Tuesday, 8 September 2020

UK Labour Market Under COVID Part 3: Layoffs, Hiring & Wages

Recessions, usually signify a rise in layoffs. Even the last recession, which was accompanied by much more wage moderation than previous ones, generated a significant rise in redundancies. This time it is a little different. Furloughing allows firms to postpone difficult decisions. This can be seen in Figure 1 below which tracks layoffs week by week through the first 26 weeks of 2020 relative to the norms of the past 5 years. While there may have been a small upturn in layoffs from around week 16, it is hard to discern a pattern to weekly layoffs that lies outside the norms of the last 5 years. And since the last 5 years were not recessions, it is hard to see much evidence of layoffs so far.

Figure 1. Layoffs in 2020 Week by Week Relative to Norm

Nor is there much evidence of wage cuts in the face of the crisis. We know hours have fallen back and this may affect weekly wages so to net this hours drop out out we can look at hourly wages. Figure 2 plots the level of hourly wages, adjusted for inflation, week by week. While there is a deal of volatility in these numbers, due to smaller samples, the overall picture ios of continued real wage growth over the course crisis.

Figure 2. Hourly Wages in 2020 Week by Week Relative to Norm
So people may not have been fired or had their wages cut so far, but there haven't been many people hired. This, in addition to workplace absences, is perhaps the most noticeable feature of the UK labour market during the crisis so far. Between 2% to 4% of the workforce is usually newly-hired in any week, with fewer hires typically in Spring and Autumn being the main hiring period. In uncertain times, many firms try to make do with their existing workforce rather than expand. They may not even replace staff who leave. Figure 3 plots the proportion of employees who have been in new jobs for under one month as a proxy for hiring. Hiring rates in the spring and summer of 2020 were some 1 to 2 percentage points below seasonal norms in the weeks leading up to the lockdown and thereafter. In short, hiring stalled over the crisis. Cumulative hires were some 40.5 percentage points lower than usual by the end of week 26. This is around 12.5 million fewer hires than might be expected. many firms are having to make do with existing staff.

Figure 3. Hiring in 2020 Week by Week Relative to Norm

Friday, 4 September 2020

What's Happening to the Labour Market under COVID-19? Part 2

So if unemployment, employment or more marginal types of working appear little changed under COVID, things are different from normal if we look elsewhere. Figure 1 shows the percentage of workers who say they had a job but were away and did zero hours during the survey week. The weekly norms show large spikes around Christmas, Easter and the summer holiday season. However, there is a notable departure from the norm that begins in week 10 of 2020 and increases rapidly to 24% of the employed by week 16 (four weeks into lockdown). This is around 5.4 million more workers away from their jobs than expected at that week in the year. By week 22, the start of the government’s gradual relaxation of lockdown, absences from work had fallen back to just under 20% of the employed workforce, but then rise again in week 26 (last week of June). There are around 31.5 million people in work in the UK. Adding up these work absences "excesses" from week 5 to week 26 suggests that have been around 69 million additional person-week absences from work since the crisis began. This is equivalent to the entire UK workforce doing nothing for 2 weeks and 1 day.

Figure 1. With Job but Not Working

The shock to output is somewhat larger than this because among the majority of employed still in work and not away, many were working fewer hours than usual. Some 48% reported working fewer hours than usual in lockdown week 12. The number of hours worked by the average (median) worker fell from the norm of 38 to 34 during week 13, (Figure 2), before gradually rising back to around 35 hours a week by week 26, still less than the norm for this week. Average hours worked usually drop at Christmas, Easter and the summer holiday weeks, but the departures from the norm in week 13 and afterwards are notable.

Figure 2. Total hours worked each week of those not away
Again we can estimate the number of hours lost among this section of the workforce by multiplying by the number of hours each person works relative to the norm. A "normal" week of work in the UK amounts to around 1000 million person-hours. If we add all the lost hours for every individual, the cumulative excess suggests a fall of 2400 million hours since the crisis began, equivalent to the entire workforce doing nothing for 2 weeks and 2 days.

Taking the job absences and the reduced hours together that's a combined loss of about 4 and a half week's output in the year so far - or a quarter of the usual amount of stuff. That, more or less, lines up with the estimated drop in GDP so far (see ONS )

But there is more happening that has the potential to affect future labour market developments. For that we need to read part 3 of this update.

