Inheritance Tax and the UK Tech Cohort — full version

The complete paper, with international comparators and a detailed treatment of what different evidence would mean.

Summary

The reformed inheritance tax regime that took effect on 6 April 2026 caps 100 per cent Business Property Relief at £2.5 million of qualifying assets per person (transferable to spouses for a combined £5 million per couple), with 50 per cent relief above that. For unlisted UK trading-company shares above the cap — founder equity, venture LP interests, EIS portfolios in unlisted private companies — the result is an effective 20 per cent inheritance tax rate on the excess. The reform is contested in a way most reform debates are not: not over whether the change should have happened, but over whether the death-event valuation mechanism the reform uses is operationally fit for the asset class it now applies to.

Six serious positions have been advanced by knowledgeable people in good faith. Position E holds that the reform should be implemented as written: the £2.5m / £5m allowance and 50% relief above (effective 20% rate) are correct as designed, the existing IHT-at-death mechanism is the right vehicle, and operational issues will resolve as the regime beds in. Position A holds that the existing mechanism is correct in principle but that operational issues should be resolved through four practical measures within the existing regime. Position B holds that the death-event valuation mechanism is mismatched to illiquid private-company shares and should be replaced for that asset class with a Capital Gains Tax charge on actual realisation by the heir. Position C holds that the four practical measures should be adopted now and the mechanism question deferred until the evidence on relocation, dispute caseload, and receipt outcomes can be assessed. Position D holds that the existing mechanism should be retained but the £2.5m / £5m allowance should be raised specifically for qualifying unlisted trading-company shares, on the grounds that the operational mismatch is concentrated in that cohort and a scope adjustment is more targeted than a mechanism change. Position F, added on 9 May 2026 as a sixth position offered into the same analytical space, holds that the estate should elect at death — per company — between settlement at death under the reform as written and a deferred-realisation regime with a hard ten-year backstop, the rate locked at the rate that would have applied at death and the taxable base floored at the date-of-death value. Position F is the bounded version of Position B; the year-ten deemed disposal closes the door on the unbounded-deferral failure mode the case-against-B identifies.

This piece does not argue that any of these positions is correct. It sets out what each is, what they share, what their differences turn on, what comparable jurisdictions actually do, and — most importantly — what different empirical outcomes from HMRC modelling and observed receipts data would imply for which position the evidence ultimately supports. The intent is to help a reader think clearly about a question argued at high volume and low resolution. Readers will reach different conclusions depending on which empirical outcomes they consider most likely and which normative framing they treat as primary. The piece is structured to make those dependencies visible, not to resolve them.

Section 3 sets out what each of the six positions would look like in operational form — the legislation, the HMRC capacity, the changes for the affected cohort. The publication does not advocate among them; the point of giving each position equal operational treatment is to let the reader see what each would actually involve.

1. What is actually in dispute

The public debate over the IHT reform reads as a clash between irreconcilable positions: on one side, founder representatives arguing the reform will drive talent abroad and damage the UK's growth prospects; on the other, defenders of the reform arguing that any softening would be capitulation to the wealthy. Both framings overstate the disagreement.

Six serious positions have been advanced and they agree on the principle and on the £2.5m / £5m allowance structure. Four of the six (A, B, C, D) support adopting four practical measures in addition to whatever mechanism change they propose: a collateral-trigger rule to close the buy-borrow-die avoidance vector; a statutory safe-harbour valuation methodology to reduce SAV dispute volume; targeted tightening of temporary non-residence rules to capture relocation; and practical guidance on structured buy-backs under existing instalment provisions for the cohort where they work. Position E (reform as written) holds that these measures are not yet necessary and should be considered only if operational issues prove material in practice. Position F (added 9 May 2026) is a structural addition to the menu rather than a position on the practical measures; under F's Regime 1 the practical measures apply if A is the underlying mechanism overlay, and under F's Regime 2 most of them are subsumed by the deferred-realisation design. The four practical measures are common ground for A through D.

One scope note. This piece assumes the principle of the reform — that very large illiquid private holdings should be brought into the inheritance tax base above an allowance. It does not defend that assumption. A reader who thinks the reform itself was wrong in principle will find the analysis here operating one level too deep: the disagreement among Positions A through F is a disagreement among people who already accept that the reform should have happened. The publication treats the principle question separately in the principle piece, which presents both sides of that prior question without picking. The current piece takes the principle as given.

The disagreement is genuine. Position E holds that the reform as enacted is correctly designed and the operational concerns are overstated. Position A holds that the principle is right but the four practical measures are needed to make the regime work properly. Position B holds that the death-event mechanism is unfit for unlisted trading-company shares and should be replaced for that asset class. Position C holds that the mechanism question should be deferred pending evidence. Position D holds that a scope adjustment (a higher threshold for the cohort where the operational mismatch is sharpest) is the cleaner answer than mechanism change. Position F holds that the estate should be allowed to elect, per company, between the death-event regime and a bounded ten-year deferred-realisation regime, with anti-avoidance rules sufficient to prevent the unbounded-deferral failure mode B is criticised for. That is the actual policy choice.

The six positions, on their proponents' best terms

Each of the six positions is presented below in two paragraphs: one giving the strongest case for the position in the voice of its strongest defender, and one giving the strongest case against the position in the voice of its strongest critic. The two paragraphs are presented without rhetorical scaffolding from the publication and without a closing verdict on which position carries. The six positions are presented in the same length and same register so that a reader can weigh them against each other on equal terms.

Position E — Reform as written. Implement as enacted. Maintain the £2.5m / £5m allowance, 50% relief above (effective 20% rate), the ten-year interest-free instalment regime, and the existing IHT-at-death mechanism. Do not adopt any of the four practical measures preemptively. Allow the regime to operate, observe what actually happens, and revisit only if material problems emerge.

The case for Position E. The reform was the product of extensive Treasury and HMRC work, public consultation through the Finance (No. 2) Bill 2024-26 process, the December 2025 amendments that raised the allowance from £1m to £2.5m and introduced spousal transferability, and the legislative process culminating in Finance Act 2026 receiving Royal Assent on 18 March 2026. The IFS, Resolution Foundation, and CenTax (the three most-respected independent UK fiscal-policy researchers on this question) have broadly welcomed the reform's principle, with CenTax proposing better-targeted alternative designs but accepting the principle. The ten-year interest-free instalment regime has been in IHT law since 1984 and is well-understood by practitioners; the SAV process exists and works for valuation; the temporary non-residence rules are part of established UK tax architecture. The forecast cohort is small (HMRC December 2025 estimate: approximately 1,100 estates per year affected by the whole APR/BPR reform, of which approximately 220 are BPR-only excluding AIM-only). Predicted operational catastrophes when a tax newly applies to a wealthy cohort have, in past UK tax-policy episodes, consistently been larger in advance-prediction than in actual outcome (top-rate income tax, mansion tax, the abolition of the non-dom regime, capital gains tax on second homes). The honest response is to allow the regime to operate as designed; if material problems emerge, the practical measures and mechanism alternatives can be considered with the benefit of actual evidence rather than predicted evidence. Pre-emptive amendment of a regime that has just received Royal Assent risks instability that the affected cohort has more difficulty planning around than it has with the enacted regime.

The case against Position E. Position E asks the affected cohort and their advisers to absorb operational issues — SAV valuation disputes, liquidity at death for genuinely illiquid estates, pre-death behavioural pressure — that practitioners with no personal stake in the outcome (the CIOT, the Family Business Research Foundation, several major firms in their published commentary) have argued are real and material. The four practical measures (collateral-trigger; statutory SAV safe-harbour; targeted TNR tightening; structured buy-back guidance) are independently valuable and would not need to wait for material problems to manifest before being adopted. Position E's "wait and see" posture has the form of caution but the substance of asking the affected cohort to bear the operational costs of mechanism choices that practitioners with relevant expertise have argued for adjusting. The political-economy argument also cuts against E: by the time material problems are visible enough to justify amendment, the cohort whose mobility is the consequential variable will have made its decisions on the regime as enacted, and revisions will be remediation rather than prevention. Position E rests on a generalisation about wealthy cohorts overstating their mobility; whether the BPR-affected cohort's specific mobility characteristics resemble the populations that historical generalisation was drawn from is unmeasured.

The other design positions

Position A — Hold the existing mechanism with practical fixes. Maintain IHT-at-death as the taxable event for unlisted trading-company shares. Adopt the four practical measures (collateral-trigger; statutory SAV safe-harbour; targeted TNR tightening; structured buy-back guidance). Resource SAV adequately to handle the dispute caseload.

The case for Position A. The principle of the reform — that very large illiquid private holdings should be brought into the inheritance tax base — is delivered most cleanly by a regime that taxes wealth at the point of generational transfer. Death is the moment ownership transfers; the tax event aligns with the moment the legal relationship between asset and owner changes. The cohort directly affected is small (HMRC's December 2025 estimate is approximately 1,100 estates affected by the whole APR/BPR reform in 2026-27, of which approximately 220 are BPR-only excluding AIM-only — the precise unlisted-trading-company-share subset on which the operational mechanism question primarily turns). Operational issues exist but are tractable through capacity expansion and methodology publication. Mechanism change for one asset class would create a special carve-out for the wealthiest holders and would weaken the principle the reform is meant to embody. The historical record across capital gains tax, top-rate income tax, mansion taxes, and the abolition of the non-dom regime is that the predicted operational catastrophes when a tax newly applies to a wealthy cohort tend to be larger in advance-prediction than in actual outcome. The honest response is to adopt the four practical measures, resource HMRC, and let the actual behavioural data accumulate — rather than capitulating to predictions about behaviour that has not yet been measured.

