The Civic Capital Compact: Securing Agency Under Unit Cost Dominance

Doctrine in one sentence

The state should stop trying to preserve the wage circuit as the sole source of mass economic agency and instead build a durable civic capital base, funded by a rebalanced tax system, that gives every household an independent economic floor and a meaningful stake in the productive assets that replace labour.

Executive summary

This brief proposes a contingency-ready policy package for the United Kingdom and the European Union that accepts the possibility of unit cost dominance in cognitive labour without assuming the collapse is either certain or immediate. The package rests on three pillars: first, a permanent rebalancing of the tax burden away from labour hiring and toward the productive inputs that substitute for labour, with the revenue supporting an expanded public wealth base; second, a citizen capital endowment and a contingent income floor that activate as labour-share indicators fall; and third, structural measures to prevent housing, energy and platform rents from absorbing any new household resources. The proposal does not require measuring the AI share of individual workflows, does not assume a global moratorium on adoption, and does not rely on preserving commercially unnecessary jobs. If the thesis is substantially wrong, most planks remain valuable as ordinary reforms to tax efficiency, housing affordability, competition and skills. The first 24 months focus on observable, legal and fiscally bounded steps. Years 3 to 5 deepen only if pre-specified triggers are met. The package is designed to be difficult to capture, hard to offshore and politically defensible to a broad coalition of workers, small businesses, savers and public-service users.

The policy package

The full package contains six operational planks and one dormant trigger.

Plank 1: Rebalance the tax base away from labour hiring. In the UK, reduce employer National Insurance contributions and fund part of the loss by ending permanent full expensing for plant, machinery and software above a modest threshold, and by introducing a compute and data-centre energy levy tied to observable commercial energy use and compute capacity. In the EU, member states would commit to a parallel shift in the mix of labour taxation and capital input taxation, coordinated through the European Semester and, where legal obstacles can be overcome, through a new own resource based on digital platform gross margins or data-centre energy. The base is observable: energy metering, server capacity, corporate accounts and customs declarations. It does not require identifying which human tasks an AI performs.

Plank 2: Citizen capital endowment and public wealth fund dividends. The UK National Wealth Fund would take equity, convertible notes or revenue-linked positions in large AI, compute, energy and platform projects that receive public support or procurement. A statutory share of returns would flow into a Civic Capital Endowment. Every resident citizen and long-term resident would receive, at adulthood, a non-transferable, means-tested-lite endowment usable for housing, education, care, starting a business or retirement savings. The EU would explore a parallel European Future Fund through the European Investment Bank and national development banks, with member states free to top up nationally. This gives households a durable asset stake rather than a merely temporary transfer.

Plank 3: Contingent civic income floor. A dormant, low-administration income floor would be legislated now but activated only if the labour share of income falls below a pre-specified threshold for two consecutive years, or if median real hourly earnings fall for three years while productivity rises. The floor would be paid through existing tax and benefit infrastructure, would not be withdrawn pound for pound against low earnings, and would be indexed to a basket of essentials. It is not a universal basic income; it is a conditional circuit-breaker.

Plank 4: Portable skills, benefits and working-time accounts. Every worker would accrue a personal lifelong learning account funded by a small automatic contribution from employers and the state. Health, pension and training entitlements would become more portable across employment, self-employment and non-employment. This preserves agency whether someone is in a conventional job, platform work, retraining or caring.

Plank 5: Rent absorption defences. A significant expansion of social and affordable housing, utility-price regulation that links allowed returns to productivity improvements, and strengthened platform and data portability would prevent landlords, energy suppliers and digital gatekeepers from capturing the value of any new transfers or productivity gains. Without this, the income floor would leak rapidly into rents and margins.

Plank 6: Procurement-conditioned transparency and interoperability. Public contracts above a threshold would require open model-evaluation standards, data portability for users, skills-transition plans for suppliers' workforces, and public reporting of workforce composition by task category. This uses existing procurement law, does not impose AI bans, and creates information that helps calibrate the triggers.

Plank 7: Dormant automation adjustment levy. A fallback levy on the gross value added of highly digitised sectors would be legislated in outline but would take effect only if labour-share triggers are met and if the primary compute and energy levy proves avoidable. This creates a credible backstop without imposing it prematurely.

United Kingdom: first 24 months

The UK can move fastest on tax and welfare parameters because these require only primary legislation and Treasury direction. The first 24 months should deliver the following.

