International Tax Rules for the Digital Era

With good policies and a willingness to cooperate
across borders, we can and should harness
these exciting technologies to improve
well-being without diminishing the energy
and enthusiasm of the digital age.

— Martin Mühleisen, director of IMF’s strategy, policy, and review department¹

Introduction

As the economy has become increasingly digital in the past three decades, pressure has increased on traditional tax and regulatory systems to keep pace. While the international community has generally formed consensus rules for international taxation, that consensus has not yet reached the taxation of the digital economy. The lack of consensus has not lessened the political will for some countries to push for a greater share of the digital economy’s taxable profits. Because some countries are moving forward with unilateral digital services taxes, now is a critical time for stakeholders to redouble efforts to reach a global consensus regarding how to tax the digital economy. The OECD has made significant and valuable progress on that front, but if the international community intends to deliver on the promises of digital technology and innovation, it is imperative that OECD members avoid a world of uncoordinated overlapping digital tax regimes.

The importance of the tax challenges of the digitalization of the economy is evident in this topic being made the first action of the OECD action plan on base erosion and profit shifting.² The 2015 action 1 report reached relatively modest conclusions, stating that it is impossible to ring-fence the digital economy for tax purposes, because it is increasingly becoming the economy itself. While it noted that features of the digital economy exacerbate BEPS risks, it concluded that those risks would be mitigated by the other BEPS action plan recommendations. Regarding broader tax challenges for allocating taxing rights between source and residence country, the action 1 report concluded that in the direct tax area, none of the options available at the time was ripe for multilateral adoption, but that further work would be done, leading to an additional report in 2020.

That limited consensus did little to assuage the concerns of countries that believed immediate action was necessary. As a result, the timeline for further work was accelerated, and under the direction of the G-20, the OECD agreed to produce an interim report in 2018. That interim report identified divisions among the countries participating in the process, but again did not recommend a particular solution or concede that action was necessary. It did, however, note that while all countries agreed that global consensus was preferable to uncoordinated unilateral action, some countries believed that interim measures were necessary. The interim report was closely followed by a proposed EU directive to impose an interim DST, which would have levied a tax of 3 percent on net revenues of targeted digital businesses.

While the push for an EU-wide DST did not succeed, several jurisdictions have adopted, or are considering, unilateral actions, including:

• DSTs and other measures targeted solely at digital businesses;

• antiabuse measures, including, for example, diverted profits taxes in the United Kingdom and Australia, and the base erosion and antiabuse tax and the global intangible low-taxed income regime in the United States; and

• increasingly aggressive or expansive interpretation of existing provisions, resulting in pressure on nexus and profit allocation standards.

Those developments are challenging the stability of the international tax system in several ways. The continued trend toward uncoordinated, unilateral tax measures that depart from the existing international consensus on the allocation of taxing rights creates risks of double or even multiple taxation.³ Taxpayers face increasing uncertainty regarding future changes to international standards; the interaction of unilateral measures with existing standards; and the application of new, more subjective multilateral measures being administered from country to country. Further, the proliferation of reforms and the unraveling of consensus is placing pressure on both governments and taxpayers in the face of growing numbers of controversies and treaty disputes.

It is becoming increasingly clear that measures to address political concerns over the tax challenges raised by the digitalization of the economy will ultimately be enacted in one form or another. One possibility is that the trend of countries adopting unilateral measures will continue unabated, resulting in a de facto abandonment of the existing international consensus. If that path is taken, the lack of coordination among those measures could result in multiple taxation of income, taxation of enterprises out of proportion to their profitability, and a chaotic transition to a new normal as taxpayers and administrations try to manage multiple uncoordinated regimes. Alternatively, countries can attempt to bypass the harmful effects of uncoordinated action by reaching agreement on a new consensus that goes beyond the BEPS measures and squarely addresses the sources of disagreement and uncertainty created by digitalization. The OECD inclusive framework is working toward the second goal. The work program issued by the inclusive framework in May outlines some of the options being considered for that purpose, discussed in more detail below.⁴

The time has come for the business community to engage collaboratively with governments and other stakeholders on these issues to reach a consensus solution to the tax challenges that the international tax system faces as a result of digitalization. Only a collaborative approach can ensure a consensus solution that squarely addresses those challenges in a way that is principle-based, administrable, and fair. As such, Uber has chosen to work with other members of the international community in support of a consensus solution to prevent the disintegration of the existing consensus and restore long-term stability. This article outlines the elements of a proposal that we believe could form the basis for a global consensus. In initial discussions with multiple countries and companies regarding this proposal, we have considered their feedback in shaping the specific design features and received broad support.

