THE GENERATIONAL RESET

Open Gaps and Stress Test Register

Version 5.0, 23 October 2026

v5.0 rewrites the register in plain prose and reviews every gap's status against the October 2026 corrections. The main changes: Gaps 1 and 8 are downgraded from CRITICAL to HIGH, because Economic Renewal no longer borrows ahead of inheritance tax revenue (see U182 in the Update Register); Gap 26 is partly resolved by the Reform Sequence Roadmap; Gap 18 is partly resolved by new evidence in Political Renewal; and several figures repeated here from the pillars have been brought into line with the corrected pillars (Gaps 2, 14, 20, 21, 29 and 31). Earlier versions: v4.1 added Section G (Gaps 27–28) on the foundation documents; v4.2 added Section H (Gap 29) on the spending table.

This register lists unresolved questions, known weaknesses and areas that need stronger evidence before the arguments in the Generational Reset can be considered robust. It is published alongside the main documents as a core part of the project, not as a footnote.

A project that hides its weaknesses can't be trusted. One that publishes them, and commits to closing them through evidence rather than assertion, is making a different kind of claim.

Gaps are closed by research and by design, not by argument. When a gap is answered with solid evidence or a workable design, it is marked resolved. Until then it stays open.

Version 4.0 widened the register from the economic model (the inheritance tax and income tax reform) to every pillar then published. Gaps 1–10 were carried forward; Gaps 11–24 came from a review of each pillar. The register now has eight sections (A–H) and 35 gaps, of which 6 are fully resolved and 7 partly resolved.

Section A: The Economic Model

Gaps 1–10 and 33–35 cover the core Economic Renewal proposal: a 95% inheritance tax at death, paying for a substantial income tax cut. Until 14 October 2026 the proposal was to abolish income tax altogether; several gaps below were written against that version and have been updated.

Income tax and NI together raise £475 billion a year (2024/25), rising to roughly £600 billion by 2030/31. Total UK household wealth is £10.8 trillion. Wealth passing between generations, at death and in lifetime gifts, is roughly £100–115 billion a year (Cebr for Kings Court Trust: £115bn in 2027; the Resolution Foundation, from HMRC data: £87bn of estates in 2014–15, expected to double over two decades). This register used to give £200–400 billion, a figure no source supports; it was corrected on 14 October 2026. At 95% on every pound, the theoretical maximum yield is roughly £95–110 billion, about a fifth of income tax and NI. Because both grow in cash terms, that share doesn't change much at the peak of the baby boomer transfer.

Economic Renewal therefore now proposes an income tax cut sized to proven inheritance tax revenue, not abolition. That removes the financing problem this gap was first written about: there is no gap to bridge, because no cut runs ahead of the revenue. This is why the gap is downgraded from CRITICAL to HIGH.

What stays open is the size and timing of the revenue itself. The proposal's income tax cut is only as large as the transfer turns out to be, and the transfer path is projected, not modelled. Earlier versions also considered transition bonds and a sovereign wealth fund as ways to finance the gap; those now have narrower roles (bringing forward cuts that proven revenue will pay for, and smoothing year-to-year swings in receipts).

Remaining requirements:

• Model the baby boomer death curve and the projected wealth transfer year by year, 2025–2060 (see Gap 35 for asset composition)
• Cost the revenue path under optimistic, central and pessimistic assumptions about behaviour and avoidance
• Set out what size of income tax cut each scenario would support, and at what pace

At 95%, contesting valuations becomes the most valuable thing any significant estate can do. Under today's 40% rate, the effective rate on estates over £30 million is about 12% once Business Property Relief is applied (CenTax, before the April 2026 reforms), and Agricultural Property Relief, spousal exemptions, lifetime gifts and pension wealth widen the gap further. Since April 2026, 100% business and agricultural relief has been capped at £2.5 million per person, with 50% relief above that. At 95%, the avoidance industry would be larger than anything that exists now.

Asset classes needing their own frameworks: private companies (valuation methods give very different answers); intellectual property (patents, music catalogues, brands); crypto and digital assets (designed to resist oversight); art and collectibles (thin markets, easy to move); and pension wealth (around a third of UK household wealth; the October 2024 decision to bring pensions into inheritance tax from April 2027 is a precedent, though how it will work is still contested).

Design responses now set out in Economic Renewal §4.3, but not yet worked through in detail:

• Safe harbour valuations: HMRC publishes standard multiples, and contesting them triggers a full audit and litigation costs
• Independent valuation panels with no financial relationship to the estate
• An equity stake mechanism: the state takes a share of ownership instead of cash for illiquid assets
• Staged payment over 10–20 years for assets that really are illiquid

A £2 million house is a dynasty in Newcastle and a family home in Kensington. Where the threshold sits isn't a technical question; it is where the political coalition holds or breaks. Questions to resolve: treatment of the main home; pension wealth; indexing the threshold; regional variation; whether the tax is per estate or per beneficiary; and whether Stage 1 keeps the spousal exemption, which changes Stage 1 revenue by about 30% (Revenue Architecture §8.1). The model has to feel clearly aimed at dynastic accumulation, not at the family that built a small business over forty years.

