Economics of data (Part 5: Tax policies)

Economics of data (Part 5: Tax policies)

Phuah Eng Chye (12 September 2020)

Data abundance is a prerequisite for and an outcome of scaling the levels of monetisation and exchange in society. But it also leads to the concentration of data ownership and the dominance of a handful of global platforms. This trend has broad implications for tax policies. The existing tax frameworks were built for a legacy industrial economy rather than a data economy. The cracks are showing up as fiscal revenues come under pressure from data-driven organisational change.

It is critical to re-align tax policies to the data economy. But what type of reforms are needed? In this regard, because platforms are ubiquitous and all kinds of new taxes have been proposed to correct a range of competition and general grievances including for privacy, income inequality and content.

Data taxes

Complaints that platforms are exploiting consumer data has given rise to proposals such as the data as labour or markets for data[1].  These concepts seek to make companies pay consumers for use of their data or at least to raise government revenues to finance redistribution. A recent proposal to impose data sales tax[2] is aimed at addressing inequalities and budget shortfalls. The four-step plan involves quantifying the personal data companies use for commercial purposes; requiring companies to register and disclose their transactions of that data; taxing those companies (i.e. through a transaction fee or as a percentage net revenues) and allocating those funds to state and local governments.

Ashley Johnson notes there are drawbacks. First, the two largest platforms do not sell data but rely on advertising income. Second, the price of data tends to be cheap. For these reasons, a data sales tax wouldn’t be able to raise sufficient revenues to fund budget shortfalls. In addition, a data sales tax could disincentive the provision of free online services or add additional layers of tax. He argues that it did not make sense for a tax to “only focus on internet data” when there are more valuable data transactions (such as online hotel or flight bookings or purchases).

There are also considerable computational challenges. Apart from the difficulties in defining and tracking data use and sales, there are valuation challenges. The value of the same types of data could vary from negative (spam) to valuable (high net worth individuals). In addition, the value of much data comes from joint contributions (leading to value attribution problems) while income from some data (e.g. influencers, gig workers) may already be subject to tax.

While it is true that concentration of data ownership funnels income from data into relatively few private hands, nonetheless it may be best conceptually to separate the data challenges from the income challenges rather than lump them together. In this regard, data taxes may have similar effects to a financial transactions or Tobin tax – i.e. it would reduce the volume of data collected. While this is intended to discourage irresponsible uses of data, it will also weaken the incentive to innovate and reduce welfare and income without necessarily addressing national inequality.

Data ownership concentration challenges should thus be addressed by policies to redistribute data. Widening ownership of data will generally result in the transfer of the benefits of income from the private to the public domain. Similarly, income concentration or inequalities should be addressed by macro policies to widen income distribution – such as through basic income or restructuring national savings and welfare schemes – or through private sector alternatives such as platform cooperatives[3].

Taxes on robots and automation

Data-driven technologies have been a major factor reducing the labour share of national income. Daron Acemoglu, Andrea Manera and Pascual Restrepo argues “the US tax code systematically favors investments in robots and software over investments in people…The result is too much automation that destroys jobs while only marginally improving efficiency…heavy taxation of labor and low taxes on capital encourage firms to automate more tasks and use less labor than is socially optimal”.

Daron Acemoglu, Andrea Manera and Pascual Restrepo note that “compared with effective labor taxes that stand above 28.5 percent, the effective tax rate on capital invested in equipment and software has declined to about 5 percent today, largely as a result of favorable depreciation provisions”. Their analysis suggests “eliminating the capital bias in the tax code would raise the number of employed people by 6.5 percent and increase labor’s income share by 1.1 percentage points.  More-modest reforms could still increase employment by 1.9 to 2.9 percentage points”. Optimally, “labor should be taxed at 9 percent and capital at 22 percent”.

