International Co-operation and Tax Administration Division
Centre for Tax Policy and Administration
Organisation for Economic Co-operation and Development
Keidanren appreciates the opportunity to submit its comments on this public consultation document.
Pillar Two seeks to address the remaining challenges facing the Base Erosion and Profit Shifting (BEPS) project and stop a harmful race to the bottom on corporate taxes. Implementing government policy measures requires financial resources so it is understandable that governments recognize these challenges as they endeavor to secure a tax revenue.
To prevent tax avoidance, however, many issues can be sufficiently addressed by the existing BEPS recommendations, which cover transfer pricing, controlled foreign companies (CFC), limits on interest deductions, neutralization of hybrid mismatches, and preventing tax treaty abuse. The Pillar Two proposals significantly overlap those recommendations.
Compliance burdens on multinational enterprises (MNEs) have increased sharply in recent years as a result of BEPS recommendations, and we are concerned about the additional burdens Pillar Two might impose. Rules that target low-tax situations, without considering economic substance, would adversely affect cross-border trade and investment. Even if a race to the bottom on corporate taxes poses problems, we are unconvinced that a global minimum rate of tax is the right solution. We should reach a consensus on long-term solutions to Pillar One issues before agreeing Pillar Two proposals. The interaction of Pillar Two and Pillar One should be addressed as it is conceivable that an adjustment under one pillar could lead to a different outcome under the other pillar. When applying the rules under Pillar Two, the concept should be streamlined; the income inclusion rule should take precedence over the undertaxed payment rule. This may reduce complexity and administrative burdens.
All jurisdictions face BEPS risks and have their own tax systems which address issues identified from the wide range of discussions to date. If the Pillar Two proposals need to be implemented, compatibility with existing tax systems should be considered. It is desirable that a wide range of jurisdictions reach agreement on the framework of Pillar Two, but the final recommendations should ensure measures to reduce administrative burdens and require coordination to eliminate double taxation among countries. A grace period should be set before the rules are introduced, and implementation should be at the discretion of each jurisdiction.
The design of any new rules should consider how to eliminate overlap with the CFC regime. One straightforward option will be to adopt a jurisdiction's CFC regime, provided that it is consistent with the recommendations in the final BEPS Action 3 report. Finally, in order to avoid double taxation and increased administrative burdens, the system should be designed so that the ultimate parent company is taxed only in its country of residence.
From these perspectives, we set out our responses to the questions below.
2. Responses to the Questions
(1) Tax Base Determination
Whether the use of financial accounts would simplify the proposal depends on the blending approach.
If an MNE's tax base is adjusted based on the accounting items presented in the consolidated financial statements, and if the subsidiaries which are excluded from those statements based on materiality may also be excluded from the blending calculation, then this approach will likely help to reduce compliance burdens and using financial accounts as a starting point may be appropriate. On the other hand, if accounting figures are allocated between the ultimate parent's jurisdiction and the other jurisdictions, an additional compliance burden will arise.
The use of financial accounts requires accounting figures to be allocated in a way that corresponds to the blending approach adopted (i.e., allocation between the ultimate parent's jurisdiction and the other jurisdictions, on a jurisdictional basis, or on an entity basis). The problem is that the effects of intra-group transactions, such as unrealized gains from product transactions, are eliminated on consolidation. Hence, these consolidation adjustments would have to be reversed to enable the accounting figures to be allocated as required. We are concerned that reversing the consolidation process to allocate accounting figures in a predetermined way might not lead to much simplification. Where a subsidiary compiles a subset of consolidated financial statements on a sub-group basis, a parent company may be unable to trace the data provided by individual entities that constitute the sub-group. It would be onerous for subsidiaries to reverse their sub-group consolidations upon the request of the parent company, and doing that work accurately would be a challenge.
Under the jurisdictional and entity blending approaches, using financial accounts as a starting point would not be beneficial because these approaches require the MNE group's accounting figures to be broken down to an entity by entity basis. There would then be no need to use the financial accounts as a starting point. Instead, in jurisdictions like Japan where an established method already exists for computing effective tax rates under CFC rules, MNEs should be allowed to decide that effective tax rates for the GloBE proposals are computed in the same way. Otherwise MNEs would need to do the same work twice which would be an undue compliance burden.