If private money is removed from elections, is it also removed from referendum campaigns? If not, the spending mechanism has the same capture problem as the political system it replaces. Specific risks: dark money in referendum campaigns; algorithmic amplification by wealthy interests; the advantage of those who can make complex issues seem simple; and manipulation of how citizens' assemblies are framed. Proposed mitigations include applying full campaign finance rules to referendums and publicly funding information campaigns for both sides of every question.

Closing the offshore routes depends on enough international cooperation, and the world is fragmenting rather than cooperating. Questions: what happens if Singapore, Dubai, Switzerland or the Cayman Islands decline to take part? What is the smallest group of countries that makes the model work? How are dual nationals and people who give up citizenship handled? There are encouraging precedents (the OECD minimum corporate tax in 2021, FATCA, the Common Reporting Standard), and the residence-triggered exit tax now proposed in Tax Avoidance and Economic Renewal §5.3 covers people who leave. But only physical assets can be fully captured without global cooperation.

Spending down by wealthy people is presented as a feature. But a 95% inheritance tax has never been tried at scale. Sweden's (a 70% top rate from the late 1950s, cut to 30% in 1992 and abolished in 2004) is instructive, but Sweden was a small, homogeneous economy. Responses that need testing: status consumption (superyachts and trophy property, which tie up capital without much economic activity); aggressive lifetime gifting; shorter investment horizons once dynasty-building is removed; entrepreneurs emigrating; and reduced appetite for risk. Evidence to gather: Swedish data from the 1970s to 2004; high-inheritance-tax countries (Belgium, France, Japan); Norway's departures after its 2022 changes; and survey research on whether dynastic ambition actually drives wealth creation.

Sharper framing, added 12 October 2026: the question isn't only whether people respond unpredictably, but whether there is a tipping point: a rate or level of valuation pressure past which spending down and emigration stop being a manageable tail and become the main response of the whole wealthy population at once. The evidence-gathering above should look for any threshold or break in the Swedish and comparator data, not just an average response.

Political Renewal comes before everything else, but the people who would have to pass it are those whose positions depend on the current system. Questions: which political circumstances make a first move most likely? Is full proportional representation necessary, or are there intermediate steps? How does a reform movement avoid being captured? How can it be ready when a crisis opens a window, without simply waiting for one?

You can't prove inheritance tax revenue will arrive without bringing in the tax, and you can't responsibly bring in the full tax without that proof. Four ways out were assessed: (A) a crisis forces the issue, for example rapid, visible unemployment from AI; (B) two-stage implementation, starting at 60–70% on large estates and building revenue data and administrative systems over 5–7 years before making the case for the full rate; (C) a political moment that tolerates open discussion of transition costs, as after the war; (D) a pilot in Scotland, Wales or another devolved area.

Economic Renewal has adopted (B), and Stage 1's revenue is now estimated at roughly £10–15bn a year from HMRC data. Since the income tax cut now follows revenue rather than borrowing against it, the original problem (borrowing against revenue nobody can yet prove) no longer applies, and the gap is downgraded from CRITICAL to HIGH. What remains is that nothing yet says what evidence from Stage 1 should trigger the move to Stage 2, as opposed to a political decision to go ahead.

Partly resolved 23 October 2026. Economic Renewal §4.5 now sets five tests that must all hold before Stage 2: the effective rate on estates above £10 million within 5 points of the legislated rate; revenue at least at the bottom of the Transition Financing Plan's range for three years; avoidance no higher than forecast; enforcement working (the authority staffed to plan, disputes settled within two years, penalties back to at least 14% of closed cases); and the rate protection and published estate data in place. The OBR and the National Audit Office would assess them and report to Parliament each year, on the Climate Change Act model, because the Treasury assessing its own tests (as with the five euro tests in 1997 and 2003) can't be separated from the politics. The section also explains why 5–7 years is roughly how long the evidence takes to exist, given HMRC's two-year lag in publishing inheritance tax figures. The tests allow the move but can't force it, since Parliament can't be bound (Gap 9).

• Test the thresholds (5 points, three years, 14%) against Stage 1 revenue modelling once Gap 1 is done
• Decide what happens if some tests pass and others fail

Winning the argument and the election isn't the same as making lasting change. Implementation can be captured by the same interests that lost at the ballot box. Roosevelt's Revenue Act of 1935, described at the time as soaking the rich, is the cautionary tale: it passed into law but produced symbolic rather than real redistribution within a decade. At 95%, every pound sheltered saves 95p, so the reward for capturing implementation is larger than for almost any tax in history.

