The low-gear labour market − speech by Clare Lombardelli

Introduction

Today I want to talk about an unusual feature of the UK labour market. It is cooling, but mainly through less hiring rather than more firing.

The economy has been more resilient to the recent energy cost shock than expected. Activity has held up better than anticipated, unemployment has risen only gradually, and redundancies remain relatively low. That may sound like good news. In lots of important ways, it is. But a low-hire, low-fire labour market creates its own problems. It protects those already in work while making it much harder for those trying to get in or move on.

This matters because the current slowdown in hiring comes on top of a longer-term decline in mobility within the labour market. People are changing jobs less often. And when they do move, they are less likely to change industry or move into work requiring different skills. The labour market has become less effective at moving people towards new opportunities. The labour market is stuck in low gear.

That has three important consequences.

First, it affects how we interpret the rise in unemployment. If those looking for work are poorly matched to the jobs available, higher unemployment may do less to ease wage pressures than it has in the past.

Second, it can weaken productivity and growth. Moving jobs helps workers find roles where their skills are better used. When that process slows, people miss opportunities to progress and firms find it harder to recruit the staff they need. The costs are falling particularly heavily on young people, for whom the first step into work has become harder.

Third, it can make the economy less able to adapt. That matters as artificial intelligence changes the tasks people perform and the skills employers need. The gains from AI will depend partly on how readily people can move towards the new opportunities it creates.

I will begin with the unusual way in which the labour market is cooling. I will then look at the longer-term decline in job mobility and the particular challenges facing young people. Finally, I will consider what this means for inflation, productivity and the economy’s ability to adapt, before turning to the outlook for monetary policy following last week’s decision.

Cooling through hiring, not firing

The UK economy has, so far, proven to be more resilient to the energy price shock than many expected. We produced our first post-conflict projections just nine weeks after the outbreak. Those were very uncertain times, so we published three scenarios and no central projection.

Chart 1 shows how the economy has evolved in terms of output and unemployment since those projections. Activity has been significantly stronger than the middle scenario from our April report and unemployment has been lower.

To date, real activity has been unexpectedly robust given the circumstances. That said, the unemployment rate still increased from below 4% in 2022 to peak above 5% late last year. It now stands at 4.7% in the latest data for the three months to August.

This rise in unemployment has been persistent but relatively gradual by historical standards as Chart 2 shows. Unemployment increased for six consecutive quarters from mid-2024 to the end of 2025 but by only around 1 percentage point in total, before falling back in the latest data. Sustained increases of this kind have historically tended to occur around recessions when unemployment typically rose much more sharply following abrupt job losses. Unemployment fell in the most recent data and in our November central projection remains broadly stable in the region of 5%, as shown in Chart 1.

The headline unemployment rate does not tell us how the labour market is adjusting. To see that, we need to distinguish between people losing existing jobs and people finding it harder to get new ones.

Labour market flows data show that the loosening has not been driven by an abrupt rise in job losses (Chart 3). Redundancies have remained relatively stable. According to the LFS^{footnote [1]}, the redundancy rate is at roughly the same level as it was before the pandemic. This picture is consistent with data on corporate insolvencies, which have been stable or even moderately declining following a slight jump after the pandemic.

In contrast, we have seen a larger adjustment in vacancies and other measures of labour demand. Vacancies have fallen and are now well below their pre-pandemic level (Chart 4). The REC measure of labour demand also remains contractionary (below 50), though it has shown signs of improvement. In the recent labour market loosening, firms have so far adjusted much more through reducing demand for new workers than through shedding existing employees.

Explicitly breaking down the actual rise in unemployment into the separate contributions from hiring and firing shows just how different the recent slowdown has been (Chart 5). Weaker hiring is by far the more significant driver.

The cooling in labour demand has also been uneven across sectors (Chart 6). Vacancy rates are now much more dispersed than before the pandemic. In some industries, including public administration and defence, vacancies remain well above their historical average. In others, notably retail, they have fallen well below it.

