Inheritance Tax and the Companies the Country Says It Wants More Of

What is being argued, what the disagreement turns on, and what different evidence would mean.

The reformed inheritance tax regime that took effect in April 2026 brings unlisted UK trading-company shares — founder equity, venture LP interests, EIS portfolios — into the inheritance tax base above a £2.5 million per-person allowance and £5 million combined allowance for couples. The reform is contested in a way that most reform debates are not: not over whether the change should have happened, but over whether the mechanism the reform uses is operationally fit for the asset class it now applies to.

This piece does not argue that any of the positions canvassed below is correct. It sets out what each is, what they share, what their differences turn on, and what different empirical outcomes would imply. The intent is to help a reader think clearly about a question that has been argued at high volume and low resolution.

What is actually in dispute

Three serious positions have been advanced by knowledgeable people in good faith.

• Position A — Hold the existing mechanism. Keep IHT-at-death, adopt practical fixes, resource HMRC adequately. The principle of the reform requires a death-event tax; the affected cohort is small (about 220 estates in year one); mechanism change risks special carve-out for the wealthiest holders.
• Position B — Switch to Capital Gains Tax on realisation. Replace IHT-at-death with CGT charged on actual proceeds when the heir realises, with no base-cost uplift on death and a long-stop deemed-disposal rule. The death-event mechanism mismatches the asset class; a realisation-based regime collects the same principle against optimised exit values.
• Position C — Hybrid, fixes only, mechanism question deferred. Adopt the practical measures both sides accept; defer the mechanism question pending evidence on relocation, dispute caseload, and receipts.

All three positions support four practical measures: a collateral-trigger rule, a statutory safe-harbour valuation methodology, targeted tightening of temporary non-residence rules, and structured buy-back guidance for cash-generative companies. These four measures are common ground. The actual disagreement is whether they are sufficient on their own.

A note on timing

Before the rest of the discussion: the behavioural response to the announced reforms began before the reforms took effect. The Office for Budget Responsibility's official January 2025 costing of the non-dom reform — the source for the £33.9 billion the reforms are projected to raise — assumes that approximately 25 per cent of non-doms with excluded property trusts (the wealthiest, most mobile cohort, most directly comparable to the IHT-affected population) will leave the UK as a result of the reform. That is not a lobby figure; it is the central UK government assumption.

The directly observable evidence is consistent with this assumption. Financial Times analysis of Companies House records shows 3,790 UK company directors changed their primary residence to abroad between October 2024 and July 2025 — a 40 per cent increase on the same period a year earlier, with director departures in April 2025 alone running 79 per cent above April 2024 and 104 per cent above April 2023. Approximately 150 directors moved specifically to the UAE in the second quarter of 2025.

The named tech-founder cases (Nik Storonsky of Revolut, Herman Narula of Improbable) and the adviser-survey reporting from Sifted (15–20 per cent of new business enquiries at one specialist firm now concern UAE relocation) point in the same direction. The widely-cited Henley & Partners figure of 16,500 millionaire departures in 2025 has been forensically critiqued by Tax Policy Associates and Tax Justice Network and should be treated with caution; the OBR figures and the Companies House data have not been similarly contested.

What this means for the rest of this piece: any analytical framework calibrated against post-April-2026 data will measure a population that has already partly responded. The early-mover cohort has already exited the dataset. So measured outcomes will understate true behavioural response. The reader should hold this in mind through the analysis below.

What the disagreement turns on

Five questions, none of which has been answered with rigour:

1. How large is the relocation channel under the existing mechanism? Empirical, answerable through HMRC dynamic modelling. Currently relying on adviser surveys and trade-press reporting, neither sufficient.
2. How quickly does SAV dispute caseload grow as the cohort grows? Empirical. Depends partly on the rate of cohort growth, which industrial strategy is trying to accelerate.
3. How effective is the collateral trigger against sophisticated structuring? Technical. Depends on drafting quality.
4. Can a 10- or 15-year long-stop be administered without recreating the original problem? Empirical. Depends on the realistic distribution of realisation events across the cohort.
5. What is the regime for? Normative. Revenue, fairness across asset classes, or industrial-strategy alignment? When they conflict, which wins? Ministers, not analysts, must answer this.

