UK job vacancies fell to 707,000 in the March to May 2026 quarter, the lowest reading since February to April 2021, according to the Office for National Statistics (ONS). The figure lands below pre-pandemic levels and arrives in the same reporting cycle in which the ONS acknowledged a level of reduced quality in the underlying data supporting its labour market bulletin. For anyone relying on that headline to make a hiring decision or a pay negotiation, the signal is real but the margin of error is wider than usual.

What the ONS actually published

Total estimated vacancies fell by 31,000 (4.2%) compared with a year earlier, with declines recorded across 10 of the 18 industry sectors the ONS tracks and across most employer-size bands. Retail, hospitality and parts of the public sector absorbed the sharpest contraction, with employers pulling back on new hiring rather than cutting existing headcount outright. Independent corroboration comes from HMRC (His Majesty's Revenue and Customs) payrolled-employee data, which has been tracking a similar cooling trend throughout early 2026, and from recruitment industry trackers reported by the Financial Times.

The data-quality caveat deserves a sentence of its own. The ONS has spent the past two years attempting to rebuild credibility in its labour market series after response-rate problems undermined the Transformed Labour Force Survey. The most recent bulletin is transparent about the fact that input quality has not yet returned to the standard the agency is targeting. In plain terms: the direction (vacancies down) is consistent across multiple independent indicators, but any single precise figure should be read as a central estimate with wider-than-usual error bars rather than a precise count.

For context on pay, ONS data shows average regular earnings rose 1.3% in real terms in the three months to April 2026 - a slower-moving series than vacancy counts, but one the ONS itself considers more statistically reliable at this juncture. The Resolution Foundation has separately noted that real-wage recovery remains uneven across income bands, with lower-paid workers in service industries still trailing pre-2022 purchasing power.

For employees and candidates

If you are job-hunting or weighing an internal move, the practical implication is not that the market has collapsed. It is that a single national vacancy number was never the right anchor for a personal career decision, and that is more true now than it was twelve months ago.

A few data points worth triangulating against the ONS headline:

• Sector-level vacancy counts, not the aggregate. The 4.2% annual fall is an average across 18 sectors; the distribution underneath is what matters for your specific role. Technology, professional services and healthcare are not behaving like retail and hospitality - and the ONS Annual Survey of Hours and Earnings (ASHE), published each October, remains the most granular public source of sector and regional pay benchmarks available to UK workers.
• Real pay growth, not raw vacancy volume. Average regular earnings rising 1.3% in real terms in the three months to April 2026 is a slower indicator than vacancy counts, but it tells you whether offers in your sector are keeping pace with prices - which is ultimately what a pay negotiation is about.
• Live job boards and verified salary submissions. Anonymised salary data contributed by workers in your industry and region - the kind surfaced by employer review and salary benchmark platforms - updates continuously and reflects the spread of actual offers rather than a quarterly survey average.

For salary negotiation specifically, a softer vacancy market does not automatically close the conversation. It means the case needs to be more concrete. Arriving with a documented pay range grounded in ASHE sector data and live salary submissions, alongside a clear account of what you have delivered in the last twelve months, is a more defensible position than relying on general market heat. ACAS (Advisory, Conciliation and Arbitration Service) guidance on pay discussions also notes that employees who raise pay in writing, with supporting evidence, tend to reach resolution faster than those who rely on verbal requests alone.

For HR and hiring managers

From the employer side, the temptation when ONS quality flags appear is to treat the macro number as noise and lean entirely on internal data. That instinct is understandable but partially mistaken. The vacancy decline is corroborated by HMRC payrolled-employee figures, by recruitment industry trackers and by consistent reporting in the Financial Times - so the directional signal is credible even if the exact level carries a wider margin of error than in a normal reporting cycle.

What changes is how you build the second-half hiring and compensation plan:

• Use multiple benchmarks, not a single ONS series. If your most recent compensation review used the ONS Annual Survey of Hours and Earnings as its sole anchor, the CIPD (Chartered Institute of Personnel and Development) Reward Management Survey is a useful second source. CIPD's annual survey of UK reward practitioners captures employer-side pay-setting intentions and profit-share take-up data that the ONS series does not. Adding at least one live source - your own offer-acceptance data, sector-level salary submissions, recruiter intelligence - before setting bands for the next cycle reduces the risk of anchoring on a figure that carries a wider-than-usual error bar.
• Reread your time-to-hire trend. A 4.2% fall in vacancies should, mechanically, ease candidate competition in most roles. If your hiring funnel has not loosened, the constraint is more likely internal - job specification, employer brand, or compensation band - rather than macro conditions.
• Be transparent with hiring managers about data quality. A finance director who has read the Financial Times coverage of the ONS quality caveat will not be reassured by a headcount forecast that treats the headline figure as definitive. Presenting the uncertainty explicitly, and showing the triangulation methodology behind your plan, is a stronger governance position than presenting false precision. For organisations covered by the Information and Consultation of Employees Regulations 2004 (ICE Regs), sharing the reasoning behind workforce planning decisions with employee representatives is not merely good practice - it is a statutory expectation where a valid request has been made.

The wider point about labour market data

The ONS is not uniquely affected here. Statistical agencies across the OECD have faced declining survey response rates since the pandemic, and the gap between headline figures and granular ground-level conditions has widened in multiple countries. The UK case is notable mainly because the ONS has handled the problem transparently, publishing the quality caveat alongside the data rather than confining it to a methodology annex. That transparency is worth acknowledging even when it makes the numbers harder to use.

The lesson for decision-makers is not to abandon official statistics. It is to stop using any single series as the only anchor. A vacancy headline from the ONS, a pay-growth series from ASHE, HMRC payroll counts, internal hiring-funnel data and worker-submitted salary benchmarks all measure different things. The most useful answer to whether now is a sensible moment to ask for more pay, or to open a new hire, sits in the overlap of those sources - not in any one figure on its own. For HR professionals, the CIPD's guidance on evidence-based people decisions offers a practical framework for exactly this kind of multi-source triangulation.

Key takeaways

• UK vacancies fell to 707,000 in March to May 2026 - a five-year low and below pre-pandemic levels - per ONS figures.
• The ONS has flagged reduced quality in the underlying data; treat the precise figure as a central estimate with wider error bars, noting that the downward direction is corroborated by HMRC payrolled-employee data and independent recruitment trackers.
• Candidates should triangulate sector-level data from ONS ASHE, real earnings growth and live salary submissions before entering a pay negotiation, rather than relying on the headline vacancy number alone.
• HR and reward teams should supplement ONS ASHE benchmarks with CIPD Reward Management Survey data and internal offer-acceptance evidence before setting compensation bands for the second half of 2026.
• Organisations with active Information and Consultation of Employees (ICE Regs 2004) arrangements should ensure their workforce planning communications reflect the current uncertainty in official data, rather than presenting headline figures as more reliable than the ONS itself claims them to be.