This uneven pattern matters because workers do not seamlessly move between industries. When opportunities are contracting in one part of the economy but remain stronger elsewhere, adjustment requires workers to move to areas where staff are more in demand.

Wage growth provides important information on labour market conditions. As Chart 7 shows, pay growth has fallen substantially from its post-pandemic peak. But the latest Bank assessment is that underlying wage growth remains at the upper end of the range consistent with the inflation target, while forward-looking indicators point to continued strength into next year. I will return later to what this combination of weak hiring and continued wage pressure may tell us about the labour market.

Why are firms cutting hiring without making more redundancies?

The labour market is responding to a range of factors. Following the pandemic, it was exceptionally tight, as strong demand ran ahead of the economy’s capacity. More recently, the energy cost shock and heightened uncertainty have affected firms’ decisions.

As the economy reopened from the pandemic, acute labour shortages were widespread. The furlough scheme and wider fiscal and monetary support prevented large-scale unemployment. The rapid expansion in demand then left firms struggling to attract the skills they needed. Bank Agency intelligence and a range of business surveys repeatedly reported acute recruitment difficulties and skills shortages, leading many firms to place a high value on retaining experienced staff (Chart 8 and Chart 4).

Memories of recent labour shortages may have led employers to be more hesitant to make workers redundant, concerned that they will struggle to find the skills they need when activity recovers.

The conflict in the Middle East has further challenged firms. Disruption to energy supplies has raised oil, gas and other commodity prices, increasing businesses’ costs and squeezing household’s real incomes. Geopolitical uncertainty has also increased, making it harder for firms and households to plan.

Businesses have responded to these pressures in a variety of ways, including absorbing some of the increase in costs in their margins (Chart 9). Profit margins remain compressed as competition and weak demand limit firms’ ability to pass higher costs through to prices. Looking ahead, our Agency network reports that business contacts expect to restore margins through a combination of price increases and further efficiencies such as through procurement and automation.

Part of firms’ adjustment has also been by cost reductions, including hiring restraint, particularly in consumer-facing sectors where intelligence suggests that forward-looking employment intentions remain negative. And while overall employment intentions have recently firmed a little, they remain below normal levels, consistent with continued caution about hiring.

So far, businesses have been able to absorb part of their higher energy costs in their margins. That has probably helped to limit the adjustment in employment, with firms responding through lower hiring, rather than higher redundancies. Heightened uncertainty may be driving the use of margins as a buffer in case the energy cost shock proves only temporary.

But that buffer cannot be used indefinitely. The longer energy costs stay high, the greater pressure there will be on businesses to restore margins through other means including higher prices, lower wages or lower employment. A more prolonged shock could therefore force firms to adjust current employment as well as hiring plans.

So the labour market has so far adjusted mainly though weaker hiring rather than higher firing. But this cyclical slowdown has come on top of a longer-term decline in the movement of workers between jobs. I’ll turn to that next.

A less mobile labour market and its effect on young people

We have seen how the current economic cycle has been characterised by unusual stability in the existing workforce. But that cyclical pattern sits alongside a longer-term trend: people are moving less often into, out of, and between jobs.

At any point in time, there is always a substantial amount of movement in the labour market. In 2025, almost 3 million people switched jobs, around one in twelve employed. A further 3.6 million people left employment at some point during the year while about 4.3 million people entered.^{footnote [2]} This underlying amount of change is higher than might be assumed from the stability of aggregate statistics, such as the headline unemployment rate.

People are making fewer job moves

The clearest long-run measure of labour market mobility is the job-to-job switching rate: the share of people moving directly between jobs without a spell of joblessness. It is cyclical, rising in booms and falling in recessions. Chart 10 shows this measure for the last 30 years.

Despite the volatility, there is a pronounced downward trend. Thirty years ago, more than 3% of employed people changed jobs each quarter. Over the past year, the quarterly rate has been a little above 2%. This is not too far off global financial crisis lows despite much less cyclical weakness.

