On 28 October 2026 Marginal Revolution featured an anonymous AI audit of research by the economists Emmanuel Saez and Gabriel Zucman. The audit lists 147 problems it calls serious enough to change a headline number
. We call each listed problem a flag; a headline number is a number in a paper’s abstract, introduction or main tables. We checked all 147 against the papers, their sources and the authors’ own data. We check the audit’s claims and take no side where researchers dispute a method. We present the facts; readers draw the conclusions.
facts all right Every fact in the flag that we could check was right.
core facts wrong 8 had smaller errors; 8 we could not check.
change a headline number Shown by a calculation or by the work’s own numbers.
depend on a disputed choice of method Reasonable researchers differ on these. We take no side.
Each is shown by a calculation or by the work’s own numbers. The numbers are ours: 1 to 30 are the random sample we checked first.
Neef et al. · EU Tax Observatory / World Inequality Lab Note, March 2022
The note says about half of Russian household wealth is held abroad; the study it cites gives about a quarter (24%) of total wealth. The note’s summary claim survives.
Gamage et al. · Policy report on New York bill S8277B (2019-20), 1 September 2020; SSRN 3766547
The bill caps the gains counted as income, not the tax. With the report’s own inputs, 2020 revenue is about $9.0 billion, not $23.2 billion; most of the rest comes due in later years.
Silva et al. · International Tax Observatory report commissioned by Brazil under the PTLAC Presidency, April 2026
The report’s own revenue figures cannot lift the top rate in Brazil to 50% (about 34%, at most 46%). The Chile part was not checked.
Alstadsæter et al. · EU Tax Observatory Note, May 2024
The authors’ September 2025 paper splits the same $98.1 billion by residence: the offshore share is about 30%, not 43.8%. Dubai stays far above the other cities.
Alstadsæter et al. · EU Tax Observatory report, October 2023
The 2021-22 points were projections; the same team’s July 2026 series gives a 6.9% loss share, not nearly 10%. The $1 trillion level survives.
Alstadsæter et al. · Journal of Public Economics 162: 89-100
The authors’ code keeps a 2006 value for Venezuela against the paper’s stated rule; on the rule, Venezuela falls from 64% to about 25% of GDP.
Alvaredo et al. · World Inequality Lab, 2017; Cambridge, MA: Harvard University Press (Belknap), 2018
An undocumented adjustment to Middle East incomes for 1980-84 sits in the published code; without it, world income growth per adult is 53%, not 60%.
Alvaredo et al. · AEA Papers and Proceedings 108: 103-108
The same code adjustment as in 41 and 45; without it, top 1% growth is 79%, not 101%, and the top 1% share rises 2.98 points, not 4.28. The elephant shape survives.
Alvaredo et al. · Economie et Statistique / Economics and Statistics 517-518-519: 41-59 (special issue Beyond GDP
)
The code sets Middle East incomes for 1980-84 below the authors’ own inputs; without that step, the rise in the top 1% share falls from 4.28 to 2.98 points.
Andreescu et al. · World Inequality Lab Working Paper 2025/25, October 2025
The released code and data do not reproduce Tables 3 and 4; the growth coefficient is 0.041, not 0.090, so 0.9 points per 10 points
becomes about 0.4.
Bachas et al. · Working paper, June 2026 (SSRN 7122974)
The abstract says enforcement explains 22-28% of the rise in reported offshore wealth. The paper defines that rise as financial wealth but takes its rate from the all-asset rows of its own Table A.6, which also has financial-asset rows. With those rows the share is about 19% (19.3% level, 19.1% log), and the combined 60-65% becomes about 57%. The change is small and inside the paper’s standard errors.
Added 9 November 2026 by a second pass, after the authors’ reply window (29 October letter, 4 November reminder, replies asked by 6 November) closed with no reply.
Piketty and Saez · IMF Economic Review 61(3): 456-478
Over 15% of national income
uses shares of fiscal income; with the paper’s own Table 3, the shift is about 10% of national income.
Saez · Tax Policy and the Economy 31: 71-120
The paper’s own method on the final 2015 IRS data gives a 31.5% revenue loss, not at most 20%
. The paper marked its 2015 data as preliminary.
Saez and Zucman · Quarterly Journal of Economics 131(2): 519-578
The authors’ own 2020 and 2022 data releases put the 2012 top 0.1% wealth share at about 20%, not 22%.
Saez and Zucman · Letter, January 18, 2019; updated letter on the Ultra-Millionaire Tax Act of 2021, February 24, 2021
The letters derive a 15% avoidance rate for a 2% tax, then apply it unchanged at 3% and 6%. Their own rule cuts the 2021 estimate of $3.9 trillion by 18-26%; the main 2019 figures move 4-6%.
