UK Tech and the IHT Reform — The Funding Stack and the Fiscal Model

Technical depth on each part of the UK tech funding stack — founders, angels, VCs, PE, EIS, LPs, early employees — and a 25-year fiscal model of what each policy option means for the Treasury, including second-order effects.

The reform, recapped

From 6 April 2026, 100 per cent Business Property Relief is capped at £2.5 million per person of combined qualifying agricultural and business property, transferable between spouses to give a combined £5 million per couple. Above the cap, 50 per cent relief applies, producing an effective 20 per cent IHT rate. The £2.5 million allowance refreshes every seven years for individuals (ten for trusts) and is set to be CPI-indexed from April 2031 subject to a future government implementing it via statutory instrument. AIM-listed shares now receive 50 per cent relief from the first pound, with no allowance; AIM-listed EIS shares fall under this AIM treatment, while unlisted-private-company EIS shares fall under the £2.5m cap-then-50% regime alongside other unlisted trading-company shares. The ten-year interest-free instalment option, previously available for some forms of business property, has been extended to all APR/BPR-eligible property. The original Autumn Budget 2024 cap was £1 million; this was raised to £2.5 million on 23 December 2025 in response to lobbying. The Finance Act 2026 received Royal Assent reflecting these changes.

The rest of this piece sets out, by cohort, what the reform actually does — and the technical mechanics that determine how hard it bites — before turning to the question of what each policy option means for the UK Treasury over a 25-year horizon, including the second-order tax base from the companies the cohort builds.

Founders

The valuation problem at death

The mechanism the reform retains for unlisted UK tech-trading-company shares is death-event valuation. HMRC's Shares and Assets Valuation team determines the open-market value of the deceased's shares as at the date of death, applying the following framework. Open-market value per the SAV manual: the price the shares would fetch on a sale between a willing buyer and a willing seller, both having access to all relevant information, neither acting under compulsion. For unlisted private-company shares, this is a hypothetical price; no actual market exists. The recent funding-round valuation is the typical anchor — SAV will look at the most recent priced equity round as a starting point, but apply adjustments for time elapsed since the round, changes in company performance and outlook, share class differences (most rounds price preference shares; the founder typically holds common stock), and the specific shareholder agreement in force.

SAV will then apply a minority discount of 15-30% for non-controlling stakes and a discount for lack of marketability (DLOM) of typically 25-40%, reflecting the absence of a secondary market. Where the cap table includes preference shares with liquidation preferences, the value of common stock is reduced by present-value waterfall analysis, modelling the distribution of expected exit values against the preference stack. The combined effect: a founder's headline paper valuation typically overstates the SAV-adjusted value by 30-60% for venture-backed equity. A £10m headline valuation may produce an SAV-agreed value of £4-7m. The negotiation takes time and the outcome varies materially with the quality of advice the estate can fund.

Funding the tax bill

For founders facing a tax bill above the cap, three primary funding routes exist. The ten-year instalment option is now extended to all APR/BPR-eligible property and interest-free for the duration. A £2m IHT bill becomes £200,000 a year for ten years. For founders of profitable companies, this is fundable from operating dividends. For pre-profit growth-stage companies, it is not — the company is reinvesting all cash and the founder's estate has no income source from which to fund the instalments. Company share buy-back under CTA 2010 s.1033 is the second route: the company purchases its own shares from the estate, providing cash to fund the IHT liability, treated as capital rather than income for the seller. The practical issue is that the buy-back requires distributable reserves at the company level; venture-backed pre-profit companies typically do not have these. The third route is life insurance written into trust, the cleanest solution for founders in good health and at younger ages.

The instalment provision, taken on its strongest terms

The ten-year interest-free instalment regime under IHTA 1984 s.227, as extended by the Finance Act 2026 to all APR/BPR-eligible property, is the single strongest objection to the publication's case for changing the death-based mechanism. An earlier version of this piece soft-pedalled the provision in places where the analysis would have been sharper if the provision had been engaged with directly. This section takes the provision on its strongest terms.

The provision's strongest version is this: an estate liable for IHT on qualifying business property may pay the tax in ten equal annual instalments, interest-free for the full ten years, with the first instalment due six months after the end of the month in which death occurred. For a £2m IHT bill — the size of bill that arises on a £15m founder holding under the new regime — the annual instalment is £200,000. For a £5m bill on a £30m holding, the annual instalment is £500,000. Spread against ten years of expected company growth, eventual exit liquidity, dividends from a profitable company, secondary share sales, or the heir's own income, this is for many cohorts a manageable obligation. It is not the death-trap the lobby framing has often suggested.

For which segments of the cohort does the instalment provision cleanly resolve the liquidity problem?