Reaching a Global Consensus

To have a realistic chance of garnering consensus in the short term and remaining stable in the long term, a viable solution must have several characteristics. First, it must be a principle-based approach. To have long-term stability, any solution must be tied to principles, which will be key to ensuring that as future business models arise, the solution will continue to be adaptable.⁵ That adaptability will help guarantee the long-term stability and sustainability of the solution, avoiding the need for it to be revisited in a “BEPS 3.0” project. Having a bedrock of identifiable, agreed principles will also help mitigate concerns that countries may have when any particular application of the proposal might result in losses of tax revenue, and it will help ensure measured and appropriate application of the proposal.

Second, the solution must squarely address key concerns. To be effective, it must focus on the specific concerns that led to the digital economy work in the first place. Uber believes the key concerns that are shaping the debate are:

• the ability to engage with customers in a market without sufficient physical presence to create effective taxable nexus in that market under existing standards;

• the ability to concentrate residual profits attributable to market intangibles in low-tax jurisdictions, thereby reducing the amount of tax attributable to market countries; and

• the ability to target advertising to nonpaying users in a jurisdiction, while receiving payments from customers that are outside that jurisdiction.

Third, the approach must be easy for companies to comply with and must be administrable by countries with widely varying levels of resources. It should be formulaic in nature, with simplifications when possible regarding:

• the scope of taxpayers covered and the threshold for nexus to ensure the burden of implementation falls primarily on businesses with the resources to manage that burden;

• the calculation of profits covered or excluded to rely on verifiable data whenever possible and minimize potential for disputes; and

• the allocation of profits to countries to minimize potential for disputes.

Fourth, the solution should preserve existing principles where possible. Any solution that reallocates income to market countries will result in some countries collecting more revenue than before, and some countries collecting less. On the other hand, a new consensus has the potential to create benefits, including reduction in disputes, increased certainty and predictability, and long-term stability. To gather broad consensus, the approach should avoid reallocating taxing rights in so drastic a way that it outweighs those long-term benefits for countries that lose revenue. If possible, it should preserve the arm’s-length standard, and when departure from existing principles is necessary to address tax challenges, the solution should provide simple and administrable coordination rules.

Fifth, the proposal should incorporate an agreement to repeal unilateral measures. Because effective development of a global consensus cannot be done overnight, political reality may mean that more countries will feel compelled to enact their own measures before negotiations are over. Because one core purpose of this work is to prevent the chaos of uncoordinated, unilateral action, any effective solution will need to address the measures that have been adopted so far.⁶ That should include both an agreement to repeal past measures (subject to a reasonable transition, as discussed below) and an agreement not to adopt future unilateral measures.

Finally, the approach should have a reasonable transition period. Any solution will require significant changes to domestic law. Before the overall solution can be effective, countries should be given sufficient time to adopt those domestic law changes in a uniform and coordinated manner. The inclusive framework should continue to provide support during that process. Also, because any solution is likely to be inconsistent with the nexus and profit allocation rules of existing tax treaties, time will be needed to allow the development and implementation of an instrument to update those treaties. Further, there will be a need to ensure that DSTs and other unilateral measures can be phased out to avoid creating a period in which double taxation arises because both proposals apply. However, that kind of approach would also need to recognize that given the political pressures involved, it would be challenging, if not impossible, for countries to agree to a transition period in which neither their existing measures nor the new proposal applies.

Uber’s Proposal

Choice of Modified Residual Profit Split

Our proposal is based on a modified residual profit split (MRPS), with simplifying measures intended to mitigate the complexity associated with profit splits while adhering to the principles identified above. In application, it results in a reasonable reallocation of profit to market jurisdictions over and above traditional transfer pricing methods. We considered various options for a proposal, including the other potential approaches identified in the OECD work program, but concluded that a modified residual profit split was the best adapted to address the largest tax challenges and the best suited to use as the basis for a principle-based solution that appropriately takes into account how value is created.

We considered a fractional apportionment method like that in the OECD work program but believe that it was unlikely to address the core tax challenges of digitalization in a way that could draw a broad consensus among countries. In particular, targeting a fraction of a multinational group’s profits without regard to whether those profits are routine or non-routine returns would likely cause a larger departure from existing principles. Doing so would also add complexity and could create more drastic changes in revenue allocations among countries that might draw taxing rights away from countries in which meaningful functions are being carried out for reasons that have little to do with taxation. Also, given the wide variation in economic profiles among the countries involved in discussing tax challenges, it is difficult, if not impossible, to identify allocation keys based on principles to which all countries would agree. Absent such principles, we are concerned that even the most formulaic approach could give rise to concerns of unfairness, resulting in inconsistent administration and increased disputes.

