ITIF Search

Digital Services Taxes: A Bad Idea Whose Time Should Never Come

May 13, 2019

Arguments used to justify digital services taxes on large Internet companies are flawed. The United States should counter these proposals and work to restore international consensus around tax treatment for multinational companies.


Countries such as the United Kingdom, France, Spain, and Italy are considering ways to tax corporate revenues of multinational Internet companies through digital services taxes, which would tax a small percent of the companies’ domestic revenues.
Proponents of digital services taxes attempt to justify this tax grab by claiming users are creating value and therefore that value should be taxed where users reside. In fact, users do not create value; companies do. Users consume.
Further, taxing profits based on where users reside would violate longstanding international agreements by taxing income more than once and imposing an ad valorem tax that primarily targets imports.
The international community should work together in a multilateral forum to reject digital services taxes. In the absence of such a process, the United States should prepare a trade complaint for the World Trade Organization.


The Current International System

The Nature of Digital Services Taxes

The Impact of Digitalization

The Flawed Rationale for a Digital Services Tax

Digital Services Taxes Introduce Market-Disrupting Distortions

The Base-Erosion and Profit-Shifting Process

The Breakdown of Multilateral Consensus

Why Not a VAT?

The U.S. Response Thus Far




Over a dozen countries have recently begun to enact new digital services taxes (DSTs), which would generate 2 to 3 percent of the countries’ domestic revenues from a narrow group of large Internet companies. Supporters justify this unprecedented change in corporate tax law by arguing that most of these companies’ value is created by users, and therefore the profits on that value should be taxed in the nations wherein the users reside. This would violate the spirit behind long-standing international agreements, because users do not create value in any significant way. DSTs should be seen for what they really are: ill-disguised, convoluted, and capricious attempts to grab tax revenues from a narrow set of large, highly profitable, and mostly American companies. The United States should strongly resist DSTs.

Digital services taxes should be seen for what they really are: ill-disguised, convoluted, and capricious attempts to grab tax revenue from a narrow set of large, highly profitable, and mostly American companies.

Over the last century, the international community has developed a broad body of law to govern the taxation of multinational enterprises that operate in two or more countries. Much of this effort concerns when countries can impose their corporate income taxes on foreign companies that do business in those countries. Two of the most important principles underlying this regime are income should be taxed wherever the value of a good or service is created, and foreign companies do not become subject to a country’s corporate income tax (CIT) until after they have created a permanent establishment (PE) there.

Lately there has been growing concern over the broad digitalization of the economy, combined with other factors, undermining the foundations of the international system. Some of these concerns focus on the increased ability of firms to restructure their internal contracts to move income out of high-tax and into lower-tax jurisdictions. Firms have also found it easier to take advantage of the rules in unintended ways. There is a wide consensus that these efforts are undermining the global tax system and therefore need to be addressed. Over the last six years, the international community has launched, completed, and enacted a base-erosion and profit-shifting (BEPS) effort. The BEPS process gives countries a prominent platform for raising issues that clearly violate the spirit of the international system. Most experts believe the implemented BEPS agreements will address the most important tax avoidance problems.[1]

However, some countries think BEPS is not enough. Now that increased digitalization of the economy has enabled more goods and services to be traded across borders—the same way manufacturing goods long have been traded—some nations are concerned about losing domestic output to, and thereby business tax income from, foreign companies using digital business models. Rather than focusing on becoming more internationally competitive, many of these nations are seeking ways to tax foreign-company income, relying on the flawed reasoning that users create value so the countries are justified in taxing the revenues those companies earn from their residents.

This dispute is eroding the international consensus around existing rules. Some nations want to address the problem by using the BEPS process to enact further changes specifically targeting “highly digitalized” companies. These proposed changes would ease the current requirement that companies have a PE in a country before they are subjected to corporate income tax and divert more corporate tax revenues to the countries in which the consumers live. Most nations, however, believe BEPS adequately addresses the most important issues, and that further changes should apply to all industries similarly without ring-fencing the digital economy by creating a separate DST. Countries have therefore continued their discussions.

It would be a mistake to ring-fence the most digital companies with a DST. The main argument for doing so—that users in these industries create significant value—has no merit.

Clearly the second group has the better argument. It would be a mistake to ring-fence the most digital companies with a DST. The main justification for doing so—that users in these industries create significant value—has no merit. Digital companies create the lion’s share of the value, which is why they invest so much in improving software code and in research and development (R&D). DSTs will not only lead to significant economic distortions, they will create more complexity in a system already overburdened with it.

Because the chance of negotiating the changes these countries favor is so low—in large part because their core argument and rationale are flawed—several countries are moving ahead on their own with plans to implement a DST. Some view unilateral action as a bludgeon with which to force other nations to agree to proposed changes in BEPS. Their frustration stems from the fact that digitalization has made it possible for companies to export more goods and services into a country without subjecting themselves to that country’s CIT. As such, they want to find ways to tax these foreign corporate profits. However, in order to get around the international tax agreements that prevent countries from taxing foreign companies unless they have a nexus in those nations and create value there, these laws tax revenues rather than profits. Those nations argue that revenue taxes are not covered by these agreements. Yet they also go to great lengths to argue that these taxes are justified under the current system in which corporate income is taxed wherever value is created, by claiming digital users create most of the value. They also target a very limited set of digital business models, arguing that those businesses are earning profits from users in, but not paying enough corporate taxes to, the nations wherein the users reside. But this argument cannot be valid, for many companies exporting goods (e.g., cars, food, washing machines, etc.) to other nations make profits from consumers but do not pay income taxes to the country they export to unless they have a PE there.

So, the key argument these nations make to justify the DST is, unlike in other industries, users of these business models create a large part of the value driving corporate profits, and therefore the nations are justified in taxing the revenue. Users supposedly create value by producing content for the platform or generating data the companies sell or use to generate ad revenue. Because many of these companies lack a PE in those countries, they pay taxes on that income in their home country, but not in the nations where the consumers are. DST supporters argue that they are adhering to the principle that value should be taxed wherever it is created. But their interpretation of value creation is illogical and ignores almost 100 years of history. Implementation of a DST therefore represents a unilateral departure from long-established international rules.

As noted, at the crux of their argument is the notion users in the destination countries create value, and therefore those nations are justified in taxing revenue generated from that value. However, the definitions of user-created value are faulty. Even in the most digital markets, the majority of the value is created by companies, not users, which is why major digital companies employ hundreds of thousands of workers—just like traditional companies—to create value. The U.S. Department of Commerce reported data processing, Internet publishing, and other information services companies created over $300 billion in value added in the United States in 2018—significantly more than the motor vehicle industry, for example.[2]

In addition, the proposals make artificial and inaccurate distinctions between ad-supported business models (wherein, they argue, local value is created) and subscription-supported business models (wherein, they argue, value is not created). Moreover, the activities asserted to create value are increasingly being used in many traditional industries, and will increase as more and more business models incorporate digital components. Just as problematic, the proposals make artificial distinctions based on industry, size, corporate business model, and nationality without any clear rationale other than to narrow the set of target companies. These distortions are not backed by sound policy, other than targeting the easiest sources of revenue: large foreign firms. The disparate treatment of foreign firms also likely violates international trade agreements because it essentially involves imposing an ad valorem tax that almost exclusively targets imports.

The narrow structure of the DST combined with the faulty justification of user-created value gives it the appearance of a naked attempt to unilaterally raise revenue from large foreign firms, most of which are based in the United States. Unlike in the BEPS process, its supporters generally do not argue that companies are abusing the rules or their overall tax burden is too low. The argument by supporters of DSTs is the existing rules do not divert enough taxable profits to them. Rather than focus on how to build a more competitive economy that exports more, they seek to tax foreign profits by unilaterally undermining long-standing rules and agreements.

So far, the U.S. response has been mild. Both Congress and the administration have voiced concerns, but much of the latter’s attention seems to be focused on trying to get multilateral agreement on the broader implications of digitalization. This needs to change. As DST efforts go from debate to actual implementation, the Trump administration needs to forcefully convey its opposition and prepare a trade complaint for the World Trade Organization.

The Current International System

Over the last century, the international community has developed a broad body of law to govern when a country may tax the income of a foreign company that sells goods or services within its borders. For a company operating in many countries, the key question is what proportion of its total profits may any one country tax? The answer lies in a collection of highly detailed multilateral agreements and literally thousands of bilateral tax treaties. The latter attempt to merge the application of each signatory’s corporate tax law to a firm that does business in both countries.

Unlike in the BEPS process, DST supporters generally do not argue that companies are abusing the rules or that their overall tax burden is too low. Their argument is that the existing rules do not divert enough taxable profits to them.

