Bloomberg Tax
July 8, 2020, 7:00 AM

INSIGHT: The Unsurprising Stalemate on Digital Taxation

John L. Harrington
John L. Harrington

Break-ups are often messy. Even in relationships that are going nowhere and where ending them is just a matter of accepting the obvious, recriminations and second-guessing are all but unavoidable.

Recent press reports indicate that a sad, but inevitable, break-up is occurring over irreconcilable tax differences. On June 12, 2020, U.S. Treasury Secretary Mnuchin sent a letter to the Finance Ministers of France, Italy, Spain, and the U.K. requesting a “pause” in ongoing negotiations on the taxation of companies that provide digital services. This project, led by the Organization for Economic Cooperation and Development (OECD), involves 137 countries and jurisdictions that call themselves the “Inclusive Framework on Base Erosion and Profit Shifting.” On June 17, 2020, U.S. Trade Representative Lighthizer confirmed this position, implying that the U.S. would no longer participate in this high-profile international project.

The Creation of the Digital Taxation Project

Perhaps understandably, many have viewed this U.S. “need for some space” as a surprise. The Inclusive Framework started the project with high hopes. The Inclusive Framework and OECD had developed and released 15 Action Items to address base erosion and profit shifting (BEPS) in 2015. The participating jurisdictions were able to claim satisfactory resolution of most of the challenges they identified as part of the BEPS project, but how to deal with income earned as a result of the so-called digital economy was one loose end that merited follow-up work. Interested countries, including the U.S., kept working to define the specific concerns raised by “digitalisation of the economy,” as the participants termed it in an interim report.

For some countries, the concern was pretty straight-forward: large Internet-based companies were earning income attributable to sales and services to consumers within their borders but not paying any income tax to the country where those consumers reside. For other countries, the concern was not specific to Internet-based companies but derived from broader business and technological changes. These countries viewed shifts in the global economy as justifying an increased ability of “market jurisdictions” (i.e., the jurisdictions where the consumer was located rather than the jurisdiction where the income was considered to be earned under international tax rules) to tax companies. Still other countries, long dissatisfied at an international tax system that they viewed as favoring countries where businesses were located (“residence countries”) over jurisdictions were income was earned (“source countries”), saw this particular digital taxation issue as grounds to revisit the long-established (but increasingly shaky) limits on the rights of source countries.

Thus, countries generally agreed as to which types of companies were in the bull’s eye of the target but disagreed as to the size and composition of the target itself. Given the different motivations, this meant that the Inclusive Framework/OECD project necessarily had amorphous boundaries and a roving focus. Because it is hard to agree on a solution if the participants cannot agree as to the problem, the description of the project and its aims evolved over time, all with the unstated goal of identifying and describing a “solvable” problem.

The unifying theme among the participants seemed to be that an agreement had to be reached. That meant—at least among the expected beneficiaries of such an agreement—that the breadth of support for the eventual agreement, not its specific terms, was the most important thing. Following this pragmatic approach, the participating jurisdictions sought to move forward by regularly issuing progress reports, announcing a “unified approach,” and building “pillars” to support this new tax approach. As with the original BEPS project, artificial and unrealistic deadlines were set to force decisions and to divert and exhaust both the proponents and opponents. The result, at least at a conceptual level, were two separate, complementary approaches. Pillar One, which dealt with nexus and allocation of profits in the new digital world, was the one most focused on market countries’ desire to tax the income of digital services. Pillar Two, which dealt with the rights of jurisdictions to tax income that other countries fail to tax either intentionally or inadvertently, addresses broader tax concerns that go well beyond digital services.

Difficulty of Reaching Agreement

Nonetheless, actions to accelerate decision-making work only if a decision is possible. When it comes to designing a specific digital taxation proposal as part of Pillar One of the Inclusive Framework/OECD project, optimism and effort have not been able so far to overcome political and policy difficulties. As noted above, the primary justification for reaching agreement on digital taxation has been the “failure is not an option” mantra. Politically, failure to reach agreement would be an embarrassment not just for the countries that have invested so much time and resources but also for the supranational organizations, such as the G-20 and European Union, that have endorsed and encouraged the project. Economically, the consequences, from both a tax and trade standpoint, of jurisdictions undertaking unilateral actions rightly strikes fear. The expectation was that failure to reach agreement was in no country’s interest and therefore should act as a strong inducement to compromise.

However appealing this line of reasoning may be, the fact remains that it is impossible to expect a diverse group of 137 jurisdictions to agree on a specific approach in the current environment. “Impossible” is a strong word, but given the choices before the jurisdictions in the Inclusive Framework, the different preferences they have, and the lack of a dominant decision-maker that could force an outcome on unwilling participants, lack of agreement on a specific proposal is inevitable. This conclusion is not based on my pessimism or cynicism, but rather on the dynamics of group decision-making. Optimism is no match for Condorcet’s Paradox and Arrow’s Impossibility Theorem.