Design responses:

• Constitutional rate protection that needs a supermajority to change
• An independent inheritance tax authority modelled on an independent central bank, with its board appointed through a citizens' assembly
• Real-time publication of every estate, assessment and relief claimed, within the limits set out in Economic Renewal §4.4 (see Gap 34)
• Automatic escalation: if the effective rate falls more than a set margin below the legislated rate, a legislative review is triggered
• A movement that stays organised through implementation instead of disbanding after winning

Historical analysis added 23 October 2026. Economic Renewal §4.4 now sets out how four strong inheritance taxes were diluted (UK estate duty, capital transfer tax, the US estate tax and Sweden's) and matches each route of erosion to a defence in the design. Two findings matter. First, most of the routes (early gifts, trusts, favoured-asset reliefs) are answered by design features already proposed, with enforcement capacity (Gap 20) the main weakness. Second, the defence against outright legislative cuts, constitutional rate protection, is weaker than the pillar implied: under UK parliamentary sovereignty a statutory supermajority or referendum lock can be removed by an ordinary Act, as happened to the Fixed-term Parliaments Act 2011 (bypassed in 2019, repealed in 2022) and the European Union Act 2011's referendum lock (repealed in 2018). The gap stays CRITICAL because the main protection against the route that ended Sweden's and the US's taxes is now known to be a political barrier, not a legal one.

• Set out the strongest procedural protection available in UK law (for example a statutory referendum requirement for any rate cut, or a supermajority rule), and say plainly what it can and can't guarantee
• Assess whether the independent authority and real-time publication make a cut politically costly enough to deter it, using the Fixed-term Parliaments Act and the EU referendum lock as test cases

Sharper framing, added 12 October 2026: capture isn't only political and administrative erosion. At 95%, the design responses themselves (equity stakes, staged payment, valuation panels) change how entrepreneurs and family businesses behave, in ways that could hold back new business formation and investment, an economy-wide effect distinct from simple avoidance. Any study of how such taxes were diluted in the past should include this channel, not only avoidance and reliefs.

Three questions, all resolved. Land value tax is a live proposal in its own right: Revenue Architecture §3 proposes it as a standalone Key Proposal, projected to raise £30–50bn a year for other spending priorities, on a separate revenue track from the inheritance tax so the two aren't double-counted. This entry used to say land value tax had been "resolved against", which was wrong; corrected 12 October 2026. How the two taxes interact on the same land is still unmodelled, and is part of Gap 25. Distinguishing "earned" from "unearned" asset classes was resolved against: applying the tax to everything is both more principled and simpler to run. Spending down by the wealthy is treated as a feature, not avoidance, because spending puts capital back into use. Systematic lifetime gifting is handled through gift thresholds and look-back rules.

The objection that inherited wealth was already taxed once, as income, so taxing it at 95% at death taxes the same money twice, is one of the most persuasive available to an opponent, because it sounds like arithmetic rather than a values argument. Until October 2026 the project hadn't answered it directly.

Resolved 23 October 2026. Economic Renewal §7 now has a counter-argument answering it. Much of a large estate was never taxed at all: capital gains tax isn't charged at death, and heirs take assets at their value on the date of death, so a lifetime's gains are wiped clean, and some assets currently escape both capital gains tax and inheritance tax (which the Office of Tax Simplification recommended changing in 2020). Where money was taxed as income, the 95% falls on a different event and a different person, the heir, who pays no income tax on an inheritance. Taxing the same money more than once is normal (income tax, then VAT, stamp duty and council tax), and double taxation in the strict sense, the same profit in the same hands, is a different thing. The reply concedes that the objection still has force for estates built from fully taxed salary, and answers that case with the value judgement in §3 rather than with arithmetic.

Gap 9's response to implementation capture is real-time publication of every estate, assessment and relief claimed, to make any erosion of the 95% rate visible and costly. Until October 2026 the project never weighed the cost: publishing a named estate exposes the heirs' privacy, creates a plausible security risk at the top of the wealth scale, and can expose a family business's finances to competitors.

Resolved 23 October 2026, in two steps. First, Economic Renewal §4.4 now states the cost, who bears it and why the project judges it worth paying, and limits the design: each estate's total valuation, reliefs, contested valuations and tax are published by name, linked to the grant of probate, which is already public; itemised holdings, addresses, where assets are kept, account details and heirs' shares are not; lookups need a verified login and are logged, as in Norway; and aggregates are published by estate size. Second, the page now sets out whether this is lawful. Data about the dead is outside UK data protection law, but anything identifying living relatives or co-owners isn't, so publication needs an explicit statutory basis and must be proportionate; HMRC's statutory confidentiality means the legislation must authorise each published field; and the EU court's 2022 ruling against unrestricted public access to ownership registers shows why the limits matter. Two safeguards were added: company holdings published as the estate's own stake, and a right for living relatives to withhold specific details where there is a real security risk.

• Remaining for whoever drafts legislation: a full legal review of the published fields against the UK GDPR, the Data Protection Act 2018 and Article 8 ECHR

Gap 1's modelling of the death curve covers timing (when people die and how much passes on each year) but not composition: what the cohorts dying in the coming decades will actually own, as distinct from today's £10.8tn snapshot. Pension, housing and business wealth differ in liquidity, in how hard they are to value and in how exposed they are to avoidance (see Gap 2). A cohort whose wealth is mostly in private business equity poses a very different challenge from one whose wealth is mostly in pensions or housing. Today's snapshot isn't a safe guide to the wealth of the cohorts who will die during Stages 1 and 2, given decades of change in pension design, home ownership and self-employment.