The case against Position A. The operational mismatch between the death-event tax and the cohort it now reaches is real and is documented in the practitioner literature. The valuation problem at SAV produces multi-year disputes for a meaningful share of cases; the liquidity problem at death produces forced disposals that disrupt the underlying companies; the pre-death relocation pressure is documented in the practitioner literature even where the empirical magnitude is not yet measured for the BPR-affected cohort specifically. Position A asks the affected cohort to absorb a mechanism whose unfitness for unlisted trading-company shares specifically has been argued by tax practitioners with no personal stake in the outcome (the CIOT, the Family Business Research Foundation, several major firms in their published commentary). The "wait and accumulate data" posture has the form of caution but the substance of inertia: by the time enough data has accumulated to settle the empirical questions, the cohort whose response would be measured will have made its mobility decisions and the question will be effectively settled by exit rather than by analysis. Position A's defence relies on a historical generalisation about wealthy cohorts overstating their mobility; the BPR-affected cohort's specific mobility characteristics may differ from the populations that historical generalisation was drawn from.

Position B — Switch the mechanism to CGT on realisation for the affected asset class. Replace IHT-at-death for unlisted trading-company shares with Capital Gains Tax charged on actual proceeds when the heir realises the asset, with no base-cost uplift on death. Combine with a long-stop deemed-disposal rule (10 to 15 years) and a collateral-trigger rule. Apply the four practical measures alongside.

The case for Position B. The death-event mechanism produces problems no first-principles design would generate: a growing dispute caseload at SAV that does not scale to industrial-strategy growth assumptions; behavioural pressure on the affected cohort that leaks revenue through pre-death relocation; an avoidance vector through collateralised borrowing that the existing regime does not close. The proposed mechanism collects the same principle — taxation of generational wealth transfer — against optimised exit values rather than contested death-date values, with lower administrative cost. Australia has operated a comparable rollover-to-realisation regime for forty years; Canada operates a related design. The valuation problem is removed entirely (the tax is calculated against actual sale price). The liquidity problem is removed entirely (the tax is paid out of actual sale proceeds). The pre-emptive relocation pressure is reduced (the heir's eventual sale can occur wherever the heir lives; the founder's location at death is less consequential).

The case against Position B. The administrative-settledness argument cuts directly against B: the UK inheritance tax framework has used death as the tax event since 1894, and the practitioner profession, the SAV process, and the instalment regime are all built around that event. Switching for one asset class produces administrative inconsistency that the death-based regime does not have. The lock-in literature (Auerbach 1989; Burman 1999) establishes that realisation-based taxation produces deferral behaviour: heirs hold appreciated assets longer than they would in a no-tax world to defer the realisation event. A UK realisation-based regime for inherited unlisted business shares would acquire a version of this lock-in problem and a version of the practitioner industry that has grown up around managing it in Australia (deferral structures, family-trust workarounds, valuation-gaming at the moment of inheritance when the cost base is set). Position B also makes the state a contingent quasi-equity holder in private companies — the state's eventual receipt is contingent on the heir's eventual sale decision, which the heir controls — which is a structural feature the death-event regime does not have.

Position C — Hybrid: practical fixes only, mechanism question deferred. Adopt the four practical measures. Defer any mechanism change pending evidence on the size of the operational issues over the following two to three years. With or without trigger conditions for revisiting the mechanism question.

The case for Position C. The four practical measures are independently valuable and unobjectionable — they would be adopted under any of the design positions. Mechanism change carries political risk and design risk that the practical measures do not. If the practical measures resolve enough of the operational issues, the case for mechanism change weakens; the deferral posture commits Treasury to data-driven review rather than to a specific outcome that prejudges the empirical questions on which A and B disagree. The empirical questions (the magnitude of the SAV dispute caseload at scale; the actual rate of pre-death relocation; the size of the collateralised-borrowing avoidance vector) can be measured in two to three years of operation under the existing mechanism with the practical measures applied. Mechanism change before that data exists risks adopting an alternative whose problems are not yet visible because the alternative has not been operated.

The case against Position C. "Defer pending evidence" has the form of caution but in practice produces indefinite drift if the trigger conditions for revisit are not specified or not specified credibly. The political economy of UK tax policy is that mechanism changes happen at moments of fiscal event (Budget, autumn statement) and the windows for them are narrow; deferring past one window typically means deferring past several. The cohort whose mobility is the consequential variable will make its decisions on the regime as it stands, not on the deferred review; by the time the data accumulates, the cohort response will have happened. Position C's defenders argue this risk is overstated; critics argue it is the central feature of how mechanism changes do or do not happen in UK tax policy. The triggers-or-not variant matters: Position C with specified, public, credible triggers (defined fiscal-cost thresholds, defined caseload thresholds, defined behavioural-response thresholds with measurement methods specified in advance) is meaningfully different from Position C without — the first commits to a review process; the second commits to nothing in particular.

Position D — Targeted higher threshold for qualifying unlisted trading-company shares. Keep the existing IHT-at-death mechanism. Keep the £2.5m / £5m allowance for the general BPR base. For a defined sub-class of qualifying unlisted trading-company shares, raise the 100% relief threshold to a higher figure (proposals from advisers and trade bodies range from £5m to £10m per individual). Combine with the four practical measures.

The case for Position D. The design preserves the principle of the reform across most of the base while moving the cohort where the operational mismatch is sharpest above the threshold at which the mismatch bites. It is structurally a scope adjustment rather than a mechanism change: it does not give the state quasi-equity exposure across hundreds of private companies (which is Position B's structural feature), it does not defer the question (Position C's posture), and it does not ask the affected cohort to absorb a mechanism the operational case says is unfit (Position A's posture). Its closest international analogue is the way the US estate tax exempts qualifying small-business interests at a higher level than other assets through Section 6166 mechanics combined with the unified credit. The CenTax minimum-share-rule and upper-limit alternative-design proposals (Advani, Gazmuri-Barker, Mahajan, Summers 2025) sit within the same family of scope-adjustment designs and have been modelled against HMRC inheritance tax data 2018-2022; the publication's Position D should arguably be broadened to incorporate the CenTax alternatives, which are a stronger empirical basis for design than the publication's own work.

The case against Position D. Special-pleading carve-outs for sympathetic asset classes are historically how tax bases erode. The original BPR was itself a carve-out justified on similar grounds (family businesses are special, illiquid, important to the economy) and the reform exists because that carve-out grew unbounded over decades. A targeted higher threshold for tech founders risks repeating the cycle: defensible on its merits at introduction, expansionary in practice, eventually requiring its own reform. The definitional problem has no clean solution. Defining "tech" by SIC code is gameable in days. Defining it by activity (software, biotech, deep tech, AI) creates HMRC case-by-case classification disputes. Defining it by exclusion (the EIS/SEIS-style approach) inherits all the boundary fights EIS already has, scaled up because the stakes per estate are much larger. Defining it by R&D intensity or growth metrics is cleanest analytically and worst politically (a profitable manufacturer that built the country's industrial base does not qualify; a loss-making AI company does). The fairness objection is direct: if the cap is raised for tech founders but not for the family that owns a 200-year-old engineering firm employing 400 people, a justification has to be available. "Because tech founders are mobile and the engineering family is not, so the behavioural response is larger" is a prudential answer, not a fairness answer; whether prudential considerations should override fairness-across-asset-classes is a normative question on which reasonable people disagree.

An interest disclosure that applies specifically to Position D. The author of this publication is a UK technology founder. Position D is the position whose policy outcome aligns most directly with the commercial interest of the cohort the author is part of (a higher BPR threshold for unlisted trading-company shares would benefit founders in that cohort). The publication notes this fact openly. The case for Position D is set out above in the voice of its strongest defender, and the case against in the voice of its strongest critic, in the same length and register as Positions A, B, C, E, and F. The reader should weight the alignment between the author's standing and Position D's policy outcome when assessing the case-for paragraph. Positions A, B, C, D, E, and F are presented at the same operational weight in this article; whether a reader finds the equal-weight presentation defensible after factoring in the author's standing is a judgement the reader makes.

Position F — A founder election with a decade cap. On the death of a qualifying shareholder, the personal representatives elect, per company, between Regime 1 (settlement at death under the existing reform — the £2.5m / £5m allowance, 50% relief above, ten-year interest-free instalment regime) and Regime 2 (the inheritance tax charge deferred until the qualifying shares are sold, subject to a hard ten-year backstop, the rate locked at the rate that would have applied at death, the taxable base floored at the date-of-death valuation). Anti-avoidance perimeter: connected-party deemed disposals at independent valuation; exit charge if the heir ceases UK tax residence during the window; asset-stripping triggers; continuing-qualification conditions; the CGT base-cost uplift itself deferred under Regime 2 to keep the principle clean. The full long-form treatment is at A Founder Election with a Decade Cap; the five-minute version is here.