We also considered a distribution-based approach like the one outlined in the work program. That kind of approach would be intended to ensure that market jurisdictions collected a specified, formula-based baseline profit for distribution-related activity, resulting in a reallocation of both routine and non-routine profits from other jurisdictions. However, we believe that kind of approach is not well suited to addressing the full range of the tax challenges identified above. In particular, it does not appear to be an effective way to address the situation in which advertising is targeted at nonpaying users in a jurisdiction, while paying customers are outside that jurisdiction. It also appeared more challenging to design a proposal like that in a way that focused on clear principles for allocating revenue between market and residence countries, rather than an arbitrary, politically driven line-drawing exercise. Thus, we were concerned about both the longevity of that kind of solution and the potential for inconsistent administration to give rise to disputes.

Overview of Considerations for an MRPS

The OECD work program identifies four steps in applying an MRPS: (1) determine total profit to be split; (2) remove routine profit, either using current transfer pricing rules or simplified conventions; (3) determine the portion of the non-routine profit (or loss) that is within the scope of the new taxing right, using either current transfer pricing rules or simplified conventions; and (4) use allocation keys to allocate the in-scope, non-routine profit to the relevant market jurisdictions. Any proposal would also need to include rules to coordinate its application with the entity-level tax results under traditional international tax and transfer pricing rules. Finally, the proposal would need to define in an administrable way which taxpayers are in scope and which characteristics will create nexus.⁷

The following example illustrates with simplified figures how our proposal would be applied to hypothetical multinational company GlobalCo, which sells products and services worldwide. Assume the following facts regarding GlobalCo’s operations:

• GlobalCo’s Country A Parent sells to third-party customers in countries A, B, and C;

• Country B Support Co provides marketing support for Country A Parent;

• Country E HQ sells to third parties in Country E;

• Country F Reseller buys from Country E HQ for sales to third parties in countries D and F;

• the group has no physical presence or dependent agents in countries C or D and therefore no tax nexus in those countries under traditional tax rules; and

• the group has economic nexus (discussed below) for purposes of pillar 1 of the OECD work program in countries A-F.

Step 1: Determine the Total Profit to Be Split

Our proposal would start from the worldwide group operating profit, derived from the group’s audited financial statements, rather than examining profitability entity by entity. Recognizing that taxpayers may have multiple lines of business that might experience very different financial results, if taxpayers desire, they should be allowed to apply the proposal separately to each line of business, determined according to the business segment reporting on their audited financial statements.⁸ Reliance on financial statement data is intended to ensure that the operating profit figures are from a verifiable source with safeguards to ensure that the data are reliable and not tax-motivated.⁹

See Table 1 for a summary of GlobalCo’s consolidated income statement, which indicates worldwide group operating profits of $1,057.

Table 1. GlobalCo Factual Data

	Country A Parent	Country B Support Co	Country E HQ	Country F Reseller	Eliminations	Consolidated
Third-party sales	1,900		200	1,600		3,700
Related-party sales/ services		165	1,300		(1,465)	-
Cost of goods sold (COGS)	(435)	-	(650)	(1,300)	1,300	(1,085)
Gross margin	1,465	165	850	300	(165)	2,615
R&D expenses	(250)					(250)
Related-party support costs	(165)				165	-
Selling, general, and adminstrative expense (SG&A)	(450)	(158)	(450)	(250)		(1,308)
Profit before tax	600	7	400	50	-	1,057

Sales by Country:
Country A	1,000	Country E	200
Country B	500	Country F	900
Country C	400	Country D	700
Total	1,900	Total	1,800

Step 2: Remove Deemed Routine Profit

From the total pool of worldwide group operating profit,¹⁰ the next step would be to remove routine profit. Rather than making a case-by-case determination of what is routine profit, as would be the case for a profit split under traditional principles, our proposal would deem a routine profit on a groupwide basis in an amount equal to the greater of 4 percent of sales or 15 percent of depreciable and amortizable assets other than goodwill. That deemed routine return is intended to act as an approximation of the amount that would result under a principle-based profit split, while dramatically reducing complexity and the potential for disputes over what constitutes a routine return in a particular case. The remainder of worldwide group operating profit after the deemed routine profit is removed is intended to approximate the portion of the group residual profit attributable to intangibles¹¹:

Routine Profit = Worldwide Third-Party Sales * 4%

Residual Profit = Worldwide Operating Profits - Routine Profit

In the current example, 4 percent of worldwide third-party sales yields $148 ($3,700 * 4%).¹²

As a result, residual profit is $909 ($1,057 - $148).

