Upload
others
View
0
Download
0
Embed Size (px)
Citation preview
© Institute for Employment Studies 1
Institute for Employment Studies 1
Labour Market Statistics, June 2020 16 June 2020
This briefing note sets out analysis of the Labour Market Statistics published this morning.
The analysis covers:
■ Claimant unemployment. The ‘Claimant Count’ is a measure of the number of people
claiming benefits principally for the reason of being unemployed, compiled from
Jobseeker’s Allowance and Universal Credit data. This is not an official measure of
unemployment, but it generally follows the same trends as the official survey-based
unemployment measure and provides more timely and granular data. Today’s data
sets out claimant unemployment as at 14 May 2020.
■ Labour Force Survey data. This is a household survey that collects official data on
employment, unemployment and economic inactivity. Today’s data covers the period
February to April 2020, drawing on approximately 33 thousand interviews. Around one
third of these will have been conducted in April, so this is the first release with a full
month of data from after the crisis started.
■ Pay As You Earn Real Time Information. These are experimental statistics on
employee levels and pay, covering the period to May 2020. The statistics are drawn
from data supplied by employers in PAYE returns. Where data is missing (which
affects approximately 10% of cases in the most recent month being reported) values
are imputed based on previous returns.
Claimant unemployment is rising faster than at any point since unemployment benefits were introduced
Between March and May, claimant unemployment rose by 1.56 million, from 1.24 million
to 2.80 million – an increase of 125%. Claimant unemployment has risen more in two
months than it did in the whole of the last two recessions, and by more than it did in the
first twelve months of the Great Depression (when it rose by 1.0 million). As Figure 1
below shows, unemployment is now as high as it was in October 1993, while Figure 2
shows that this is now the biggest year-on-year rise in unemployment benefits since their
introduction in 1922.
It is important to note The Office for National Statistics has emphasised this morning that
the claimant count may be over-stating the true rise in unemployment, as some of those
being treated as unemployed in the benefits system may in fact still be working a small
Labour Market Statistics, May 2020: Claimant count analysis
number of hours. While there does appear to be a discrepancy between the claimant
count and other measures, our view is that short-hours working cannot explain most of
the rise that has been observed in the last two months. This is explored in more detail
later in the report.
Figure 1: Administrative and Claimant Unemployment, 1922-present
Source: IES analysis of Bank of England and Office for National Statistics data. The figure presents
Administrative Unemployment for 1922-1971, and the Claimant Count for 1971 onwards.
Figure 2: Year-on-year increase in claimant unemployment, 1922-present
Source: IES analysis of Bank of England and Office for National Statistics data. The figure presents
Administrative Unemployment for 1922-1971, and the Claimant Count for 1971 onwards.
Claimant unemployment is highest for young people, but is rising across all age groups
As Figure 3 below sets out, the claimant count has doubled across all age groups. It has
risen by 122% for those aged 25-34, although as we set out last month the large rise in
unemployment for older people – up from 300 thousand to 600 thousand – is very
-
500,000
1,000,000
1,500,000
2,000,000
2,500,000
3,000,000
3,500,000
Dec-2
2
Mar-
26
Jun-2
9
Sep-3
2
Dec-3
5
Mar-
39
Jun-4
2
Sep-4
5
Dec-4
8
Mar-
52
Jun-5
5
Sep-5
8
Dec-6
1
Mar-
65
Jun-6
8
Sep-7
1
Dec-7
4
Mar-
78
Jun-8
1
Sep-8
4
Dec-8
7
Mar-
91
Jun-9
4
Sep-9
7
Dec-0
0
Mar-
04
Jun-0
7
Sep-1
0
Dec-1
3
Mar-
17
-2,000,000
-1,500,000
-1,000,000
-500,000
0
500,000
1,000,000
1,500,000
2,000,000
Nov-2
3
Oct-
26
Sep-2
9
Aug-3
2
Jul-35
Jun-3
8
May-4
1
Apr-
44
Mar-
47
Feb-5
0
Jan-5
3
Dec-5
5
Nov-5
8
Oct-
61
Sep-6
4
Aug-6
7
Jul-70
Jun-7
3
May-7
6
Apr-
79
Mar-
82
Feb-8
5
Jan-8
8
Dec-9
0
Nov-9
3
Oct-
96
Sep-9
9
Aug-0
2
Jul-05
Jun-0
8
May-1
1
Apr-
14
Mar-
17
Feb-2
0
© Institute for Employment Studies 3
Institute for Employment Studies 3
concerning, as previous recessions have shown that older people are much less likely to
leave unemployment and get back to work than others.
