Giulia Varaschin, Quentin Parrinello, Gabriel Zucman · International Tax Observatory Note/Policy Brief, December 2025

Wealth Taxes and High-Net-Worth Individuals in Europe: Five Lessons for the Twenty-First Century

18 problems found: 6 moderate · 12 minor

What the paper claims

Extreme wealth concentration has grown dramatically in Europe.
High-net-worth individuals face lower effective tax rates than the average citizen.
Historical European net wealth taxes had low thresholds that created liquidity pressures and 'prompted carveouts'; exemptions, valuation discounts and caps removed high-net-worth wealth from the base.
Wealth taxes raised modest revenues that did not rise with household wealth.
Headline: earlier wealth taxes failed 'because of the narrow tax bases and low thresholds, not because of large-scale mobility'; relocation responses are real but small, concentrated and have negligible economic effects.
Five design lessons: (1) broad base without carve-outs or preferential valuation; (2) high thresholds targeting extreme wealth, which 'addresses regressivity gaps' and avoids liquidity pressures; (3) a minimum-tax top-up crediting taxes already paid; (4) credible anti-exile rules (exit taxes, trailing residence) to 'neutralise relocation incentives'; (5) integration with transparency tools.
Alternatives (higher capital-income taxes, regulating holdings, property taxes) fall short on their own.

Every problem found

1moderateconfirmed · newly identified here · misrepresentation of source

The brief says a French study found that a one-point tax rise drives out only 0.003-0.03% of rich taxpayers a year, and puts this next to a Scandinavian figure of about 2%. The two numbers measure different things (a yearly flow per point of income tax versus a long-run total per point of wealth tax), and the French study itself says French reactions are about as strong as the Scandinavian ones. The conclusion that emigration was not the main problem survives, but the French evidence is made to look far more reassuring than it is.

Technical detail and evidence

Why it matters, in detail. The CAE semi-elasticity is per point of the effective rate on reported income, not on wealth. For a top-1% capital-income group with reported income equal to 2-5% of wealth, 1 pp of wealth tax is 20-50 pp of the income rate. A linear rescaling, for illustration only, turns the CAE's long-run stock effect into roughly 0.5-12% per wealth-tax point, bracketing Jakobsen et al.'s 1.77%; the CAE's own comparison of elasticities is 0.3-3.8 times the Swedish one. Read side by side, the brief's numbers (0.003-0.03% a year versus less than 2% stock) imply a French response 59-590 times smaller. The bias runs towards understating French mobility; it does not reverse the ranking of migration below other behavioural responses.

Evidence. Brief p.4 (s.2.3). CAE Focus 118 (Bach, Bozio, Grimprel, Guillouzouic, Landais & Malgouyres, July 2025): p.5 (sensitivity statement; 0.02-0.23% long run), Constat 1 p.9 (+32-79% net departures), p.10 (0.003-0.03% a year; 0.3-3.8 times Swedish), p.17 (asymmetry; 'legitime l'utilisation de la borne haute'), Annex A3 pp.25-26 (definition of the shock; table: flow semi-elasticity 0.003-0.029%, stock 0.023-0.232%). Jakobsen et al., June 12 2025 version, p.23 (stock elasticity 1.77). Adjudicator: scripts/arith_checks.py ([2]).

Why this severity. It distorts one of the brief's three pieces of mobility evidence by about two orders of magnitude in the implied comparison. But the CAE's own bottom line (small flows, small aggregate economic effects) still supports the qualitative claim that migration did not drive the failure.

2moderateconfirmed · previously known · selective presentation

The studies the brief relies on do say that the rich moving abroad matters less than their other reactions. They also say it still costs about a fifth of the extra revenue a wealth tax would raise, and that all reactions together eat up about three-quarters of it. The brief leaves these numbers out, and it skips European studies (Spain, Switzerland) that show large moves between regions. Its core point, that the rich did not flee en masse, stands, but the evidence is more mixed than 'consistently contradicts' suggests.