Mobility is not only about people moving directly between jobs. People also move into and out of employment altogether. Chart 11 combines job-to-job moves with entries to and exits from employment. This more comprehensive measure also shows a long-term slowing.

And job moves are becoming narrower

Not only are people moving jobs less often, the moves they make are also becoming narrower (also see Haldane, 2019). In the early 2000s, more than half of job switches crossed industry boundaries. That share fell by around ten percentage points during the financial crisis and has not recovered (Chart 12). Moves between broad occupational skill groups have also declined, while moves within the same skill level have become more common (Chart 13).

Why has job mobility declined?

The decline in labour market mobility reflects a combination of factors.

Structural change across the economy has slowed since its peak in the 1980s Cominetti et al. (2022) which may have reduced the opportunities for workers to move between jobs and sectors. ONS evidence finds that job reallocation was lower in 2024 than in 2001 in every industry, showing the trend is broad based rather than concentrated in a few sectors. Research has linked slowing labour market dynamism to a shift towards older, more established businesses (Davis and Haltiwanger, 2014).

At the same time, the workforce has aged. This reduces mobility because older workers change jobs, occupations and locations less frequently than younger workers. Rising inactivity associated with ill health also hampers labour market movement. And the move from EU free movement to a migration system more reliant on employer sponsorship and eligible occupations has changed how readily migrant labour can respond to shifts in demand.

These factors may reinforce one another. Fewer fast-growing businesses generate fewer opportunities for people to move into, while a less mobile workforce makes it harder for labour to follow the opportunities that do emerge.

Young people are bearing the greatest cost

The effects of these changes have been highly uneven. Younger workers and new entrants have borne a disproportionate share, even as aggregate employment has remained relatively resilient.

The difficulties facing young people have been well documented. Unemployment among 16 to 24-year-olds has risen much faster than for other age groups (Chart 14). The youth unemployment rate is now more than three times the aggregate rate, its highest relative level since the early 1990s. Despite accounting for only around 12.5% of the labour force, 16 to 24-year-olds account for more than half of the rise in aggregate unemployment since mid-2022 (Chart 15).

The extent of this rise in youth unemployment is similar to that seen in the aftermath of the global financial crisis, even though the wider labour market has deteriorated by far less. Young people now account for around 40% of all unemployment, up sharply since 2021. That is close to the share seen during the aftermath of the 2008 crisis.

The problem goes beyond unemployment (Milburn, 2026). Nearly one million 16- to 24-year-olds in the UK are now not in education, employment or training, equivalent to around 13% of that age group. And the problem is entrenched, 45% of people who are NEET at the age of 24 have never worked.

Youth unemployment is not exceptionally high in the UK by international standards. Some other European countries have much higher rates. What stands out more is the gap between young people and the rest of the labour market. Relative to unemployment among prime-age workers, youth unemployment is unusually high in the UK and that gap has widened over time (Chart 16).

So the labour market is behaving very differently for young people than it is for the wider population. Job opportunities for young people have deteriorated by more relative to the national labour market than in other countries. The reasons are complex, but four factors appear particularly important: the low-hire low-fire pattern of adjustment; the concentration of weaker hiring in sectors that employ many young people; the deterioration in young people’s health; and the rising cost and perceived risk of employing inexperienced workers.

By definition, a low-hire, low-fire labour market hurts those entering work for the first time. The composition of labour demand has also moved against some of the routes through which young people traditionally enter employment. This is not simply the unwinding of the exceptional and temporary post-pandemic vacancy boom. Compared with immediately before the pandemic, vacancies are around 14% lower across the economy, but around 19% lower in accommodation and food services and 46% lower in retail. Job opportunities have therefore contracted by more in two sectors that provide important entry routes into work.