Ryan Abbott and Bret Bogenschneider expressed concerns that “robots are not good taxpayers…automation significantly reduces the government’s tax revenue since most tax revenue comes from labor-related taxes”. In most developed nations, “the bulk of taxes are currently remitted by workers either through wage withholding, taxation of labor income, or indirect taxation of workers as consumers. Since robots are not subject to these types of tax regimes, automation reduces the overall tax base”. “If all workers were to be replaced by machines tomorrow most of the tax base would immediately disappear. As a matter of taxation, automated workers represent a type of capital investment, and capital income is currently taxed at much lower rates than labor income”. “When a machine replaces a person, the government loses a substantial amount of tax revenue – potentially hundreds of billions of dollars a year in the aggregate… changes to tax policies are necessary to account for the loss of government tax revenue due to automation. This is particularly critical because the education and social benefit reform necessitated by automation will only be possible with more, not less, tax revenue”.

Robot taxes are increasingly being contemplated. Ryan Abbott and Bret Bogenschneider notes “in February 2017, the European Parliament rejected a proposal to impose a robot tax on owners to fund support for displaced workers, citing concerns of stifling innovation…Bill Gates stated…governments should tax companies’ use of robots to slow the spread of automation and to fund other types of employment…In August 2017, South Korea announced plans for the world’s first tax on robots by limiting tax incentives for automated machines” – i.e. by reducing the tax deductible for investments in automation equipment by up to two percent.

Ryan Abbott and Bret Bogenschneider propose tax policies should aim to, at least, achieve tax neutrality “between human and automated workers” to avoid tax distortions and other negative effects and to lower administration and compliance costs, promote distributional fairness, increase transparency and broaden the tax base with lower rates. They suggest “this could be achieved through some combination of disallowing corporate tax deductions for automated workers, creating an automation tax which mirrors existing unemployment schemes, granting offsetting tax preferences for human workers, levying a corporate self-employment tax, and increasing the corporate tax rate”.

In my view, technology adoption and labour employment should not be viewed as competing trade-offs. Society needs to progress along the technological path. Imposing data- and automation-taxes can lead to a cascading of costs and impair future competitiveness. Hence, tax reform may be needed to rebalance the tax burden between labour and capital. But rather than penalise investments in technology (robots or data), the rebalancing can be achieved by paring deductibles, closing loopholes, increasing corporate tax rates and penalising concentration directly.

Cross-border digital taxes

Datafication of economic activities is having a negative impact on fiscal revenues at the local and national levels. Global platforms have emerged as the main vehicle for organisational disruption; facilitating the digitalisation of physical products and the bypassing of traditional intermediaries and physical stores. This has put downward pressure on the prices of physical products and squeezed related tax revenues.

Aqib Aslam and Alpa Shah notes “seeing that large highly-digitalized multinational enterprises are paying minimal tax in the jurisdictions in which they provide services, policymakers are increasingly more sensitive to the growing inadequacies of the current international corporate income tax system when it comes to generating a sufficient level of tax revenues from these businesses. The debate on international taxation has now coalesced around whether and how governments should be taxing these businesses and what the appropriate distribution of those revenues across countries should be. Furthermore, the rapid growth of digital service providers over the last decade has made them an increasingly popular target for special taxes – similar to wealth and solidarity taxes – which can also help mobilize much-needed revenues in the wake of a crisis”.

Several governments, particularly the EU members, are considering imposing digital service taxes (DSTs). This is controversial due to the cross-border implications. Joe Kennedy points out DSTs will likely only apply to companies with significant global operations and as a result will mainly target the large US internet firms while avoiding the domestic firms. He notes the targeting of US firms is attractive as “an easy source of money” and it is politically attractive as it makes it easier for domestic digital firms to compete with American firms.

Joe Kennedy disputes the logic that “a fair proportion of companies’ revenues should be subject to taxation where their users reside” is faulty. He argues DSTs are discriminatory as it exempts “many similar domestic companies”. Hence, “DSTs are trade barriers that violate international trade agreements. These agreements are premised on the principle that imported goods and services should receive the same treatment as domestic ones. Although on the surface DSTs offer equal treatment to all firms, distinguishing only according to their size and business model, most of the burden falls on a small set of foreign companies, mostly based in the United States. Structurally, the decision to tax these revenues makes the practice equivalent to an ad valorem tax on imports, something that violates WTO rules”. At the moment, the US has threatened retaliatory actions against countries imposing DSTs and discussions are still ongoing.

