As the mass proliferation of digital services taxes becomes increasingly certain, the United States may be wise to reassess its categorical opposition to them and rededicate its resources to developing multilateral DST standards.
DSTs are still relatively novel tax policy instruments that have yet to win consensus acceptance and, therefore, are not subject to the kind of uniform multilateral standards that drive convergence in transfer pricing.
There is thus no OECD-endorsed definition of a DST other than article 38 of the draft multilateral convention on pillar 1, which defines “DST or relevant similar measure” to broadly include any tax assessed based on market criteria, applicable solely or almost entirely to nonresidents, and treated as a tax other than an income tax under the jurisdiction’s domestic law or its bilateral income tax treaties.
However, many DSTs are modeled on the proposal outlined by the European Commission in a 2018 proposed EU directive (COM(2018) 148 final) that never gained unanimous support among member states.
The proposal would have imposed a 3% tax on the proceeds from offering advertising space targeted at users of a website or any other online application (a “digital interface”), providing access to a multi-sided digital interface that allows buyers to sell to customers or different users to interact with each other, or selling personal data collected from users of the interface.
The EU proposal and the national DSTs it inspired apply to revenue from ads that appear on internet search engines sites (Google) next to the search results, commissions charged by online marketplace platforms (Amazon.com, Apple’s App Store, Airbnb, and eBay) or ride-sharing applications (Uber), and revenue from ads targeted at users of social media platforms (Facebook and X, formerly known as Twitter).
They do not include fees charged by online payment processors (PayPal), revenue from subscription-based video (Netflix) and music (Spotify Premium) streaming services, or cloud services revenue (Amazon Web Services).
They also exclude online sales by traditional retailers through their own websites and any revenues earned by an entity that isn’t part of a multinational group with more than €750 million in annual global turnover.
As the European Commission explained in its proposal, the idea was to target value generated by local users’ interaction with each other or with the digital interface as opposed to value attributable to the digital interface itself. National legislation modeled on the commission’s proposal, including the DST recently enacted by New Zealand, retain this focus on data monetization and digital intermediation services.
Other DST measures define scope differently. Article 12B of the U.N. model tax convention allows the imposition of a tax on gross income from “automated digital services” that require minimal human involvement from the service provider and generally applies to the activities targeted by the EU proposal.
However, article 12B covers a broader range of activities than the EU proposal and EU-inspired national DST measures, including revenue from cloud services and subscription-based streaming services. India’s equalization levy, which applies to any nonresident’s supply of e-commerce services to Indian consumers if the value of its services exceeds a de minimis monetary threshold, is even broader.
The DST Menace
Whatever their specific form, DSTs have drawn the relentless opposition of U.S. government officials and legislators like no other anti-profit-shifting measure has. Officials from two different administrations and lawmakers from both parties have tenaciously opposed DSTs and the logic underlying their enactment since their inception.
The Office of the U.S. Trade Representative consider DSTs to be sufficiently discriminatory and burdensome for affected U.S. companies to justify the threat of draconian retaliatory tariffs. The unambiguous exclusion of DSTs from creditable foreign taxes was the purpose of a currently suspended update to the foreign tax credit regulations.
The preamble to the 2022 FTC regulations (T.D. 9959) aptly conveys the U.S. level of disdain for DSTs:
“Recently, many foreign jurisdictions have disregarded international taxing norms to claim additional tax revenue, resulting in the adoption of novel extraterritorial taxes that diverge in significant respects from U.S. tax rules and traditional norms of international taxing jurisdiction. These extraterritorial assertions of taxing authority often target digital services, where countries seeking additional revenue have chosen to abandon international norms to assert taxing rights over digital service providers.”
Although the IRS has offered taxpayers temporary relief from the 2022 FTC regulations by allowing them to apply former regulatory standards to determine an FTC’s creditability, even the relief notice (Notice 2023-55) states that DSTs aren’t creditable “because the base of the tax is gross receipts or gross income and does not consist solely of investment income that is not derived from a trade or business, or wage income.”
The singular intensity of U.S. opposition to DSTs has led to the acceptance of measures that would otherwise have held little appeal. Treasury officials’ unenthusiastic acquiescence to amount A of pillar 1, which reallocates an arbitrarily determined 20% of in-scope multinationals’ taxable profit to market jurisdictions, never would have been possible without the looming threat of mass DST proliferation.
The objections to DSTs raised by U.S. officials and lawmakers over the years may be overstated, but they aren’t baseless. Beyond the risks of triggering a trade war and the economic distortions inherent in any attempt to target a specific industry sector, DSTs have problems tied to basing the assessment on gross revenue instead of net income.
