Changing jobs?

Policy debates are dominated by strongly held misconceptions about economic change such as: it’s speeding up; it takes place mainly by workers in shrinking sectors losing their jobs; and that it undermines job quality for some via a ‘polarisation’ of the labour market into bad and good jobs. There certainly are bad outcomes for those workers who do face involuntary job loss, and periods of faster economic change have been associated with higher rates of such job losses. But we need a richer understanding of how economic change happens, especially if we are to make a success of navigating the faster change we expect to see in the 2020s. In particular, we highlight three key facts about the UK’s recent experience of economic change:

• Structural change (the reallocation of labour across different sectors) has been slowing down in recent years, not speeding up.[1] Alongside this, the rate at which workers move between jobs and sectors has also slowed down.
• Sectors can decline through older workers leaving and fewer younger workers joining, as well as through workers in the middle of their careers being forced out. To a significant extent, and as far as it is possible to tell with the available data, this was the case for the fall in manufacturing employment since the 1980s. Involuntary job losses do happen, however, particularly in declining sectors, and these can have serious negative repercussions for the workers involved.
• In recent years, and in contrast to some other countries, occupational change has tended to involve ‘upgrading’ (the growth of higher paid occupations) more than it has ‘polarisation’ (the growth of high and low paid jobs), especially for women.

[1] Of course, there are other ways in which the labour market is changing besides the sectoral composition of employment. The growth of alternative forms of employment (such as zero-hour contracts and self-employment) and market concentration are also important. One justification for focusing on sectoral reallocation here is that this is the type of change we might expect to be triggered by the ‘shocks’ facing the country in the 2020s.

There have been significant changes in the sectoral composition of the UK labour market in the past half century. The most significant shift has been the fall, in absolute terms and as a share of jobs, of manufacturing, and the rise of several service sectors, including business services. Employment has also grown significantly in parts of the public sector – especially in healthcare and education. In 1970 there were 7.7 million jobs in manufacturing, accounting for 29 per cent of the total. By 2021 that number had fallen to 2.5 million jobs, 8 per cent of the total. Over the same period, the number of jobs in professional services, education and health rose from 3.7 million (14 per cent of the total) to 10.6 million (31 per cent of the total). Other service sectors which have seen significant jobs growth are hospitality and administrative services. The structural transformation from manufacturing to services was driven by technical change and the automation of manufacturing jobs; by globalisation, which meant manufactured goods could be more readily imported and production moved overseas; and more generally by changes in consumption patterns.

Alongside sectoral change there has also been occupational change, where the main trend has been one of occupational ‘upgrading’, meaning jobs growth has been highest in higher-paying occupations. In the 1980s and 1990s this was accompanied by growth in low-paid occupations, giving rise to a story of jobs ‘polarisation’. Since the 2000s, jobs growth in the lowest-paying occupations has been negative, meaning ‘upgrading’ rather than ‘polarisation’ is a better description. The consistent trend, though, has been stronger jobs growth of higher-paying occupations. Occupational upgrading has occurred for both sexes but has been more pronounced for women than men. Total female employment grew by 4.3 million between 1992 and 2019, almost all of which (3.9 million) is accounted for by employment growth in the ‘top’ three occupation groups (managers and directors, professional and associate professional occupations).

It may feel like change is speeding up, with stories of ‘robots taking our jobs’ appearing in the press, and new ways of work, such as the gig economy, emerging. But when it comes to the sectors we work in, which ultimately stems from consumption patterns, productivity and the types of goods and services the UK produces, the pace of change has been slowing down. The 1970s and 1980s was a time of rapid change – this was when the shift from employment in manufacturing to services was at its most intense. But since then, the rate at which different sectors have grown and shrunk (underneath changes in overall employment) has trended downwards. In 2021, the reallocation of labour across 21 industry sectors, compared to a decade ago, was equivalent to 7 per cent of total employment. This is about one-third as high a rate of reallocation as the 1980s peak. Looking further back, the period around World War 2 was also a period of fast change. Although we don’t have data as far back as we do for employment by sectors, the pace of occupational change measured by the dispersion in employment growth rates doesn’t appear to be slowing down. This suggests that occupational upgrading isn’t wholly dependent on sectoral change and occurs in part within sectors.

Structural change – the growth of employment in some sectors at the expense of others – can happen in a number of ways. We might first of all think about workers moving from shrinking sectors (like manufacturing) to growing sectors (like professional services). Another route to the reallocation of labour is via the net effect of workers entering and exiting work altogether. If more workers enter employment (from non-employment) than leave employment (to non-employment) in a given sector, employment in that sector will grow even without any direct job moves between sectors. We find that the movement of workers between jobs in different sectors only accounts for around one-third of the fall in manufacturing employment from the 1980s to the 2000s, the rest coming from the net effect of workers joining and leaving the workforce.