The dilemma is greater for developing countries. The Pathways for Prosperity Commission note “developing countries represent a large share of digital services’ user base, but are unable to collect taxes from their profits. For example, almost 1.4 billion people in developing countries are Facebook users, representing almost 70% of active users worldwide (although they account for a smaller share of global revenue). It is common to see firms for which the only parts of their business that exist in a developing country are their customer base and a facility to receive payments. It is still possible to tax the transaction when money changes hands (several countries apply their regular goods-and-services consumption tax to digital goods), but the profit and the rest of the business remain abroad. This is not a problem if we assume that the product is created entirely in a foreign country and merely imported whole, but that is not necessarily the case with digital services that, for instance, rely on local data. As a result of such tax arrangements, technology companies often fail to contribute a fair share to national revenues, fuelling further economic inequality, and limiting funds available for education, health, and infrastructure”.

“Developing countries have – understandably – been implementing measures to try to capture some of the wealth generated by digital transactions, but this has caused considerable controversy”. The Pathways for Prosperity Commission note “in 2018, Uganda introduced a social media tax, which charged users 200 Ugandan shillings (around US$0.05) per day for the use of a number of internet applications, including popular services such as Facebook, Twitter, WhatsApp, and Instagram. The new tax adds up to about US$1.50 per month or US$19 per year, in a country where many people live on less than US$1 a day. In the period after the introduction of the tax, data use and mobile money transactions decreased in Uganda, and internet user penetration dropped from 47% to 35%…the Uganda Communications Commission (UCC) announced that, following the imposition of the social media tax, the number of over-the-top (OTT) subscriptions had declined by more than 2.5 million in the last quarter of 2018. The tax also disproportionately affected marginalised users – the cost of the social media tax represents 2.4% of average individual income in metropolitan Kampala, but 22.6% of the average individual income in rural Bukedi.

“India implemented an equalisation levy on cross-border digital advertising in June 2016. The 6% levy applies to payments made by companies based in India to a foreign company for online advertisements, if the annual payments exceed Rs.100,000 (approximately US $1,450) in one financial year. The levy, which is only applicable to cross-border business-to-business (B2B) transactions, is withheld at the time of payment by the purchaser of the services (ie the Indian firm hiring the advertisement services), and subsequently paid to the government. The measure is controversial because it puts an extra burden on local firms using foreign platforms for advertisements, and is especially heavy for startups. However, it has contributed to an increase in tax revenue: the Indian government reportedly collected approximately US$76 million between 2017 and 2018 through the equalisation levy. A government committee has been analysing measures for other types of cross-border digital transactions, but the equalisation levy has not yet been expanded to other sectors. However, the Indian government has considered other measures to tax digital transactions, such as the introduction of a significant economic presence (SEP) concept, which would allow the government to tax income of foreign companies based on their virtual economic presence”.

The Pathways for Prosperity Commission conclude “reforms are required in international taxation to ensure that developing countries share in the benefits of global technological progress in an inclusive way. Current tax treaties prohibit the taxation of business profits of companies without a physical establishment in a country”. “Developed countries – through fora such as the G7 and the Organisation for Economic Co-operation and Development (OECD) – are starting to respond to this problem. The OECD’s recommendations include adopting the concept of a non-physical taxable presence, and efforts to identify and define income derived from a particular source in a jurisdiction. Another measure could be a global tax. This would tax multinational enterprises on their global income at a minimum rate, regardless of where they are headquartered, and distribute the revenue according to the proportion of the profits generated in each country”.