First, the Chancellor should commission an Office for Budget Responsibility review of the labour-share elasticity of employer National Insurance and of full expensing, with a view to reducing the former and narrowing the latter. The direction should be clear in the first Budget and implemented in the second. Employer NICs above the threshold would be reduced by 2 to 3 percentage points, funded by limiting full expensing to the first £1 million of qualifying expenditure per firm per year and by introducing a data-centre energy and compute levy. The levy would apply to commercial facilities above a capacity threshold, measured in kilowatt-hours and server-equivalent capacity, with imports of AI services subject to a withholding or reverse-charge mechanism. This is administrable by HMRC and Ofgem.

Second, the National Wealth Fund should adopt a mandatory equity or convertible participation in any AI, compute or energy project receiving more than a defined amount of public support, procurement preference or subsidy. The Chancellor should direct the Fund to ring-fence a share of realised returns for a Civic Capital Endowment. Legislation would be needed to protect the ring-fence from future raids.

Third, the Department for Work and Pensions should design a contingent civic income floor as a statutory instrument, with a trigger tied to ONS data on the labour share of income and median real hourly earnings. The floor would be tested in pilot form in two regions before national activation. Universal Credit capital limits and taper rates should be eased as an immediate, bounded-regret step that helps households build modest savings regardless of AI outcomes.

Fourth, the government should launch portable lifelong learning accounts, funded initially by redirecting a small slice of existing apprenticeships levy underspend, with top-ups for workers in sectors where hiring is falling.

Fifth, housing and energy measures should include a social housing expansion target funded by land-value capture and by allowing local authorities to borrow against future rental income, plus an Ofgem review of whether utility network returns should be indexed to economy-wide productivity gains.

European Union and member states: first 24 months

The EU cannot move as fast on direct taxation or permanent transfers because of unanimity requirements, the ordinary budget constraint and Article 123 TFEU. Its first 24 months should therefore focus on legal scaffolding, competition and investment architecture, while member states pursue national income measures.

First, the Commission should issue guidance confirming that the AI Act and procurement rules permit contract-linked transparency, interoperability, skills-transition and evaluation conditions for public-sector AI procurement. This gives member states legal certainty to implement Plank 6 immediately.

Second, the Commission should propose a Digital Infrastructure Contribution as a potential EU own resource. The base would be the gross margin of very large digital platform enterprises and the energy use of hyperscale data centres. Because direct tax own resources require unanimity, the proposal should be accompanied by a political commitment from member states to introduce mirror national levies if unanimity is not reached within 24 months. This preserves pressure and jurisdictional coverage even if the EU route stalls.

Third, the European Investment Bank should design a European Future Fund vehicle that can take equity or revenue-linked positions in AI gigafactories, energy projects and platform interoperability infrastructure, with returns directed to a member-state-distributed reserve. Member states could opt into a common legal structure or replicate it nationally.

Fourth, the European Semester should be used to coordinate a shift in the tax mix away from labour and toward capital inputs, with country-specific recommendations urging member states to review employer social contributions, capital allowances and data-centre energy taxation. This does not require harmonised rates, only a common direction.

Fifth, the Commission should legislate or encourage through the AI Act and Digital Markets Act stronger data portability and interoperability for workplace AI systems, so that workers and firms are less locked into single providers. This is ordinary competition policy with bounded-regret value.

Sixth, member states should legislate contingent income floors and portable skills accounts at national level, using the EU scaffolding where available. The EU should not attempt a uniform basic income; the diversity of welfare states makes that politically and legally implausible.

Years 3 to 5 and dormant triggers

If the triggers described below are not met, the package should remain in its first-phase form. If they are met, the following deeper measures would activate.

First, the civic income floor would move from pilot to national operation, with the level set at a defined share of median disposable income, and with its financing anchored in the rebalanced tax base rather than debt.

Second, the UK National Wealth Fund and any European Future Fund would begin distributing realised returns as a periodic citizen dividend, distinct from the income floor and protected against fiscal raids.

Third, the dormant automation adjustment levy would take effect if the primary compute and energy levy is shown to be avoidable through offshoring or corporate restructuring. The fallback base would be UK or EU value added of firms in highly digitised sectors, defined by observable standard industrial classification codes and turnover thresholds.

Fourth, housing and energy rent defences would scale: social housing build rates would rise, local authority housing companies would be empowered, and utility price controls would be tightened if essential-service inflation exceeds wage growth for two consecutive years.

