Implementing employment support schemes in response to the COVID-19 pandemic — National Audit Office (HC 862)
Summary
A report by the Comptroller and Auditor General of the National Audit Office, "Implementing employment support schemes in response to the COVID-19 pandemic," HC 862, Session 2019–2021, dated 23 October 2020, examining HM Treasury and HM Revenue & Customs. It reviews the implementation of the Coronavirus Job Retention Scheme (CJRS) and the Self-Employment Income Support Scheme (SEISS). Its key facts report £52.7bn in total costs up to 20 September 2020, 9.6 million jobs furloughed, and an HMRC upper estimate of £3.9bn of fraud and error on CJRS. The summary finds that the schemes were implemented quickly and ahead of schedule and states that as many as 2.9 million people were not eligible. The document closes with a correction dated 12 October 2020 changing a figure in paragraph 2.19 from one-third to 12%.
Summary drafted by a model from the document's text below and checked by script against that text before publication. It is a navigation aid, not a reading of what the document proves. Where AI is used
Full text
A picture of the National Audit Office logo
Report
by the Comptroller
and Auditor General
HM Treasury, HM Revenue & Customs
Implementing employment
support schemes in response
to the COVID-19 pandemic
HC 862 SESSION 2019–2021 23 OCTOBER 2020
We are the UK’s independent public spending watchdog.
We support Parliament in holding government to account and
we help improve public services through our high-quality audits.
The National Audit Office (NAO) scrutinises public spending for Parliament and is independent
of government and the civil service. We help Parliament hold government to account and
we use our insights to help people who manage and govern public bodies improve public
services. The Comptroller and Auditor General (C&AG), Gareth Davies, is an Officer of the
House of Commons and leads the NAO. We audit the financial accounts of departments and
other public bodies. We also examine and report on the value for money of how public money
has been spent. In 2019, the NAO’s work led to a positive financial impact through reduced
costs, improved service delivery, or other benefits to citizens, of £1.1 billion.
HM Treasury, HM Revenue & Customs
Implementing employment
support schemes in response
to the COVID-19 pandemic
Report by the Comptroller and Auditor General
Ordered by the House of Commons
to be printed on 21 October 2020
This report has been prepared under Section 6 of the
National Audit Act 1983 for presentation to the House of
Commons in accordance with Section 9 of the Act
Gareth Davies
Comptroller and Auditor General
National Audit Office
16 October 2020
HC 862 | £10.00
This report examines the role of HM Treasury and HM Revenue
& Customs (HMRC) in implementing the Coronavirus Job
Retention Scheme (CJRS) and the Self-Employment Income
Support Scheme (SEISS). The report considers how well the
schemes were implemented, recognising the need to deliver
at speed in response to the COVID-19 pandemic.
© National Audit Office 2020
The material featured in this document is subject to
National Audit Office (NAO) copyright. The material
may be copied or reproduced for non-commercial
purposes only, namely reproduction for research,
private study or for limited internal circulation within
an organisation for the purpose of review.
Copying for non-commercial purposes is subject
to the material being accompanied by a sufficient
acknowledgement, reproduced accurately, and not
being used in a misleading context. To reproduce
NAO copyright material for any other use, you must
contact copyright@nao.org.uk. Please tell us who you
are, the organisation you represent (if any) and how
and why you wish to use our material. Please include
your full contact details: name, address, telephone
number and email.
Please note that the material featured in this
document may not be reproduced for commercial
gain without the NAO’s express and direct
permission and that the NAO reserves its right to
pursue copyright infringement proceedings against
individuals or companies who reproduce material for
commercial gain without our permission.
Links to external websites were valid at the time of
publication of this report. The National Audit Office
is not responsible for the future validity of the links.
009194 10/20 NAO
Contents
Key facts 4
Summary 5
Part One
HM Revenue & Customs’ implementation
of the employment support schemes 14
Part Two
Beneficiaries of the employment
support schemes 25
Part Three
Fraud and error affecting the employment
support schemes 42
Appendix One
Our audit approach 59
Appendix Two
Our evidence base 61 The National Audit Office study team
consisted of:
Appendix Three James Ball, David Betteley,
International comparisons 65 Anthony Pitt, Tom Tyson, Nigel Vinson
with support from Andrea Jansson
Appendix Four and Stephen Jobling, under the
Evolution of HM Revenue & Customs’ direction of Andy Morrison.
employment support schemes 67
This report can be found on the
National Audit Office website at
www.nao.org.uk
For further information about the
National Audit Office please contact:
National Audit Office
Press Office
157–197 Buckingham Palace Road
Victoria
London
SW1W 9SP
Tel: 020 7798 7400
Enquiries: www.nao.org.uk/contact-us
Website: www.nao.org.uk
If you are reading this document with a screen reader you may wish to use the bookmarks option to navigate through the parts. If
you require any of the graphics in another format, we can provide this on request. Please email us at www.nao.org.uk/contact-us
Twitter: @NAOorguk
4 Key facts Implementing employment support schemes in response to the COVID-19 pandemic
Key facts
£52.7bn £69.7bn £3.9bn
Total costs reported for Combined total forecast HM Revenue & Customs’
the Coronavirus Job spend for CJRS and first (HMRC’s) upper estimate
Retention Scheme (CJRS) two grants of SEISS of fraud and error on
and the Self-Employment CJRS to 20 September,
Income Support Scheme based on 5% to 10%
(SEISS) up to 20 fraud and error levels on
September 2020 £39.3 billion of payments.
The 5% estimate equates
to £2.0 billion
9.6 million total jobs furloughed with 1.2 million employers (61% of those
eligible) making at least one CJRS claim1
2 million estimate of workforce (to nearest million) remaining furloughed
based on employers surveyed, 7 to 20 September 2020
9% proportion of furloughed employees in our survey telling us
that they worked at their employer's request while furloughed
2.6 million self-employed individuals made a first grant claim to the SEISS
scheme (77% of those potentially eligible) to 31 July 2020
99.5% CJRS claims paid within six working days
97.5% SEISS claims paid within six working days
£278 million amounts companies repaid voluntarily to HMRC for CJRS
payments they did not need or took in error
£275 million overpayments HMRC estimates it could recover on
10,000 CJRS payments it believes are at high-risk of fraud
Note
1 The latest HMRC COVID-19 statistics are available at: www.gov.uk/government/collections/hmrc-
coronavirus-covid-19-statistics
Implementing employment support schemes in response to the COVID-19 pandemic Summary 5
Summary
Introduction
1 On 20 March 2020 government announced the Coronavirus Job Retention
Scheme (CJRS), followed on 26 March by the Self-Employment Income Support
Scheme (SEISS) as part of its economic response to the COVID-19 pandemic.
Our report refers to these as ‘the schemes’.
2 Initially government’s overriding ambition for both schemes was to provide
financial support to businesses and individuals as quickly as possible in the wake
of the COVID-19 pandemic to protect jobs. In summary:
• CJRS: The initial aim was to provide rapid financial support to help firms
continue to keep people in employment. Employers could put workers on
temporary leave and government would pay them cash grants of 80%
of employees’ salaries, up to £2,500 a month (CJRS 1). From July and
under a ‘flexible furlough’ phase (CJRS 2), the objectives changed to give
firms flexibility to support the recovery and in August, by introducing an
employer contribution.