The system is built on several main principles.[3] The first is all income should be subject to tax, but only once. The division of total profits among different countries is driven by the principle of value being taxed wherever it is created.[4] If 10 percent of the total value of a car is produced in France, France should get to tax the profits associated with the activity performed there—presumably 10 percent of the total. France gets to choose the tax rate.[5] The determination of how much value is created in a country can often be made by market prices. But in the case of internal transfers within a company, market prices may not always exist. In these cases, a large body of guidance tries to ensure prices reflect the arms-length standard: the price two independent firms would agree to if they were both pursuing their own interests.

Another principle is when dividing taxing rights between countries, no value is attached to the mere sale of a product. If a car is produced in Germany and then shipped over the border for sale to a French citizen, the German carmaker does not suddenly become subject to the French corporate income tax. All of its corporate profits from French sales will be subject to the German corporate income tax. France can, however, levy a sales or value added tax (VAT) on the transaction. To the extent permitted by international agreements, France might also apply a tariff to the foreign product. Similarly, if a Canadian citizen travels to the United States to buy a computer from a U.S. company, the company will not have to pay Canadian CIT on the profits from the sale. The same result holds if the purchaser stays in Canada and orders the computer over the phone or Internet. The result is not changed if there is only a minimal domestic presence.

The final guiding principle of the corporate tax system is a company must have a PE in a country before it becomes subject to the CIT. This rule partly ensures administrative costs are proportionate to the revenue raised. Without the PE rule, a country could still only tax the proportion of profits associated with the value created within its borders. Because the mere sale is not considered to add value, this amount might be too low to justify the administrative costs to both the country and the firm. In order to tax more income within the existing rules, a destination country would have to both weaken the PE rule and reinterpret the value-creation rule to assign more value to activity within its borders. A recent Organization for Economic Cooperation and Development (OECD) document offers options for doing both.[6]

Once taxable corporate profit is allocated to each, the different countries are free to define the taxable base and set the tax rate. For example, a country may choose to allow a deduction for interest costs or provide a credit for research spending. Companies are free to place their management, manufacturing, sales, and service functions wherever they want. Although many factors, including infrastructure, the presence of trained workers, protection of intellectual property, and the legal system, influence this decision, corporate tax rates play a large role in certain conditions.

The system has several implications. One is that very little CIT revenue is currently directed to the destination countries (where the goods or services are sold). Most of the value gets taxed in the source countries where productive activity occurs. Another implication is companies can sometimes lower their taxes by placing productive activity in low-tax countries. This places pressure on national governments to make their tax laws more efficient by lowering the statutory rate or offering more tax incentives. This tax competition has caused many countries to significantly lower their corporate tax rates over the last few decades, and has led to calls for measures to impose a de facto minimum global CIT rate.[7] For some countries, the main response has been a DST.

The Nature of Digital Services Taxes

The European Commission launched the first major effort to enact a DST. In the spring of 2018, it brought forward two proposals to change the tax rules of every member of the European Union (EU). The first was a DST that would have levied a tax of 3 percent on the revenues of a narrow set of companies.[8] The second proposal would have replaced the DST with a new law that would have loosened the PE requirement for companies selling in the EU. This would have lowered the threshold for applying CIT to companies that export goods and services to their citizens, thereby erasing the need for a DST. However, this second law would not come into effect until similar reforms were made to the multilateral system.[9] The intention was to make the DST seem temporary and not be an impediment to multilateral negotiations. But, as the near-term outlook for multilateral agreement was dim, the DST was likely to stay in place for a long time.

The Commission’s proposals required unanimous consent in order to move forward. By December 2018, it had become clear many EU countries, including Ireland and Sweden, were intent on defeating them. Since then, other EU countries, including the United Kingdom, France, Spain, and Italy, have moved forward to enact their own versions of the DST. These have been joined by a number of other nations outside of the EU, including Australia, Chile, and South Korea.

Although each nation’s proposal differs in minor ways, they all share the same general features. First, they impose a relatively small tax, usually 2 to 3 percent, on the domestic revenue of firms subject to the tax. The focus on revenues was meant to avoid bilateral tax treaties that prohibit taxing corporations’ profits. The low rate disguises the DST’s impact. A 3 percent revenue tax, for example, is equal to a 23 percent income tax for a business with a markup of 15 percent (a company with $115 in revenues would have $15 in profits, and 0.03*115 = 15*0.23). For companies with very low margins or no profits, the effective rate can be much higher. For example, this would be equivalent to a 63 percent income tax on a company with a 5 percent rate of profit. Countries are in fact imposing the equivalent of a major income tax even though current rules prohibit them from doing so directly. The focus on revenues also likely means companies cannot deduct the DST from the CIT in their source country. As a result, DSTs will significantly increase the total taxes these companies and industries pay, thereby reducing overall global digital innovation.

DSTs will significantly increase the total taxes these companies and industries pay, thereby reducing overall global digital innovation.

Second, DSTs only target multinationals in a narrow range of business models meant to reflect the claimed sources of user value. The Commission’s proposal was aimed at activities wherein “the participation of a user in a digital activity constitutes an essential input for the business carrying out that activity and which enable that business to obtain revenues therefrom.”[10] Although the description could be read broadly, the Commission identified only three revenue sources that would meet this test: (1) placing ads targeting users of a digital interface based on their data profiles; (2) providing a multisided digital platform that allows users on one side of a transaction to interact with users on other sides, including the purchase of goods and services; and (3) selling users’ data.[11]

Third, the proposals include thresholds that would have eliminated all but the largest firms. The Commission’s proposal would have applied once a firm made €750 million in worldwide revenue and €50 million in revenues from within the EU. The U.K. proposal only would apply once a company has £500 million in global annual revenues. It would then subject all annual U.K. revenue above £25 million to a tax of 2 percent, although there would be a special provision for companies that have low profit margins or suffer losses. These thresholds would limit the number of foreign companies brought into the system while also largely exempting domestic firms.

The Impact of Digitalization

For most of the last century, the footprint of foreign subsidiaries tended to mirror that of their parents. Rather than a truly globalized business, a multinational company tended to look like a lot of national companies tied together through common ownership. In order to capture a large portion of foreign markets, multinationals usually found it necessary to have a significant presence within each target market, thereby exposing themselves to local income taxes. Several major trends have altered this pattern and allowed firms to become much more integrated worldwide, adjusting the structure and activity of each business to maximize the efficiency of both their supplier bases and consumer markets.

Some of these trends affect all industries.[12] Declines in the cost of transportation and communication have made it profitable to reach more foreign markets without basing production there. The increase in both the number and scope of trade agreements has made it cheaper to import both goods and services, and reduced the cost advantage of domestic suppliers. The relaxation of currency rules has made it easier to invest capital in foreign markets and take profits out. Firms now run complex supply chains spanning dozens of countries.[13]

This ability to serve global markets from only a few locations has been expanded by digitalization. Digitalization is the process of putting information into an electronic form so that it can be cheaply collected, transmitted, stored, and analyzed. Rapid improvements in both the capacity and cost of computer and communications systems, along with the rise of the Internet, have increased both the feasibility and the value of this process. While this has also affected goods exports, its main impact has been to make more services tradeable across borders.[14]

Rather than respond to the forces of globalization by competing to offer the best business environment and supporting exporters, many countries simply want to rewrite the rules, proposing changes that seek to tax at least a share of foreign activity enabled by digitization.

All of these trends, including digitalization, make it easier for companies to export products and, especially, services rather than produce them in the destination countries. Under existing rules, this reduces the importing countries’ corporate income tax base. But rather than respond to these forces of globalization by competing to offer the best business environment and supporting exporters, many countries simply want to rewrite the rules, proposing changes that seek to tax at least a share of this foreign activity.

To What Extent Are Digital Companies Unique?

Virtually all companies are increasingly digital. But some are more digital than others, and these “highly digital companies,” an OECD report argues, differ from traditional companies in three ways that supposedly make them more disruptive to the international system.[15] However, the first two don’t apply only to digital companies and the third is becoming a feature of an increasing number of companies regardless of industry.

The first is by making it easier for companies to sell in foreign markets without having much of a physical presence there (“scale without mass”). Companies have always had some ability to export goods and services to other nations. However, digital technologies are making it easier for a larger share of goods and services to be exported. Companies can increasingly reach customers around the world through Internet sales, relying on third parties for shipment, service, and returns. Companies also sell services online across borders, including Internet search and social networking services. But this is true for traditional companies with a web presence, not just “pure-play” Internet companies.

The second trend—the rise of intangible assets, especially intellectual property, as a source of company value—is also not strictly related to digitalization or limited to digital companies. An increasing share of the value of services, and even goods, is in intangible aspects such as intellectual property, design, and brands. Intangible assets are much easier to transfer between jurisdictions, but harder to value when being sold or leased.[16] Yet, the current arms-length standard provides at least some guidance on how to value these assets. Moreover, the BEPS process has tightened the rules governing this standard by insisting ownership and valuations reflect economic reality. Finally, the move toward intangibles is not solely, or even principally, a function of digitalization; increased R&D and marketing have also played a large role.