What’s Happening to the UK Labour Market under COVID-19? Looking for signs of excess and what it shows (so far). Part 1

Six months into the worst pandemic of the modern era, the worst recession of the modern era and yet and yet ....
the labour market is proving to be remarkably resilient (so far) – at least along the dimensions normally used to measure labour market performance. Recessions are usually accompanied by some combination of job loss, hiring freezes, wage cuts or hours reductions. But this of course is a recession with few precedents (the 1918-1920 flu pandemic and aftermath of world war 1 perhaps comes closest). As we shall see the furlough policies put in place have almost certainly helped prevent major job shedding seen elsewhere. In a rapidly evolving crisis, there is a need for regular and timely information to assess the course of labour market performance throughout the crisis and develop strategies to address the problems that emerge. Administrative data, like the claimant count measure of unemployment, are timely but not generally rich enough to provide the comprehensive view of labour market performance needed to inform and develop policy. Household surveys, like the Labour Force Survey (LFS) are much richer but hitherto have been used/made available at more discrete intervals, making assessments of labour market performance better, but less contemporary.

A couple of recent aricles (see Paper 1  or Paper 2 ) have shown how the weekly information in the LFS can be used to estimate a broad range of labour market indicators at much higher frequencies than usual. We can compare the current weekly performance of, say, the unemployment rate with the equivalent weekly norm observed over the previous five years. Any sustained departures from the norm is something to take heed of. In what follows we update a broad range of labour market performance indicators to the latest available dates (last week in June).

The graphs that follow plot the five-year average (dotted line) for an indicator in each week of the year - the “norm” - alongside its five-yearly maximum and minimum values for the same week (grey shade). The equivalent 2020 weekly estimates (solid line) are then overlaid. The “excess” on the graphs is the difference between the solid and dotted lines. Any 2020 data outside the grey range is a notable departure from the norm. Adding these weekly excesses gives the “cumulative excess” of, say, unemployment. The dashed vertical lines on the graphs highlight the first registered Corona related death (week 5 of the calendar year) and the week the UK went into lockdown (week 12). The 2020 data currently run to week 26, the last week of June.

Using the most common metric of labour market performance - unemployment - not much can be observed over the first 26 weeks of 2020 that is unusual. Figure 1 shows that UK unemployment is typically lower in the weeks that run up to Christmas and higher immediately after. However, the weekly unemployment rate in the first 26 weeks of the Corona crisis was well below the average of the last five years.

Figure 1. Unemployment in 2020 Relative to the Norm

A similar story applies to the employment rate and self-employed and temporary jobs. No obvious departures from previous norms. The government's furloughing and other related schemes appear to have helped contain any widespread job letting up to this point.
But there have been changes. And for this we need to read part 2 of this posting

Figure 2. Employment in 2020 Relative to the Norm
Figure 3. Self-Employment in 2020 Relative to the Norm
Figure 4. Temporary Job Share in 2020 Relative to the Norm

Thursday, 24 January 2019

Measure for Measure: Has the Government Already Met its Net Immigration Target Without Noticing?

It may well be that we no longer care about estimating the number of immigrants in the UK (though this is doubtful). Government policy (at the moment) still appears to be based around the notion that getting net immigration down below 100,000 is the thing to do. As such it needs a measure of how well it is doing on this front.

If so it may well be that the government has already achieved this without noticing it or indeed doing anything much to immigration policy - other than have a Brexit vote.

To see why this may be so, it is important to note that there are at least 2 different data sources, both based on surveys, with which to estimate changes in the numbers of immigrants coming to and leaving the UK. Neither source is perfect, but the government (and the ONS) refer to the net changes as estimated by the International Passenger Survey (IPS).  This estimates the number of non-British nationals arriving minus the number of non-British nationals leaving in any year to get an estimate of net-immigration.

There is however an alternative source, the Labour Force Survey, (LFS), which can estimate the numbers of immigrants living in the UK at any time. The LFS is used by the ONS to estimate the numbers of employed and unemployed in the UK and as such is deemed to be a reliable and timely source for population numbers. The difference between the LFS immigrant estimates at two points in time is also an estimate of the net change in immigration.  Figure 1 tracks the LFS estimates of the total immigrant population (based on self-reported country of birth) from 2010 to the present.

Figure 1: LFS Estimates of UK Immigrant Population
The dotted lines represent 95% confidence intervals and so map out the likely extent of sampling error in the estimate.  The essential point is that the LFS indicates that the immigrant population (measured here as those born outside the UK) in the UK has been falling since 2017 q1. The dashed lines indicate that this is unlikely to be the result of sampling error.