• SEISS: The aim was to provide fast grant payments to self-employed
individuals whose businesses had been ‘adversely affected’ by the pandemic.
The grant was intended to help support self-employed individuals with
living costs, but they could also continue to work, start a new trade or take
up a new employment. A second SEISS grant was available to claim from
17 August 2020.
3 The schemes were only open to existing taxpayers who met certain
eligibility criteria. At 20 September 2020 CJRS had supported 1.2 million
employers and 9.6 million jobs, with claims totalling £39.3 billion. The Office for
Budget Responsibility (OBR) expects CJRS claims to reach £54.5 billion in total.
By 20 September 2020 the SEISS scheme had at least 2.6 million claims, totalling
£13.4 billion. OBR forecasts SEISS claims will total £15.2 billion, bringing total
forecast spending for these schemes to nearly £70 billion. The CJRS scheme will
end on 31 October. On 24 September 2020 government announced a new Job
Support Scheme (JSS) intended to provide help for short-time working (reduced
hours) and an extension to SEISS until April 2021. On 9 October government
announced an extension to JSS (referred to as ‘expanded JSS’) to support
companies whose businesses are legally required to close as a direct result of
coronavirus restrictions.
6 Summary Implementing employment support schemes in response to the COVID-19 pandemic
4 HM Revenue & Customs (HMRC) and HM Treasury – collectively referred
to as ‘the Departments’ – were responsible for advising ministers on the design of
the schemes. HM Treasury led on policy design and HMRC led on administrative
design and then the implementation and administration of the schemes.
Scope of this report
5 The scale and nature of the current COVID-19 pandemic, and government’s
response to it, is unprecedented in recent history. This report considers how
well HM Treasury and HMRC have managed risks thus far in implementing these
schemes. This report considers whether the Departments have:
• managed design and delivery risks effectively in implementing the
schemes. Part One assesses the Departments’ implementation of these
schemes against evaluative criteria that draw upon our work analysing
the government’s response to other crises;
• understood whether the schemes are reaching the people intended.
Part Two examines the Departments’ approach to ensuring that the
schemes were reaching their intended recipients, whether they had a good
understanding of the consequences of their design decisions and what the
impact of those decisions has been; and
• managed fraud and error risks effectively. Part Three considers how far
fraud and error risks have been addressed.
6 This report does not consider HMRC’s other COVID-19 interventions
designed to support businesses, including the Coronavirus Job Retention Bonus
and Eat Out to Help Out, or the relationship between the schemes and wider
government support such as business loans and benefits. We have previously
reported on the Bounce Back Loans Scheme.1
7 Our audit approach is described in Appendix One and the evidence base we
used is in Appendix Two. Appendix Three provides international comparisons and
Appendix Four shows how the schemes have evolved over time.
1 Comptroller and Auditor General, Investigation into the Bounce Back Loan scheme, Session 2019–2021,
HC 860, National Audit Office, October 2020.
Implementing employment support schemes in response to the COVID-19 pandemic Summary 7
Key findings
Designing and implementing the employment support schemes
8 The Departments implemented the schemes quickly and ahead of schedule.
Ministers set clear objectives that both schemes should be delivered quickly.
Government announced the CJRS scheme on 20 March 2020. HMRC planned
to make initial CJRS payments by the end of April 2020, but made it available
to employers from 20 April, just a month after the government’s announcement.
Government announced the SEISS scheme on 26 March. HMRC intended to
make SEISS payments by the beginning of June but made the scheme available
from 13 May, two weeks early. HMRC accelerated the original timetable for the
SEISS scheme, recognising the need to provide financial support to eligible
customers as quickly as possible (paragraphs 1.1 to 1.4 and Figure 2).
9 The scale of the challenge was potentially increased by the lack of
pandemic contingency planning and existing employment support schemes
the Departments could easily adapt. Instead, HM Treasury told us it drew on
economic contingency planning designed for financial rescues, developed
following the credit crisis; and draft policy work on wage subsidy schemes
and lessons learned from other countries, such as Germany, in implementing
short‑time working schemes (paragraph 1.7).
10 Given the compressed timeframe to design each scheme, the Departments
could not follow standard processes comprehensively. In the circumstances, the
Departments had insufficient time to produce detailed documentation – such as
business cases, options appraisal and detailed cost-benefit analysis – that we would
normally expect to be available to support key investment decisions. Instead, policy
and operational officials worked closely together to rapidly develop employment
support schemes that they could implement quickly. HMRC agreed clear principles
for both schemes, including that the claim process should be simple and the grant
calculation straightforward. HMRC used existing supplier relationships and contracts
to develop the IT solution (paragraphs 1.5 to 1.6, 1.8, 1.9).
11 HMRC implemented the schemes quickly through strong project
management, risk management and service testing. Detailed planning and project
management were central to rapid delivery. The Departments took a structured
approach to identify and manage risk, using government’s Orange Book on risk
management. HMRC’s IT staff took four weeks to implement the CJRS scheme,
compared with an average of around 18 months that they normally need to deliver
major IT projects. HMRC tested customer journeys and developed guidance for
customers and training for staff to enable the effective operation of the schemes
(paragraphs 1.10 to 1.18 and Figure 3).
8 Summary Implementing employment support schemes in response to the COVID-19 pandemic
12 Other countries, at the start of the pandemic, were able to adapt existing
schemes to deliver support more quickly. We contacted 20 national audit
institutions to understand how their respective governments designed and
implemented similar employment support schemes. Most countries implemented
their schemes faster than CJRS, but almost all had pre-existing arrangements
they could adapt. Countries such as Germany and France already had short‑time
work schemes in place to support companies with salary costs at times of
economic crisis (paragraphs 1.19 to 1.23 and Figure 4).
Supporting the people intended
13 The schemes have been largely successful in protecting jobs through the
lockdown period, with at least 12.2 million people benefitting from support.
The CJRS scheme supported 9.6 million jobs and, at its peak in May, around
30% of the workforce eligible for the scheme across the UK were furloughed.
The SEISS scheme supported at least 2.6 million self-employed people, around
77% of the 3.4 million people potentially eligible for the scheme. The number of
jobs furloughed fell to around five million by the end of July, while unemployment
levels remained broadly stable at around 4% of the workforce. This suggests
the schemes provided an effective bridge during the early phases of the
pandemic, allowing some people to return to work when the national lockdown
eased. However, the number of people on payrolls fell by 0.5 million between
March and April. We found one in five people we surveyed who were in paid
employment were not furloughed but had their pay or hours reduced, presumably
because they could continue to work in some capacity through lockdown without
their employer drawing on CJRS (paragraphs 2.12 to 2.16, Figure 7 and Figure 9).
14 A combination of policy decisions and constraints in the tax system
meant that as many as 2.9 million people were not eligible for the schemes.
People were excluded from the schemes either because of ministerial decisions
about how to target the schemes, or because HMRC did not have data needed
to properly guard against the risk of fraud. The precise number of people needing
help is uncertain because not everyone will have been sufficiently affected by the
pandemic to need financial support. Groups ineligible for support were as follows:
• CJRS: an estimated 1.1 million people were ineligible because HMRC had
limited data to verify claims. HMRC has not estimated the number affected
but third-party estimates suggest around 0.4 million short-term contractors
moving between jobs were ineligible. Additionally, 0.7 million limited company
directors could not claim for company dividends paid instead of salaries.
