When Progress Becomes a Risk

The policy question

The central policy question is not whether people in genuine need should be supported. They should. Nor is it whether work should be expected from everyone regardless of health, disability, caring responsibilities or labour-market opportunity. It should not.

The more difficult question is whether the design of support makes progress sufficiently safe, worthwhile and sustainable for people who can work, increase their hours or build a viable enterprise.

The United Kingdom has moved away from some of the sharpest historic benefit cliffs. Universal Credit normally falls by 55 pence for each additional pound of relevant earnings, rather than ending as soon as work begins. That matters: it would be inaccurate to claim that taking work generally leaves claimants worse off. Yet the formal taper does not capture the whole lived decision. Tax, National Insurance, council tax support, childcare, transport, housing exposure, fluctuating hours and administrative uncertainty can leave a household retaining only a limited and unpredictable share of progress. For self-employed claimants, the Minimum Income Floor can calculate support using assumed earnings that have not actually been realised.

This creates a policy vulnerability. A system intended to support movement into work may produce conditions in which caution, limited progression or remaining within a predictable entitlement feels rational. The behaviour is then labelled as dependence even where it is an adaptation to risk created by the system itself.

The practical proposal is a time-limited Stability Runway: a structured transition for eligible claimants entering work, increasing earnings or starting a credible business. Support would taper in stages, volatile self-employment income would be assessed over a more realistic period, and childcare, coaching and essential passported support would be aligned around sustained progression. Withdrawal would follow demonstrated stability rather than the first appearance of income.

Beyond the false choice between compassion and contribution

Public debate often collapses into two moral positions. One treats welfare as evidence of collective decency; the other treats it as a potential substitute for effort. Both contain part of the truth and neither is sufficient for policy design.

Income protection prevents illness, disability, unemployment, caring responsibilities or economic shocks from becoming destitution. It also supports children whose life chances should not depend entirely on the immediate earning power of their parents. These are legitimate public purposes.

But support also operates within a behavioural system. People make decisions in response to what they can see, what they believe will happen and what they can afford to risk. A claimant deciding whether to increase hours is not comparing only a wage with a benefit payment. She may be comparing a predictable monthly position with variable shifts, childcare deposits, travel, possible loss of council tax support, delayed adjustments and the risk that a job or business will not last.

If the gain is small but the exposure is immediate, hesitation is not necessarily laziness. It may be prudent household risk management. Conversely, it would be equally mistaken to deny that systems can sometimes normalise low expectations or permit long-term disengagement where work is feasible. A credible policy must hold both realities at once: protection without abandonment, and expectation without punishment detached from circumstance.

The question for government is not simply, ‘Did we pay the correct entitlement?’ It is, ‘Did the design increase the likelihood of sustainable participation, capability and independence?’

What the UK system does, and where risk remains

Universal Credit was designed partly to simplify multiple working-age benefits and improve incentives to enter work. Department for Work and Pensions evaluations have reported positive employment effects relative to Jobseeker’s Allowance for some claimant groups, including an early finding that UC claimants were around four percentage points more likely to have been in work within six months than comparable JSA claimants. More recent DWP work has also reported positive employment effects among single parents.

That evidence should be acknowledged. It prevents a critique from becoming a caricature. Nevertheless, four design issues remain.

A taper is better than a cliff, but retention can still be weak

Universal Credit is reduced by 55 pence for each additional pound of relevant earnings above any applicable work allowance. Claimants responsible for a child or assessed as having limited capability for work may receive a work allowance; many other claimants do not.

UC is calculated using net take-home earnings. For a taxpayer paying basic-rate income tax and employee National Insurance, an illustrative additional £1 of gross pay can first fall to about 72 pence, after which the UC taper withdraws 55 per cent of that net amount. The household retains roughly 32 pence before considering any reduction in council tax support or additional costs of working. The exact result varies by household, earnings and locality, but the illustration shows why ‘always better off’ can be technically true while the felt reward from progression remains modest.

Separate systems can recreate cliffs

Council tax support operates outside UC and varies across English local authorities. Institute for Fiscal Studies analysis found that many councils use income bands in which awards fall in steps when a household crosses a threshold. This can layer local cliff edges onto UC’s national taper.

The behavioural issue is not only how much is eventually retained. It is whether the person can predict it. In DWP’s 2025 UC survey, only 57 per cent of claimants eligible for a work allowance and 52 per cent of those without one correctly recognised the 55 per cent taper; around four in ten were unsure. A financial incentive that cannot be confidently understood is a weak behavioural signal.