On the LFS measure therefore the government has met its target of getting net immigration into the tens of thousands for more than a year now (a falling population means net immigration is in the minuses)

To see how this conflicts with the estimates regularly published and commented on by government and the media, use the fact that any change in a population (be it people with freckles, goldfish etc) equals the number of arrivals minus the number of leavers. In the case of the immigrant population this means

Inflows into UK – Outflows from UK + Births - Deaths = Net Change in Immigration

The left hand panel of Figure 2 tracks the LFS estimate of net immigration - as measured by the 12-month change in the immigrant population numbers at a point in time - along with a (generous) allowance for sampling error. This again makes it clear that net immigration - according to the LFS - has been falling for over a year.  More people are leaving than arriving (or dying than being born).  The government's target has been reached - and more.

In contrast the IPS measure of immigration in the right hand panel shows net immigration happily growing at around 300,000 a year - and so way above the 100,000 target, with little change since the Brexit vote.

So how can we reconcile these two patterns? The IPS just measures inflows and outflows not births and deaths, so it is just possible that the LFS numbers are in negative territory because more immigrants are dying than being born - though this seems highly unlikely given the age profile of immigrants (immigrants are typically younger than the rest of the population and so less likely to be at risk of death). The IPS counts immigrants as "Non-British nationals" who are expecting to stay in the UK for more than 12 months, while the LFS estimates above estimate immigrants based on country of birth. So it is conceivable that the two different measures will produce different trends (though they didn't in the run-up to the Brexit vote - see forthcoming Blog post). So do we just surrender to the fact that surveys measure different things and give up? Or should we worry more about the different signals coming from the different surveys and the consequent uncertainty for policy making.

This is important because immigration is likely to remain an issue at centre of public/political/academic debate for the foreseeable future. As such would seem important to try to get good estimates of the level as well as the change in the immigrant population in order to inform policy. The two principal UK data sets that measure aspects of immigration currently seem to be saying different things.
We probably need more regular open discussion/dialogue between users and providers of the data about what can be done. And some recognition from policy makers that there is a discrepancy in the signals from data sources that may be worth thinking about.

Monday, 7 August 2017

Brexit and the UK's Immigration Needs

The call to action from the UK government to the Migration Advisory Committee went out last week urging the MAC to look at, among other things, the sectoral spread of EU migrants with a view to informing  the government  about "aligning the UK immigration system with a modern industrial strategy". In essence this suggests that the MAC will be asked to think about whether there are any sectors of the economy that are particularly reliant on migrant labour from the EU that might be put on some sort of expanded shortage list in the so-called "third phase" ie after or if the, as yet undefined, "transitional period" that follows Brexit in March 2019 has ended.

Immigrants make up, on average, around 15% of the employed workforce and EU migrants comprise around 5% of the workforce, but these proportions rise significantly for certain agricultural and manufacturing sectors where, as the Table below shows, the immigrant share is close to 50% and EU workers often comprise more than a third of the workforce. 

Most of these sectors are small in terms of total number employed but their reliance on immigrant labour is clear.  They are also, as the Table shows, typically low wage, high labour turnover sectors.

Even if the UK ends up with a hard Brexit, then most of the existing EU workforce in these sectors - and elsewhere -  would still be eligible for British citizenship or leave to remain having the requisite years of residence in the UK. So the workforce is unlikely to disappear overnight.

It is at the hiring margin that employers in these industries may face difficulties after Brexit without some sort of sectoral quotas, seasonal schemes or adjustment period.  The annual hiring rate in the UK typically hovers around 15% of the total workforce, though goes up and down depending on the state of the economic cycle. But low wage sectors such as those in the Table tend to have higher turnover. Annual labour turnover in the hotels sector is close to one third of the workforce.

Table: Sectoral Distribution of EU immigrants, 2016

Total employment

Percentage of sector who are immigrants
Percentage of sector who are EU immigrants
Average hourly Wages
Hiring rate1

Meat Processing

Fish processing

Veg. processing

Crop growing



Food Manfcture



UK Average


Source: LFS. Note 1. Hiring rate is % of workforce in employment for less than 1 year. Average wage is median hourly wage of all employees in the sector

How might Brexit affect  hiring? Unless there is an eventual agreement to stay in the single market recruitment from the EU is unlikely to continue at similar volumes after Brexit. Moreover, if, as might be possible at time of writing, the UK economy stalls while many of the EU economies begin to grow more rapidly, then there may be further falls in the net migration figure – for reasons beyond government control. Migrants seeking work will be attracted to the best employment opportunities.

