Unilateral Taxation of the Digital Economy
The fight is not over yet—it’s only beginning

print this article
An abstract illustration of a phone on a blue background

Notwithstanding the recent efforts of the G20/OECD Inclusive Framework on BEPS (hereafter the IF) to develop a uniform, multilateral, and consensus-based solution for taxing the digital economy, legislatures and tax administrations around the world continue to propose and enact a host of largely uncoordinated digital services taxes (DSTs) and other unilateral digital tax measures. This article examines the current status of the IF’s efforts to thwart unilateral measures through a consensus-based solution, some of the proposed unilateral measures, the problems inherent in these measures, and what multinational enterprises (MNEs) can do now to prepare for the possibility of unilateral measures.

The Current Status

On January 31, 2020, the Organisation for Economic Co-operation and Development (OECD) published a statement (the IF statement) announcing that the IF, which currently includes 137 member countries, endorsed the “Pillar One” approach to taxing the digital economy and approved an architecture for finalizing the Pillar One principles by the end of 2020.1 While much work remains, the IF’s endorsement of high-level Pillar One principles was a significant accomplishment, considering that this approach promises the most extensive changes to the international tax system in a century.2

The IF statement also sets forth an ambitious revised work plan for delivering a detailed consensus-based solution to Pillar One by the end of 2020. As identified in the IF statement, additional work remains in eleven key areas, some of which are fundamental to the operation of Pillar One.3 This work includes evaluating the feasibility and impact of the United States’ proposal to make Pillar One a “safe harbor” that MNEs could elect to apply rather than a mandatory feature of the international tax system.

In terms of timeframe, the IF statement sets the objectives for the IF to approve “the key policy features of a consensus-based solution to the Pillar One issues” by July 2020 and to deliver a final report on Pillar One that sets out the “technical details of the consensus-based solution” by the end of 2020.4 On a parallel track, the OECD continues to work on an economic analysis and impact assessment of Pillar One, the results of which will inevitably be a key factor determining whether the IF can meet its year-end objective of a consensus-based solution.5 After the publication of the final report, implementation by domestic legislation and a new multilateral instrument (MLI) would follow in 2021 and beyond.

Risk of Unilateral Taxation Remains Real

As discussed below, unilateral measures are already in force or soon will be in a number of countries. Notwithstanding the impressive progress that the IF has made toward a consensus-based solution, the risk of further unilateral digital tax measures remains very real. To start, there is an obvious risk that the IF will not result in a consensus-based solution by the end of 2020, which could open the floodgates to unilateral measures. While the IF’s recent agreement on the “architecture” of Pillar One demonstrates the IF’s ability to reach a high level of consensus quickly, reaching agreement on specific policy features and technical details of a consensus-based solution is a much taller order, given the diverse and sometimes diverging interests of the IF’s 137 members. Among possible hurdles, a failure by IF members to address the United States’ proposal to make Pillar One a “safe harbor,” or various countries’ sovereignty-related concerns related to arbitration-like binding dispute resolution, present considerable challenges to consensus.

Reaching consensus will depend largely on whether the taxes collected through Pillars One and Two will match taxes collected through DSTs and other unilateral measures. A country can be expected to give up a DST only if it can collect a similar amount of tax through measures enacted through Pillars One and Two. Accordingly, the OECD’s ongoing work on the economic analysis of Pillars One and Two is critical.

On February 13, 2020, the OECD presented a preliminary high-level assessment of the economic impact of Pillars One and Two. That report estimated that annual corporate tax receipts would increase by roughly four percent or $100 billion based on certain broad assumptions and estimates. Notably, less than twenty-five percent of expected revenue was estimated to come from Pillar One. This may not be enough to persuade countries to forgo unilateral measures.6

Further, even if a consensus-based solution is reached, there is a risk that not all countries will withdraw their unilateral measures as currently contemplated, or that the consensus-based solution could be interpreted in ways that allow for supplemental unilateral measures. In this regard, the IF’s statement indicates that “it is expected that any consensus based agreement must include a commitment by [the IF] members…to withdraw relevant unilateral actions and not adopt such unilateral actions in the future.”7 Although this is certainly a step in the right direction, the modifier “relevant” could open the door to an argument that a particular unilateral tax is not a relevant unilateral action that must be withdrawn.