It is only the third factor OECD identifies—the involvement of users in certain digitalized business models—that is relatively unique to narrowly defined digital companies that provide only certain kinds of digital services, such as Internet search and social media applications. However, even here, this distinction is likely to erode. Indeed, OECD has written that these digital factors, including user involvement, “will become common features of an even wider number of businesses as digitalization continues.”[17] Moreover, it rightly points out that “since the degree of user participation may not closely correlate with the degree of digitalization, a pure focus on data and user participation without reference to other characterizing factors may imply that the tax challenges affect only a specific, more limited group of digitalized businesses.”[18]

The OECD report is clear in pointing out there is no agreement among members on whether this kind of user involvement represents value creation. Indeed, it states, “There are differences of opinion on whether and the extent to which data and user participation represent a contribution to value creation by the enterprise.”[19]

OECD does note, however, that for some country members:

The role of user participation is seen as a unique and important driver of value creation in digitalized businesses. These countries point to the participation and sustained engagement of users which allows digital businesses to collect large amounts of data through the intensive monitoring of users’ active contributions and behavior. These countries also point to the contribution of content by users, which can be central to a digital business’ offering and central to attracting other users and generating network effects.[20]

The core argument made by DST proponents is that these kinds of digital companies rely more heavily on value being created by the user and are therefore more like the German car company opening a factory in France, than a Canadian consumer importing a computer from the United States. This report discusses how user participation specifically has become a main argument used to justify the DST, and why it is fundamentally flawed. Here we will just point out that in most cases, the value does not come from the consumer, but from the business, just as it is from the German carmaker. Moreover, many businesses are increasing the data they collect, including from their customers. Customer feedback and sales lists have always been important to most traditional firms. But the Internet of Things increasingly allows them to put sensors into their products and collect detailed information on use and performance. This in turn has increased the value of the intellectual property contained in the product and made it easier for companies to shift more of their focus toward maintenance, service, and product improvement, and away from the initial sale.

Most nations recognize that digitalization is affecting all industries and has largely been a positive thing. A recent OECD report lists some of these benefits, including bringing platform markets into the formal economy, and improving both tax administration and compliance.[21] OECD also pointed out that it is impossible and would be unwise to try to ring-fence the digital economy by applying special rules to it.[22] Given that each of the three trends increasingly affects most industries, the case for a new tax that would only apply to a few firms is weak at best.

Do Highly Digitalized Companies Pose a Unique Threat to the International System?

Despite the fact that traditional industries increasingly incorporate digital traits, OECD has argued that highly digitalized companies may pose a special challenge to national tax bases and therefore to support for the international corporate tax system. A recent OECD report mentions three main policy challenges posed by these digital firms.[23] First, for some countries, the growing potential of digital technologies and the reduced need for an extensive physical presence calls into question whether the existing nexus rules for creating a taxable presence remain valid.[24] If companies can increasingly capture a large portion of a market from outside their country’s borders, fewer will conduct any substantial activity there, and some domestic firms may decide to move productive activity to countries with more competitive tax systems. We have already seen the Internet and other trends allowing industries to reduce their physical presence in a destination country. But this argument is flawed because the same thing could have been said about the rise of mass production industrialization in the late 1900s, as it also enabled companies to capture a large portion of a market from outside their country’s borders. If the issue is production, and exporting from tax havens, the solution is measures to limit tax havens, as BEPS is doing.

Second, OECD states that digital companies gather and use information across borders. This raises the issue of whether value is created from the generation of this data, and, if so, whether to characterize the supply of data as a free good or a barter transaction.[25] If it is the former, the company would normally not have to report taxable income. In the latter case, it might be viewed as having received value, and therefore taxable income, in return for its service, although in practice this is seldom done. But this assumes users are creating value whenever they provide data. The gathering and use of increased amounts of data, including that generated by customers, is not unique to highly digitalized companies. Financial services companies, manufacturers, and retailers, for example, also gather data and transfer it abroad.[26] Although it may be difficult to verify in which county the value behind this data is being created, the traditional principle that the sole act of collecting data does not create a great deal of value remains. This does not change whether the transaction is viewed as the provision of free labor or as the purchase of services with data instead of money.

Finally, the OECD report states that digital products and delivery services create uncertainty about whether a given transaction should be viewed as the sale of a service, a royalty for using intellectual property, or a provision of technical services, each of which might receive different treatment under existing tax laws.[27] In the case of a royalty, for example, the destination company is allowed to tax the transaction and enforce it using a withholding requirement. If it is a sale of a service, CIT may be collected only if value is added in that country, which it typically is not. The report mentions that this uncertainty is particularly true of computer services, such as processing power and storage, sold over the Internet, as their tax treatment apparently is still in doubt. Although the tax characterization of remotely delivered services to consumers may be an important technical question about how to tax the value of intangibles (and something the BEPS process can address), this does not imply these consumers are creating value when they merely “buy” free services.

Ring-fencing a few of the most digital business models is unlikely to solve the real problems facing the system and may even worsen them by making necessary reforms more difficult to achieve.

Continued changes, including digitalization, will no doubt require adaptation of the old rules to new developments, which has always been the case. The complexity of modern business simply makes it more difficult but does not call into question the fundamental rules of the international system. To the extent digitalization requires them, these reforms should reflect digitalization being a broad trend affecting most industries. Countries should be wary of introducing more complexity for the wrong reasons. An increasing volume of trade can now be conducted across borders, thereby creating competition between business climates—which is positive. Countries should not be allowed to stop it just because they fear losing revenue. Moreover, ring-fencing a few of the most digital business models is unlikely to solve the real problems facing the system, and may even worsen them by making necessary reforms more difficult to achieve. None of the trends through which digital services are thought to pose a unique problem seem to be new or even particular to the digital sector.

The Flawed Rationale for a Digital Services Tax

It is in this context that several countries have begun to enact a digital services tax to collect revenue from a limited number of large, mostly American, companies. The crux of their argument is international tax codes allow nations to collect CIT if there is value added in those nations. The DST advocates believe customers create a great deal of value in some, but not all, highly digitalized business lines. They point to three ways domestic users add significant value to specific digital services: user-created content, user data, and patronage of Internet platforms.[28]

DST advocates claim that, without these contributions, companies would make far fewer profits. But at its core, this is like arguing exporting manufacturing firms would make far fewer profits if foreign customers did not contribute money (e.g., pay for the products). Not to be deterred, policymakers in these nations contend that because the value associated with user content is provided domestically, the destination country should be able to tax the profits resulting from it, even if no other productive activities occur in the country and the company lacks a PE. These arguments are mistaken. In each case, it is difficult to argue users are creating great value.

DSTs only target multinationals in a narrow range of business models, which reflects the sources of user value previously mentioned. According to the United Kingdom’s HM Treasury, “The DST is focused on those business models for which the participation of a user base can reasonably be considered a central value driver, critical to the success or failure of the business.”[29] Most proposals would only cover profits from three business models: (1) placing ads targeting users of a digital interface based on their data profiles; (2) providing a multisided digital platform that allows users on one side of a transaction to interact with users on other sides, including the purchase of goods and services; and (3) selling users’ data.[30] But, not only are users not creating significant value, none of the business practices that supposedly lead to user value is unique to digital businesses. They increasingly characterize many traditional industries. There is no economic reason to target a narrow group of Internet companies for DSTs.

There is no economic reason to target a narrow group of Internet companies for DSTs.

User-Created Content

DST advocates argue that many social networking sites rely on user-created content to attract other users. Good examples are Facebook, YouTube, and Twitter. But this overlooks the fact that users create this value mostly for themselves and, to some extent, for friends and others in their network. In doing so, they largely fulfill their own desires. People want to post pictures of what they eat, descriptions of where they go, and their thoughts. Platforms let them do that for free on sophisticated, ever-evolving networks the companies add value to. Posting is consumption, not production.

It is helpful to divide users into three groups. Many people post very little on social media sites but enjoy consuming the content of others. This group often derives great value from their friends’ or relatives’ Facebook pages and Twitter postings about the latest news. But they add very little content. If they were to pay money for the service, the transaction would be treated as a normal purchase of a service. If they receive it for free, they pay for it by seeing ads. The substance of the transaction is the same, despite the fact that most DST proposals would only tax the second form of consumption. Most users consume more data than they create, and this trend seems to be growing.[31]

A second group may post a lot, but their postings have little value to the vast majority of users. For example, a user’s Facebook page might contain a great deal of content but have virtually no visitors—aside from their spouse—because so few other people are aware it even exists. These users provide content for free because they want to give something to their network. Companies are unwilling to pay them anything because of the low value their individual content creates—and the consumer is already getting the company’s services without paying.