• Extend the death-curve modelling in Gap 1 to cover asset composition by cohort, not just total value and timing
• Check whether ONS Wealth and Assets Survey data can be broken down by birth cohort and asset class, for an early read before commissioning new modelling

Section B: Public Spending Pillars (S1)

Gaps 11–17 cover the original Section One spending pillars. These pillars mainly describe what Britain spends and what it gets for it, so the gaps here are mostly about data quality, attribution and the limits of international comparison.

Partly resolved in August 2026. The cost of delayed discharge was updated to the King's Fund's current £2.7bn (2025/26, up 7.5% on the year), and the pillar now uses the King's Fund's three-part explanation (NHS community services, social care capacity and administrative bottlenecks), not social care alone. For delayed discharge specifically there is now quasi-causal evidence, and it is cited: a panel study of English local authorities (2013/14–2018/19) linking social care cuts to delayed-discharge days, with controls, and a study finding that each extra home care provider per 10km² reduced delayed transfers of care by 14.9%.

The broader claim, that social care failure is a main driver of the wider NHS crisis (waiting lists and admissions in general), doesn't meet the same standard, and the pillar now says so. A peer-reviewed study of 132 English councils (2005–2016) found no statistically significant link between social care cuts and emergency admissions. The project cites that null finding rather than leaving it out. What remains open: no natural experiment or trial has yet tested the specific reforms the pillar proposes (the Dilnot cap and changes to means-test thresholds).

• If such a study is published, use it. Until then, the pillar's proposals should be read as consistent with the evidence, not proven by it

The Education pillar reports that 47% of school leaders use Pupil Premium money to plug general budget gaps, up from 23% in 2019. The gap is in enforcement. The proposal to ringfence the Pupil Premium, with mandatory spending categories and a separate Ofsted measure, doesn't yet say how misuse would be detected, what would happen when it was, or whether those consequences would change behaviour given the financial pressure driving the misuse.

The deeper problem is that schools redirect the money because they have no other way to balance their budgets. Ringfencing without easing that pressure may simply add compliance costs without helping disadvantaged pupils. The pillar touches on this in its cross-pillar dependencies, but the enforcement design needs more work.

• Design how misuse would be detected and what the consequences would be, and how this interacts with schools' financial pressure
• Assess whether ringfencing alone would improve outcomes without better general school funding

The Welfare pillar is careful about disability benefits and argues that reform means treating causes, not cutting support. The gap is that the evidence on how many current claimants have untreated capacity to work, which the proposed interventions (clearing NHS waiting lists, employer return-to-work duties, community mental health services) would unlock, is weaker than the pillar implies. The Dutch and Norwegian comparisons are real: both countries have higher employment among disabled people and better-resourced support. But how much of that comes from the support itself, rather than the health system or the labour market, is less certain than the pillar suggests.

• Find the best natural-experiment or quasi-experimental evidence on the employment effects of the specific interventions proposed
• Estimate the share of current PIP and Universal Credit health-element claimants whose employment the proposed interventions could plausibly change

Resolved in August 2026. The pillar's one-line reference to Germany and Singapore was replaced with real institutional detail. Germany's municipal housing companies (for example Berlin's HOWOGE) now carry the design weight: a two-board governance structure, a negotiated agreement setting social and growth targets, and financial independence through borrowing on its own credit rating (S&P A long-term and A-1 short-term; corrected 13 October 2026 from "AA1") rather than relying on central grants. Singapore's HDB is kept only as evidence that state delivery can work at scale, and is flagged as not transferable given its unusual preconditions (near-total state land ownership, leasehold-only tenure, financing linked to compulsory savings). An earlier assumption was also corrected: UK land acquisition powers aren't obviously weaker than Germany's, since the 2023 Levelling-up and Regeneration Act's removal of "hope value" from compulsory purchase compensation arguably goes further than Germany's narrower pre-emption rights. The real gap is the delivery vehicle and its financing, not legal powers over land.

On the UK track record, HS2, Hinkley Point C and Universal Credit are each addressed, and their failures found less relevant to housing than often assumed (cost escalation on a rail corridor, first-of-a-kind nuclear technology, bespoke digital delivery). The most relevant precedent, previously missing, has been added: Croydon's Brick by Brick, a council housing company whose collapse (governance capture, non-standard delivery) is the direct warning for the Corporation's design. A positive precedent is also cited: Homes England's programmes delivered 40,332 completions in 2025-26. The pillar is clear that building directly is a further capability a Corporation would still have to show it has.

The Criminal Justice pillar treats the fall in the charge rate, from 11.1% of victim-based crime in 2015–16 to 4.6% in 2021–22, recovering to 7.3% in 2025–26, as the main measure of system failure. The causes are only partly established. The pillar identifies CPS underfunding (£1.1 billion against £19.5 billion for policing) as a main driver. How much of the fall reflects (a) limited policing and prosecution resources, (b) a shift towards offences that take more resources to investigate, such as fraud and sexual offences, (c) changes in reporting by victims, and (d) changes in evidential standards hasn't been separated out.