In addition, if the IF were to adopt (or permit member jurisdictions to adopt) the U.S. “safe harbor” proposal as part of its consensus-based solution, then countries could presumably apply unilateral measures to those MNEs that do not elect into the Pillar One safe harbor. While MNEs not electing into Pillar One would effectively be choosing to subject themselves to unilateral measures, the availability of this option could nevertheless defeat the Pillar One objective of ensuring the withdrawal of unilateral measures.

Finally, even if the IF agrees to a consensus-based solution, it may not be enacted soon enough through domestic legislation or a multilateral instrument in enough jurisdictions to be effective. In such a case, unilateral measures can be expected to make a comeback.

Overview of Specific Unilateral Measures

EU Proposals

On March 21, 2018, the European Commission of the European Union (EU) presented two proposals for council directives introducing special tax regimes for digital companies. One proposal, which is intended as a long-term solution, would expand the permanent establishment (PE) concept to include “significant digital presence.” This expanded PE definition would allow a source state to tax nonresident companies with substantial business activity in the state, even absent a physical presence or an agent.

Specifically, a company would have a “significant digital presence” if:

  • it had annual revenues in an EU country in excess of €7 million; or
  • more than 100,000 users in an EU country accessed the company’s digital services in a year; or
  • it concluded more than 3,000 business contracts for digital services in a year.

The other proposal, the digital services tax (DST), is an interim solution conceptually intended to work as an excise tax levied on turnover. Specifically, the DST would apply to entities with consolidated worldwide revenues exceeding €750,000,000, of which at least €50,000,000 are qualifying taxable revenues (defined below) earned within the EU. Based on these thresholds, it appears that the DST is targeted to apply in particular to U.S. tech giants, although some EU multinationals should still fall within its scope.

The DST applies a three percent tax to taxable revenues, generally defined as gross revenues derived from any of the following activities:

  • the placing on a digital interface of advertising targeted at users of that interface;8
  • the making available to users of a multisided digital interface that allows users to find other users and to interact with them, and that may also facilitate the provision of underlying supplies of goods or services directly between users;9 and
  • the transmission of data collected about users and generated from users’ activities on digital interfaces.

In each case, the focus of the DST is on value generated at the location of the end user, who is deemed to be located in “the Member State where a user’s device is used […] by reference to the Internet Protocol (IP) address of the device, or, if more accurate, any other method of geolocation.” For example, users liking a page on social media will allow the social media operator to monetize their preferences for targeted advertising.

As noted, the DST is presented as an interim measure, but the current text does not contain a sunset clause. The DST would cease to have effect only when a long-term digital tax measure enters into force, at least among EU member states. However, the DST may remain in effect indefinitely as applied to non-EU countries, resulting in discriminatory treatment between EU and non-EU jurisdictions.

In March 2019, the EU member states failed to reach a unanimous consensus to implement the DST on an EU-wide basis. As a result, the European Commission decided to defer further consideration of potential EU-wide digital tax measures while the OECD seeks a global consensus-based solution.

Unilateral DSTs and Similar Measures

The EU member states’ inability to reach the consensus required to implement an EU-wide DST has led several EU member states to introduce their own initiatives for national taxes on digital companies. These unilateral DSTs and similar measures generally include “sunset clauses” and expire if an agreement on digital taxation is reached at the OECD or EU level. Even though most of these unilateral measures are based on the draft EU proposal, differences in the applicable rules may lead to double taxation and differing views of what transactions are covered.

In Europe, DSTs have been proposed, announced, or implemented by Austria, Belgium, the Czech Republic, France, Hungary, Italy, Poland, Slovenia, Spain, Turkey, and the United Kingdom. DSTs have also been proposed by numerous other countries beyond Europe, including Canada and Australia.

A summary of certain of these proposed DSTs appears below.

