Policy recommendations

  • Progress is being made on combating corporation tax avoidance at international level. After a pause pending the results of the 2020 presidential election in the USA, discussions have resumed in an ‘inclusive framework’ within the OECD. The recent decisions by the Finance Ministers of the G7 and by the ‘Inclusive framework at the OECD’ confirm the political momentum to combat tax avoidance.
  • There are two conflicting models: continuing to rely to the utmost on a model that regards the various entities of a multinational group as separate entities, with their tax base set by transfer pricing rules and based on physical presence, or moving over to unitary taxation, with the tax base allocated within the group – and among countries – on the basis of tangible factors, such as jobs or sales. The preferred option is the second one, and the proposals under discussion are a first step in this direction.
  • The introduction of a minimum level of taxation is certainly a step forward. However, care must be taken to ensure that the level and scope are right in order to put an end to tax competition, which is detrimental in both economic and budgetary terms. 


On 5 June 2021, the G7 reached agreement on a minimum taxation rate for multinational companies. This was described by the press as historic, albeit with the qualification that this was only an initial stage. The second stage has just taken place: on 1 July 2021, 130 of the 139 members of the ‘inclusive framework’ approved the agreement. Among those that did not approve it were Estonia, Hungary and Ireland. The ‘inclusive framework’ was established in 2016, on the initiative of the ECD (Organisation for Economic Co-operation and Development) and the G20 (group of the 20 richest countries in the world), to discuss reforming the taxation of companies at international level. Among these 139 members are 66 developing countries and most tax havens.

A few weeks earlier, on 18 May, the European Commission had published a Communication on business taxation for the 21st century (European Commission 2021), which also supports fairer taxation and a reduction in tax competition. The lines are definitely shifting. It is worth mentioning that, at the start of this century, the European Commission confined its comments on taxation to stigmatising ‘tax obstacles’, such as barriers to the internal market (Valenduc 2002).

In this publication, we give a brief overview of developments in recent years and the discussions in progress within the OECD and the G20, which it is hoped will be completed in July 2021. First we set out the background to the issue and then go on to present some initial evaluations.


International taxation is based on very old rules, some of which date back to 100 years ago this year. These rules come from a time when the proportion of goods traded within multinational companies was low and trade was a matter of clearly identifiable intermediate goods or finished products. It was easy to establish where a product was manufactured, where an economic activity gave rise to a profit and the residence of a shareholder. A company established itself physically in a country in order to sell and make profits there. These days, a company can make profit in a country without being established there. A given final product incorporates parts that may be manufactured in different places. Intangibles have become significant, in terms of both company assets and in the design and manufacturing processes of a good. A large proportion of trade takes place between companies in the same group, and applying the market price to them – the basic rule for transfer pricing – has become a challenge, when the main point is the singularity of a particular part of the product. In this context, it is an economic fiction to regard the various companies within a multinational group as separate entities, even if this is still in line with the legal reality: they are parts of a whole.

Tax avoidance has swept into the gaps in the current system. By exploiting transfer prices, the location of intangible assets and their income stream, and the financing structures of the group’s businesses, multinational companies shift the tax bases to low-tax jurisdictions and/or take advantage of preferential regimes established in countries with ‘normal’ tax regimes, primarily patent boxes. In this way, they can significantly reduce their effective taxation. This leads to downward pressure on corporation tax revenue and unfair taxation. Beyond these strictly fiscal aspects, their dominant position may be economically ineffective (Sorbe and Johansson 2017).

From the BEPS plan to minimum taxation

In 2013, the OECD and the G20 countries took stock of the situation (OECD 2013a) and adopted an action plan against ‘Base Erosion and Profit Shifting’ (OECD 2013b), known as the BEPS plan. This plan was based on 15 actions, intended to address the various aspects of the problem: taxation of the digital economy, transfer pricing rules, preferential regimes, the improper use of bilateral agreements to secure zero taxation, excessive interest deductions and hybrid arrangements. The action plan also provided for an economic evaluation of the extent of tax avoidance and its impact.

In 2015, an agreement was reached (OECD 2015) comprising three decision-making levels, four binding minimum standards, an upgrading of the transfer pricing rules and basic rules for international agreements and recommendations for best practices, in particular to neutralise the effect of hybrid arrangements and counter excessive interest deductions. The European Commission took this up and, soon afterwards, put forward a proposal for an Anti-Tax Avoidance Directive (ATAD), which went further in some areas, including as regards interest deductions (European Commission 2016). The proposal for a Directive was adopted within just a few months, which is an extremely rare event in tax matters.