The tax system treats company dividends as investment income and
HMRC cannot separately identify those payments from other investments.
However, company directors could still claim for earnings registered with
the Pay-As-You-Earn scheme and apply for bounce back loans.
Implementing employment support schemes in response to the COVID-19 pandemic Summary 9
• SEISS: HMRC estimated that around 1.6 million self-employed people did not
meet the scheme’s policy criteria. Of these, 1.4 million people had trading
profit that was less than their non-trading income; 0.5 million people had
a trading profit of £0 or made a loss; and 0.2 million people were ineligible
because their trading profits exceeded £50,000.2
• SEISS: third-parties estimated a further 0.2 million people who were newly
self-employed in 2019-20 were ineligible because they had not yet submitted
a Self Assessment return. Therefore, HMRC did not have verified records on
which to confirm their activity and estimate their income. This figure could
have been greater had lockdown occurred further from the January 2020
deadline for annual tax Self Assessment returns. HMRC intends to introduce
more frequent (quarterly) reporting for self-employed people under Making
Tax Digital from April 2023 (paragraphs 2.6 to 2.11 and Figure 6).
15 The Departments considered the equality implications of their design
decisions. The Departments carried out equality impact assessments for both
schemes. They identified that 12% of self-employed workers had a Black,
Asian or minority ethnic background and undertook work to raise awareness of
the SEISS scheme with stakeholder groups. HMRC’s monitoring data provide
information on the age and gender profiles of people covered by the schemes but
do not report on other protected characteristics, such as ethnicity, as these data
are not necessary for the administration of taxes. The data show that a greater
proportion of younger workers were furloughed. A greater proportion of men were
furloughed initially, but this reduced over time and by September there was little
difference between men and women. In August 2020 HMRC began tendering
for survey and qualitative research to gain additional feedback on the schemes
(paragraphs 2.17 to 2.22 and Figure 10).
16 In the long-term the number of jobs protected will depend on wider
government support. Around two million (9%) workers remained furloughed in
mid-September. Retail, accommodation and food services, manufacturing and
construction have claimed the most financial support from the schemes to date.
Many of these sectors have also utilised bounce back loans. The long-term
impact of the schemes is likely to be difficult to disentangle from the effects of
wider government support and the ongoing impact of COVID-19. The new JSS is
intended to support short-time working between November 2020 and April 2021.
Government has also announced an extension to SEISS over the period and
further furlough support (expanded JSS) for businesses legally required to close
due to lockdown restrictions (paragraphs 2.23 to 2.28 and Figure 11).
Post publication this page was found to contain an error which has been corrected (Please find Published Correction Slip)
10 Summary Implementing employment support schemes in response to the COVID-19 pandemic
Managing fraud and error
17 In implementing the schemes, the Departments accepted there may be a
higher risk of fraud and error than normal, in order to provide rapid financial support
and protect jobs. Recognising the priority placed on speed, the Departments have
tolerated greater risk than normal. Limiting the schemes only to taxpayers with
current records has helped to reduce the risk of certain types of fraud significantly
because applicants had to be known to HMRC and had to be paying tax. However,
the aim to make payments within six working days limited HMRC’s ability to carry
out pre-payment checks. For example, HMRC did not validate some data upfront
or require details of the amounts claimed for every employee before making
payments. It made a systematic assessment of the risks and prospects of recovery
from different groups and prioritised its response for the time constraints it faced.
HMRC did not have the enforcement powers to recover overpayments by the time
the schemes went live. It took a calculated risk that powers would be granted and
received these in July 2020 (paragraphs 3.5 to 3.9 and Figure 15).
18 There is evidence that significant levels of furlough fraud occurred, with
limited controls over employers’ arrangements with employees. Employers
committed furlough fraud if they claimed furlough payments but kept employees
working for them against CJRS rules. HMRC’s fraud hotline has received more
than 10,000 reports, mainly of furlough fraud, but it has not yet carried out survey
work or random sampling to estimate the scale of the issue. Of furloughed people
responding to our survey, 9% admitted to working in lockdown at the request
of their employer, and against the rules of the scheme. Other surveys indicate
between 7% and 34% of furloughed employees surveyed worked at the request
of their employer while furloughed. HMRC concluded it would tackle fraud through
whistleblowing and retrospective compliance work. However, employees would not
have known if their employer was part of the government furlough scheme unless
their employer had informed them. Controls such as contacting employees directly
or publicising which companies claimed furlough payments were considered but
rejected. HMRC concluded it would have been unrealistic to contact employees
because of the large numbers involved, and that publishing a list of employers
risked deterring too many legitimate claims. HMRC intends to publish the names
of employers claiming the JSS scheme and to notify employees through their
personal tax accounts when an employer has claimed JSS (paragraphs 3.10
to 3.21 and Figure 18).
Implementing employment support schemes in response to the COVID-19 pandemic Summary 11
19 The scale of total fraud and error is likely to be considerable, particularly for
CJRS, but HMRC will not know the actual levels for some time. HMRC’s planning
assumption was that total fraud and error could range from 5% to 10% on CJRS,
which would equate to £2.0 billion to £3.9 billion based on payments made by
mid-September. For the first SEISS grant, HMRC’s planning assumption was that
fraud and error could range from 1% to 2%. Both these estimates were largely
assumption-based rather than evidence-based. At the end of September 2020,
HMRC was developing a programme of work to understand the full scale of fraud
and error. It aims to refine its provisional estimates again by the end of 2020 and in
spring 2021 as it undertakes more compliance work and receives more operational
intelligence. HMRC does not expect to have a complete assessment of the total
fraud and error it needs to tackle until the end of 2021 at the earliest. HMRC is
monitoring the levels of organised criminal activity including the stealing of taxpayer
identities and coercion of taxpayers to make fraudulent claims. To date, HMRC has
blocked only £10 million of CJRS claims. It intends to measure its effectiveness in
mitigating the risk in due course (paragraphs 3.20 to 3.28 and Figure 16).
20 HMRC’s initial assessment was that it must divert resources from tax
compliance activities to tackle fraud on these schemes. HMRC estimates it could
deploy around 500 staff to recover £275 million on 10,000 of the most high-risk
CJRS grants awarded. While this offers a high rate of return, HMRC estimates
that redeploying staff will come at a cost to tax revenue because HMRC’s
tax compliance work offers even higher rates of return. In June 2020 HMRC
concluded that its only option to address grant fraud and error was to redeploy
existing, trained staff. It concluded it could not bring in additional staff because
they would take up to 18 months to recruit and train to undertake complex
compliance work, and this would be too long. In October 2020 HMRC told us
that with the announcement of new schemes, and better understanding from the
operation of SEISS and CJRS, it was now planning to use private contractors to
supplement its compliance capacity where necessary. HMRC does not yet know
the scale of fraud and error it needs to tackle to any degree of certainty, and it is
not yet clear for how long its compliance activity will be affected by the COVID-19
pandemic (paragraphs 3.29 to 3.31 and Figure 19).