France

On July 11, 2019, the French Parliament passed the first unilateral DST among EU member states. The French DST applies a tax rate of three percent on revenues from certain services provided by digital means for which the contribution of the users is seen as key in value creation. Specifically, the French DST applies to:

  • intermediary services where a digital interface allows users to enter into contact with each other and to interact for the purpose of delivering goods or providing services directly between users, including:
    • sites or applications that allow users to meet each other (e.g., buyers and sellers, even when payment is not organized through the website);
    • services that allow interaction aimed at directly providing goods and services between users, such as marketplaces for the sale of goods or services; and
    • digital interfaces that provide reservation services or that collect sums on behalf of creators of applications and withhold a commission;
  • targeted advertisements, such as services sold to advertisers or their agents that are aimed at placing targeted advertisements on a digital interface based on the personal data of users who consult or interact with or on the digital interface. Services can consist of the sale, the storage and the distribution of the advertisement messages, the control of the advertisement, and performance measures as well as the sale and management of the personal data of the users. Such services include:
    • services that allow the optimization of advertisements through improvement of their performance (such as analysis of the views of the publication, etc.); and
    • services linked to the sale of users’ personal data (note that collecting data for internal purposes, such as for loyalty programs, is not in scope).
  • The French DST does not apply to:
  • non-targeted advertisements or services not rendered to advertisers;
  • direct online sale of goods and services (i.e., e-commerce);
  • digital interfaces that directly provide users digital content, communication services, or payment services, e.g., direct messaging services or online gaming services;
  • digital interfaces for financial services or that put buyers and sellers of advertisement spaces in contact with each other; and
  • digital interfaces used to manage any of a number of systems and services related to financial instruments, trading platforms and systems, and investment advisory activities.

The French DST applies to services provided in France, which in turn are determined by the use of a terminal (e.g., a computer, telephone) located in France. Whether a French terminal is used is determined in the first instance by the IP address and in the second instance by all other means, such as the physical address, Wi-Fi connection, or GPS coordinates of the user (subject to limitations under the General Data Protection Regulation).

This following secondary set of rules determines whether the service is provided in France in specific circumstances:

  • Intermediary services that allow the users to deliver goods or provide services are considered provided in France when the transaction is concluded by a user located in France (buyer or seller).
  • Intermediary services that are not targeted at delivering goods or providing services, e.g., where the buyer and seller perform the transaction outside the scope of the website, are considered to be provided in France if at least one user has, during the year, opened from France an account to access the digital services.
  • Advertisement services are considered to be provided in France if the message is published on a digital interface and is based upon personal data of a user that accessed the interface when located in France. It does not matter whether the user effectively viewed or clicked on the advertisement.
  • Sale of users’ data is considered to take place in France if the data sold is compiled from a user located in France.

The French DST applies to all companies, regardless of tax residence, that earn revenue from taxable services in excess of €750 million globally and in excess of €25 million in France. The revenue test is conducted on a consolidated, groupwise basis where the revenues of related entities are aggregated if certain ownership tests are met.

The tax liability under the French DST is based on revenue (excluding value-added tax, or VAT, amounts) from the payments in exchange for taxable services in France. Mechanically the tax amount is calculated by applying the following formula:

taxable revenue = worldwide amounts received x (number of users in France / total amount of users)

 

Similar to the EU DST, the French DST is a gross basis tax that applies without regard to whether the in-scope company is profitable or operating at a loss. The French DST would be deductible against French corporate income taxes but not against other taxes.

Italy

On October 15, 2019, the Italian government followed France in enacting a DST effective January 1, 2020. The Italian DST would apply to foreign and Italian companies that, individually or as a group,

  • record total worldwide revenues equal to or greater than €750 million; and
  • obtain total revenues from domestic digital services equal to or greater than €5.5 million.

Revenues subject to the DST include those derived from advertising services, intermediation and marketplace, and data transmission. For this purpose, “advertising” refers to the placing of at least one advertisement on a digital interface, targeting users of that interface; “intermediation and marketplace” refers to those platforms that offer a multilateral digital interface allowing users to contact and interact with each other and facilitating the direct supply of goods or services; and “data transmission” refers to the transmission of data collected by users and generated by the use of a digital interface. The DST also taxes transactions carried out in the marketplace, including intermediation in the sales of goods, whereas transactions concluded directly with final consumers and pure e-commerce transactions still seem to be out of scope. Business-to-business transactions relating to digital services are excluded from the scope of the tax.

Revenues subject to taxation are generally linked to the location of the users of the services; revenues are considered taxable if the user of a taxable service is located in Italy during a specific tax period. The localization rules vary by the type of service. Like the EU DST, the Italian DST will apply at a rate of three percent and will be based on revenues generated each quarter.

