In 2017, Spain was one of the signatories to the Political Statement addressed to the European Commission, which recommended that an “equalization tax” designed to achieve “fair taxation” of the “digital giants” should be explored. In the absence of a consensus with respect to the shape of long-term measures, the signatory countries launched a number of short-term, interim measures, putting them at the top of the EU tax agenda.

It should come as no surprise, then, that Spain may soon become one of the first European countries to approve a digital service tax (DST), the features of which substantially replicate the content of the Council Directive Proposal on DST. (Proposal for a Council Directive on a common system for a digital services tax on revenues resulting from certain digital services 2018/0073 (CNS) (Mar. 21, 2018).)

The Spanish government expects to collect around 1,200 million euros from this new tax and has already approved a draft bill that has to be passed by Congress. (Spain recently underwent presidential elections where the former president was the more voted candidate. However, he still needs support of other political parties for being appointed president by the parliament). There are a number of factors driving the push for Spain to be one of the first countries to approve a unilateral measure of this type: the length of time that has passed without alternative practical solutions having been found; social pressure; the concept of “fairness in taxation” and the “sustainability” of the tax system; and, one might add, compromises made by the government with other political parties. In other words, there is nothing new under the sun.


In line with the Council Directive Proposal, the Spanish DST is defined as an indirect tax, compatible with value added tax (VAT). As such, it is outside the scope of the tax treaties that Spain has concluded for the avoidance of double taxation. As a consequence, the DST will likely generate double taxation issues that Spain’s tax treaties cannot resolve.

The scope of application of the DST does not encompass all digital services—only those in which the user’s role in value creation is considered essential. According to the bill, the targeted services cannot exist in their current form without implicating the user. Even though Action 1 of the OECD’s BEPS Project acknowledged that the digital economy could not be ring-fenced from “the economy itself,” the DST does exactly that and targets three specific types of services:

  • The placing of advertisements on a digital interface—“online advertising;”
  • The making available of multi-sided digital interfaces—“online intermediation;” and
  • The transferring of users’ data—“data transfer.”

The digital interface is defined, for the purpose of the DST, as any software, including a website or part of a website, and applications, including mobile applications, accessible to users that enables digital communication. The Spanish DST is expressly intended not to apply to activities such as financial intermediation and crowdfunding.

The first group of services (online advertising) would include those provided by Facebook, YouTube, and search engines such as Google. The second group of services (online intermediation) would capture those provided by, for example, Airbnb, Uber, and eBay. However, the scope of the DST does not extend to platforms providing audiovisual content (Netflix) and music (Spotify), or to e-commerce platforms (Amazon and Alibaba). The DST, thus, only affects half of the tech giants popularly known as the “FANG” (Facebook, Amazon, Netflix and Google).

The subjects of the DST are both resident and nonresident companies (both inside and outside the European Union) whose revenues, when computed on a consolidated basis, exceed the following thresholds:

  • Worldwide revenues: 750 million euros (a “strong market position”); and
  • Taxable revenues in Spain: 3 million euros (a “significant digital footprint” in Spain).

The 750 million euro threshold applies for purposes of country-by-country (CbC) reporting obligations, from which the tax authorities expect to obtain information to help them identify companies that should be subject to the DST. The bill indicates that Spanish taxable revenues are to be calculated in accordance with the rules for determining the DST taxable base. The Council Directive proposal sets the EU digital footprint for inclusion in the scope of the DST at 50 million euros.


Nexus with Spain for the purpose of allocating DST taxing rights to Spain is based on the location or “geolocalization” of the user’s device:

  • Online advertising would be subject to the Spanish DST when the user’s device on which the advertising appears is located in Spain;
  • Online intermediation would be regarded as occurring in Spain and therefore subject to the Spanish DST when: either the underlying transaction takes place through a digital interface using a device located in Spain; or the user’s account is opened with a device located in Spain; and
  • Data transfer would be subject to the Spanish DST when the data transferred was generated by a user through a digital interface using a device located in Spain.