This matters for the reforms. If the fall is mainly about resources, investing in the CPS and in police investigation should restore the charge rate. If it is mainly about the mix of crime, a different response is needed. The pillar leans towards the resource explanation, but the question deserves more rigorous treatment.

• Find research that separates the drivers of the fall by offence type, period and police force area

Since 2023 all nuclear spending has been grouped under the Defence Nuclear Enterprise heading, so Parliament and the public can no longer easily see what the deterrent costs. The pillar's estimate that it takes about 18% of the defence budget is reasonable but uncertain. This is a problem for democratic accountability and for the pillar's ability to make specific, costed proposals. The gap is political rather than analytical: the data exists but isn't published in a form that can be checked independently.

• Submit FOI requests or find parliamentary written answers with more detailed nuclear cost data
• Engage with RUSI, IISS or similar institutions that may have better cost modelling

Resolved in August 2026, then corrected again on 13 October 2026 because the August fix had introduced a new error. The pillar originally said debt would reach 275–325% of GDP by the 2070s, which was close to right. The August fix replaced this with "about 200%", wrongly believing it to be the OBR's revised central case. The OBR's own Table 5.1 (Fiscal risks and sustainability, July 2026), read directly, gives 300.4% of GDP in 2075-76. The pillar now cites that figure from the primary table. The sensitivity scenarios were corrected from the same data: the higher-productivity case reaches about 182% by 2075-76 (not about 80%), and the OBR's conclusion that even this eventually becomes unsustainable, about twenty years later, still stands; the lower-productivity case reaches about 706%. The OBR's summary is quoted accurately: "almost all of these scenarios suggest that the public finances will at some point move onto an unsustainable path." A separate, shorter-range OBR finding (a swing of about £90bn in 2030-31 borrowing between optimistic and pessimistic AI scenarios) is included and kept distinct from the 50-year projection. No migration-specific sensitivity is included, because the project hasn't found one.

Section C: Political Structure Pillars (S2)

Gaps 18–19 cover Political Renewal and the Public Office Covenant.

The Political Renewal pillar makes a strong case for proportional representation, citing Germany and Scandinavia. The counter-evidence is also real: proportional systems can produce weaker governments, longer coalition talks and more difficulty passing reform in a crisis.

Partly resolved in September 2026. The pillar now engages much more of that counter-evidence. It leans on the UK's own proportional systems (the Scottish Parliament and London Assembly since 1999–2000, the Northern Ireland Assembly since 1998, Scottish local elections since 2007), which operate under the same legal and civil-service culture a Westminster reform would inherit, and which neither Israel nor Italy can offer as a comparison. It says that the Senedd's fully proportional system, first used in May 2026, is too new to count as evidence either way. It treats the Netherlands as a counter-case rather than a success story (a very low threshold and coalition talks of 223 to 299 days). And it answers, in its Steel Man, the case for ranked-choice voting in single-member seats as an alternative to full proportional representation, conceding that STV meets part of that objection.

What remains open is the broader empirical claim that proportional systems deliver more far-reaching reform than first-past-the-post. Some evidence supports it (welfare states, pension reform, inequality outcomes in Scandinavia and Germany), but the political science literature is divided.

• Engage with the most rigorous comparative literature on reform delivery under proportional and majoritarian systems, including cases where proportional systems blocked reform

The Covenant's strongest argument is self-selection: rules known in advance attract public servants and deter people with something to hide. That is intuitively powerful, but the evidence that strict transparency attracts better candidates, rather than deterring everyone with complicated finances, is limited.

The available comparisons are partial. Many OECD democracies require some financial disclosure, but none as complete as this. The US STOCK Act, for example, requires disclosure but hasn't obviously improved financial decision-making in Congress. The argument rests more on reasoning than on evidence. That may be unavoidable for a new proposal, but it should be said openly.

• Survey the evidence on disclosure requirements and candidate quality in comparable systems
• State the limits of the self-selection argument more openly in the pillar and here

Section D: Tax and Economy Pillars (S3)

Gaps 20–22 cover tax mechanics, tax avoidance and the wider economy.

The Tax Avoidance pillar shows the gap between the headline and effective inheritance tax rates at 40%: the largest estates (over £30m) paid about 12% before the April 2026 relief reforms. At 95% that gap would grow enormously. Neither Tax Avoidance nor Economic Renewal has yet fully addressed whether HMRC could cope. Its wealthy individuals unit already struggles with the complexity it faces at 40%. A 95% rate, with a far larger incentive to avoid and far more sophisticated structures, would need a transformation in HMRC's capacity, not just more staff.

The question isn't whether HMRC could in principle be resourced for this. It is whether the culture, skills, IT systems and protection from political interference could realistically be built within the timetable.