United Kingdom

The U.K.’s DST is set to become effective in April 2020, although observers believe that international pressure (including the negotiation of a U.S.–U.K. trade agreement) may defer its entry into force. The U.K. DST is a two percent tax on the revenue earned from social media platforms, online marketplaces, search engines, and related advertising businesses. The tax is directed at large players in the digital space, namely businesses with revenues above GBP 500 million of which at least GBP 25 million is earned from U.K. users.

In proposing the tax, the U.K. government noted that it “believes in international co-operation regarding the future of tax on large international businesses. The DST, and progress made in international discussions on the future of corporation taxes, will be reviewed in 2025. The outcome of that review will be reported to Parliament.”10 This statement is consistent with the U.K. government’s view that it will repeal its DST if a consensus on Pillar One is reached, provided that the consensus achieves a result acceptable to the United Kingdom.

Spain

On February 18, 2020, the Spanish government approved a bill for a DST for discussion by the Parliament. The current government is very likely to pass the bill, since it has a majority consensus on the new tax with the other political parties. Unlike Spain’s initial DST proposal, collection of the tax would be deferred until the end of the year. In 2019, the government estimated that the new DST would generate approximately €2 billion in revenues, although the Spanish Independent Fiscal Authority has since indicated that this sum was overestimated.

The Spanish DST would apply to Spanish and foreign entities with 1) total net revenues during the prior calendar year exceeding €750 million and 2) total revenue derived from activities subject to DST in Spain exceeding €3 million. In the case of corporate groups, these revenue thresholds would be applied at the group level.

With the Spanish DST, taxable revenues, which would be subject to a three percent tax, include revenues arising from 1) online advertising services; 2) data transmission services; and 3) intermediation services. Digital services for purposes of the DST are deemed as having been carried out within Spanish territory whenever the user is located in Spain. Specific rules for identifying locations of users apply for each type of taxable digital service.

The Spanish DST bill acknowledges that these rules are intended to apply only until the European DST directive is finally approved, after which the local legislation would be amended accordingly.

Advertising-Type Taxes—India’s Equalization Levy

In 2016, India enacted a six percent tax (the “equalization levy”) on B2B payments for online advertising. The equalization levy applies when an Indian resident advertiser (including a taxable PE held by a nonresident) makes a payment to a non-Indian supplier for an online advertisement, the provision of digital advertising space, or any other service or facility that enables online advertising. At its most general, the equalization levy would apply to an Indian taxpayer buying advertising space on a digital marketplace where the owner of the digital marketplace is not otherwise taxable on the advertising revenue. The equalization levy became effective on June 1, 2016, and by its terms applies broadly to any digital advertising services, although questions remain as to the scope of services subject to tax.

The equalization levy is collected through a withholding mechanism where the Indian advertiser is required to withhold and remit the tax periodically. In this manner, the Indian government effectively put the enforcement mechanism in the hands of the Indian advertiser. As the withholding agent, the Indian advertiser is secondarily liable for under-withheld tax. As a result, an Indian advertiser is likely to take a conservative position on whether a payment is subject to withholding.

Value-Added Taxes

Much as U.S. states have responded to Wayfair,11 many countries have already modified their local VAT regimes to account for digital sales. In most cases, online retailers are now required to register and collect VAT on online sales to local customers. Some countries apply a sales threshold below which registration is not required, but in many countries registration is required on the first dollar of online sales.

What Is So Bad About Unilateral Measures?

DSTs and other unilateral digital tax measures present numerous well-known problems. Compliance with DSTs will be inherently challenging and burdensome, since they will require MNEs subject to DSTs to design and implement new systems to track user activity by IP address and/or to develop alternative ways to identify the geolocation of end users.

Furthermore, the increased prevalence of virtual private networks (VPNs), which protect data privacy by disguising the IP address and location of the end user, will make enforcement and administration of DSTs particularly challenging. Indeed, the most recent Global Web Index identified that an average of twenty-six percent of internet users globally used a VPN in the past month (including eighteen percent in Europe and North America and thirty percent in the Asia-Pacific region).12 VPN usage will likely result in a significant portion of DST-compliance location tracking being incorrect.