In the case of online intermediation, it is irrelevant from a nexus point of view where payments are made and where the underlying transaction takes place. For example, if two individuals agree, through eBay, on the sale of a picture, it is not relevant where the sale of the picture occurs, but only whether the transaction was conducted using a device located in Spain. Also, e.g., if an individual is on vacation in Spain and opens a Google account while the individual’s device is in Spain, the device will be regarded as a Spanish device for DST purposes, even though the user is only visiting Spain briefly.

The geolocalization of the user’s device in Spain is the key element for allocating DST taxing rights to Spain, once the digital footprint threshold is reached. Under the bill, for purposes of determining whether a device is located in Spain, reference will be made to internet protocol (IP) addresses, unless other types of proof are adduced (for example, in the form of geolocalization instruments). “Internet Protocol Address” is defined as the codes assigned to networked devices to facilitate their communication through the internet.

In light of the above, it is not hard to imagine an increase in new types of tax fraud involving redirecting or disguising the geolocalization of devices. It is also worth asking whether the proposed nexus is even compliant with the General Data Protection Regulation (GDPR), .that entered into force last May 25, 2018.

In line with the Council Directive Proposal, the DST tax rate is set at 3% of gross revenues (net of VAT) arising from the supply of the digital services concerned. The fact that the tax is computed on gross profits has been the subject of sharp criticism on the grounds that it benefits companies with larger margins and harms those that have lower margins or are even in a loss position. In addition, this gross basis computation is seen by some as an indication that the DST is not the indirect tax it appears to be.

Under the “sunset clause” in the Council Directive Proposal, the DST is designed to be a temporary, short-term measure that is doomed to disappear once the new permanent establishment (PE) concept based on a “significant digital presence” (SDP) under the Directive proposal becomes a reality (or a new global solution to the nexus problem is found). There is no doubt that the new PE concept is conceived of as operating within the framework of income tax treaties and, more generally, direct taxes on income. This fact raises even more questions as to the nature of the DST, specifically, as to the compatibility of the notion of a “temporary,” “indirect” tax with its replacement by the new direct tax “virtual PE” concept.

In this context, it is worth highlighting the fact that, since the sunset clause was eliminated from the final version of the Spanish bill, technically, the new PE definition and the DST could co-exist, as there is no express provision to the effect that the DST will be repealed once the new PE definition becomes effective. Thus, besides the controversy surrounding its indirect nature, the temporary nature of the DST is also uncertain.

As regards the DST base, the allocation keys would depend on the targeted digital service:

In the case of online advertising, the taxable base would be computed by reference to the proportional number of times the advertising is shown on Spanish located devices to the number of times the advertising appears on all devices (apparently irrespective of where they are located);

  • In the case of online intermediation, the taxable base would be computed by reference to the proportion of Spanish users to the total number of users and the proportion of income resulting from users’ accounts opened in Spain to the income from the total number of accounts opened; and
  • In the case of data transfer, the taxable base would be computed by reference to the proportion of Spanish users generating the transmitted data to total users.

Although the bill does not expressly indicate this, there seems to be a consensus that the DST will be considered a deductible expense for corporate income tax purposes, regardless of whether the DST concerned is a Spanish domestic DST.

In summary, the determination of where and what is to be taxed revolves around the users and where they are located.


Where the old paradigm in international taxation was the arm’s-length principle, the new paradigm is the alignment of taxation with “value creation.” (Action 1 and Actions 8-10 of the BEPS Project).

Since it is easy to achieve consensus on the proposition that taxing powers should be granted to the country where thevalue is created, the concept of “value creation” makes its appearance in nearly all discussions and analyses relating to the taxation of the digital economy.

However, discrepancies begin to emerge when attempts are made to define value creation and to determine where it is generated. Even though value creation seems to be the key concept for revolutionizing the rules of international taxation, which are deemed ill-suited to the digital world, none of the most relevant documents released by the OECD or the European Union on this matter provide a single definition of value creation.

The reason for this deficiency may be that it is not economically possible to determine accurately where value is created, though, generally, a business may be said to create value when its revenues exceed its costs. (Andersson, United States/European Union/International—Should We Use Value Creation or Destination as a Basis for Taxing Digital Businesses?, 72 IBFD Bulletin for Int’l Tax’n 12 (2018); Marcel and Spengel, International Taxation in the Digital Economy: Challenge Accepted?, 9 World Tax J. 1 (2017).)