Technical detail and evidence

Why it matters, in detail. Scandinavia: dB/dM is 0.22 from migration and 0.54 from intensive responses, so about 24 cents of each marginal mechanical euro is retained (MCPF about 4.2); migration is about 29% of the total behavioural loss. Spain: induced income-tax losses are six times the direct wealth-tax losses. Switzerland: taxable-wealth semi-elasticity of at least 43. None of the cited studies estimates counterfactual revenue under a broad base, so the attribution of low revenue 'primarily' to narrow bases rests on the French effective-rate evidence (IPP 2021), not on a decomposition.

Evidence. Brief p.2, p.4 (s.2.3), Conclusion p.6. Jakobsen, Kleven, Kolsrud, Landais & Munoz, June 12 2025 version: Intro p.4; s.7.2 pp.36-37 (dB/dM 0.22 and 0.54); s.7.3 p.37 (Denmark taxed business assets); p.39 MCPF about 4.2. CAE Focus 118 pp.16-17. Agrawal, Foremny & Martinez-Toledano, WIL WP 2020/26 abstract (Madrid +10%); AEJ: Applied 17(1):402-30 (2025) abstract (income-tax losses six times direct wealth-tax losses). Brulhart, Gruber, Krapf & Schmidheiny, AEJ: Policy 14(4):111-150 (2022) (corpus optimal-tax-behavior-19). OECD (2018) p.17. Adjudicator: scripts/arith_checks.py ([3]).

Why this severity. The ranking the brief draws (migration smaller than other responses, no 'mass flight') is the cited authors' own conclusion and survives. But 'consistently contradicts ... eroding revenue' and 'negligible' overstate evidence showing a non-trivial fiscal cost of migration and large total behavioural losses even where business assets were taxed.

The authors' response. No response specific to this brief. As recorded in the corpus (offshore-havens-17), Zucman argued that past wealth taxes were 'truffes d'exonerations' while a no-exemption floor leaves no escape (France 5, quoted by BFM TV, 16 Sep 2025). Camille Landais, co-author of both cited studies, reportedly said the Danish case 'n'a en pratique rien a voir avec la taxe Zucman' (France 24, 27 Sep 2025; not read directly). Jakobsen et al. conclude that migration is 'far less important for welfare and policy design than intensive margin responses'.

3moderateconfirmed · newly identified here · misrepresentation of source

The brief says, citing the OECD, that wealth-tax revenues did not grow even as household wealth grew. For France and Switzerland, the two countries it names, the OECD says the opposite: revenues rose. The general point that these taxes raised little money still stands.

Technical detail and evidence

Why it matters, in detail. Revenue as a share of GDP rose 140% (France) and 74% (Switzerland) over 1990-2016. 'Disconnected from wealth' holds for Spain and Norway and, after 2000, roughly for France (revenue relative to household wealth -12%). The French revenue path, with the revenue-to-GDP ratio more than doubling, is also hard to reconcile with exemptions having 'gradually hollowed out' the base; the hollowing-out shows in top effective rates (IPP Fig. 3.7), not in aggregate revenue.

Evidence. Brief p.4 (s.2.2), Conclusion p.6. OECD (2018) p.19 and Fig. 1.4. OECD SDMX DF_RSOECD T_4210 PT_B1GQ (file fetched by Auditor A: /mnt/zdata/downloads/oecd-revstats-4210-claude/rev4210_PT_B1GQ.csv); WID mhweal and mgdpro. Adjudicator: scripts/revenue_trends.py (output saved).

Why this severity. The cited source is reversed for exactly the two countries named, in the direction of the narrative, and the error is carried into the Conclusion. The separate point that revenues were small is unaffected.

4moderateconfirmed · newly identified here · internal inconsistency

The brief says a tax on fortunes above about EUR 100 million would fix the problem of the rich paying lower tax rates. Its own chart shows that in Italy and the Netherlands rates already start falling for the top 5% of earners, far below that level, and the authors' earlier report said extra measures would be needed there. The proposal fixes only the very top.