Saez and Zucman · AEA Papers and Proceedings 112: 58-62
On the World Inequality Database figures available in October 2021, benchmark revenue for Spain, Norway and Switzerland is 0.71, 0.45 and 1.16, not the printed 0.96, 0.86 and 1.36.
Wright and Zucman · NBER Working Paper 24983, September 2018
Running the counterfactual the paper describes on the authors’ workbook gives 3.2 to 3.8 points of the 14.5-point fall, not almost half
.
all in this together?
Zucman · CEP Election Analysis Paper EA030, Centre for Economic Performance, LSE (#ElectionEconomics series), April 2015
The summary box says 1.6 to 3 times; the briefing’s own text and Figure 2 give 1.65 to 1.97 times.
Zucman · Economics for Inclusive Prosperity (EfIP) Research Brief, September 2018
The brief’s 40% is higher than its source’s 36.2%. A small difference.
Zucman · Report commissioned by the Brazilian G20 Presidency, June 25, 2024 (EU Tax Observatory)
On the report’s own figures, 7.1% is total group wealth growth; per member it is about 5.4%, implying a return near 5.8%, not 7.5%. Revenue is unaffected.
Zucman et al. · Regards croisés sur l’économie no. 24: 10-17 (2019)
The interview says the share doubled from 20% to 40%; the authors’ own series give about 24% to 40%, a rise of about 1.6 times.
After the authors’ reply window closed, we went back to the 99 flags first counted as depending on a disputed choice of method: the 98 still counted that way, plus No. 54, which this pass moved. For each we asked how standard the choice is. We wrote the rules first and recorded their hash before reading any flag: a fingerprint of the rules file, so anyone can confirm the rules did not change. Two AI readers from different model families graded each flag: Claude (Anthropic), from the paper, its appendix and code and the works it cites, and Gemini (Google), from the pages the first reader cited, blind to its grades. They agreed on 61 of 99; we settled the rest from the cited pages. A red team, a separate check that tries to break a finding by redoing it, re-tested every possible change to a headline number.
| Grade | Flags | 95% range |
|---|---|---|
| Standard | 11 | 6.3-18.8% |
| Defensible, stated in the paper | 61 | 51.8-70.6% |
| Defensible, not stated | 9 | 4.9-16.4% |
| Questionable | 11 | 6.3-18.8% |
| Cannot tell | 6 | 2.8-12.6% |
| Contested | 1 | 0.2-5.5% |
Standardmeans used in the work the paper cites, not
correct. Nothing here speaks to anyone’s intentions.
Facts: are the flag’s statements about the work, its sources and its numbers right? Headline number: does the problem change one, as the audit says? Our note sits under the title where we have one. Second pass: the grade for the flags that depended on a disputed choice of method.
| No. | Work and our note | Facts | Headline number | Second pass |
|---|---|---|---|---|
| 1 | Zucman, Taxing across Borders: Tracking Personal Wealth and Corporate Profits | right | depends on a disputed choice of method | defensible, stated in the paper |
| 2 | Saez and Zucman, The Rise of Income and Wealth Inequality in America: Evidence from Distributional Macroeconomic Accounts | right | depends on a disputed choice of method | defensible, not stated |
| 3 | Johannesen and Zucman, The End of Bank Secrecy? An Evaluation of the G20 Tax Haven Crackdown | right | depends on a disputed choice of method | defensible, stated in the paper |
| 4 | Piketty and Zucman, Wealth and Inheritance in the Long RunThe flag says the authors gave no response; the chapter itself answers the gift-incentive point (p. 1337). | partly right | depends on a disputed choice of method | standard |
| 5 | Saez and Zucman, Top Wealth in America: A Reexamination | right | depends on a disputed choice of method | defensible, stated in the paper |
| 6 | Neef et al., Effective Sanctions against Oligarchs and the Role of a European Asset RegistryThe note says about half of Russian household wealth is held abroad; the study it cites gives about a quarter (24%) of total wealth. The note’s summary claim survives. | right | changes (shown at the time) | not graded |
| 7 | François et al., A Modern Excess Profit Tax | right | depends on a disputed choice of method | standard |
| 8 | Zucman, Global Wealth Inequality | right | depends on a disputed choice of method | defensible, not stated |
| 9 | Saez and Zucman, The Triumph of Injustice: How the Rich Dodge Taxes and How to Make Them Pay | right | depends on a disputed choice of method | contested |
| 10 | Atkinson et al., Top Incomes in the Long Run of HistoryCanada’s rise is about 7.4 to 7.75 points, not 7.8, so the US rise is marginally larger, not no larger. Minor. | partly right | depends on a disputed choice of method | standard |
| 11 | Landais et al., Réponse de Camille Landais, Thomas Piketty et Emmanuel Saez | could not check | could not check | not graded |