Evidence added 23 October 2026. Economic Renewal §5.4 now sets out HMRC's current capacity from the National Audit Office's May 2025 report: about 910 staff for 850,000 wealthy individuals, a compliance yield of £5.2 billion for about £350 million of cost, and roughly 4,000 inheritance tax investigations a year. It also sets out three warning signs: the wealthy tax gap is probably larger than HMRC estimates, penalties fell from 14% to 5% of closed cases between 2018-19 and 2023-24, and there is no full strategy for the very wealthy. It lists what the independent authority would need before Stage One (a dedicated unit for the largest estates, penalty targets, funding set by its own tax gap assessment, published data). The gap stays CRITICAL because nobody has costed that authority or shown the specialist valuers and litigators can be recruited in time.

• Cost the independent authority's staffing and budget for Stage One (60–70%) and the full rate
• Review how high-capacity tax authorities elsewhere are designed (IRS Criminal Investigation, France's DNEF) and what they spend per case on the largest estates

Resolved in August 2026. The pillar now names the specific failure behind each UK precedent instead of blaming "institutional inadequacy" in general. Green Investment Bank: the "special share" kept after privatisation protected only the bank's charter purpose, not its investment decisions, so it constrained nothing in practice (the bank was sold in 2017). British Business Bank: capitalised at roughly 1% of KfW's scale. KfW's borrowing doesn't count towards Germany's public debt because the national accounts classify it as a financial institution rather than part of government (corrected 20 October 2026 from a reference to a legal exemption), so any UK plan to match KfW's scale has to deal with how its borrowing would be counted. UK Infrastructure Bank: launched in 2021 with £22bn to deploy, it committed about £1bn across ten deals in roughly three years, with no performance measure beyond a financial return. The pillar also fixed a live error: it had proposed reforming the UK Infrastructure Bank, which was absorbed into the National Wealth Fund in October 2024. The National Wealth Fund deploys capital faster (£3.6bn in its first year) but, according to the Treasury Committee, still can't show that its investment is additional any better than its predecessor could.

The tax pillars argue that AI is putting pressure on the income tax base, because it can replace the work that pays most income tax. This is a central reason for moving towards taxing assets and inheritance. The direction is plausible, but the pace and scale are uncertain, and income tax receipts are still rising in cash terms, helped by frozen thresholds.

Estimates of AI-driven displacement vary enormously by method and timeframe. McKinsey's 2017 upper estimate of 800 million workers displaced worldwide by 2030 is one end; the IMF estimates that about 60% of jobs in advanced economies are exposed, with roughly half likely to benefit rather than lose. The pace matters: fast erosion strengthens the case for urgency, while slow erosion weakens it and makes sustained political will harder.

• Set out the range of credible estimates for AI's effect on the income tax base, with timeframes and confidence ranges
• Model what fast, central and slow scenarios would mean for the timing and size of the income tax cut

Sharper framing, added 12 October 2026: the uncertainty also cuts in a particular direction for this project's urgency argument. AI could raise productivity and wages in ways that grow the income tax base for a time before any net erosion appears. The modelling above should trace that chain explicitly rather than assuming that AI leads straight to erosion.

Section E: Context Pillars (S4)

Gaps 23–24 cover Energy and Immigration.

Partly resolved in August 2026. The pillar now states, citing NESO directly, that GB balancing costs were £2.7bn in 2024/25 and £3.1bn in 2025/26, and are forecast to reach £6.4–8.3bn by 2030 without faster transmission investment, with transmission constraints (curtailing wind and replacing it with gas) making up 71% of the total and rising. That closes the immediate gap: the generation cost per kWh is now paired with the current cost of actually delivering that electricity on today's grid.

The harder question remains: the cost of integrating the whole system at very high (80–100%) renewable shares in the 2030s and beyond. NESO's Clean Power 2030 Annex 4 is reported to find that overall system costs "should not increase" on its central pathway, but this project couldn't check that against the primary document, only secondary summaries. The pillar shouldn't cite a specific cost at high renewable shares until the primary source has been read.

• Read NESO's Clean Power 2030 Annex 4 directly and verify or correct the "system costs should not increase" claim with an actual figure
• Track whether the £6.4–8.3bn 2030 balancing-cost forecast falls as new transmission capacity comes online, as NESO expects

The Immigration pillar is open that social cohesion is where quantitative evidence is thinnest and hardest to interpret. It cites Robert Putnam's research (ethnic diversity associated with lower social trust in the short to medium term, fading over longer periods) and notes that the UK has had high immigration without treating integration as a policy priority in its own right.

The gap is that its integration proposals (early-years English as an additional language, language support, investment in integration) are supported by evidence of effects at the margin but not costed as a programme. The pillar says integration needs active investment without saying what that would look like at scale, what it would cost or what results to expect. This is the most under-specified proposal in the pillar.

• Develop a costed integration programme (early-years language provision, language support, community infrastructure), drawing on comparable programmes in Germany, the Netherlands and Canada
• Engage with Migration Advisory Committee and Migration Observatory research on which integration measures work

Section F: Cross-Pillar Gaps

Gaps that cut across several pillars and can't be resolved within any one document.