Chart 17 shows the distribution of job seekers and available jobs across industries, splitting job seekers into those who have never worked before and those who have. There is a relative undersupply of jobs in the accommodation, wholesale and retail sectors – the parts of the labour market with the highest concentration of inexperienced job seekers.

Those industries matter because they provide many of the starter jobs through which young people acquire workplace skills and start to build an employment history. Online job posting data shows a similar picture (Indeed, 2026). Graduate vacancies are around 7% lower than a year ago and at their lowest level for this time of year since 2020. Summer jobs, another common point of entry into employment, are also at their lowest level in four years.

Young people have experienced a greater deterioration in their health than the wider population, particularly in their mental health. The proportion of adults aged 16-64 meeting the criteria for a common mental disorder increased from 15.5% in 1993 to 22.6% in 2024. Among 16 to 24 year olds that increase was from one in six people to one in four (Fonagy Review, 2026). And they face further challenges. This cohort also entered adulthood carrying the effects of the pandemic. Disrupted education reduced learning and work experience, and school absence has remained higher than pre-pandemic levels.

The relative cost of employing younger workers has also risen because of changes to the National Living Wage. Minimum-wage rates for under-18s and 18 to 20 year-olds have increased faster than the adult rate. In the Bank’s DMP survey 29% of employers affected by the National Living Wage said they would lower overall employment in response.

The drivers of weaker labour market outcomes are difficult to disentangle. The evidence does not allow us to separate precisely effects of weaker demand for young workers from changes in their availability for work.

Through our Agency Network we regularly hear that uncertain demand and higher business costs, particularly energy costs, have depressed overall hiring since the Middle East conflict began. And that higher labour costs are incentivising firms to adopt automation and AI which is further pushing down on labour demand. But for young people in particular, the greater barrier may not be cost alone, but perceived risk. For an employer, taking on a young person involves uncertainty and investment. New entrants need training, supervision and we hear increasingly about work readiness and the need for pastoral support.

Those costs can be worthwhile for firms and transformative for individuals. But when firms face uncertainty, the risk of inaction in terms of pausing hiring new workers is often less than the risk of letting experienced employees go. That asymmetry has contributed to lower hiring of young people.

The human consequences are severe. Work and training provide income, purpose, skills, confidence and social connection. Losing access to them narrows opportunity and damages wellbeing. The effects can persist: unemployment early in working life leaves lasting scars on employment and earnings (Gregg 2001; Bell and Blanchflower 2011). The size of the affected population means these effects will have a macro impact for the wider economy.

A labour market that is less effective at moving people into work and towards more productive opportunities has implications not only for those directly affected, but also for wage pressure, productivity and the economy’s ability to adapt. I turn to those wider economic consequences next.

Macroeconomic implications

In headline terms, the labour market has proved more resilient than expected, during both the recent energy shock and previous shocks. But resilience in employment does not mean that the labour market is working well for everybody. A labour market can preserve existing jobs while dropping down a gear under the bonnet. Adjustment through weaker hiring rather than higher firing can conceal pressure on young people and new entrants. And it builds on trends of weaker mobility, poorer matching and slower reallocation.

This has three broad macroeconomic consequences.

Higher unemployment may do less to ease wage pressures

First, the interaction between the cyclical and structural stories matters for how we interpret the rise in unemployment.

The cyclical loosening in the labour market has been gradual, occurring mainly through slower hiring rather than abrupt layoffs. But it has happened against the backdrop of a labour market in which people are moving less readily between jobs, industries and skill levels. The combination of those factors presents a risk that jobseekers are not matched quickly enough with the opportunities that are available.

That risk is greater because the slowdown in labour demand has been uneven. Vacancies have fallen much further in some sectors than in others. Adjustment therefore requires more workers to move between industries and occupations at a time when such moves have become less common.

In a more mobile labour market, imbalances between labour demand and supply can be resolved more quickly. Displaced workers find new jobs and compete for available positions. Unemployment then represents a readily available pool of labour, which restrains wage growth.