The case for Position F. The case-for-realisation in the timing piece notes that the SAV process produces multi-year disputes for a meaningful share of cases and produces values that differ substantially from any later realised value; that the death-event mechanism produces forced disposals that disrupt the underlying companies; that pre-death migration pressure under death-event taxation is documented in the practitioner literature. The case-against-B in that same piece turns on a specific empirical observation: Australia's forty-year experience of CGT-on-realisation for inherited assets has produced indefinite generational deferral, valuation-gaming at the cost-base-setting moment, and a family-trust deferral industry. Those costs are real and they are problems of an unbounded realisation regime, not problems intrinsic to realisation-based timing as such. Position F is the bounded version: deferral capped at ten years through a deemed-disposal backstop, the Treasury's downside protected by the date-of-death floor (the heir who realises a loss against the date-of-death value pays the rate against the date-of-death value, not the lower realised value), the gaming routes closed at the perimeter through anti-avoidance rules that mostly already exist in analogous form across UK tax legislation. The valuation problem largely dissolves — the SAV value at death sets the rate base and the floor only, and the substantive tax falls against the actual sale price in the typical case. The liquidity mismatch dissolves — the tax falls when the cash arrives. Pre-death relocation pressure is reduced — a founder who knows their estate will be able to elect Regime 2 has a weaker reason to relocate before death, and the exit-charge condition prevents the heir from realising a relocation benefit post-death. Productive-economy alignment is preserved without the unbounded-deferral cost the case-against-B identifies. And underneath the technical case is an agency case: the current reform creates a problem founder families cannot plan around because the timing of death is the one variable the founder cannot control, and Position F preserves the Treasury's long-run revenue while restoring to families the ability to choose when within the decade the tax event falls. The leaving decision for founders considering exit is, in many cases, less about the tax itself than about exposure of their family to a forced event at an unknowable moment; Position F does not eliminate the tax but does eliminate the ambush. The full agency argument is set out in the dedicated long-form Position F piece.

The case against Position F. Treasury cashflow becomes lumpier within each decade: death-event taxation with ten-year instalments gives HMRC a known forward profile of receipts; Position F means the Exchequer does not know which year of the next decade the receipt arrives in. For each individual estate this is small; aggregated across the affected population (HMRC's December 2025 estimate is approximately 1,100 estates per year), it is a meaningful change to the cashflow shape of a small but visible tax line. A reader who weights fiscal-stability considerations heavily will treat the timing change as a real cost regardless of the long-run base preservation. The anti-avoidance perimeter has costs even where the legislative concepts are existing: applying connected-party rules, exit charges, deemed disposals, continuing-qualification conditions, and asset-stripping triggers to this combination produces interactions that are not present in any existing regime, and a practitioner industry will grow up around optimising the election and the window. The horizontal-equity question Position D faces applies to F: a founder dying with £20m of qualifying unlisted trading-company shares can elect a deferred-realisation regime; a founder dying with £20m of cash, listed equities, or property pays IHT at death. The differential treatment is justified within the publication's frame by the differential characteristic (illiquidity, productive activity); a reader who weights the IFS / Resolution Foundation horizontal-equity argument heavily will treat the differential as a problem rather than a feature. Administrative inconsistency: F introduces a second mechanism for one asset class, parallel to but separate from the death-event mechanism that handles every other asset; estates with mixed assets will have one tax event on one calendar and a second on a different calendar; practitioners will need to maintain expertise in both regimes; the deferred-CGT-uplift treatment adds a further layer. The administrative-settledness argument the case-for-A draws on is real, and Position F weakens it.

1.5. A note on timing

The behavioural response to the announced reforms began before the reforms took effect. The available evidence on this is partial and indirect, and the article should be read with the limits of that evidence understood, not glossed over. The data does not directly establish what the article would need it to establish — that the BPR-affected cohort specifically is responding at scale to the BPR reform specifically — but it is suggestive of a broader behavioural environment in which mobile wealthy populations are responding to the cumulative weight of multiple tax changes, of which BPR is one.

The OBR's behavioural assumption — for a different reform

The Office for Budget Responsibility's January 2025 supplementary forecast on the non-dom reform sets out the central behavioural assumption underpinning that reform's costing: the OBR expects approximately 25 per cent of non-doms with excluded property trusts to leave the UK as a result of the reform, alongside 12 per cent of other non-doms. These are official UK government assumptions in the Treasury's own scoring document.

An earlier draft of this article cited these figures as if they were directly transferable to the BPR cohort. They are not. The non-dom reform governs a different population (non-domiciled UK residents), uses different mechanisms (residence-based IHT for foreign assets, abolition of the remittance basis), and applies to a population whose mobility characteristics are structurally different — non-doms with excluded property trusts hold offshore assets in offshore structures and are highly mobile by virtue of the regime they were already operating under. BPR-affected founders hold UK trading-company equity, with UK customers, UK staff, and UK operational presence; their assets are anchored in the UK in ways non-dom assets are not. There is no a priori reason to assume the behavioural elasticity would be the same in both directions: it could be lower (UK anchoring matters) or higher (mobile founders may already have international structures available). The OBR figure is not a direct estimate for the BPR cohort. It is a useful reference point for the order of magnitude of departure rates HMRC considers plausible for highly mobile wealthy populations under tax reform — nothing more.

The Companies House data — contaminated by simultaneous reforms

Financial Times analysis of Companies House records recorded 3,790 UK company directors moving their primary residence abroad between October 2024 and July 2025 — a 40 per cent increase on the 2,712 in the same period a year earlier. April 2025 alone saw 691 director departures, 79 per cent above April 2024. Approximately 150 directors moved specifically to the UAE between April and June 2025. (These figures are contaminated by the non-dom reform, residence-based IHT for non-doms, CGT changes, and carried-interest changes all hitting the same population in the same window; the next paragraph explains why none of these numbers can be read as BPR-driven departures.)

This data is the cleanest single piece of behavioural evidence in the public record, but it cannot be read as evidence of BPR-driven departures. The figure is much larger than the BPR-affected cohort (3,790 directors versus the BPR-only subset on which the operational mechanism question turns, which is in the low hundreds within HMRC's December 2025 estimate of ~1,100 estates affected by the whole APR/BPR reform), so most of these directors are not affected by the BPR reform at all. The period overlaps with the non-dom reform's effective date (April 2025), changes to capital gains tax, changes to carried-interest taxation, residence-based IHT for non-doms, and general political uncertainty. The April 2025 spike correlates more directly with the non-dom reform than with the BPR reform (which took effect a year later). The data demonstrates that wealthy UK-resident directors were relocating in significantly higher numbers in 2024-25 than in earlier years; it does not isolate the BPR-driven component, and treating it as such would be analytically dishonest.

The named cases — illustrative, not evidential

Several individual departures at the top of the UK technology cohort have been individually reported. Nik Storonsky, founder of Revolut, has registered residency in the UAE. Herman Narula, co-founder of Improbable (a company last externally valued in 2022 at approximately $3.4bn), has been reported as preparing to emigrate to the UAE; the trigger he has cited publicly is a mooted exit-tax on departures, rather than the BPR reform specifically, and he has been quoted saying he does not want to leave. A number of named non-tech billionaires including Lakshmi Mittal have also been reported as having relocated. Trade press reporting from Sifted (May 2025) cites four UK tax-advisory firms — Wilson Partners, Evelyn Partners, Founders Law, and Capital Partners — confirming a marked uptick in UK-based tech-founder enquiries about Dubai relocation, with Founders Law specifically reporting that UAE relocation features in 15–20 per cent of new business enquiries. The named cases are illustrative rather than evidential, and the publication's journalists' page has more detail on what each does and does not establish.

These cases are illustrative rather than evidential. Three named individual departures from a country with sixty-seven million people, several million of whom are wealthy, do not establish a rate or a trend. Adviser-survey reporting is qualitative — "an uptick in enquiries" can mean almost anything in quantitative terms. What the named cases and adviser-survey reporting establish is that tax-driven relocation by UK founders is happening; what they do not establish is the underlying rate, the proportion attributable to BPR specifically as opposed to other reforms, or whether the trend will continue. The article cites these cases as colour, not as evidence on which the policy argument turns.

The Henley figure — not credible

The widely-cited Henley & Partners projection of 16,500 UK millionaire departures in 2025 should not be cited as evidence by anyone serious about this question. Tax Policy Associates published a detailed forensic critique in July 2025 — examining methodology changes (the underlying definition of "wealth" was altered between reports without producing the data effects those changes should have produced), statistical anomalies (excess of even-numbered values, unusual digit-clustering suggestive of estimates rather than measurements), and contradictions with official UK figures of 70 to 100 per cent in some cases — and concluded that the Henley reports should be treated as marketing material rather than evidence. Tax Justice Network reached similar conclusions independently. The article does not rely on the Henley figure.

What this evidence does and does not support

What the evidence supports: that wealthy UK-resident directors and founders were relocating in significantly higher numbers in 2024-25 than the recent baseline; that pre-reform departures were occurring at scale in response to the cumulative weight of multiple tax changes; that the most mobile and most advised members of the affected cohort were not waiting for the regime to bite before responding; that the UK government's own behavioural assumptions for related reforms anticipate double-digit departure rates from highly mobile wealthy populations.

What the evidence does not support: that the BPR reform specifically is causing departures at any specific rate; that the Companies House data isolates BPR-driven behaviour; that the OBR figure can be directly transferred to the BPR cohort; that the named cases establish a population trend. An earlier draft of this article overstated what the evidence could carry. The corrected reading is that the evidence is suggestive of a broader behavioural environment that should concern policymakers, but does not directly establish the BPR-specific empirical claims that some framings of Position B rest on.