Fifth, lifelong learning accounts would become universal and automatically topped up at key life transitions, such as redundancy, caring exits, mid-career and age 50.

Funding and fiscal arithmetic

The package requires explicit funding. In the UK, reducing employer NICs by 2 to 3 percentage points would cost roughly £10 billion to £15 billion per year, depending on thresholds and behavioural response. Limiting full expensing to the first £1 million per firm would raise several billion pounds annually, but the exact amount depends on investment forecasts and is currently unmodelled here. A data-centre compute and energy levy could plausibly raise £2 billion to £5 billion annually if applied to large commercial facilities, but the base is narrow and avoidance risks are real. The honest position is that the package is not fully funded by these measures alone; it would require either additional revenue, a slower reduction in NICs, or a smaller initial income floor.

The civic income floor, if activated nationally at a meaningful level, could cost tens of billions of pounds annually. It should therefore be scaled to the revenue available from the rebalanced tax base and from returns on public wealth fund holdings. If the thesis is wrong and labour share stabilises, the floor is never activated and the fiscal exposure is limited to the first-phase measures. If the thesis is right, the scale of the problem justifies larger fiscal adjustment, including possible reductions in tax expenditures, carbon pricing revenue, wealth-transfer taxation and public-asset returns.

In the EU, the ordinary budget cannot run a deficit, so any EU-level transfer must be funded by new own resources or by reallocating existing spending. The more plausible route is member-state financing with EU coordination. The Commission should publish a public funding model before any EU-level entitlement is legislated.

The National Wealth Fund has £27.8 billion in capacity, but this is not free money; equity returns are uncertain and may take years to materialise. It should be treated as seed capital for a civic capital base, not as a current revenue source.

Political coalition and public case

The ordinary-language public case is that the economy is becoming more productive but the rewards are not reaching working households, and that the government should ensure everyone owns a share of the productivity gains rather than merely taxing work more heavily than machines. The framing is not anti-technology; it is pro-shareholder democracy and pro-fair transition.

A plausible coalition includes: workers in administrative, legal, financial and public-sector roles who fear disruption; small and medium enterprises that currently face high employer NICs and expensive housing for staff; savers and pensioners who would benefit from a public wealth fund approach; housing and energy bill payers; and technology firms that prefer a predictable, observable tax base to ad hoc AI bans or sector-specific witch hunts.

Losers include: large firms that have structured their investment around permanent full expensing; commercial landlords in areas with expanded social housing; energy and data-centre operators facing new levies; and high-income households if capital allowances or wealth-transfer taxes are tightened. Compensation for losers should be transitional: phased implementation, grandfathering of existing investment commitments, and explicit tax relief for energy-efficient or domestically located compute.

Sequencing matters. The government should begin with visible, bounded-regret measures: lower employer NICs for small firms, portable skills accounts, and social housing expansion. These build trust before the more controversial compute levy and dormant income floor are activated. The public case should be tested through citizens' assemblies in the UK and through the European Citizens' Panels in the EU.

Durability and anti-capture design

Durability requires more than good intentions. Each plank should contain anti-capture mechanisms.

The Civic Capital Endowment should be held by an independent statutory body with a mandate to maximise long-term real returns for citizen beneficiaries. Raiding the fund, changing its mandate or diverting returns to the Treasury should require a supermajority in Parliament or, in the EU, a reinforced qualified majority plus member-state unanimity on the use of returns.

The income floor trigger should be defined in statute and based on published official statistics, not ministerial discretion. Parliament should review it every five years, but the activation decision should be automatic unless both Houses vote to override it on grounds of statistical error.

The compute and energy levy should apply to imports through a reverse charge or border adjustment, and should be harmonised in outline across the UK and EU to reduce arbitrage. Standard industrial classification codes and capacity thresholds should be updated by an independent technical panel, not by political appointment.

Public procurement conditions should be designed to be non-discriminatory under EU and WTO procurement rules, with evaluation standards set by independent standards bodies. This reduces the risk that the policy becomes a disguised industrial subsidy for favoured firms.

Local and devolved governments should receive a statutory share of any new revenues to prevent centralisation of administrative power. In the EU, member states should retain implementation flexibility to protect subsidiarity and democratic legitimacy.