There are those who hold the view that, in the case of the new business models where there is often no direct charge on the user and where the network effects are undeniable, value creation should be assigned to the demand side, that is, to the market jurisdiction. This view thus entails the shifting of taxation from the residence country to the country where the users are located and establishes the grounds for the existence and particular configuration of the DST.

The Spanish DST bill states that Spain is adopting this unilateral measure to allow it “to exercise its legitimate taxing rights as the source of the data and user contributions that create value for the enterprises.” In other words, Spain is claiming taxing powers over the profits generated by the “tech giants” that are the target of the DST, because the data and contributions of users located in Spain are “creating value” for the targeted taxpayers.

In contrast to the SDP Directive, where the targeted services are broader and the user participation may be marginal (at least in some cases), the DST rules focus on services that are purported to have a close connection with user value creation. (Nieminen, Martin, The Scope of the Commission’s Digital Tax Proposals, IBFD Bulletin for Int’l Tax’n 11 (2018)) Users come accompanied by data. Consequently, data is always present in the value creation process (at least in its early stages).

A headline in The Economist stating that “The world’s most valuable resource is no longer oil but data” inevitably raises the question of how much data is worth. An online article published in 2012 stated that the value of a person’s data was somewhere between US $0.05 and US $1,200. It is beyond dispute that tech companies like Facebook and YouTube profit from the data and content provided by their users, and even from the mere presence of the users connected to the platform (the network effect).

However, is one to infer from this that these companies’ value is created by the users and where the users are? Conversely, should one take into account the value destroyed by the users? (Bal, Aleksandra, (Mis)guided by the Value Creation Principle—Can New Concepts Solve Old Problems?, 72 IBFD Bulletin for Int’l Tax’n 11 (2018).) For example, in determining the taxable base for DST purposes, is it appropriate for Spanish devices used to post statements that damage a company’s reputation to be included? Even though it would appear illogical, every device geolocalized in Spain is taken into account, irrespective of the value the user creates or destroys with it: the DST counts devices, not users.

In various ways, consumers have always provided data that is potentially useful to a business. The key question is whether value has to be attributed to such data: the answer to that question depends partly on the extent to which such data is monetized. Referring to traditional business models, Becker astutely points out that “one could argue that hanging around at a bar contributes to the lively atmosphere which attracts other pub crawlers.” (Becker, Johannes, EU Digital Services Tax: A Populist and Flawed Proposal, Kluwer Int’l Tax Blog (Mar. 16, 2018).) The author would add: what is Facebook if not people (virtually) hanging around and contributing to a lively (online) atmosphere that attracts other users?

Another point to consider in this analysis is that raw data has little or no value and is extremely cheap to acquire. It is only useful to companies that have the ability to collect it, store it, classify it and address it with the appropriate algorithm.

If you were asked to pay 1 euro per month to prevent your social network from storing and using your data, would you be willing to pay it? On the other hand, if your social network offered you 1 euro per month to use your data, would you take it? These questions demonstrate that the value of data for the user depends largely on the user’s perception of the value of privacy and how the user’s profile data affects it. A Spanish online newspaper recently published a story relating how a Spanish bank charged clients that were unwilling to release their data a 5 euro (penalty) fee.

The DST is based on the assumption that each piece of data provided by the user has the same value and the same quality. Thus, for purposes of calculating the Spanish piece of a tech company’s “profit pie,” the DST takes the Spanish devices generating the transferred data as a proportion of all other devices generating the transferred data wherever located. For purposes of the DST, data generated in Spain (or in Europe generally) where the GDPR limits the possibilities of data collection, storage, and transfer, is worth the same as data generated, for instance, by Chinese devices where these limitations do not apply.

This leads, in turn, to the question of whether the true value creation lies in users giving their data “for free” or in the investment in advanced computer programming, software, and skilled personnel (human capital) that create the appropriate algorithms. in the author agrees with the opinion of Van Horzen and Van Esdonk that data only has value once it is in the hands of a party that can do something with it or can process it. (Van Horzen, Fred and Van Esdonk, Andy, Proposed 3% Digital Service Tax, IBFD Int’l Transfer Pricing J. 4 (2018), at 269.) The author shares this view.