Step 3: Determine In-Scope Residual Profit

Step 3a: Split Residual Profit

The remaining group residual profit after Step 2 will be split between profit attributable to product intangibles (product intangible profit, or PIP) and profit attributable to market intangibles (market intangible profit, or MIP). The overall approach is to determine the taxation of a portion of MIP under the new rules described below, while continuing to subject PIP and the remaining portion of MIP to traditional tax rules. The reason for that approach stems from the origin of the work to address the broader tax challenges created by the digitalization of the economy, which was that existing tax rules failed to take into account the role that market externalities play in creating value in a digitalized economy.¹³ Examples of that include levels of user participation and engagement that were less common in the predigital age, including the use of data gathered from nonpaying users to support advertising by paying customers. They also include the ability in some cases to locate key business functions outside a market while selling into that market remotely.¹⁴ Those concerns relate primarily to market intangibles, rather than product intangibles. As a result, our proposal is designed to split the two types of intangibles — excluding PIP entirely and subjecting only a portion of MIP to the new regime.

PIP is determined by multiplying the residual profit by a PIP-split percentage from Table 2 (the PIP Table). The appropriate PIP-split percentage for a particular group is identified by computing that group’s ratio of R&D Plus (explained below) over selling, general, and administrative expense (SG&A)¹⁵ plus R&D Plus as reflected in the group’s audited financial statements (R&D Plus Ratio),¹⁶ and then referencing the PIP-split percentage assigned to the tiering band the R&D Plus Ratio falls in. Uber has performed economic analyses confirming that the PIP-split percentage per the PIP Table closely correlates to the result that would have been determined if an actual residual profit-split computation had been completed using intangible development costs in the audited financial statements of a particular group. Uber advocates using the PIP Table (over an actual MRPS computation) because it provides simplicity without sacrificing a principled outcome. It is also expected to substantially reduce tax controversies and risk of double taxation.

R&D Plus consists of all costs related to product intangibles as disclosed in the group’s audited financial statements. They will include all costs separately disclosed as research and development, product or content development, and in-process R&D. R&D Plus costs also include amortization of any of those costs that were capitalized (for example, capitalized product development or in-process R&D costs). To ensure reliability, those costs would be included only if they were identified and disclosed in financial statements. While many of those costs might not currently be so identified, we believe that if those costs are material, separate disclosure of them in future audited financial statements will not raise major concerns for companies.

The PIP Table would assign taxpayers to one of eight tiers.¹⁷

Table 2. PIP Table

Tier	R&D Plus Ratio	PIP-Split %
1	< 3%	30%
2	3-14%	38%
3	14-18%	45%
4	18-27%	50%
5	27-42%	56%
6	42-50%	62%
7	50-65%	68%
8	≥ 65%	77%

The PIP-split percentage derived from the table would then be multiplied by the total group residual profit to arrive at PIP. Any group residual profit that is not PIP would be treated as MIP. Calculating MIP as the residual amount after removing deemed routine return and PIP is intended to eliminate potential disputes about how to determine the scope and value of market intangibles and the profit they attract.¹⁸

GlobalCo had $250 of R&D Plus expenses and SG&A of $1,308. As a result, the R&D Plus Ratio is:

R&D Plus/(R&D Plus + SG&A) = R&D Plus Ratio

$250/($250 + $1,308) = 16%

That ratio places GlobalCo in Tier 3 of the PIP Table for a PIP-split percentage of 45 percent. GlobalCo therefore calculates its PIP as:

45% * Residual Profit = PIP

45% * $909 = $409

MIP is then identified as the remaining residual profit after the PIP has been removed, or:

Residual Profit - PIP = MIP

$909 - $409 = $500

Step 3b: Determine Taxable MIP

After arriving at MIP, the total MIP will be further segregated between the portion that is properly attributable to market externalities (which would be subject to reallocation under the proposal) and the portion properly attributable to the functions supporting the development, enhancement, maintenance, protection, and exploitation (DEMPE) of market intangibles (which would be subject to existing tax rules as modified by the BEPS measures). Our proposal would treat 20 percent of total MIP as taxable MIP under the new taxing right. We derived that 20 percent from economic analysis of the portion of profit appropriately allocable to market externalities.¹⁹