Still, collectively, this group creates a great deal of content, which in turn draws many users. Governments normally do not tax the value associated with volunteered labor, whether it is someone posting a picture of a line at an airport or someone washing their clothes. Nor do they count as income the value users get from receiving free services, such as when someone eats dinner at a friend’s house or is given a ride to work. In part, this is because it is virtually impossible to measure the value of the volunteered labor, which in most cases is small anyway. DST proposals skip over this problem by arbitrarily taxing a fixed percentage of revenues derived from users in a given country. But if they are to do this, then, by definition, nations should tax as income all the volunteer work its citizens do, including the imputed value of watching ad-supported television, reading ad-supported newspapers, and interacting with ad-supported websites.

Finally, some users do create value for social networks. For example, President Trump’s tweets make Twitter a more valuable network. In this case, the user’s contribution can be loosely judged by the number of followers they have. Although they are adding value to the site, they are also being compensated for that value by being able to participate for free. Moreover, in some cases, the compensation is in the form of notoriety, fame, and influence, all of which may lead to higher income from other sources. In this sense, one could imagine the most popular social media users actually paying for the privilege of getting all that free exposure. In other cases, such as on YouTube and other content sites, the poster can receive direct payments from a share of the ad revenue generated by the viewers.[32] But in these cases, if the YouTube artist is a resident of France, the money they make is taxed in France.

Digital companies employ tens of thousands of workers and invest billions of dollars to support and continually improve their services. They, not users, are the ones adding value.

It is more accurate to view, as the OECD report does, the users’ supply of data as another input into the firm’s supply chain, similar to its purchases of data storage and broadband access, but wherein the purchase price is the free use of the platform or service. Firms do not pay corporate income taxes on the income others derive from selling them inputs, only on profits from the value they themselves add. Moreover, while this type of purchase theoretically could be subject to tax, countries rarely capture barter agreements when there is no cash payment on either side of the transaction.[33]

Finally, the overall argument that users create most of the value overlooks the fact that the distinguishing value lies in the technology, customer service, and business model of the social networking site, not the user content. Facebook did not beat MySpace because it had more users or better content; it provided a better platform. Moreover, digitalized companies employ tens of thousands of workers and invest billions of dollars to support and continually improve their services. They, not users, are the ones adding value.

User Data

Advocates of a digital services tax also argue that user data creates value. The United Kingdom’s consultation report claims, “A key driver of revenue will often be advertising based on data provided by the user.”[34] Many businesses, including digitalized businesses, collect data about their customers. But this is not value added. Rather, it is payment. Data is being provided in exchange for receiving the “free” service. There is no reason to think the data is worth any more than the value of the service it is being exchanged for. Some people have even argued that because users are creating value, they should get paid for it. It is not clear whether this would affect goods and services tax (GST) payments. It might, however, clarify the fact that users create little value because the amounts would be de minimis.[35] If companies are unwilling to pay, and users earn only a few dollars per year, users cannot create much value relative to the free services they get.[36]

The question of value creation can be clarified by looking at two alternative transactions. In the first, the transaction resembles a simple import of the service, with the user paying for the digital service with currency. The destination country could levy a VAT, but would not be able to collect CIT. The subscription revenues would also not be subject to a DST. DST proposals usually exempt paid digital services such as Spotify and Apple Music.

Alternatively, the user might pay for the service by providing their personal information and receiving more targeted and valuable ads. But this transaction is not substantively different from the first one: In both cases, the company is being paid to sell a service—in the former, with cash from the customer; in the latter, with cash from the advertiser data or attention from the user. As such, the tax treatment should be the same.

This DST approach creates an artificial distinction between similar business models. In each case, users are paying for the service by giving something of value to the company. In the case of Spotify’s subscription service, for example, the user is paying in money. In its free service, the user is paying by (hopefully) listening to ads. The ads are able to be better targeted to the customer based on data they provide. The data has no value to the customer, but presumably the company values it. Either way, the company receives value and the user gives up something (either money or time listening to ads). Based on actual practice, for many Internet applications, it seems both users and companies prefer the free model over the subscription service—otherwise the latter would be more widespread. Nonetheless, the DST subjects the free model to a revenue tax, but does not apply to the subscriptions. One result of a DST could be more companies switching to a subscription model, thus widening the digital divide.

In addition, the mere collection of data should not be viewed as creating value that justifies a radically different tax treatment. Companies collect data for many reasons, such as to improve the services they offer. A search engine’s ability to see what users are searching for and the sites they select from the search results allows companies to not only place ads they think are the most relevant, but also improve their algorithms. But the user is not creating value; the company is. Moreover, most traditional companies also try to collect data from their customers in order to refine their products, evaluate sales efforts, set prices, etc. Companies using the data for ads or selling it to third parties should not suddenly subject them to a DST.

One result of a DST could be more companies switching to a subscription model, thus widening the digital divide.

The Internet of Things will dramatically increase the amount of data each user generates, in virtually every industry. Under existing rules, this would not expose companies to the CIT. One of the newest trends in manufacturing is to use this data to improve products. For example, Tesla collects lots of information on the performance of every car it sells. Although the data can be important, its objective value is small. But if a DST is implemented—which ITIF believes should not be the case—to be consistent, it should not discriminatorily apply only to a narrow group of companies, but rather should apply to all revenues from data or user-contributed content.

User Patronage of Internet Platforms

Some Internet platforms exist in large part to connect different sides of a market with each other in order to transact. Examples include Amazon and Uber. DST supporters claim most of the value from these platforms comes from users on both sides of the market rather than the companies that create and run them. But the types of platforms covered would extend far beyond this. The U.K. proposal, for example, mentioned consumer-review and dating platforms. It also included search engines on the theory:

Much of the content of a search engine is delivered, directly or indirectly, by users of that platform. The intensive monitoring of user data also allows the platform to tailor experiences to individual users, and also to indirectly improve platform performance for other users.[37]

But surely most of the value of search engines lies in their ability to quickly deliver the most relevant sites to virtually any search request, which is a function of its software and algorithms. This is also true of other platforms. Uber, for example, would not be worth very much if drivers did not use it to earn money and riders did not use it to get places. But both drivers and users get value out of the experience. They presumably use a ride-sharing app rather than hail a regular taxi because the technology, coverage, and user experience offer them the best overall value. In fact, the DST argument gets it backward. The growth in users is a consequence of the value created by companies, not a cause of it. Granted, data about users, such as their location, can sometimes help tailor users’ search results (e.g., a search for an Italian restaurant). But in such cases, providing information benefits only the users—not the companies—by making their search results more valuable.

Of course, market platforms are neither new nor limited to digital companies. Newspapers, shopping malls, and dating sites, for example, have long matched users on different sides of the same market. Many of these platforms offer their services free to at least one group of users. In the Internet age, many of these platforms have moved to digital interfaces. The DST would apply to only those that are large and digital.

Even for the largest platforms, the argument for user-created value is flawed. In a case such as Uber’s, each side pays or receives payment for the narrow service each receives or supplies, nothing more. Uber riders pay approximately the same in fares as they would for taxis, and the drivers make about as much money, if not more. The extra value derived from users is hard to identify, especially if the platform is not profitable. Uber drivers are better off when more riders use Uber, but that is the same as shoppers benefitting from a department store because it offers products from more companies. The uniqueness is even more remote when parties interact with more than one platform.

More significantly, users are not creating most of the value for the business. It is much more likely the particular business offering, which may include software, sales strategy, user support, or pricing models, creates a service that distinguishes the firm from its rivals, and its users value. Google, for example, has a better system for recycling user searches to improve future results. Amazon’s ease of ordering, wide selection, and quick delivery increase consumer value in a way many other sellers have not been able to match. Moreover, the competitive battle between Uber and Lyft will not be settled by their individual riders or users, who often switch between the two platforms, but by the strength of the overall service the companies offer to both groups. In each, the technology, service, and business model—not users—account for the value added.

Finally, in the case of ride sharing (e.g., Uber), renting of properties (e.g., Airbnb), and other similar business models, companies have figured out a way to digitize a portion of the overall value created and “export” it to foreign nations. Before the Internet, taxi companies provided a range of services including driving and matching the driver and customer when the latter called to reserve a cab. But with the Internet, portions of that value (ride matching, payment, ranking of drivers, etc.), can be provided separately and remotely. The result is some nations want to be able to tax this new value creation even though it does not occur in their nation (the value creation of driving still does and can be taxed by taxing the drivers’ incomes).