Figure 3: Claimant unemployment by age, March-May 2020
Source: IES analysis of NOMIS claimant count data
Looking at the proportion of the workforce unemployed, in Figure 41 below, we can see
that claimant unemployment is highest for young people – now standing at 12.8% of the
18-24 workforce, a rise of 6.7 percentage points since the crisis began. Given that
around half of young unemployed people do not claim benefits (usually because they are
not eligible, due to their age or studying) it is likely that the official measure of
unemployment could be on its way towards 25% in the coming months, or one million.
This would be a higher rate than was reached in the aftermath of the 2008/9 recession.
1These figures should be treated with caution, as they use estimates of the economically active population
between February and April 2020 which are likely to over-state economic activity and so under-estimate the
unemployment rate. This is because some of those who left employment since March 2020 will have
become economically ‘inactive’ rather than unemployed.
0
100,000
200,000
300,000
400,000
500,000
600,000
700,000
800,000
900,000
18-24 25-34 35-49 50+
Mar-20 Apr-20 May-20
Labour Market Statistics, May 2020: Claimant count analysis
Figure 4: Estimated claimant unemployment rate by age, March-May 2020
Source: IES analysis of NOMIS claimant count and ONS Labour Force Survey data
Claimant unemployment is highest in many ex-industrial, inner city and coastal areas
Figure 5 below illustrates that claimant unemployment rates are highest in the North East
(10.1%), West Midlands and North West (both 9.0%). The percentage point rises in
unemployment however are now fairly consistent across regions, rising by around four
percentage points since March. The exception to this is London, where the rate has risen
by 5.1 percentage points – going from having one of the lowest rates to being among the
highest.
0.0%
2.0%
4.0%
6.0%
8.0%
10.0%
12.0%
14.0%
18-24 25-34 35-49 50+
Mar-20 Apr-20 May-20 Change Mar-May (percentage point)
© Institute for Employment Studies 5
Institute for Employment Studies 5
Figure 5: Claimant unemployment rate by region or nation, March-April 2020
Source: IES analysis of NOMIS claimant count data
The fact that unemployment rates have increased by similar amounts between regions
also means that the percentage rise has been greater in areas where unemployment was
previously lower. So as Figure 6 shows, claimant unemployment has risen by over 170%
in the South East and South West, while it is up by 72% in the North East (albeit from a
far higher pre-crisis rate than elsewhere).
Figure 6: Percentage increases in claimant unemployment by region, Mar-May 2020
Source: IES analysis of NOMIS claimant count data
0.0% 2.0% 4.0% 6.0% 8.0% 10.0% 12.0%
South West
East
South East
Northern Ireland
East Midlands
Scotland
United Kingdom
Wales
London
Yorks and Humber
North West
West Midlands
North East
Mar-20 May-20 Change Mar-May (percentage point)
0%
20%
40%
60%
80%
100%
120%
140%
160%
180%
200%
Labour Market Statistics, May 2020: Claimant count analysis
Figure 7 looks beneath the regional level at individual Local Authorities, grouping them
according to the decile of unemployment immediately before the crisis began. Those
areas in the first decile had the lowest unemployment (averaging just 1.3% of residents),
while those in the tenth had the highest unemployment (averaging 5.3% of residents).
This analysis shows that the rises in claimant unemployment rates have generally been
greater in areas that had higher unemployment before the crisis began. Places in the
eighth to tenth deciles saw increases of 3.7 to 3.8 percentage points. However there
have been very large rises in areas that had fairly average levels of employment before
the crisis began, emphasising the broad based nature of the labour market impact so far.
Areas with lower unemployment have seen smaller increases in unemployment rates,
although as with the regional analysis above this disguises large percentage increases –
with unemployment more than trebling in the lowest decile, albeit from a low base.
Figure 7: Percentage point change in proportion of residents claimant unemployed, by
decile of unemployment in March 2020 (1=lowest)
Source: IES analysis of NOMIS claimant count data
Looking at individual local authorities, those areas with the highest unemployment are
dominated by ex-industrial areas, coastal towns and inner cities. Table 1 below illustrates
this, showing areas where more at least one in eleven working age adults (9%) are
unemployed.
It is worth noting that several London boroughs now feature among the local authorities
with the highest rates, which have all seen unemployment more than double since March.
The highest pre-crisis rate among these boroughs of 4.4% in Barking and Dagenham put
it 39th in the list of all local authorities sorted from the highest to lowest claimant rate.