The system can mistake turnover for stability

Self-employed UC claimants report income and expenses monthly. During an eligible 12-month start-up period the Minimum Income Floor does not apply. Afterwards, UC may be calculated as though the claimant earned a specified minimum even when actual monthly earnings are lower.

The purpose is legitimate: government should not subsidise indefinitely an activity that cannot become a viable livelihood. The problem is temporal fit. Many genuine businesses have seasonal receipts, uneven invoices, front-loaded expenditure and long routes to profitability. A rule intended to test viability can instead reduce support during the period in which viability is being built. Parliamentary committees have previously recommended longer start-up periods and assessment windows of up to a year for irregular earners.

Conditionality can produce exits without earnings

Conditionality has a legitimate place where capability, opportunity and expectations are reasonably aligned. But sanctions are not equivalent to employment outcomes. DWP’s causal analysis found that sanctions shortened UC spells mainly by increasing exits into states with no PAYE earnings, while exits into PAYE earnings decreased. A reduction in the claimant count can therefore be recorded as an administrative benefit without producing the intended value of sustainable work.

This is the distinction between delivery and value: an intervention can be implemented correctly, change a measured indicator and still fail to improve the outcome that justified it.

A behavioural value perspective

The Behavioural Theory of Value Realisation, or BTVR, is an emerging practitioner-led perspective, not an empirically validated universal law and not a substitute for economics, benefits management or programme evaluation. Its contribution is to examine the mechanism between policy intent and experienced value.

The intended value is not merely fewer benefit payments. It is greater sustainable participation, improved household resilience, productive capability and a fair reduction in public dependence where that is possible.

The organisational conditions include taper rates, assessment periods, work-coach capacity, childcare availability, local council tax rules, reporting burdens, labour demand and the credibility of assurances about what will happen when earnings change. These conditions shape what feels rational, legitimate, defensible and possible.

Three forms of quiet divergence can follow. First, a claimant limits hours or avoids uncertain enterprise because predictable support has greater stability-adjusted value than a fragile increase in gross income. The measured benefit remains, but the intended value of progression is only partially realised.

Second, a claimant moves off UC after a sanction or administrative interruption without moving into secure work. The caseload falls, but value has not necessarily been created. The cost may transfer to debt, housing insecurity, family support, local government, health services or charities. The value did not disappear. It transferred.

Third, a start-up produces early turnover but not yet stable profit. Support falls, the venture fails and the person returns to benefits. The system records an initial transition and a later claim as separate events, obscuring the failure to sustain value across time.

BTVR therefore changes the accountability questions.

• Why did the claimant’s behaviour make sense within the conditions created?
• Did support withdrawal follow demonstrated stability or merely visible activity?
• Who experienced the intended value, and who absorbed transferred risk or cost?
• Did the intervention build capability that endured after support ended?

What international comparisons can and cannot tell us

Cross-country comparisons are hazardous. Welfare systems operate within different tax bases, labour markets, household structures, informal economies, public services and cultural expectations. A low benefit bill can reflect high employment, but it can also reflect low coverage, family absorption of hardship or unmet need. The purpose of comparison should therefore be to identify mechanisms worth testing, not to declare one national model morally superior.

Mauritius: supplement low-paid work rather than penalising effort

Mauritius offers a directly relevant mechanism. Its Negative Income Tax provides government support to eligible low-paid employees rather than reserving support only for those outside work. When the scheme was adjusted, government based eligibility on basic salary rather than total earnings so overtime would not be penalised and the scheme would not discourage additional work. Conditions excluding new and part-time workers were also relaxed.

This does not make Mauritius a complete template for the UK. Its economy, wage structure and wider benefit system differ substantially. But the design principle is clear: assistance can reinforce earned income and additional effort, rather than make the first gains from work the trigger for immediate loss.

The Netherlands: support the transition into viable enterprise

The Dutch Bbz scheme allows municipalities to support people starting businesses while receiving social assistance. Subject to an approved and credible business plan, support can include a temporary supplementary benefit, start-up capital or a loan, a starter programme, advice and coaching.

The notable feature is not generosity alone. It is the explicit recognition that enterprise requires a transition architecture. Income support, viability assessment, capital and capability are treated as connected conditions rather than separate policy silos. The business is tested, but it is also given a route to become testable.