A U.K. request for consultation raises some of the difficult judgments the definition of user content requires concerning which business models and revenues are in and which are out. After stating the sale of online games that allow users to play with other users in an online network would resemble the direct sale of online content and therefore would probably not be taxed, it points out:

There are online games that share similar features to social media and online marketplace business models e.g. those that benefit from the sustained engagement of a large user base that is encouraged to build networks, communicate and enter into exchanges. The government believes that there is a need to reflect further on how such online games should be characterized against the definitions above.[38]

With all the interactive games in the market, how could the government determine which have sufficient “user content”? Would massive multiplayer online games be in or out? What about games that let users buy and sell tokens?

Each of the three ways in which users participate with Internet businesses does not create much value for those firms. Participation is the benefit users receive from the Internet service.

In summary, each of the three ways in which users participate with Internet businesses does not create much value for those firms. Participation is the benefit users receive from the Internet service. While more users can increase network effects, users are drawn to Internet platforms largely by the value created by the digital company, not the actions of other users. Whether it is users paying for services with data and attention rather than money, using the service to post and enjoy content that benefits them, or purchasing products on a website rather than a shopping mall, users create little value. According to long-standing principles, the mere act of exporting these services should not expose a company to CIT in the users’ home countries. The fact that these user roles are growing in other industries further belies the argument that a narrow set of digital companies is unique and should be subject to a DST.

Proponents of a DST recognize that user participation is less central to certain business models. Examples include the direct sale of goods online and the provision of hardware and software.[39] Users do not contribute to the business offering, and their participation is not vital to its success. But the proponents never explain why this should be a determiner of who pays a DST. There are many markets wherein network and scale effects are critical to success. Microsoft Word is valuable largely because so many people use it. Car companies rely heavily on economies of scale to make their cars affordable. In other words, having many users (e.g., customers) makes their products cheaper and more valuable. Yet neither would be subject to a DST.

Digital Services Taxes Introduce Market-Disrupting Distortions

The implementation of a revenue tax, especially one structured like the DST, raises a host of policy problems, some of which are specific to the provisions of a particular bill, while others are inherent to the broad outlines of DSTs in general.

With regard to the latter, a tax on revenues inevitably has a disproportionate impact on firms with low profit margins or losses. Yet the large upfront costs required to build out a network, and the practice of pricing low in order to attract more users on every side of the market, means many digital companies experience large losses prior to becoming profitable. DSTs would make this growth strategy significantly more difficult by increasing the “burn rate” of cash. DSTs partially address this by not imposing the tax until a firm reaches a certain size in terms of revenues over which it might be presumed to be profitable. The U.K. proposal promised a separate safe harbor for firms with low profit margins or losses. But this is not well defined, and other proposals omitted it entirely.

Aside from this, DSTs violate several goals of good tax policy. At the 1998 Ottawa Ministerial Conference, the international community listed a set of overarching traits that should characterize tax policy: neutrality, efficiency, certainty and simplicity, effectiveness and fairness, and flexibility.[40] Of course, these goals often conflict with each other, making it hard to achieve all of them. Although trade-offs must be made, neutrality is particularly important because treating similar transactions differently distorts economic decisions. This can be desired if tax policy effectively corrects for public costs and benefits that are not otherwise reflected in the market price. But in the absence of externalities, taxing similar transactions differently usually reduces national income and welfare because it deters parties from choosing the cheapest—and therefore most efficient—options.

Ideally, tax policy would be neutral with respect to several factors, including industry, size, business function, and nationality, except when clear market differences justify different treatments. DSTs violate neutrality in each of these ways, creating incentives that tend to slow growth and innovation.


DSTs obviously discriminate against a narrow set of highly digital industries. For example, the United Kingdom's proposed DST targets the following three business models: search engines, social media platforms, and online marketplaces.[41] But why these—and not others such as subscription music services or Internet sites that only sell products from one manufacturer?

This explicitly violates OECD’s admonishment against trying to ring-fence the digital economy with special rules. This advice was based on the fact that most industries are becoming more digitalized, and the nature of the most digitalized industries continually undergoes rapid change, causing special rules to rapidly become out of date.

The Commission originally justified this selectivity by alleging the largest digital companies were not paying their “fair share” of taxes. At one point, the Commission estimated domestic digitalized companies paid an effective tax rate of only 8.5 percent, compared with 23.2 percent for the “traditional international business model.”[42] This estimate was criticized by the Commission’s Regulatory Scrutiny Board, which pointed to “significant shortcomings,” including the fact that it did “not show the urgency for the EU to act, before global progress is achieved at the OECD/G20 level.”[43]

At least two studies have shown that, even before the recent reforms, large digital companies often paid higher effective tax rates than their peers in more traditional industries. Although he found, “The current system of international corporate taxation, as it is enforced in most countries inside and outside the EU, is fundamentally flawed and in need of substantial reform,” economist Matthias Bauer has argued that digital industries neither call for nor deserve separate tax policies based on their alleged avoidance of significant taxation.[44] His analysis shows digital companies often pay far more in taxes than many large and well-known traditional companies based in the EU. A study by Copenhagen Economics made a similar finding, pointing out digital companies often benefit from tax provisions meant to encourage large research expenditures, which economists have shown benefit society as much, if not more, than the companies doing the research.[45] In addition, these companies are also more likely to rely on equity funding, which raises their effective rates.

The argument that digital companies do not pay enough taxes in total is false. The real argument is over the distribution of that tax revenue among different countries.

The argument that digital companies do not pay enough taxes in total is false. The real argument is over the distribution of that tax revenue among different countries. Because some countries are dissatisfied with the current distribution, they are acting unilaterally to increase their tax revenues.

Even if the substance of the argument about user-created value is accepted—which it should not be—the decision to apply DST to only a narrow set of largely U.S. companies is flawed. The basic argument that users create value applies to many industries. Although BMW and Tesla, in addition to other firms, regularly collect data from the consumer use of their products, they would not be covered. Similarly, if Harley-Davidson were to build an Internet platform wherein users in a foreign nation could meet eachother, discuss their love of motorcycles, and make suggestions for improvements, it also would escape the tax. Because virtually all industries are becoming more digitized, and new business models are forming, there is likely to be continued pressure to apply DST to more companies, especially if the tax produces less revenue than anticipated.

Another problem with discriminating by industry is all the largest targets of the DST also have business lines that do not fall under the tax, such as artificial intelligence, drones, driverless cars, Internet delivery, medical records, and cloud services. Collectively, these companies are a major source of investment and innovation in each area. This further complicates the question of how much value users contribute to companies’ profits.


DSTs penalize companies that get too big. For example, once a company passes the U.K. threshold of £500 million in global revenues and £25 million in U.K. revenues, they suddenly become subject to the tax. Because there is nothing unique about this size level, its choice is arbitrary and unfair. It is not clear why users suddenly create more value when a company gets beyond this size.

The threshold clearly punishes firms for getting larger and reduces their incentive to grow. Moreover, it could easily reduce the attractiveness of mergers, especially when the combined entities would cross the revenue threshold. Although some might welcome less concentration, it could just as easily harm competition—by removing one of the main exit strategies for venture capitalists that invest in new firms—while closing off one of the principal ways promising technologies achieve scale.

Finally, the size threshold implicitly overrules the PE test, which exempts these companies from local CIT. Advocates of the DST argue that the PE test is irrelevant because the DST is not a corporate income tax. But the focus on revenues is little more than an end run around the international standards and bilateral treaties governing the corporate income tax.


DSTs also treat corporate functions differently, even when they have the same basic form. This benefits some types of activity over others, despite there being no clear policy rationale for doing so. Three examples illustrate this problem. First, with regard to online marketplaces, a platform that sells only products from one supplier would not be subject to the tax, while platforms that match consumers with a variety of sellers would be taxed. Thus, a site that offers products from many clothing manufacturers would have to pay the tax, while a company selling only its own label would not, even if third parties actually made the clothing, and both websites featured consumer product reviews. The implicit reason for this—users creating significant value for a multi-seller platform, but not a single company site—is not logical.

DSTs treat corporate functions differently, even when they have the same basic form.

This gives single-vendor sites a clear price advantage, and platform owners an incentive to narrow competition. For example, a purchase from Amazon’s website would be subject to a DST if it were from a third-party vendor, but not if it were from a site that only sells Amazon products—even if the product were the same. That would give Amazon an incentive to expand its product offerings at the expense of other suppliers. It would also give other online sellers an advantage over Amazon, provided they used their own websites to sell only their own products, and would harm the many independent suppliers that rely on Amazon to be their main source of marketing and sales, and to connect them to a vast number of potential customers. Of course, Amazon might decide to get around the tax by simply purchasing those products outright before selling them on its site.

This distinction makes little sense. The platform owner’s role in creating value is not substantively different when it is limited to one vendor rather than several. Even—perhaps especially—in the multivendor model, the lion’s share of the competitive advantage is in the platform, execution, and overall business model of the firm, not the participants.