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
1 2 3 4 5 6 7 8 9 10
© Institute for Employment Studies 7
Institute for Employment Studies 7
Table 1: Local authorities where at least one in eleven working age adults are claimant
unemployed
Local authority Region Proportion of 16-64 population % increase
May 2020
March 2020
Change (percentage point)
In claimants
Blackpool North West 12.3 7.2 5.1 72
Thanet South East 10.8 5.7 5.1 91
Birmingham West Midlands 10.6 6.7 3.9 57
Wolverhampton West Midlands 10.3 6.4 3.9 62
Middlesbrough North East 10.1 6.2 3.9 62
Haringey London 10.1 3.8 5.7 165
Barking and Dagenham London 10.1 4.4 6.3 133
Kingston upon Hull, City of Yorks and Humber 9.7 5.8 3.9 66
South Tyneside North East 9.5 6.2 3.3 53
Oldham North West 9.5 5.2 4.3 83
Hartlepool North East 9.4 6.0 3.4 56
Newham London 9.4 3.3 5.9 185
Waltham Forest London 9.4 3.5 6.1 173
Brent London 9.3 3.6 5.7 161
Bradford Yorks and Humber 9.1 5.2 3.8 75
Sandwell West Midlands 9.1 5.3 3.9 72
Hastings South East 9.1 4.9 4.2 87
Burnley North West 9.0 5.4 3.6 66
Knowsley North West 9.0 4.8 4.2 88
Lewisham London 9.0 4.0 5.0 128
Source: IES analysis of NOMIS claimant count data
Vacancies have more than halved, with eight claimant unemployed people chasing every job opening
The ONS Vacancy Survey reports that there were on average 476 thousand vacancies
over the three months to May 2020. The single month estimate of vacancies for May
however was 320 thousand. Our analysis of Adzuna online vacancy data for Joseph
Rowntree Foundation suggests that June vacancies will be around 350 thousand,
meaning that by next month the three month average being reported by the ONS will be
around 330 thousand. As Figure 8 below shows, this will be the lowest level of vacancies
in the twenty years that the survey has been running (although longer-run data suggests
that vacancies were lower in the 1990s).
Labour Market Statistics, May 2020: Claimant count analysis
Figure 8: Estimated number of vacancies
Source: IES analysis of ONS Vacancy Survey. Dotted line indicates IES estimates based on analysis of
online vacancy data supplied by Adzuna for Joseph Rowntree Foundation
Record low vacancies combined with very high claimant unemployment means that there
are now 8.5 claimant unemployed people chasing every job advert. This has risen from
just 1.5 claimants per vacancy two months ago and is already well over double the rates
in the last downturn. This is shown in Figure 9 below. Our separate analysis of local
trends in this data, which we will be updating later this week, suggests that in ex-
industrial, coastal and inner city areas these ratios are significantly higher still.
Figure 9: Number of claimant unemployed per job vacancy
Source: IES analysis of NOMIS claimant count and ONS Vacancy Survey. Dotted line indicates IES
estimates based on analysis of online vacancy data supplied by Adzuna for Joseph Rowntree Foundation.
-
100
200
300
400
500
600
700
800
900M
ay-0
1
Jan-0
2
Sep-0
2
May-0
3
Jan-0
4
Sep-0
4
May-0
5
Jan-0
6
Sep-0
6
May-0
7
Jan-0
8
Sep-0
8
May-0
9
Jan-1
0
Sep-1
0
May-1
1
Jan-1
2
Sep-1
2
May-1
3
Jan-1
4
Sep-1
4
May-1
5
Jan-1
6
Sep-1
6
May-1
7
Jan-1
8
Sep-1
8
May-1
9
Jan-2
0
Thousands
Axis Title
-
1.0
2.0
3.0
4.0
5.0
6.0
7.0
8.0
9.0
May-0
1
Dec-0
1
Jul-02
Feb-0
3
Sep-0
3
Apr-
04
No
v-0
4
Jun-0
5
Jan-0
6
Aug-0
6
Mar-
07
Oct-
07
May-0
8
Dec-0
8
Jul-09
Feb-1
0
Sep-1
0
Apr-
11
Nov-1
1
Jun-1
2
Jan-1
3
Aug-1
3
Mar-
14
Oct-
14
May-1
5
Dec-1
5
Jul-16
Feb-1
7
Sep-1
7
Apr-
18
Nov-1
8
Jun-1
9
Jan-2
0
© Institute for Employment Studies 9
Institute for Employment Studies 9
Real-time Pay as You Earn data shows that the number of employees has fallen by at least 600 thousand
Figure 10 below shows the latest estimates of employee numbers from PAYE data up
until May 2020. This covers employees only, and is likely to have missing data for around
10% of staff (with these values imputed based on previous trends). This suggests that
employee numbers have fallen by approximately 600 thousand between March and May,
from 29.0 to 28.4 million.
