Brian Galle, David Gamage, Emmanuel Saez, Darien Shanske · Response note, 17 March 2026; SSRN 6433598

Response to 'The Net Present Value of the Billionaire Tax Act' March 4, 2026

9 problems found: 2 moderate · 7 minor

What the paper claims

Rauh et al. (2026) estimate billionaire mobility as though taxpayers faced a recurring, permanent 5% annual wealth tax. They count the resulting California income-tax losses in perpetuity but the wealth-tax revenue only once. This 'unreasonable asymmetry is the largest flaw' and is what produces their negative net present value.
On the critics' own terms, a permanent 5% tax would raise 'hundreds of billions', far more than the income tax lost. With their elasticity the revenue-maximising permanent rate is ~10% (1/semi-elasticity). Counting losses as permanent at a 1.5% discount rate magnifies the annual income-tax loss 67-fold, which 'explains' their negative NPV.
The one-time 5% CBTA (payable in five installments plus an interest charge, 'roughly equivalent to a 1% annual tax' for five years) raises about $100bn with 'comparatively very modest' income-tax losses. The authors' prior analysis and the LAO each 'conclude' it would 'on net raise tens of billions'.
The critics' assumption that 'nearly 30%' of the base has already departed rests on press reports of attempted moves (6 pre-snapshot billionaires, 4 later) that are unlikely to sever California residency as of 1 Jan 2026. 'There is no real evidence that any significant number' of end-2025 Forbes California billionaires will avoid the tax. A challenge to the 1 Jan 2026 date is 'unlikely' to succeed.
Smaller points. Billionaires pay ~$3bn/yr of California income tax, not the critics' $4.55bn (0.24% of wealth); the gap arises 'primarily' from assuming 78% rather than 50% of top-212 earners' income. Page, Brin and Zuckerberg (>90% of the departed wealth) paid only $269m in 2025 (0.04% of wealth; $222m/yr in 2019-25), about 0.2% of $120bn of state income tax. Ellison ($245bn, >10% of the base) may not be a resident. Both sides omitted 24 non-US-citizen residents worth $110bn (~5%).

Every problem found

1moderateconfirmed · previously known · methodology

The note says it makes no sense to count lost income taxes forever while counting the new tax only once. That logic works against the critics' guess about extra unannounced departures, which assumed the tax would be charged every year. It does not settle the case of people who genuinely moved to dodge the one-time tax. Moving back is costly, so they may stay away and their California income tax may be lost for good, and the note gives no evidence that they will return. This does not overturn its conclusion, because the people reported to have left pay relatively little California income tax.

Technical detail and evidence

Why it matters, in detail. Critics' March framework with WT fixed at observed departures ($67.51bn, f = 0.283) and perpetual losses: mean NPV +$20.3bn, 13.9% of draws negative. The sign is negative only for low r-g with proportional allocation (-$18.4bn at 1.5%; +$24.5bn at 3%; C = $4.55bn). With person-specific income tax (Page $87m, Brin $56m in 2025), the six movers' loss is ~$0.24-0.32bn/yr and NPV ranges from +$46bn to +$62bn (mean +$57bn) on the critics' revenue base, and +$49-62bn on the GGSS base. Persistence therefore moves the magnitude of the net by up to ~$10-20bn but not its sign on the note's own evidence. The note's case rests on the elasticity mismatch and on person-specific income tax, not on losses being temporary.

Evidence. Note pp.1-2, 4. Codex counterexample: $1bn of wealth, $50m levy, $25m relocation cost each way and a $1m/yr amenity cost discounted at 5%. Moving costs $45m < $50m, so the person moves; returning gains $20m < $25m, so the person does not return, even with no future wealth tax. Critics' later rejoinder (Jaros & Rauh 2026, NBER w35899, sec. 5.2 and App. D 'Resolution I'): 'a pre-announced one-time levy with an avoidable snapshot date can generate large behavioral responses even if taxpayers expect no future wealth taxes'. Adjudicator figures (scripts/adjudicator_reproduction.py): the six reported movers' annual loss with person-specific income tax (BSZ Table 3 for Page and Brin; average rate for the rest) is ~$0.28bn, PV $18.7bn at r-g = 1.5% and $9.3bn at 3% if perpetual, and $4.3bn / $3.5bn with a 5%/yr return hazard. Under the critics' wealth-proportional allocation the loss is $1.29bn/yr, PV $86bn / $43bn. BSZ (Aug 2026, Table 5) label the leavers' loss 'temporary' without evidence on return rates.