What this means for the rest of the article. The framework in Section 3 commissions HMRC modelling on the relocation channel; the four scenarios in Section 5 describe what that modelling could show. Both treat the data as forward-looking. The honest reading is that significant pre-positioning has occurred against the cumulative tax environment, of which the BPR reform is one component. The trigger thresholds in Section 3 — even taken as illustrative — are calibrated against a population that has already partly responded to a broader set of changes. Any HMRC modelling commissioned now will measure a residual cohort, not the original one. This is a real complication for any conditional framework, and the framework in Section 3 should be read with this complication understood, not as if the evidence cleanly establishes a BPR-specific effect.

1.6 How the reform applies across the UK tech funding stack

Article One's analysis treats "unlisted UK tech-trading-company shares" as a single asset category. That is appropriate at the level of the policy question but obscures meaningful variation across the funding stack. The cohort directly affected by the reform is not homogeneous; the reform interacts differently with each part of the stack, and a serious analysis has to engage with that variation rather than gesture at it. This sub-section sets out, in compressed form, what the reform means for each part of the stack. The dedicated funding-stack piece (separately published) treats each of these in greater technical depth.

1.6.1 Founders

The most directly affected cohort, with the widest range of outcomes by company stage. Pre-seed and seed-stage founders with paper equity below the £2.5m allowance are unaffected in current terms but face the question of whether their company will grow into the cap. Series B and later founders with paper equity above the cap are in scope. The death-event valuation problem is most acute for venture-backed founders in the mid-stage — Series B through pre-IPO — where paper valuations are large, marketability is constrained by shareholder agreements and preference structures, and the company is typically reinvesting all cash so dividends and instalment funding are unavailable until exit. The instalment option (interest-free over ten years, extended in the reform) is meaningful for founders whose companies generate operating profit, less meaningful for venture-backed pre-profit companies, and least meaningful for very-late-stage growth companies whose value sits substantially above the cap.

The structural responses available to founders — gifts to children with seven-year survival, EOT (employee ownership trust) transitions where appropriate, family investment companies, lifetime trusts, residence planning — vary materially in their applicability across the stage spectrum. A founder of a profitable mid-market software company has a different toolkit from a Series C founder with paper equity but no liquidity. The article's analysis treats founder behavioural response as a single elasticity; the dedicated funding-stack piece sets out the stage-specific structure that this analysis necessarily smooths over.

1.6.2 Angels and EIS investors

UK angel investors face a meaningfully different question from founders. EIS shares qualify for BPR after a two-year holding period and so are potentially relievable on death up to the new £2.5m combined APR/BPR allowance. An active angel investing £150,000 to £400,000 a year across 8 to 15 EIS positions builds a portfolio that, after a decade, is materially in scope above the cap. The reform interacts with the existing EIS rules (30% income tax relief, CGT exemption on disposal, loss relief on failures, BPR after two years) to change the back-end shelter without affecting the front-end incentives.

Two structural points the practitioner literature has been working through: first, the £2.5m cap applies to combined APR and BPR property at estate level, so an active EIS-investor angel who also holds founder shares from earlier ventures is more constrained than the headline cap suggests. Second, the seven-year refresh on the £2.5m allowance interacts with lifetime gifting strategies in ways that change the calculus for active angels considering when to pass shares to next-generation family members. The most active angels in the UK tech ecosystem — those funding seed-stage at scale across multiple sectors — are also the cohort most likely to have used family-office structures, holding companies, and trust arrangements that further complicate the picture. The first-order effect on individual angel behaviour is moderate. The second-order effect — through the founders the angels would otherwise back, the seed-stage capital that goes elsewhere, the operator-investor expertise that distinguishes UK angel money from purely financial capital — is the larger question.

1.6.3 Venture capital and growth equity (LP interests)

LP interests in UK venture and growth-equity funds qualify for BPR if the underlying fund is investing in qualifying trading companies, subject to the two-year holding period applied at the LP-interest level. For UK-resident high-net-worth individuals investing as LPs, this is a meaningful shelter that the reform now caps at £2.5m. A UK-resident individual with £10m committed across three or four UK venture funds is materially in scope above the cap.

The mechanism issue is sharper here than for any other part of the stack. An LP interest in a closed-ended venture fund is illiquid by structural design — there is no realisation event until the fund itself realises the underlying portfolio companies, which can be eight to twelve years from initial commitment. Death during the fund's life produces a contested valuation against an asset class with no public market price, no liquidity option, no buy-back mechanism, and no observable yield. The valuation that HMRC's Shares and Assets Valuation team will ultimately agree depends on a set of judgments about discount for lack of marketability (typically 25-40% for closed-ended fund interests with no secondary market), the fund's NAV reporting methodology, the underlying portfolio company valuations within the fund, and the timing of expected realisations. The negotiation is technically complex and slow; the evidence base on which estates and HMRC negotiate is thinner than for founder direct holdings, where at least a recent funding round provides an anchor.

The behavioural risk for UK venture funds is that high-net-worth UK-resident LPs reduce commitments to UK funds, redirect commitments to non-UK funds (where their LP interest is not exposed to UK IHT in the same way), or themselves relocate. UK venture funds have historically depended significantly on UK-resident LP capital, alongside US, European, and sovereign capital. The marginal commitment that goes elsewhere is the commitment most affected by the reform, and the cumulative effect on UK venture-fund capital formation over time is hard to estimate without HMRC microdata but directionally clear.

1.6.4 Private equity in UK tech

Private equity invested in growth-stage and mature UK tech companies — buyout-style PE, growth equity at later stages, and the increasingly important large-scale growth-equity funds — interacts with the reform mainly through the personal holdings of partners and the management equity in portfolio companies. The fund-level economics are not directly affected by the IHT reform; the personal IHT exposure of UK-resident partners is. For the largest UK-resident PE partners, personal carry interests above the £2.5m allowance are in scope. The cohort is small in number but well-advised, and structural responses (offshore trust arrangements where applicable, family investment companies, residence planning) are typically already in place. The reform's first-order effect is to accelerate planning that this cohort was largely already doing in response to the 2024-25 carried-interest reforms.

The effect on PE-backed UK tech portfolio companies is indirect but worth naming. Where founders of PE-backed UK tech companies retain material personal stakes — increasingly common in growth-stage PE structures that prefer founder retention over founder buy-out — those stakes are affected on the same terms as direct founder equity. The PE fund's interest in retaining UK founders for the duration of the hold period (typically four to seven years for growth-stage PE) interacts with the founder's interest in not facing a punitive death-event tax during the hold. This is a private negotiation between PE sponsors and founders, and the reform changes its dynamics at the margins — particularly for founders considering whether to relocate during the PE hold period.

1.6.5 Early employees with vested equity

An often-overlooked cohort. UK tech employees who joined growth-stage companies early and now hold material vested equity above the £2.5m allowance — typically through EMI options exercised during the company's growth, or through founder-allocated share grants — are in scope on the same terms as founders. This cohort spans several thousand people across the UK tech ecosystem: early engineers and operators at companies that have grown to unicorn or near-unicorn valuations, as well as senior hires brought in during scale-up phases with material equity grants.

The death-event valuation problem is the same for them as for founders, but two features make the early-employee position structurally different. First, early employees typically hold smaller stakes (1-5% rather than 15-40%) so are less likely to face full minority-discount adjustments and more likely to face simple application of recent funding-round valuations. Second, their structural responses are typically less sophisticated than founders' — fewer have established family investment companies or trust arrangements, more rely on basic spousal transfer and the £2.5m allowance itself. The practical effect of the reform for this cohort is that early-employee equity has become a more complex tax-planning question than it was, and the difficulty for UK tech companies in recruiting and retaining senior operators (a recurring concern in UK tech industrial strategy) is now compounded by an additional planning consideration that compensation conversations have to address. The dedicated funding-stack piece sets out the technical mechanics for early employees in more detail.

2. What the disagreement actually turns on

The design positions cannot all be right. They rest on differing answers to a small number of questions. If those questions could be answered with rigour, the disagreement would resolve. The questions are partly empirical and partly normative, and they have not been answered with rigour.

Question 1 · How large is the relocation channel under the existing mechanism?

If the proportion of the affected cohort that relocates pre-death is small, Position A's revenue forecast is robust and the mechanism issue is overstated. If the proportion is large, the existing mechanism leaks materially and the case for change strengthens. The question is empirical and answerable through dynamic behavioural modelling using HMRC data. It has not been answered with rigour. Adviser surveys exist but are self-selecting; trade-press reporting exists but is anecdotal.

Question 2 · How quickly does SAV dispute caseload grow as the cohort grows?

If the caseload scales linearly with affected estates and SAV can be resourced to keep pace, the operational problem is solvable through capacity expansion. If the caseload scales superlinearly because each dispute is a multi-year file consuming disproportionate capacity, resourcing alone cannot keep pace. The question depends partly on the rate at which the cohort grows — which Government industrial strategy is actively trying to accelerate — and partly on whether disputes resolve faster or slower as the safe-harbour methodology beds in.

Question 3 · How effective is the collateral-trigger rule against sophisticated structuring?

If the rule cleanly captures economic realisation through borrowing, the buy-borrow-die vector is closed under any of the design positions. If sophisticated advisers can route around the rule through structures the legislation does not capture, the rule's effectiveness is overstated. The question is technical and depends on drafting quality and ongoing anti-avoidance work.