I am a tax attorney, and not a social scientist or game theory expert, and so my explanation of Condorcet’s Paradox and Arrow’s Impossibility Theorem will be simplistic. In short, however, they address situations in which there are multiple parties and at least three options. In such a scenario, if each option lacks majority support and there are significant differences in the parties’ second and third choices, then agreement will be elusive. Any tentative agreement reached will be unstable because another alternative will be equally or more appealing to a majority of the parties.

To illustrate Condorcet’s Paradox, take the simple case of three alternatives: red, blue, and yellow. Assume that red is the first choice of 1/3 of the parties, the second choice of 1/3 of the parties, and the third choice of 1/3 of the parties. The same is true of blue and yellow. At first blush, red should have the support of 2/3 of the parties: it is the first choice of 1/3 of the parties and the second choice of 1/3 of the parties. However, the 1/3 of the parties of which red is their last choice are unlikely to accept red as the group’s choice. Instead, they will recommend that the group adopt blue or yellow. That way the anti-red parties get their first or second choice, and the group should be no worse off because either blue or yellow should similarly be supported by 2/3 of the parties. But, then the process would endlessly repeat itself, as those stuck with their last choice will advocate a color that has 2/3 support. The result is stalemate because, unless one group of parties is powerful enough to force their choice on the group, each choice is equally supported and opposed. Arrow’s impossibility Theorem, reflecting the formidable intellect of Kenneth Arrow, demonstrates the parties’ inability to resolve this situation, at least under the assumptions identified.

Granted, this is a gross simplification and, for purposes of simplicity, assumes an even split in support. In real life, if one option is much more popular than the other two, its opponents may find opposition too costly to continue to oppose it. Similarly, if one option is much less popular than the other two, its proponents may give up.

Nonetheless, the paradox or impossibility theorem—whatever one wants to call it—is applicable in this digital taxation case because there are at least three viable options:

  • One option (the “no new digital tax option”) is to refrain from adopting any new tax rules specific to digital companies. This option has one big supporter, the U.S., but not much public support among other jurisdictions.

  • The second (the “digital services tax option”) is to adopt a specific tax on companies that provide digital services in that jurisdiction. This is what the jurisdictions seeking to impose digital services taxes have been advocating and which they say they will do if the Inclusive Framework/OECD project fails.

  • The third (the “change taxing rights option”) is to modify a jurisdiction’s existing domestic tax and treaty rules so that jurisdictions can tax what they consider to be a more appropriate amount of income of companies providing digital services in that jurisdiction. This is what Pillar One of the “Unified Approach” of the Inclusive Framework/OECD project has been seeking to do.

The U.S.’ first choice, the no new digital tax option, is most jurisdictions’ third choice. The U.S. cannot agree to the second option, the digital services tax option, which is many jurisdictions’ first or second choice. So, from the U.S.’ standpoint, the only option that it can accept that would be the first or second choice of other jurisdictions is the third option, the change taxing rights option. But, unlike the other two options, the change taxing rights option has to be developed. It is complicated, novel, and being constructed in a very short time frame, a combination that is creating unease for both the jurisdictions involved and the affected companies. So, even if there was conceptual agreement as to the desirability of this option, there is significant execution risk given that it is still under-development and will have to be released with limited ability to refine it as a result of public response.

So, Arrow’s Impossibility Theorem suggests that participating jurisdictions will not be able to reach agreement on a specific approach as long as levels of support among the options do not change. Further, even if the participating jurisdictions could agree to adopt the change taxing rights option, the only option likely to be supported by both the U.S. and a large number of other countries, that option is yet to be unveiled and may lose support as details are fleshed out. So, even if the U.S. can convince (or impose its will on) a sufficient number of jurisdictions, its version of the change taxing rights option would have to be transformed into a detailed set of rules that work across jurisdictions.

Falling Back to a More Limited Agreement

If agreement on a single approach at the Inclusive Framework is (paradoxically) impossible, the participants in the Inclusive Framework/OECD project still have the incentive to seek some kind of agreement. The participants have warned that the likely result of no agreement is the uncontrolled chaos of unilateral action: countries enacting their own digital services taxes, countries like the U.S. retaliating through tariffs or other sanctions, trade and tax wars as one country or group of countries matches or raises defensive or provocative actions taken by another.

If each of the options will have a hard time finding a stable majority and the deadline for making a decision is fast-approaching, the issue becomes what kind of agreement can be reached. Even though all of the members of the Inclusive Framework may not be able to agree on an option, a subset of members may be able to agree on an option. In that case, a group, rather than all, of the jurisdictions would reach agreement on an option and agree to follow that option. How well this “coalition of the willing” approach would work depends on how amicable the split-off is and how the excluded parties decide to react. One possibility is that jurisdictions split off in a coordinated way: the participants sanction a menu of alternative, authorized approaches. Different—but specific—approaches of taxation are described and best practices identified, and the separate groups follow their options in an mutually-agreed-upon way. Even though real agreement is not possible for the reasons noted above, the jurisdictions could agree on how to handle their disagreements. As long as countries stay within the agreed-upon constraints, they can legitimately say that they are acting in accordance with international agreement even though they are following different approaches. This compromise, though less desirable than agreement on a specific outcome, allows them to claim victory and implement their decisions in a more orderly way.

