South Centre Comments on Progress Report on Amount A of Pillar One
The South Centre is the intergovernmental organization of developing countries that helps developing countries to combine their efforts and expertise to promote their common interests in the international arena. The South Centre has 54 Member States coming from the three developing country regions of Africa, Asia, and Latin America and the Caribbean. It was established by an Intergovernmental Agreement which came into force on 31 July 1995. Its headquarters are in Geneva, Switzerland.
The South Centre in 2016 launched the South Centre Tax Initiative (SCTI). This is the organisation’s flagship program for promoting South-South cooperation among developing countries in international tax matters.
The South Centre offers its comments to the OECD Inclusive Framework’s Task Force on Digital Economy (TFDE) on the Progress Report on Amount A of Pillar One.
In June 2022, the Coalition for Dialogue on Africa (CODA), a Special Initiative of the African Union, and the South Centre, jointly released country-level revenue estimates from Amount A compared with Article 12B of the UN Model Tax Convention, for the 84 combined Member States of the African Union and the South Centre.
The revenue estimates showed that for the majority of the 84 Member States, all of whom are developing countries, the maximum revenue would be gained by the UN Model Tax Convention’s Article 12B, applied on a gross basis with a 4% rate, to a broad scope of Automated Digital Services (ADS). By contrast, Amount A had minimal benefits for most developing countries, providing two to five times less revenue compared to Article 12B.
This is the world’s first such set of country-level estimates, and it is worrying that the OECD, has not publicly released data on the potential impact of Amount A. Countries should not sign a ‘blank cheque’ by signing up to an agreement when there is no information on what revenues are likely to be generated for them.
To address this situation, CODA and the South Centre have produced the above-mentioned estimates and provided a set of recommendations to developing countries on the taxation of the digitalized economy.
The Progress Report on Amount A, the latest version of the OECD’s proposed solution for taxation of the digitalized economy, makes it clear that the revenues expected for developing countries will dwindle even further than estimated by CODA and the South Centre.
With each successive update of the rules, the proposed solution is becoming increasingly less appealing to the developing countries. The OECD must, at a minimum, release revenue estimates for the 141 jurisdictions of the Inclusive Framework such that each can take an informed decision in the national interest. As an organization that sets ‘transparency’ standards, OECD must itself be transparent and provide countries with the essential information needed for making what may become a historic decision for the international taxation regime.
III. Specific Comments
- National Tax Measures
The Overview states that the Multilateral Convention (MLC) through which Amount A will be implemented will “include a commitment not to enact Digital Services Taxes (DSTs) or relevant similar measures, provided they impose taxation based on market-based criteria, are ring-fenced to foreign and foreign-owned businesses, and are placed outside the income tax system (and therefore outside the scope of tax treaty obligations).”
- Commitment Not to Enact
This is a restriction on the power of legislatures to make future laws on taxation, which is an essential component of State sovereignty. The prohibition will also likely run into constitutional limitations in several developing (and developed) countries and further reduce the acceptability of the MLC.
Recommendation: The commitment not to enact DSTs or relevant similar measures must be political rather than legal and kept out of the MLC.
It is also proposed that the MLC include a “definitive list” of such national measures. Such a list is likely to become a de facto ‘blacklist’ and be used to force countries to eschew the measures contained within. Accordingly, great care must be taken as to which measures are placed on the list. Given that Amount A is a foray into hitherto unexplored territory, it is unclear how the implementation experience will be. To expect developing countries to give up their policy space and pin their hopes on an experimental and untested approach, foregoing the possibility of all other alternatives, is an excessive demand that is unlikely to be met. This is exacerbated by the ongoing global economic crisis which has increased the need for tax revenues.
- Definitive List of Measures
To bring developing countries on board it is advisable that an open approach be taken which avoids straitjacketing their sovereign powers of taxation through an expansive list of prohibited measures. Such a limitation on taxing sovereign rights, advocated by some tech companies, may end up backfiring upon its proponents and result in the rejection of the deal, which will mean the continuation of uncoordinated national measures such as DSTs.