Legal reform

Aqib Aslam and Alpa Shah notes “policy proposals have been put forward which seek to limit the scope of tax avoidance and tax competition” but face legal definitional and computational challenges. First, the internet “facilitated a surge in remote cross jurisdictional sales, decoupling economic and physical presence…displacing traditional physical stores…This has strained the traditional concept of permanent establishment, which relies on a fixed physical presence as a precondition for governments to exercise their right to tax”. The permanent establishment requires a “significant economic presence” above a threshold “typically specified in domestic legislation and/or double tax treaties”. Aqib Aslam and Alpa Shah explain “while activities such as local data collection or warehousing were previously considered to be of a merely auxiliary nature…they now form core elements of digitalized business models. In other words, previously simple routine functions for which it was generally acceptable to allocate only a small share of the overall business profits have arguably become key activities for many firms, with a pivotal role for data analytics…this surge in cross-border business-to-consumer sales as a result of digitalization has taken us into a grey area”.

Aqib Aslam and Alpa Shah points out the OECD acknowledges “the current definition of permanent establishment would need to be modified to establish the right for user-jurisdictions to tax multinational enterprises that collect data on their citizens…Some countries have also already taken steps to change the permanent establishment threshold tests in domestic legislation to include” online retailers, warehouses, and distribution networks. In June 2018, the US Supreme Court ruled that “physical presence should not be required for a state to compel out-of-state sellers to collect state sales tax on sales to customers in their state”.

Second, Aqib Aslam and Alpa Shah notes multinational enterprises derive “a larger share of their value from intellectual property that is both easy to shift across borders and hard to value for transfer pricing purposes…frustrated the arm’s length principle and exacerbated opportunities for profit shifting”. The current system of arm’s length pricing and the growing importance of near-impossible-to-value intangibles – including user data – has allowed opportunities for corporate income tax avoidance to proliferate. As a result, many 52 countries are seeing profits shifted out of their reach by companies using transfer pricing tactics, while those without any permanent establishment-based taxing right are missing out on the opportunity to tax these companies altogether”.

There have been proposals focusing on “marketing intangibles” – such as brands, goodwill, trademarks and user data – as a basis for claiming more taxable income. This requires rights to be assigned to the market country as the source of value and a formula to “allocate profits to the various newly empowered jurisdictions”. Multinational companies contend the application of “marketing intangibles to extend taxing rights” would move a disproportionate share of profits away from research and development towards populous end markets. There are also risks governments may discriminate by targeting and ringfencing certain types of businesses.

Third, while online users are the critical driver of the value of digital services, Aqib Aslam and Alpa Shah points out “user participation is not recognized under the existing international tax framework as a source of taxable value. As a result, the blurred line between their role in both supply (production) and demand (sales) has…opened the door to an important conversation about the concepts of source, destination, taxable presence and profit attribution”.

Aqib Aslam and Alpa Shah points out there are policy proposals that attempt “to pre-determine a distribution of taxable profits across countries…predicated…on the idea that the user has a role to play in value creation, justifying the designation of source-based taxing rights to the jurisdiction in which they are located”. Some “countries have begun to implement user-based turnover taxes, targeting specific digitalized industries and activities. However…recent proposals have tended to restrict the scope of special tax treatment, singling out activities or business models which are seen as particularly data intensive…the collection and use of potentially monetizable information is so pervasive in today’s economy, that drawing a line between cases in which users are and are not material contributors is inevitably fraught. It would seem that if user data is indeed being exploited at a scale large and wide enough for it to be both recognized as an economic input to production and protected on behalf of the user, it should be recognized for all businesses”. However, it is a challenge to “design revised profit allocation rules which reflect user contribution…there are no available market-based prices or established benchmarks to guide how user data should be measured and valued…In other words, the issue of user data has added to the scale of the existing valuation issues posed by intangible assets, while also expanding the potential number of countries claiming the right to tax. Both issues are heavily politicized and influenced by countries jockeying for position”.

