Labour market Understanding the labour market: pandemic not pandemonium The labour market is normalising, not overheating 28 June 2021 Torsten Bell Kathleen Henehan Krishan Shah Hannah Slaughter Greg Thwaites The mild euphoria phase of the economic commentary cycle has arrived somewhat earlier than after previous downturns. This is a big change from the excessive pessimism of late 2020. According to some, we have “eye popping growth” to look forward to as “Brexit Britain Booms”. For the labour market, this turn to optimism has seen a swift shift away from people worrying about unemployment. Now the most discussed ‘danger’ is that we have a tight labour market where it is hard to hire. Headlines like “A shortage of workers is driving up wages” are common. While some worry these labour shortages are holding our recovery back, others claim this is the beginning of a welcome new era of worker power. On a quick count, there’s already been 50 articles referencing “labour shortages” in the national media so far in June, with warnings that this may affect everything from meat production to construction to cannabis growing. But while some firms, especially in hospitality, are facing difficulties recruiting, the real danger is of us losing the plot. The only bigger nonsense than thinking this is in aggregate a tight labour market is believing a tight labour market would be a bad thing. Crucially, most of the unusual developments we are currently seeing reflect one big fact that rather risks being lost: we’re in a pandemic. The key features of a labour market during a pandemic, as opposed to just any old recession, are that the impacts are very sectorally unequal (focused on sectors that involve social contact) and that falls (and then rises) in activity happen much more rapidly. Together these two features mean labour market rune-readers should expect a bumpy ride; focus more than usual on levels rather than rates of change; and avoid drawing general conclusions about the labour market from specific sectors. Caution should also be exercised in interpreting aggregate data that can be distorted by the speed and/or unequal nature of the pandemic’s effects. These issues are just as important in pondering the strength of this year’s recovery as they were when faced with last year’s recession. The good news The good news is that there is a lot of good news about labour market-wise. Last month, 197,000 more employees were on payroll than the month before, with the result that half of the pandemic damage to employee numbers has now been undone. Vacancies now exceed pre-crisis levels (Figure 1). Figure 1: Vacancies are now higher than pre-crisis levels Analysis from a new survey commissioned by the Resolution Foundation and conducted online by YouGov in the first week of June 2021 also shows that 80 per cent of those on furlough during the first and second lockdowns have now returned to work (see Figure 2).1 Moreover, 69 per cent of those who were still furloughed in October 2020 (the trough for furlough rates in between lockdowns) have also returned. This group is who we might expect to be the least likely to return to work post-furlough. So, the rate of return is encouraging. Figure 2: The majority of those who were furloughed have now returned to work Alongside this, the unemployment rate has been falling and employment rate rising. The former now stands at 4.7 per cent, down from its peak of 5.1 per cent at the end of 2020. The 2021 unemployment rates are, so far, much lower than forecast back in 2020, thanks to the extension of the Job Retention Scheme (JRS) and the fact that we’ve managed to produce more GDP during the pandemic than previously thought. Is the labour market tight? This is all great news. Unemployment of 4.7 per cent is what full employment looked like in the mid-2000s, so maybe it is not surprising that some talk as if this is now a tight labour market. There are also some other signs of patterns that we would expect to see in a tight labour market world, including the talk of worker shortages and data showing significant wage growth. As Figure 3 shows, average weekly earnings were up 5.6 per cent in the 3 months to April compared to a year earlier, the fastest increase this millennium. Figure 3: Earnings appear to be growing rapidly Dig beneath these headlines though, and remembering there is still a pandemic on, and it is clear that although the labour market is tightening, it is very far from tight. A key feature of a tight labour market is that we would all be doing lots of work. This is absolutely not what is happening. Total hours worked in the UK are still down 5 per cent on pre-crisis (Figure 4). Talk of a hospitality boom is focused on the rates of change in activity levels, but if we look at levels instead, it is clear hours worked in the sector are hugely down. Figure 4: Hours worked point to an economy with slack The number of hours worked still looks much more like a recession than a full recovery. And it is the right measure of labour market health in a period where the JRS makes traditional measures of employment and unemployment far less useful (those furloughed count as employed even if doing no work). For those who prefer to think in terms of people not hours, we can calculate a ‘Covid employment gap’ – made up of furloughed staff, plus the fall in employees and self-employed workers since the start of the pandemic. This reinforces the idea that the labour market is recovering not recovered. As Figure 5 shows, the gap more than halved between February and May 2021 to fall below three million for the first time since the start of the crisis, and more recent furlough estimates suggest that the gap stood at around 2.3 million in early June (before accounting for any further changes to employment since the latest HMRC and ONS data). But this still means there are over 2 million people who are not working as they were pre-crisis. Our survey also shows that 10 per cent of those who were self-employed pre-crisis are not working now. Figure 5: Over two million people are not working as they were pre-crisis On pay, while average pay growth is strong, the key measure of a tight labour market is actual people getting pay rises. The data hugely overstates how much of that is going on for two different reasons: a compositional effect and a base effect. The former stems from the fact that low-paid workers have been much more likely to be hit by this crisis than higher- paid ones – either by losing work or not getting a job in the first place. This pushes up measures of average pay without anyone earning any more. Think of a world with