Legal and institutional obstacles

In the UK, the main obstacles are parliamentary time, Treasury resistance to narrowing capital allowances, and OBR scoring uncertainty around new levies and behavioural effects. The government should address these by commissioning independent modelling before enactment and by phasing in changes.

In the EU, unanimity on direct taxation and own resources is the central obstacle. The package therefore avoids relying on EU-level taxation for the core income measures. If unanimity fails, member states should proceed nationally with EU coordination through the Semester and competition rules. Article 123 TFEU rules out ECB monetary financing, so any fund must be equity-based and deficit-compliant.

Procurement conditions must relate to the contract, be verifiable and non-discriminatory. Requirements about parent-company workforce composition or equity participation must be carefully drafted to avoid challenge under EU procurement law or WTO Government Procurement Agreement obligations.

State aid rules limit selective national support for AI projects. Public equity participation through the National Wealth Fund or a European Future Fund must be on market terms or notified and approved.

Data protection and worker consultation rules, including GDPR and national works-council laws, may slow some procurement and workplace transparency measures. These are manageable but require early legal review.

Failure modes, review and exit rules

The package should be treated as a contingency experiment, not as permanent ideology. Review points and exit rules are essential.

Indicator 1: Labour share of income. If the UK and EU labour shares remain within 2 percentage points of their 2015 to 2019 averages for three consecutive years, the civic income floor remains dormant and the automation adjustment levy is not activated.

Indicator 2: Median real hourly earnings. If median real hourly earnings grow in line with productivity for three consecutive years, the urgency of scaling the income floor diminishes.

Indicator 3: Adoption without displacement. If AI adoption rises but hiring in exposed occupations stabilises, the rebalancing of labour taxation can proceed more slowly.

Indicator 4: Rent absorption. If housing costs and utility prices rise faster than general inflation while the income floor is active, the rent-absorption planks must be scaled up before the floor is increased.

Indicator 5: Avoidance and relocation. If the compute levy base erodes by more than 20% through avoidance or offshoring within three years, the dormant automation adjustment levy activates.

Exit rules. Any plank that fails to deliver its intended effect after a full review cycle should be scaled down or repealed. The full expensing reform and the compute levy should be reviewed after three years. The civic income floor, if activated, should sunset after ten years unless explicitly renewed by Parliament. The public wealth fund endowment should be perpetual, but the distribution rate should be adjustable.

Feasibility table

The reasoning behind the feasibility ratings is as follows. UK feasibility is highest where Parliament alone can legislate and existing institutions can administer the change. EU-level feasibility is highest on competition, procurement and investment architecture, and lowest on direct taxation and permanent transfers because of unanimity, the no-deficit rule and Article 123. Member-state feasibility is generally high for national tax, welfare and housing measures, provided funding is found. The main blockers are fiscal credibility, legal constraints and organised opposition from incumbent rent extractors, not technology adoption itself.

What is genuinely new here

The proposal does not simply call for a robot tax, a basic income or retraining. Its distinctive elements are three.

First, it replaces the search for an AI percentage in each workflow with an observable macro trigger: the labour share of income. This accepts the thesis's argument that task-level measurement is administratively fragile and invites arbitrage.

Second, it links productivity taxation to asset ownership through a Civic Capital Endowment, so that households receive a durable stake rather than a stream of transfer payments that can be cut by a future government. Agency is built through capital, not only through consumption.

Third, it treats rent absorption as co-equal with income provision. Most income-floor proposals assume the new money will be spent in the economy; this proposal assumes, more realistically, that landlords, energy suppliers and platforms will try to capture it, and therefore legislates the defences in advance.

Bottom line

Unit cost dominance, if it materialises, cannot be answered by pretending that human labour will remain commercially necessary or by measuring AI's precise share in every office. The durable response is to make households owners of the productive base, to tax the inputs that substitute for labour rather than labour itself, and to prevent the new resources from leaking into rents. The package proposed here is politically difficult but institutionally feasible. It does not depend on global coordination, precise workflow measurement or preserving unnecessary jobs. If the thesis is wrong, most of its planks remain worthwhile reforms to tax, housing, competition and skills. If the thesis is right, it gives the UK and the EU a fighting chance to preserve mass economic agency, democratic legitimacy and state capacity in an economy where wages alone no longer guarantee independence. The honest caveat is that full fiscal arithmetic is still missing; the next step is to model the revenue, cost and behavioural effects before any large permanent transfer is enacted.