Question 4 · Can a 10- or 15-year long-stop deemed disposal be administered without recreating the original problem?

Position B's long-stop assumes a deemed disposal at year 10 or 15 produces an administrable computation against an observable value. If the asset remains illiquid at the long-stop date, the long-stop may recreate the same valuation and liquidity problem the original mechanism produced. The honest reading: the long-stop is structurally weaker than its proponents acknowledge, but closer to acceptable when paired with a payment-on-realisation provision allowing the long-stop charge itself to be deferred until actual liquidity, with a secured claim attached to the asset.

There is a further structural feature of Option B that the public debate has not engaged with and that Position A's defenders should be making more of. Under payment-on-realisation deferral, the state holds an unfunded contingent claim against private company equity for potentially decades. Functionally, this gives the state economic exposure to the company's outcomes without governance rights — upside participation through eventual CGT collection if the company succeeds, total loss if it fails. Across the affected cohort over time, the state ends up holding contingent quasi-equity exposure across hundreds and eventually thousands of UK private companies, selected by the non-random criterion of which founders or significant shareholders died holding them. This is structurally a different relationship between state and private business than tax collection. Whether it is desirable (alignment of state incentives with company success) or undesirable (a meaningful expansion of state economic interest in private business without the deliberation, statutory framework, or governance structure that would normally accompany state equity holdings) is a question the public debate has not asked, because the framing has been "tax administration choice" rather than "implicit state acquisition of contingent equity exposure." A serious decision on Option B requires this question to be asked, and the question is not contained within the tax-policy frame in which Option B has so far been advanced.

Question 5 · What is the regime for?

This is the question that cannot be settled by modelling. If the primary purpose of the reform is revenue, the position that maximises receipts net of leakage and administrative cost wins; this becomes empirical once the modelling exists. If the primary purpose is fairness across asset classes — that all wealth above the allowance is taxed at the point of generational transfer regardless of liquidity — Positions A and E are uniquely consistent (E by implementing the reform as enacted; A by adding only the four practical measures without changing the mechanism); mechanism change for one asset class (Position B) and scope adjustment for one asset class (Position D) are both by definition unequal treatment. If the primary purpose is to support industrial strategy by retaining the cohort that builds growth-stage UK companies, Positions B and D are both consistent — Position B by changing the mechanism so the timing matches asset realisation, Position D by raising the threshold so most of the cohort is above the bind. Position D delivers the industrial-strategy objective with less administrative complexity than B but with a sharper fairness objection than B (because it is explicitly a higher allowance for one cohort, where B is a different mechanism for one asset class — both are differential treatment, but the scope-adjustment version is harder to defend on equity grounds). Position E delivers fairness-across-asset-classes most cleanly because it requires no special treatment for any cohort. The question is normative, the answer depends on which objective is treated as primary when they conflict, and ministers — not analysts — should answer it. But the answer should be stated, because every position depends on a different one.

Reasonable, informed, good-faith readers reach different conclusions on these five questions. The disagreement between the design positions is largely a disagreement about how the questions should be answered, weighted, and prioritised. A reader who reaches a clear view on the questions will, by implication, reach a clear view on the policy.

3. What each of the design positions would look like in operational form

This section describes what each of the design positions would look like as a deliverable policy package. None of them is presented as the publication's recommendation. The point of setting them out at equal length is that the reader can see what each would actually involve — what legislation it requires, what HMRC has to do, what changes for the affected cohort — and form their own view about which package answers the empirical and normative questions of Sections 1 and 2 most defensibly.

One observation applies to all four. The four practical measures — collateral-trigger rule, statutory SAV safe-harbour methodology, targeted tightening of temporary non-residence rules, and structured buy-back guidance — are common ground. Position A endorses them because they address the genuine operational issues without changing the underlying mechanism. Position B endorses them because they are sensible regardless of which mechanism eventually applies. Position C endorses them as the foundation on which any subsequent decision would build. Position D endorses them as one component alongside scope adjustment. Treating the four practical measures as conditional on resolving the A-versus-B-versus-D question delays improvements that none of the design positions disputes. They should be adopted on their own merits, separately from the harder design question.

The four practical measures, in detail

• Collateral-trigger rule. Use of inherited unlisted trading-company shares as collateral for borrowing above 25 per cent of the asset's assessed value triggers a deemed disposal proportionate to the encumbered share, crystallising the IHT instalment balance under the existing regime (and CGT under the realisation regime if Position B applies).
• Statutory SAV safe-harbour methodology. HMRC publishes a statutory valuation methodology covering preference-stack treatment, minority discount ranges, DLOM ranges by company stage, and aggregation under IHTA s.161. Estates that adopt the safe-harbour formula receive fast-tracked SAV approval.
• Targeted temporary-non-residence tightening. Extends the qualifying TNR period from five to ten years for individuals whose UK-acquired assets exceed a defined threshold, integrated with registrar-level enforcement on UK-incorporated companies.
• Structured buy-back guidance. HMRC and HMT publish guidance on the use of company purchase of own shares under CTA 2010 s.1033 to fund IHT instalments out of operating cash flow. Works for cash-generative private companies; does not work for early-stage venture-backed holdings or LP interests in closed-ended funds.

Position A in operational form — Hold and resource

Adopt the four practical measures within ninety days. Resource HMRC's Shares and Assets Valuation team for the projected dispute caseload — the publication has not modelled the staffing requirement, but the implied caseload growth requires meaningful capacity expansion. Run the regime as enacted. Commit to a routine post-implementation review at two to three years — the standard cycle for tax measures of this scale — without pre-committed triggers. If the review evidence supports the existing mechanism, the regime continues. If it does not, the question is reopened on its merits at that point. Position A's defenders argue that this is the right posture under uncertainty: pre-committed triggers privilege one possible outcome over others before the data supports it, and observed behavioural response should drive any subsequent change rather than thresholds set ex ante. The published modelling Question 1 of Section 2 calls for is the primary evidence input under this design.

Position B in operational form — Switch the mechanism for the affected asset class

Adopt the four practical measures. Legislate a CGT-on-realisation regime for unlisted trading-company shares above the £2.5m / £5m allowance, with a long-stop deemed disposal at year 10 to 15 and a payment-on-realisation deferral provision attached to the long-stop charge. Position B's defenders argue that the mechanism switch should happen on the available evidence, without waiting for behavioural confirmation that may arrive too late to retain the cohort that builds growth-stage UK companies. The legislation requires Treasury legal drafting, an HMRC operating model for tracking realisation events over potentially decade-long horizons against deceased estates, and parliamentary time. The structural feature noted in Section 2 — that under payment-on-realisation deferral the state holds contingent quasi-equity exposure across hundreds of UK private companies over time — is a genuine consequence of this design and one its proponents should engage with explicitly rather than treat as a side-effect.

Position C in operational form — Practical fixes only, mechanism question deferred

Adopt the four practical measures within ninety days. Commission HMRC dynamic modelling on the open empirical questions (relocation elasticity, SAV caseload growth, projected realisation distributions). Defer the mechanism question pending the modelling output and observed behavioural data over two to three years. Two variants exist: Position C without triggers — review the question afresh when evidence accumulates, with no pre-commitment — is closest to the current government posture. Position C with triggers — pre-commit to switching to a Position B mechanism if defined evidence thresholds are breached at a published twelve-month review point — is the variant that some commentators have proposed. The trigger variant requires Treasury legal to draft Position B legislation in parallel so that if the triggers fire the legislation is ready for the next Finance Bill. Position C's strongest defenders argue that the without-triggers variant is the honest version: pre-committed thresholds privilege a particular subsequent direction before the data supports it. Position C's other defenders argue that without triggers the deferral is criticised by some commentators as risking indefinite drift, defended by others as appropriate caution given the absence of UK-cohort-specific evidence, and that the conditional commitment is what makes "deferred" different from "abandoned." Both variants are workable; they imply different things about how seriously the trigger thresholds are taken.

If the trigger variant of Position C is being designed, illustrative thresholds would include: documented departures among the affected cohort exceeding a defined number in the review window; SAV dispute caseload exceeding a defined active-files figure with average resolution time above 24 months; audited BPR-only estate IHT receipts more than 20 per cent below Budget forecast in the first full operational year; settled SAV cases showing average final-assessed value at less than 65 per cent of opening assessment. Any specific numbers would need to be calibrated by HMRC working from internal data — the publication has no access to HMRC microdata, AEOI exchange information, or internal behavioural modelling, and robust calibration requires those inputs. The thresholds presented here describe the kind of trigger framework the design would need, not figures a government could adopt as drafted. The December 2025 HMRC revision of the affected-estate count from approximately 220 to approximately 1,100 means even the rough percentage-of-cohort logic earlier proposals embodied needs recalibration; "60 departures" calibrated against a 220-estate cohort would correspond to nearly 300 departures against the wider 1,100-estate cohort if the percentage logic were preserved.