1. Founders of cash-generative private companies above the cap. Where the company generates distributable profits in the £1m-£10m+ annual range, the founder's heirs can fund the instalments through dividends. The corporation tax effect on the dividend funding is real but manageable. The publication regards the instalment provision as cleanly sufficient for this cohort and concedes that Position B's case (death-event mechanism is broken) is materially weaker for cash-generative private companies than for pre-profit growth-stage companies. This cohort is the largest single sub-segment of the BPR-affected population by estate count.
2. Founders of public-stage or near-public-stage growth companies. Where the company is at IPO-readiness and secondary market liquidity exists, the heir can fund instalments through staged secondary sales of inherited shares. The provision works cleanly because the asset is liquid even if the liquidity is structured. SAV will engage with the secondary-market evidence in valuation, which mitigates the valuation-uncertainty problem described elsewhere in this piece.
3. Estates with diversified asset bases. Where the founder's estate includes substantial liquid assets — public-equity portfolios, property, deposits, life-insurance proceeds — alongside the BPR-eligible business holding, the heirs can fund instalments from non-business assets without forced disposal of the business holding. The instalment provision functions as an effective deferral mechanism. The death-event nature of the tax becomes a timing inconvenience rather than a liquidity crisis.
4. Founders with adequate life insurance written into trust. Where life insurance proceeds are held outside the estate and sized to cover the IHT liability, the instalment option becomes optional rather than necessary, and the heirs can either pay the IHT in full from insurance proceeds or use insurance proceeds to fund instalments while preserving estate liquidity. The market for term and whole-of-life policies sized to cover IHT liabilities for founder estates is well-developed.

For which segments does the instalment provision not resolve the liquidity problem?

1. Pre-profit venture-backed founders. Where the company has no distributable reserves, no s.1033 buy-back capacity, no near-term exit liquidity, and the founder's estate has no income source from which to fund instalments, the instalment provision provides time but not money. The heirs face a choice between forced secondary sale (often impossible at any price for a private growth company between rounds), heir-funded debt against the inherited shares (a structurally difficult borrow given the asset's illiquidity), or default on instalments — which compounds the IHT with interest from the date of default and ultimately triggers HMRC enforcement against the estate's other assets if any exist. This is the cohort the publication's case for mechanism change is strongest for.
2. LP interests in closed-ended funds. Already discussed in the LP section. The fund cannot buy back the LP interest. Distributions arrive on the GP's exit timetable, which can be 8-12 years post-commitment and is unpredictable within that window. The instalment provision's timing exactly does not match the fund's distribution timing, and the heir owes tax against a paper valuation of an asset that may produce no cash for years.
3. Concentrated EIS portfolios in early-stage companies. Where the holdings are individually too small to interest a secondary buyer, individually unable to support a buy-back, and the portfolio's diversification cannot offset the per-holding illiquidity. EIS-portfolio investors typically expect a multi-year holding pattern with substantial loss-tolerance built in; an unscheduled IHT event mid-portfolio creates the same problem the venture-LP cohort faces, in a more fragmented form.

The honest finding is therefore narrower than Position B's strongest framing has often suggested. The ten-year interest-free instalment provision resolves the liquidity problem cleanly for perhaps half of the BPR-affected cohort by estate count — the cash-generative private companies, the public-stage growth companies, the diversified estates, and the well-insured founder estates. For these segments, Position B's mechanism-change case is materially weaker than the publication's earlier framing suggested. The death-event mechanism, with the instalment provision attached, is operationally tolerable.

The residual case for mechanism change therefore concentrates on the segments where the instalment provision does not cleanly work — the pre-profit venture-backed founders, the closed-ended LP interests, the early-stage EIS portfolios. These are smaller subsets of the BPR-affected cohort by estate count but are the segments that the public-debate has rightly focused on, and they are the segments where the publication's case for mechanism change retains force. The publication's revised position, after engaging the instalment provision on its strongest terms: the death-based mechanism with the instalment provision attached works for most of the BPR-affected cohort. It does not work cleanly for the pre-profit venture-equity, LP-fund-interest, and concentrated-EIS-portfolio segments. The case for either mechanism change (Option B) or carve-out treatment (a narrower Option C) is strongest for these specific segments rather than for the full cohort.

This is a genuine concession. An earlier framing of this piece treated the case for mechanism change as applying with roughly equal force across the cohort. After engaging the instalment provision properly, the case is materially narrower than that. Position A's defenders are correct that, for the largest segment of the affected cohort by estate count, the instalment provision substantially mitigates the liquidity case. Position B's defenders are correct that, for the pre-profit venture-equity and LP-interest segments, the instalment provision does not solve the problem. The honest version of the analysis acknowledges both findings and points toward a narrower carve-out architecture — possibly the "two-track" design referenced elsewhere in the publication — rather than a wholesale mechanism shift.