Technical detail and evidence

Why it matters, in detail. In two of the three countries shown, the regressive range spans roughly P95-P99.99 of the income distribution, far below a EUR 100m net-wealth threshold (about the top 0.005% in France), so a high-threshold floor leaves most of it untouched. The statement holds for France, where rates decline from about P99.99.

Evidence. Brief s.2.1 p.2, Fig. 1 p.2, s.3.2 p.5. Zucman (2024) G20 blueprint pp.11-12 and p.28. Adjudicator: scripts/fig1_vectors.py (output saved).

Why this severity. An important limit on Lesson 2, which the authors themselves acknowledged elsewhere, is omitted. The lesson still targets the extreme top, so the recommendation is incomplete rather than wrong.

The authors' response. The caveat is stated in Zucman (2024, pp.11-12), the brief's own source; it is not repeated in the brief.

5moderatelikely · newly identified here · overclaiming

The brief says old wealth taxes failed partly because they started at too low a level, which supposedly pushed governments to add loopholes. It gives no evidence for this, and the facts it relies on point the other way: Switzerland taxes ordinary households and raises the most, and France had its big loophole for business owners from the very start. The claim that loopholes undermined the old taxes holds; the claim that low thresholds caused the loopholes is unproven.

Technical detail and evidence

Why it matters, in detail. The mechanism (low threshold, then liquidity pressure on house-rich or SME owners, then carve-outs, then a narrow base and low revenue) is an untested political-economy story. In the available cross-section, low-threshold Switzerland collects about 76% of the theoretical potential of a 1% tax on the top 1%, while high-threshold France and Spain, which had full business exemptions, collect 19-27%. Thresholds and carve-outs plausibly moved together for political reasons; the data cannot assign causation in either direction, and n is about 9.

Evidence. Brief Abstract p.1, s.2.1 p.2, Conclusion p.6. OECD (2018) pp.17-18, Table 4.2 p.81, pp.81-82, Table 4.3 p.84. Saez & Zucman (2022) AEA P&P 112:58-62, Table 1 p.59. Bach, Bozio, Guillouzouic & Malgouyres (2021), IPP report 36, pp.22-23. Adjudicator: scripts/arith_checks.py ([4]). Auditor B noted the same OECD Swiss passage in its assumptions table without counting it as an issue.

Why this severity. It removes the evidential basis for one of the two causes named in the headline, and for part of Lesson 2's rationale. The narrow-base half of the diagnosis remains well supported.

6moderatelikely · previously known · overclaiming

Two of the brief's five lessons run into legal obstacles it does not mention: taxing people when they leave or for years afterwards, and a wealth floor with no cap linked to income. EU law limits taxing people who move within Europe. Weeks before publication, France's highest administrative court advised that a 10-year post-departure tax and an uncapped wealth levy were likely unlawful. The ideas are not ruled out, but they are harder to implement than the brief suggests.

Technical detail and evidence

Why it matters, in detail. Within the EU an exit tax must allow deferral to realisation, so it does not reach owners who hold until death. A multi-year trailing liability is of doubtful compatibility with Art. 49 TFEU and with treaty allocation of taxing rights over capital to the residence state. An uncapped floor at the top would have to overcome French ability-to-pay case law that has historically required a cap, and the IPP evidence shows the cap alone cut the top 0.001%'s potential ISF rate by roughly an order of magnitude. The revenue and feasibility consequences are unquantified.

Evidence. Brief s.3.3-3.4 pp.5-6, Conclusion p.6. CJEU press release 13/04 (11 Mar 2004), C-9/02. Hourani & Perret (2025), OECD Taxation WP 72, s.3.6, p.42, fn 52. Public Senat, 'Rejet de la taxe Zucman a l'Assemblee nationale : est-elle inconstitutionnelle ?', 3 Nov 2025 (saved at /mnt/zdata/downloads/vpz2025-five-lessons-final/web/publicsenat.html). IPP (2021) Fig. 3.7 p.56 (read visually). Corpus offshore-havens-21 (Senate report n. 689, 2024-2025).