Reasonable readers reach different conclusions on these questions, and that is most of what produces the disagreement between the three positions. A reader with a clear view on the questions has, by implication, a clear view on the policy.

One possible framework

Some commentators have proposed a fourth design: adopt the four practical measures within ninety days; commission HMRC dynamic modelling within thirty days; set a twelve-month formal review against five calibrated trigger criteria; if any two of the criteria breach defined thresholds at the review point, a pre-drafted Capital Gains Tax mechanism takes effect at the next Finance Bill.

This is one workable design for acting under uncertainty. It is not a recommendation of this article. Position A's defenders argue, with force, that no trigger is needed — observed data should drive any subsequent change, not pre-committed thresholds. Position B's defenders argue that the mechanism switch should happen now rather than wait for confirmation that may arrive too late to retain the cohort. Each design is internally coherent. Each rests on different assumptions about behavioural response and triggered review.

The full version of this paper sets out the conditional framework in detail. This summary moves on to what different evidence would actually imply, because that is what determines which design is right.

What different evidence would mean

Four scenarios, ordered by what HMRC modelling and observed receipts data could plausibly show. The reader should hold the timing point above in mind: each scenario describes what the data could measure, but the data will measure a population that has already partly responded. A small measured response could mean either a small underlying response or a large response already absorbed into the pre-reform departures.

If the relocation channel is small and SAV caseload manageable — annual departures below ~30, caseload growth tracking estate count, receipts within forecast — Position A is broadly vindicated. The four practical measures are sufficient. Mechanism change is unnecessary. Treasury collects what was forecast; the affected cohort finds the regime more navigable than the announced response suggested; the principle of fairness across asset classes is preserved. The venture-stage cohort still faces residual operational difficulty, but too small a group to drive policy change.

If the relocation channel is meaningful but bounded — annual departures in the 30–80 range, caseload somewhat faster than estate count, receipts 5–15% below forecast — the picture is mixed. Holding the regime delivers most of the revenue but loses some; switching may collect more but at the political cost of being seen to capitulate. The regime is workable for most but pressure persists for the segment most exposed to mechanism mismatch. Whether this is acceptable depends on Question 5: revenue, fairness, or industrial-strategy alignment as primary.

If there is substantial capital flight and operational pressure — annual departures above 80, caseload superlinear with multi-year disputes, receipts more than 20% below forecast — Position B's case strengthens substantially. The regime in its current form is not collecting what was forecast and is producing a measurable behavioural response. Mechanism change recovers some receipts going forward, but the cohort that left does not return. The principle of fairness is sustained nominally but undermined operationally. The damage to industrial strategy is significant and persistent — affecting not just current founders but the next cohort considering the calculation.

If the data is contested or delayed — the likeliest real-world outcome, on the historical record of UK tax policy reviews — any conditional framework's 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 regime drifts in its current form for several years, regardless of the underlying empirics. This is the scenario that makes both Position A's and Position B's defenders most uncomfortable with conditional frameworks, in different ways.

The limits of this analysis

Three limits are worth naming. The analysis underplays strategic behaviour — sophisticated holders restructure to avoid the question rather than respond to it, the OBR's 25 per cent behavioural assumption and the Companies House director-departure data both point to this happening already, and any HMRC modelling commissioned now will measure a depleted cohort. The analysis is asymmetric in its quantification — friction is measurable, outcomes are not, and the reader should weight accordingly. The normative question (what the regime is for) is a political choice, not an empirical finding, and no analysis can settle it.

Closing

The reformed inheritance tax regime is the right reform in principle in the view of all three 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. Three 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 — not a failure of analysis, but what the question actually looks like.