Why this severity. It removes one of the note's arguments against perpetual losses but not its conclusion. The negative sign of the critics' NPV comes from the elasticity-predicted departures, which the argument does defeat. For the observed movers, losses are small on the note's person-specific income-tax evidence even if perpetual (PV $9-19bn against at least $68bn of revenue). They decide the sign only under the critics' proportional allocation and a low discount rate (observed-only NPV -$18.4bn at r-g = 1.5%, C = $4.55bn). Codex's 'major' is not supported because the conclusion survives through the other two arguments.

The authors' response. No specific response found as of Oct 2026. BSZ (Aug 2026, Table 5) continue to describe the leavers' income-tax loss as 'temporary'.

2moderatelikely · previously known · overclaiming

The note repeats that the tax would raise $100 billion and says there is no real evidence that billionaires have escaped it. But two of its own admissions push the figure down to about $92 billion. One very rich man was counted who may not live in California, and a group of foreign-citizen residents was left out. If several reported moves out of the state hold up in law, the figure could be $70-90 billion. The tax would still raise a lot of money, but $100 billion is the hopeful end of the range.

Technical detail and evidence

Why it matters, in detail. Static score = 0.05 x (1 - 0.10) x W. W = $2,182bn gives $98.2bn; the note's own concessions give $92.1bn (authors' later tracker: $93.5bn on 17 Mar 2026; $86-100bn range Jan-Apr 2026). Six reported moves effective: $68.0bn (-31% vs $98bn). Authors' aggressive case: $84bn (May) / $88.8bn (Aug, at July 2026 prices, from a $103.8bn benchmark). The 10% haircut is assumed, not estimated. Snapshot-date litigation risk (sec. 50311(b) reformation) is unquantified. The critics' later NBER version uses q = 0.5, an accounting-standard threshold rather than an estimate. At March 2026 prices and before litigation risk, the defensible range is roughly $68-100bn, not a $100bn point. The sign of the net fiscal effect does not change.

Evidence. Arithmetic (scripts/adjudicator_reproduction.py): 2,182 x 0.05 x 0.9 = $98.2bn; minus Ellison ($245bn, 11.2% of the base) plus non-US citizens ($110bn) = $92.1bn. The authors' own tracker series (cabillionairetracker.org; first archived 26 May 2026, i.e. reconstructed after the note) gives $93.5bn on 17 Mar 2026. The GGSS April 2026 report (fn 6) says the score 'hovered in the range $86 billion to $100 billion' after 1 Jan 2026. If the critics' six pre-snapshot moves ($536bn) were legally effective, the same method gives $68.0bn. Boll, Saez & Zucman (NBER w35218, May 2026, PDF p.23, and Table 5 row 3; PDF p.26 of the Aug 2026 version) say Page 'could possibly successfully argue he left California before January 1, 2026' and Thiel, Hankey and Kalanick 'may have moved before January 1'. They score that case at $84bn (May; $88.8bn in Aug at July prices). Before or around the note's date, Forbes had re-listed Snyder (12 Jan), Hankey (15 Jan) and Kalanick (16 Mar) outside California (tracker movers file). Forbes's own Forbes 400 2025 list already placed Ellison in Lanai, Hawaii. The April 2026 report removed him as 'very likely a Florida resident since March 2023'. Forbes's Oct 2026 re-listing of Page (Florida) and Brin (Nevada) shows where they live now, not their legal residence on 1 Jan 2026; BSZ consider Brin very likely still a resident. Codex's sensitivity table reproduces the $92.1bn and the effect of 20-40% erosion ($87/76/65bn) but does not treat it as a finding.

Why this severity. The level of a reiterated headline is overstated by ~6-8% on the note's own concessions, by ~10-15% in the authors' later 'aggressive leavers' case, and by up to ~30% if all six reported pre-snapshot moves hold. The qualitative claim that the tax raises tens of billions net survives every scenario examined: +$49-62bn NPV even with all six moves effective and perpetual person-specific income-tax losses. It qualifies a secondary, reiterated claim, not the note's critique of the critics.