DSTs may also add imbalance in the split of taxable revenues among jurisdictions. Indeed, in jurisdictions with higher purchasing power, users generally own multiple interconnected devices (e.g., smartphones, tablets, laptops, multimedia hubs), which may in turn lead to double-counting when one user, for example, streams content initially read on one device to another. In such cases, an advertisement may appear to the user only once, before the content is played, but may be counted twice, since the content is digitally transferred to another device with another IP address. The compatibility of IT systems that potentially must be designed for DST purposes would also have to be assessed vis-à-vis the (new) privacy regulations in place in many jurisdictions. Further, unlike Pillar One, which would have general applicability to large numbers of in-scope MNEs, many DSTs would apply more narrowly to only a handful of digital companies. DSTs discriminate particularly against foreign companies providing digital services in the enacting jurisdiction. For example, it has been estimated that the French DST would affect only twenty-seven companies, only one of which is French (and seventeen of which are U.S.-based MNEs).13 In addition, DSTs would appear to benefit the larger enacting jurisdictions, which by nature have more extensive user bases than smaller member states. To the extent that DSTs are determined to be deductible against corporate income taxes, smaller enacting jurisdictions would potentially lose part of their tax revenue, whereas larger enacting jurisdictions would likely benefit from the additional DST income.

In addition, DSTs have inherent shortcomings as gross basis taxes. Because DSTs apply to revenue, not net profit, DSTs could impose a significant financial burden on companies in a loss position. As a result, DSTs may have a disproportionate adverse impact on high-investment and low-margin businesses, and thus run contrary to the EU’s and other governments’ policy aims to provide for growth-friendly tax rules. For this reason, DSTs are not generally considered taxes on “income,” and therefore may not be creditable under either domestic law (Internal Revenue Code Section 901 in the United States) or income tax treaties. In fact, the United Kingdom’s DST was designed specifically not to be an income tax so it would not be subject to the nondiscrimination rules of the U.S./U.K. treaty.

Thus, DSTs are likely to result in significant double taxation of income already subject to corporate income tax in other jurisdictions. Furthermore, as purely unilateral domestic measures, disputes regarding the imposition of DSTs—including their double-tax effect—are likely outside the scope of both the Mutual Agreement Procedure of existing treaties as well as the dispute resolution provisions of any MLI that may be implemented as a result of Pillar One. Thus, MNEs facing disputes with respect to the imposition or calculation of DSTs will likely have no avenue for challenge other than domestic administrative procedures and litigation.

Furthermore, while only a limited class of digital companies will likely be directly subject to DSTs, such taxes may have a ripple effect throughout the entire economy. Costs of DSTs will likely be passed to consumers. Moreover, it can be expected that governments that oppose DSTs may respond with trade policies that impose costly, retaliatory measures on governments implementing DSTs, just as the United States continues to threaten through its Section 301 investigation of the French DST.

What Can MNEs Do Now?

Given the inherent problems of unilateral measures and the risk that they may still proliferate and become the international norm, what can MNEs do now to prepare?

First and foremost, MNEs can and should stay engaged in the Pillar One process. Although Pillar One may seem undesirable to the extent that it increases tax costs, compliance burdens, and the risk of disputes, a consensus-based solution by the IF on Pillar One is nevertheless the most realistic path to eliminating or mitigating the impact of uncoordinated unilateral measures. For this reason, MNEs may find it in their best interest to participate in the process and to provide input that helps the IF attain a consensus-based solution that is as modest in scope and as administrable as possible. It is particularly important for MNEs to make their views known to the OECD, Treasury, and other local country tax administrations in the small window of time that remains before the full IF meets in July 2020.

Second, given the current level of uncertainty with respect to the future of digital taxation, internal messaging by corporate tax departments is of paramount importance for corporate tax departments. At the moment, MNEs need to be prepared for outcomes ranging from the full implementation of Pillar One and the withdrawal of all unilateral measures to a complete failure of Pillar One and the widespread proliferation of unilateral measures (not to mention an infinite number of possibilities in which Pillar One and at least some unilateral measures will coexist). Given the range of possibilities, it is important for corporate tax departments to keep key stakeholders from being blindsided and to secure the resources needed to comply with the future of international taxation in the digital era—whether that future ultimately means Pillar One, a host of different unilateral measures, or some combination thereof.