This is a very large fall, as the figure sets out, but it is far smaller than the 1.6 million rise
in unemployment recorded in the claimant count. There are three plausible explanations
for this:
■ The employee data below is understating the fall due to incomplete returns. This
seems highly plausible, as the estimated monthly fall for previous month (March) was
revised up this morning to 450 thousand from a flash estimate of 20 thousand reported
last month. The flash estimate for April was 160 thousand, and we may well see this
also revised up next month.
■ Falls in the number self-employed. This is explored below, but it appears certain that
at least 300 thousand fewer self-employed people were in work in April than in March.
■ Claimant unemployed people are working short hours, and so not officially
unemployed. This will certainly explain part of the difference, as those claiming
Universal Credit and earning below a set threshold (roughly equivalent to sixteen hours
a week) are required to search for work and so may be counted as unemployed.
All told, our view is that the figures in the coming months will confirm that employment has
fallen by an unprecedented amount, but that the official measure of unemployment may
continue to lag the claimant count for some time.
Figure 10: Employee numbers reported through Pay As You Earn data
Source: IES analysis of Real Time Information Pay As You Earn data
25,500,000
26,000,000
26,500,000
27,000,000
27,500,000
28,000,000
28,500,000
29,000,000
29,500,000
Jul-14
Sep-1
4N
ov-1
4Jan-1
5M
ar-
15
May-1
5Jul-15
Sep-1
5N
ov-1
5Jan-1
6M
ar-
16
May-1
6Jul-16
Sep-1
6N
ov-1
6Jan-1
7M
ar-
17
May-1
7Jul-17
Sep-1
7N
ov-1
7Jan-1
8M
ar-
18
May-1
8Jul-18
Sep-1
8N
ov-1
8Jan-1
9M
ar-
19
May-1
9Jul-19
Sep-1
9N
ov-1
9Jan-2
0M
ar-
20
May-2
0
Labour Market Statistics, May 2020: Claimant count analysis
Labour Force Survey statistics are showing large rises in inactivity, but not yet unemployment
Turning to the main Labour Force Survey measures, as noted above the labour market
impacts of the crisis are not yet apparent in the headline measures of employment,
unemployment and inactivity. Overall in the three months from February to April,
employment fell by 150 thousand compared with the figures reported last month,
unemployment was broadly unchanged and economic inactivity (those not looking nor
available for work) rose by 190 thousand. However these are three month average
figures, so the single-month changes for April were larger than this.
ONS publishes single month estimates as experimental statistics, and Figure 11 below
shows the reported changes – employment appears to have fallen by just over 300
thousand while worklessness rose by the same amount.
Figure 11 Change in single-month estimates of employment, unemployment and inactivity
– March to April 2020
Source: IES analysis of Labour Force Survey data.
Looking in more detail at reasons for inactivity, it appears that the rise is more than
explained by those saying that they are inactive for ‘other’ reasons than being too sick to
work, looking after their family or home, retired, students or discouraged workers. Overall,
inactivity for these reasons actually fell in today’s data. The most plausible ‘other’ reasons
for being economically inactive would be that people were not seeking work while they
waited for their claims to Universal Credit to be processed, and/ or were self-employed
and had temporarily stopped working.
Tellingly, the number of people economically inactive who report that they ‘want a job’ has
also increased significantly compared with the data reported last month – which suggests
that these large rises in inactivity will feed through directly into increases in the headline
measure of unemployment in the months ahead.
-400
-300
-200
-100
0
100
200
300
400
Apr-20
Thousands
Employment Unemployment Inactivity
© Institute for Employment Studies 11
Institute for Employment Studies 11
Most of the reported fall in employment is explained by fewer people self-employed
As noted, the three-month average for employment was 150 thousand lower this month
than the figure reported last month. Figure 12 shows that two thirds of this fall is
explained by lower self-employment.
Given that this is a three-month average, the implied single-month figure for April is likely
around 300 thousand lower than pre-crisis levels. It is also likely that this understates the
real fall in self-employment, as anyone reporting having spent any time working for their
business in the reference week.
Figure 12: Number of people self-employed, Mar-May 2019 to Feb-Apr 2020
Source: IES analysis of Labour Force Survey data.