These include an acute level of double taxation risk: Taxpayers subject to DST by multiple jurisdictions on the same revenue generally do not have access to anything analogous to mutual agreement procedures and, unlike a tax on corporate profit, a U.S. tech company’s foreign DST payments aren’t creditable against its U.S. corporate tax liability.
Without a profit-based alternative along the lines of article 12B’s net income option, which allows taxpayers to be taxed at a 30% rate on the profit attributable to automated digital services, imposing the tax on turnover also raises the possibility that loss-making companies with no taxable profit to shift will nonetheless be subject to DST.
Defenders argue that overriding tax policy concerns outweigh these shortcomings, including tech companies’ aggressive tax planning practices, their ability to derive revenue from local consumers without creating a nexus for corporate tax purposes, and the value generated by local user engagement and user data. But none of these arguments really holds up as a justification for DSTs as they are typically designed.
DSTs are far too narrow in scope to be justified by aggressive intellectual property planning practices. It’s true that tech giants like Google, Amazon, Facebook, and Apple heavily rely on intangible assets that have no inherent physical location and that intragroup transfers of unique and valuable IP can be powerful profit-shifting tools.
However, as cases like Medtronic Inc. v. Commissioner, T.C. Memo. 2022-84, and the European Commission’s state aid investigations concerning IKEA, Nike, and Starbucks clearly demonstrate, multinationals in more traditional sectors can be equally adept at shifting profit through intangible transfers.
Most DSTs are also too narrowly targeted to be justified as permanent establishment antiavoidance measures, which the OECD’s draft multilateral convention on pillar 1 expressly excludes from the scope of “DSTs or relevant similar measures.”
It’s true that in-scope multinationals that provide online marketplace platforms, social media services, and internet search engines can derive revenue from consumers in a market without creating a local PE.
But the same goes for the online payment processors and subscription-based streaming services that are exempt under most national DST regimes. Simultaneously, they’re overbroad: DST legislation based on the 2018 EU proposal and its progeny offer no exemption for multinationals that have a local PE.
For user participation and the new business models that exploit it, it’s never been clear what unique properties justify a derogation from the treatment of other multinationals through a targeted turnover tax.
There is a questionable claim (made by the European Commission in its 2018 DST proposal and recently echoed in the commentary on New Zealand’s DST legislation) that a high level of local user engagement means that less value should be attributed to the underlying digital platform. A digital platform cannot foster local user participation and engagement without offering features that entice local users, and the monetization of user data through targeted advertising is only profitable through sophisticated algorithms.
Put another way, the premise that the technology underlying Netflix’s streaming services generates revenue in a way and to an extent that the technology behind Google’s search engine does not is doubtful on its face.
If the issue is specific to the monetization of user data, then most DSTs are both too broad and too narrow. The unique properties of user data don’t explain why sales commissions charged by online marketplace platform operators should be lumped in and taxed in the same way as revenue earned by directly monetizing personal data, for example by the sale of user-targeted digital ads.
And if the extraction of local user data had some fundamental tax policy significance, it would be hard to justify excluding other tech multinationals, including most developers of Chromium-based web browsers other than Google’s Chrome, which derive the vast majority of their revenue from monetized user data, because of a global revenue threshold.
But recognizing and drawing attention to DSTs’ many flaws is one thing; making the total eradication of DSTs enacted by other tax-sovereign jurisdictions a top national policy priority is another. It’s important to distinguish the interests of the United States from the interests of a handful of its best-known multinationals.
Although it’s perfectly reasonable for U.S. government officials and lawmakers to seek to deter or stop foreign discrimination against domestic business interests, it’s unclear that any compelling national interest justifies a policy of relentless confrontation backed by threats of self-destructive, retaliatory trade policies.
For one thing, there’s something odd about the U.S. government, which itself has been or remains entangled with top tech giants in high-stakes international tax disputes, going to such lengths to shield a particularly quarrelsome group of domestic taxpayers from foreign tax liability.
But even if there’s nothing wrong with Treasury’s zealous advocacy of the foreign tax policy preferences of multinationals that aggressively interpret Treasury’s section 482 regulations, the alignment between U.S. interests and U.S.-based multinationals’ interests isn’t as obvious as is often assumed.
The U.S. fisc’s level of exposure to DSTs lies somewhere between minimal and nonexistent: Because DSTs aren’t assessed based on net income, they cannot be credited against U.S. corporate tax liability. The only way foreign DST liability could affect domestic tax revenue would be through an attenuated and likely minor reduction in a U.S. tech multinational’s pretax income.