Recommendation: The “definitive list” of DSTs and relevant similar measures must be kept as limited as possible.
The Overview proposes three conditions of taxation based on 1) market-based criteria, 2) ring-fenced to foreign and foreign-owned businesses, and 3) placed outside the income tax system and therefore outside the scope of tax treaty obligations.
- Definition of Prohibited Measures
Recommendation 1: Any measure considered for inclusion should meet all the three conditions unambiguously and comprehensively.
Recommendation 2: There must be a system of frequent and periodic review of the list. If it is found that a measure no longer qualifies or its inclusion is unjustified in principle, it must be easily removable. The MLC must contain a simple decision-making process through which Parties can remove measures from the list. Conversely, a stricter decision-making process must be involved for adding measures to the list.
It must be clarified what will happen in case a country goes ahead and implements national measures, despite being a Party to the MLC. This information is required upfront so countries can make an informed decision based on the opportunities and threats from joining the MLC. Further, it should be clarified that consequences should be contained within the MLC and should not emanate from other treaty or statutes whether international or domestic.
- Consequences of violation
The Overview mentions the relation of withholding taxes (WHTs) to the elimination of double taxation and double counting.
- Withholding Taxes
This reflects the ongoing trend of de facto reductions to Amount A being “slipped in” to detailed rules in contradiction with the Inclusive Framework’s political Statements. Examples abound throughout the Model Rules, such as the introduction of the prior period test, the proposal to apply averaging to revenues in addition to profitability, pre-implementation loss carry forward, and so on. Similarly, there was no mention whatsoever of withholding taxes in the July and October 2021 IF Statements.
However, the issue of withholding taxes is qualitatively different, and poses a far more serious challenge than the other mentioned additions. It would mean that a country would have to adjust 100% of a withholding tax for the minuscule quantum of Amount A, which is 25% of residual profit. This has grave implications, and as categorically stated by the G-24, “will lead to erosion of existing taxing rights and will make Pillar One unattractive and meaningless for the developing countries.”
Beyond the practical implications, the inclusion of withholding taxes in the Amount A allocation or elimination mechanisms is not justified in principle. The assertion that it will lead to double counting is logically flawed.
Firstly, the argument of double counting presumes double taxation on an existing taxing right, but Amount A is a new taxing right. Withholding taxes are levied as per existing bilateral tax treaties whereas Amount A is an overlay over the existing rules. The two are mutually exclusive.
Secondly, double counting presumes that withholding taxes will exclusively affect the quantum of Amount A, which is a minuscule portion of residual profit. But withholding taxes apply to the gross payment on a transaction basis and do not make artificial distinctions between routine and residual.
Thirdly, it must be reiterated that the distinction between routine and residual is arbitrary. The definition of residual profit as profit in excess of 10% of revenue was politically negotiated and agreed upon. Developing country groupings such as the G-24 and ATAF had demanded that a portion of total profits be reallocated instead.
Lastly, withholding taxes are not a tax in themselves; they are only a means of collecting taxes. Double counting assumes two separate taxes are being levied on the same income; it is thus irrational to compare Amount A, which is a tax, with a withholding tax, which is a collection mechanism.
Recommendation: Withholding taxes must be completely removed from the Amount A allocation and elimination mechanisms.
- MLC Design
The MLC must contain an exit clause through which countries can leave the agreement if they see it is not working for them. The withdrawal period must be kept to a minimum so they can leave as soon as they need. It will also make the MLC more acceptable for countries considering to join it, if they know the agreement does not lock them in and there is an expeditious way out procedure.
- Exit Clause
Recommendation: The MLC must contain an explicit exit clause, with a minimal withdrawal period, preferably not more than sixty days after the notice has been submitted to the Depositary.