| 12 | Landais et al., Pour une révolution fiscale : un impôt sur le revenu pour le XXIe siècle | could not check | depends on a disputed choice of method | defensible, stated in the paper |
| 13 | Chetty et al., Where is the Land of Opportunity? The Geography of Intergenerational Mobility in the United States | right | depends on a disputed choice of method | defensible, stated in the paper |
| 14 | Zucman, Taxing across Borders: Tracking Personal Wealth and Corporate Profits | right | depends on a disputed choice of method | defensible, stated in the paper |
| 15 | Kleven et al., Migration and Wage Effects of Taxing Top Earners: Evidence from the Foreigners’ Tax Scheme in Denmark | right | depends on a disputed choice of method | defensible, stated in the paper |
| 16 | Gidron et al., Global Profit Shifting, 1975-2023 (earlier: Global Profit Shifting, 1975-2019)No headline number changes. The abstract’s peaked, then declinedtiming is contradicted by the paper’s own Table A1. | right | does not change; a claim is overstated in words | not graded |
| 17 | Bustos et al., Challenges of Monitoring Tax Compliance by Multinational Firms: Evidence from ChileNo number changes. Effective tax rateis not defined in the paper, and on a payroll reading the claim holds. | right | does not change | not graded |
| 18 | Landais et al., Pour une révolution fiscale : un impôt sur le revenu pour le XXIe siècle | right | depends on a disputed choice of method | defensible, stated in the paper |
| 19 | Saez, Taxes and International Mobility of Talent | could not check | could not check | not graded |
| 20 | Bachas et al., Automatic Exchanges of Information and Offshore Tax Compliance: Evidence from Colombia | right | depends on a disputed choice of method | defensible, stated in the paper |
| 21 | Saez, Should We Tax Wealth? | right | depends on a disputed choice of method | cannot tell |
| 22 | Boas et al., Taxing Capital in a Globalized World: The Effects of Automatic Information Exchange | right | depends on a disputed choice of method | defensible, stated in the paper |
| 23 | Alvaredo et al., Global Inequality Dynamics: New Findings from WID.world | right | depends on a disputed choice of method | defensible, not stated |
| 24 | Chetty et al., Economic Mobility | right | depends on a disputed choice of method | standard |
| 25 | Piketty et al., Simplified Distributional National AccountsTwo figures reproduce, but the key re-ranking term does not: we get +0.64 to +0.97 points for 2016, not +2.68, and an overlap of 81-84%, not 63-79%. The flag does not state its data version, so part of the gap may be method. | partly right | depends on a disputed choice of method | defensible, stated in the paper |
| 26 | Card et al., Inequality at Work: The Effect of Peer Salaries on Job Satisfaction | right | depends on a disputed choice of method | defensible, stated in the paper |
| 27 | Bachas et al., Automatic Exchanges of Information and Offshore Tax Compliance: Evidence from Colombia | right | depends on a disputed choice of method | defensible, stated in the paper |
| 28 | Jakobsen et al., Wealth Taxation and Wealth Accumulation: Theory and Evidence from DenmarkThe flag says the paper reports only the top of the range; Table II prints the raw estimate (0.144) beside the adjusted one (0.312). | partly right | depends on a disputed choice of method | defensible, stated in the paper |
| 29 | Gamage et al., The NY Billionaire Mark-to-Market Tax Act: Revenue, Economic, and Constitutional AnalysisThe bill caps the gains counted as income, not the tax. With the report’s own inputs, 2020 revenue is about $9.0 billion, not $23.2 billion; most of the rest comes due in later years. | right | changes (shown at the time) | not graded |
| 30 | Silva et al., Progressive Taxation and Inequality: Minimum Taxation of High-Net-Worth Individuals in Latin America and the CaribbeanThe report’s own revenue figures cannot lift the top rate in Brazil to 50% (about 34%, at most 46%). The Chile part was not checked. | right | changes (shown at the time) | not graded |
| 31 | Alstadsæter et al., Foreign Investment in the Dubai Housing Market, 2020-2024The authors’ September 2025 paper splits the same $98.1 billion by residence: the offshore share is about 30%, not 43.8%. Dubai stays far above the other cities. | right | changes (shown by later data) | not graded |
| 32 | Alstadsæter et al., Global Tax Evasion Report 2024The 2021-22 points were projections; the same team’s July 2026 series gives a 6.9% loss share, not nearly 10%. The $1 trillion level survives. | right | changes (shown by later data) | not graded |
| 33 | Alstadsæter et al., Global Tax Evasion Report 2024 | right | depends on a disputed choice of method | questionable |
| 34 | Alstadsæter et al., Global Tax Evasion Report 2024 | right | depends on a disputed choice of method | defensible, stated in the paper |