But if jobseekers are not well matched to the vacancies available, a given rise in unemployment may exert less downward pressure on wages. Unemployment could remain higher for longer without producing the easing in wage pressure that we would ordinarily expect.

That may help explain why wage growth has remained relatively strong even as vacancies and other quantity indicators have weakened. As I showed earlier, pay growth has fallen substantially from its post-pandemic peak. But underlying wage growth remains at the upper end of the range consistent with the inflation target, while forward-looking indicators point to continued strength into next year. That degree of wage pressure sits somewhat uneasily alongside the apparent weakness of the labour market. It reinforces the case for looking closely at whether all the weakness in the quantity indicators reflects a general shortfall in labour demand, or whether some reflects poor matching between jobseekers and the opportunities available.

In the language of our forecast, a persistent reduction in the labour market’s ability to match people with jobs would mean a higher equilibrium unemployment rate (or NAIRU), at least in the short-term. The gap between unemployment and its equilibrium rate would then be smaller than the rise in headline unemployment would suggest.

Given the relatively moderate rise in unemployment so far, I do not think that the equilibrium has changed materially. But these dynamics affect how I interpret the loosening we have seen. If unemployment were to rise further, I will pay close attention to whether that is easing wage pressures or instead revealing the limits of the labour market’s ability to move people quickly into new jobs.

Weaker mobility can reduce productivity and growth

Second, slower labour market reallocation can lower productivity and growth. This has long been recognised in the academic literature. Reallocation enables workers to move to the firms, sectors and occupations where their skills are most highly valued.

Moving up the job ladder is also an important source of earnings growth. Job switchers tend to receive significantly larger pay rises than those who stay with the same employer (Chart 18). The prospect of workers moving can also improve the bargaining power of those who stay, because retaining staff becomes more difficult for employers when alternative opportunities are readily available.

At the macroeconomic level, job moves can generate productivity gains by producing better matches between workers and employers. Where pay rises reflect those gains, they do not create additional inflationary pressure. Research suggests that moves near the bottom of the job ladder are primarily productivity-enhancing, while moves near the top are more likely to add to wage pressure without generating equivalent productivity gains (Moscarini & Postel-Vinay, 2023).

When reallocation slows, matching deteriorates, human capital accumulates more slowly and productivity suffers. Declining mobility may therefore reflect a labour market that has become less effective at allocating people to the work in which they can be most productive. Allocative efficiency plays an important role in long run productivity growth. Decker et al., (2017)

Lower turnover is not unambiguously bad. A slower pace of structural change can bring with it fewer involuntary job losses (Cominetti et al., 2022). It is also possible that turnover is lower because better recruitment technology may have reduced the number of very short-lived and unsuccessful employment relationships. Some research has argued that much of the decline in turnover reflects fewer jobs lasting only one or two quarters (Pries and Rogerson, 2022).

That may be part of the story, including for the UK. But it sits uncomfortably with wider evidence that matching efficiency has if anything declined over time (Barlevy et al., 2024, or Pizzinelli and Speigner, 2017 for the UK). And with newer evidence suggesting that fewer moves do not simply reflect workers becoming more content to remain where they are. Workers report lower job satisfaction and fewer opportunities for progression (Slaughter, 2026). Workers still want better jobs, but a less mobile labour market is providing fewer opportunities to obtain them.

Lower productivity resulting from a less mobile labour market would reduce the economy’s supply capacity over time. That would mean lower growth and, for a given level of demand, greater inflationary pressure. Of course, all else is not equal. How the longer-term effects of lower mobility will sit against the potential productivity gains from the adoption of AI across the economy is uncertain.

The long-term effects on growth are particularly important because they are concentrated among young people. Missing the first steps on the job ladder delays the accumulation of skills and experience. If a large cohort remains detached from work for longer, future labour supply and productivity will be lower^{footnote [3]}. A cyclical shortage of entry-level jobs can therefore become a structural loss of productive capacity.