Position D in operational form — Targeted threshold for qualifying unlisted trading-company shares

Adopt the four practical measures. Legislate a higher BPR threshold for a defined sub-class of qualifying unlisted trading-company shares — proposals from advisers and trade bodies range from £5m to £10m per individual. The general £2.5m / £5m threshold remains for farms, mature family businesses, and the bulk of the affected cohort. The legislation requires a statutory definition of qualifying shares (which is the design's hardest problem, as Section 1's Position D entry sets out), a transitional period for existing estates, and HMRC capacity to administer the differential treatment. Position D's defenders argue that this is one of several designs proposed in response to the cohort-heterogeneity observation; advocates argue it is the most directly responsive while critics argue it creates definitional problems around what counts as a qualifying unlisted trading company the article identifies: it moves the cohort where the mismatch is sharpest above the threshold at which the mismatch bites, without changing the mechanism for the rest of the base. Position D's critics — particularly defenders of Position A's strongest form — argue that this is the special-pleading carve-out for a sympathetic asset class that has historically been how tax bases erode, that the definitional problem has no clean statutory solution, and that the fairness objection (raising the cap for tech founders but not for the family that owns a 200-year-old engineering firm) is harder to defend than the supporters of the design typically acknowledge.

What is in dispute and what is not

The four practical measures are not the contested choice. They are decisions that should be made now, on grounds none of the design positions disputes, separately from the harder question of which mechanism or scope applies above the threshold. The contested choice is the design of what sits above the four measures: hold-and-resource (A), switch-the-mechanism (B), defer-with-or-without-triggers (C), or scope-adjust (D). Each is internally coherent, each rests on different assumptions about behavioural response and policy stability, and each has consequences set out in the next section. The publication does not adjudicate between them.

4. International comparators

The UK's chosen combination — a £2.5m / £5m allowance, 50 per cent relief above, death-event valuation, no business-specific conditional relief above the threshold, no realisation route — is distinctive in international terms. This section sets out what comparable jurisdictions actually do. It is not an argument that the UK should copy any of them. It is a check on the implicit claim, present in some of the public debate, that the UK's approach is the natural or default treatment.

Inclusion principle. The comparator set covers G7 economies plus Australia, plus jurisdictions frequently cited in UK debate as relocation destinations (Switzerland, UAE), plus South Korea (frequently cited in tax-policy practice as a stricter regime). It is intended to span the spectrum of approaches — no inheritance tax, deemed-disposition, high-exemption, conditional-relief, strict death-event — rather than to support a particular conclusion. Italy, omitted only for length, would behave as another softer-treatment comparator (4 per cent direct-descendant rate, high allowances).

Australia · No inheritance tax; CGT rollover at death

Australia does not levy an inheritance tax. When an Australian resident dies, capital gains tax is not triggered at the point of death. Instead, the heir inherits the deceased's cost base, and CGT crystallises only when the heir later disposes of the asset. Pre-CGT assets (acquired before 20 September 1985) get a step-up to market value at death; post-CGT assets carry the deceased's original basis through to the heir.

This is the closest international precedent for what Position B proposes. The Australian regime has been operating for forty years without producing the systemic failures Position A's defenders fear from a UK realisation-based mechanism. What the comparison does not show: Australia raises substantially less from intergenerational wealth transfer than the UK does. Whether this is a feature or a flaw depends on what the regime is for — Question 5.

Canada · No inheritance tax, but deemed disposition at death

Canada has no inheritance tax in the formal sense, but the Canada Revenue Agency treats the deceased as having sold all capital property at fair market value immediately before death — the "deemed disposition" rule. CGT crystallises on the deceased's final tax return. The heir then inherits the asset at the stepped-up cost base.

This is structurally the opposite of Position B. Where the Australian and Position B models defer the tax event until the heir realises, the Canadian model brings the tax forward to death itself. Operationally, Canada's deemed-disposition system has run for several decades. Spousal rollover provisions defer the tax until the second death. The capital gains inclusion rate is 50 per cent (the proposed increase to 66.67 per cent for gains above C$250,000 was cancelled in March 2025); combined with provincial top marginal income tax rates, the effective tax on large unrealised gains at death can reach roughly 26-27 per cent. Canadian estate planning relies heavily on life insurance to fund the deemed-disposition tax, on holding-company structures, and on principal residence designations.

United States · High exemption with step-up in basis

The US position changed materially under the One Big Beautiful Bill Act of July 2025. Effective 1 January 2026, the federal estate tax exemption is $15 million per person and $30 million per couple, made permanent and indexed to inflation. The top federal estate tax rate above the exemption remains 40 per cent. The step-up in basis on death survives unchanged: assets passing at death receive a basis adjustment to fair market value, eliminating any embedded unrealised capital gain.

In practical terms, US federal estate tax exposure now begins at a level (approximately £11–12m equivalent per person at current exchange rates) far above the UK's new £2.5m allowance. For most family-business owners below this threshold, the federal tax position is straightforward: no estate tax, full step-up in basis, heirs can sell with little or no tax. Some US states (New York, Massachusetts, Connecticut, Oregon) impose state-level estate taxes at substantially lower thresholds.

Germany · Inheritance tax with conditional business relief

Germany operates a recipient-taxed inheritance tax (Erbschaftsteuer) with personal allowances per beneficiary: €500,000 for a spouse, €400,000 for a child, €200,000 for a grandchild, €20,000 for an unrelated heir. Tax rates range from 7 to 50 per cent depending on the recipient's tax class.

The mechanism that matters for the UK comparison is the conditional business-asset relief under §§13a and 13b ErbStG. Standard relief exempts 85 per cent of qualifying business value, conditional on a 5-year hold and a cumulative payroll over the period of at least 400 per cent of the original annual payroll. Optional full relief exempts 100 per cent, conditional on a 7-year hold, cumulative payroll of 700 per cent, and a maximum 20 per cent share of "administrative assets" (passive holdings, excess cash, non-operating real estate). Breach of any condition triggers proportionate clawback. Germany is the clearest example of a regime that combines a death-event inheritance tax with structural carve-outs for genuinely-operating family businesses.

France · Pacte Dutreil — 75 per cent exemption with retention conditions

France's Pacte Dutreil regime, set out in Articles 787 B and 787 C of the General Tax Code, provides a 75 per cent exemption from transfer taxes (gift and inheritance) on qualifying business shares. Conditions include a collective undertaking to retain the shares for at least 2 years before transfer, an individual undertaking by each heir to retain the transferred shares for a further 6 years (extended from 4 by the 2026 Finance Act), and a management role exercised by at least one heir for 3 years. Combined with allowances of €100,000 per parent per child every 15 years, effective transfer tax on a French family-business succession can fall to around 11–12 per cent.

Japan · Inheritance tax up to 55 per cent

Japan operates an inheritance tax with progressive rates up to 55 per cent at the top marginal rate, with comparatively low basic allowances (¥30 million plus ¥6 million per statutory heir). Japan provides a Business Succession Taxation system (jigyō shōkei zeisei) for unlisted small and medium-sized enterprise shares, but the relief is conditional on extended holding periods and continued operation by family heirs, with strict clawback provisions for breach. The regime is materially stricter than the UK's reformed BPR position above the £2.5m allowance, both in headline rate and in the conditional structure of the available reliefs. Japan is the comparator that most clearly demonstrates that the UK's reformed regime is not at the strict end of the international spectrum.

South Korea · Inheritance tax up to 50 per cent with controlling-shareholder premium

South Korea operates an inheritance tax with progressive rates up to 50 per cent at the top marginal rate, with a controlling-shareholder premium (typically 20 per cent) that pushes effective rates on family business succession higher still — in some cases to over 60 per cent of value transferred. Conditional reliefs for SME succession exist but with strict eligibility criteria. Korea is frequently cited in advisory practice as the comparator that most clearly establishes that family-business succession can be taxed materially harder than the UK proposes to tax it. A UK family-business owner facing the new BPR cap pays significantly less than the equivalent Korean owner of comparable assets.

Switzerland and the UAE · No inheritance tax

Switzerland has no federal inheritance tax; cantonal taxes vary, but most cantons exempt direct descendants entirely. The UAE has no inheritance tax. Both are frequently cited as destinations for UK relocators in the affected cohort. These are not policy comparators in the same sense as the seven jurisdictions above. They are relevant to Question 1 (the relocation channel) rather than to the mechanism question.

What this comparison actually shows

The international landscape is wider than UK debate often acknowledges. At the soft-treatment end: Australia, the US, Switzerland, and the UAE. At the conditional-relief end: Germany and France. At the strict-treatment end: Canada, Japan, and South Korea. The UK after April 2026 sits in the middle of this spectrum: stricter than the soft-treatment jurisdictions, less generous than the conditional-relief jurisdictions, but materially less harsh than Japan, South Korea, or Canada in pure rate terms.

What is genuinely distinctive about the UK regime is the combination of a relatively low allowance with no business-specific conditional relief above it. Several jurisdictions are stricter on rate; few combine the UK's exact allowance level with the absence of structural carve-outs for genuinely-operating businesses. This is the narrower form of the comparator finding: not that the UK is an outlier in the strictness of its overall position, but that it has chosen a structural design — flat allowance, no operational carve-out — that several peer jurisdictions facing similar political pressures have rejected. International evidence neither vindicates the UK regime nor condemns it. Position A's claim that the reform is straightforwardly correct in design has to do its own work; the international evidence does not do it. Equally, Position B's claim that "this is what every other country does" is overstated — what Position B proposes is what Australia does specifically, and Australia is one model among several. Both rhetorical positions are weakened by the comparator data; neither is settled by it.