| 35 | Alstadsæter et al., Global Financial Transparency and Offshore Wealth AccumulationThe numbers reproduce. The before-and-after difference is never tested and is not significant, so calling it established is overstated. | right | does not change; a claim is overstated in words | not graded |
| 36 | Alstadsæter et al., Global Financial Transparency and Offshore Wealth AccumulationThe estimates stand. The paper’s Table A.2 contradicts its sentence that the non-haven effects are almost identical. | right | does not change; a claim is overstated in words | not graded |
| 37 | Alstadsæter et al., Global Financial Transparency and Offshore Wealth Accumulation | right | depends on a disputed choice of method | defensible, stated in the paper |
| 38 | Alstadsæter et al., Who Owns the Wealth in Tax Havens? Macro Evidence and Implications for Global Inequality | right | depends on a disputed choice of method | defensible, stated in the paper |
| 39 | Alstadsæter et al., Who Owns the Wealth in Tax Havens? Macro Evidence and Implications for Global Inequality | right | depends on a disputed choice of method | defensible, not stated |
| 40 | Alstadsæter et al., Who Owns the Wealth in Tax Havens? Macro Evidence and Implications for Global InequalityThe authors’ code keeps a 2006 value for Venezuela against the paper’s stated rule; on the rule, Venezuela falls from 64% to about 25% of GDP. | right | changes (shown at the time) | not graded |
| 41 | Alvaredo et al., World Inequality Report 2018An undocumented adjustment to Middle East incomes for 1980-84 sits in the published code; without it, world income growth per adult is 53%, not 60%. | right | changes (shown at the time) | not graded |
| 42 | Alvaredo et al., Distributional National Accounts | right | depends on a disputed choice of method | cannot tell |
| 43 | Alvaredo et al., The Elephant Curve of Global Inequality and GrowthThe same code adjustment as in 41 and 45; without it, top 1% growth is 79%, not 101%, and the top 1% share rises 2.98 points, not 4.28. The elephant shape survives. | right | changes (shown at the time) | not graded |
| 44 | Alvaredo et al., Towards a System of Distributional National Accounts: Methods and Global Inequality Estimates from WID.worldThe numbers reproduce. The Gini is flat, so global inequality has increasedholds only for measures sensitive to the top. | right | does not change; a claim is overstated in words | not graded |
| 45 | Alvaredo et al., Towards a System of Distributional National Accounts: Methods and Global Inequality Estimates from WID.worldThe code sets Middle East incomes for 1980-84 below the authors’ own inputs; without that step, the rise in the top 1% share falls from 4.28 to 2.98 points. | right | changes (shown at the time) | not graded |
| 46 | Alvaredo et al., The Top 1 Percent in International and Historical Perspective | right | depends on a disputed choice of method | standard |
| 47 | Andreescu et al., Equality and Development: A Comparative and Historical Perspective 1800-2025 | right | depends on a disputed choice of method | defensible, stated in the paper |
| 48 | Andreescu et al., Equality and Development: A Comparative and Historical Perspective 1800-2025The released code and data do not reproduce Tables 3 and 4; the growth coefficient is 0.041, not 0.090, so 0.9 points per 10 pointsbecomes about 0.4. | right | changes (shown at the time) | not graded |
| 49 | Andreescu et al., Equality and Development: A Comparative and Historical Perspective 1800-2025 | right | depends on a disputed choice of method | defensible, stated in the paper |
| 50 | Andreescu et al., Equality and Development: A Comparative and Historical Perspective 1800-2025 | could not check | could not check | not graded |
| 51 | Andreescu et al., Equality and Development: A Comparative and Historical Perspective 1800-2025 | right | depends on a disputed choice of method | questionable |
| 52 | Bachas et al., Globalization, Capital Taxation, and Development: Evidence from a Macro-Historical Database | right | depends on a disputed choice of method | defensible, stated in the paper |
| 53 | Bachas et al., Automatic Exchanges of Information and Offshore Tax Compliance: Evidence from Colombia | right | depends on a disputed choice of method | defensible, stated in the paper |
| 54 | Bachas et al., Automatic Exchanges of Information and Offshore Tax Compliance: Evidence from ColombiaThe abstract says enforcement explains 22-28% of the rise in reported offshore wealth. The paper defines that rise as financial wealth but takes its rate from the all-asset rows of its own Table A.6, which also has financial-asset rows. With those rows the share is about 19% (19.3% level, 19.1% log), and the combined 60-65% becomes about 57%. The change is small and inside the paper’s standard errors. | right | changes (shown at the time) | questionable |
| 55 | Baraké et al., Revenue Effects of the Global Minimum Tax Under Pillar Two | right | depends on a disputed choice of method | defensible, stated in the paper |