GlobalCo’s taxable MIP is calculated as 20 percent of the MIP:

20% * MIP = Taxable MIP

20% * $500 = $100

Step 4: Allocate Taxable MIP

Once taxable MIP is identified on a worldwide basis, the next step would be to identify the jurisdictions that are entitled to tax a portion of the group’s taxable MIP. That would be done by determining the net revenues that are treated as arising from sources in each jurisdiction (market-sourced net revenues) and, for each jurisdiction that has sufficient market-sourced net revenues to create nexus, allocating taxable MIP pro rata based on the ratio of that jurisdiction’s market-sourced net revenues to the group’s total net revenues.

Determine Market-Sourced Net Revenue

In general, market-sourced net revenue is the worldwide group’s net revenue sourced to a particular jurisdiction based on where the customer is located. Customer location would generally be determined by reference to the customer’s billing address, unless the taxpayer has reasons to know that products or services are for ultimate use and consumption in a jurisdiction other than that of the customer’s billing address. In that case, the group’s net revenue would be sourced to the jurisdiction of ultimate use and consumption.

That general sourcing rule is intended to create a bright-line, administrable rule that will reach an appropriate result in most cases. To ensure that the sourcing rules reach an appropriate result in cases that have been identified as presenting particular concerns, however, special rules would apply to subscription-based digital services, in which customers may be mobile, and digital advertising, in which the target of the advertisement and the paying customer may be in different jurisdictions. Subscription-based digital services would be sourced based on the primary location of the user when the services are consumed, and digital advertising would be sourced based on the location of the users when the revenue-generating activity occurs (for example, when the user clicks on an ad, or when an impression is delivered). Location would generally be determined based on the user’s IP address, or, if using that information is impractical, based on any other reliable data available to the taxpayer. Our view is that those specific fact patterns are at the heart of the concerns over taxing the digital economy. As a result, our proposal assumes that the incremental increase in complexity caused by collecting that information is necessary to reach a solution that addresses the tax challenges head-on. However, because those special sourcing rules impose an administrative burden, they would apply only if the total net revenue from digital advertising or subscriptions is greater than 25 percent of the worldwide group’s net revenue.

Determine Nexus

Our proposal would allocate taxable MIP only to jurisdictions where a taxpayer has nexus. We propose allowing countries to tax their pro rata share of taxable MIP if they have annual market-sourced net revenues exceeding €25 million. That threshold is intended to create a new nexus that is entirely independent from and does not modify permanent establishment standards for income tax or VAT purposes. It is intended to ensure that only companies with significant economic engagement with a jurisdiction are covered by the proposal and that only countries with significant market activities are entitled to impose tax.

Although we use €25 million as the threshold for our proposal, as have other recent proposals,²⁰ the specific threshold is intended as an example. While we believe that a de minimis threshold is necessary to reduce complexity, we recognize that countries may prefer a lower threshold, and that ultimately the choice of threshold will be a matter for negotiation. However, when revenues fall below the minimum threshold, the portion of worldwide taxable MIP that would be allocated to that country would not be subject to tax under the proposal. We recognize that countries might therefore prefer a lower threshold to avoid concerns that companies might earn large amounts of revenue without the associated profits being subject to market country taxation under the proposal.

For GlobalCo, we assume the economic nexus threshold is met in each country where it sells (countries A-F). As a result, GlobalCo would allocate its taxable MIP as shown in Table 3.

Table 3. Taxable MIP Allocation

Market Country	Sales by Market Country	Market/Worldwide Sales	Taxable MIP	Tax Rate	MIP Tax
A	1,000	27%	27	21%	5.7
B	500	13.5%	13.5	25%	3.4
C	400	10.8%	10.8	28%	3
D	700	18.9%	18.9	28%	5.3
E	200	5.4%	5.4	25%	1.4
F	900	24.3%	24.3	17%	4.1
Total	3,700	100%	100		22.9

Step 5: Where Does MIP Come From?

International tax and transfer pricing rules calculate taxable income entity by entity. It is therefore necessary to have a mechanism to identify which specific entities will be taxpayers of the new tax and will therefore surrender a portion of their taxable income to that new taxing right. There also needs to be a mechanism to ensure that the reallocated income is not taxed under both the existing system and the new taxing right.