Entry and exit can be further separated into the entry and exit of workers in the middle of their working lives (‘within-career entry/exit’) and the entry of young workers and the exit of older workers (‘lifecycle entry/exit’). We find that in the 1980s to 2000s, lifecycle entry/exit played a bigger role in the fall in manufacturing employment than within-career entry/exit – this goes against the idea some may have that such sectoral decline happens via mid-career workers being forced out of their jobs. In professional services, where employment has grown since the 1980s, the picture is more mixed. Employment growth has been driven by different factors (job moves, and lifecycle and within-career entry and exit) in different periods. Overall, job moves played a larger role in the growth of employment in a growing sector – professional services – than they did in the fall of employment in manufacturing.

Worker job mobility has slowed down, roughly in line with the slowdown in the rate of structural change

Even if not the main way in which labour reallocates across sectors, job mobility does make an important contribution to sectoral reallocation, particularly outside of downturns. And it’s important for other reasons too: it is a route through which workers can improve their pay, as we will later show. Despite the measurement challenges of long-term time series, we can confidently say that job mobility rates are lower now than they were in the 1980s. We have better quality data from 1992 onwards, when it becomes possible to measure high-frequency changes in labour market status. Looking through the ups and downs of the economic cycle suggests the fall in mobility has been significant. In 2000, 3.2 per cent of workers moved jobs per quarter, and 1.7 per cent of workers moved jobs to a different sector per quarter. In 2019, the rate of job mobility was 2.4 per cent, 25 per cent lower than in 2000, and the sector mobility rate was 1.1 per cent, 35 per cent lower than in 2000.

It is not clear what is behind this fall. Compositional factors only play a small role (more on which below). The returns to moving jobs, compared to staying put, have also not changed dramatically during this period, suggesting workers aren’t responding to changing incentives. The incentive to move job and home has likely been dampened due to faster rising housing costs in higher paying areas, but job-and-home moves only account for a small minority of all job moves. One factor might be the rules and institutions governing the labour market, although it’s not clear why the changes in this period (towards weaker regulations around hiring and dismissal, and curbs on union power) would have led to slowing job mobility. A reasonable hypothesis is that the rate at which workers move jobs is related to the pace of economic change, and indeed the slowdown over the past three decades in the pace of structural change and the rate at which workers move jobs have been similar. It might therefore be that workers are moving jobs less because there is less demand for them to do so, although establishing a causal link between these two trends would require further analysis.

Compositional changes in the workforce – specifically the fact that it is becoming older – have had a small negative effect on the average mobility rate. This is because older workers have lower rates of job mobility than younger workers. In 2015-19, 5.8 per cent of 16-24-year-olds moved jobs per quarter, five times the rate of 55-to-64-year-olds (1.2 per cent). There is an even bigger age skew when it comes to sectoral job mobility: 3.5 per cent of 16-24-year-olds move job to a new sector each quarter, compared to 0.5 per cent among 55-to-64-year-olds.

This may be a cause for concern if structural change is set to accelerate – an older workforce is one that is less likely to make voluntary job moves, which may in turn inhibit the pace at which the labour market adjusts to change, or lead to a bigger role for involuntary rather than involuntary moves. However, the effect of compositional changes on aggregate job mobility has been small, and only accounts for one-fifth of the slowdown in job mobility since the turn of the century. This suggests that composition of the workforce won’t be a significant impediment to faster labour market change.

Age is not the only factor along which the rate of job mobility varies – other features such sector and occupation also matter. For example, workers from hospitality and retail move jobs at around three times the rate of workers from public administration. Together, in 2015-19, workers from hospitality and retail accounted for 26 per cent of all job moves and 32 per cent of all sector moves, but only 19 per cent of all employment. There are also big differences in the mobility of workers at different occupation levels: workers from lower-paid occupation groups (‘elementary’ and ‘service and care’ occupations) move jobs at twice the rate of workers from the high-paid ‘managers and directors’ group. Such high rates of job and sector mobility show that mobility is partly about ‘churn’ between jobs and sectors, and not always related to structural change.

Young workers are more likely to change occupations when they move jobs, but when older workers do change, the changes are bigger

As well as measuring the rate at which workers move jobs, we can also, by looking at the task content of the jobs workers move between, capture the ‘distance’, in terms of task similarity, of job moves being made. This provides some important insights. Looking at all the job moves made over the past 20 years, a third (36 per cent) were to a job in the same occupation, hence involving zero ‘distance’. For those moves that involve a job with new tasks, we might expect the relationship to be straightforwardly negative, with fewer moves made the larger the ‘distance’ between jobs. However, that is not the case. Instead, the most common move is not over the smallest possible distance but over a medium distance – equivalent, say, to the difference (in the task content of their jobs) between a nurse and a factory worker. A significant minority of moves are over a larger ‘distance’: of job moves involving a change of occupation, one-third involve moving a distance (in task content) equivalent to the difference between a nurse and a solicitor.