| 56 | Blanchet et al., Real-Time InequalityThe numbers stand. The authors’ own real-time releases showed +0.25 points against +1.42 final, so the claim of real-time accuracy is overstated. | right | does not change; a claim is overstated in words | not graded |
| 57 | Boas et al., Taxing Capital in a Globalized World: The Effects of Automatic Information ExchangeThe 70% central estimate stands. The 56-95% range leaves out the uncertainty in one input, so the stated precision is overstated. | right | does not change; a claim is overstated in words | not graded |
| 58 | Bustos et al., The Race Between Tax Enforcement and Tax Planning: Evidence From a Natural Experiment in ChileThe estimates reproduce. The confidence intervals cannot rule out recovering about half the gap, so shown ineffectiveis overstated. | right | does not change; a claim is overstated in words | not graded |
| 59 | Bustos et al., The Race Between Tax Enforcement and Tax Planning: Evidence From a Natural Experiment in ChileThe estimates stand. Reject any substantial reductionholds only for some specifications; others cannot reject reductions of 50-100%. | right | does not change; a claim is overstated in words | not graded |
| 60 | Chancel et al., World Inequality Report 2022 | right | depends on a disputed choice of method | defensible, stated in the paper |
| 61 | Chetty et al., Is the United States Still a Land of Opportunity? Recent Trends in Intergenerational Mobility | right | depends on a disputed choice of method | defensible, stated in the paper |
| 62 | Chetty et al., Is the United States Still a Land of Opportunity? Recent Trends in Intergenerational Mobility | right | depends on a disputed choice of method | defensible, stated in the paper |
| 63 | Chetty et al., The Determinants of Income Segregation and Intergenerational Mobility: Using Test Scores to Measure Undermatching | right | depends on a disputed choice of method | standard |
| 64 | Chetty et al., Income Segregation and Intergenerational Mobility Across Colleges in the United States | right | depends on a disputed choice of method | standard |
| 65 | Diamond and Saez, The Case for a Progressive Tax: From Basic Research to Policy Recommendations | right | depends on a disputed choice of method | defensible, stated in the paper |
| 66 | Diamond and Saez, Taxing High Earnings | right | depends on a disputed choice of method | standard |
| 67 | François et al., A Modern Excess Profit Tax | right | depends on a disputed choice of method | defensible, stated in the paper |
| 68 | Galle et al., Expert Report on Proposition 40, the California 2026 Billionaire Tax: Revenue, Economic, and Constitutional Analysis | right | depends on a disputed choice of method | defensible, stated in the paper |
| 69 | Galle et al., Expert Report on Valuation of Controlling Shares of Publicly Traded Companies Under the California Billionaire Tax Act (CBTA)The memo gives no number. It leaves out conditions in the statute ( clear and convincing evidence, substantially), so the legal claim is overstated. | right | does not change; a claim is overstated in words | not graded |
| 70 | Galle et al., Expert Report on Valuation of Controlling Shares of Publicly Traded Companies Under the California Billionaire Tax Act (CBTA) | right | depends on a disputed choice of method | defensible, stated in the paper |
| 71 | Garcia-Bernardo et al., Did the Tax Cuts and Jobs Act Reduce Profit Shifting by US Multinational Companies? | right | depends on a disputed choice of method | defensible, stated in the paper |
| 72 | Guyton et al., Tax Evasion at the Top of the Income Distribution: Theory and Evidence | right | depends on a disputed choice of method | defensible, stated in the paper |
| 73 | Jakobsen et al., Wealth Taxation and Wealth Accumulation: Theory and Evidence from Denmark | right | depends on a disputed choice of method | defensible, stated in the paper |
| 74 | Jakobsen et al., Wealth Taxation and Wealth Accumulation: Theory and Evidence from DenmarkThe 77% reduction comes from several controls together, not income deciles alone, and earlier drafts found much smaller effectsholds only for some comparisons. | partly right | depends on a disputed choice of method | defensible, stated in the paper |
| 75 | Jakobsen et al., Wealth Taxation and Wealth Accumulation: Theory and Evidence from Denmark | right | depends on a disputed choice of method | questionable |
| 76 | Johannesen and Zucman, The End of Bank Secrecy? An Evaluation of the G20 Tax Haven CrackdownThe numbers reproduce. Treaty-induced outflows are too small to show in the totals, so rather than repatriatingis stronger than the evidence. | right | does not change; a claim is overstated in words | not graded |
| 77 | Kleven et al., Taxation and International Migration of Superstars: Evidence from the European Football Market | could not check | could not check | not graded |