Under our proposal, taxpayers would be group entities that are considered to earn residual profits (residual profit entities, or RPEs). An entity’s residual profit would equal its operating profit reduced by its deemed entity-level routine profit, which would be equal to the greater of 4 percent of the group net revenue or 15 percent of depreciable and amortizable assets (other than goodwill) multiplied by the ratio of the entity’s value-added costs to the total of the worldwide group’s value-added costs. That would guarantee that only entities that earn greater than a routine return (determined under a simplified approach intended to ensure administrability and maintain overall consistency) would be subject to the tax. Because taxable MIP will exist only when cumulative residual profits exceed cumulative residual losses on a worldwide group basis, there will always be at least one RPE in any group that has taxable MIP.

Each RPE will pay tax on its share of taxable MIP at the normal corporate rate in each country to which taxable MIP is allocated. The share of taxable MIP attributable to each RPE will be determined pro rata based on its residual profits relative to the residual profits of the group’s other RPEs. To avoid double taxation, each RPE would be entitled to a deduction for its share of taxable MIP against its taxable income under the conventional tax rules applicable in its country of residence (a surrender deduction). We considered whether the proposal should instead provide a credit against the tax that would otherwise be due in the RPE’s country of residence but concluded that a deduction system would be more equitable and substantially easier to administer.

Returning to GlobalCo, the RPEs would be identified as described below.

Step 5a: Deemed Routine Profit Percentage

The deemed routine profit percentage is reflected in the following formula as applied to GlobalCo:

(4% * Group Sales)/Total Group Costs = Deemed Routine Profit Percentage

(4% * $3,700)/($1,085 + $1,308 + $250) = 5.6%

Step 5b: Entity Routine and Residual Profit

Once the deemed routine profit percentage is identified, GlobalCo can then calculate the deemed entity routine profit for the relevant legal entities in its group. Residual profit by entity (and the associated percentage of group residual profit earned by each legal entity) can then also be determined (as illustrated in Table 4). Specifically, the routine profit by entity is determined by multiplying the deemed routine profit percentage by the entity’s value-added costs. The residual profit by entity is identified as the remainder. In the GlobalCo group, Company B Support Co is identified as having no residual profit and therefore is not considered an RPE.

Table 4. Residual Profit by Entity

Entity		% of Total Residual Profit
Country A Parent
Operating profit	600
Deemed entity level routine profit	(63.6)
Residual profit	536.4	58.9%
Country B Support Co
Operating profit	7
Deemed entity level routine profit	(8.8)
Residual profit	-	0%
Country E HQ
Operating profit	400
Deemed entity level routine profit	(61.6)
Residual profit	338.4	37.2%
Country F Reseller
Operating profit	50
Deemed entity level routine profit	(14)
Residual profit	36	4%
Total RPEs’ deemed residual profit	910.8
Note: Country F Reseller deemed entity level routine profit does not consider a markup on COGS as these are not value-added costs (as discussed above).

Table 5. RPE’s Taxable MIP by Country

Entity	% of RPE’s Residual Profit	Taxable MIP Surrender Deduction	Taxable MIP by Country
			Country A	Country B	Country C	Country D	Country E	Country F
			27.0%	13.5%	10.8%	18.9%	5.4%	24.3%
Country A Parent	58.9%	58.9	15.9	8.0	6.4	11.1	3.2	14.3
Country E HQ	37.2%	37.1	10.0	5.0	4.0	7.0	2.0	9.0
Country F Reseller	4.0%	4.0	1.1	0.5	0.4	0.7	0.2	1.0
		100.0	27.0	13.5	10.8	18.9	5.4	24.3
Note: Percentages identified above reflect sales by country as a percentage of worldwide group sales.

Table 6. RPE’s MIP Tax Liability by Country

Entity	% of RPE’s Residual Profit	MIP Tax Liability	Taxable MIP by Country
			Country A	Country B	Country C	Country D	Country E	Country F
			27.0%	13.5%	10.8%	18.9%	5.4%	24.3%
Country A Parent	58.9%	13.5	3.3	2.0	1.8	3.1	0.8	2.4
Country E HQ	37.2%	8.5	2.1	1.3	1.1	2.0	0.5	1.5
Country F Reseller	4.0%	0.9	0.2	0.1	0.1	0.2	0.1	0.2
		22.9	5.7	3.4	3.0	5.3	1.4	4.1
Note: Percentages identified above reflect sales by country as a percentage of worldwide group sales.

Step 5c: Determine Surrender Deduction

GlobalCo now allocates the taxable MIP and the MIP taxes to each of the identified RPEs based on their relative ratios of the residual profit (see tables 5 and 6).