The Government has recognised these challenges and is considering the interaction between young people’s health, skill, work and opportunities. Addressing them would strengthen the economy’s productive capacity and resilience. The forthcoming report from the Alan Milburn’s Young People and Work Review will be an important milestone.

A less mobile labour market is less able to adapt

The third macroeconomic consequence is that lower mobility makes the economy less resilient to shocks and structural change. A fluid labour market absorbs change by moving people from contracting activities towards expanding ones. A rigid one turns change into longer spells of unemployment, lost skills and persistent mismatches.

That matters particularly at a time of rapid technological change. Over time, artificial intelligence is likely to alter the demand for labour across sectors, the tasks people perform and the skills employers need. The eventual adjustment is therefore likely to involve more than simply fewer jobs in certain occupations (Acemoglu and Restrepo, 2019; Autor and Thompson, 2025).

The contribution of AI related sectors is increasingly visible in UK growth. Bank estimates show that AI-related activity has accounted for around 0.5 percentage points of GVA growth since the beginning of 2026.^{footnote [4]} After allowing for these sectors’ usual contribution, the excess is around 0.3 percentage points – just under one fifth of total GVA growth (Chart 19). Using a more recent historical benchmark gives a similar estimate. This suggests that AI-related activity is becoming increasingly apparent in the UK growth data.

This is consistent with the results in the Bank’s DMP survey. Business leaders are increasing their investment in AI, and expect AI to raise productivity meaningfully over the next few years. They expect this effect to be split broadly equally between increasing output and reducing employment.

Again the effects may fall disproportionately on young people. Some emerging evidence points to weaker hiring of younger workers in occupations most exposed to AI (Brynjolfsson, Chandar and Chen, 2025), although other changes, including remote working, may also be contributing (Lambert and Schindler, 2026).

Adjustment through fewer vacancies rather than redundancies may be less visible, but it is not necessarily less consequential. If AI takes on tasks traditionally performed by junior staff, it could remove not only an entry-level job but also the opportunity to learn, acquire experience and progress to more skilled and higher-paid work.

On the flip side, there is a more positive possibility. Future cohorts of young people may become more sought after as AI changes the nature of work and demand shifts to AI native employees. But that will depend on the skills they acquire and the opportunities available to put those skills to use.

For the wider economy, the nature and timing of the adjustment matters as much as its eventual scale. If productivity gains come through quickly, they could support growth and real incomes while easing inflationary pressure. But if displacement comes first, while investment, new tasks and retraining take longer, the transition could involve higher unemployment without a corresponding easing in inflation. Who gains matters too. If lower income workers do not benefit, then overall demand may go down. Higher income gainers tend to save more of their income.

Even if new jobs ultimately offset those displaced, they may not emerge at the same time, in the same places, or require the same skills. The labour market’s ability to adapt and move people towards new areas of opportunity will be critical. Greater labour market mobility would reduce the risk that the pace of technological change outruns the capacity of workers and firms to respond.

Monetary Policy

Since I’m speaking at the Bank of England Watchers Conference at a challenging time for inflation, I want to finish by setting out how I am thinking about policy and the outlook following last weeks’ decision.

The framework we are using to think about the impact of the global energy price shock remains the same (Table A). The shock affects inflation through four main channels: direct, indirect, second-round and demand effects. Those channels operate at different speeds and can point to different policy responses. The appropriate stance of monetary policy depends on their balance.

We continue to learn about both the size of the shock and the way it is passing through the economy. Since our assessment in July the energy price shock has proved larger and more persistent. That has increased the direct pressure on inflation, raised the prospect of greater indirect effects as higher costs pass through supply chains, and increased the risk of second round effects through wages, prices and inflation expectations.

At the same time, the offset from weaker demand has so far been smaller than expected. Household spending and business activity have held up better than I anticipated. That resilience is welcome. But it means the shock is producing less disinflationary pressure through the demand channel.