Lifetime structures

Beyond death-event funding, founders have three primary lifetime structures: potentially exempt transfers (PETs) where the gift falls outside the IHT estate after seven years, lifetime trusts (with a 20% immediate IHT charge above the £2.5m trust allowance plus subsequent ten-year periodic charges), and family investment companies (FICs) where the founder transfers shares to a UK-incorporated company held by family members. The seven-year refresh of the £2.5m allowance interacts with PETs in ways that create planning capacity over time — multiple smaller PETs can deploy more allowance than a single death-date claim. FICs were under HMRC scrutiny during 2020-2024 but no major restrictive legislation has emerged; the structure remains usable but increasingly compliance-heavy.

Residence planning

The most consequential structural response, and the one the relocation evidence base engages with most directly. The April 2025 non-dom reforms substantially changed the UK's tax-residence and IHT-exposure framework. From 6 April 2025, the historical concept of domicile was replaced for tax purposes by a residence-based long-term-residence framework. Under the new framework, an individual is a "long-term UK resident" — and so liable to UK IHT on worldwide assets — once they have been UK-resident in 10 of the previous 20 tax years. The pre-2025 domicile rules continue to have legacy effects in transitional cases, but the operative test for IHT exposure post-April 2025 is the long-term-residence test, not domicile.

The Statutory Residence Test (SRT) under FA 2013 Schedule 45 remains the operative test for whether an individual is UK-resident in any given year — a complex multi-factor test based on UK days, ties to the UK (family, accommodation, work, 90-day rule, country tie), and prior UK residence history. A founder seeking to escape UK long-term-resident status must accumulate enough non-UK-resident years to fall outside the 10-of-20 test. The "tail" exposure for IHT after losing UK residence varies depending on prior length of UK residence and is set out in detail in the post-2025 legislation; it is not a simple three-year window in all cases. The temporary non-residence rules under TCGA 1992 s.10A continue to charge gains realised during a period of non-residence on return to the UK if the period is five years or less. The practical implication: a founder relocating must commit to several years of genuine non-UK residence to escape long-term-resident status, and must plan for the temporary-non-residence rules on any UK-asset disposals during the period. The post-2025 framework is more complex than the pre-2025 domicile-based framework, and the working-out of the new rules in practice is still being settled by HMRC guidance and (eventually) tribunal decisions.

Angels and EIS investors

EIS qualification and BPR interaction

EIS shares in unlisted private companies qualify for BPR after a two-year holding period, provided the issuing company continues to qualify as a trading company under BPR rules. The two-year clock runs from the date of issue, not the date of EIS reinvestment certification. For most unlisted-private-company EIS positions, the BPR qualification is straightforward: tech-trading companies typically qualify clearly. Edge cases arise where companies pivot toward investment-heavy operations post-investment. The reform caps 100 per cent BPR at £2.5m of combined APR/BPR property (with the allowance set to be CPI-indexed from April 2031, subject to future statutory instrument). For an EIS investor whose primary BPR-qualifying assets are unlisted-private-company EIS shares, the £2.5m cap applies to the aggregate value of those shares at death. The 50 per cent relief above the cap reduces the IHT impact relative to no-relief assets but does not eliminate it. AIM-listed EIS shares, by contrast, receive 50 per cent relief from the first pound regardless of the £2.5m allowance — the same treatment as other AIM holdings.

The compound interaction of EIS reliefs

EIS provides several reliefs that interact in non-obvious ways with the new BPR regime. Income tax relief at 30 per cent on the amount invested, up to £1m per year (£2m for knowledge-intensive companies), is unaffected by the BPR reform. CGT exemption on disposal of EIS shares held for at least three years is unaffected. Loss relief on failures is unaffected. CGT deferral relief for gains reinvested through EIS is unaffected. BPR on death is the change — capped at £2.5m. The compound implication: the front-end EIS reliefs preserve most of the economic case for active angel investing through EIS in any single year; the BPR cap matters cumulatively over time, when the portfolio reaches material scale. For an active angel investing £150,000 to £400,000 a year across 8-15 EIS positions, the front-end reliefs typically dominate the economic calculation in any single year, and the BPR cap becomes the binding constraint after 8-12 years of active investing.

Portfolio dynamics

The diversified nature of EIS portfolios changes the death-event valuation problem materially. An angel with twenty EIS positions across ten years of investing has twenty separate valuation determinations to make at death, each subject to its own SAV negotiation. The administrative burden is non-trivial but the diversification smooths the outcome distribution: where a founder estate's IHT liability is highly sensitive to a single company's death-date valuation, an EIS angel's liability is the sum of twenty smaller, less correlated valuations. The death-event valuation of EIS positions held in companies that have themselves not yet had a recent priced round is harder to anchor than the founder case. SAV's approach in these cases varies; in practice, prior round valuations adjusted for time elapsed and company-specific factors are typically the basis. The negotiation is more contested than the founder case but the lower individual stakes reduce the cost of mis-pricing any single position.