If there are suddenly more opportunities in mainland Europe we would expect more migrants to choose Germany, France or Spain rather than the UK. The 15% fall in sterling relative to the Euro over the last year also makes UK wages 15% less attractive relative to a job in the Eurozone to prospective migrants. This means that even in the absence of Brexit, a seemingly never-ending supply of European labour – faced with increasing better alternatives elsewhere - must be in doubt.

Non-EU work visas to the UK are currently restricted to “graduate-level” jobs. Most of the sectors in the Table (excepting translation services) are non-graduate jobs – although there are some graduates working in these sectors. Any quotas on work visas on EU nationals after Brexit are also likely to favour graduate sector jobs because the net fiscal benefit is higher. High paid workers tend to put in more than they take out in benefits and public services. So certain firms and sectors would also have to look around for different sources of labour, raise wages or change their methods of working (though any firm that relies on a never ending supply of EU workers in an environment of free movement has an unstable business model). 

 Business has long argued for the freedom to hire the best people for the job – though no country in the world allows business unfettered access to migrant labour or allows them complete freedom to set wages or other employment conditions. In the UK, at the less skilled end of the labour market, the national minimum wage, the soon to be revamped Labour Inspectorate ensure that wages and working conditions are regulated.   The issue for a minority of employers going forward is how to adjust to a labour market in which one particular source of less skilled workers is likely to be restricted.

Tuesday, 25 July 2017

Is There a Public Sector Pay Premium?

The issue of public sector pay went centre-stage recently in the debate over whether to lift the public sector pay cap. Presumably one of the key factors, alongside where the revenue to pay for any uplift would come from, would be whether public sector salaries had fallen behind those in the private sector which could make it harder to recruit, retain and motivate staff.
The problem is how to measure any public sector pay gap. Because public sector jobs often have different characteristics from those in the private sector (there are relatively more graduate jobs in the public sector for example) a simple comparison of average pay between the public and private sector can be misleading.  The difficulty is what to net out when doing comparisons. Education ? Almost certainly because of different entry requirements. Age? Almost certainly because longer work experience tends to be reflected in higher pay and any differences in experience between the public and private sectors needs to be netted out in making comparisons. What about job characteristics? Here economists are divided. Some argue that people choose jobs on the basis of its features (eg unionised work) and this choice partly depends on wages and not the other way round. Others argue that there is a going rate for a particular kind of job and so job characteristics should be controlled for.  In the end what matters is whether this makes any difference to the conclusions regarding the public sector pay gap. And the answer is yes it does.

The solid blue line in the graph below shows estimates of the average (mean) public sector hourly pay gap over time measured 3 ways a) the raw percentage difference between public and private without netting out anything  b) the gap net of experience, education and gender c) the gap  net of experience education, gender and a control for whether the job is manual or not
The Public Sector Pay Gap Measured 3 Ways
(Technical note the blue line in panel 3 is the estimate from an OLS regression of log hourly wages on public sector work net of a quartic in experience interacted with each of 7 qualification dummy variables, plus controls for gender and manual work. The dotted lines are the 95% confidence intervals around the estimates. The regression is run for each year. All data are taken from 4 quarters of the UK Labour Force Survey in each year).

Whichever way pay is measured there is a common pattern. Public sector pay has been falling since 2011 and is now - in relative terms -  the lowest it has been compared to wages in the private sector for 25 years. If we believe the 2nd panel there was still an average public sector premium of around 3% in 2016 (the latest full run of data we have). If we believe panel 3, average public sector pay is now lower than that in the private sector and has been since 2015.

Another feature of the public sector pay debate is that the pay gap varies across different areas of the UK.
Public Sector Pay Gap Across the UK
The graph shows the evolution of the public sector pay gap as measured according to the 2nd methodology (net of experience, education and gender). Even on this conservative estimate it is clear that the public sector pay gap in London and the South-East of England is much lower than in the rest of the UK. Indeed according to this estimate average pay in the public sector in London and the South East is lower than in the private sector and has been falling further - to around 5% below the private sector average - since 2011. In contrast public sector pay in the rest of the UK, whilst falling in relative terms, is still around 7 % above that of the average private sector wage.


Tuesday, 4 July 2017

Interesting Times: Immigration and the UK Election

Interesting Times: The Election and UK Immigration
Had things gone as most commentators expected, the UK would now be entering hard Brexit talks with the near certainty of leaving the single market and/or customs union and the consequent ending of free movement of people from the European Union. A few weeks later and that near certainty no longer seems as certain, with murmurings of a Softer Brexit and the implication that allowing freer movement of labour from the EU may now be up for discussion.  Yet any rowing back on ending free movement would immediately conflict with the net migration target of 100,000 - a key feature of the Conservative party manifesto. If we assume that this is still in play, any allowance of free movement from the EU – around half of the net migration number - would obliterate any attempts to hit such a target within a parliament.

