UK Labour Market Evaluation Fund: findings 2023-25

Ever wondered what actually works to help people get into jobs, stay in work, or earn more? The UK government set up the Labour Market Evaluation and Pilots Fund after the Spring Budget 2023 to test ideas properly, not just talk about them. Together we’ll walk through what departments tried, how they measured it, and what the early findings tell you as a learner, teacher or curious citizen.

In plain terms, HM Treasury and the Cabinet Office’s Evaluation Task Force ran three competitive rounds where departments could bid to evaluate existing programmes or pilot new ones. Round 1 awards landed in August 2023, Round 2 in December 2023, and Round 3 in May 2024. The common thread is rigorous methods: randomised trials, quasi-experiments, and linked datasets that connect health, tax and census records while protecting privacy.

A quick methods guide you can use in class: a randomised controlled trial (RCT) compares outcomes for people randomly assigned to get an intervention versus those who do not. Quasi-experimental studies attempt a fair comparison by matching similar people. Linked administrative data lets researchers follow earnings and employment over time without relying on memory. When you see ‘percentage points’, think of a change in probability, not a percent of a percent.

Childcare costs shape careers. HMRC commissioned Ipsos to survey around 3,600 parents and interview 60 in depth to understand how Tax‑Free Childcare influences work decisions. Parents told researchers that TFC can shift hours for some families, but it sits alongside other big factors like childcare availability, employer flexibility and household finances. Results are expected to be published in November 2025; the work also feeds into the expanded free hours policy announced at Spring Budget 2023.

On the numbers side, HMRC linked Tax‑Free Childcare, Child Benefit, PAYE and Self Assessment records to study labour supply changes, especially for mothers of one‑ to two‑year‑olds and three‑ to four‑year‑olds. By identifying similar non‑users through Child Benefit, the team created a counterfactual to estimate TFC’s impact on earnings and the timing of return to work from pregnancy until a child turns two. This analysis supports the Department for Education’s evaluation of the early years expansion.

Health conditions can tilt career paths. Using census, hospital and tax data, the Office for National Statistics (ONS) looked at endometriosis. It estimated around 2% of women of reproductive age in the linked population had a recorded diagnosis between 2011 and 2021, with an average diagnosis age of 35. A follow‑up analysis found monthly earnings fell on average one to five years after diagnosis, and the chance of being a paid employee dropped by 2.7 percentage points. That is a measurable hit to participation and pay.

ONS also studied two routes to diabetes prevention: bariatric surgery and the NHS Diabetes Prevention Programme. The first publication reported a sustained rise in monthly earnings from six months to five years after surgery, mostly because more people entered work. A second paper, including the prevention programme, is due in December 2025, working with the University of Leicester. The project also examined who is able to access these services, which matters for fairness.

Workplace flexibility is often promised but not always visible. The Department for Business and Trade (DBT) tested a ‘flexible schedule’ badge on sponsored job adverts on Indeed. It nudged up views by 0.5%, but applications did not rise significantly. What it means: simply labelling a benefit may not move behaviour if underlying job design, pay or hours do not match what applicants need.

DBT also ran an RCT with new parents. One group received letters signposting their flexible working rights; another group was offered personalised career support sessions. Six months on, neither intervention produced a statistically significant boost in returning to work, and there was no difference between mothers and fathers. Still, parents who took up the support rated it highly. In evaluation, satisfaction and impact can diverge.

Small firms often lack HR capacity. A DBT‑backed trial of one to two days of bespoke HR support pivoted to an implementation and process study after low take‑up. Interviews with SMEs, HR consultants and local partners suggested the support helped with policies, recruitment basics and legal compliance. It was less suited to deeper staffing challenges that typically need longer engagement. That’s a practical lesson for programme design.

Money rules can discourage work in supported housing. The Department for Work and Pensions (DWP) and the West Midlands Combined Authority tested a rent simplification and support offer for 18‑ to 24‑year‑olds. Participants paid a portion of rent from earnings, with at least 87% topped up for up to six months. Take‑up was too low for a robust statistical read, and interviews revealed mixed views on the generosity and duration. Even so, the pilot adds real‑world learning about incentives and administrative complexity.