Why this severity. The constraints do not invalidate the lessons: deferral-based exit taxes and shorter trailing rules exist, and the constitutional question is unresolved. But they bear directly on whether Lessons 3-4 are feasible in the EU and in France, they were salient when the brief was published, and 'neutralise' overclaims.

The authors' response. No response specific to this brief. As recorded in the corpus (offshore-havens-21), Zucman said the tax would bring French law into line with constitutional equality before taxation (on X, reported by franceinfo, 3 Sep 2025), and rapporteur Eva Sas argued that the EUR 100m threshold and new evidence distinguish it from the ISF precedents.

7minorconfirmed · previously known · documentation gap

The brief now says corporate and inheritance taxes would count towards the minimum tax, whereas the authors' earlier proposals counted only personal taxes. Since most of what billionaires pay is corporate tax, this could shrink the revenue a lot. The brief neither flags the change nor explains what 'structural corporate taxation' means.

Technical detail and evidence

Why it matters, in detail. If all corporate tax attributable to the owner were credited, total taxes would be about 1.25-1.9% of wealth at returns of 5-7.5%, so a 2% floor would raise only 0-26% of the personal-tax-only top-up; critics' French estimates are EUR 1.25-5.5bn against EUR 20-24bn (offshore-havens-18). If 'structural' means only holding-level taxes, the effect is small. Because X is unspecified, the brief's revenue claim cannot be checked either way.

Evidence. Brief s.3.3 pp.5-6, Conclusion p.6. Parrinello, Varaschin & Zucman (March 2025), 'Resources for a Safe and Resilient Europe', pp.2-3 and Table 1 notes. Zucman (2024) pp.14-15 (French billionaires' 27% = 1.7% individual income tax plus about 25% corporate tax) and p.20. Adjudicator: scripts/arith_checks.py ([5]), checking Auditor A's scripts/cit_credit_sensitivity.py.

Why this severity. The design is coherent with the brief's own all-taxes diagnosis (Fig. 1), and the brief gives no revenue figure that the change would contradict. The issue is disclosure, but it matters for anyone carrying over the authors' earlier revenue estimates.

The authors' response. Zucman (Le Grand Continent, 2 Jul 2025; corpus offshore-havens-18) acknowledged that most of billionaires' all-in rate is corporate tax but defended crediting only personal taxes. The December 2025 wording moves towards crediting corporate tax without comment.

8minorconfirmed · newly identified here · data error

Several numbers in the brief's summary table are wrong or mix up years: Denmark's rate cut came in 1989-91, not 1996; Finland's and the Netherlands' rates were lower than shown; and Austria's tax, listed as having no exemptions, had in fact exempted bank savings and bonds in its last year. These are small errors that do not change the brief's argument.

Technical detail and evidence

Why it matters, in detail. Rate-threshold pairs mix different years (Finland, Netherlands), and the Danish chronology misdates the 2.2% to 1% cut by five years (it happened over 1989-91). Austria, presented as the one exemption-free tax, had removed bank deposits and domestic bonds from its base by 1993, just before its 1994 abolition.

Evidence. Brief Table 1 p.3. Danish Skatteministeriet tables 'Centrale skattesatser' 1987-1993 and 1994-2001, row 'Formueskat, trin 1' (/mnt/zdata/downloads/dk-skm-rates-claude/); JJKZ (2020) QJE p.342. Finnish HE 91/2004 (/mnt/zdata/downloads/finlex-varallisuusvero-claude/he-154115.pdf). Kamerstuk 36748 nr. 3. Ley 19/1991 art. 30.2 as amended by Ley 11/2020 (BOE). OECD (2018) Fig. 4.2 p.88, Table 4.3 p.84. Endbesteuerungsgesetz s.1 (Wayback copy of RIS, /mnt/zdata/downloads/vpz2025-five-lessons-final/web/ris_wb.html).