ten workers in it, five earning £1 and five earning £2. If three of the low earners lose their jobs (say because the hospitality sector is shut down), average pay for the remaining seven workers rises by 14 per cent, but none of them have actually seen so much as a 1-pence pay rise. These compositional effects have accounted for around half of the measured pay growth at times during this pandemic, although that effect should weaken as the economy reopens.2 The base effect, however, is actually strengthening recently. Annual measures of average pay growth depend not just on what average pay is today, but what it was this time last year. And this time last year average pay growth fell like a stone to zero, or even negative on some measures, as the pandemic hit and furlough saw millions of workers’ wages cut by 20 per cent. Because of this temporarily low baseline, measures of annual pay growth a year on have surged. But this kind of base effect does not tell us that bumper pay rises are taking place right now. In order to (partially) address the compositional effect we can focus on the typical pay growth individuals have seen (see Figure 6). And to look through last year’s temporarily slumped pay levels we can examine pay growth over two years. On this latter measure pay growth is relatively weak, down by around one-third (from just over 3 to just over 2 per cent) from pre-pandemic levels. Different pay measures show different results, but taking them together it’s clear that headlines implying pay is currently rising faster than it was pre-crisis are misleading.3 Figure 6: Pay growth is weaker than it appears To examine how widespread are the difficulties firms face in finding staff we can also look at measures of the time it takes to fill each vacancy (see Figure 7, where we have nowcasted the most recent datapoint to provide a more timely, albeit more uncertain, estimate).4 In an improving labour market we should see more vacancies and it taking longer to fill them, which broadly is what is happening. But that increase is from a very low level last year. While it is unsurprising that firms facing challenges finding staff are vocal, the reality is that overall it remains significantly easier to fill a vacancy today than in 2019 (when we certainly did have a tight labour market). Figure 7: Vacancies are still easier to fill than they were pre-crisis Taken together, the evidence is fairly conclusive that our labour market is improving but far from recovered. Claims of a new dawn for worker power with widespread labour shortages are wide of the mark. Indeed, recent experience in the UK and in other flexible labour markets should teach us that a labour market needs to be very tight indeed for workers’ bargaining power to materialise. What’s going on? A pandemic. If we do not have a tight labour market in aggregate, how do we explain the challenges some firms are facing to recruit? By recognising that sectorally concentrated hiring challenges is exactly what we should expect as we emerge from the pandemic. The totally exceptional turning on and off of sectors, as human behaviours and government regulation respond to high virus prevalence, will see more acute short-term staffing challenges in the likes of hospitality than we see in a normal labour market recovery. To understand why, think back to when you last experienced a fire alarm test. Work is switched off as everyone troops out of the building and mills around outside for a while. Then the decision is made to restart work. Everyone goes to head back into the building and you get a queue forming at the entrance. That queue forms despite there being a perfect match between the supply of workers and demand for those workers (i.e. desks), because it takes time to overcome the entrance bottleneck and get everyone back to work. That might be frustrating for the employer, but it does not show that an economy-wide labour shortage exists. This is a big part of what is happening today. Hospitality firms are swiftly moving from no economic activity to lots of economic activity. But finding workers is not the same thing as switching the restaurant’s electricity back on. Bottlenecks form because workers are not necessarily in the same places as the firms that want them; may not be ready to start straight away; and the ones available may not be a great match for the specific job. These frictions in what economists call matching of workers to vacancies put a speed limit on aggregate employment recoveries after normal recessions. But when a pandemic tries to pack a year’s worth of a normal recovery into one week for a specific sector, it is not surprising those speed limits really bite and feel like acute labour shortages. In fact, the economics of pandemics are riddled with bottleneck problems well beyond the labour market exactly because of swift behavioural change.5 Current crunches reportedly include cardboard boxes and warehouse space as online retail surges. Scrapping furlough is not the answer to labour supply worries The pandemic has not only driven these weird patterns of labour demand, it is also affecting labour supply. Some argue that the existence of the JRS is the big problem, with workers happy to earn 80 per cent of their previous wages on furlough rather than fill one of the new vacancies being advertised. But while the existence of the furlough scheme will have some impact on the incentive for such workers to apply for new jobs, it is hard to conclude it is the principal cause of sectorally-focused hiring challenges. To begin, that is because the scheme very much puts employers in charge: they, not the employee, decide when someone should come back to work. Hence the swift falls in furlough rates in reopening sectors in recent months, and the fact that the majority of those remaining on the JRS are now partially furloughed (i.e. doing some work). The scheme is also about to enter its final three months, which is very material to any impact it might have on labour supply. In three days’ time, firms will have to cover 10 per cent of the wages of their furloughed staff. This will strengthen their incentive to bring back any workers where there is demand to be met, or release employees where that is unlikely to be the case. Workers’ incentive to stay furloughed will also fall as the scheme nears its end: if you do not think you have a job to go back to, the promise of one or two months of 80 per cent wages is not a reason not to take a new permanent job, unless you have a very high discount rate. There are other pandemic-related impacts on labour supply that will also be playing a role. The