Every pillar names dependencies on others: the NHS depends on housing stability, housing reform on adequate welfare, education outcomes on child poverty, criminal justice on mental health provision. These are correctly identified, but none is quantified in a way that would let the whole programme be costed and sequenced.

So each pillar makes credible proposals, but the interactions, where reform in one area saves or costs money in another, haven't been modelled. A serious child poverty programme might reduce emergency admissions, demand on criminal justice and special educational needs provision all at once. Leaving these out probably undervalues the programme as a whole.

• Commission or build a model of the fiscal and outcome interactions between the main proposals
• At minimum, quantify the three or four most valuable interactions (for example social care and the NHS; early years, the attainment gap and lifetime earnings) with the evidence available

Two interactions added 12 October 2026, following the Gap 10 correction:

• Land value tax and the inheritance tax. Land value tax is a live proposal (Revenue Architecture §3), so its interaction with the 95% inheritance tax is a real, unmodelled risk. An annual tax on land value during life and a 95% tax on all wealth, including land, at death both fall on the same asset. Whether together they change how land is held (speeding up sales, or distorting the split between use value and speculative value that land value tax depends on) hasn't been examined in either pillar.
• Asset prices moving in anticipation. Once a 95% inheritance tax is credible and expected, it would change asset prices and holding behaviour before it applies: selling in advance, revaluations and capital moving elsewhere. Neither Economic Renewal nor Revenue Architecture's Numbers Ledger models this effect on the full-rate revenue ceiling (about £95–110bn at today's values), which assumes wealth keeps growing at its recent rate regardless. Revenue Architecture §8.2 Assumption 3 already flags both this and the land value tax interaction as open, but doesn't model either.

The project argues convincingly that Political Renewal must come first, and that argument is accepted. What it lacked was a sequence for the reforms that follow: which spending reforms come first, which economic reforms depend on others, and how the move from the current system to the proposed one would be ordered in practice. Without that, critics could say the programme is a list of good things without a theory of delivery.

Partly resolved in October 2026. The Reform Sequence Roadmap working paper sets out a three-phase sequence (Foundation, Build-Out, Transition), using each reform's own timeline instead of forcing one scale on all of them. A first set of proposal-level dependencies, each traced to a sentence in a pillar, is published as structured data alongside the proposals. What remains open: the roadmap is a working paper, held to the lighter standard of evidence, and the ordering it proposes isn't costed or tested. Most cross-pillar relationships describe connections, not a required order, and only a handful have been classified.

• Extend the dependency data beyond the Economic Renewal proposals and the few stated hard dependencies
• Cost the first phase of the roadmap and test whether its ordering holds once Gap 25's interactions are modelled

Two pillars propose different fixes for the same problem, and the project has deliberately left them unreconciled rather than picking one. Welfare keeps the State Pension's funding priority where it is and reforms only how it is uprated, replacing the triple lock with a smoothed earnings link. Fiscal Hierarchy, in the Options stage of its argument chain (§1.2), goes further and moves the State Pension down to a lower, conditional Tier Six, below children, NHS care, working-age welfare and education.

Both are evidenced, both are live proposals, and neither is treated as more authoritative. This isn't an oversight: the two pillars were cross-referenced to each other on 8 October 2026. But it is a fair question for a reader to ask: does the project think the pension's claim on public spending should change, or only how it is calculated?

• Decide, or decide to leave this open with a stated reason. Options: one pillar takes precedence; a hybrid is developed; or the question is settled later by evidence, using Fiscal Hierarchy's own falsification test (whether the Tier Six re-ranking "proves cosmetic")

Section G: Foundation Document Gaps

Gaps 27–28 cover the foundation documents that come before the pillars, What It Means to Be British and Britain by the Numbers. They don't make reform proposals, but their claims have to meet the same standard of evidence as everything else.

In important places, the identity essay relies on data for England and Wales rather than the whole UK. The social mobility figures (the OECD's estimate of about five generations for the poorest families to reach average income; corrected 14 October 2026 from "second lowest in the OECD", which the OECD data doesn't support), the class data, the data on religion and the census figures mostly come from ONS data for England and Wales. Scotland and Northern Ireland have different stories on each of these.

Scotland has its own educational tradition, a class culture less defined by accent, and a politics in which Scottish and British identity are in more open tension than anywhere in England. Northern Ireland's identity question isn't a variant of the English one: it has a different history, a different population and a constitutional status that is still contested in ways with no parallel elsewhere in the UK.

The essay recognises that the Union is recent and contingent, and treats Northern Ireland's position as precarious. But it doesn't fully bring Scottish and Northern Irish experience into its account of British identity. It mainly describes an English experience of Britishness and generalises from it.

• Find comparative research on how people in Scotland and Northern Ireland describe their relationship to British identity, and whether the essay's core arguments (absorption, fairness, scepticism of power) hold across all four nations
• Add a section stating the limits of the essay's mainly English frame and what a four-nation account would need

Section 06 of the identity essay surveys how Britain is seen abroad: in the United States, Europe, South Asia, Africa and the Caribbean, and the older Commonwealth. It is rich and frank about the gap between how Britain sees itself and how others see it, but it relies more on qualitative judgement than any other part of the essay.