Second, companies that gather user data for internal purposes, such as improving product quality, offering maintenance service, or building a customer list, would not be covered by the DST. But if a company either sold that data to others or used it to sell ads targeted to its users, DST would apply. This would be true even if the companies collected the same data from the same users, and the data were worth the same to the two companies.

Finally, companies that monetize their service by selling subscriptions would escape having to pay DST because users pay them in currency rather than in data or content. But free streaming services that sell ads would have to pay. A U.K. request for consultation states that a key element for taxing an online platform such as a search engine is that “it generates revenue by monetizing users’ engagement with the platform and with other closely integrated functions e.g. websites accessed through a web browser.”[46] A website that makes money from both ads and subscriptions would have to pay DST on the former revenues but not on the latter.

Why should one business model, especially one that is more attractive to higher-income households, be preferred over another? The difference in treatment also raises problems for companies that use a freemium model wherein users can choose a basic form of the service for free or pay a subscription to receive expanded services.


Finally, the DST clearly discriminates against foreign companies—although its defenders argue this is only by coincidence. Some Europeans have mentioned their surprise when American officials assert the DST is aimed at U.S. companies. Yet, given the structure of the tax and its actual coverage, it is difficult to believe the nationality of various prospective targets does not come into play.

On one level, the DST clearly discriminates against foreign firms that are not otherwise subject to a given country’s corporate income tax, presumably due to a lack of PE. To avoid double taxation, some countries allow firms to deduct DST payments from CIT or VAT liabilities. But these deductions are of little use to foreign companies that would otherwise not pay the CIT. In the case of a corporate income tax, foreign companies would normally receive at least a partial credit against the corporate tax paid in their own country. Because the DST is levied on revenues rather than income, it is very likely foreign companies will not be entitled to a credit in their home jurisdiction. The reduced ability to offset the DST against other taxes places foreign companies at a disadvantage, as their overall tax burden becomes higher

The U.K. has designed its DST to “ensure it is established tech giants—rather than our tech start-ups—that shoulder the burden of this new tax.”

A recent article about the U.K. tax points out the only two search engines likely to qualify for the tax are Google and Microsoft’s Bing. Among social media firms, it cites only Facebook. Two other sites, Twitter and Snapchat, earn enough revenue to qualify but are not yet profitable. And the coverage of online marketplaces will likely be limited to eBay and Amazon. The article concludes that, for now, only American tech firms are likely to qualify for the tax.[47] The government has admitted as much. According to the chancellor of the exchequer, “[the UK DST] will be carefully designed to ensure it is established tech giants—rather than our tech start-ups—that shoulder the burden of this new tax.”[48]

The threshold is therefore likely to protect most domestic companies from having to pay DST. Although they will pay the United Kingdom’s corporate income tax, their foreign competitors will likely have to pay both the DST and CIT in their home country, giving the former a clear advantage in domestic markets—although it may diminish their value by constricting their growth. As DSTs are very likely to fall most heavily on American companies, they have the definite look of a revenue grab. This should directly concern U.S. policymakers.

Precisely because it falls mainly on U.S. companies, the DST also likely violates existing trade agreements.[49] Countries are allowed to tax domestic transactions, provided they treat domestic and foreign suppliers the same. A tax determined by a portion of the market price has the feel of a tariff, especially when the vast majority of the revenue comes from foreign firms. The United States Trade Representative recently took this approach in a report on trade relations, stating:

The United States opposes proposals by any country to single out digital companies. Such proposals are based on an unprincipled and unsupported distinction between digital companies and non-digital companies. In addition, U.S. companies have expressed concerns that the specific services included in the proposal along with the thresholds for global and EU-wide revenues appear to target almost exclusively U.S. companies and so would have a discriminatory effect on U.S. suppliers participating in EU markets.[50]

There are other problems with the DST. A number of commentators have questioned whether DSTs will raise as much revenue as promised. A report by Copenhagen Economics analyzed the Commission’s estimate that the DST would raise 4.7 billion each year and found that an estimate of 2.5 billion to 2.7 billion was more likely, and that the take could go even lower, if the distortions introduced by the tax cause a large amount of revenue to shift away from companies subject to it.[51] Low revenue could create pressure to expand DST’s scope.

Determining the location of users (and therefore which government is entitled to the revenue tax) is also likely to be a huge problem, especially for mobile sales (in which users’ locations can change significantly, even within a single session) and large transactions (when users might have an incentive to hide their location in order to avoid GST, if part of it were passed on in the form of higher prices or fewer services). GST proposals generally support utilizing users’ IP addresses, but also allow companies to propose a better identification method. But this does not always provide a clear answer. The U.K. government, for example, recognizes the following “difficult cases”:[52]

  • Users whose location is not tracked, possibly due to privacy concerns (e.g., they are using a virtual private network, or VPN);
  • Users whose IP address contradicts the intended destination of advertising;
  • Users who frequently cross borders;
  • Users who register for a platform while traveling; and
  • Users whose foreign employer purchases access to the platform on their behalf.

The tax could also reduce a nation’s ability to compete in digital markets. Although the tax protects smaller domestic firms, it also restricts their incentives to grow, and makes them less attractive as merger targets. The tax is also likely to reinforce the current pattern of highly digitalized companies starting up in the United States—where they benefit from a large national market, less regulation, and fewer controls—and expanding to Europe only after they have attained a large scale.

The Base-Erosion and Profit-Shifting Process

The international community has been trying to address legitimate international tax concerns. There are clear cases wherein companies have used the international rules to reduce their taxes in ways which, though legal, were not intended by policymakers. Many of these strategies also violate the spirit of the laws. A 2015 OECD report mentioned several ways in which companies could shift profits to low-tax countries or generate artificial deductions by using loopholes in individual tax treaties.[53] In some cases, these practices could result in profit not being taxed anywhere.

The BEPS project began in 2012 when the G20 directed OECD to develop an agreement on the issue of tax erosion. This plan was approved the next year, and by 2015, the G20 agreed to an action plan consisting of 15 action items. Action item 1 is specifically devoted to the broader tax issues surrounding the digital economy. OECD is directed to “identify the main difficulties that the digital economy poses for the application of existing international tax rules and develop detailed options to address these difficulties.”[54] The 2013 Action Plan mentioned many of the arguments being made by today’s supporters of reform. Since then, the international community has made significant progress in implementing specific action items that address issues that were widely perceived to be problems. But it has not resolved the broader issue of whether the tax system itself needs broader reform.

Collectively, the action items aimed to accomplish three overarching goals.[55] The first was to introduce coherence into the domestic rules and bilateral tax treaties that affect international transactions. This eliminated the possibility of companies finding idiosyncrasies in the interaction of specific bilateral treaties that allow income to go untaxed by either jurisdiction. The second goal was to reinforce the substance requirements of the current standards so taxation occurs wherever the economic activities and value creation occur. By requiring transfer prices to reflect both the economic substance of the transaction and arm’s-length pricing, countries hope to minimize the ability to shift profits between countries. Finally, the action items attempted to increase transparency and improve legal certainty for both companies and countries.

A 2018 OECD report found that, while still early in the process, countries had gone a long way in achieving widespread implementation of the BEPS measures, and their efforts were having an impact.

Collectively, these reforms made it harder for companies to use the flexibility inherent in digitalization to change tax results. A 2018 OECD report found that, while still early in the process, countries had gone a long way in achieving widespread implementation of the BEPS measures, and their efforts were having an impact.[56] Perhaps most important, a significant number of multinationals have already taken steps to align their corporate structures with actual economic activity.[57] The OECD report specifically mentions Amazon, eBay, Facebook, and Google. This should go a long way toward addressing areas wherein there were legitimate problems and a consensus for reforms. Given this progress, unilateral moves seem especially unwarranted.

The Breakdown of Multilateral Consensus

The BEPS process does not, however, solve the question of whether digitalization represents a broader challenge to the international tax system. Among experts, there seems to be wide agreement that the existing system is unstable due to growing complexity and declining support for the current allocation of CIT among source and destination countries, even as reformed by BEPS. Much of the latter is caused by concern that the existing rules do not adequately protect countries from tax competition. Now that companies cannot reduce their taxes by taking advantage of paper transactions, they may decide to lower them by moving substantive activity. A recent report by the International Monetary Foundation explicitly discusses “tax competition” as being an arguably greater threat to the system than tax avoidance.[58] The global intangible low-taxed income and base-erosion anti-abuse tax provisions in the U.S. tax reform legislation reflect this concern by ensuring U.S.-headquartered firms pay a minimum tax on all of their worldwide income.

The 2018 OECD report found there is no agreement among countries over the tax implications of scale without mass and the greater role of intangible assets. It also found no consensus on whether, and the extent to which, data and user participation contributes to value creation.