| 78 | Kleven et al., Taxation and International Migration of Superstars: Evidence from the European Football Market | could not check | could not check | not graded |
| 79 | Kleven et al., Taxation and International Migration of Superstars: Evidence from the European Football Market | could not check | could not check | not graded |
| 80 | Kleven et al., Taxation and International Migration of Superstars: Evidence from the European Football Market | could not check | could not check | not graded |
| 81 | Kuziemko et al., How Elastic Are Preferences for Redistribution? Evidence from Randomized Survey ExperimentsThe effects stand. The prime also moved a second attitude, but the paper itself calls its analysis of the mechanism not definitive. | right | does not change | not graded |
| 82 | Landais et al., Le modèle de micro-simulation TAXIPP - Version 0.0 | right | depends on a disputed choice of method | defensible, stated in the paper |
| 83 | Landais et al., Le modèle de micro-simulation TAXIPP - Version 0.0 | right | could not check | not graded |
| 84 | Landais et al., A Progressive European Wealth Tax to Fund the European COVID Response | right | depends on a disputed choice of method | questionable |
| 85 | Murphy and Saez, Income and Wealth in America | right | depends on a disputed choice of method | cannot tell |
| 86 | Novokmet et al., From Communism to Capitalism: Private versus Public Property and Inequality in China and Russia | right | depends on a disputed choice of method | defensible, stated in the paper |
| 87 | Novokmet et al., From Soviets to Oligarchs: Inequality and Property in Russia, 1905-2016 | right | depends on a disputed choice of method | defensible, stated in the paper |
| 88 | Novokmet et al., From Soviets to Oligarchs: Inequality and Property in Russia, 1905-2016 | right | depends on a disputed choice of method | defensible, stated in the paper |
| 89 | Parrinello et al., Resources for a Safe and Resilient Europe: The Case for Minimum Taxation of Ultra-High-Net-Worth Individuals in the EURevenue figures are unaffected. Neutralize the regressivityis contradicted by caveats in the note’s own source. | right | does not change; a claim is overstated in words | not graded |
| 90 | Piketty and Saez, Optimal Labor Income TaxationThe regressions reproduce. The chapter offers above 80%only as a possibility, with hedges. | right | does not change | not graded |
| 91 | Piketty and Saez, Top Incomes and the Great Recession: Recent Evolutions and Policy Implications Over 15% of national incomeuses shares of fiscal income; with the paper’s own Table 3, the shift is about 10% of national income. | right | changes (shown at the time) | not graded |
| 92 | Piketty et al., Optimal Taxation of Top Labor Incomes: A Tale of Three Elasticities | right | depends on a disputed choice of method | defensible, stated in the paper |
| 93 | Piketty et al., Distributional National Accounts: Methods and Estimates for the United States | right | depends on a disputed choice of method | defensible, stated in the paper |
| 94 | Piketty et al., Simplified Distributional National AccountsThe arithmetic reproduces and no reported number changes. The exercise cannot show where the two estimates differ, so its unrealisticverdict is not shown. | right | does not change; a claim is overstated in words | not graded |
| 95 | Piketty et al., Rethinking Capital and Wealth Taxation | right | depends on a disputed choice of method | questionable |
| 96 | Piketty et al., Income Inequality in the United States: A Comment | right | depends on a disputed choice of method | defensible, stated in the paper |
| 97 | Piketty et al., Capital Accumulation, Private Property, and Rising Inequality in China, 1978-2015 | right | depends on a disputed choice of method | questionable |
| 98 | Piketty and Zucman, Capital is Back: Wealth-Income Ratios in Rich Countries, 1700-2010 | right | depends on a disputed choice of method | defensible, stated in the paper |
| 99 | Piketty and Zucman, Wealth and Inheritance in the Long Run | right | depends on a disputed choice of method | questionable |
| 100 | Piketty and Zucman, Wealth and Inheritance in the Long Run | right | depends on a disputed choice of method | standard |
| 101 | Reck et al., Response to a Comment by Auten and Splinter onThe reply states +0.9 to +1.2 points (p. 23), not +1.5. No printed number changes.Tax Evasion at the Top of the Income Distribution: Theory and Evidence If anything too lowis not borne out by later versions (0.8 to 1.1 points). | partly right | does not change; a claim is overstated in words | not graded |
| 102 | Saez, Striking it Richer: The Evolution of Top Incomes in the United States (annual updates) | right | depends on a disputed choice of method | defensible, not stated |
| 103 | Saez, Taxes and International Mobility of Talent | right | depends on a disputed choice of method | standard |