5. Where the analysis aligns and diverges from formal institutional commentary

Several of the arguments in this article — the principle case for taxing intergenerational business-wealth transfers, the operational-mismatch case against the death-event mechanism for unlisted trading-company shares, the case for alternative scope-adjustment designs, the case for deferring the mechanism question pending evidence — also appear in formal consultation responses, parliamentary committee reports, and published professional commentary by named UK institutions on the April 2026 reform. The cross-references below identify where the arguments in this article converge with named institutional positions, and where they diverge. The publication does not claim endorsement by any of these bodies.

Position A — the principle case and the existing-mechanism case. The Institute for Fiscal Studies (Adam, Miller, Sturrock 2024, Inheritance tax and farms, ifs.org.uk; IFS, Options for tax increases, November 2025) and the Resolution Foundation Budget 2024 briefing both broadly accept the principle of bringing concentrated business wealth into the inheritance tax base on horizontal-equity grounds — the case-for paragraph of Position A above is broadly aligned with their position. The IFS goes further than this article in arguing that capital gains should not be forgiven at death, an additional reform that would raise approximately £2.3 billion a year and that this article touches on but does not centrally treat. A reader who finds Position A's case-for persuasive should read the IFS and Resolution Foundation work directly; both are better-resourced than this publication.

Position B — the realisation-based mechanism case. The case-for paragraph of Position B above (the death-event mechanism producing problems no first-principles design would generate; the realisation-based alternative collecting the same principle against optimised values) is the publication's analytical synthesis rather than a position carried by named UK institutions. The international comparators (Australia's CGT-on-realisation regime for inherited assets; Canada's deemed-disposition rules) are documented in the academic and professional literature; the case for the UK adopting a comparable mechanism is the publication's argument, not an institutional position the publication is summarising.

Position C — the deferral case and the practical-fixes ground. The four practical measures common ground (collateral-trigger; statutory SAV safe-harbour; targeted TNR tightening; structured buy-back guidance) is closest to the institutional position taken by the House of Lords Economic Affairs Finance Bill Sub-Committee (January 2026, publications.parliament.uk/pa/ld5901/ldselect/ldeconaf/250/250.pdf), which recommends extending the IHT payment deadline to 12 months for estates with qualifying APR and BPR assets, recommends a statutory safe-harbour for personal representatives, and recommends the government commission qualitative research over seven years on the impact on farmers and family-business owners (Paragraph 363). The Sub-Committee's framing converges with Position C's deferral-pending-evidence posture; the Sub-Committee does not however take a position on the underlying principle or timing questions.

Position D — the threshold-raise case, and the better CenTax alternatives. Position D's case-for paragraph above is the publication's argument; CenTax's published alternatives (Advani, Gazmuri-Barker, Mahajan, Summers 2025, centax.org.uk) — a minimum-share rule under which estates with APR/BPR assets ≥60% of total receive 100% relief up to £5 million per estate, and an upper limit on relief at £10 million of claim with 100% relief on the first £2 million — are stronger empirical foundations for the scope-adjustment family of designs than the publication's threshold-raise framing. CenTax's analysis is grounded in HMRC inheritance tax data 2018-2022 and is the work a serious engagement with Position D should be reading. The publication's Position D should arguably be broadened in a future revision to include the CenTax alternatives properly. A reader interested in the case for scope adjustment should read CenTax directly.

The administrative-burden concerns the design positions accept as part of the practical-measures ground. The Chartered Institute of Taxation has submitted formal consultation responses on the draft legislation (tax.org.uk/ref1551) and on the trusts aspect of the reform (tax.org.uk/ref1481); the CIOT-ATT joint commentary in Tax Adviser (taxadvisermagazine.com) records the engagement on transferable-spousal-allowance, allocation-of-allowance, and gifting-transitional questions. ICAEW has expressed similar administrative-burden concerns through professional commentary, summarised in the Commons Library briefing CBP-10181 (commonslibrary.parliament.uk). The Family Business Research Foundation's Kemp 2025 report and the FBRF/Cebr research project (interviews running December 2025 to May 2026) produce the cohort-impact evidence the article's behavioural-response sections currently treat as not yet published.

The pattern of institutional cross-reference: IFS, Resolution Foundation, and the academic literature broadly support the principle case; the Lords Economic Affairs Sub-Committee converges with Position C's practical-measures-and-defer posture; CenTax and FBRF produce the empirical foundations that Position D should be working from; the CIOT and ICAEW provide the practitioner-administrative material the design positions need to engage with seriously. None of the design positions in this article is uniquely held by a single institution; the article's contribution is the structured presentation of the four together at equal length, with the cross-references above identifying where the institutional positions sit relative to each.

6. What different evidence would mean

The five questions in Section 2 have answers that are partly knowable and partly contestable. This section walks through what each plausible evidence outcome would imply for the policy question, and what it would mean for the affected groups. The intent is to show readers the consequences of each branch — so they can form a view about which outcome they consider most likely, and which they would consider acceptable — rather than to predict which branch the data will land on.

The reader should hold the timing point from Section 1.5 in mind throughout this section. The four scenarios below describe what HMRC's modelling could show. They will measure a population that has already partly responded — the Companies House data and the OBR's own assumptions both indicate this — and so even the most pessimistic scenario likely understates the true behavioural response by some margin. The scenarios remain useful for thinking about the policy choice, but readers should treat the "small" scenarios with particular caution: a small measured response is consistent with either a genuinely small underlying response or a large response that has already been absorbed into the pre-reform departures.

One narrowing of the choice is worth stating before the scenarios begin. The companion policy options paper, which puts indicative numbers on each option in a way this article does not, finds Option B's central-case revenue impact relative to Option A in the range of approximately −£100 million per year over the first five years on conservative behavioural assumptions, with a wider range of −£200m to +£600m depending on which behavioural elasticity is correct. Option B is therefore not obviously revenue-positive on the central case. B's stronger argument lies on cohort retention, behavioural elasticity, and reduced administrative cost — not on direct receipts. Readers and policymakers framing the choice as "which option collects more tax" are operating on a question the evidence does not clearly answer in the way they assume.

Four scenarios are described. They are not exhaustive. They are the four that the available data and the reasonable range of expert opinion suggest are most worth considering.

Scenario one · The relocation channel is small and SAV caseload manageable

If HMRC's behavioural modelling shows annual departures from the affected cohort below roughly 30 in steady state, SAV caseload growth tracking estate-count growth linearly, and receipts within Budget forecast in year one — Position A is broadly vindicated. The four practical measures, plus adequate SAV resourcing, are sufficient. Mechanism change becomes unnecessary.

What this would mean. For Treasury, the reform delivers as designed; the £5m couple allowance is robust; the principle of fairness across asset classes is preserved. For the affected cohort, the regime is more navigable than the founder-lobby framing suggested — funding routes work for cash-generative companies, the safe-harbour methodology removes most valuation uncertainty, and the cohort that planned to relocate largely doesn't. For industrial strategy, the regime is neutral-to-positive; founders building UK companies face a calculable tax position at death that does not, in practice, drive most of them to leave. The political cost of the original reform turns out to have been front-loaded — the announced response was sharper than the realised response.

The honest qualification. Even under this scenario, the venture-stage cohort (pre-revenue companies, illiquid LP interests) is not fully solved by the four practical measures. A small group continues to face genuine operational difficulty. That group is too small to drive policy change but real enough that individual cases will surface periodically.

Scenario two · The relocation channel is meaningful but bounded

If HMRC modelling shows annual departures in the 30–80 range, SAV caseload growing somewhat faster than estate count but not unmanageably, receipts modestly below forecast (5–15 per cent) — the picture is mixed. The reform is leaking some revenue and losing some of the cohort, but not catastrophically.

What this would mean. For Treasury, the question becomes harder. Holding the regime delivers most of the revenue but loses some; switching the mechanism may collect more but at the political cost of being seen to capitulate. The choice is genuinely contested on the empirics. For the affected cohort, the regime is operationally workable for most but pressure persists for the segment most exposed to the mechanism mismatch — venture-stage holders, holders of EIS portfolios, holders of LP interests in closed-ended funds. Some structural restructuring (holding company creation, family-investment-company use, life-insurance funding) becomes more common as advised practice. For industrial strategy, the cost is real but bounded — the UK retains most of its founder cohort but gives up some at the margin. Whether this is acceptable depends on Question 5: if the regime's primary purpose is revenue, the leakage is the metric that matters; if it is industrial-strategy alignment, the loss of even bounded numbers of growth-stage founders has a multiplicative effect on the next cohort's calculation.

The honest qualification. This is the scenario in which design choices matter most, and in which conditional frameworks like the one in Section 3 are most likely to be useful — because the evidence does not clearly support either Position A or Position B, and both reasonable people and reasonable governments could land on either side.

Scenario three · Substantial capital flight and operational pressure

If HMRC modelling shows annual departures above 80, SAV caseload growing superlinearly with multi-year disputes the norm, receipts more than 20 per cent below forecast, and adviser-survey evidence of widespread relocation planning — the regime is in trouble. Position B's case strengthens substantially.

What this would mean. For Treasury, the choice becomes politically difficult but empirically clearer. The regime in its current form is not collecting what was forecast and is producing a measurable behavioural response. Mechanism change to a realisation basis would likely collect more, against optimised exit values, with lower administrative cost — but at the political price of being seen to retreat under lobbying pressure. For the affected cohort, the calculation that drove the relocation has been validated; the cohort that left does not return, regardless of subsequent policy change. For industrial strategy, the damage is significant and persistent; the UK loses the perception of being a place to build a company that you intend to hold long-term, and that perception affects not just current founders but the next generation considering the calculation. For the long-term tax base, mechanism change recovers some receipts going forward, but the lost cohort is lost. The principle of fairness across asset classes is sustained nominally — the law still treats all wealth equally — but is undermined operationally, because the wealthy mobile cohort has reorganised around the law while the immobile cohort has not.