Although views span a broad spectrum, the report generally describes them as falling into three groups.[59] The first group of countries believes some characteristics frequently observed in highly digitalized business models can lead to misalignments between the location in which profits are taxed and where value is created. This group believes reform is needed, but that it should be limited to a very small set of highly digitalized Internet business lines because the source of their value is created by users in each country rather than by the companies themselves. Under the principle that profits should be taxed wherever value is created, they believe these companies should pay CIT to the country wherein the users live, rather than where the company operates. The Commission and many of its individual members belong to this group. These are the countries that are implementing DSTs.

The second group of countries believes the ongoing digital transformation of the economy, and the broader increase in globalization, challenge the continued effectiveness of the existing international corporate tax framework, but are not exclusive or specific to highly digitalized businesses. Broader reform is therefore needed. The United States belongs to this group, and has proposed the taxation of so-called “marketing intangibles” as a possible solution. This group apparently disagrees with the assertion that users create significant value in highly digitalized firms, but does believe existing rules do not adequately measure where some value is created. Finally, a third group of countries believes the BEPS process has largely solved the problem of double nontaxation and do not see the need for significant reform to the existing rules.

Although OECD is committed to developing further agreement on these issues, the chances of a quick agreement seem slim. On January 23, 2019, the BEPS Project issued a policy note soliciting views on three alternative reforms.[60] Each proposal would significantly modify both the arm’s-length test for transfer pricing and the need for a physical presence. They would also move the system toward some type of apportionment formula in which at least part of the global tax base is distributed among countries based on a fixed formula involving some combination of sales, employment, and assets in each country. The plan is to have a detailed program of work OECD could agree to at its May meeting, with the goal of reporting progress at the G20 meeting of finance ministers in June 2019, and delivering a solution in 2020.[61] This is highly ambitious.

Why Not a VAT?

The Spanish government has apparently argued that its DST would be compatible with the European VAT.[62] The statement is revealing because the DST is basically a VAT in disguise. Whereas a DST is levied on revenues, the VAT is applied to sales prices. But in most cases, the marginal revenue from a transaction is the same as the sales price. Because international tax treaties clearly give the destination countries the right to apply VAT to transactions within their borders, including sales involving imports, countries would be well within their rights to tax Internet transactions with a VAT the user pays. So why do they not?

There are several reasons. One that is not particularly convincing is VAT might be difficult to collect from foreign companies that have only a minimal presence in a given country. DSTs have the same problems. The BEPS project developed guidelines for collecting VAT on business-to-consumer cross-border transactions.[63] The guidelines affirm that the country wherein the consumer has their normal place of residence has the right to collect VAT on remote supplies of services and intangibles. They also recommend establishing a simplified registration and compliance regime under which the foreign seller is required to register and collect VAT at the same rate as domestic suppliers. OECD concluded that in countries implementing these measures a significant number of suppliers comply by either registering in the VAT jurisdiction, collecting and paying tax on their remotely delivered services, or by setting up a PE in the country.[64] So far, over 50 jurisdictions have adopted rules in accordance with these guidelines.[65] It is therefore possible under existing agreements to increase tax revenues from distant firms with a VAT.

Other reasons undermine the merits of DSTs. The first is many Internet services are provided for free, so there is no sales price to attach the tax to. DST supporters get around this problem by slicing off part of a company’s ad and other revenues and arbitrarily assigning them to the destination country. The focus on user-created value thus ends up justifying an arbitrary tax on revenues.

Second, international rules, including trade agreements, clearly state domestic producers cannot receive preferential tax treatment. Under a VAT, the high thresholds would have to go, exposing domestic companies to the same tax. Third, a VAT would make it clear that at least part of the tax burden would fall on consumers. How much depends upon the elasticities of the supply and demand curves. But as the marginal price of adding consumers is almost zero, much of the burden would likely be borne by users. Although a VAT would clearly be within a country’s rights, it would weaken the claim that DSTs are aimed at large companies that are not paying their fair share.

The U.S. Response Thus Far

Prior to recent tax reform, the United States was often viewed as part of the BEPS problem. Although it subjected all foreign income to an extremely high tax rate of 35 percent, the tax was not due unless companies brought the income back to the United States. Many companies, including some of the largest Internet firms, accumulated large amounts of foreign profits overseas. European countries often failed to consider this deferred U.S. tax liability when assessing whether U.S. firms were fairly taxed.

Tax reform dramatically lowered the U.S. corporate income tax. But it also contained several elements that prevent or deter companies from engaging in BEPS. One could argue that the United States has done its part to minimize BEPS. The United States has also tried to engage in the broader concerns regarding the international system by proposing at least some of the profits companies derive from “marketing intangibles” (e.g., brand names and customer relationships) be directed to destination countries. The Department of the Treasury has promised to release a report providing more details on the proposal.

The United States needs to defend its national and global interests with more vigor. The DST seems like an ill-disguised effort to target companies that are thought to be too powerful, too profitable, and too American.

The United States must remain engaged in these issues. The country continues to benefit from being home to many of the largest Internet companies. The DST would raise revenue by taxing many of its largest firms. If the United States allowed the countries to deduct DST payments from their U.S. tax liability, its own revenues would decrease.

Although the United States has opposed unilateral changes to the existing system, its public involvement has been minimal. In March 2018, Secretary of the Treasury Steven Mnuchin issued a statement that the United States “firmly opposes attempts to single out digital companies, while fully supporting international efforts to address broader tax challenges arising from the modern economy.”[66] Last October, the chairman and ranking member of the Senate Finance Committee wrote a letter urging both the Commission and the European Council to cease consideration of the digital tax on revenues, arguing that it would violate the principle that taxes on multinationals should be based on profits, not revenues, and that it discriminates against U.S. companies.[67]

These two cursory actions have done little to influence either the debate over, or the momentum toward, unilateral action. Although neither Commission directive received the unanimous support it needed to go into effect, the failure was mainly due to the opposition of EU countries, including Ireland, rather than U.S. pressure. Meanwhile, individual countries are moving toward enacting DSTs on their own. The United States needs to defend its national and global interests with more vigor. The DST seems like an ill-disguised effort to target companies that are thought to be too powerful, too profitable, and too American.


For almost a century, both the question of when a country may tax foreign corporations and the distribution of taxable profits among countries have been guided by an international consensus on basic principles. These principles are reflected in a large number of multilateral and bilateral tax treaties, and generally require a company to have a PE in a country and add value domestically before it becomes subject to that country’s corporate income tax. The determination of how to distribute profits among countries generally reflects the amount of value created in each country, with the caveat that the mere sale of a good or service neither defines a PE nor creates value. Finally, the treatment of transactions within a company, while often subjective, should reflect an arms-length bargain.

These principles worked well when a multinational’s operations within each country were largely isolated from each other. Over the last several decades, a number of developments have allowed multinationals to create globally integrated companies in which cross-border transactions are both more numerous and less tangible. The rise of information technology, including the Internet, has challenged the consensus even further by making it much easier for companies to earn revenue from a country’s consumers without having a physical presence there.

Since 2012, the international community has made significant progress addressing the legitimate problems of base erosion and profit shifting. The basic agreements emerged within three years, and over the last three years, countries have made good progress implementing them. The result is a wide consensus that the most serious problems have been addressed.

However, a consensus has yet to be formed on some of the broader implications of digitalization. Some countries are concerned they are losing revenues as more domestic output is traded over the Internet. Rather than accept that reality and raise other domestic taxes (e.g., VAT) or work to boost their own exports (the profits of which would be taxed domestically), they have sought simply to tax foreign corporate revenues.

The United States must continue engaging in these issues to find a way of restoring the consensus around how to treat multinationals. But it must make clear that DSTs are not the answer. Policymakers advocating for DSTs rely on unconvincing explanations of how users in a few business models create value. As such, DSTs are not only capricious and unfair, they introduce a number of destructive distortions into the economy. The net effect of these distortions will likely be a reduction in consumer surplus, economic growth, and innovation. Countries should not be allowed to impose unilateral taxes on foreign companies simply because they think the current rules need reform.


[1]. It is important to note that most of these strategies were legal under the existing rules. The main complaint against them was they violated the “spirit” of the law. The Tax Justice Network states, “Tax avoidance complies with law, but goes against the spirit of what our legislators intended.” Tax Justice Center website, Tax Avoidance, (accessed April 25, 2019).

[2]. Bureau of Economic Analysis, “Interactive Access to Industry Economic Accounts Data: Value Added by Industry,”, (Accessed April 23, 2019).