Finland: removing insecurity helps wellbeing, but is not sufficient for work

Finland’s 2017–18 randomised basic-income experiment paid 2,000 unemployed people €560 per month regardless of other income or active job search. The employment effect was small: in the principal second-year reference period, recipients worked about six more days than the control group, while the first year showed no group-level employment effect. Recipients nevertheless reported better economic security, mental wellbeing, trust and confidence.

This is an important counterweight to simplistic arguments on both sides. Removing withdrawal and conditionality did not produce a large employment transformation, suggesting that bureaucracy and financial incentives were not the only constraints. But greater security may still create legitimate value in wellbeing, autonomy and trust. Policy must define which value it seeks and avoid treating one outcome as proof of all others.

Denmark: security and reciprocal activation can coexist

Denmark’s model has historically combined substantial income security, relatively flexible hiring and dismissal, and active labour-market policies. OECD analysis notes that generous, long-duration benefits without effective activation contributed to rising unemployment in the 1980s, while reforms combining reduced duration with stronger job-search and activation requirements preceded falling unemployment in the 1990s.

The lesson is balance rather than severity: security can preserve mobility and legitimacy, while reciprocal obligations and active services maintain connection to work. Conditionality is most credible when the system can offer something practical in return.

Graduation programmes: do not withdraw the platform before capability forms

Economic-inclusion and ‘graduation’ programmes across African and other lower-income settings commonly combine predictable transfers with assets or capital, savings, training, coaching and market links for a time-limited period. World Bank reviews and multi-country randomised evaluations have found sustained improvements in income, assets, consumption and resilience in many settings, though effects and cost-effectiveness vary.

Their central insight is behavioural and systemic: poverty and weak labour-market participation rarely reflect a single missing input. Removing cash alone does not create capital, confidence, childcare, skills, customers or opportunity. The platform and the progression mechanism must operate together long enough for capability to stabilise.

A Stability Runway pilot for the UK

Government does not need to redesign the entire welfare state to test a different transition logic. It could run a bounded pilot across a diverse set of Jobcentre Plus and local-authority areas.

Purpose and eligible groups

The purpose would be to test whether staged, predictable support for people entering work, increasing earnings or starting a credible business produces more sustained employment, earnings progression and business survival than standard withdrawal arrangements.

The pilot should focus on claimants for whom work or enterprise is a realistic route, while retaining separate protections for people whose health, disability or caring circumstances make conventional progression unsuitable.

• People moving from no earnings into employment.
• Low-paid workers increasing hours, responsibilities or skills.
• Self-employed claimants with an approved viability plan.
• Parents returning to work where childcare is a material constraint.

Core design

1. A visible transition guarantee. Before changing work or starting to trade, the claimant receives a personalised statement showing expected net household income under realistic scenarios, including UC, tax, council tax support and childcare. A capped guarantee protects against a material, unexpected loss during the first three months while the position is corrected.
2. A staged earnings runway. For a limited period, an enhanced earnings disregard or transition supplement allows the claimant to retain more of early earnings. Support then steps down transparently over 12 to 18 months, converging gradually with ordinary UC.
3. A realistic enterprise clock. Approved start-ups receive up to 24 months before full application of the Minimum Income Floor, with formal reviews at 6, 12 and 18 months. Continuation is earned through evidence of activity, learning, demand and a credible route to viability. Seasonal and volatile income is averaged quarterly, with annual reconciliation and safeguards against manipulation.
4. Support that addresses actual constraints. The package includes up-front childcare where required, business coaching or skills provision, and referral to finance or procurement opportunities. Participation requirements follow the agreed progression plan rather than generic activity counts.
5. Temporary alignment of passported support. Council tax support and selected passported benefits are protected or tapered according to a common schedule during the runway. Local authorities receive compensating funding and shared data standards so national incentives are not quietly reversed locally.
6. A stabilisation test before exit. ‘Ready to move on’ means sustained net income and reduced volatility across a defined period, not one high-earning month or a temporary contract. The test considers household costs and resilience as well as gross earnings.

Evaluation

A stepped-wedge or randomised phased rollout could compare eligible participants receiving the runway with similar claimants under standard arrangements. Participants should be followed for at least 24 months after entry.