If the MNE's consolidated financial statements are to be used as a starting point to determine the tax base, adjustments to comply with accounting standards that are acceptable for the purpose of the income inclusion rule may be reviewed by accounting auditors. The rules should be as simple as possible to avoid additional complexity and cost being introduced into the review process.
Constituent entities of Japan headquartered MNE groups often use local accounting standards in their respective jurisdictions to prepare their financial statements. When these entities are included in the MNE's consolidated financial statements, financial data is adjusted to conform with the ultimate parent's accounting standards. However, this data is in many cases not managed separately from other data adjusted for consolidation, such as the elimination of intra-group transactions. Thus, it would introduce a substantial compliance burden and involve a considerable risk of error to identify and extract individual entity data requiring adjustment and obtain accounting figures in accordance with the ultimate parent's accounting standards. Adopting a jurisdictional or entity blending approach requiring the use of the ultimate parent's accounting standards would not be realistic. We therefore believe that the consequences of using the accounting standards applicable to the ultimate parent entity—in other words, using the consolidated financial statements—would essentially be to adopt the worldwide blending approach.
The treatment of subsidiaries that are excluded from the MNE's consolidated financial statements based on materiality should be carved out from the Pillar Two rules, otherwise compliance burdens would increase and additional fees may be charged by accounting auditors to review adjustments that are made solely for tax purposes. MNEs should be allowed to use the local accounting figures of their subsidiaries—the only figures available regarding those subsidiaries—to minimize compliance burdens.
While there are items that are treated differently depending on the accounting standard adopted, such as the amortization of goodwill, a company's accounting profits under any accounting standard will match its free cash flow over the life span of the business. Ultimately, all differences among accounting standards can be reduced to timing differences. Over the long term there should be no concern that allowing more than one accounting standard will place some MNEs at a competitive advantage.
To ensure comparability, however, accounting standards should be limited to those commonly used, including IFRS, US GAAP, and Japanese GAAP. Consideration must also be given to consistency with the accounting standards used for the determination of Amount A under Pillar One.
Even if the number of accounting standards is appropriately limited, reliability should still be considered further. For example, the interpretation of IFRS is principle based and differs in practice by jurisdiction. Further scrutiny may be needed to determine whether there is material divergence that cannot be dismissed for the purpose of determining tax base under the income inclusion rule.
To alleviate compliance burdens, it is crucial to reduce the number of items requiring adjustment to the extent possible. An important source of permanent differences that is common across many jurisdictions is the exclusion of dividend income. To eliminate double taxation, dividend income should also be excluded from the tax base. Careful consideration should also be given to the treatment of capital gains and losses, which may be taxable or tax-exempt, depending on the jurisdiction.
Adjustments should be limited to those that MNEs can make based on the information they can obtain from the consolidated financial statements. It is essential to specify the scope and take a pragmatic approach. An appropriate method will be to subtract permanent differences from profit before tax.
As the treatment of dividend income varies among jurisdictions, adjusting based on individual tax laws would be an onerous task. Accordingly, simplified calculation methods should be allowed, with one option being to deduct all dividend income recognized as accounting income.
We do not endorse any particular approach at this stage, but deferred tax accounting will be preferable from the perspective of compliance burdens. Our observations follow.
- Carry-forward of losses and excess tax
Based on the premise that temporary differences will eventually reverse over the course of the MNE's lifespan, the use of carry-forward of losses and excess tax is understandable as a concept, so long as the figures used are limited to those normally used for consolidated financial accounts. There is, however, concern that this approach would significantly increase compliance burdens because carry-forwards must be tracked over a relatively long period of time in order for this approach to work.
Attention should also be given to the impact of accelerated depreciation of noncurrent assets, which could increase the likelihood of the MNE's effective tax rate falling below the minimum tax rate and rendering it subject to tax. Carrying forward losses and excess tax might also risk the tax point being brought forward, thereby effectively negating the benefit of accelerated depreciation. This is a serious concern for leasing and other sectors in which business models based on accelerated depreciation have been established. The approach might render businesses in these sectors unfeasible, causing asset liquidity to deteriorate and financial markets to stagnate on a global scale.