The authors' response. The note argues that announced moves are unlikely to be legally effective (pp.3-4) and that a challenge to the 1 Jan 2026 date is unlikely to succeed (fn 2). Later, the GGSS expert report (Apr 2026) removed Ellison and added non-US citizens. BSZ (May/Aug 2026, Table 5 row 3) added an 'aggressive leavers' scenario ($84-89bn), which they call an aggressive assumption.

3minorconfirmed · newly identified here · misrepresentation of source

The note mainly blames the critics' negative result on a very low discount rate that multiplies yearly losses 67 times. The critics' headline actually averaged a range of rates, which gives a multiplier of about 37, and their result stays negative at a middle rate. What really drives it is their assumption of many extra unannounced departures, which the note criticises correctly elsewhere. The misstatement is in the argument, not in the conclusion.

Technical detail and evidence

Why it matters, in detail. The perpetuity multiplier at the critics' lower-bound rate (66.7x) is ~1.8x the effective multiplier behind the headline (36.6x); the authors later use 33x at 3%. The critics' headline is robust to the discount rate (mean -$17.9bn at 3%) but not to removing the elasticity-predicted departures (+$20.3bn). Observed departures are ~62% of the mean departure fraction, so 'almost the entirety' overstates.

Evidence. NPV_dist.R (r_discount <- runif(n, 0.015, 0.045)); NPV_calculations_5.2.xlsx D35-D50; E[1/r] = ln(3)/0.03 = 36.62; scripts/adjudicator_reproduction.py. Co-author Saez's later paper restates the critique as 'magnified by a factor 33 (using their central discount rate 3.0%)' (Boll, Saez & Zucman, Aug 2026, fn 29). In the authors' favour: the critics' later NBER version anchors r-g on a 1.5% dividend yield (sec. 5.2), so '1.5%' has some basis as 'their' rate.

Why this severity. A rhetorical attribution error. The note's substantive diagnosis, that the elasticity step drives the sign, is correct, and correcting the multiplier changes no conclusion.

The authors' response. Implicitly corrected later: Boll, Saez & Zucman (Aug 2026, fn 29) give the multiplier as 33 at the critics' 'central discount rate 3.0%'.

4minorconfirmed · newly identified here · misrepresentation of source

The note says the state's nonpartisan Legislative Analyst's Office concluded the tax would raise tens of billions of dollars on net. The LAO actually listed two separate effects: a temporary gain of tens of billions, and a lasting loss of hundreds of millions of dollars or more each year. It never added them up. The note's own conclusion may still be right, but the LAO did not reach it.

Technical detail and evidence

Why it matters, in detail. LAO: one-time 'tens of billions' versus an ongoing loss of 'hundreds of millions or more' per year. The break-even perpetual loss for $20-40bn is $0.3-0.6 / $0.6-1.2 / $0.9-1.8bn per year at r-g = 1.5 / 3 / 4.5%, so the LAO statement is agnostic on the sign of the NPV, the very quantity in dispute.

Evidence. LAO/DoF letter to the Attorney General, 11 Dec 2025, 'Fiscal Effects' and 'Summary of Major Fiscal Effects' (https://oag.ca.gov/system/files/initiatives/pdfs/fiscal-impact-estimate-report%2825-0024A1%29.pdf; re-downloaded, identical to the cached copy). The perpetual loss that would offset $20-40bn is $0.6-1.2bn/yr at r-g = 3% ($0.3-0.6bn at 1.5%), inside the LAO's open-ended 'hundreds of millions ... or more' range. The same attribution recurs in the Feb, Apr and Jul 2026 expert reports ('a point noted by the LAO').

Why this severity. It misstates a third party's conclusion that the note uses as corroboration, but it does not affect the note's own numbers or its critique. The LAO's ranges are compatible with, though they do not establish, a positive net. Codex graded it moderate; the misattribution is clear, but its consequence for the note's findings is small.

5minorconfirmed · newly identified here · documentation gap

The note says the three richest people reported to be leaving paid only $269 million in California income tax in 2025. That is an outside estimate built from public stock filings, it counts only income from their main company shares, and it came from a paper not yet published when the note appeared. Even allowing for their other income, their tax is small relative to their wealth, so the note's point survives.