Countries would agree that the taxable MIP would be allowed as a deduction against local legal entity taxable income (the surrender deduction). While the example addresses the application of all five steps to a hypothetical fact pattern, the proposal also has several features, discussed below, intended to address other common fact patterns or to otherwise simplify the proposal and make it easier to administer. As illustrated in the GlobalCo example, because all countries have market-sourced net revenues, all countries are allocated a portion of taxable MIP.

Other Important Features

Treatment of Losses

Our proposal recognizes that not every enterprise will be profitable and that losses must be addressed. It would subject a worldwide group to reallocation only after the group has earned taxable MIP on a cumulative basis. That would avoid separate allocation to, and tracking of, losses on a country-by-country basis under a parallel tax system. Instead, taxable MIP losses — that is, negative amounts of MIP resulting from applying steps 1-3 above — would be held in abeyance at the group level until the group earned cumulative taxable MIP. After that, for simplicity, the proposal would allocate profit based on the ratio of current-year market-sourced net revenue (discussed below).

For example, assume that a worldwide group’s taxable MIP reflects a residual loss of $100 each in years 1 and 2. In Year 3, it earns taxable MIP of $300. Under our proposal, no reallocation would occur in Year 1 or 2. In Year 3, the $300 would first be applied to eliminate the $200 in taxable MIP losses from years 1 and 2. The remaining $100 would then be allocated to market jurisdictions based on current-year market-sourced net revenues.

While it would be more accurate to require a blending of the market revenue allocations applicable in each of the years in the cumulative pool, we believe that the administrative difficulty of performing that kind of tracing exercise for the source of revenue is not justified by the incremental improvement in accuracy. However, if there were concerns that changes in revenue sources over time might lead to unfair results for taxpayers earning long-term losses, the OECD inclusive framework could consider applying a three-year averaging approach rather than an allocation based solely on revenues earned in the current year.

Finally, consistent with the idea that taxable MIP is a cumulative concept, preenactment losses should be considered in quantifying post-enactment cumulative taxable MIP.²¹

Impact of Transfer Pricing Adjustments

Step 5 above identifies RPEs that will act as taxpayers under the proposal and gives each RPE a deduction against its residence-country taxable income for its share of the taxable MIP of its worldwide group. Because the results of applying the proposal globally are then reconciled with the results under traditional entity-by-entity application of tax and transfer pricing rules, the impact of transfer pricing adjustments must be considered.

For example, suppose in years 1-4 of applying the proposal, a worldwide group’s subsidiary in Country X is not an RPE. In Year 5, X makes an upward transfer pricing adjustment of the subsidiary’s profits that covers the previous four years. If the subsidiary’s income were increased in each of the previous years for Step 5 purposes, it would become an RPE (and would reduce the share of “surrender deduction” borne by each other RPE in the group). Carrying the results of the transfer pricing adjustment back through each of the previous four years could cause major disruption. Treating the results as relating entirely to Year 5 could cause significant distortion. On the other hand, wholly ignoring transfer pricing adjustments that could affect who bears the burden of the new tax in previous years would create unfortunate incentives to make upward adjustments with retroactive effect. To avoid those incentives while not disturbing past years, our proposal would smooth the impact of the above adjustment by carrying the results forward to the following four years (years 5-8) for determining which entities are RPEs and what share of worldwide taxable MIP they are allocated.

Designated Agent

While each RPE would be a taxable MIP taxpayer, for administrative convenience each group would be entitled to designate a single agent (or multiple agents, for example, by region) to file and pay taxes on behalf of all group members. RPEs would be able to reimburse the agents for the taxes paid without additional tax consequences.

Worldwide Revenue Threshold

In addition to the economic nexus threshold for each jurisdiction described above, a taxpayer would not be subject to the proposal unless it had annual worldwide net revenues exceeding €750 million. While the proposal is intended to be simple and administrable, managing two parallel tax regimes will inevitably introduce complexity. Focusing solely on larger taxpayers will ensure that the tax is focused on companies most likely to have the resources necessary to manage that complexity. Focusing the proposal in that way will also simplify administration and enforcement.

Implementation and Timeline

Our view is that implementing any global consensus is likely to require a new multilateral agreement. While a full discussion is beyond the scope of this article, that kind of agreement should apply between countries in the absence of a preexisting tax treaty relationship. As a result, our view is that the OECD’s Multilateral Convention to Implement Tax Treaty Related Measures to Prevent BEPS is unlikely to be a suitable model for the implementation agreement.