Second-order effects on the seed-stage ecosystem

The harder question for UK tech is not the first-order effect on individual angels but the second-order effect on seed-stage capital formation. UK angel capital — particularly the operator-investor segment, where successful former founders invest in next-generation founders — is structurally important to UK seed-stage tech. The marginal active angel deciding whether to commit £200,000 to a new EIS position considers, among other factors, the expected lifetime portfolio value and the IHT shelter applicable to it. If the most active angels (those at or above the £2.5m cap on EIS holdings alone) reduce activity at the margins, or relocate, the seed-stage ecosystem loses both capital and the operator-investor expertise that distinguishes UK angel money from purely financial capital. The aggregate effect is hard to estimate without sector data, but the direction is consistent across the practitioner reports the author has reviewed.

Venture capital and growth equity (LP interests)

BPR qualification at the LP-interest level

Limited partner interests in UK venture and growth-equity funds qualify for BPR if the underlying fund is investing predominantly in qualifying trading companies, with the two-year holding period applied at the LP-interest level (not the underlying portfolio-company level). The technical detail matters: for funds investing across multiple holding-company and SPV structures, the BPR qualification depends on the substance of the underlying investments, not the formal LP-interest characterisation. For most UK venture and growth funds investing primarily in UK and other-jurisdiction tech-trading companies, BPR qualification is clear. For some structures (real-asset-heavy, late-stage growth funds investing in companies with significant non-trading activities, or funds with material public-equity sleeves), BPR qualification is contested.

The illiquidity problem at death

The death-event valuation problem is sharpest for LP interests because the asset is illiquid by structural design. A UK-resident LP with £10m committed across three or four UK venture funds has, at any given moment, capital deployed across underlying portfolio companies whose values are reported by the funds at NAV but not realised. NAV reporting methodologies vary: some funds use last-round-price for portfolio companies; some use marked-to-model approaches that adjust for time elapsed and company performance; some apply standardised discounts. SAV's valuation of an LP interest at death typically considers the fund's most recent NAV report, a discount for lack of marketability of typically 25-40% for closed-ended fund interests with no secondary market, adjustments for time elapsed since the most recent NAV report, company-specific information for material underlying portfolio positions, and the fund's expected liquidation horizon. The valuation negotiation is technically complex and slow. Estates often retain specialist valuation advisors at material expense to argue for higher discounts; SAV resourcing constraints mean the negotiation can extend over years.

Funding the LP interest's tax bill

The LP cohort's funding options are more constrained than founders' or angels'. The fund itself does not provide a buy-back mechanism for LP interests. Distributions from the fund are unpredictable in timing — they arrive when the GP exits underlying portfolio companies, which the LP cannot control. The instalment option helps but does not solve the underlying problem: the heir owes tax against a paper valuation of an asset they cannot sell and from which they may not see distributions for years. The LP-interest secondary market exists but typically requires a 15-25% discount to NAV for venture-fund interests, and even larger discounts for early-life-cycle funds or specialty strategies. Practical funding routes include selling the LP interest into the secondary market at a discount, selling other estate assets to fund the IHT, life insurance written into trust, and the ten-year interest-free instalment option spread against expected fund distributions.

UK venture-fund capital formation

The behavioural risk for the UK venture industry is that high-net-worth UK-resident LPs reduce commitments to UK funds. Several cumulative effects: marginal commitments redirected to non-UK funds, some LPs relocating, structural responses through offshore vehicles, family-office structures, or trust arrangements. The aggregate effect on UK venture-fund capital formation over time is hard to quantify without sector data but directionally clear: at the margin, less UK-resident LP capital flows into UK venture funds than would have done under the prior regime. The funds themselves remain operationally viable — most large UK venture funds raise materially from non-UK LPs — but the UK-resident LP base shrinks at the margin, and the connection between UK private wealth and UK tech funding weakens commensurately.

Private equity in UK tech

Personal partner exposure

Private equity invested in growth-stage and mature UK tech companies interacts with the reform mainly through the personal IHT exposure of UK-resident partners. Carried interest, by its nature, vests over time and accrues against successful fund performance; for UK-resident PE partners with material vested carry above the £2.5m allowance, the IHT exposure is real. The cohort is small in number — perhaps a few hundred UK-resident senior PE professionals at this scale — but well-advised, and structural responses are typically already in place from prior carried-interest reforms. The reform's first-order effect is to accelerate succession planning that this cohort was largely already doing.

Co-investments and direct holdings

Beyond carry, UK-resident PE partners typically hold direct equity in portfolio companies through co-investment programmes. These holdings are unlisted UK trading-company shares and qualify for BPR subject to the new £2.5m cap. For the largest UK-resident PE partners with significant direct co-investment exposure, the IHT impact is material; the structural responses (offshore trust arrangements where eligible, family investment companies, residence planning) mirror founders' but are typically more sophisticated.