At the OECD/G20 level, the 2015 agreement left the issue of the digital economy unresolved. Work on this subject resumed, and an interim report was published (OECD 2018) which, among other things, included a very interesting economic analysis of business models. This report raised some important points regarding taxation in the digital economy, in particular the role of the market in value creation and the economic rent received by the sector, but did not draw any conclusions about reforms. The discussions stalled but later resumed, structured around two pillars: Pillar One on profit allocation to the country of the market and Pillar Two on minimum taxation for multinational companies.

Pillar One is designed to address the problem of the digital economy. One of the latter’s characteristics, highlighted in OECD (2018), is the capacity to generate profits in a given country without the company having a physical presence there, in particular by creating value from the user data that it collects. The basic rules for allocating taxing rights are based on the concept of ‘permanent establishment’, and the fundamental principle of the BEPS project is that, to counter profit-shifting, the tax must be levied where the value was created. Allocating part of the tax base to the country of the market is therefore justified.

But how is this to be determined? The basic idea is to tax part of the ‘non-routine profit’ in the country of the market. Discussions on Pillar One swing between two models: building on the transfer pricing rules or opting for the unitary taxation model and fixed apportionment of profits (formulary apportionment). Behind this seemingly highly technical alternative lies a fundamental choice.

The distinction between ‘routine profit’ and ‘non-routine profit’ is a transfer pricing concept, which therefore belongs in the taxation specialist’s toolbox (Devereux et al. 2019). It is a concept that is applied on a case-by-case basis and is certainly difficult, if not impossible, to translate into the basis for a new taxation system. An economist tends rather to speak of the normal return on capital (the long-term risk-free interest rate plus the risk premium) and ‘excess return’ (remaining component of the rate of return). Adopting this approach means carrying on with the traditional view: that the companies of a multinational group are separate entities. Conversely, opting for unitary taxation means starting from the principle that the multinational is a single entity and that it is not possible to proceed as though the transactions between its various constituent parts can be valued at market price.

Consensus is based on a compromise: adopting formulary apportionment but not in so many words – that is to say keeping the concept of ‘non-routine profit’ as a front but determining it on a fixed basis, as well as the share allocated to the country of the market.

The core option is to confine it to multinational companies with a consolidated turnover of at least €750 billion. Controversy is primarily about the sectors concerned. Initially, there was no distinction based on the nature of the economic activity, but later it was proposed to limit the application of Pillar One to two types of activity: the GAFA (Google, Apple, Facebook, Amazon) companies, renamed ADS (automatised digital services), and CFBs (consumerfacing businesses), a concept which is nothing if not vague. CFBs are defined as ‘businesses that generate revenue from the sale of goods and services of a type commonly sold to consumers, including those selling indirectly through intermediaries and by way of franchising and licensing’ (see OECD 2020, pp. 32-33).

Pillar Two is intended to establish a minimum taxation level without, however, removing the right for each country to determine its tax rate. The basic idea is as follows: if, in a given country, the effective taxation of a subsidiary of a multinational group is below a critical threshold (say 12.5%), the country of the parent company may tax these profits up to the level of the difference between the effective rate and the 12.5% threshold, even if the profits have not been allocated (distributed) to the parent company. There are two rules to achieve this: an ‘income inclusion rule’ and a ‘denying payment rule’, according to which payments from a high-taxation country to a lower-taxation country can be denied.

Discussions stalled in late 2020 as a result of the position of the Trump Administration, which wanted Pillar One to be optional and for the USA to be able to keep their own system of minimum taxation instead of Pillar Two.

The change of presidency opened up the way again. The Biden Administration was very quick to table new proposals. These accept the rationale of Pillar One but aim to limit the scope to the 100 largest multinationals, irrespective of the nature of the economic activity, with a threshold of US$ 2 billion of consolidated turnover and a profitability threshold of 10% to determine the excess partially attributable to the country of the market. For Pillar Two, the Biden Administration proposes a minimum tax rate of 21%. The agreement reached within the inclusive framework refers to a rate of ‘at least 15%’ for Pillar Two and a threshold of €20 billion for Pillar One.

Initial evaluation

It is not possible to evaluate these proposals fully within the limited scope of this publication. We will simply discuss the three key points here.

Taxing in the country of the market?

Taxing in the country of the market had already been proposed by economists under the name ‘destination base cash-flow tax’, and it was part of the Trump Administration’s initial tax reform proposal, but was later abandoned (OECD 2007, pp. 93 et seqq., and Auerbach et al. 2017). This taxation system incorporates
two fundamental changes compared with a corporation tax system: taxation of cash flow rather than profits in the usual sense of the term and taxation in the country of destination rather than in the country of production.