Conclusion on value for money
21 HM Treasury and HMRC met their objective to rapidly implement the
schemes and the Departments should be commended for making these available
ahead of schedule. The schemes were relatively straightforward to apply for, and
payments quickly reached those who applied. Indications are that this has helped
to protect jobs in the short term and the numbers of people moving from furlough
arrangements back to work are encouraging. However, many other people have
lost earnings and have not been able to access support. The long-term impact
of the schemes will also depend on wider financial support and the ongoing
impact of COVID-19.
12 Summary Implementing employment support schemes in response to the COVID-19 pandemic
22 A key value-for-money test for these schemes will be how far the
Departments can mitigate fraud and error. The pace at which the schemes
were designed and implemented meant the Departments had to accept a greater
risk than normal. As such, there are likely to be considerable amounts of fraud
and error, particularly on the furlough scheme. Limiting applications to existing
taxpayers should have reduced the fraud risk, but HMRC could have done more to
make clear to employees whether their employer was part of the furlough scheme.
In future, the Departments should do more while employment support schemes
are running to protect employees and counter acts of fraud. The Departments will
need to ensure sufficient resources are committed to recover money where it is
cost-effective to do so.
Recommendations
23 To learn from their experience in implementing the employment support
schemes, and to protect taxpayer interests HM Treasury and HMRC should:
a consider how to ensure that reliable data covering as many people as
possible can be used to determine eligibility so that fewer people suffering
loss of income are excluded from future similar schemes;
b monitor how far employment support schemes protect jobs, recognising that
the approach may need to adapt rapidly in response to how the pandemic
evolves over the coming months;
c increase the emphasis on using preventative controls for tackling fraud and
error in the new schemes. Where appropriate for future schemes, carry out
more direct work with employees to ensure employers treat them according
to scheme rules, and increase visibility of which employers use employment
support schemes;
d more quickly assess the total value of error and fraud; and explore the
feasibility of commencing assessment activity earlier for future schemes so
that some testing is undertaken while schemes are live;
e review whether a faster programme of recruitment and training can be
provided for grant compliance staff, recognising that the activity may differ
to tax compliance work; and
f review how to organise HMRC’s compliance response to ensure that
sufficient resources are committed to recover overpayments and fraudulent
payments on both schemes where it is cost-effective to do so.
Implementing employment support schemes in response to the COVID-19 pandemic Summary 13
24 As the COVID-19 pandemic continues, leading to ongoing uncertainty,
the Departments may need to develop longer-term plans to support businesses
and jobs which will involve balancing speed of response with risks to value for
money. This includes targeting support to those who need it, treating employees
in accordance with the scheme rules, and reducing fraud and error. In balancing
these, the Departments should:
g ensure that their consideration of options, including under the JSS and
extended schemes, are sufficiently well-documented to demonstrate how
risks to value for money have been considered and resultant risks are clearly
understood and managed;
h specify how performance and value for money will be judged as the
schemes progress, monitoring outcomes and adapting arrangements quickly
if required; and
i consider how HMRC should organise its systems and capabilities to provide
this targeted support, which may require different choices in digitising the
tax system, more frequent filing of tax return data, better linking of customer
records, and changes to customer services.
14 Part One Implementing employment support schemes in response to the COVID-19 pandemic
Part One
HM Revenue & Customs’ implementation of the
employment support schemes
1.1 In late March 2020 the government announced two UK-wide employment
support schemes in response to the coronavirus crisis. The schemes,
administered by HM Revenue & Customs (HMRC) following HM Treasury’s
direction, were:
• The Coronavirus Job Retention Scheme (CJRS): to provide grant
payments to employers to help firms continue to keep people in employment.
Employers could put workers on temporary leave and the government would
pay them cash grants of 80% of employees’ salaries, up to £2,500 a month.
Phase 1 (CJRS 1) of the scheme operated to the end of June 2020. From
July to October 2020 a modified, second phase (CJRS 2) allowed employers
to bring employees back to work part-time.
• The Self-Employment Income Support Scheme (SEISS): to provide a
grant payment to self-employed individuals whose business has been
‘adversely affected’ by the pandemic. The grant was intended to help
support self‑employed individuals with living costs, and they could continue
to work, start a new trade or take up a new employment. A second SEISS
grant was available to claim from 17 August 2020.
1.2 The schemes are part of a wider government response targeting businesses
and individuals (Figure 1).3
3 Our May 2020 report on government’s response to COVID-19 gives a fuller description of its crisis response.
See Comptroller and Auditor General, Overview of the UK government’s response to the COVID-19 pandemic,
Session 2019-21, HC 366, National Audit Office, May 2020.
Implementing employment support schemes in response to the COVID-19 pandemic Part One 15
Figure 1
The schemes are part of a wider portfolio of government support
The schemes provided grant support for businesses and for individuals during the pandemic
Wider context of government support for businesses and individuals
Support for businesses Support for individuals
Loans and liquidity. Benefits and sick pay.
Grants and other funding support. Direct support to individuals and households.
Additional reliefs. Deferring tax payments.
Support for retaining jobs – Support for self-employed people –
Coronavirus Job Retention Scheme (CJRS). Self-Employment Income Support Scheme (SEISS).
HM Revenue & Customs
employment and income
support interventions
announced March 2020
Coronavirus Job Retention Scheme Self-Employment Income Support Scheme
Phase One (CJRS 1) March to June 2020 Two grants paid between April and October 2020 to
eligible self-employed people who had:
Employers received grant payments for their furloughed
employees up to 80% of their salary, capped at £2,500 • traded in the 2018-19 tax year;
per month. Employer can top up voluntarily.
• trading profits up to £50,000;
Phase Two (CJRS 2) July to October 2020
• at least half of their income from trading sources; and
Employers were expected to contribute an increasing
proportion of costs: • continued to trade in 2019-20 and expected to
continue in 2020-21.
• July and August 2020: 80% of salary, capped
First grant was for 80% of three months of an individual’s
at £2,500 per month. From August, employer
pays employers National Insurance and trading profits capped at £7,500.
pension contributions. Second grant was for 70% of three months of an
individual’s trading profits capped at £6,570.
• September 2020: 70% of salary, capped at £2,187.50
per month. Employer tops up to at least 80%.
• October 2020: 60% of salary, capped at £1,875 per
month. Employer tops up to at least 80%.
Notes
1 For information on the government’s initial response to the Coronavirus pandemic, see Comptroller and Auditor General, Overview of the UK
government’s response to the COVID-19 pandemic, Session 2019-21, HC 366, National Audit Office, May 2020.
2 CJRS 2 is also referred to as the Flexible Furlough scheme.
3 On 24 September 2020, the Government announced that it was creating the Job Support Scheme to replace the Coronavirus Job Retention
Scheme and was extending the Self-Employment Income Support Scheme. Further information on these developments can be found at:
www.gov.uk/government/publications/winter-economy-plan/winter-economy-plan
Source: National Audit Office analysis of HMRC and other government departments’ information
16 Part One Implementing employment support schemes in response to the COVID-19 pandemic
The Departments implemented the schemes quickly and ahead of
their schedule
1.3 Ministers asked HM Treasury and HMRC to develop schemes to protect
jobs and for these to be operational very quickly. The Departments were set clear
targets for each scheme:
• CJRS: to build and launch a claims service before the end of April 2020,
with claims being processed and paid within six working days.