Place matters too. Jobs Plus, a long‑running US model, is being tested across ten UK social housing sites that went live in summer 2024. The Learning and Work Institute is leading the evaluation, with an implementation report due in summer 2025 and the first impact results scheduled for spring 2026. Expect this to test whether wrap‑around support can lift employment in communities facing multiple barriers.

Access to childcare is uneven. Ofsted and ONS created a hyper‑local picture of ‘childcare deserts’ using transport links and the number of local children to map access across about 180,000 neighbourhoods in England. An interactive ONS tool and Ofsted visuals show where provision has changed since March 2020. That evidence is already shaping conversations about parent choices and local labour supply.

Skills for the AI era came under the microscope. The Department for Science, Innovation and Technology’s Flexible AI Upskilling Fund launched in April 2024 to co‑fund training in small and medium‑sized professional services firms. Demand came in lower than planned, with 327 successful applications. Ipsos’ process evaluation pointed to low awareness, the need for clearer guidance on courses, and lingering cost barriers even with matched funding. The impact evaluation runs to 2027.

Two further health‑and‑work studies will add detail. ONS is linking records to show how musculoskeletal conditions and orthopaedic surgery relate to work, pay and hours, with publication expected in 2025. Another ONS analysis has already examined the aftermath of adverse pregnancy events. Compared with the year before an event, total earnings losses over several years ranged from about £3,500 after a missed miscarriage to around £13,600 after a stillbirth. Following an ectopic pregnancy, earnings were lower for at least three and a half years, with an average loss of about £2,040. Employment probabilities typically recovered within one to two years, depending on the event.

Housing support reform in Kirklees offers a different lens. The Government Outcomes Lab at Oxford evaluated the Kirklees Better Outcomes Partnership, which shifts to person‑centred support across housing, health and work. The August 2025 impact report found participants were on average 3 percentage points more likely to sustain a job or start one within six months, compared with matched pre‑programme cohorts. Participants were also more likely to sustain non‑receipt of the Universal Credit Housing Element for up to 18 months, with effects between 6% and 10%, and the approach was more cost‑effective per outcome.

A quick tour of the rest. ONS analysed NHS Talking Therapies for anxiety and depression: seven years after starting treatment, the likelihood of being a paid employee was 1.5 percentage points higher, with a peak average earnings uplift of about £17 per month two years after completion. Gains were larger for people actively seeking work and for adults aged 25 to 34. ONS has also linked data on major cardiovascular events such as stroke, heart failure and heart attack, and on NHS Health Checks’ role in prevention, with findings due in 2025, and mapped ‘transport deserts’ to quantify how getting to jobs varies by place and job type.

The Cabinet Office commissioned several evidence reviews and trials. A July 2025 review found strong links between skills and productivity, especially where management is effective and innovation is encouraged. An RCT of the Digital Excellence Programme for senior officials improved behaviours and attitudes but not a new self‑reported productivity measure, with a full report expected in September 2025. Another review on civil service learning design favoured blended, interactive approaches such as peer coaching and workshops, and a study on supporting sick‑listed public‑sector workers highlighted the role of supportive policy, positive delivery culture and coordinated management. Meanwhile, DWP is comparing quasi‑experimental methods with past RCTs to judge robustness. The STEP Ukraine programme, delivered by World Jewish Relief and the British Council to 7,805 people, combined ten weeks of daily online English with twelve weeks of employment support; evaluations on impact and value for money are due in early 2026. Outside Whitehall, Northern Ireland’s JobStart 50+ pilot supported 180 economically inactive people into six‑month roles with wage and onboarding support; an independent evaluation by Ulster University has been completed.

What you can do with this as a learner or teacher: use these studies to practise reading an RCT chart, to debate what counts as a meaningful effect size, and to spot where good intentions met practical limits. Null results are not failures; they narrow the field and save money. Keep an eye on pending reports due in November and December 2025 and into 2026, and ask the crucial question every time: who benefits, who misses out, and how do we know?