If U.S. taxpayers aren’t footing the foreign DST bills of domestic multinationals, U.S. opposition to DSTs could be justified by the general economic interests of U.S. citizens and residents. But many investors in Google, Amazon, Facebook, and Apple aren’t U.S. citizens, and those who are will only be harmed to the extent that the costs of DSTs can’t be credited against other taxes or passed on to consumers or other businesses.
There is no definitive empirical verdict on the incidence of DSTs, but it seems likely that the economic burden will generally be shared with foreign consumers, foreign businesses, or others whose well-being should be of no particular concern to the U.S. government.
Anecdotal evidence supports this assessment, including Google’s proclamation that it would pass all of its United Kingdom and Austrian (and most of Turkey’s) DST costs to local advertisers and Amazon’s announcement of similar DST-linked fee increases. Apple later did essentially the same thing in response to India’s equalization levy.
These may have been, at least in part, symbolic gestures or handy justifications for price increases that would have taken place anyway, and not all U.S. multinationals subject to foreign DSTs responded in a similar manner.
Many have predicted that the taxpayers specifically targeted by DSTs would not bear the full burden of DSTs, including Deloitte’s French member firm Taj. A 2019 study by Taj estimated that tech giants would bear only 4% of France’s DST burden, while local consumers would bear 57% and businesses that use digital platforms would bear the remaining 39%.
These were only the ex ante predictions from an organization that arguably had an interest in depicting DSTs as self-defeating, but they generally align with research on other consumption taxes. It’s fair to tentatively assume that a material share of U.S. tech multinationals’ DST costs is shifted to foreign consumers and businesses and, in the absence of any firm evidence to the contrary, a U.S. policy of anti-DST fundamentalism is hard to justify.
To the extent that the burden of a foreign country’s DST is largely borne by the foreign country itself, alleviating that burden probably shouldn’t be an urgent U.S. policy priority. In contrast, U.S. consumers would almost certainly bear the brunt of any tariffs imposed by the United States in retaliation for DSTs.
There is also a broad U.S. interest in a principled and reasonable multilateral tax system, but the existing system is far from perfect; the United States isn’t particularly well situated to serve as the enforcer of multilateralism and principle in international tax affairs. The United States has tolerated flawed unilateral tax measures in other contexts and, as recently as 2017, enacted some of its own.
It’s also important to assess the shortcomings of DSTs, considering the flaws of the plausible alternatives. For DSTs to warrant the level of concentrated U.S. opposition they currently receive, they should be materially and unambiguously worse than anything else on the table.
But the concession that U.S. Treasury officials were willing to make in exchange for the rollback of DSTs, amount A, is no model of perfect tax policy either.
Under established transfer pricing principles, allocating an arbitrarily selected 20% of profit in excess of an equally arbitrarily selected 10 percent routine return to market jurisdictions would be a strong candidate for the world’s dumbest application of the residual profit-split method. Amount A may be free from discriminatory tendencies and the problems specific to turnover taxes, but its mechanics have no firmer basis in established tax policy principles than DSTs.
Even if the economic burden of DSTs did fall almost entirely on U.S. interests and DSTs were significantly and unambiguously worse than all competing proposals, it doesn’t make sense to invest finite government resources in a campaign to prevent something that the United States is clearly powerless to stop.
Neither the United States’ participation in the two-pillar international tax reform project nor its threats of self-destructive trade policy have succeeded in deterring countries from enacting DST legislation. DSTs appear to be here to stay, and there’s little if anything the United States can do about it.
If the United States’ global anti-DST campaign is doomed, U.S. officials’ zeal for sound tax policy would be better dedicated to an effort to establish multilateral standards that round some of the sharper edges of DSTs.
The United States could contribute to an OECD-led effort to develop model DST legislation, something that regional multilateral organizations like the African Tax Administration Forum have already done. Making DSTs creditable against local corporate tax liability, exempting taxpayers with a local PE, offering a net income option consistent with article 12B, and providing MAP-like double taxation dispute resolution mechanisms in accordance with OECD-endorsed model rules would go a long way toward easing the burdens and limiting the distortions of DSTs.
The development of model DST rules would also provide an opportunity to reassess the scope of DSTs, including some of the more dubious exclusions associated with EU-inspired DSTs.
Improving, rather than eliminating, DSTs in this way may not be U.S. policymakers’ first choice. But now that DSTs have proliferated despite the United States’ determined effort to stop them, it may be the only rational choice left.