Why this severity. Descriptive errors of 0.1-1.2 pp in rates and dates; none changes the base-erosion argument, and the Austrian correction, if anything, supports it.

9minorconfirmed · newly identified here · arithmetic or transcription

The brief slightly understates French wealth-tax revenue as a share of the economy (about 0.23%, not 'less than 0.2%'); it was still small.

Technical detail and evidence

Why it matters, in detail. 0.229% versus 'less than 0.2%', about 13% above the stated bound.

Evidence. Brief p.4. Saez & Zucman (2022) pp.59-60. INSEE Premiere 1650. Adjudicator: scripts/arith_checks.py ([1]).

Why this severity. A rounding or wording slip of about 0.03 pp of GDP; the conclusion that revenue was small is unaffected.

10minorconfirmed · newly identified here · documentation gap

The chart of tax rates differs slightly from its stated source for French billionaires (25% instead of 27%), and it does not say that the Italian billionaire figure comes from one person. The overall picture, with the richest paying lower rates, is the same.

Technical detail and evidence

Why it matters, in detail. The French billionaire endpoint is 7.4% below the cited source, slightly widening the displayed gap. The Italian billionaire rate is a single named taxpayer, so it is not a distributional estimate.

Evidence. Brief Fig. 1 p.2, Fig. 2 p.4. Adjudicator: scripts/fig1_vectors.py and scripts/fig2_vectors.py (outputs saved). March 2025 note chart (embedded image, /mnt/zdata/downloads/vpz2025-five-lessons-final/euto_img-000.png). Zucman (2024) pp.10-11. IPP (2021) s.3.3 and Fig. 3.7 p.56.

Why this severity. A 2 pp difference on one data point plus undisclosed caveats; the pattern of falling rates at the very top is unchanged.

11minorconfirmed · newly identified here · documentation gap

The brief's numbers on the growth of top wealth in Europe roughly match the World Inequality Database, but the richest 0.1% hold about 3.6 times (not almost 4 times) what the poorest half hold. The rise looks larger because the comparison starts in a low year.

Technical detail and evidence

Why it matters, in detail. Top 0.1% share (QE-PPP): 9.76% (1980), 8.45% (1985), 11.41% (2024); bottom 50%: 3.17% (2024); ratio 3.60 (PPP) or 4.17 (MER). The 'dramatic' growth is +2.96 pp from the 1985 trough but +1.65 pp from 1980.

Evidence. Brief p.1 and fn 1. Adjudicator: scripts/wid_check.py (output saved; raw JSON at /mnt/zdata/downloads/vpz2025-five-lessons-final/wid/), scripts/arith_checks.py ([7]).

Why this severity. The levels reproduce; the ratio is overstated by about 10% and the start year flatters the trend, which matters little for the brief's argument about tax design.

12minorconfirmed · newly identified here · documentation gap

Several references are wrong (a working-paper number, the report editors, an author's name) and one study is cited in an earlier version than the numbers quoted, which makes the claims harder to check but does not change them.

Technical detail and evidence

Why it matters, in detail. The version mismatch means a reader checking the cited NBER WP finds different (bound-type) numbers; the substantive estimates are those of the June 2025 revision.

Evidence. Brief References p.7 and in-text citations pp.2-6. Du Rietz & Henrekson WP 1000 cover (/mnt/zdata/downloads/durietz-henrekson-2014-verify/wp1000.pdf). GTER 2024 title page. CPB DP 436 title page. Jakobsen et al. w32153 (Feb 2024) abstract versus the June 12 2025 version abstract.

Why this severity. Citation accuracy only; the cited content is generally represented.

13minorconfirmed · newly identified here · misrepresentation of source

The brief slightly misdescribes why UK tax revenue from wealthy foreigners who stayed went up: it was because more of their foreign income became taxable, not because they brought money home. The key numbers are reported correctly.