Effects on PE-backed portfolio companies

The most consequential effect on PE-backed UK tech companies is indirect, through the founders the PE fund has retained or backed in. Growth-stage PE structures increasingly prefer founder retention with rolling equity, particularly in tech. Where founders of PE-backed UK tech companies retain material personal stakes, 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.

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 is now compounded by an additional planning consideration.

Modelling the second-order fiscal effects

The analysis above sets out what the reform does to the cohort. The corresponding question — what does the reform do to the UK Treasury, and what do the three policy options mean for UK fiscal position over the long term — has not been put to numbers in the public debate. This section sets out a 25-year fiscal model that attempts that. The full model, with every assumption editable, is provided as a downloadable spreadsheet alongside this piece.

What the model is and is not

The model estimates the 25-year present-value fiscal effect on the UK Treasury under each of the three policy options (Hold the system, Change the mechanism, Wait and decide), under three scenarios (Best, Central, Worst). For each of the nine outcomes, it produces two figures: the direct fiscal effect (IHT or CGT collected from the cohort) and the indirect fiscal effect (the corporation tax, employment-related tax, VAT, and exit-event taxes from the companies the cohort builds, including the next companies that successful founders build after exit).

The model is most useful for showing the structure of dependencies. Which assumptions, when changed, change the answer. Where the actual fiscal stakes are largest. Whether the direct revenue effect or the indirect effect dominates under each option. Whether the worst case for one option is bigger than the worst case for another. The model is least useful as a source of definitive numbers. Every input is uncertain. The behavioural-response figures are uncertain. The company-tax-base figures are uncertain. The founder-repeat-rate figures are uncertain. The network-multiplier is contested. Years 15-25 of the projection are highly speculative because the next-company effect plays out over decades and the assumptions become correspondingly less reliable. The spreadsheet companion is the more useful artefact: any reader can substitute their own assumptions and watch the conclusion move.

What the model assumes

The cohort: approximately 50 UK tech founders enter the IHT-exposed cohort each year, with average paper equity above the £2.5m allowance of approximately £8m per founder. (The total HMRC December 2025 estimate is approximately 1,100 estates affected by the whole APR/BPR reform; the unlisted tech-trading-company subset is in the low hundreds within that, and the model focuses on the annual flow into that cohort.) Annual mortality among founders, weighted to typical founder age range, is approximately 0.8% per year.

Behavioural response varies by option and scenario. Under Option A (Hold), departure rates are 10% (Best), 20% (Central), 35% (Worst) — anchored by the OBR's 25% non-dom EPT figure with adjustments for population differences. Under Option B (Change), departure rates are lower because the death-event pressure is removed: 5% (Best), 10% (Central), 20% (Worst). Under Option C (Wait), the rates are intermediate and the model assumes a 40% probability that the trigger fires in year 3 and switches the regime to B from that point.

Direct tax: under Option A, IHT collected at the effective 20% rate on the equity above the £2.5m cap, less an SAV admin cost of approximately 6%. Under Option B, CGT collected on actual realisation at the prevailing 24% main rate, with an exit-multiplier of 1.6 (reflecting that successful exits typically realise above death-date value), an average lag of 8 years from death to realisation, and a 70% probability that the long-stop fires productively (some companies never realise). Admin cost on Option B is lower (approximately 3%) because realisation-event valuations are observable rather than negotiated.

Indirect tax — the second-order channel. The model estimates that a typical in-scope UK tech founder's current company contributes approximately £4m per year in UK tax base (corporation tax, PAYE/NICs employee and employer, VAT generated on UK operations) over a 12-year horizon. A network multiplier of 1.20 applied to retained-founder tax base captures cluster effects: capital retained, talent retained, peer founders retained. When a founder departs, approximately 30% of their current company's tax base follows them; the remaining 70% stays in the UK at least temporarily. A network damage term of 0.10 reflects that each departure pulls some additional value with it (other founders considering, capital reallocating, talent following).

The next-company channel is the largest second-order effect. Successful UK tech founders build second companies at an estimated 45% probability within 10 years of exit. Of those, 30% reach scale (defined as cohort-relevant scale). A scaled second company contributes approximately £5m per year in UK tax base over 10 years — slightly higher than the first company because experienced founders typically have more capital and execute faster. If the founder is UK-resident when building the second company, 100% of that tax base is UK; if they have relocated, only 5% of that tax base is captured by the UK (a small residual through any continuing UK linkage). The next-company effect ramps in linearly between years 5 and 15 of the projection as first-cohort founders begin to exit and start building again.

The discount rate is 3.5% (HMT Green Book social discount rate, real). The horizon is 25 years.