Technical detail and evidence

Why it matters, in detail. The person-specific allocation cuts the six reported movers' annual income-tax loss from ~$1.29bn (pro rata to wealth, C = $4.55bn) to ~$0.28bn, or ~$0.3-0.4bn with the omitted-asset adjustment. This step makes the observed-only NPV robustly positive (+$57bn, no negative draws in the critics' framework). The scope caveat moves that by a few billion of PV at most.

Evidence. BSZ May 2026 workbook Tab3!B12:D12 = 87.07 + 56.25 + 125.98 = $269.30m; 2019-25 averages 75.83 + 51.70 + 95.09 = $222.6m; $269.3m / $695.3bn company wealth = 0.039%; / $120bn = 0.22%. The Table 3 notes say 'focusing solely on their wealth from their company stakes'. Residual Forbes wealth is $25.1bn: at the critics' 0.24% ratio it adds +$60m (+22%); at a 4% taxable yield taxed at 13.3% it adds +$134m (+50%) (Codex). Consistency checks by both auditors: Brin's 2024-26 Form 4s are gifts only; Alphabet's FY2025 dividend was $0.83/share; Brin's holdings x dividend x 13.3% after modelled giving is about $55.5m vs BSZ's $56.25m.

Why this severity. A problem of verifiability and wording only. Adding plausible income on the omitted assets leaves the trio at ~0.05-0.06% of wealth, far below the critics' 0.24%, so the conclusion that their departure has small income-tax consequences stands. Codex's 'moderate' overstates its effect on the note's conclusions.

The authors' response. BSZ (NBER w35218, May 2026; revised Aug 2026) later published the estimates and stated their scope in the Table 3 notes.

6minorconfirmed · newly identified here · internal inconsistency

The note says billionaires pay about $3 billion a year in California income tax, compared with the critics' $4.55 billion. The $3 billion is an average over several earlier years, while $4.55 billion is for 2025, a strong year. The authors' own later estimate for 2025 is about $4.1-4.35 billion, close to the critics'. The note itself calls this difference modest.

Technical detail and evidence

Why it matters, in detail. C enters the critics' NPV linearly: using $4.35bn instead of $3bn raises the PV of any given departure fraction by ~45%. The 2025 versus multi-year difference is of the same order as the 78%/50% difference that the note credits with the whole gap. The 50% ratio itself is well supported (45-51% nationally; SCF ~50%) and if anything generous.

Evidence. Critics' NPV workbook B6:B8 (5.8, 3.3, midpoint 4.55 = 78.4% of 5.8); 4.55/1,894.8 = 0.24%. BSZ May/Aug workbooks Tab2!C13-C14; billionairesCAinctax!O81, O85, O88. Codex reproduced the $3bn only after applying adjustments first disclosed in the later BSZ workbook (fixed N=200: 2.49% of PIT = $2.98bn). A literal Pareto reading of the Feb report's description gave 3.27% ($3.93bn).

Why this severity. It affects a secondary claim that the note itself calls a 'relatively modest' discrepancy. For a perpetuity calculation a multi-year average is arguably the right concept, so no conclusion changes.

The authors' response. BSZ (May/Aug 2026, Table 2) report both the 2025 figure and the 2019-25 average; BSZ fn 13 repeats the 78.4% vs 50% explanation.

7minorconfirmed · previously known · arithmetic or transcription

The note says Forbes switched Larry Ellison's home from California to Florida on 10 January 2026. Archived Forbes pages show it first switched to Hawaii, then to Florida in March, and another Forbes list had already put him in Hawaii in late 2025. The slip changes no number, but it shows his inclusion in the $100 billion was shakier than the note implies.

Technical detail and evidence

Why it matters, in detail. The timeline error has no direct effect. Including Ellison is worth ~$11bn of the $98bn score (5% x 90% x $245bn), offset by about half by the omitted non-US citizens ($110bn, ~$5bn). See F1.

Evidence. scripts/ellison_residence.py on 15 Wayback captures (Jun 2025-Mar 2026) in /mnt/zdata/downloads/ggss2026-rauhreply-claude/ellison/. scripts/forbes_lists.py: Billionaires 2021-2023 = Lanai, Hawaii; 2024-2025 = Woodside, California; 2026 = Manalapan, Florida; Forbes 400 2021-2022 and 2025 = Lanai, Hawaii; Forbes 400 2023-2024 = Woodside. $245bn / $2,182bn = 11.2%.