Technical detail and evidence

Why it matters, in detail. The stayers' revenue gain reflects a change in the tax base (the end of the remittance basis), not a real relocation of income or investment.

Evidence. Brief p.4. Advani, Burgherr & Summers (May 6 2025), 'Taxation and Migration by the Super-rich', abstract.

Why this severity. A loose paraphrase of the mechanism; the headline numbers are accurate.

14minorconfirmed · newly identified here · misrepresentation of source

The brief slightly overstates two background facts: undistributed company profits are a big part, but not the biggest part, of the very top incomes in the Dutch study, and the 11% property figure comes from billionaire family offices rather than rich people in general. Its points still hold.

Technical detail and evidence

Why it matters, in detail. Retained earnings plus capital income make up about 74% of top-0.01% pre-tax income in the Netherlands. The family-office allocation excludes the operating business, so it probably overstates real estate's share of total UHNW wealth while understating it for ordinary HNWIs.

Evidence. Brief Info box p.5. CPB Discussion Paper 436, Fig. 5 p.17 and text p.16 (read visually; consistent with Auditor A's digitisation). UBS Global Family Office Report 2025, pp.2, 10-13.

Why this severity. Wording overstatements; the qualitative points (large untaxed retained earnings at the top; real estate a small share of the largest fortunes) hold.

15minorlikely · previously known · overclaiming

The brief says a high threshold avoids the problem of people owing tax without the cash to pay it, because very large fortunes earn high returns. But high returns on paper are not cash: a founder whose company pays no dividends could owe millions a year. This affects relatively few people and can be handled by letting them pay in shares, but the brief does not say so.

Technical detail and evidence

Why it matters, in detail. Liquidity depends on cash yield and borrowing capacity, not on the expected total return. At a 2% floor, owners with payout yields below 2% must sell, borrow or pay in kind. Prevalence and revenue effects are unquantified; payment in shares solves liquidity at the cost of cash revenue and raises governance questions.

Evidence. Brief s.3.2 p.5, Conclusion p.6. IPP (2021) p.53. OECD (2018) pp.91-92. Corpus offshore-havens-20 (Senate report n. 689, IV.B; Aghion/Jaravel). Auditor B scripts/reproduce.py (liquidity_counterexamples).

Why this severity. The gap concerns a narrow subset of taxpayers (high-valuation, low-payout founders) and can be closed by design, so it does not undermine the lesson; the problem is overstatement ('avoid').

The authors' response. Zucman (Challenges debate, 17 Sep 2025; corpus offshore-havens-20) said payment could be made 'en nature' with company shares and that hard cases must be studied but should not be an 'alibi'.

16minorplausible · previously known · methodology

Whether billionaires pay 'half' the tax rate of ordinary people depends on what counts as a tax (for example pension contributions and VAT). Under every reasonable choice they still pay less, but the gap is smaller than the headline comparison suggests, and the brief does not say so.

Technical detail and evidence

Why it matters, in detail. Average French rate 52% (all levies), about 41% excluding pension contributions, about 32% also excluding VAT, about 28% net of pension spending, against about 26% for billionaires (Zucman's reply). Against millionaires (IPP: about 46%) the gap is smaller; counting future transfer taxes raises the billionaire rate to about 32% (Geerolf).

Evidence. Brief Abstract p.1, Fig. 1 p.2, Info box p.5. Corpus offshore-havens-19 (Geerolf, L'Express, 18 Sep 2025; Zucman, Le Grand Continent, 2 Jul 2025); us-wealth-taxes-12, -13, -14 (Splinter 2020; Kopczuk 2019; Gleckman 2019).

Why this severity. The gap narrows under alternative conventions but survives in every variant examined; the regressivity at the very top is not contested.

The authors' response. Zucman, 'Comment taxer les ultrariches ? Cadrer le debat sur l'impot plancher', Le Grand Continent, 2 Jul 2025: the gap persists after excluding pension contributions and VAT, and net of transfers (as recorded in corpus offshore-havens-19).

