Digital Services Taxes and Their Incidence — the Location-Specific-Rent Rationale and the Amazon Pass-Through

Summary

A digital services tax (DST) is a levy on the gross revenue a large digital platform earns from a jurisdiction's users — typically from online advertising, online intermediation (marketplaces), and the sale or transmission of user data. Since 2019 the UK (2%), France (3%), Austria (5%), Turkey (7.5%), Spain, Italy, India and others have enacted DSTs, largely as a stop-gap while the OECD/G20 negotiated the "Pillar One" reallocation of taxing rights. DSTs were explicitly motivated by the perception that a handful of platforms earn large returns from a country's users while paying little profit tax there — the European Commission's 2018 figure was an effective rate of 9.5% for digital businesses versus 23.2% for traditional firms.[1]

For a Geoist reference the DST matters for two connected reasons. First, the most careful theoretical defence of the DST — Wei Cui and Nigar Hashimzade's — rationalises it as a tax on location-specific rent, the digital analogue of a natural-resource royalty: this is a genuine attempt to capture a platform rent rather than to tax production.[2] Second, the best empirical study of an actual DST — Rohit Reddy Muddasani and Dominika Langenmayr's study of Amazon — finds that the UK DST was largely passed forward to third-party sellers and then to consumers, so the dominant platform bore only a small part of the burden.[3] The gap between those two findings is the point: a tax intended to fall on rent that is designed to fall on revenue does not behave like a rent tax. The DST is the wiki's cleanest worked example of a badly-aimed rent tax.

The Theory: a DST as a Tax on Location-Specific Rent

Cui and Hashimzade (2019) offer what they describe as the first affirmative economic rationale for the DST — noting that before governments proposed it "the DST had no intellectual proponent."[2] Their claim is that a dominant platform earns rent that is specific to a location (the users in a given country), even though the underlying intangible asset is mobile and non-rival, and that the traditional international income-tax paradigm — keyed to physical presence, source of payment, and profit allocation among related entities — systematically fails to assign that rent to the place it arises. Their framing is explicitly Georgist in structure:

"We offer a rationalization of the DST as a tax on location-specific rent (LSR). That is, just as many countries already levy royalties on rent from extracting natural resources, one can think of the DST as levied on rent earned by digital platforms from particular locations."[2]

This maps the DST directly onto the wiki's resource-rents template — a royalty on a rent tied to a jurisdiction — and Cui's companion paper argues the DST is best understood not as a protectionist tariff or a cascading consumption tax (the standard critiques) but as a rent tax with an efficiency and fairness case.[4] On this reading a well-designed DST would sit on the rent-capture side of the Geoist menu alongside spectrum auctions and resource royalties.

The Practice: the Amazon Pass-Through

Whether a real DST behaves like a rent tax is an empirical question about incidence, and the record is not encouraging. When the UK's 2% DST received royal assent in July 2020, Amazon announced within weeks that it would raise UK seller charges to cover it:

"from September 1, 2020, the fee types mentioned above will increase by 2%" — applied to referral fees, Fulfilment-by-Amazon fees, monthly FBA storage fees, and (from 15 September) Multi-Channel Fulfilment fees.[5]

Amazon had done the same in France a year earlier, raising French seller fees by 3% to match that country's 3% DST.[6] Google likewise added a 2% "DST Fee" to UK advertisers' invoices from November 2020, and comparable surcharges for Austria's 5% and Turkey's 7.5% DSTs — passing the levy on "at the same rate the government takes."[7] The Tax Foundation used the Amazon announcement as a textbook illustration that the statutory payer of a tax is often not the economic bearer: "just because a 2 percent revenue tax applies to large digital companies does not mean that the companies will bear the entire cost."[1]

Muddasani and Langenmayr (2025) put numbers on this using Amazon marketplace data across DST-adopting countries. Their finding:

"on average, Amazon increased its fees by roughly half the amount of the digital service tax. Firms using Amazon as a platform have largely passed these increased fees on to consumers. Large digital firms thus bear only a small part of the tax burden, but the tax may nevertheless succeed in making them less competitive relative to brick-and-mortar competitors."[3]

So on the best available evidence the dominant platform — the party actually earning the rent the DST was meant to reach — absorbed only a minority of a tax nominally levied on it, with the rest travelling down the two-sided market to sellers and shoppers.

Relation to the Georgist Case — the Gradient at Its Sharpest

This case belongs on the contested frontier of the rent gradient, and it is the sharpest single illustration of why the frontier is contested. The DST is a deliberate attempt to capture platform and data rents, and Cui and Hashimzade give it a rigorous rent-tax rationale. Yet its real-world incidence is almost the opposite of the land case the wiki treats as clean.