The balance of the four channels has therefore become less favourable for inflation. The inflationary effects of the shock are larger, while its disinflationary effect through weaker demand has been more limited.

Table A: Global energy shocks can affect UK inflation through several channels 

Taking these developments together, at our meeting last week, the MPC voted by a majority of 8–1 to increase Bank Rate by 25 basis points to 4%. The scale and duration of the energy shock have generated sufficient inflationary pressure in the pipeline to warrant an increase now. The resilience in activity pushes in the same direction for inflationary risks.

Looking forward, the appropriate path for policy will depend on how the shock continues and how it propagates through the economy. As our scenario analysis illustrates, a variety of alternative paths for the economy are plausible, with different policy implications.

We must be mindful of the uncertainties and risks. The absence of second round effects may be telling us something about the ability of businesses to pass on costs in a weakening economic environment. There is a risk that households and firms exhaust their short-term means of adjusting to cost pressures as the cumulative effects of higher energy costs and tighter financial conditions build. Consumer demand and business activity could then weaken, increasing unemployment and reducing inflationary pressures. The evidence of economic resilience to date sits counter to this, but we cannot dismiss this risk, and it forms the basis of our downside scenario.

My own judgement is that, at present, the balance of risks to inflation remains to the upside. The energy shock has lasted longer and become larger than expected, activity has been more resilient and recent data suggest that underlying disinflation may have stalled. The conditions for persistent inflationary dynamics over the medium-term are crystallising. That warranted an increase in Bank Rate last week. In the absence of evidence of weaker inflation pressures, I continue to think there is likely to be a case for a further increase in rates in the near term.

But that judgement is conditional. This is not just about the conflict and its consequences for energy costs. But also how those feed through to price and wage setting more broadly across the economy. The policy judgement does not depend on a single indicator or variable. It will depend on how the combination of data on pricing, activity and slack come together alongside energy and related prices.

That, together with the considerable uncertainty around the outlook, is a reason to proceed gradually rather than trying to anticipate today the full policy adjustment that will ultimately be needed to return inflation sustainably to target.

Conclusion

Let me conclude.

The UK labour market has been more resilient than expected. Unemployment has risen only gradually and redundancies remain relatively low. But that aggregate resilience conceals a more troubling picture. Firms are hiring less, people are moving between jobs less often and young people are finding it particularly difficult to gain a foothold on the job ladder.

That matters for monetary policy. If those looking for work are poorly matched to the jobs available, higher unemployment may do less to ease wage pressures than it has in the past. It also matters for the economy’s longer-term prospects. A less mobile labour market makes it harder for people to find work in which their skills are best used, weakening productivity and growth.

And it matters for the economy’s ability to adapt. Artificial intelligence could bring substantial gains in productivity and living standards. But realising those gains will require people to move towards new tasks, new skills and new opportunities. If that adjustment is slow, the costs could arrive before the benefits.

The labour market is not standing still. Millions of people still find work and change jobs every year. But it is moving more slowly. Too many people are struggling to get in or move on. The challenge is not simply to keep unemployment low. It is to help the labour market shift up a gear.

In setting monetary policy, I will continue to look beyond the headline numbers: at whether hiring is recovering, whether people can move towards the jobs available, and whether labour market weakness is genuinely easing wage and inflation pressures. That evidence will shape my judgement of how much further monetary policy needs to go to return inflation sustainably to the 2% target.

I would like to thank Maria Balgova, Fabrizio Cadamagnani, Stuart Campbell, Gokce Karasoy Can, Douglas Rendle, Michal Stelmach, Jagdish Tripathy, Hassana Babangida and Brad Speigner for their work in preparing this speech. I am also grateful to Adrian Paul, Chris Young, Harvey Daniell, Ambrogio Cesa-Bianchi, James Bell, Ken Clark, Swati Dhingra, Sarah Breeden, Rohan Churm and Rebecca Maule for helpful comments.