What the model produces

Under the central assumptions specified above, the 25-year NPV of the total UK fiscal effect under each of the nine outcomes is approximately:

Option	Scenario	Direct (IHT/CGT)	Indirect	Total fiscal NPV	vs A-Central
A — HOLD	Best	£86m	£44.4bn	£44.4bn	+£2.3bn
A — HOLD	Central (baseline)	£77m	£42.1bn	£42.1bn	—
A — HOLD	Worst	£62m	£38.6bn	£38.7bn	−£3.5bn
B — CHANGE	Best	£53m	£45.5bn	£45.6bn	+£3.4bn
B — CHANGE	Central	£50m	£44.4bn	£44.4bn	+£2.3bn
B — CHANGE	Worst	£44m	£42.1bn	£42.1bn	~£0
C — WAIT	Best	£74m	£45.1bn	£45.1bn	+£3.0bn
C — WAIT	Central	£68m	£43.2bn	£43.3bn	+£1.1bn
C — WAIT	Worst	£56m	£39.7bn	£39.8bn	−£2.3bn

All figures 25-year NPV at 3.5% real discount rate. Spreadsheet companion contains every assumption editable.

The headline finding — stated as a range, not a point

An earlier version of this section led with the central-case figure ("indirect fiscal effects are roughly 500 to 1,000 times larger than direct fiscal effects"). That framing was rhetorically effective and analytically misleading. A model whose output spans two orders of magnitude across plausible assumption sets should lead with the range, not the central case. This version corrects the framing.

Across the publication's central case, indirect-to-direct fiscal effects run at roughly 500x. Across the plausible range — running the same model with the four most consequential assumptions varied between defensibly-conservative and defensibly-optimistic settings — the ratio runs between approximately 50x and 1,000x. The headline finding the publication makes is therefore not "indirect fiscal effects dominate direct by 500-1,000x" (the earlier framing). It is: indirect fiscal effects from cohort retention plausibly dominate direct receipts at the moment of transfer by between 50x and 1,000x, depending on inputs the publication does not have empirical anchors for, with a central estimate at approximately 500x. The qualitative finding (indirect dominates direct, by a large multiple) survives every reasonable assumption set. The precise magnitude does not.

Stated as a sensitivity table across the four load-bearing assumptions:

Sensitivity case	Indirect:Direct ratio (25-year NPV)	What this scenario assumes
P10 — defensibly conservative	~50x	Current-company tax base £1m/year per founder (vs £4m central). Network multiplier 1.00 (no cluster effect). Departure rate at upper end of plausible range. Next-company probability 30% (vs 45% central).
P25	~150x	Current-company tax base £2m/year. Network multiplier 1.10. Departure rate near central. Next-company probability 35%.
P50 — central	~500x	Current-company tax base £4m/year. Network multiplier 1.20. Central departure rates. Next-company probability 45%.
P75	~750x	Current-company tax base £5m/year. Network multiplier 1.30. Central departure rates. Next-company probability 50%.
P90 — defensibly optimistic	~1,000x	Current-company tax base £6m/year. Network multiplier 1.40. Lower departure rates. Next-company probability 55%.

P10 / P25 / P50 / P75 / P90 are the publication's subjective probability intervals across the four load-bearing assumptions, not statistical confidence intervals. They reflect what the publication considers the defensibly-conservative through defensibly-optimistic range; a specialist reader may reasonably bracket more widely. The intervals are uniform-prior subjective ranges, not derived from sampling distributions over the inputs.

The four assumptions ordered by sensitivity weight (impact per unit change on the 25-year NPV):

1. Current-company UK tax base per typical in-scope founder (corporation tax + employment tax + VAT contribution from the founder's current company). Central £4m/year. Plausible range £1m to £6m. Largest single lever; a £1m change moves the 25-year NPV by approximately £8-10bn across all scenarios.
2. Network multiplier on retained-founder tax base (the cluster effect that retained founders generate beyond their direct contribution — through hiring, mentoring, follow-on investment, ecosystem density). Central 1.20x. Plausible range 1.00 to 1.40. Most contested input; the publication does not have a defensible empirical anchor for this and a reader who sets it to 1.00 should expect Option B's advantage to shrink materially.
3. Departure rate under each option (annual share of the in-scope cohort relocating in response to the regime). Plausible range varies by option; the publication uses 8-10% for Option A, 5-7% for Option B, 6-8% for Option C in the central case. Empirically anchored to the OBR's non-dom 12-25% range with substantial downward adjustment for the BPR-cohort being differently exposed; a 10 percentage-point change moves the NPV by £1.5-3.5bn.
4. Next-company probability (probability that an in-scope founder, conditional on remaining, builds a second company in the UK). Central 45%. Plausible range 30% to 55%. Empirically anchored to UK tech serial-founder rates; a 15 percentage-point change moves the NPV by £700m-£1bn.