• SEISS: to build and launch a claims service by the beginning of June 2020,
with at least 80% anticipated to claim, and claims processed and paid
within six working days.
1.4 The Departments implemented both schemes ahead of schedule, with CJRS
available to employers from 20 April and SEISS to the self-employed from 13 May.
HMRC accelerated the original timetable for the SEISS scheme by almost two
weeks, recognising the need to provide financial support to eligible taxpayers
as quickly as possible. The two new digital services could be accessed from the
GOV.UK website.
Developing the strategic response
1.5 Initial strategy work focused on CJRS. HMRC and HM Treasury officials
worked collaboratively to develop the response under lockdown conditions,
engaging regularly with senior ministers. Policy and operational staff in both
Departments worked closely to ensure that policy choices were feasible, seeking
to balance the need to implement support for people quickly with the need to
guard against fraud.
1.6 HMRC agreed clear design principles for the schemes including that the
claim process should be simple, that the data used to determine eligibility should
be easy to obtain and the grant calculation straightforward. The Departments
recommended that only people and businesses with an existing taxpayer record
could apply for support to reduce the risk of fraud.
1.7 The Departments had little contingency planning for a pandemic and
limited experience of designing schemes of this type and scale. Contingency
planning for pandemics had focused on the public health response and not
on the economic response. HM Treasury told us it drew instead on economic
contingency planning for financial rescues, draft policy work on wage subsidy
schemes and short-time working schemes, and lessons learned material provided
by German counterparts.
Implementing employment support schemes in response to the COVID-19 pandemic Part One 17
Devising robust plans and governance
1.8 Given the compressed timeframe to design each scheme, the Departments
could not follow standard processes comprehensively. In the circumstances,
the Departments had insufficient time to produce the detailed documentation
– such as business cases, options appraisal and detailed cost-benefit
analysis – that we would normally expect to be available to support key
investment decisions.4 HMRC’s Accounting Officer accepted the Department’s
responsibility for delivering CJRS grants following verbal advice on the legality,
propriety, deliverability and value for money of the scheme from his officials on
20 March 2020, just prior to the Chancellor’s announcement of the scheme.
However, due to time constraints, HMRC’s Accounting Officer did not receive
formal written advice on the legality, propriety, deliverability and value for money
of the scheme until after the Chancellor’s announcement.
1.9 HMRC moved quickly to put effective governance arrangements in place, with
core project teams set up to work on both schemes. It used its experience from
EU Exit work and the Statutory Sick Pay Rebate scheme to inform its approach.
HMRC deployed staff familiar with EU exit governance arrangements, with high
levels of engagement from senior staff. HMRC appointed a senior responsible
officer at director-general level and created a dedicated director‑led COVID-19
response unit. There was a clear escalation route to HMRC’s senior management
with its Accounting Officer involved in decision-making. HM Treasury officials
worked with HMRC programme leads and attended meetings. HMRC’s Internal
Audit function which provided support and assurance to the schemes noted no
significant issues with the Departments’ governance approach.
Implementing the schemes
1.10 HMRC implemented a structured approach to managing and recording
project progress, risks, issues and decision-making for the schemes, making
use of the government’s Orange Book on risk management best practice.
HMRC also consulted with other government departments familiar with good
practice in administering grant programmes.
1.11 In challenging circumstances, HM Treasury and HMRC staff worked long
hours, at weekends and through the Easter period to develop the schemes.
HMRC made use of its existing supplier contract to build the digital systems
for the schemes. The demands on the Departments’ core project teams were
considerable with its IT project team reporting it completed the work in just
four weeks. HMRC’s IT staff told us major IT projects normally take 18 months
to 21 months to deliver, while small projects normally take between 8 months
and 13 months.
4 See HM Treasury, Guide to Developing the Business Case, 2018 at: https://assets.publishing.service.gov.uk/
government/uploads/system/uploads/attachment_data/file/749086/Project_Business_Case_2018.pdf
18 Part One Implementing employment support schemes in response to the COVID-19 pandemic
1.12 HMRC developed comprehensive customer journey mapping which it tested
internally and with tax agents and developed guidance for customers and staff,
providing additional online training.
1.13 HMRC identified that developing the schemes at pace had caused some
difficulties. For example:
• Record-keeping of key documents such as process maps was not always
controlled and internal reviewers found gaps in the logging of project risks,
although HMRC told us these omissions were minor.
• Guidance to front-line staff was sometimes issued with little notice due to
the delivery timescales and HMRC identified issues with staff compliance
with the guidance. Changes to guidance were not always communicated
clearly, meaning staff used it inconsistently, which increased the risk of
fraud and miscommunication. Some staff failed to take customers through
security processes or read out key statements to customers.
• The planned model for processing claims did not work as anticipated at first.
HMRC planned a three-tiered approach in which certain staff would process
simple claims, with more experienced staff at a second and third tier dealing
with more complicated or higher risk claims. Roles and responsibilities
for each tier were unclear, resulting in customer calls being unnecessarily
escalated and placing pressure on upper tiers. At one point, 80% of SEISS
calls were handed to a higher tier. HMRC believes that earlier involvement
of front-line operations teams in the work of the project development teams
would have improved implementation. It revised its three-tier model from
June 2020 and told us that it became more efficient. HMRC has also since
taken steps to clarify advisor roles and responsibilities.
Measuring and evaluating performance
1.14 HMRC developed performance dashboards for reporting CJRS and SEISS
progress to support management decision-making, and it published statistics
on the take-up of schemes. Dashboards were reported regularly to HMRC’s
executive management with information on scheme take-up, claim values,
project risks and ‘risk profiling’ of fraudulent claims.
1.15 HMRC achieved most of its early ‘ambition criteria’ for the schemes
(Figure 2). The vast majority of payments were processed within target
timeframes, meeting the key objective to get financial support to claimants
quickly. HMRC was able to maintain processing performance and meet payment
targets despite, in the week of 11 May 2020, receiving over 2.5 million claims for
support under the two schemes.
Implementing employment support schemes in response to the COVID-19 pandemic Part One 19
Figure 2
HMRC performance against its early performance criteria
HM Revenue & Customs (HMRC) achieved most of its early performance criteria
Coronavirus Job Retention Scheme (CJRS)
Criteria Performance
Build, test and launch a claims service by CJRS launched on 20 April 2020.
20 April 2020.
Make vast majority of payments within six 99.5% of claims were processed within three
working days, ensuring a risking window of working days.3
no more than 72 hours.1
Deliver payments by 30 April 2020 for those 99.8% of claims were paid.
who claimed by 23 April 2020.
Enable access for agents.2 Agents able to apply for CJRS on behalf
of employers.
Self-Employment Income Support Scheme (SEISS)
Criteria Performance
Build and launch a claims service by SEISS was launched on 13 May, well ahead
early June. of schedule.
80% take-up from eligible customers. 77% claimed under SEISS 1 by 31 July.
Customers paid within six working days. 97.5% of SEISS payments were made within
six working days.3
Notes
1 A ‘risking window’ is the period allowed for HMRC to assess the risk of fraud attached to a claim and decide
whether to investigate the claim further.