Both elements are partially to be found in the Pillar One proposal. In the case of the former, we are not talking about taxation based on cash flow, but a mechanism that ends up with a partly similar result. A cash-flow tax has the effect of exempting the normal return on capital from tax and limiting taxation to the part of the return constituted by the economic rent (‘Excess return’). Pillar One simply separates the return into two components on a fixed basis, with the ‘rent’ aspect sheltered in a tax-efficient location, transferring it to the country of the market. As regards the second aspect, taxation in the country of destination has two interesting properties. First, it recognises the principle of value creation independently of a physical presence of an economic entity in the country of destination. This point is particularly welcome for taxation of the digital economy. This kind of taxation formula can be better estimated than taxation on sales, which, for its part, is independent of  profitability. The other advantage is lower mobility of this tax base: sales to final consumers are less relocatable than production, which, for its part, is less relocatable than the location of the intangible assets, currently a determining factor for the taxation of companies in the digital economy.

Too small a step towards unitary taxation?

By defining the rate of return separating ‘routine profit’ from ‘non-routine profit’ and the proportion allocated to the country of destination on a fixed basis, Pillar One is taking a step towards unitary taxation. This comes down, in actual fact, to a formula separating profitability into two components, allocating the first on the basis of the transfer pricing rules and the second on a fixed basis according to sales by destination.

There is no denying the conceptual leap: Pillar One introduces a fundamental innovation, in that it proposes to apply the principles of consolidation and fixed apportionment to some of the profit of a multinational group at a broader level than a federal (or confederal) state or than a supranational body such as the EU.

Its excessively limited scope, with very high consolidated turnover and profitability thresholds, is regrettable, but the reform’s supporters will respond that the main thing is to take the conceptual leap and, once this has been done, the scope could be extended. In any event, it marks the end of the absolute reign of separate taxation of the entities of a multinational group and the establishment of their profits by the rules on transfer prices.

Is the minimum tax set too low?

In the context of tax competition that has prevailed over the past few decades, the introduction of a minimum tax by Pillar Two is decidedly welcome. But the devil is in the detail, so it is worth being aware of two points: the way in which effective taxation is calculated and, above all, the scope. A much-discussed question is the possibility of leaving certain preferential regimes out of scope. Patent boxes, which are still compatible with the rules of the BEPS agreement, were safeguarded in this way for a long time. For patent box regimes, the 2015 BEPS agreement required a mathematically quantified link with research expenditure (the ‘nexus’ approach). The safeguarding of patent boxes was abandoned in favour of a formula making it possible to leave regimes granting a low rate for real economic activities out of minimum taxation. This option is less detrimental, but it can be argued that it leaves the way open to tax competition.

Pillar Two also has a welcome collateral benefit. A common definition of effective taxation requires a common definition of the denominator of the fraction giving the effective tax rate, and hence the profit. One of the objections of the opponents of unitary taxation was the impossibility of securing a sufficiently broad agreement (in terms of the number of countries) on the definition of the tax base. On this point, the technical discussions on Pillar Two are a very promising step forward.

The last discussion point is clearly the level of the minimum rate. Whereas the initial US proposal was 21%, the G7 agreement sets it at 15%, and the inclusive framework agrees on ‘at least 15%’. Compared with the average taxation rates that prevailed 20 or 30 years ago, it is decidedly low. Politically, the question is whether it is necessary to put an end to tax competition or to take a step backwards, eliminating some of its effects. If the rate is too low, it could be the signal for tax cuts. From a pragmatic viewpoint, it is a choice between what may appear to be a minimum agreement or the absence of an agreement that would keep a more detrimental situation in place. Other, more radical, proposals have been made, such as those of Barake, Zucman et al. (2021) and Cobham et al. (2021).

In this debate, it is essential, at all times, to distinguish between taxation of corporations and taxation of shareholders. The economic impact of corporation tax is much disputed: its final economic impact may be passed on to the end consumer as a price increase or to employees in the form of lower investment and hence job cuts. The final impact of taxation on shareholders is distinctly less controversial: in a globalised economy, it is not normally passed on in the cost of capital. If a country regards the minimum tax as too low, it can always raise taxation on shareholders.


The agreement reached at the G7 and supported within the OECD inclusive framework by 130 of the 139 members is certainly a positive point. In considering that limits need to be placed on tax competition and that companies must, in the words of President Joe Biden himself and the European Commission, pay their fair share, it constitutes a political shift. In terms of principles, it is the first time that a formulary apportionment approach has, albeit partially, replaced transfer prices in the determination of taxable profits. These two points are to be welcomed. Ideally, further progress would have been desirable, in particular as regards the scope of Pillar One, the level of minimum taxation and taxation in the digital sector. The recent Communication of the European Commission seems to support this, with a new approach to the common tax basis and specific taxation for the digital sector, which may be seen as taxation of the economic rent. Something to keep an eye on: the debate is far from over.

References see PDF


Buy printed version
€ 5.00
en fr

Table of contents

International taxation-Biden opens up the way to reform_2021.pdf