It is of vital importance that the Amount A MLC be easily amendable. What must be avoided is an MLC which locks Parties into a rigid framework unable to be adapted to changing circumstances. This will only increase the chances of it becoming rapidly irrelevant and being jettisoned, heralding a return to uncoordinated national measures.
The amendment clause should be such that each and every aspect of Amount A, such as the scope, quantum, tax base determinations, etc. could be reviewed if enough countries decide so. The decision-making process must be kept simple, transparent and democratic. Simple majority voting by the Conference of Parties must suffice for most, if not all, amendments. This will ensure the Amount A MLC can continuously adapt to changing circumstances and become a durable instrument.
Recommendation: The MLC must contain an amendment clause through which most, if not all, changes can be carried out through simple majority voting by the Conference of Parties.
- Averaging Mechanism
The October 2021 Statement mentioned an averaging mechanism to be applied to determine if the Group has profitability exceeding 10%. As mentioned in prior South Centre comments, the averaging mechanism must function only as an entry-level test, meaning that once the Group is in-scope, the averaging mechanism need not apply any more. Trying to make it permanent, i.e applying it each time the Group is out of scope, goes beyond the intent of the October Statement and is purely constructed to keep the MNEs out of Amount A. It must be reiterated that the scope was restricted only to the largest and most profitable MNEs, for whom such fine-tuning is not necessary.
- Permanent or Entry Level Test
Recommendation: The averaging mechanism must function only as an entry-test, meaning once the Group is in-scope, it will not apply ever again.
It is welcome that Article 1(2)(b) provides a revised version of the averaging mechanism, responding to the aforementioned concern previously raised by the South Centre. As a compromise, this version can be considered. However, the requirement of being out of scope for two consecutive years must be increased to a longer duration. Further, the prior period test, which was ‘slipped in’ despite not being mentioned in the October 2021 Statement, must be eliminated.
- Revised Averaging Mechanism
Recommendation 1: A Group must be out of scope for at least five consecutive years to be able to benefit from the average test.
Recommendation 2: The duration of the average test must be for three years instead of five as is presently proposed.
Recommendation 3: The prior period test must be removed.
The rules provide for 3-year loss recognition of pre implementation loss and 10 year carry forward period for both pre and post implementation losses.
- Loss Carry Forward
As mentioned in previous comments, pre-implementation losses are unjust in principle. The MNE should be allowed to share losses only if it is sharing profits.
A ten-year loss carry forward period is far too long. This implies reduced revenues for the developing countries. It is also difficult from an administrative standpoint.
Recommendation 1: Pre-implementation loss carry forward should be disallowed.
Recommendation 2: The post implementation losses should be carried forward for a maximum of five years, preferably even lesser.
- Regulated Financial Services
As mentioned in prior South Centre comments, reinsurance and asset management both derive assets primarily from cross-border activities and excluding them would deny market jurisdictions their fair share of taxes.
- Reinsurance and Asset Management
Recommendation 1: Reinsurance must be in scope, and the revenue must be sourced to the jurisdiction where the underlying assets which are being insured are located, and from where the payment arises.
Recommendation 2: Asset management must be in-scope and the revenue must be sourced to where the investor is, because they are the source of the asset managers’ revenues.
It is welcome that the revised rules have addressed the concerns previously raised by the South Centre about the low deposits : liabilities ratio requirement and the special situation created by tax haven jurisdictions who do not have central banks, like Jersey. The revised rules have created a bifurcation with different ratio requirements for jurisdictions with and without central banks. While the differentiated approach is welcome, it is unfortunate to see that for jurisdictions with central banks the deposits : liabilities ratio has been reduced even further, to 10% while for the latter it remains at the earlier level of 20%. This is too low and the risk remains that jurisdictions especially without central banks, can become a “sink” through which financial institutions can take advantage of the exclusion rules and escape the scope of Amount A. The rise of “fintech” exacerbates this risk.