• Sustained employment at 6, 12, 18 and 24 months.
• Earnings growth and hours, not merely benefit exit.
• Business survival, profit and volatility.
• UC re-entry after apparent exit.
• Disposable household income after childcare, housing and travel.
• Debt, arrears and use of crisis support.
• Health, confidence, trust and perceived legitimacy.
• Distributional effects by disability, sex, ethnicity, parenthood and locality.
• Total fiscal effects across DWP, HMRC, councils and other public services.

Risks, safeguards and boundaries

The proposal has real risks. It could subsidise low-productivity employment or businesses that will never become viable. It could create unequal treatment between pilot participants and other claimants. Longer support could increase short-term expenditure, attract gaming or encourage employers to hold wages down. Work coaches and local authorities may lack the capacity to make consistent viability judgements. A guarantee could also become too complex to understand, defeating its purpose.

These are reasons to test carefully, not reasons to preserve the current design without examination. Safeguards should include capped support, published eligibility criteria, independent review of adverse decisions, common viability standards, audit of employer behaviour, and automatic sunset unless the evidence supports continuation. Administrative burden should be measured as an outcome, not treated as a free input.

There is also a deeper warning. Some people will not progress because suitable jobs, health support, childcare, transport or demand do not exist. No taper reform can solve a weak local economy. If the pilot works only where opportunity is already strong, that boundary condition must be reported rather than hidden in a national average.

Conclusion: support should be a platform, not a perimeter

The UK does not face a choice between unlimited welfare and abandonment. It faces a design challenge: how to protect people from hardship while making participation, effort and enterprise progressively safer and more worthwhile.

The evidence does not justify describing claimants generally as lazy, nor does it justify assuming that any increase in conditionality creates work. Finland shows that security alone does not remove every barrier. Denmark shows that support and reciprocal activation can coexist. Mauritius demonstrates the deliberate use of in-work support to avoid penalising additional effort. The Netherlands and graduation programmes show the value of combining temporary income protection with capital, coaching and a credible route to viability.

If the answer is caution, concealment, limited hours or premature business exit, the problem is not resolved by repeating the policy intention more firmly. Government must examine the conditions translating that intention into lived decisions.

Support should not be withdrawn at the first sign of activity. It should reduce as capability and stability become real. That is not dependency by design. It is value realisation by design.

Sources

• Department for Work and Pensions, Universal Credit and earnings ↗
• Department for Work and Pensions, Universal Credit: how your wages affect your payments ↗
• Department for Work and Pensions, Understanding the impact of Universal Credit on the labour market ↗
• Department for Work and Pensions, Estimating the employment impact of Universal Credit among single parents ↗
• Institute for Fiscal Studies, Universal credit: incomes, incentives and the remaining roll-out ↗
• Institute for Fiscal Studies, How does council tax support shape household incomes and work incentives? ↗
• Department for Work and Pensions, The Universal Credit Survey 2025 ↗
• Department for Work and Pensions, Self-employed earnings: Guidance V17.0 ↗
• Work and Pensions Committee, Universal Credit for self-employed people ↗
• Department for Work and Pensions, The impact of benefit sanctions on employment outcomes ↗
• Government of Mauritius, Negative Income Tax ↗
• Republic of Mauritius, Budget Speech 2018/19 ↗
• Netherlands Enterprise Agency, Starting a business while on social assistance benefits ↗
• Finnish Ministry of Social Affairs and Health, Results of the basic income experiment ↗
• OECD, Reforms of active labour market policies in Nordic and OECD countries ↗
• OECD Economic Surveys: Denmark 2008 ↗
• World Bank, Assessing the Impact and Cost of Economic Inclusion Programs ↗
• Innovations for Poverty Action, The Ultra-Poor Graduation Approach ↗

Material limitations

• This analysis is based on published evidence and policy design rather than personal experience of claiming Universal Credit. The proposed mechanisms and assumptions should therefore be tested with claimants and informed by lived experience before any pilot is designed.
• Cross-country mechanisms are illustrative and not estimates of what identical policies would produce in the UK.
• UC interactions depend on household circumstances, tax status, housing, childcare and local council tax rules; the 32-pence illustration is not a universal retention rate.
• Some comparator evidence is historical because it concerns the evaluated introduction or reform of a policy.
• BTVR is used as an analytical perspective and source of testable questions, not as established causal proof.
• The proposed Stability Runway requires microsimulation, costing, legal analysis, claimant co-design and a formal equality impact assessment before implementation.