| 104 | Saez, Income and Wealth Inequality: Evidence and Policy ImplicationsThe values reproduce. A 1962-2014 US series is used to rule out avoidance over a centuryand across countries, which overstates its reach. | right | does not change; a claim is overstated in words | not graded |
| 105 | Saez, The Research Agenda after Capital in the Twenty-First Century | right | depends on a disputed choice of method | standard |
| 106 | Saez, Taxing the Rich More: Preliminary Evidence from the 2013 Tax IncreaseThe paper’s own method on the final 2015 IRS data gives a 31.5% revenue loss, not at most 20%. The paper marked its 2015 data as preliminary. | right | changes (shown by later data) | not graded |
| 107 | Saez, Taxing the Rich More: Preliminary Evidence from the 2013 Tax Increase | right | depends on a disputed choice of method | defensible, stated in the paper |
| 108 | Saez and Seim, Wealth Tax Enforcement in Sweden: Filing Requirements and Pre-Populated Returns | right | depends on a disputed choice of method | defensible, stated in the paper |
| 109 | Saez and Stantcheva, Generalized Social Marginal Welfare Weights for Optimal Tax TheoryA published theorem (Sher 2024) shows these weights produce preference cycles, so the unifyclaim is overstated. No number changes. | right | does not change; a claim is overstated in words | not graded |
| 110 | Saez and Zucman, Who Benefits from Tax Expenditures on Capital? Evidence on Capital Income and Wealth Concentration | right | depends on a disputed choice of method | defensible, stated in the paper |
| 111 | Saez and Zucman, Wealth Inequality in the United States since 1913: Evidence from Capitalized Income Tax DataThe authors’ own 2020 and 2022 data releases put the 2012 top 0.1% wealth share at about 20%, not 22%. | right | changes (shown by later data) | not graded |
| 112 | Saez and Zucman, Progressive Wealth Taxation | right | depends on a disputed choice of method | defensible, stated in the paper |
| 113 | Saez and Zucman, Response to Summers and Sarin,A wealth tax presents a revenue estimation puzzle | right | depends on a disputed choice of method | defensible, stated in the paper |
| 114 | Saez and Zucman, Scoring of Senator Sanders’ Wealth Tax ProposalThe $4.35 trillion and the simulation stand. Fully eliminate the gapis contradicted by the note’s own figure: the gap shrinks by about two-thirds. | right | does not change; a claim is overstated in words | not graded |
| 115 | Saez and Zucman, Scoring of Senator Sanders’ Wealth Tax Proposal | right | depends on a disputed choice of method | defensible, stated in the paper |
| 116 | Saez and Zucman, The Triumph of Injustice: How the Rich Dodge Taxes and How to Make Them Pay | right | depends on a disputed choice of method | cannot tell |
| 117 | Saez and Zucman, The Triumph of Injustice: How the Rich Dodge Taxes and How to Make Them Pay | right | depends on a disputed choice of method | defensible, stated in the paper |
| 118 | Saez and Zucman, Scoring of Senator Warren’s Wealth Tax Proposal (letters to Senator Elizabeth Warren)The letters derive a 15% avoidance rate for a 2% tax, then apply it unchanged at 3% and 6%. Their own rule cuts the 2021 estimate of $3.9 trillion by 18-26%; the main 2019 figures move 4-6%. | right | changes (shown at the time) | not graded |
| 119 | Saez and Zucman, Scoring of Senator Warren’s Wealth Tax Proposal (letters to Senator Elizabeth Warren) | right | depends on a disputed choice of method | defensible, stated in the paper |
| 120 | Saez and Zucman, The Rise of Income and Wealth Inequality in America: Evidence from Distributional Macroeconomic Accounts | right | depends on a disputed choice of method | defensible, stated in the paper |
| 121 | Saez and Zucman, Trends in US Income and Wealth Inequality: Revising After the RevisionistsThe flag says the residual is mostly not partnership depreciation; the paper’s 1.6% and the flag’s own 1.42% put that item at about half. | partly right | depends on a disputed choice of method | defensible, stated in the paper |
| 122 | Saez and Zucman, How to Get $1 Trillion from 1000 Billionaires: Tax their Gains Now | right | depends on a disputed choice of method | defensible, not stated |
| 123 | Saez and Zucman, Top Wealth in America: A Reexamination | right | depends on a disputed choice of method | defensible, stated in the paper |
| 124 | Saez and Zucman, Wealth Taxation: Lessons from History and Recent DevelopmentsOn the World Inequality Database figures available in October 2021, benchmark revenue for Spain, Norway and Switzerland is 0.71, 0.45 and 1.16, not the printed 0.96, 0.86 and 1.36. | right | changes (shown at the time) | not graded |
| 125 | Silva et al., Progressive Taxation and Inequality: Minimum Taxation of High-Net-Worth Individuals in Latin America and the Caribbean | right | depends on a disputed choice of method | defensible, not stated |
| 126 | Silva et al., Progressive Taxation and Inequality: Minimum Taxation of High-Net-Worth Individuals in Latin America and the Caribbean | right | depends on a disputed choice of method | defensible, stated in the paper |