• Why land rent can't be shifted, but DST revenue can. An LVT falls on the value of a fixed-supply site; the owner cannot pass it forward because the quantity of land does not change and the tax is capitalised into a lower land price (the incidence result the wiki documents at length). A DST, by contrast, is levied on gross revenue in a two-sided market, not on the rent or the fixed factor, so it acts like an excise: the platform re-optimises its fees, and standard tax-incidence logic pushes much of the burden onto whichever side of the market is less elastic — here, sellers and consumers. Taxing revenue is not the same as taxing rent.
• The "is it even rent?" dispute is unresolved. Even granting Cui and Hashimzade's LSR framing, how much of a platform's return is location-specific rent versus the quasi-rent that rewards genuine product quality, logistics, and intangible capital is exactly the dispute the wiki refuses to flatten (see data-rents, Crouzet & Eberly's intangibles counter, and the steelman at taxing quasi-rents kills innovation). A revenue tax makes no attempt to separate the two, which is part of why it misfires.
• The honest Geoist lesson. The DST does not show that platform rents are uncapturable; it shows that a rent tax aimed at the wrong base behaves like an ordinary turnover tax. Muddasani and Langenmayr note a partial silver lining — the DST may still shift competitive balance toward brick-and-mortar rivals — but as a rent-capture instrument it is badly aimed. This is the empirical anchor for the wiki's standing warning never to let the airtight land case lend its certainty to the digital frontier.

Nuances and Limits

• Incidence is not zero-on-the-platform, and it is contested. Muddasani and Langenmayr estimate Amazon passed through roughly half, not all; the split between sellers and consumers, and how it varies by market elasticity, is model-dependent. Their study centres on Amazon's marketplace, the clearest case; advertising DSTs (Google, Meta) and data DSTs may have different incidence.
• DSTs were designed as an interim measure, not an optimal rent tax. They were explicitly a placeholder pending the OECD/G20 two-pillar deal; the Cui–Hashimzade LSR rationale is a reconstruction of a defensible logic, not necessarily the design intent. A tax deliberately built to capture location-specific rent might be structured very differently (e.g., on excess returns rather than gross revenue).
• The critics' file is real. The Congressional Research Service and others characterise DSTs as poorly-designed cascading taxes or as protectionist measures aimed at U.S. firms; the U.S. threatened retaliatory tariffs over France's DST. Cui and Hashimzade contest these readings, but the wiki should carry the DST as contested, not settled.[2][4]
• Source provenance. The Amazon and Google pass-through announcements are primary corporate communications; the incidence magnitudes come from a 2025 working paper (Muddasani & Langenmayr) that had been presented at CESifo, IIPF, EEA and OECD seminars but whose final peer-reviewed version should be re-checked before the numbers are hardened on any claim page. The LSR framing (Cui & Hashimzade) is a 2019 working paper.

See Also

• Korinek–Stiglitz: AI and income distribution — the AI-rents theory (non-reproducible factors absorb the gains; taxing them is non-distortionary) · DMA interoperability — the dissolve pole as legislated
• Platform and Data Rents — the contested frontier the DST tries to reach
• Resource Rents — the royalty template Cui & Hashimzade invoke
• Economic Rent · Quasi-Rent
• Land Value Tax — why a rent tax on a fixed factor can't be shifted, unlike a DST
• Objection: Taxing quasi-rents kills innovation
• Romer's digital advertising tax — a rival, revenue-based instrument aimed at the same firms
• Geoism — the rent-domain program and its gradient

Sources

1. European Commission (2018) effective-rate figure and the incidence framing are reported in Tax Foundation, "Who Will Ultimately Pay the UK Digital Services Tax? Amazon Passes the Cost Along to Sellers" (2020) — Tax Foundation. Used for the 9.5%/23.2% effective-rate contrast and the statutory-vs-economic-incidence point. (Policy institute; used for framing, not as the primary incidence estimate.)
2. Wei Cui & Nigar Hashimzade (2019), "The Digital Services Tax as a Tax on Location-Specific Rent," Allard School of Law (UBC) Faculty Publications / Durham University Business School, working paper (18 Nov 2019). Allard PDF — used for the location-specific-rent rationale, the resource-royalty analogy, and the verbatim rationalisation quote (C/D-claims; fetched and read this session).
3. Rohit Reddy Muddasani & Dominika Langenmayr (2025), "Navigating the Amazon: The Incidence of Digital Service Taxes," WU Vienna / KU Eichstätt-Ingolstadt / CESifo (12 June 2025). WU PDF — used for the empirical incidence finding (Amazon raised fees ~half the DST; sellers largely passed the increase to consumers; large firms bear a small share) and the verbatim abstract quote (B-claims; fetched and read this session; a working paper — re-verify the final published version before hardening figures).
4. Wei Cui (2019), "The Digital Services Tax: A Conceptual Defense," prepared for the NYU Tax Policy and Public Finance Colloquium (later published, Virginia Tax Review). NYU PDF — used for the argument that the DST is a rent tax rather than a protectionist tariff or cascading consumption tax, and for the survey of the critics' case (D/E-claims).
5. Amazon (2020), "Fee changes in the UK following introduction of the Digital Services Tax," Amazon Seller Central (UK). Seller Central — primary corporate announcement; used for the 2% UK seller-fee increase effective 1 September 2020 and the verbatim quote (A-claim).
6. Andersen LLP / Accountancy Daily (6 Aug 2020), "Amazon hikes UK seller fees in response to digital services tax." Andersen — used for the France 3% DST precedent (Amazon raised French seller fees 3% in 2019) and contemporaneous expert commentary (A-claim).
7. Simon Sharwood, The Register (2 Sept 2020), "Google, Amazon pass on UK Digital Services Tax by hiking ad prices, fees." The Register — used for Google's 2% UK "DST Fee" on advertisers (from Nov 2020) and the Austria (5%) and Turkey (7.5%) DST pass-throughs (A-claim; trade press).