It is these algorithms that reinforce the network effects and allow the raw data to be combined in sophisticated ways that create value for decision-making processes, tailored-made targeted advertising and, ultimately, allow the raw data to be monetized. If one accepts this view, value is created not in the market jurisdiction, but where the computer programming, software and the skilled personnel are located, which may be determined using the DEMPE (Development, Enhancement, Maintenance, Protection, or Exploitation) approach.

In the author’s view, while the DST is targeted at taxing services where the value creation relies heavily on the user, it fails to provide a clear definition of what value creation is. The device geolocalization formula can lead to unjustified distortions and illogical and even unfair outcomes, depending on the business models concerned and whether the data and the user’s participation are being monetized.

It may be that a more fundamental approach is needed based on a better understanding of the newly emerging and constantly evolving business models, which are driven by artificial intelligence, deep learning and even block-chain technologies, if we are to grasp clearly where the value lies and to what jurisdictions it is “fair” to allocate taxing rights. (CFE fiscal committee. Opinion Statement submitted to the European institutions. FC 1/2018 on the European Commission Proposal of March 21, 2018, for a Council Directive on the Common System of a Digital Service Tax on Revenues Resulting from the Provision of Certain Digital Services -May 201-).) The main challenge is to understand how data is monetized and what functions within a multinational enterprise create this value.


A glance at the Spanish DST allocation keys will make immediately obvious the enormity of the task the tax authorities will face in attempting to verify the accuracy of the figures provided for purposes of the calculating the DST taxable base. This is probably why the Spanish bill provides for the creation of a special register for all companies subject to the DST. The bill also provides that these companies will be obligated to file periodic returns of information requested by the authorities relating to the computation of the DST and to keep registers that will be subject to regulation. Non-EU resident taxpayers will have to appoint a representative for the purpose of complying with the obligations laid down in the bill.

To encourage taxpayers to comply and cooperate in fulfilling these obligations, the DST system features a rigorous penalty regime: penalties can be as high as 0.5% of the turnover of a company that attempts to disguise the geolocalization of the relevant devices.

Like VAT, the DST accrues on a transaction-by-transaction basis (or at the time when there is an advance payment for the provision of a digital service that is within its scope), although the corresponding tax returns will be filed on a quarterly basis.


The DST is an extremely complex, “sub-optimal” and purportedly temporary measure that creates a degree of legal uncertainty for companies potentially affected by it. It seems to be a politically acceptable patch for what is probably the most important tax challenge of the 21st century: taxing the new scale-without-mass business models for which the existing rules have proven to be obsolete.

The classic dispute between the OECD (residence-favoring) countries and the UN (source-favoring) countries seems to be over in the world of the digital economy, with many of the countries in the first group claiming taxing rights with respect to the profits made by digital tech giants solely on the basis of where the users are located (a typical source-country taxation approach). For example, it is quite understandable that China, India, and Brazil should be pushing hard for this approach. On the other hand, the European Union, which has a much smaller and aging population, seems to be willing to lead the way in this direction.

More prosaically, implementing, computing, collecting, and monitoring the DST promises to be extremely challenging and complex for the tax authorities, unless taxpayers and the groups to which they belong give their full cooperation by providing the information needed to calculate the relevant amounts.

In the author’s view, by the time the tax authorities figure out the exact parameters of the DST (assuming they will do so at some point), it is possible that the scope of the DST may already have become obsolete and some “re-ring fencing” of the targeted digital services and a redefinition of the nexus rules will be needed. Additionally, when this article was being written, the OECD took a step forward and released a Public Consultation Document with a view to reviewing, among other things, the “user participation” proposal for revising the profit allocation and nexus rules for the digital economy. The conclusions reached could lead to a review of the proposed interim measures.

Before further steps are taken, maybe somebody should explain to the DST supporters the implications of Moore’s law. (Moore’s Law is the observation made by Intel co-founder Gordon Moore that the number of transistors on a chip doubles every two years while the costs are halved (1965). It is used to illustrate the exponential growth of information technology development)).

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

Guadalupe Díaz-Súnico is an international tax partner at Lener in Barcelona, Spain.