Why this severity. The date (10 Jan) and the substance (Forbes stopped listing him in California) are right, and the slip has no numerical effect. Its consequence for the base is covered under F1.

The authors' response. The GGSS expert report (Apr 2026) removed Ellison as 'very likely a Florida resident since March 2023'.

8minorconfirmed · newly identified here · internal inconsistency

The note calls the tax 'roughly 1% a year'. In fact it is 5% charged once, which taxpayers may pay in five yearly instalments. Spread over the long run, a single 5% charge is worth much less than 1% a year, so the loose wording actually understates the note's own point.

Technical detail and evidence

Why it matters, in detail. With the Swiss-calibrated semi-elasticity (10.32), base erosion would be 51.6% at 5%/yr (linear), 10.3% at 1%/yr, and 0.8-2.3% at r x 5% (r = 1.5-4.5%), against 28.3% observed. The note's 1% understates how far apart the critics' pricing and literal one-time pricing are.

Evidence. Initiative sec. 50301(b)-(c). Installments per 100 of base: 1, 1.3, 1.225, 1.15, 1.075 (nominal 5.75); PV 5.00 at 7.5%, 5.42 at 3% (scripts/adjudicator_reproduction.py). The p.1 description, 'roughly equivalent to a 1% annual tax assessed for only five years plus an interest-deferral charge', is accurate.

Why this severity. Terminology only. The imprecision cuts against the note's own case, so no conclusion is affected.

9minorplausible · previously known · overclaiming

The note assures billionaires that the tax can never exceed 5% of what their holdings are really worth. For shares traded on the stock market that is the most natural reading. For private companies, though, the text starts from a voting-control rule that the owner must disprove with strong evidence, so 'never' promises more than the text guarantees.

Technical detail and evidence

Why it matters, in detail. Under (c)(3)(C) a founder whose voting share exceeds their economic share is presumed to own the voting share of (book value + 7.5 x average book profits). Until rebutted, the effective rate on economic wealth is 5% x voting share / economic share. Exposure is mainly for private dual-class companies; no revenue effect is claimed.

Evidence. Initiative sec. 50303(b)-(c), sec. 50308(j). GGSS memo on valuation of controlling shares (version dated 5 Feb 2026), which says taxpayers 'can submit an appraisal' but does not mention the clear-and-convincing threshold. Jaros & Rauh (2026) fn 11 and App. A.

Why this severity. Peripheral to the revenue claims. It bears only on the note's suggestion that some departures were driven by a misreading of the valuation rules.

The authors' response. The GGSS memo on controlling shares (first issued Jan 2026; version of 5 Feb 2026) argues that the voting-share presumption does not apply to publicly traded companies and can be displaced by a certified appraisal.

Claims examined and rejected

Problems one auditor raised that did not survive checking.

I7: the note's 'nearly 10%' revenue-maximising permanent rate depends on a functional form the critics did not use; under the critics' linear March implementation it is 4.7%. — Not an error in the note. Footnote 1 states its convention explicitly ('Formally, the revenue maximizing rate ... is given by the inverse of the semi-elasticity'), and that is the standard meaning of a constant semi-elasticity: 1/10.32 = 9.7%, or 9.45% net of the income-tax spillover. Applied globally, the critics' linear form implies the whole base disappears at a 9.7% tax. Under the note's convention the critics' own elasticity-based revenue would be higher ($56.2bn rather than $45.6bn at 10.32; $50.4bn rather than $35.3bn at 1.25x), so the convention works against the critics, not the note. The note's conclusion, that a permanent 5% tax raises far more than the income tax lost ($45.6-56.3bn/yr vs $1.8-2.4bn/yr), holds under every form. Codex also declined to treat this as an error.
Within I1: because the critics' 'most favorable' workbook scenario has a positive NPV, the note's statement that the critics present 'only the worst case' (p.4) is too broad. — Misreads the context. The sentence concerns departures: 'At best, Rauh et al. offer a kind of worst-case scenario for the CBTA's revenue'. Every wealth-tax value in the critics' March model (WT from $35bn to $67.51bn, 'most favorable' scenario included) assumes that all six reported pre-snapshot moves are legally effective; none considers moves failing. As a description of the departure assumption, the note's characterisation is accurate.
I1 graded 'major': the asymmetry argument is a confirmed logical overclaim that materially undermines C1-C3. — Partly rejected (severity). The logical point is retained as F2 (moderate, confirmed). It is not major. The note's 'largest flaw' claim concerns mobility estimated as if the tax were a permanent 5% annual levy, and for that component the inconsistency the note identifies is real. Removing that component flips the critics' mean NPV from -$24.7bn to +$20.3bn. Persistent losses from genuinely effective one-time moves would change the net by at most ~$10-20bn of PV on the note's person-specific income-tax evidence, without changing its sign. Codex's own illustration (a 20% permanent loss of $3bn/yr = $600m/yr) is about twice the evidence-based figure for the six reported movers (~$0.28bn/yr).
I2 and I3 graded 'moderate' (LAO attribution; scope of the $269m). — Both findings are retained (F4, F5) but downgraded to minor. Neither affects the note's own estimates or its critique of the critics. The LAO ranges are compatible with a positive net, and the omitted-asset adjustment leaves the founders' tax at ~0.05-0.06% of wealth, far below the 0.24% that matters for the argument.
I2(b), first draft: the note's rational-expectations argument 'ignores the participation margin'. — Superseded. The note recognises the one-time levy's 'obvious and powerful incentive' to establish or claim an abode elsewhere (p.4). Auditor A's final report corrected this to 'treats the incentive only as a source of sham or temporary moves', which is the version retained in F2.