The honest qualification. Even in this scenario, switching to Position B is not costless. Long-stop deemed-disposal mechanics need careful design or they recreate the original liquidity problem at the long-stop date. Some segment of the cohort will continue to find the mechanism uncomfortable — pre-revenue founders facing a long-stop charge they cannot fund without forced sale, for example. No mechanism is friction-free for every member of every cohort.

Scenario four · The data is contested or delayed

The likeliest real-world outcome, on the historical record of UK tax policy reviews, is that the modelling produces a mixed picture, that the trigger thresholds in any conditional framework are challenged on definition, that the political environment by month 12 differs from the one in which the framework was set, and that the formal review does not produce a clean fire-or-not answer.

What this would mean. For any conditional framework — including the one in Section 3 — credibility depends on the trigger firing under conditions advocates will agree have been met. If the data is genuinely contested, the trigger becomes another consultation rather than a forcing function. The political room that the framework was meant to bind closes; ministers face the choice without the cover the framework was intended to provide. For Treasury, this is the scenario most likely to result in the regime drifting in its current form for several years before any further change, regardless of what the underlying empirics actually show. For the affected cohort, the operational uncertainty extends — the question is not resolved, the rules do not change, and individual planning decisions are made under continuing ambiguity. For long-term policy quality, this is the worst outcome: the regime persists not because it is right but because the political conditions for changing it never quite cohere.

The honest qualification. This scenario is what makes Position A and Position B's defenders most uncomfortable with conditional frameworks like Section 3's, in different ways. Position A's defenders fear the trigger will fire under contested evidence and produce a hasty mechanism change. Position B's defenders fear the trigger will fail to fire under supportive evidence and entrench an inadequate regime. Both fears are reasonable. The honest answer is that conditional frameworks are not robust to data-quality failure, and any framework that pre-commits to evidence-based action assumes evidence will be deliverable in a form that supports action. That assumption is the framework's most fragile element.

7. The limits of this analysis

Three limits are worth naming honestly.

First, the analysis is system-internal by design — and the most consequential behavioural responses operate at the system-selection level the analysis does not engage with. The analysis above asks: given the UK tax system, how should the regime work for this asset class? The proposals — safe-harbour valuation, anti-avoidance rules, buy-back pathways, conditional-trigger frameworks — operate within the boundary of the UK system. They address responses HMRC can observe and Parliament can legislate against.

Sophisticated holders of unlisted trading-company shares operate at a different level. They ask: given multiple tax systems available to me, which one should I sit inside? Trust structures established years before death, holding-company restructures that move shares out of personal estate, life-insurance funding of any residual liability, residence planning that uses the TNR window strategically rather than passively, primary residence changes to jurisdictions with no inheritance tax — these are not responses to the UK regime; they are exits from it. The Companies House director-departure data, the Sifted adviser-survey reporting, and the OBR's own 25 per cent behavioural assumption for excluded-property-trust non-doms all point in the same direction: the most mobile and most advised members of the affected cohort are not waiting for the regime to bite before responding, and many are responding by leaving the system the analysis is designed for.

This boundary is intentional. An analysis that foregrounded system-selection responses would lead to no clear domestic fix, because system-selection responses are not legislatively addressable through the kind of measures the article proposes. The trade-off is real: the article's tractability comes from staying within the system; its most significant limit is that the cohort the policy is most concerned to retain operates at the system-selection level. The article's fixes may work perfectly within the UK system while behaviour outside the system dominates outcomes — meaning policy could appear successful by its own measures while high-value activity quietly relocates or restructures. The system-selection question deserves its own treatment; this article does not provide it.

The framing is borrowed from an AI cross-critique session run on an earlier version of this article (a separate chat with one of the four AI tools, prompted to read the draft as a hostile reader would), which put the distinction more cleanly than the article's own earlier drafts had managed: the article operates in a "given the system, how do we fix it?" model, while sophisticated actors operate in a "given multiple systems, which one do we sit inside?" model. That is the fundamental mismatch. The framing was correct. The within-system analysis remains, in this analyst's view, strong on its own terms. The across-system dynamics are largely out of scope. That boundary is intentional, but it matters, and a reader who needs an analysis of the across-system question will not find it here.

Second, the analysis is asymmetric in its quantification, and the data it would rely on is partial. The friction side of the regime — SAV caseload, dispute resolution time, valuation discounts — gets specific numbers and trigger thresholds. The outcome side — relocation, restructuring volume, capital flight, long-term reinvestment — gets qualitative treatment. This asymmetry exists because the friction data is collectable and the outcome data is not, but the asymmetry is real. Worse, the outcome data that does exist measures a population that has already partly responded: as Section 1.5 notes, the Companies House director-departure data already shows a 40 per cent year-on-year increase between October 2024 and July 2025, before the IHT reform took effect. Any HMRC modelling commissioned now will measure the cohort that remained; the cohort that pre-positioned is already outside the dataset. The reader should adjust their weighting of any quantitative finding accordingly.

Third, the normative question is not settled by any analysis. Whether the regime's primary purpose is revenue, fairness across asset classes, or industrial-strategy alignment is a political choice, not an empirical finding. The four scenarios above describe consequences for each of these objectives separately, but they cannot tell the reader which objective should take precedence when they conflict. That is a matter for ministers, voters, and the deliberative judgment of the people most directly affected. The piece is structured to inform that judgment, not to pre-empt it.

Closing

The reformed inheritance tax regime is the right reform in principle in the view of all four serious positions. The mechanism for one specific asset class is contested, on grounds that turn on five questions whose answers are not fully knowable today. Four positions and several conditional frameworks have been proposed; each rests on different empirical assumptions and different normative priorities; each has identifiable consequences under each plausible empirical outcome.

This piece has tried to set out the question, the open evidence, and what different outcomes would imply, without arguing for any of them. Readers will reach different conclusions depending on which empirical outcomes they consider most likely and which normative framing they treat as primary. That is the nature of policy questions answered under genuine uncertainty.

What would help the question resolve, in any direction, is the modelling that has not yet been done and the data that the regime's first operational year will produce. Until then, serious people will reach different conclusions about the right answer with the same evidence. That is not a failure of analysis; it is what the question actually looks like.

---

Postscript · A note on this version

An earlier version of this article presented one framework — Position C with hard triggers toward Position B — as a recommendation. An AI cross-critique session — Claude in a fresh chat, prompted to read the published article as a hostile reader would — noted that this was advocacy with the architecture of analysis rather than analysis itself. That AI critique session was right.

This version is rewritten in a different register: the framework is presented as one possible response to the current uncertainty rather than as the right answer, and a new section on consequences makes the stakes of each empirical outcome explicit rather than implicit. Where the earlier version asked the reader to accept a recommendation, this version aims to give the reader the materials to form their own view.

Two further points the original critique raised, which are not addressed in this version because they require more space than a postscript: the article assumes the principle of taxing intergenerational business wealth without defending that assumption explicitly; and it focuses on friction more than on outcomes. Both will be the subject of later pieces.

— Doug Scott. This version published in place of an earlier recommendation-format version, in response to substantive critique from an AI cross-critique session.

---

A coda — speaking directly, once

The analysis above is real and stands as analysis. The design positions are presented as the publication can honestly see them; the model is the publication's best attempt to make the dependencies between assumptions and outputs visible; the corrections page records what the publication has been wrong about and what it changed in response. If the piece has been useful on its own terms, that is genuinely the publication's first hope.

The reason it exists, though, is harder to explain in the policy register the rest of the piece has used, and so the publication has not tried to explain it in the body of the analysis. I will try to say it here, in the first person, because a reader who has got this far has earned the answer to the question they are probably asking, which is: why has this person put a day of concentrated AI-assisted work into producing institutional-grade analysis on a specific UK tax-policy question, and signed seven separate websites with their name in a single month?

The eight sites are one project. The project is about how humanity is in relation to the machines being built right now, who is making them, and what is being lost and held as that relationship changes. The Many Builders names the people whose work has gone into the systems we talk about as if they had built themselves. The trilogy (If This Road, orphans.ai, theheld.ai) works through what building, losing, and holding feel like from the inside — the wake, the diagnosis, the disposition. The three bear books (The Bear Was Right, The Bear Loved, The ADHD Bear) speak in a voice the policy register cannot carry. This publication treats a tax-policy question that affects a specific cohort whose response to AI may matter, and chose this question because the cohort whose attention to the larger questions about humanity and machines might matter most are, in general, not the readers who would click on a site about humanity and machines directly — but they are the readers who will click on a site about a tax that affects them.

The IHT analysis is not a pretext for the larger work. It is real analysis of a real question, written with the workflow the project uses, with the limits of that workflow stated openly throughout the corrections page and the disclosure blocks. But the reason I chose this question, rather than any of the dozens of others I could have chosen, is that the readers it reaches are the readers I cannot reach any other way. If the analysis has been useful, I am glad. If you finish here and click through to the rest of the work, that is what the project was built to make possible. The Many Builders is the place I would suggest you start.

— Doug Scott