fact that we are currently in a third wave of this pandemic will have an impact on the desirability of jobs in hospitality requiring close contact. Even those of us who enjoyed years of work in pubs in our youth might be less keen on a job behind a bar right now. And the people most likely to work in that sector are also exactly those least likely to be vaccinated (see Figure 8). Thought of in this regard, the pressure to raise wages in hospitality could be seen as a form of danger pay rather than telling us very much about aggregate labour market tightness. Figure 8: Those most likely to work in hospitality are also the least likely to be vaccinated Consistent with a picture of the relative attractiveness of sectors having been, at least temporarily, changed by the crisis, we see from our own survey that those who previously worked in the hardest hit sectors (hospitality, retail and leisure) are more likely to have taken jobs elsewhere since the crisis started (see Figure 9). And while they are more likely than others who have moved sector recently to want to move on again, that is disproportionately a wish to move to yet another different type of work than back into the likes of hospitality. Figure 9: Hospitality workers are more likely to have moved sector and want to move again Significant falls in migration during the pandemic have also meant a reduction in effective labour supply. While initial reports of 1.3 million migrants having left overstated the case, our own analysis and that of the ONS using HMRC data points to a possible reduction of mid-hundreds of thousands. In the long term, fewer migrants reduces demand as well as supply for labour in our economy, but in the short term this has clearly contributed to difficulties hiring for sectors most reliant on such labour. This includes haulage as well as hospitality. The fact of such shortages should not come as a surprise. The sectors currently finding it difficult to hire are exactly the ones we identified as needing to make significant adjustments (be that via higher prices, lower output or business model change) following major changes to the UK’s migration regime post-Brexit. The pandemic’s effect has been to speed up the process through which changes in patterns of migration feed through to the labour market, causing shortages via this route rather than because the labour market as a whole is tight. Of course, labour supply can rise as well as fall. The financial crisis taught us an important lesson in that regard: by making us all poorer, recessions can also boost labour supply as people (generally women) choose to work more to compensate for income shocks. There are signs that this major feature of the 2010s, enabled but not caused by our flexible labour market, has continued during the pandemic, with a large three percentage-point rise in the share of women in full time employment since February 2020. The recovery has a long way to go While an exact quantification of labour supply changes is very hard at present, it is very clear that our recovery is far from complete. The Bank of England believes the economy currently remains 2.5 per cent below its pre-crisis level. That is small compared to the 20 per cent plus drop last year but it remains greater than the peak-to- trough fall in GDP seen in the 1990s recession and is a very long way from a boom. The recent rate of GDP growth is also unlikely to be maintained for two reasons. First, the impact of the current third wave of the virus is likely to have contributed to some plateauing, or even falls, in timely measures of economic activity such as retail footfall and credit card spending. Second, because we are recovering from a pandemic rather than a normal recession, that recovery should not be thought of as a smooth process where all sectors grow together in order to fill the GDP gap that has opened up. Some sectors need to shrink significantly while the likes of hospitality and leisure grow, as major pandemic driven switches in consumption unwind. Looking at the retail sector as we do in Figure 10, for example, shows that household goods sales are still 33 per cent up on their pre-pandemic levels as richer households who cannot go on holiday buy things instead. A full recovery will actually see big falls in this area of economic activity. Figure 10: Change in retail sale volumes are very different between sectors How far the recovery has to go is also clear if we look beneath the aggregates at those people and places hardest hit so far. The number of employee jobs amongst young people and in London remain hugely down (see Figure 11). This is not what a tight labour market looks like. Figure 11: Employment for under 25s and Londoners remains low Conclusion The labour market is improving fast, but it is doing so from an exceptionally weak starting point. The challenges those improvements bring for some sectors should not cloud us from the reality that in aggregate we have a very long way still to go. A tightening labour market is not the same thing as a tight one. Labour shortages in hospitality are neither a total disaster for our economy, nor evidence of a new dawn of power for low-paid workers. Instead, they are part of exactly the bumpy ride we should expect as our badly-needed recovery gets going so long as we pay attention to one key fact: we are still living through a pandemic. 1 The survey was undertaken by YouGov from the 3rd – 8th June 2021 of 8,030 adults aged 18+. Results are weighted so as to be representative of the population of that age group. Respondents were asked to describe their labour market status during 24-30 May 2021. 2 For more discussion of this see, H Slaughter, The Resolution Foundation Earnings Outlook Q1 2021, Resolution Foundation, March 2021. 3 The Bank of England’s Monetary Policy Committee concluded last week that “On an underlying basis, pay growth appeared close to pre-pandemic levels.” 4 Job-to-job moves forecasted for Q2 on the basis of May unemployment rate. Inflows from non-employment forecast for Q2 holding May RTI number fixed and adjusting downward for typical revisions to first estimate, then applying growth rate to ONS labour market flows (X02) Vacancies for June forecast using Adzuna weekly data and historical relationship with ONS vacancies. The interval around the final data point represents alternative scenarios for filled vacancies. 5 The supply chain disruption caused by the normalisation of economies and why it is likely to be temporary is explored in S Helper & E Soltasa, Why the pandemic has disrupted supply chains, White House Council of Economics Advisors, June 2021.