Default in Absence of an Agreement

But, suppose the participating jurisdictions cannot agree on a specific menu of choices. In that case, the real question is what is the default outcome if the jurisdictions cannot agree.

One possible default outcome is the “I’m feeling lucky” scenario. If a majority of jurisdictions—specifically excluding the U.S.—can reach agreement on an option, they may think they have sufficient safety in numbers to prevent (or at least blunt) the U.S. from retaliating. Whether they guess correctly will determine whether the result is the chaotic outcome identified as the motivation for the Inclusive Framework/OECD project. Not being a gambler, this choice scares me. I understand why it is often raised as a foil to encourage agreement between the participants, always based on the assumption that saner heads will prevail. But, if agreement is not possible, no carrots or sticks will work as an inducement to reach agreement. Rationality leads to the stalemate, not the way out of it.

That leaves us with what is hopefully more likely the default outcome: some kind of controlled chaos. One scenario is that countries take the position that they tried to reach a compromise, and given the failure of the U.S. to make concessions, they are free to go their own way. If they believe that the U.S. will not follow through on its threats, this is likely the preferred outcome for quite a few countries.

Many of the countries that have taken steps to enact a digital services tax will find it difficult to accept some compromise in the Inclusive Framework/OECD project that falls short, from a political and/or revenue standpoint, of the digital services tax they have announced. If their choice is between accepting a scaled-back version in the Inclusive Framework and letting their digital service tax enter into force due to a failure of the members of the Inclusive Framework to agree to a compromise, it is clear the default outcome is preferable to them, as long as they are not punished by opponents of the digital service tax. If there are enough countries imposing such a tax in a coordinated way, they may be able to prevent opponents (whether they be countries like the U.S. or affected companies that are subject to the tax) from effectively punishing them.

Trade sanctions or economic boycotts can be quite effective if aimed at one or two countries; threats against half the world is another matter. So, if enough countries think that they can act unilaterally because the threats of punishment are unlikely to be carried out effectively, they may be unwilling to accept any compromise that is not as advantageous to them as this default of their digital services taxes. Indeed, the very possibility of this outcome is one of the factors that keeps agreement in the Inclusive Framework impossible.


If the current Inclusive Framework/OECD project is stuck within Condorcet’s Paradox and Arrow’s Impossibility Theorem, then the current impasse on digital taxation cannot be solved simply by the U.S. being more flexible or other jurisdictions being more cooperative. As noted above, agreement is impossible in the current situation due to the participants in the Inclusive Framework, the number of options, and their preferences. The options are unlikely to go away, but the participants and their preferences could change. In that happens, an option might command a majority preference.

Accordingly, decisions by one or more participants can have some interesting consequences. For example, if the U.S. were to cease to be an active participant in the Pillar One discussions, then agreement among the remaining participants might become possible. A “coalition of the willing” agreement among remaining participants would just be one that the U.S. strongly dislikes, and an outcome made possible only by the U.S.’s withdrawal. That leaves the U.S. with a rather unsatisfactory choice: continue participation, only to see some sort of controlled chaotic outcome, or leave the negotiations, only to see agreement on an outcome it does not like. Further complicating things is the U.S.’s apparent desire to continue to participate in one aspect of the ongoing Inclusive Framework/OECD project (the Pillar Two discussions) while disengaging from another (the Pillar One discussions). It remains to be seen how amenable the other participants will be to the U.S. “participating under protest” when and how the U.S. sees fit and how those actions will affect the Pillar Two discussions.

As long as the problem is unclear, it is hard to identify good outcomes, much less plot a path toward them. Fundamentally, this is an evolving situation, as demonstrated by the various proposals and trial balloons coming from the OECD and Inclusive Framework over the past few years. As long as technological changes continue to impact our lives and cause us to change the way we do business, countries’ tax systems will become out-of-date and cause the process to begin all over again as new options are identified and preferences change. Looked at this way, the U.S. is not making a one-time decision about participation in this project. Its reaction, and the response of other jurisdictions to its reaction, will color the relationship going forward. The U.S. can take a pause, but it can’t leave the relationship it has with other countries on digital taxation. Even if this is a marriage of convenience, they are bound together, for better or for worse.

This column does not necessarily reflect the opinion of The Bureau of National Affairs, Inc. or its owners.

Author Information

John L. Harrington is one of the co-leaders of Dentons U.S. tax practice.

Learn more about Bloomberg Tax or Log In to keep reading:

Learn About Bloomberg Tax

From research to software to news, find what you need to stay ahead.

Already a subscriber?

Log in to keep reading or access research tools.