The reader is strongly encouraged to run the interactive model with their own assumptions rather than take any single number from the table above as decisive. The model's value is showing what the answer is sensitive to, not delivering one number; the table above is itself a summary of model runs, with the underlying calculations in the downloadable spreadsheet.

The qualitative finding the publication makes is therefore narrower than the earlier framing suggested: indirect-to-direct dominance is highly likely (the publication regards the qualitative finding as surviving most defensibly-conservative assumption sets), but the magnitude is uncertain across roughly 1.3 orders of magnitude. The most important inputs to a reader's own assessment are the current-company tax base and the network multiplier; substituting more conservative figures for either compresses the ratio toward the lower end of the range above. The publication's claim is not that the ratio is 500x, or 1,000x, or any specific number; it is that the ratio is large enough — across all defensible assumption sets — to make the indirect channel the dominant fiscal consideration. Whether "large" means 50x, 500x, or 1,000x is a separate question that the available evidence does not yet settle.

The direct revenue from the IHT mechanism is, in fiscal terms, a rounding error compared to the indirect tax base the cohort generates — at any point in the range above. Which option collects more in direct revenue is, on the model, almost irrelevant to the total fiscal position. What matters — by at minimum 50x in the most conservative case — is which option retains more founders.

This finding inverts the common framing of the debate. The public conversation has been conducted as if the central question is "how much tax will the Treasury collect from the cohort under each option." The model suggests the central question is "which option keeps the cohort building UK companies." On the central assumptions, Option B (Change the mechanism) outperforms Option A (Hold) by approximately £2.3 billion in 25-year NPV, almost entirely through the indirect channel: lower departure rates under Option B mean more founders stay, more companies stay, more next companies get built in the UK, and more cumulative tax base gets generated.

Where the result is most sensitive

The model's headline finding is robust to most assumption changes but particularly sensitive to four inputs. The current-company tax base is the largest single lever: a £1m change in the assumed annual UK tax base contributed by a typical in-scope founder's current company (the central assumption is £4m per year) moves the 25-year NPV by approximately £8-10bn across all scenarios. The departure-rate assumption is the second largest: a 10 percentage-point change in the central case moves the NPV by approximately £1.5-3.5bn. The network multiplier is the third: a change from 1.20 to 1.40 (or back to 1.00) moves the NPV by approximately £4bn. The next-company probability is the fourth: a change from 45% to 30% moves the NPV by approximately £700m-£1bn.

The most contested assumption is the network multiplier. There is real evidence for cluster effects in tech ecosystems — Silicon Valley, Tel Aviv, Stockholm, London itself — but quantifying the multiplier on retained-founder tax base requires inferences from comparable ecosystems and is unavoidably uncertain. A reader who sets the multiplier to 1.00 (no cluster effect) and the network damage to 0.00 will see Option B's advantage over Option A shrink materially but not disappear; the departure-rate gap alone produces meaningful divergence. A reader who sets the multiplier higher (say 1.40) will see the gap widen further. The underlying point — that retention dominates direct revenue — is robust to a wide range of assumptions about the cluster effect.

What the worst cases show

Each option has a defensible best case and a defensible worst case. Reading the worst-case row by row is useful because it shows what kind of empirical world each option is most exposed to.

Option A's worst case (£38.7bn NPV, £3.5bn below baseline) is the world where the death-event mechanism produces a 35% departure rate over time, the cohort meaningfully thins, and the second-company channel is correspondingly weakened. This is the world Position B's defenders are most worried about: holding a regime that has already shown signs of producing the behavioural response, and watching the indirect tax base shrink while the direct revenue stays flat.

Option B's worst case (£42.1bn NPV, approximately at baseline) is the world where the long-stop fails on a meaningful fraction of cases, the deferred CGT collection arrives smaller than expected, and even with reduced departures the Treasury position does not improve over Option A. This is the world Position A's defenders are most worried about: switching to a regime that delivers less than its proponents promise, in exchange for a behavioural improvement that may not materialise at the assumed scale.

Option C's worst case (£39.8bn NPV, £2.3bn below baseline) is the world where the trigger never fires (only 40% probability assumed) and the departure rate sits at the worst-case 30% throughout, producing most of Option A's worst-case downside without the eventual mechanism correction. This is the world Position B's defenders most fear about Option C: that conditional frameworks fail to fire when the evidence supports them, and the regime drifts in its current form while the cumulative damage accumulates.