| 127 | Tippet and Zucman, Taxing Extreme Wealth in the United Kingdom: Revenue Estimates for a 2% Minimum Tax on Ultra-High-Net-Worth Households | right | depends on a disputed choice of method | questionable |
| 128 | Tippet and Zucman, Taxing Extreme Wealth in the United Kingdom: Revenue Estimates for a 2% Minimum Tax on Ultra-High-Net-Worth Households | right | depends on a disputed choice of method | defensible, stated in the paper |
| 129 | Tørsløv et al., Externalities in International Tax Enforcement: Theory and Evidence | right | depends on a disputed choice of method | defensible, stated in the paper |
| 130 | Tørsløv et al., The Missing Profits of NationsThe 36% headline stands. A units mismatch in the replication spreadsheet means that taking tax-haven data at face value cuts shifted profits by about half, not 18%, and the stated lower bound falls from 22% to 10-12%. | right | does not change; a claim is overstated in words | not graded |
| 131 | Tørsløv et al., The Missing Profits of Nations | right | depends on a disputed choice of method | defensible, stated in the paper |
| 132 | Tørsløv et al., The Missing Profits of Nations | right | depends on a disputed choice of method | defensible, stated in the paper |
| 133 | Wright and Zucman, The Exorbitant Tax PrivilegeRunning the counterfactual the paper describes on the authors’ workbook gives 3.2 to 3.8 points of the 14.5-point fall, not almost half. | right | changes (shown at the time) | not graded |
| 134 | Zucman, La richesse cachée des nations: enquête sur les paradis fiscaux / The Hidden Wealth of Nations: The Scourge of Tax Havens | right | depends on a disputed choice of method | questionable |
| 135 | Zucman, La richesse cachée des nations: enquête sur les paradis fiscaux / The Hidden Wealth of Nations: The Scourge of Tax Havens | right | depends on a disputed choice of method | cannot tell |
| 136 | Zucman, The Missing Wealth of Nations: Are Europe and the U.S. net Debtors or net Creditors? | right | depends on a disputed choice of method | defensible, stated in the paper |
| 137 | Zucman, The Missing Wealth of Nations: Are Europe and the U.S. net Debtors or net Creditors? | right | depends on a disputed choice of method | questionable |
| 138 | Zucman, The Missing Wealth of Nations: Are Europe and the U.S. net Debtors or net Creditors? | right | depends on a disputed choice of method | defensible, stated in the paper |
| 139 | Zucman, Taxing across Borders: Tracking Personal Wealth and Corporate Profits | right | depends on a disputed choice of method | defensible, stated in the paper |
| 140 | Zucman, Taxing across Borders: Tracking Personal Wealth and Corporate Profits Overstated by about a quarterholds only if shifted profits would otherwise pay the 23% average rate; at the 35% statutory rate, the paper’s about 20 percentsurvives. | partly right | depends on a disputed choice of method | defensible, not stated |
| 141 | Zucman, Taxing across Borders: Tracking Personal Wealth and Corporate Profits | right | depends on a disputed choice of method | defensible, stated in the paper |
| 142 | Zucman, Inequality: Are we reallyThe summary box says 1.6 to 3 times; the briefing’s own text and Figure 2 give 1.65 to 1.97 times.all in this together? | right | changes (shown at the time) | not graded |
| 143 | Zucman, Taxing Multinational Corporations in the 21st CenturyThe brief’s 40% is higher than its source’s 36.2%. A small difference. | right | changes (shown at the time) | not graded |
| 144 | Zucman, Can a Wealth Tax Work? | right | depends on a disputed choice of method | cannot tell |
| 145 | Zucman, A Blueprint for a Coordinated Minimum Effective Taxation Standard for Ultra-High-Net-Worth IndividualsOn the report’s own figures, 7.1% is total group wealth growth; per member it is about 5.4%, implying a return near 5.8%, not 7.5%. Revenue is unaffected. | right | changes (shown at the time) | not graded |
| 146 | Zucman, The Billionaire Tax: A (Modest) Proposal for the 21st Century | right | depends on a disputed choice of method | defensible, not stated |
| 147 | Zucman et al., Introduction : Où est l’argent ?The interview says the share doubled from 20% to 40%; the authors’ own series give about 24% to 40%, a rise of about 1.6 times. | right | changes (shown at the time) | not graded |
could not check.
changes a headline numberverdict against our written rules again. Of those, 4 did not meet them: 62, 84 and 127 rest on a disputed choice of method, and in 130 the headline figure does not change.
Changesneeds a calculation or the work’s own numbers.
Depends on a disputed choice of methodmeans reasonable researchers differ. Where the authors’ own data releases disagree with each other, we count the flag there too.
newly foundand
previously knowntags, and says no finding overturns a core result outright.