Its descriptions of how American policy elites, European publics and people in former colonies see Britain are plausible and often backed by specific examples (the Liz Truss episode, Utsa Patnaik's estimate of wealth drained from India, the reparations debate). But they aren't backed by cross-national polling (Pew Global Attitudes, Ipsos Global Advisor, YouGov International) that would let them be tested rather than asserted. The essay doesn't say explicitly that these are impressions rather than systematic findings, so a reader could take them as more firmly established than they are.

• Find and cite the best cross-national polling on how Britain is seen in the United States, the main EU countries, India, Nigeria and Australia
• Add a note to Section 06 saying the survey is qualitative, and showing where polling supports or complicates its descriptions

Section H: Data Integrity Gaps (added 30 July 2026)

Resolved in August 2026. The cause was double counting. Local government and devolved government spending were listed as if they were two more functional categories alongside Health, Education and the rest. But the Treasury's functional tables (what the money is for) already include the NHS, education and social care spending that councils and devolved governments deliver; the institutional figures (who spends it) are the same money seen from a different angle, not extra money. The functional rows add up to £1,171bn, £112bn short of the £1,283bn headline, which confirms double counting rather than an under-count.

The table was corrected by replacing the two institutional rows with a single "Local and devolved services not captured above" row of £112bn, which brings the total to £1,283bn, about 44.4% of GDP. This £112bn is local-only spending (roads, waste, libraries, local administration) with no UK-wide functional line of its own. Its exact make-up hasn't yet been itemised against the Treasury's institutional tables, so the mechanism and total are resolved but the breakdown of the £112bn isn't.

Two later corrections to the institutional figures don't reopen this gap, because neither figure was ever added into the £1,283bn: the Barnett block grant is £83bn, not £95bn (13 October 2026), and the Local Government pillar's headline is now PESA 2025's £239bn total local government expenditure, replacing a £98bn figure that matched no Treasury table (14 October 2026).

• Itemise the £112bn residual from PESA 2025 Tables 7.4–7.8 instead of leaving it as a balancing figure
• The table also feeds the budget tool on the Me page, which takes its totals from it, so any correction carries through automatically

The Welfare pillar's breakdown of the £313bn benefit bill shows that Jobseeker's Allowance, the narrow benefit for people who are unemployed, able to work and looking, costs £0.31bn for 68,000 claimants: a rounding error against the popular image of the "workless claimant". What would sharpen this is DWP's breakdown of the much larger Universal Credit caseload by conditionality regime: how many claimants must search for full-time work with no exemption, how many are exempt because of ill health, disability or caring, and how many are already in work and receiving a top-up. DWP publishes this only through Stat-Xplore, an interactive tool with no static export this project could use; nine attempts through gov.uk, the Commons Library and independent analysts failed. The £7.3bn Universal Credit health element (paid to claimants assessed as unable to work) is the best available proxy, and points the same way as the Jobseeker's Allowance figure, but it isn't the same statistic.

• Retrieve the "People on Universal Credit" table from Stat-Xplore by conditionality regime, to give the share actively looking for work directly

The same breakdown established the legal starting point for the claim that immigrants drive the welfare bill. No Recourse to Public Funds is the default condition on most temporary visas (work, study and family visas in particular), and it is a legal bar, not a guideline. It blocks Universal Credit, Pension Credit, PIP, Attendance Allowance, DLA, Carer's Allowance, Housing Benefit, Income Support, income-based ESA and homelessness assistance, with Child Benefit as a named exception. What the project couldn't establish was the picture by claimant: the benefit caseload by nationality or country of birth, and how the foreign-born share of claimants compares with their share of the population and workforce.

Partly resolved in September 2026. DWP published Universal Credit claimants by nationality for the first time in July 2025, covering June 2025: about 1.3 million non-UK and non-Irish nationals, 16.4% of the 7.9 million caseload, all of whom had passed the Habitual Residence Test. This is cited in Welfare §1.1 with its definition and its limits: it covers Universal Credit only, roughly half of working-age welfare spending, and can't be applied to the £313bn total. What remains open: DWP hasn't published a nationality breakdown for PIP, Housing Benefit, Child Benefit or legacy tax credits, so no whole-bill figure exists; and the academic comparisons of claim rates for migrants and UK-born people haven't been freshly checked.

• Find DWP or ONS nationality data for PIP, Housing Benefit, Child Benefit and legacy tax credits, and the most recent Migration Advisory Committee or academic comparison of claim rates, before publishing any claim-rate comparison beyond Universal Credit

Summary Status Table

Every gap by severity, with its current status. Gap 10, fully resolved and described in Section A, isn't repeated here.

The Generational Reset: Open Gaps Register v5.0

generationalreset.org | Updated 23 October 2026 | Published under open licence for public discussion and adaptation

Gaps are closed by research and design, not by argument.

The Generational Reset | Open Gaps Register v5.0 | For public discussion. Not affiliated with any political party. | generationalreset.org