- Depository Institutions
Recommendation: The deposits: liabilities ratio must be higher than 20%, and for jurisdictions without central banks there must be clarity on how the rules will be enforced.
The use of the ‘predominance’ test means that if an MNE qualifies as an Extractives or Regulated Financial Services (RFS) Group, it will be completely out of scope. The predominance test says at least 75% of its revenues must come from the excluded activities.
- Predominance Test
The key problem identified by the South Centre remains in the new rules: even if a Group derives the bulk of its revenues from in-scope activities, the application of the exclusion rules will mean it will be out of scope. It is reiterated that as a matter of principle, once the Group is in-scope, only the revenues from excluded activities must be removed from the Amount A allocation so they can be taxed by the source jurisdiction. The Group must continue to remain in-scope.
As a compromise, if the revised approach is to be followed, then it still means the exclusion of significant in-scope revenues from taxation. The following example illustrates this.
Assume a Group with revenues of EUR 20 billion and which derives 75% of its revenues from excluded activities such as extractives or RFS. It would be a qualifying Group and be completely excluded from Amount A. However, the remaining 25% of the Group’s revenues are from in-scope activities. Since 25% of EUR 20 billion is EUR 5 billion, it would mean a significant amount of revenue is excluded from taxation.
In other words, the design of the predominance test will mean that a minimum of EUR 5 billion of in-scope revenues can be excluded from Amount A. This number is too large. The predominance test must be higher and can be set at 90%. That would reduce the in-scope revenues that are excluded from Amount A to a minimum of EUR 1 billion.
Recommendation: The predominance test must be increased to 90% for both extractives and regulated financial services.
Schedule D provides for transferred losses at the segmental level. This will require the creation of new accounting standards, as presently financial accounts do not provide granular detail about segment level losses. This will also be difficult for MNEs to implement and for tax administrations to audit. It would also go against the spirit of creating rules that are easy to simple and easy to administer.
- Covered Segment Losses
Recommendation: Segmental transferred losses should be disallowed.
- Revenue Sourcing
Article 4 provides for sourcing of “revenues derived from the licensing, sale or other alienation of user data”. This may not account for situations where even the supply, rather than outright sale, of user data generates revenues. For instance, para 6 of Article 12B of the UN Model Tax Convention uses the phrase “supply of user data” when defining Automated Digital Services.
- Revenue from Sale of User Data
Recommendation: The term ‘supply’ can be added to sourcing user data, so the definition would broadly read: “Revenues derived from the licensing, supply, sale or other alienation of user data.”
The definition, provided on page 82, is: ““Intangible Property” means property which is not in tangible form and which is capable of being owned or controlled for use in commercial activities but does not include Real Property, financial assets, Digital Content, User data or the right to use computer programs. It includes copyrights, trademarks, tradenames, logos, designs, patents, know-how and trade secrets.” (emphasis added)
- Definition of Intangible Property
As mentioned earlier by the South Centre, this aspect of the Model Rules contradicts international law, specifically Article 2 of the Berne Convention for the Protection of Literary and Artistic Works (1971) and Article 4 of the WIPO Copyright Treaty (1996). It also has negative implications on existing tax treaties which provide for the taxation of payments for computer software as royalties under Article 12.
Recommendation: The definitions of copyright, copyrighted work and intangible property should include the right to use computer programs.
The various allocation keys in the sourcing rules use GDP. However, nominal GDP (US$) by itself is an insufficient measure as it does not account for the differences in purchasing power between countries.
- Use of GDP PPP
Recommendation: For more accurate revenue sourcing, GDP (PPP) must be used, so countries are allocated the rightful share of revenues through the allocation keys.
Article 5 on Determination of the Adjusted Profit Before Tax of a Group proposes the “treatment of profit attributable to non-controlling interests” as one of the book-to-tax adjustments.
- Non-Controlling Interests
There is no rationale behind this specific adjustment. Profits are ultimately generated by the Group and should remain within the ambit of Amount A.