What the best cases show

Each option has a meaningful best case. Option A's best case (£44.4bn NPV, +£2.3bn) requires a low (10%) departure rate to be realised under the current regime — which would mean the OBR's 25% figure for non-doms substantially overstates what UK tech founders specifically will do, and the Companies House director-departure data is largely uninformative about the BPR cohort. Option B's best case (£45.6bn NPV, +£3.4bn) requires a very low (5%) departure rate plus the long-stop firing reliably plus the exit-multiplier holding at 1.6 — a coherent but optimistic combination. Option C's best case (£45.1bn NPV, +£3.0bn) requires both a low departure rate and the trigger firing in year 3, capturing the upside of B without the political cost of preemptive mechanism change.

The best-case ordering — Option B (+£3.4bn) best, Option C (+£3.0bn) second, Option A (+£2.3bn) third — mirrors the central-case ordering. Under all scenarios where the cohort is largely retained, Option B and Option C outperform Option A. The differences between them are small relative to the differences between scenarios.

What the model does not resolve

The model does not resolve the question of which empirical world we are in. It cannot. The behavioural-response data needed to discriminate between scenarios does not yet exist. HMRC has not published the modelling that would isolate the BPR-specific elasticity from the wider tax-policy environment. The OBR's 25% figure is for a different population. The Companies House data is contaminated by simultaneous reforms. The named cases are illustrative rather than evidential. A reader who believes the central case is the right central case will conclude, from the model, that Option B outperforms Option A by approximately £2-3bn in 25-year NPV. A reader who believes the worst-case behavioural assumptions are central will conclude that the gap is larger. A reader who believes the best-case assumptions are central will conclude that the differences between options are modest and the publication's case for mechanism change is weaker than it might first appear.

The model also does not capture certain effects worth naming. Tax revenues from non-tech parts of the economy that benefit from a stronger UK tech ecosystem (catering, real estate, professional services, suppliers) are not modelled and would add to the indirect-fiscal figure under retention scenarios. Negative-feedback effects — where a thinning UK ecosystem makes more founders consider leaving — are not modelled and would amplify the downside of the worst-case scenarios. Macroeconomic counterfactuals — what would have happened to UK tech under different reforms or no reform — are not modelled. The model's central finding is robust to these omissions but the absolute numbers under each scenario should be read with that caveat.

The honest summary

The fiscal model produces three findings that are robust to wide variation in the assumptions and that deserve weight in the policy discussion.

First, the direct revenue effect of IHT or CGT on the cohort is fiscally trivial compared to the indirect tax base the cohort generates over its working life. Treasury revenue from the IHT/CGT mechanism, under any option, is in the £40-90m range over 25 years. Treasury revenue from the companies the cohort builds, under any option, is in the £38-46bn range. The mechanism choice matters mostly through its effect on retention, not through its direct revenue effect.

Second, retention is the dominant variable. Under the central assumptions, Option B retains approximately 10 percentage points more of the cohort than Option A; this difference compounds over 25 years through more current-company tax base, more next companies built in the UK, and more cluster effect. The cumulative NPV difference is approximately £2.3bn — small in macroeconomic terms but material as a fiscal contribution from a cohort of 50 founders entering scope per year.

Third, the worst-case asymmetry favours Options B and C. Option A's worst case produces a £3.5bn shortfall against baseline; Option B's worst case is approximately at baseline; Option C's worst case is a £2.3bn shortfall. If a policymaker were to reason from worst-case downside alone — a defensible posture under uncertainty — Option B has the smallest downside, Option C the second smallest, Option A the largest. This is the principal reason the model points toward mechanism change rather than preservation, despite the publication's general posture of declining to recommend.

The model does not tell the policymaker what to do. It tells them where the fiscal stakes actually sit. The headline question — "what does the IHT mechanism cost or save the Treasury, and how does that compare to the indirect tax base from the cohort the country says it wants to grow" — has an answer that the model puts numbers to, where previously it has been argued without numbers. The numbers are uncertain. The structure of the dependency is clear.

If you want to engage with the model directly

Interactive model. The publication has built an in-browser version of the model with every assumption available as a slider. The results recalculate as you move them. You do not need Excel or any setup. Open the interactive model →

The full Excel version is provided as a downloadable spreadsheet alongside this piece. Every assumption is editable in blue text on the Assumptions tab; the four most consequential assumptions are highlighted in yellow. Substituting different assumptions and watching the headline figures move is, in the author's view, the most useful way to engage with the model. A reader who substitutes more pessimistic figures than the central case — higher departure rates, lower current-company tax base, lower next-company probability, no network effect — will see the indirect-fiscal totals fall but the relative ordering of options remain similar. A reader who substitutes more optimistic figures will see the totals rise. The qualitative finding — that indirect dominates direct, and retention dominates mechanism — survives most reasonable substitutions.

The model is one input among several to a serious policy discussion, not a replacement for the formal modelling that HMRC and HMT could in principle conduct. If those institutions have done such modelling, it has not been published. Until it is, models like this one — explicit about every assumption and structured for public engagement — are a more honest basis for the public discussion than the implicit modelling that has driven the debate to date.