Would this matter? The net migration target is one of the strangest political fetishes of modern political history. While sensible arguments could be made for placing some restrictions on immigration – no country allows unrestricted migration - It is far from obvious why any government would seek to target something over which it has very little control. Simply put the net target is the result of the difference between the numbers of migrants coming in to the UK and the number leaving. Government can currently control, at best, around half of the inflow – that from outside the EU. Even that is moot since student numbers, the majority of entrants from outside the EU, are not regulated. With a hard Brexit it could potentially control the other half of the inflow. With or without Brexit it can not and will never have, any control over the numbers leaving. Yet changes in these uncontrolables can and do affect the target.

A good example of this underlies the recent fall in net migration, shown in Figure 1 below. Undoubtedly net migration has fallen considerably (by around 100,000) over the past year. The irony is that this fall is mostly driven by factors over which the government has no control – namely a rise in the number of EU (mainly A8) migrants leaving the UK and a rise in emigration of British citizens (the green line in the second panel is below zero which means net emigration of British citizens).
Figure 1: UK Immigration Flows, 1991-2016
Source: ONS (2017) .

Going forward, if the UK economy stalls while many of the EU economies grow more rapidly then there may be further falls in the net migration figure – again for reasons beyond government control. Migrants seeking work will be attracted to the employment opportunities that are best for them. If there are suddenly more opportunities in mainland Europe we would expect more migrants to choose Germany, France or Spain rather than the UK.  (the 15% fall in sterling relative to the Euro also makes UK wages 15% less attractive relative to a job in the Eurozone to prospective migrants). So the inflow to the UK could fall over the next few months reducing the net migration numbers without the government having done anything other than preside over a stalled economy.
Even if migration form the EU fell to zero, the net migration numbers would be way above the 100,000 target and require any government intent on trying to hit the target to make some (infeasible) restrictions on student numbers from outside the EU.
But the real question is whether it matters if immigration comes down a lot or a little. With regard to pay and job prospects of the UK-born population the answer is probably not. As Figure 2 below illustrates, large rises in immigration had very little effect on jobs (and pay) of either skilled or less skilled UK-born workers. The solid red line summarises the relationship between immigration and UK-born unemployment rates. If immigration increased unemployment, we would expect a strong upward sloping line: more immigrants would mean more unemployment for local workers. It is clear from the graph that there is no positive relationship between immigration and unemployment rates of those born in the UK. If anything, the relationship is negative, suggesting areas with more immigration experienced larger falls in unemployment for the UK-born over this period. Look at two areas – dots A and B in Figure 2. Both have had increases in the immigrant share well above the national average for this period. In area A unemployment for the UK-born has risen by over 1 percentage point, which is also above the national average. So in area A it feels like immigrants are bad for jobs. But area B has had a similar increase in immigration, while unemployment here has fallen by 2 percentage points. Therefore, just because immigration and unemployment both go up in an area does not mean that immigration is the reason for rising unemployment, since it is quite easy to find areas where immigration went up and unemployment fell. Something else must underlie the prospects of UK-born individuals in areas with rising unemployment.

Figure 2: Unemployment rates of UK-born and immigration
Notes: Each dot represents a UK local area. The solid line is the predicted ‘best fit’ from a regression of changes in unemployment on the change in share of immigrants in each UK local area. These are weighted by the sample population in each area.
Source: CEP analysis of Labour Force Survey.

And if immigration had little effect when it went up a lot, it is hard to think why or how a large fall in immigration would have any large effects either.
And that is probably the crux of the immigration issue. Immigration seems to matter much more politically than it does economically. All the empirical work that has been done on UK immigration shows very small effects, either positive or negative, though its influences in cultural matters probably extends much further. True any reduction in EU immigration would make paying the deficit off a little harder – since EU migrants pay more in taxes than receive in benefits and public services. (whereas UK-born and non-EU migrants receive more in benefits than they pay in taxes), but these net payments are not very big. Certain firms and sectors -see forthcoming blog post - would also have to look around for different sources of labour or methods of working (though any firm that relies on a never ending supply of EU workers in an environment of free movement has an unstable business model), but the number of sectors affected is relatively small and the adjustment will always be at the hiring margin not the entire workforce. 
Perhaps one of the best things that could come out of the election is that we learn to stop worrying so much about immigration. Or if we can't do that then try to stop worrying about its effects on jobs and pay.

Jonathan Wadsworth