17minorplausible · previously known · speculative assumption

The brief says new information-sharing between tax authorities has cut offshore tax evasion to a third. That figure is an estimate built on assumptions and is disputed; evasion probably did fall, but by how much is uncertain.

Technical detail and evidence

Why it matters, in detail. Untaxed offshore wealth goes from 90% before the CRS to 27% in the central scenario (15% low, 37% high), which is a reduction by a factor of 3.3, 6.0 or 2.4. TJN's alternative assumptions give about 9% of world GDP undeclared against GTER's 3.2%.

Evidence. Brief s.3.5 p.6. GTER 2024 executive summary and pp.27-29 (Auditor B's scenario arithmetic). Corpus offshore-havens-07 (TJN, State of Tax Justice 2024).

Why this severity. The brief's 'suggest' hedges the claim, and the direction (evasion fell after the CRS) is not in dispute; only the magnitude is assumption-driven.

The authors' response. GTER 2024 labels the non-reporting parameter provisional and presents low and high scenarios. Faye, Godar, Moura & Zucman (2025) later report that offshore wealth stayed broadly stable as a share of GDP (corpus offshore-havens-07).

18minorplausible · previously known · overclaiming

The brief says all assets should be taxed at market value but does not explain how to value private companies that are never traded, which is where much of the largest fortunes sit. Tax authorities elsewhere have often got these values badly wrong.

Technical detail and evidence

Why it matters, in detail. Valuation shortfalls would reduce the administrable base relative to rich-list market values; the effect is unquantified, and the Norwegian evidence concerns mostly small firms, not billionaire groups.

Evidence. Brief s.3.1 p.5. Zucman (2024) p.20 and s.4.1. Corpus offshore-havens-23 (Skar, Traini, Koerver Schmidt & Solheim 2025, Oxford CBT WP 25/06).

Why this severity. A feasibility caveat that a 7-page brief can reasonably leave to companion work; it does not affect the historical diagnosis.

The authors' response. Zucman (2024, s.4.1) proposes valuing large private firms with listed-peer multiples and observed transactions (corpus offshore-havens-23).

Claims examined and rejected

Problems one auditor raised that did not survive checking.

The brief 'reverses' the CAE's characterisation of French mobility (title of A's I1). — The CAE says both that French high-wealth households are 'relativement sensibles a la fiscalite' and that 'les flux totaux d'expatriations restent tres faibles en part de la population' (p.5). The brief selects the second half; that is selective, not a reversal. The underlying unit and juxtaposition problem is upheld as F1.
Table 1's Finnish entry ('valuation discounts on selected business assets') conflicts with OECD (2018) Table 4.3, which records Finnish business assets as fully taxed. — OECD Table 4.3's own note says 'No information for Denmark, Finland and Iceland', even though a Finnish column is printed, so the table cannot establish the point; Finnish valuation law was not checked by either auditor.
Norway's '1.1%' in Table 1 is an error because the standard rate is 1.0%. — 1.1% is Norway's current top rate, and Table 1 gives single rates for most countries; this is a simplification, not an error.
Table 1 is in error for not listing income-based caps for the Netherlands and Swiss cantons. — The column is explicitly labelled 'non-exhaustive' (fn 3). The facts are correct (68% samenloopregeling; seven cantons), but the omission is not an error.
The cross-country rank correlations between thresholds and business-asset relief (+0.61; +0.43 with earlier thresholds) show that the cross-section contradicts the low-threshold mechanism. — With n=9, ordinal coding, Finland coded from an OECD column flagged as 'no information', and plausible reverse causation, the correlation is not probative. F3 rests on documentary evidence instead (OECD p.18 on Switzerland; the 1982 French exemption).
Hourani & Perret (2025) state that no empirical research on the effects of exit taxes is yet available. — Imprecise: the OECD sentence (p.42) concerns 'the impact of exit taxes on inward migration and entrepreneurship'. The narrower statement is used in F6.
The Advani et al. evidence is misused because it concerns an income-tax (non-dom) reform rather than a wealth tax. — The brief explicitly describes the evidence as concerning 'the removal of preferential tax treatment for long-term foreign-domiciled residents'; the scope is disclosed. Only the mechanism paraphrase is upheld (F14).
The change in creditable taxes under the minimum tax is a moderate issue that could remove most of the revenue. — Severity downgraded to minor (F7). The brief sets no floor rate X and gives no revenue estimate, 'structural corporate taxation' may refer only to holding-level taxes, and the design is coherent with the brief's all-taxes diagnosis; the undisclosed change itself is upheld.
The WID 1985 and 2025 endpoints cannot be reproduced (R1). — Superseded. The WID API (current vintage, Europe PPP aggregate) gives 8.45% for 1985 and 11.41% for 2024, matching the brief's levels. Only the '2025' label, the vintage and the ratio remain problems (F12).
The high-threshold liquidity overclaim is a moderate issue. — Severity downgraded to minor (F8). The problem concerns a narrow set of high-valuation, low-payout owners, can be addressed by payment in kind or deferral, and the brief's main text says 'mitigate'.