Audit notes

What reproduced. Reproduced independently (scripts/adjudicator_reproduction.py) from the critics' March replication files (GitHub bjaros20/wealth_tax, 6 and 10 Mar 2026 commits, byte-identical workbooks). The critics' mean NPV of -$24.71bn (71.0% negative) and their nine-cell scenario table match exactly. Decompositions: observed departures only +$20.3bn; r-g fixed at 1.5/3/4.5% gives -$87.0/-$17.9/+$5.2bn; person-specific income tax for Page and Brin is about zero over the full draw (+$3bn here, -$3.4bn in Auditor A's allocation) and +$57bn with observed departures only; perpetual-equivalent one-time pricing implies 0.8-2.3% erosion. The note's numbers reproduce: 1/0.015 = 66.7 (effective E[1/r] = 36.6); permanent-tax revenue $45.6-56.3bn/yr vs $1.8-2.4bn lost; revenue-maximising rate 9.45-9.69% (constant semi-elasticity); installments nominal 5.75, PV 5.00 at 7.5%; $98.2bn static score ($92.1bn with Ellison removed and non-citizens added; $68.0bn if the six reported moves hold); 4.55/5.8 = 78.4%; 0.5 x 5.8 = $2.9bn; $269.3m, $222.6m, 0.039%, 0.22% (BSZ Table 3); 91.8% / 91.4% departed-wealth shares. The Ellison Forbes timeline, Forbes California counts (197 in 2024) and non-citizen counts (25/23 in the 2025/2026 lists) were checked against archived Forbes data. Not reproduced: the exact 24/$110bn non-citizen figure at end-2025 (no person-level 31 Dec 2025 snapshot), and the Feb 2026 report's original 2.6% income-tax share without adjustments first disclosed later (Codex reaches $2.98bn with them). Person-level income-tax figures could be checked only for consistency with SEC filings, not against returns.

Limits of this audit. The critics' 4 March 2026 PDF (SSRN 6340778; Hoover) could not be obtained: SSRN returned 403 and Wayback holds only the abstract page. Its claims were checked against the SSRN abstract (18 Mar 2026 capture), the authors' Liberty Lens summary and Hoover press release (5 Mar 2026), and their March replication files. The note's references to 'Table 6' and 'Table 7' were matched to the workbook's 6-person 'Preferred' and 10-person 'Expanded' lists, not to the PDF. The SSRN copy of the note was not downloaded; the abstract page and page count match. Legal questions (whether the reported moves were effective, whether the retroactive 1 Jan 2026 snapshot is constitutional, how far the voting-share presumption reaches) were assessed only against the statutory text and FTB guidance, not adjudicated. Litigation risk is not quantified. The person-level income-tax figures are SEC-based estimates by the authors' own later paper. The tracker series and BSZ tables postdate the note and are used only to test its claims, not as evidence available to the authors on 17 Mar 2026. The non-citizen count at end-2025 could not be reproduced exactly. All revenue figures depend on market values at the 31 Dec 2026 valuation date.