“With largely undefined property rights over personal data, and in the absence of legislation governing the collection of or access to data”, the practical approach is to consider governments as holding the “resources in trust on behalf of their citizens and collect payment for extraction in the form of government revenues…data on the behavior and preferences of a country’s citizens would be seen as a collective national asset with compensation for its use payable to the government. Indeed, a government is better placed to exert collective power on behalf of individuals who cannot capture their rents by levying a tariff (for example, an export tax) on the extraction of data by foreign digital service providers”.


Overall, Aqib Aslam and Alpa Shah argue that “as businesses in almost every sector take advantage of the benefits of digitalization, the need to address the shortcomings and unresolved issues within the international corporate income tax and VAT systems has become that much more urgent”. In searching for potential solutions to remedy shortcomings in the international tax system, it is noted that “measures are not without their potential pitfalls, particularly if uncoordinated. Possible drawbacks include the creation of undesirable disincentives for investment, and the proliferation of double taxation and excessive tax burdens due to multiple royalty rates and bases, administrative complexity, and the risk of misclassified activities. The resulting variation in effective tax rates across jurisdictions could have a tangible impact on global resource allocation – all in return for potentially very little tax revenue. There are also political economy costs of retaliation. The absence of international consensus and a lack of clear data valuation methods have already pushed many countries to impose targeted turnover-based measures on varying subsets of digital services. In some cases, indirect proxies for user value are used as the tax base, for example, revenues from online advertising or from remote selling through online platforms, where the transaction value is observable. Other governments have side-stepped the valuation issue by basing their taxes on directly observable indicators, such as the volume of users. While these are all practical solutions, the risk remains that they do not sufficiently address the issue of taxing profits from value created – in this case by users. Furthermore, such an inconsistent approach could leave countries with a patchwork of overlapping measures that create more problems than they solve”.

Aqib Aslam and Alpa Shah concurred with the view[4] that “there should be no special tax regime for more digitalized companies”. “Attempts to design policies specific to a so-called digital economy are potentially flawed – it is generally better for policy makers to ignore the distinction and take a holistic approach, particularly when it comes to areas such as international taxation…general rules should be applied or adapted so that there is equal treatment for all businesses…the digital economy is increasingly becoming the economy itself”.


Ashley Johnson (8 June 2020) “A data sales tax would change the internet for the worse”. Information Technology and Innovation Foundation (ITIF).

Aqib Aslam, Alpa Shah (29 May 2020) “Tec(h)tonic shifts: Taxing the digital economy”. International Monetary Fund (IMF).

Daron Acemoglu, Andrea Manera, Pascual Restrepo (2020) “Does the U.S. tax code favor automation?” BPEA Conference Draft.

Hans-Jörg Naumer (30 November 2019) “How the Facebook cooperative could point the way for the network economy”. Voxeu.

Joe Kennedy (4 June 2020) “Two reasons digital services taxes are attractive – and five reasons they’re still wrong”. Information Technology and Innovation Foundation (ITIF).

Pathways for Prosperity Commission (2019) “Digital diplomacy: Technology governance for developing countries”. Oxford.

Phuah Eng Chye (18 July 2020) “Economics of data (Part 1: What is data?)”.

Phuah Eng Chye (1 August 2020) “Economics of data (Part 2: Market approach to valuing data)”.

Phuah Eng Chye (15 August 2020) “Economics of data (Part 3: Relationship between data and value and the monetisation framework)”.

Phuah Eng Chye (29 August 2020) “Economics of data (Part 4: The data economy)”.

Ryan Abbott, Bret Bogenschneider (2018) “Should robots pay taxes? Tax policy in the age of automation”. Harvard Law & Policy Review.

[1] See “Economics of data (Parts 1 & 2)”.

[2]Proposal by Andrew Gounardes and Eric Adams published in the Wall Street Journal. See Ashley Johnson.

[3] See Hans-Jörg Naumer.

[4] View expressed in the European Commission expert group on taxation of the digital economy (2014 report). See Aqib Aslam and Alpa Shah