2 HMRC permits agents to manage tax affairs on behalf of an individual or employer. An agent can be a
professional accountant or tax adviser, a friend or relative, or someone from a voluntary organisation.
3 Payments made by bank transfer (BACS) have a 72-hour processing window which meant that most claimants
would have received their payment up to six working days from application.
4 Performance data is for the period from scheme launch to August 2020.
Source: National Audit Office analysis of HM Revenue & Customs information
20 Part One Implementing employment support schemes in response to the COVID-19 pandemic
1.16 HMRC carried out surveys of employers, employees and self-employed
people to understand levels of scheme awareness and understanding, and
customer experience. Most (94%) respondents to HMRC’s survey of just over
2,000 employers who had applied to the scheme said they were very or fairly
satisfied with the time taken to receive their money and only 1% described their
experience of applying for CJRS funds as very poor. HMRC’s weekly customer
satisfaction data for SEISS up to August 2020 found that 95% of claimants
were satisfied or very satisfied with their experience, with only 2% being
very dissatisfied.
Lessons learned
1.17 HMRC undertook lessons learned exercises for both schemes in late spring
and used findings to inform its ongoing delivery. Figure 3 uses our own analysis
of learning from previous crisis responses to assess the Departments’ approach
to designing and implementing these schemes. We saw particular strengths in the
following areas:
• close working between policy development and operational experts, enabling
better understanding of design parameters and inclusion of operational
aspects of policy in plans from the outset;
• clear governance structures, drawing on proven models;
• detailed control frameworks, used to assess and prioritise controls through
the design and implementation of the schemes (albeit with some gaps);
• detailed project management plans, which had milestones specified to the
hour at critical points; and
• early mapping of customer journeys and testing of the approach through
process walkthroughs and beta-testing with customers.
1.18 HMRC is commissioning further research on its delivery of the
schemes and expects to use these results to inform its design of future services.
The Departments have not said whether they will commission an economic
evaluation of the additional impact of the schemes. HM Treasury told us that it
is conducting its own lessons learned exercises to identify how such employment
support schemes could be delivered at speed in the future.
Implementing employment support schemes in response to the COVID-19 pandemic Part One 21
Figure 3
Assessing the Departments’ implementation of the schemes using our core
management cycle, adapted for a crisis scenario
The Departments achieved a lot in a short time, but lessons can still be learnt
Stage Factors indicating an effective response in a crisis2 NAO assessment
Strategy Contingency plans in place before crisis.
Clear objectives and success criteria.
Use evidence and work with others.
Develop exit strategy early.
Planning and governance Assure business plans where possible.
Establish clear ownership and governance.
Promote stability in, and manage demands on,
the core response team.
Clear and consistent communications.
Implementation Robust project management.
Effective controls and risk management,
document decisions.
Robust guidance for frontline staff and resources
deployed responsively.
Measurement Appropriate indicators of performance.
Collect data, improve arrangements over time and
use several sources.
Use data continuously in a crisis.
Evaluation and feedback Evaluate during and after the crisis, identify
lessons learnt and use these to modify crisis
intervention and planning for next time.
Notes
1 Our classification of the Departments’ performance uses a three-point scale: Two ticks = good evidence,
one tick = partial evidence. A cross indicates that we have not seen clear evidence or have seen evidence
of poor performance in relation to the evaluative criteria.
2 We have used the NAO’s core management cycle to organise our evaluative criteria drawing from our past
reviews of government responses to humanitarian and financial crises. The management cycle sets out the
main stages that departments should consider in designing and implementing programmes and highlights
the importance of using evaluation and feedback.
Source: National Audit Office analysis of HM Revenue & Customs information
22 Part One Implementing employment support schemes in response to the COVID-19 pandemic
Most other countries had existing employment support schemes,
and were able to implement support more quickly
1.19 The Departments looked at the experience in other countries to inform the
design of the schemes. The experience of other countries can also be used to
assess how the schemes could have been implemented more quickly, and the
types of support arrangements that enable a more rapid response.
1.20 In developing CJRS, the Departments referred to responses in other
countries in their discussions with ministers. They looked at several schemes
including Germany’s short-time work scheme, schemes in France, Canada,
Denmark, Sweden, New Zealand and Ireland but noted their direct application
to the UK might be limited given the different labour market and welfare contexts.
1.21 Many countries including Germany and France already had short-time
work schemes in place that could be built on in response to the pandemic.
These schemes allowed firms to temporarily furlough or reduce the hours of staff
at times of economic difficulty, with governments providing income support for
employees when not working. Based on correspondence with 20 international
audit institutions and other supporting evidence, we looked at how quickly a
sample of nine OECD countries with economies of comparable size to the UK
had implemented their job retention schemes (Figure 4). We found that other
countries implemented employment support schemes more quickly, but most
of these already had an existing scheme in place. Eight of the nine had schemes
they could adapt, while six reported setting up new schemes.
1.22 Our analysis suggested that the UK had relatively few eligibility criteria
for CJRS compared with other countries. Most countries required employers to
demonstrate a specified fall in turnover, and many required employers to top-up
wages or prohibited redundancies while the scheme was active. The Netherlands
prohibited executive bonuses, payment of dividends and repurchase of
own shares. The UK CJRS scheme did not make any such requirements of
employers (Appendix Three).
1.23 On 24 September 2020 the Chancellor announced a Winter Economy Plan
which included a six-month Jobs Support Scheme (JSS), starting in November, to
replace the CJRS scheme. The JSS scheme aims to provide help for short-time
working to support ‘viable jobs’. To qualify, employees must be working at least
one-third of their usual hours. The government will pay one-third of hours not
worked up to a cap, with the employer also contributing one-third. This will mean
employees earn a minimum of 77% of their normal wages, where the government
contribution has not been capped. The level of grant will be calculated based
on employee’s contracted salary, capped at £697.92 per month. The JSS will
be open to businesses across the UK even if they have not previously used the
furlough scheme.
Figure 4
International comparison: time to implement job retention schemes
Other countries implemented support more quickly, but most already had pre-existing schemes they could adapt
Country Scheme details
UK New short-time
UK UK UK
work scheme
Australia New wage
AU AU AU
subsidy scheme
Canada New wage subsidy
CA CA CA scheme (pre-existing
short-time work scheme)
Denmark New short-time work
DK DK DK scheme (pre-existing
short-time work scheme)
Ireland New wage subsidy
IE IE IE scheme (pre-existing
short-time work scheme)
Netherlands New wage subsidy
NL NL NL scheme (pre-existing
short-time work scheme)
US New wage subsidy
US US US scheme (pre-existing
short-time work scheme)
France Pre-existing short-time
FR FR FR
work scheme
Germany Pre-existing short-time
DE DE DE
work scheme
Sweden
Pre-existing short-time
SE SE SE
work scheme
w/c 9 March w/c 16 March w/c 23 March w/c 30 March w/c 6 April w/c 13 April w/c 20 April w/c 27 April
Announced Passed into law Open to applicants
Notes
1 ‘Short-time work schemes’ and ‘wage subsidy schemes’ are different types of job retention scheme, designed to mitigate unemployment increases. Short-time work schemes subsidise employers
for hours not worked by employees, while wage subsidy schemes subsidise employers for hours that are worked by employees or top-up the earnings of employees on reduced hours.