Third, unilateral digital tax measures, if enacted, will likely result in a new wave of controversies. To prepare, MNEs would be well served to follow time-tested best practices. For example, to the extent that an MNE cannot track with certainty the location of its end users by IP address or other means, the MNE could likely reduce its future exposure by investing in robust recordkeeping systems that provide an audit-ready trail of its attempts to comply. Similarly, rationales for uncertain positions taken with respect to uncertain digital tax measures should be documented up front. Further, although DSTs likely fall outside the scope of all income tax treaty provisions, including the exchange of information articles, the possibility that tax administrations will find a way to cooperate and share information in the administration of their unilateral measures cannot be dismissed. Accordingly, MNEs should assume that information provided to one tax authority with respect to a unilateral measure will be made available to other tax administrations, and thus should avoid taking inconsistent positions with respect to unilateral measures in different jurisdictions.


Jason Osborn is a partner in the Washington office, Michael Lebovitz is a partner in the Los Angeles office, and Astrid Pieron is a partner in the Brussels office of Mayer Brown.


Endnotes

  1. Statement by the OECD/G20 Inclusive Framework on BEPS on the Two-Pillar Approach to Address the Tax Challenges Arising from the Digitalisation of the Economy (January 2020), www.oecd.org/tax/beps/statement-by-the-oecd-g20-inclusive-framework-on-beps-january-2020.pdf. Hereafter “IF statement.”
  2. Notably, Pillar One would create a new nexus for market jurisdictions to tax multinational enterprises on a deemed residual profit on in-scope activities (Amount A), less a fixed return for routine marketing and distribution functions taking place in that country (Amount B). Nexus to tax Amount A would generally be determined by reference to whether the MNE has a “sustained and significant” involvement in a jurisdiction (likely based on in-scope revenue over a period of years), without regard to whether the MNE has a permanent establishment. Furthermore, the measurement of Amount A would be determined by a to-be-agreed-upon formula, which is not intended to comport with the arm’s-length principle. Adjustments between Amount A and Amount B and other appropriate adjustments (collectively referred to as Amount C) would be made through a new, binding dispute resolution mechanism to ensure proper allocations of Amounts A and B to each country and to allocate additional profit to a country where local functions exceed the baseline amount compensated under Amount B.
  3. IF statement, Annex A, at 24.
  4. IF statement, Annex A, at 22-24
  5. On February 13, 2020, the OECD provided an update on its economic analysis and impact assessment which preliminarily shows that most tax jurisdictions, except certain “investment hubs,” can expect an increase in corporate income tax revenue from the adoption of Pillar One and Pillar Two, but the specific impact (including impacts on individual countries) will depend on the ultimate Pillar One and Pillar Two design. See www.oecd.org/tax/beps/webcast-economic-analysis-impact-assessment-february-2020.htm.
  6. It is worth noting in this regard that the United Kingdom’s new digital services tax (discussed further below) is expected to raise GBP 400 million annually once fully phased in.
  7. IF statement, ¶9.
  8. For advertising revenue, the users shall be deemed to be located in a specific member state when “the advertising in question appears on the user’s device at a time when the device is being used in that member state [in that tax period] to access a digital interface.”
  9. For so-called “multisided digital interfaces,” the place of taxation is deemed to be the member state in which a user uses a device to access the digital interface and to conclude an underlying transaction on that interface (applicable to multisided digital interfaces facilitating the provision of supplies of goods or services directly between users), or, for “other” interfaces, the member state in which a device was used to create an account on said interface.
  10. Draft Clauses for [U.K.] Digital Services Tax, Explanatory Notes, https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/816359/Digital_Services_Tax_-_ENs.pdf.
  11. South Dakota v. Wayfair, Inc., 585 U.S. ___ (2018).
  12. Global Web Index, “VPN Usage Around the World,” 2018, https://blog.globalwebindex.com/chart-of-the-day/vpn-usage-2018/.
  13. See United States Trade Representative, Report on France’s Digital Services Tax Prepared in the Investigation under Section 301 of the Trade Act of 1974, at 27 (December 2, 2019), https://ustr.gov/sites/default/files/Report_On_France%27s_Digital_Services_Tax.pdf.

Leave a Reply

Your email address will not be published. Required fields are marked *

XHTML: You can use these tags <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>