Recommendation: The adjustment for profits attributable to non-controlling interests must be removed.
- Marketing and Distribution Safe Harbour (MDSH)
The MDSH is designed as a capping mechanism to artificially limit routine profit allocated to countries under the existing transfer pricing mechanism, rather than a safe harbour. Further, the rules indicate that some countries want it to go beyond marketing and distribution and include all activities.
This would mean the MDSH would neither involve marketing and distribution nor be a safe harbour, meaning it would be transformed into something else entirely.
This feeds into the pattern of the detailed rules containing elements that go beyond the IF’s Statements. Both the July and October Statements clearly mentioned that the scope of the MDSH was only marketing and distribution activities.
Recommendation: The scope of the MDSH should be restricted to marketing and distribution activities.
The proposal to use Return on Depreciation and Payroll (RODP) for the MDSH is inappropriate, as often marketing and distribution entities have low assets and payroll costs. The result of this will be a high RODP, which will mean residual profits may end up being located in developing countries, implying reduced Amount A allocation. This will defeat the very purpose of Amount A, which was to result in increased revenue collection for the developing countries.
Recommendation: As stated by the G-24, a composite metric based on a return of 30% of global profit margin on sale (revenue as per revenue sourcing rules) and 40% RoDP on a jurisdictional basis may be considered to be an alternative metric.
The MDSH is in essence a deduction from the Amount A allocation to a jurisdiction. The deduction applies only where the Elimination Profit of the Group in that jurisdiction is greater than a threshold to be defined. Thus, the higher the threshold is, the better it would be for developing countries.
- De minimis threshold for MDSH and elimination of double taxation
A similar threshold is proposed for the elimination of double taxation.
Recommendation: The de minimis should not be an absolute value, as that would mean it may soon become inadequate, but rather should be applied to profit in excess of the respective thresholds applied for the elimination of double taxation and the MDSH.
The rules propose to allow a reduction from the tax base for fines and penalties below EUR 50,000. It is unclear why illegal payments such as bribes are disallowed but these are allowed. For example some major tech giants have been slapped with fines in the past for anti competitive practices and under the proposal these could be reduced from the tax base if they are below EUR 50,000. The distinction between “illegal” payments and “fines and penalties” is not clear, because by definition a fine or penalty is levied when an illegal act is committed. Both must be disallowed.
- Policy Disallowed Expenses
Recommendation: There must be no adjustments for fines and penalties, regardless of the amount.
The rules mention there is disagreement over the definition of the Arm’s Length Principle (ALP), with some preferring it be restricted to the OECD’s narrow definition as contained in the Transfer Pricing Guidelines (TPG).
- Definition of Arm’s Length Principle
It must be mentioned that the TPG were designed to primarily allocate profits to the industrialized OECD countries through the so-called Functions Assets and Risks (FAR) approach. There are other forms of transfer pricing apart from those contained in the TPG, such as in the UN Practical Manual on Transfer Pricing and the varying practices by developing countries.
Recommendation: The ALP definition should be inclusive and not restricted to the OECD TPG.
The rules try to restrict profit attribution to a Permanent Establishment as defined by Article 7 of the OECD Model Tax Convention (MTC).
- Definition of Profit Attribution
Again, there are various Model Tax Conventions in the world such as those of the UN, African Tax Administration Forum (ATAF), Association of Southeast Asian Nations (ASEAN), Common Market for Eastern and Southern Africa (COMESA), etc. There is no reason why the narrow definition of only one Model Convention must be followed, that too when it was designed to benefit the developed countries.
Recommendation: The approach to profit attribution should be inclusive and not restricted to that of the OECD MTC.
SOUTHNEWS is an e-newsletter service of the South Centre providing information and news on topical issues from a South perspective.
To view other articles in SouthNews, please click here.
For more information, please contact Anna Bernardo of the South Centre: Email firstname.lastname@example.org, or telephone +41 22 791 80 50.