Fiscal Policy / WORLD

Analysis: why global minimum-tax estimates are not collected revenue

Primary OECD documents and reproducible arithmetic separate historical Pillar Two projections from actual tax receipts and later 2026 implementation changes.

The source does not state the event date; the period shown is the source review period.

Reading sizeSources ↓Share story ↗
Conceptual illustration of public finance and resource allocation.Illustration
Symbolic illustration

A global minimum-tax revenue estimate answers a conditional question: how much additional revenue might a specified set of rules generate under a specified model? It does not answer how much governments have already received. That distinction matters for fiscal reporting, company analysis and any attempt to connect tax policy with bond markets. Newsverum’s review of primary OECD documents separates two historical estimates, later implementation material and the evidence that would be needed to establish actual collections. We make no new global revenue forecast.

The starting document is the OECD release dated 18 January 2023. It described expected additional annual global revenue of approximately USD 220 billion from the proposed minimum tax. Calling that amount an audited collection total changes both the basis and the time meaning of the figure. An expected annual effect is not a bank balance, a cumulative receipt or a measurement for November 2026. Our separate correction withdraws that earlier characterization; this analysis explains how to read the underlying numbers without recreating the error.

A January 2024 taxation working paper gives a different modelled annual corporate-income-tax gain: USD 155–192 billion, corresponding to 6.5–8.1% of global corporate income tax revenues. Its authors state that the distribution depends on implementation choices. The paper is research and carries a disclaimer that its views do not necessarily represent the OECD or member governments. We retain that attribution rather than presenting every result as an institutional finding about realized fiscal outcomes. The range is an analytical result, not a confidence interval identified by this review.

Our reproducible arithmetic starts with the published endpoints. Subtracting 155 from 192 gives a USD 37 billion span. Their arithmetic midpoint is USD 173.5 billion. Neither operation adds evidence: the midpoint is not an OECD point forecast, and the span is not an independently calculated margin of error. Readers can reproduce these results with decimal arithmetic using the accompanying inputs. We preserve the endpoints because replacing a range with its midpoint can hide the variation the source deliberately retained.

Comparing the older USD 220 billion figure with the later endpoints yields numerical gaps of USD 28–65 billion. Relative to 220, these are approximately 12.7–29.5%, rounded to one decimal place. Those comparisons describe different analytical vintages. They do not measure lost collections, a policy failure or a decline in a government’s receipts. Treating the gaps as a time series would silently assume that the two models share identical inputs, scope and assumptions. This review does not establish that equivalence, so it draws no fiscal-performance conclusion from the subtraction.

The percentage range requires equal care. A gain expressed as 6.5–8.1% of corporate income tax revenue uses that revenue as its denominator. The 15% minimum effective tax rate uses a tax base defined by the rules. These percentages cannot be substituted for each other. A tax rate, a percentage change in tax revenue and a share of a revenue total answer different questions. An analyst needs the denominator and the period before comparing them, just as a Markets reader needs the currency pair and observation date before interpreting an exchange-rate move.

The implementation documents also prevent a simple worldwide-enactment narrative. The OECD’s current minimum-tax overview describes the GloBE rules as a common approach: jurisdictions need not adopt them, but those choosing implementation agree to consistency with the agreed outcomes and rule order. An international agreement therefore does not by itself document identical domestic law, collection capacity or commencement dates in every jurisdiction. Our inference is that a credible country comparison must identify the applicable domestic instrument and effective period before assigning any estimated gain to that country.

Policy design continued after the historical estimates. The Side-by-Side package records approval and declassification by the Inclusive Framework on 5 January 2026. The OECD’s announcement describes simplifications, safe harbours, a stocktake process and the continuing role of qualified domestic minimum top-up taxes. Agreement on that package is a documented policy step. It does not convert a 2024 modelling exercise into a measurement under every subsequent rule. This analysis does not re-estimate the package’s revenue effect, assign its gains to individual countries or determine a company’s liability.

A further announcement dated 11 September 2026 covers legislative peer-review methodology, an updated GloBE Information Return and administrative guidance. The return is described as a means of collecting information needed to evaluate multinational tax liabilities. That purpose distinguishes a reporting instrument from a treasury receipt. The September announcement said a revised XML schema was being developed at that time; this review does not treat that historical status as evidence that the schema remains unfinished today. It also does not claim a comprehensive audit of all later technical material.

What would establish collections? For a country-level account, our reporting method would require a named official fiscal release, the tax line covered, its accounting basis and its reference period. A declared liability may differ from cash received; an annual amount cannot be added to a cumulative total without checking overlap. Refunds, adjustments and revisions need their own treatment. These are requirements of our proposed evidence comparison, not assertions that the reviewed documents supply those observations. No missing receipt is recorded as zero in this analysis.

The same discipline limits market interpretation. A modelled global revenue gain cannot independently establish a change in a sovereign yield, an exchange rate or a company valuation. Such an account would need dated market observations, an identified mechanism and evidence separating the tax effect from competing developments. We therefore link the policy analysis to the economic questions it can inform without inventing a trading implication. The accompanying table stores historical estimates as estimates and leaves actual collections absent. It is not a live market feed or an observed training series.

This review’s original contribution is a reproducible comparison of meanings and arithmetic, not a new estimate of the tax’s effectiveness. Source publication dates, model vintages, policy approval dates and Newsverum’s publication time remain distinct. The analysis supports scrutiny of a revenue claim while retaining uncertainty about realized outcomes. Updating it with official collection data would require a new source review and a transparent revision, rather than replacing missing evidence with the most prominent historical projection.

Data

Table

Historical estimates; not actual collections

Modelled additional annual global corporate income tax revenue. Distinct analytical vintages; actual receipts are not supplied.

2023220220Approximate projected annual gain-
2024155192Modelled annual gain range-
Methodology

Transcribed 2023 projected value and 2024 modelled range. Equal 2023 endpoints represent one approximate estimate, not zero uncertainty. Bounds are not a confidence interval. Null collections mean unestablished, never zero. These rows are not an observed time series.

Updated:

Structured coverage

Subjects and places

Taxonomy version: IPTC Media Topics 2026-Q2

Background

Continue reading

Fiscal Policy

Correction: OECD carbon-price coverage was misstated

The reviewed 2025 report covers 79 countries and reports 44% coverage in 2023 when fuel excise taxes are included, not our claimed 64% across seventy advanced economies.

OECD — Effective Carbon Rates 20252 min read Data included
Fiscal Policy

Jordan’s domestic public revenue rose 2% through July

Revenue reached JD5.625 billion in January–July 2026. Non-tax receipts made the larger contribution to the increase, while capital spending also rose.

Jordan News Agency (Petra)4 min read