There are striking differences in the types of job move made by old and young workers. Younger workers are more mobile in terms of changing occupations than older workers, but when older workers do change occupations they are more likely to move to occupations with significantly different tasks. Young workers are much more likely to move occupations when they move jobs – from 2002 to 2020, 70 per cent of 25-year-olds’ job moves involved changing occupation, compared to 55 per cent of 60-year-olds’ moves. But when older workers do move occupations, the distance moved doesn’t have the relationship with age we might expect. From 2002 to 2020, the average job distance of an occupation mover was highest for those age 25 to 35 but also for those age 55 and above (and low for those in their early 20s, as well as for those aged 40 to 50). The fact that some workers do manage to make significant occupational transitions later in their careers is encouraging, boding well for our ageing society’s ability to cope with periods of faster structural change.

Labour market change can be positive for workers, with those moving jobs and sectors seeing greater pay growth than those who stay put

An important question, as we face up to a decade that is likely to bring greater levels of economic change, is how workers experience this change. It’s important to recognise that there are positives. Most obviously, workers enjoy much stronger pay growth when they move jobs than when they stay put. On average, from 1975 to the present day, individuals who moved jobs enjoyed typical pay growth 4 percentage points higher than individuals who stayed in the same job. The ‘movers’ bonus’ of transitions tends to rise and fall with the economic cycle (it’s lower in downturns), but beyond the cycle it has been remarkably consistent. The movers’ bonus is even higher when workers move to a job in a different sector or region. This suggests that economic change, and changing patterns of demand for workers across sectors, is likely to create opportunities for some workers to move to where demand and pay is higher (albeit where they may also encounter higher housing costs). Whether UK workers possess the right level of transferable skills that will enable them to make the most of these future opportunities remains an open question that has recently received significant attention from both research and policy and will be further explored in future reports of this Inquiry.

But there are also downsides to a changing labour market: previous periods of faster change came with higher rates of involuntary job loss

But there is of course also a downside to economic change: not all workers in shrinking sectors will make pay-enhancing moves elsewhere – some face involuntary job exits. These are very negative experiences for workers. They are clearly painful experiences in and of themselves, but we find that they also have further negative effects. Workers who experience involuntary job loss will take longer to return to work than those leaving work voluntarily (half of those experiencing involuntary job loss have returned to employment after six months, compared to two-thirds of those leaving their previous job voluntarily). Moreover, when they do return to employment, it is on average to a job which pays less than the one they left. From 1995 to 2020, median real hourly pay growth was -1.1 per cent among those who had experienced an involuntary period out of work within the past year, compared to 2.1 per cent among all workers.

This partly reflects the nature of workers’ job changes. We find that 40 per cent of workers in declining sectors who experience involuntary job loss either return to work in the same sector, or in another declining sector. We also find (using the same job distance analysis mentioned above), that those returning to work after involuntary job separations typically move ‘away’ from jobs placing emphasis on analytical and personal tasks, and ‘towards’ jobs involving manual work. This is against the grain of workers making voluntary job moves, where the direction is, on average, towards more analytical jobs.

Importantly, previous periods of faster change have come with higher levels of involuntary job loss. In the early 1990s, 0.8 per cent of all workers faced involuntary job loss per quarter, compared to 0.4 per cent of workers per quarter in the pre-Covid period. Rates of involuntary job loss were particularly high in declining sectors in the early 1990s. This suggests the pace of structural change really matters for the volume of these negative outcomes – a warning sign if, as expected, change speeds up in the 2020s.

What does the above analysis of past periods of economic change tell us about what we can expect in the 2020s? The combined effects of Brexit, Covid-19, and the transition to net zero are likely to trigger a greater level of structural change than we have seen in recent years. Future Economy 2030 Inquiry research will attempt to quantify this – should we expect structural adjustment on the scale of the 1980s? More? Either way, with the country’s trading relationships fundamentally changing, with Covid-19 leaving lasting impacts on how and where we work and consume, and with the next phase of the net zero transition set to demand changes in consumption patterns as well as investment in carbon-saving activities, the rate of structural change is likely to increase in the coming years.

One positive lesson from the above is that the rate of sectoral structural change and job mobility appear to be related – when significant industrial reallocation was happening faster, so was the rate at which workers moved jobs. Although the current low rates of change and mobility might suggest an economy ill-prepared for faster change, it’s reasonable to expect that rates of worker mobility will rise again as the need for labour reallocation increases.

The negative lesson from the above is that faster economic change is likely to lead to greater levels of job separations. While some workers will be able to turn economic change to their advantage and make pay-enhancing job moves, many won’t. And involuntary job losses are costly for workers, both in the immediate sense, but also because those experiencing them are likely to make ‘bad’ job moves on re-entering employment. As discussed above, sectors can grow and shrink in a number of ways – broadly, through job moves, through the entry and exit of workers in the middle of their working lives, and through ‘natural’ entry and exit. ‘Natural’ entry and exit (the balance, at sector level, of young workers joining, and older workers retiring) is clearly the least painful way of managing change, in that it minimises the risk of workers being pushed out via involuntary separations.

Future reports within the Economy 2030 Inquiry will explore in detail how policy makers should go about managing what we expect to be greater levels of change this decade, where the challenge will be to promote the positive opportunities such change presents some workers, while mitigating the negative impact on others.