2 Where a country has both new and existing schemes, the dates relate to the implementation speed of the new scheme.
3 For the UK, the Coronavirus Act 2020 was passed into law on 25 March 2020, which empowered the Chancellor to issue a Direction on 15 April 2020 from
HM Treasury to HM Revenue & Customs to develop and implement the Coronavirus Job Retention Scheme.
Source: National Audit Office analysis of data provided by the Supreme Audit Institutions contacted. OECD (2020) report: Job retention schemes during the COVID-19 lockdown and beyond.
Institute for Government report (2020): Coronavirus and unemployment: a five nation comparison. Government websites
Implementing employment support schemes in response to the COVID-19 pandemic Part One 23
24 Part One Implementing employment support schemes in response to the COVID-19 pandemic
1.24 On 9 October 2020, the government announced an expansion to the
JSS to provide extra support to businesses legally required to close as a direct
result of coronavirus restrictions set by one or more of the governments in
the UK. Expanded JSS is broadly akin to the original CJRS with employers
expected to fully furlough employees. Employers are not required to contribute
to their employees’ wages apart from national insurance and minimum pension
contributions, where applicable. The government will pay up to two-thirds of
employee wages, capped at £2,100 per month.
1.25 In September, the government also announced that it would provide a further
taxable grant to self-employed people currently eligible for SEISS who continue
to actively trade and face reduced demand due to coronavirus. The initial lump
sum will cover three months’ worth of profits for the period from November to the
end of January 2021. The grant is worth 20% of average monthly profits, up to a
total of £1,875. An additional second grant, which may be adjusted to respond to
changing circumstances, will be available for self-employed individuals to cover
the period from February 2021 to the end of April 2021. Appendix Four sets out
the different levels of support for these schemes.
Implementing employment support schemes in response to the COVID-19 pandemic Part Two 25
Part Two
Beneficiaries of the employment support schemes
2.1 This part considers how far the employment support schemes have
supported businesses and people. It examines:
• how the schemes aimed to support business and individual taxpayers;
• why some existing taxpayers were ineligible for the schemes;
• the cost and impact of the schemes thus far;
• the extent to which the schemes have supported different demographic
groups, regions and sectors; and
• the number remaining furloughed as the current schemes wind-down.
How the schemes aimed to support business and individual taxpayers
2.2 On 20 March, the Chancellor announced a package of economic
support aimed at protecting jobs through the lockdown period, including the
Coronavirus Job Retention Scheme (CJRS). This was followed on 26 March by
his announcement of the Self-Employment Income Support Scheme (SEISS).
The broad aim of the employment support schemes was to provide security for
employers to retain employees and to protect the self-employed so that there
would not be a large spike in unemployment and to give the UK the best chance
of a rapid economic recovery.
2.3 Figure 5 overleaf summarises the scope of the schemes. For the first phase
of CJRS (CJRS 1), furloughed employees were not allowed to work, to avoid
government supporting those employers not adversely affected by COVID-19.
For CJRS 2, from July 2020, these aims shifted to allow furloughed workers
back on a part‑time basis, encouraging firms to increase production and, where
necessary, make redundancies if jobs had become unsustainable. Employers not
already in the CJRS 1 scheme were not allowed to claim CJRS grants during this
second phase.
26 Part Two Implementing employment support schemes in response to the COVID-19 pandemic
Figure 5
Policy objectives for CJRS and SEISS and their key eligibility criteria
Eligibility rules vary across the HM Revenue & Customs’ (HMRC’s) schemes
Coronavirus Job Retention Scheme (CJRS) Self-Employment Income Support Scheme (SEISS)
Policy objective
CJRS 1: Provide rapid financial support to help firms Income support for self-employed people adversely
continue to keep people in employment. affected by the pandemic.
CJRS 2: Give firms flexibility to support the recovery and
from August, introduce an employer contribution.
Target private sector employers.
Cut-off date to be
registered in HMRC 19 March 2020 23 April 2020
tax records to Real Time Information employer return including Extended deadline for 2018-19 Self Assessment tax
be eligible furloughed employees. return. Original deadline was 31 January 2020.
Permitted working
CJRS 1 to 30 June 2020: Furloughed employees should Self-employed people can continue to work and claim a
not work. SEISS grant as long as their trade is adversely affected.
CJRS 2 from 1 July 2020: Furloughed employees can work
for any amount of time or shift pattern, with employers
claiming CJRS grant for employee hours not worked.
Employers only claim under CJRS 2 if there was a valid
claim under CJRS 1.
Income or
No limit on employee earnings but maximum employer Trading profits of up to £50,000 and maximum
trading profit
grant is capped. grant capped.
Employee can receive multiple support across
several employments.
Excluded income
Discretionary elements of pay such as tips, bonuses Trading profits only – excludes other income from
and commission. earnings, property, dividends, savings, pension etc.
Non-monetary benefits like benefits in kind or benefits Trading profits at least 50% of non-trading income.
received under salary sacrifice schemes that reduce
taxable pay.
Extensions
to eligibility Employee must have been furloughed under CJRS 1 to be Can claim the second SEISS grant even if did not claim
furloughed under CJRS 2 from 1 July 2020. the first.
Deadline extended for employees to be registered on HMRC Parents whose trading profits were affected in 2018-19
records from 28 February to 19 March 2020. following recent birth or adoption.
Parents returning to work after extended leave and military Military reservists carrying out reservist activities for at
reservists returning to their employment after 10 June 2020 least 90 days in 2018-19 tax year.
(usual deadline for CJRS 1) allowed to be furloughed for first Individuals subject to the loan charge arrangement.
time under CJRS 2.
Notes
1 Example eligibility criteria shown only. Other eligibility rules apply for both schemes.
2 Taxpayers previously subject to the loan charge arrangement and who had been granted until 30 September to submit their
2018-19 tax return were allowed to apply for SEISS. Also see Gov.uk, Update on the implementation of the loan charge, August 2020 at:
www.gov.uk/government/news/update-on-the-implementation-of-the-loan-charge
Source: National Audit Office analysis of HM Revenue & Customs’ and other government departmental information
Implementing employment support schemes in response to the COVID-19 pandemic Part Two 27
2.4 The SEISS scheme was open to the self-employed and members of a
partnership who claimed to have been ‘adversely affected’ by coronavirus and:
• had submitted their Self Assessment tax return on or before
31 January 2020 for the 2018-19 tax year;5
• continued to trade in 2019-20; and
• intended to trade in 2020-21.
2.5 Those claiming the second grant had to confirm that their business had
been adversely affected on or after 14 July 2020. Unlike the first phase of CJRS,
SEISS recipients were allowed to continue working.
A combination of constraints in the tax system and policy decisions
restricted eligibility for the schemes
2.6 Combined, the employment support schemes had supported at least
12.2 million jobs by 20 September 2020. HM Revenue & Customs’ (HMRC’s) data
show that 1.2 million employers – 61% of those eligible – claimed a grant under
CJRS. By 20 September 2020, CJRS had supported 9.6 million jobs. Of 3.4 million
people potentially eligible for the SEISS scheme, 2.6 million (77%) claimed the
first grant and 2.2 million had applied for the second grant by 20 September on the
basis that the pandemic adversely affected their self-employed incomes.