[4]. Although the Organization for Economic Cooperation and Development has emphasized this criterion as the central principle for allocating international profits, others have argued that it has never been the sole or best criteria. For a criticism, see John Vella, “Taxing the Digitalized Economy: Targeted or System-Wide Reform,” British Tax Review, No. 4, 2018, 306–14. For example, there is growing support for having home countries impose a minimum tax on all global revenues of multinationals domiciled in their territory. This tax would apply even if all value creation occurred and were already taxed in another country.

[5]. International agreements commonly require a country to apply the same rate to all companies, regardless of whether they are domestic or foreign.

[6]. OECD/G20 Base Erosion and Profit Shifting Project, Addressing the Tax Challenges of the Digitalization of the Economy, Public Consultation Document, 2019,

[7]. International Monetary Fund, “Corporate Taxation in the Global Economy,” IMF Policy Paper, March 2019, 11–12,

[8]. European Commission, “Proposal for a Council Directive on the Common System of a Digital Services Tax on Revenues Resulting From the Provision of Certain Digital Services,”

[9]. European Commission, “Proposal for a Council Directive Laying Down Rules Relating to the Corporate Taxation of a Significant Digital Presence,” March 21, 2018, COM(2018) 147,

[10]. European Commission, “Proposal for a Council Directive on the Common System of a Digital Services Tax on Revenues Resulting From the Provision of Certain Digital Services,” 7.

[11]. Ibid, 17.

[12]. Organization for Economic Cooperation and Development, Addressing the Tax Challenges of the Digital Economy, Action 1: Final Report (OECD/G20 Base Erosion and Profit Shifting Project, 2015, 92, 143,

[13]. McKinsey Global Institute, Globalization in Transition: The Future of Trade and Value Chains, January 2019,

[14]. Services are becoming a growing source of value added in many industries. See letter by Nigel Cory, Trade Policy Analyst for the Information Technology and Innovation Foundation, to Ed Gresser, chair of the Trade Policy Staff Committee, Office of the United States Trade Representative, June 17, 2017,

[15]. Organization for Economic Cooperation and Development, Tax Challenges Arising from Digitalisation–Interim Report 2018: Inclusive Framework on BEPS, 24–25.

[16]. Carol A. Corrado, Charles R. Hulten, and Daniel E. Sichel, “Intangible Capital and Economic Growth” (National Bureau of Economic Research, Working Paper No. 11948, January 2006),

[17]. Organization for Economic Cooperation and Development, Tax Challenges Arising from Digitalisation–Interim Report 2018: Inclusive Framework on BEPS, 24.

[18]. Ibid, 26 (endnote omitted).

[19]. Ibid, 25.

[20]. Ibid, 25.

[21]. Ibid, 194–96.

[22]. Organization for Economic Cooperation and Development, Addressing the Tax Challenges of the Digital Economy, Action 1: Final Report, 11.

[23].  Ibid, 99.

[24]. Ibid, 100–02.

[25]. Ibid, 102–04.

[26]. Daniel Castro and Alan McQuinn, “Cross-Border Data Flows Enable Growth in All Industries” (Information Technology and Innovation Foundation, February 24, 2019), /publications/2015/02/24/cross-border-data-flows-enable-growth-all-industries.

[27]. Organization for Economic Cooperation and Development, Addressing the Tax Challenges of the Digital Economy, Action 1: Final Report, 104–06.

[30]. European Commission, “Proposal for a Council Directive on the Common System of a Digital Services Tax on Revenues Resulting From the Provision of Certain Digital Services,” 17.

[31]. Shane Greenstein, “Six Infrastructure Trends” March 6, 2019, Technology, Academics, Policy Blog,

[32]. Caitlin O’Kane, “Top 10 Highest-Paid YouTube Stars of 2018, According to Forbes,” CBS News, December 4, 2018,

[33]. Organization for Economic Cooperation and Development, Tax Challenges Arising from Digitalisation–Interim Report 2018: Inclusive Framework on BEPS, 25–26.

[34]. HM Treasury, Digital Service Tax: Consultation, 8.

[35]. Daniel Castro and Alan McQuinn, “No, Internet Companies Shouldn’t Have to Pay You for Your Data,” The Sacramento Bee, March 14, 2019,; Eline Chivot, “Paying Users for Their Data Would Exacerbate Digital Inequality,” Center for Data Innovation Blog, January 11, 2019,

[36]. HM Treasury, Digital Service Tax: Consultation, 11.

[37]. Ibid, 7–8.

[38]. Ibid, 13.

[39]. Ibid, 8.

[40]. Organization for Economic Cooperation and Development, Addressing the Tax Challenges of the Digital Economy, Action 1: Final Report, 20–21.

[41]. HM Treasury, Digital Service Tax: Consultation, 46–48.

[42]. European Commission, A Fair and Efficient Tax System in the European Union for the Digital Single Market, Communication from the Commission to the European Parliament and the Council, Brussels, COM(2017) 547, September 21, 2017, 6,

[43]. European Commission, Regulatory Scrutiny Board, Opinion: Impact Assessment/Fair Taxation of Digital Economy, Brussels, SEC(2018), January 24, 2018, 1,

[44]. Matthias Bauer, “Corporate Tax Out of Control: EU Tax Protectionism and the Digital Services Tax” (European Policy Information Center and European Center for International Political Economy, February 2019), 9,

[45]. Helge Sigurd Næss-Schmidt et al., “The Proposed EU Digital Services, Tax: Effects on Welfare, Growth and Revenues” (Copenhagen Economics, September 2018),

[47]. Rory Cellan-Jones, “Budget 2018: Who Will Pay the Digital Services Tax,” BBC News, October 30, 2018,

[48]. James Hill and James Harrison, “INSIGHT: U.K. Digital Services Tax—How Will It Work in Practice?” Bloomberg Tax, December 5, 2018,

[49]. Gary Clyde Hufbauer and Zhiyao Lu, “The European Union’s Proposed Digital Services Tax: A De Facto Tariff” (Peterson Institute for International Economics, Policy Brief 19–15), June 2018,

[50]. United States Trade Representative, 2019 National Trade Estimate Report on Foreign Trade Barriers, 2019, 212,

[51]. Helge Sigurd Næss-Schmidt et al., “The Proposed EU Digital Services Tax: Effects on Welfare, Growth, and Revenues,” 20.

[52]. HM Treasury, Digital Service Tax: Consultation, 20.

[53]. Organization for Economic Cooperation and Development, Addressing the Tax Challenges of the Digital Economy, Action 1: Final Report, 167–80.

[54]. Ibid, 16.

[55]. Ibid, 3.

[56]. OECD, Tax Challenges Arising from Digitalization–Interim Report 2018, 106.

[57]. Ibid, 91.

[58]. International Monetary Fund, “Corporate Taxation in the Global Economy,” 11.

[59]. OECD, Tax Challenges Arising from Digitalization–Interim Report 2018, 171–72.

[60]. OECD/G20 Base Erosion and Profit Shifting Project, Addressing the Tax Challenges of the Digitalization of the Economy—Policy Note, as approved by the Inclusive Framework on BEPS, January 23, 2019,

[61]. OECD/G20 Base Erosion and Profit Shifting Project, Addressing the Tax Challenges of the Digitalization of the Economy, Public Consultation Document.

[62]. E&Y, “Spain Releases Draft Bill on Digital Services Tax,” October 25, 2018,

[63]. OECD, Addressing the Tax Challenges of the Digital Economy, Action 1: Final Report, 126–29.

[64]. Ibid, 122.

[65]. OECD, Tax Challenges Arising from Digitalization–Interim Report 2018, 103.

[66]. U.S. Department of the Treasury, “Secretary Mnuchin Statement On OECD’s Digital Economy Taxation Report,” March 16, 2018,

[67]. Joe Kirwin, “Senate Tax Writers Push EU to Ditch Digital Tax,” Bloomberg, October 19, 2018,


The author wishes to thank Robert D. Atkinson and Eline Chivot for providing input to this report. Any errors or omissions are the author’s alone.


Joe Kennedy is a senior fellow at ITIF. For almost 30 years, he has worked as an attorney and economist on a wide variety of public policy issues. His previous positions include chief economist with the U.S. Department of Commerce and general counsel for the U.S. Senate Permanent Subcommittee on Investigations. He is president of Kennedy Research, LLC, and the author of Ending Poverty: Changing Behavior, Guaranteeing Income, and Transforming Government (Rowman & Littlefield, 2008). Kennedy has a law degree and a master’s degree in agricultural and applied economics from the University of Minnesota, and a Ph.D. in economics from George Washington University.


The Information Technology and Innovation Foundation (ITIF) is a nonprofit, nonpartisan research and educational institute focusing on the intersection of technological innovation and public policy. Recognized as the world’s leading science and technology think tank, ITIF’s mission is to formulate and promote policy solutions that accelerate innovation and boost productivity to spur growth, opportunity, and progress.

For more information, visit us at

Back to Top