While employment has not fallen, working hours appear to have collapsed
The Labour Force Survey data does however show very large falls in average working
hours in the three months to April. Figure 13 below sets these out. Overall working hours
are down by 7.3% on the figure reported last month, with average hours lower than at any
previous point. Again, because this is a three-month average, the implied single month
fall is likely to be three times this – i.e. above 20%.
This is explained by the Job Retention Scheme, with the ONS reporting that there are at
least six million people reporting that they are temporarily away from work but still
employed.
4,500,000
4,600,000
4,700,000
4,800,000
4,900,000
5,000,000
5,100,000
Labour Market Statistics, May 2020: Claimant count analysis
Fewer hours also feeds through into lower pay, although clearly falls in pay have been
cushioned by the 80% salary replacement paid through the furlough scheme.
Nonetheless, nominal pay recorded a small year-on-year fall in April, for the first time
since pay survey data was collected in 2000.
Figure 13 Average hours worked per week
Source: IES analysis of Labour Force Survey data.
Conclusions and implications
Today’s labour market data presents a somewhat confusing picture of the impacts of the
crisis over recent months.
It appears to us that the very large rises in claimant unemployment are largely reflecting
what is happening in the real economy – with HMRC data recording a fall in employee
numbers of at least 600 thousand (which may well be revised up) and the Labour Force
Survey suggesting that self-employment fell by at least 300 thousand in April. We would
expect to see this translating into rising unemployment, rather than economic inactivity, as
those out of work start looking for new jobs (as has likely already happened). Non-
claimant youth unemployment is also likely to significantly add to the numbers seeking
work. Nonetheless it is likely that the claimant unemployment rises somewhat overstate
the real level of unemployment as it was in April and May.
Looking ahead, it is certain that claimant unemployment will rise above three million next
month and may well be higher than the peak of 3.1 million reached in the 1980s
recession.
Rising unemployment has also been accompanied by a collapse in job vacancies – from
around 800 thousand before the crisis to just 320 thousand in the single-month estimate
for May. Putting this together with the unemployment data, this means that there are now
27.5
28.0
28.5
29.0
29.5
30.0
30.5
31.0
31.5
32.0
32.5
Apr-
18
May-1
8
Jun-1
8
Jul-18
Aug-1
8
Sep-1
8
Oct-
18
Nov-1
8
Dec-1
8
Jan-1
9
Feb-1
9
Mar-
19
Apr-
19
May-1
9
Jun-1
9
Jul-19
Aug-1
9
Sep-1
9
Oct-
19
Nov-1
9
Dec-1
9
Jan-2
0
Feb-2
0
Mar-
20
© Institute for Employment Studies 13
Institute for Employment Studies 13
8.5 unemployed people competing for every job opening, compared with just 1.5 before
this crisis began. This picture also varies hugely across the country – with our analysis of
online adverts for the Joseph Rowntree Foundation suggesting that in many ex-industrial,
inner city and coastal areas the picture is far worse.
Looking ahead, new claims to Universal Credit appear to have fallen back to close to pre-
crisis levels, while data that we will publish on Friday will show that the number of new
vacancies has doubled in the last fortnight (albeit from a very low base). Both of these
are positive signs. However on the other hand unless people are confident to start
spending again, and businesses are able to reopen, then there is a very real risk that a
‘second wave’ of unemployment could hit us as the Job Retention Scheme starts to
unwind. This could push unemployment up by at least a further two million, and see it
rising inexorably towards four or even five million by the end of year. As with everything
in this crisis, all roads lead back to successfully suppressing the virus.
In the meantime though, our view is that we need to act now to deal with this jobs crisis.
As we set out over the weekend with over 200 business leaders, politicians and charities,
we think that this response needs three parts:
■ An urgent and temporary stimulus to kick-start hiring and support job retention over the
summer, ideally by cutting employer National Insurance;
■ Doubling of the size of our employment services, so that everyone gets the help that
they need to find new work; and
■ Investment now so that we are ready for record long-term unemployment next year – in
particular can guarantee that all young unemployed people will have the opportunity of
an education place, apprenticeship or job.
About IES
The Institute for Employment studies is an independent, apolitical centre of research and consultancy in employment policy and human resource management. It works with employers, government departments, agencies and professional and employee bodies to support sustained improvements in employment policy and practice.
Institute for Employment Studies, City Gate, 185 Dyke Road, Brighton, BN3 1TL United Kingdom
www.employment-studies.co.uk
@EmploymtStudies
01273 763400