can rationalize their engagement in pillar 1 tax games using national sovereignty and protectionist arguments. That was true for old amount A, and new amount A is even less straightforward: It separates who wins from who pays, making it much more difficult to determine the net impact of amount A and encouraging governments to game the formula.

New amount A is more likely to raise the tax burden on MNEs and discourage investment. The predicted net impact of old amount A on government revenues was small.²⁴ The tax base gained by market jurisdictions would be exactly offset by the tax base lost by residence and source jurisdictions – in other words, amount A would be a zero-sum game with winners offsetting losers.

I believe that conclusion is far too optimistic and even less likely under new amount A. National sovereignty is a zealously guarded entitlement of nation states. The new amount A tax-base-relieving mechanism is unlikely to provide full tax base relief without an omniscient dictator to enforce the rules. The costs of moving from two principles (residence and source) to include a third principle (destination) for taxing MNE profits would likely be borne not by residence and source jurisdictions but by MNEs in the form of double taxation. Who wins becomes a positive-sum game for jurisdictions if they can shift who pays to MNEs.

Further, both versions of amount A distort market signals and MNE decision-making — at a minimum between companies and industries that are in scope for amount A and those that escape it. Differences among in-scope MNEs also affect their relative tax risks, with some MNEs finding their risks skyrocketing, depending on the aggressiveness of tax-base-receiving and tax-base-relieving jurisdictions. As a result, high administrative and governance costs would be required to implement amount A, deter gaming of the formula by both tax jurisdictions and MNEs, and settle interjurisdictional disputes. Greater economic distortions and double taxation, especially during a global recession, could also significantly depress foreign direct investment, much more so than as predicted in the 2020 OECD impact assessment.

Developing countries are less likely to benefit from new amount A. Given the added complexities of new amount A, it is now harder for low-income jurisdictions to assess the potential revenue effects. If developing countries sign onto the new amount A and its proposed multilateral agreement, they will be giving up tax sovereignty and taking on new administrative burdens for what appears to be very little reward. That is especially true for low-income jurisdictions that would be required under amount A to provide tax base relief by giving up their right to tax profits earned by foreign affiliates and permanent establishments within their borders.

One observer has spoken directly on that issue,²⁵ and I echo his two core questions: If developed countries want to support the economic well-being of less-developed nations, are there better means to do so other than the large-scale repurposing of corporate income taxation and international tax jurisdiction proposed by amount A? Would those ways be better targeted to achieve their objectives and be more within national sovereignty and administrative feasibility for developing countries?

I believe the answer to both questions is yes. It is time for a fundamental rethink of the amount A proposal.