- Deferred tax accounting
This approach is simple and practical as it is consistent with the use of financial accounts as a starting point and allows existing data to be utilized.
Still, the treatment of valuation allowances as deferred tax assets should be considered further. For financial accounting purposes, a valuation allowance is recognized as a deferred tax asset according to its recoverability. This allowance may create volatility as it fluctuates based on the estimated recoverability which is independent from the temporary differences arising. Thus, valuation allowances may need to be excluded from the computation of effective tax rates. If some portion of the deferred tax assets are ultimately unrecoverable, however, it may be appropriate to reflect the valuation allowance for the unrecoverable portion in the computation of effective tax rates. A more detailed examination should be made.
Unless the worldwide blending approach is adopted, calculating the effective tax rate based on deferred tax accounting would be difficult. In that event, a pragmatic method, such as adjusting temporary differences over a specified number of years, might be necessary.
- Multi-year averaging
If an MNE's temporary differences mostly reverse over a relatively short period of time, this approach will work. In reality, however, MNEs may have long-term temporary differences, such as those arising from the taxable write-off of investment securities. The multi-year averaging approach would have a limited effect in levelling off tax rates and would be insufficient as a mechanism addressing temporary differences. Averaging over a longer period would also increase compliance burdens.
An MNE's share in an investee's profit or loss gives rise to a temporary difference at the parent level, which often will not reverse until the investee distributes dividends or is divested. Under any of the three approaches, an MNE's tax base should always exclude its share in an investee's profit or loss (as well as the tax effect thereon if the deferred tax accounting approach is adopted).
Due consideration should also be given to different financial year-ends. The calculation of GloBE should only be performed on an annual basis and after the year end of the ultimate parent company.
Calculations based on information that is not currently compiled by the corporate group would be extremely burdensome and will lead to disputes. We do not endorse an approach at this stage, but worldwide blending appears to be basically viable as the administrative burden of its application is relatively low. Conversely, blending by jurisdiction is problematic due to the administrative requirements. As the impact will vary depending on the minimum tax rate, we expect the Inclusive Framework to indicate the proposed rate at an early stage. Keidanren has evaluated each blending approach below.
- Worldwide blending
If financial accounts are used as a starting point, the worldwide blending approach would result in the lowest compliance costs among the three approaches. We expect this approach to also resolve issues regarding the treatment of partnerships. To prevent excessive tax avoidance at an MNE group level, the worldwide blending approach is again deemed the most appropriate approach. However, if this approach is designed such that it requires MNEs to aggregate the income and corporate taxes of all foreign subsidiaries with few exceptions, such as the US global intangible low-taxed income (GILTI) regulations, this approach would generate the highest compliance costs. Thus, if this approach is adopted, it is essential to create a system based on information presented in the consolidated financial statements.
Whereas concerns about tax avoidance in foreign jurisdictions are understandable, to prevent significant increases in compliance burdens allocations between the ultimate parent's jurisdiction and the other jurisdictions should not be required. If the effective tax rate of the entire group exceeds the minimum tax rate by more than a certain level, it does not seem necessary to perform calculations that distinguish the jurisdiction where the ultimate parent company is located from other jurisdictions. Comparing numbers based on the consolidated financial statements with the minimum tax rate is the most feasible method, and the minimum tax rate should be set at a low level with reference to the GILTI level.
- Jurisdictional blending
Whether using financial or tax accounts as a starting point, the jurisdictional blending approach would be inappropriate due to high compliance costs. This approach requires an MNE to obtain the individual financial accounting figures of all group entities and then to aggregate and apportion income and taxes on a jurisdictional basis, including apportionment between branch and head-office jurisdictions and use of deferred tax accounting figures that are not currently computed. Investment in system renovation would also be required (refer to 1a) above again).
- Entity blending
In a jurisdiction that adopts the entity approach under the CFC rule, MNEs should be given a choice to use the effective tax rates computed under their existing CFC rules. The purpose of Pillar Two and the CFC tax regime are similar; both systems require a parent company to pay additional taxes when the effective tax rate is low, especially when entity blending is adopted. Because of this similar nature, and as mentioned in "1. Introduction", the target company of Pillar Two should be exempted from applying the CFC tax system, or if a sufficient CFC tax system has been introduced in response to tax avoidance, the introduction of Pillar Two should be postponed, and measures to replace it with the CFC tax system should be considered.