The binding agreement to implement the global consensus should include rules for coordinating with existing tax treaties, including nexus and profit allocation rules and provisions to relieve double taxation. It should also include a commitment to effective dispute resolution, including mandatory binding arbitration. The ability to quickly and effectively resolve disputes will be essential to ensuring certainty and eliminating double taxation as a result of the new standard. While we have made every effort to make our proposal as simple and formulaic as possible, any new standard is likely to generate increased controversy. That risk is compounded by the fact that the proposal, like others under consideration, reflects reallocations of worldwide profits. As a result, adjustments by one country could affect all other countries, and not merely a single treaty partner. Traditional dispute resolution and arbitration mechanisms should therefore be adapted to the multiparty context.

As discussed above, the binding agreement should also include transition rules that coordinate the repeal or phaseout of existing unilateral measures, including DSTs, to ensure there is no overlap between the two proposals that could result in double taxation, as well as no ability to elect which tax will apply.

Conclusion

We think our proposal addresses the concerns over taxing the digital economy in a way that is well-suited to achieving a stable long-term consensus. It is formulaic — simplifying compliance without sacrificing reliability. Its scope is appropriately targeted to the MIP that is at the core of the digital tax debate and limited to the taxpayers most likely to have the resources necessary to manage any additional burden that it may create. When possible, it uses publicly available verified data maintained for nontax purposes to ensure confidence and administrability by both governments and taxpayers. The proposal is principle-based and underpinned by economic analysis. Simplifying measures are used only when doing so would not create arbitrary outcomes that undermine the goal of reaching principled results.

Our proposal is broadly applicable and is not ring-fenced to a single sector. While its application might vary depending on structures used, profit margins, R&D spending, or other features, those variations are neutral for specific business models. The proposal is thus able to reach principled, explainable results across different industries, profit levels, and tax profiles. As a result, it would be much more adaptable than a proposal that attempts to target specific business models to ring-fence their relevant characteristics.

While the proposal is formulaic and as simple as possible, it is still squarely targeted at the key concerns underlying the debate on the tax challenges of digitalization. In particular, it addresses the ability to concentrate residual profits attributable to market intangibles in low-tax jurisdictions by ensuring that the market jurisdiction is able to tax an appropriate portion of those returns. The revenue sourcing rules ensure that both remote sales to paying customers and remote targeting of advertising to nonpaying customers in a market jurisdiction will be appropriately addressed.

Because it is appropriately targeted, the proposal is intended to accomplish a reasonable reallocation of taxing profits that is focused on the specific tax challenges raising the greatest concerns. The shift of incremental taxing rights is aimed at the portion of residual profits from market intangibles that is reasonably tied to the market. The proposal also provides an appropriate way to address losses that gives credit for those losses without asking market countries to allocate losses against income that is currently taxable under the traditional tax system. Because the proposal’s focus is limited to the portion of profits that is properly considered attributable to market externalities, it should not result in a massive shift in taxing rights that could disrupt the ability to achieve consensus. Also, all countries, including those that may lose incremental revenues, should see improvements in simplicity and administrability as a result of achieving a stable consensus for an approach that eliminates the subjectivity and complexity that has characterized recent unilateral efforts to address tax challenges.

As for companies, increasing the share of their profits taxed under this proposal does not necessarily mean the company’s tax bill correspondingly increases. Instead, because the proposal allocates profits from jurisdictions where residual profits are earned to market jurisdictions, companies with less aggressive transfer pricing policies, as well as significant domestic business or headquarters in jurisdictions with high taxes relative to their primary market jurisdictions, might see decreased tax bills. In contrast, companies that have centralized DEMPE functions in a way that attracts large shares of global residual profits to jurisdictions with favorable tax rates, or that have more aggressive transfer pricing policies, might see increased overall tax liabilities.

Implementing any proposal that significantly changes source and residence allocations of taxing rights in ways that were not contemplated by the existing tax system and are not reflected in the tax treaty network will be challenging. There will be a need for a new instrument to memorialize the agreement and appropriately coordinate with tax treaties. That instrument will raise novel questions, and time will be required for its development, negotiation, ratification, and entry into force. After political agreement is reached on a consensus solution, transition rules must allow for a clear pathway to that solution without barring countries from taking appropriate measures for taxpayers in jurisdictions whose governments take too long to implement the solution. As a result, despite the urgency of the topic, addressing these tax challenges should be viewed as a multiphased effort. The goal of this work is not merely to patch a leak in the system; it is to create a consensus solution that will stand the test of time and prevent the need to return to the negotiating table. Uber believes its proposal can form the basis of that consensus.