The Point Logo

Blame Canada

Posted: Jul 8, 2024


Canada’s Digital Services Tax

Earlier this month, Canada enacted a long-promised digital service tax (DST). The DST would apply an excise tax of 3 percent on the gross revenue from digital exports to Canada for firms with revenue over 750 million Euros. It would also be retroactive to 2022.

In general, taxes impose a burden on society greater than the revenue raised. This is the “excess burden” from taxation, and it stems from the distortions in behaviors that arise in response to a given tax policy. For example, sales taxes incrementally discourage sales of a given good by making it more expensive, and thereby reduce demand for the good. That taxes distort behavior and give rise to efficiency costs is not in dispute – the relative magnitudes of these effects, on the other hand, are the subject of significant debate among economists and policymakers.

There are principles of tax policy design that can make a tax more or less efficient. High tax rates on a narrow tax base, for example, tend to distort behavior more than lower rates on broader bases. Indeed, the basic theory of deadweight loss in taxation observes that the excess burden increases as the square of the rate. So if a tax rate doubles, the associated deadweight loss will quadruple. Sometimes, distortion is the point of the tax, such as taxes on cigarettes. But note that any tax high enough to significantly discourage an activity like smoking will necessarily come at the expense of revenue considerations.

Efficiency can be in tension with other policy goals, such as distributional considerations: how a tax will affect taxpayers up and down the income distribution. Other key elements of a tax include timing. A retroactive tax can’t affect past behavior (though it can inform future behavior). Somewhat more abstracted from pure tax policy design is the legal framework in which it is enacted. The U.S. Constitution effectively forbade income taxes until the ratification of the 16th Amendment – the implications of which continue to be refined and tested.

In terms of efficiency, timing, and legality, Canada’s DST is poor tax policy. The 3 percent levy is applied to gross revenue sourced from certain kinds of digital services. Specifically, tax would apply to four broad revenue streams: online marketplaces, online advertising, social media services, and user data. Firms with global revenues in excess of 750 million euros and with at least 20 million in revenues sourced to Canada from the four digital revenue categories would be subject to the tax.

The first strike against the Canadian DST is in its tax base. The tax applies to gross revenue, rather than net income or profits. This tax base excludes expenses, meaning that the tax applies even if the underlying business activity is not profitable. To the extent that the revenue reflects an input to a business such as advertising or data collection, the tax may applied several times throughout the business’s supply chain, with no system for crediting tax already paid. Accordingly, taxes on gross revenues can lead to high effective tax rates if they are applied multiple times, known as tax pyramiding.

The timing of the tax is also unusual in its retroactivity. In general, certainty is an important principle of sound tax policy, and is broadly recognized as such in international tax fora. DSTs in general require new record-keeping, reporting, and audit obligations. For novel tax policies such as DSTs, some of these obligations are untested. Retroactivity presumes these obligations can be met immediately, which may or may not be feasible. Both the tax itself and the associated compliance burdens were necessarily not known to firms in the past, and they made business decisions unmindful of those considerations. If firms had known how the DST would apply, they might have made alternative business choices.

Lastly, it is important to note that peripheral to the proliferation of unilateral DSTs, such as Canada’s, is an ongoing international process to apportion taxing rights for digital services. Under prevailing tax law, nations do not have taxing rights on foreign firms unless they have a physical presence in that jurisdiction, or “nexus.” Unilateral DSTs violate that principle. The United States is currently a party to a multilateral negotiation to devise new mechanisms for taxing digital services – known as Pillar One. This new mechanism would formalize the apportionment of new taxing rights on digital services sourced to foreign jurisdictions. U.S. participation in the Pillar One process had been predicated on the elimination of unilateral DSTs. Despite that ongoing process, there is a risk that DSTs will remain. Given how the digital economy is substantially dominated by U.S., these taxes are also highly discriminatory, and may violate treaty and trade agreements.

There should be no mistaking the goal of DSTs generally, and the Canadian DST’s in particular. It violates major precepts of tax design norms, is applied retroactively, and runs afoul of international norms. As a tax policy it is poorly designed; as an outright theft of the U.S. tax base – it’s perfect.