Audit notes

What reproduced. No author data or code exist. Reproduced from public sources: (1) The WID Europe (PPP) top-0.1% wealth share is 8.45% in 1985 and 11.41% in 2024, matching the brief's ~8.5% and ~11%; there is no 2025 value, and the ratio to the bottom 50% is 3.60 (4.17 at market exchange rates). (2) Fig. 1 vector values were extracted: French billionaire point 25.0% against about 27% in the cited source; Dutch and Italian profiles match. (3) Fig. 2's 31 vector values (peak 0.45%, last 0.04%) match IPP (2021) Fig. 3.7, after-cap series, 2016. (4) EUR 5.1bn is 0.229% of 2016 GDP, not under 0.2%. (5) The ISF-to-IFI cost is EUR 3.44bn (Paquier & Sicsic Table 2), and the OECD shares of total taxes are 0.48% (France) and 3.70% (Switzerland). (6) OECD 4210 revenue as a share of GDP rose in France (0.090% to 0.215%) and Switzerland (0.594% to 1.034%) over 1990-2016. (7) The CAE 0.003-0.03% range is reproduced in its income-rate units (Annex A3). (8) The Jakobsen et al. figures match the June 2025 revision (stock elasticity 1.77; -0.02/-0.07/-0.10%; fiscal externalities 0.22 and 0.54). (9) Advani et al.'s 0.26 elasticity is confirmed. (10) Danish, Finnish, Dutch and Spanish statutory rates were checked against primary sources. Not reproduced: the authors' December 2025 WID vintage and series, and the confidential microdata behind the IPP, CAE, Scandinavian and HMRC estimates, which were checked only at the level of published tables.

Limits of this audit. The brief is a non-technical policy note with no data or replication files; its WID vintage (accessed 3 Dec 2025) could not be retrieved, and the October 2026 API vintage was used. The confidential microdata behind the IPP, CAE, Scandinavian and HMRC estimates were not accessed; checks reach published tables and figures only. The Conseil d'Etat opinion of 30-31 Oct 2025 is known only through the minister's reading as quoted by Public Senat (3 Nov 2025): it is advisory, was reportedly given on a 'light' variant, and was not published. The Austrian finding rests on the constitutional provision (Endbesteuerungsgesetz s.1, Wayback copy) rather than the implementing statute, because RIS blocked direct access. The German translation (FES, Jan 2026) was not obtained. Readings of raster figures (IPP Fig. 3.7, Bruil Fig. 5, the March 2025 chart) are visual, with about 0.5 pp precision. The threshold cross-section and the corporate-tax-credit illustration are small-sample and average-parameter exercises. No author response specific to this brief was located.