Although the worldwide blending approach could theoretically reduce volatility, describing it as effective at managing volatility may be an overstatement. For example, if temporary differences arise in some entities from the recognition of impairment losses on goodwill, provision for retirement benefits, and other items, without a mechanism addressing temporary differences, volatility could not be fully absorbed, producing unreasonable results.
In principle, there should be no problem if these apportionments follow what is done for tax purposes. However, the taxation rules for overseas branches differ from country to country, and the tax treatment of pass-through entities is stipulated in individual tax treaties. The administrative burden of accurately following these rules and treatment at the parent company will increase. Jurisdictional blending is difficult.
If an MNE has paid sufficient tax on particular income, it would be unreasonable for this income to be subject to further tax under the GloBE proposal and this must be prevented. Ideally, the approach should attribute tax to the jurisdiction or entity where the corresponding income was recognized. In practice, however, this would be highly difficult to implement and requires further examination. The determination of tax attribution should also be considered with respect to withholding tax and where double taxation occurs due to transfer pricing.
While MNEs can identify the taxes paid by the ultimate parent on the income of the subsidiaries or branches, at present they cannot identify the detailed items of taxes paid in intermediate jurisdictions, or paid by intermediate entities, and by subsidiaries resident in the ultimate parent's jurisdiction. We are concerned that crediting taxes paid in intermediate jurisdictions or intermediate entities might significantly increase compliance burdens as this requires MNEs not only to gather information on taxes paid but also to understand local systems, laws, and regulations. It is important to set clear rules for calculating jurisdictional blending. We are also concerned that the burden of business operations to collect and understand information will be very large. Consider the following example.
Example: A Japanese company has a holding company in Country X, which holds an overseas subsidiary in Country A (in Europe), which has branches in Country A and Country B (in South East Asia). Assume that the overseas subsidiary is taxed under the CFC tax system in Country X. To which country should tax paid under that CFC tax system be allocated?
(3) Carve-Outs and Thresholds
The BEPS final reports reiterate the project's fundamental principle that "profits are taxed where economic activities take place." Considering this principle, taxing the parent on the income generated by a group entity with economic substance would not be justified. Thus, carve outs are unavoidable.
In theory, we consider that the following should be carved out provided that the activities concerned have economic substance: tax credits for research and development expenditures and capital investments; and tax breaks aimed at encouraging certain investments, in the form of reduced tax rates or deductions from income. In principle, tax practices that are deemed not harmful in accordance with BEPS Action 5 should also be carved out.
The problem, however, is that carve-outs like these—namely, carve-outs based on facts and circumstances—would increase compliance burdens. In the worldwide blending approach, which uses consolidated financial statements as a starting point, we are also concerned that extracting the accounting figures of the subsidiaries with economic substance could be an extremely onerous task. In order to reduce administrative burdens, a worldwide blending approach should not require specific exclusion items to be individually extracted and calculated to eliminate the impact. Even if exemptions are set, a more suitable method might be to deduct a fixed percentage of fixed assets.
Furthermore, consideration should be given to the treatment of loss-making entities within the scope.
To reduce compliance costs, thresholds based on taxpayer size must be set. At least in the beginning, the scope should be limited to MNE groups with substantial revenue or profit on a consolidated basis. Whereas the €750 million revenue threshold applied to country-by-country reporting requirements may be used as a reference, narrowing the scope further may be an option. For simplicity, it may be appropriate for the threshold for Pillar Two to match the threshold set for Pillar One.
In many cases most of the income of an MNE group may be generated by a small number of group entities. When adopting the entity blending approach, it might therefore simplify clerical work to establish a de minimis threshold for individual entities that facilitates making determinations.
Moreover, given that subsidiaries excluded from the consolidated financial statements are done so in terms of monetary materiality, such subsidiaries should be excluded from the calculations required by the GloBE proposal as well.
Subcommittee on Taxation