2.7 HMRC has estimated that 1.6 million people were ineligible for SEISS
because of the decisions ministers took to restrict support to people whose
main income came from self-employment and who earned up to £50,000.6
However, other groups of people were rendered ineligible for support because
the way the tax system has been set up limited the data HMRC could use to help
it guard against fraud (Figure 6 overleaf). Third-party estimates suggest a further
1.3 million people were excluded, although some may not have been adversely
affected by the pandemic.
2.8 To manage the risk of fraud, access to both schemes was restricted to
taxpayers with current tax records, which meant that people who were between
jobs or had recently become self-employed were not eligible. HMRC extended the
cut-off date for CJRS to people being on an employer’s payroll from 28 February
to 19 March 2020 – the day before the details of the scheme were publicly
announced. HMRC estimates this benefitted at least 0.2 million extra employees.
5 A self-employed person or partnership could be adversely affected by coronavirus for a variety of reasons such
as illness or suspended trading, but no minimum level was specified. HMRC later extended the Self Assessment
tax return deadline to 23 April 2020 – see paragraph 2.9.
6 £50,000 is the Higher rate income tax threshold and the threshold at which Child Benefit tapers take effect.
28 Part Two Implementing employment support schemes in response to the COVID-19 pandemic
Figure 6
Groups excluded from the schemes
Significant numbers of people were potentially not able to access the schemes, many because of tax system constraints
Scheme Ineligible - Policy1 Ineligible - tax system constraints2 Total
Self-Employment 1.4 million 0.2 million
Income Support (less than 50% income from self-employment) (newly self-employed)
Scheme
(SEISS) 0.5 million
(self-employment was loss making)
0.2 million
(income more than £50,000)
1.6 million3 0.2 million 1.8 million
Coronavirus Job 0.7 million
Retention Scheme (limited company directors paid
(CJRS) by dividend)
0.4 million
(short-term or freelancer workers
missing payroll)
1.1 million 1.1 million
Total 1.6 million 1.3 million 2.9 million
Notes
1 Policy decisions are those taken by the Departments to meet scheme objectives, including to target those people most likely to have
been adversely affected by the pandemic. For example, for SEISS the decision that eligible people should earn 50% or more of
their income from self-employment.
2 The principle that HMRC should use existing data on taxpayers for administrative ease and to help manage fraud risk also meant
that some people were not eligible for support because HMRC did not hold sufficient tax records to support a claim.
3 For SEISS: The 1.6 million customers ineligible for SEISS on policy grounds is an HMRC estimate. A self-employed person could have multiple
reasons for not being eligible for SEISS and could be in more than one of the three categories identified by HMRC. The 0.2 million newly
self-employed ineligible is based on Office for National Statistics (ONS) data on the number of individuals becoming self-employed in 2019-20.
4 For CJRS: The 0.7 million limited company directors is an estimate provided in evidence to the June 2020 Treasury Select Committee and
incorporated into its report. However, company directors could still claim for Pay-As-You-Earn registered earnings and apply for bounce back loans.
Meanwhile, in its June 2020 report the Treasury Select Committee estimated the population of short-term or freelancer workers to be 0.8 million.
ExcludedUK estimated 0.4 million of these were excluded.
5 A proportion of the groups listed here would not have claimed because they were able to continue to work.
Source: National Audit Office analysis of HM Revenue & Customs, Office for National Statistics, Treasury Select Committee and
ExcludedUK information
Implementing employment support schemes in response to the COVID-19 pandemic Part Two 29
2.9 For SEISS, HMRC used self-assessment records up to and including
the 2018-19 tax year, which meant that claims from around 0.2 million newly
self‑employed people were not allowed. Around 47,000 self-employed people who
had not submitted their 2018-19 Self Assessment tax return on time were allowed
to claim provided they submitted their return by 23 April 2020. However, people
who had already submitted their Self Assessment returns were not allowed to
amend them in order to increase their SEISS payment.7 In addition, a significant
number of limited company directors who were mainly remunerated through
company dividends were excluded because this is not treated as earnings by
the tax system and tax returns do not disaggregate these data from investment
dividends. Very small (micro) businesses and private limited companies have
taken up the majority of business bounce back loans, suggesting many company
directors have been able to access alternative government support on behalf of
their companies, albeit in the form of repayable loans rather than income.8
2.10 The number of self-employed people excluded could have been greater
had lockdown occurred further from the annual tax return deadline. Lockdown
occurred just two months after the deadline to submit a Self Assessment return,
which meant the time period affecting the newly self-employed was relatively
small. HMRC’s Making Tax Digital programme should enable more frequent
(quarterly) reporting of income and expenditure. However, modernising the
tax system is an extremely complex undertaking. HMRC was originally due to
implement the programme by April 2019 but it is now scheduled to be in place
from April 2023 for the self-employed.
2.11 Given the speed of their initial response to the crisis, the Departments
were not able to fully work through the implications of their design decisions on
different groups of taxpayers. Subsequently, the Departments adjusted processes
and eligibility criteria, including:
• updating the online ‘eligibility checker’ – used by the self-employed to
determine whether they would be eligible for SEISS. The details of some
people who were late in filing tax returns were not available when the
tool went live, and the checker concluded, wrongly, that some people
were ineligible;
• allowing people on parental leave and around 200 military reservists to
apply for both SEISS grants, even if they had not submitted a tax return
in 2018-19. HMRC estimated that 10,000 people on parental leave may
potentially be newly entitled by this adjustment but by late September only
around 500 had applied; and
• HMRC manually entered the trading profits of 800 self-employed people
to avoid a known technical issue preventing submission of their tax returns.
7 Taxpayers could amend their returns for the calculation of tax due but amendments were disregarded for grant
calculations to counter the risk of fraud.
8 Comptroller and Auditor General, Investigation into the Bounce Back Loan Scheme, Session 2019-2021,
HC 860, National Audit Office, October 2020.
30 Part Two Implementing employment support schemes in response to the COVID-19 pandemic
The cost and impact of the schemes
2.12 HMRC forecasts total grant spending for the CJRS and SEISS
schemes to reach £69.7 billion by the end of October 2020. It paid a total of
£39.3 billion in CJRS grants to 20 September 2020 and the Office for Budget
Responsibility (OBR) expects this to reach £54.5 billion by the end of the
scheme.9 By 20 September 2020, SEISS claims had reached £13.4 billion.
OBR forecasts total claims for the first two SEISS grants will reach £15.2 billion.
2.13 HMRC expects to spend around £98 million administering the schemes up
to March 2021. The CJRS scheme will cost around £53 million to administer, and
the SEISS scheme £45 million.
2.14 Data suggest the CJRS scheme has been successful in helping to protect
jobs in the short term:
• The number of people in employment was broadly stable between April and
August 2020 (Figure 7).
• The number of redundancies between April and June 2020 was far lower
than in the last global financial crisis: around 134,000 people were made
redundant between April and June 2020 compared with 311,000 people
made redundant between February and April 2009.
• The number of furloughed jobs under the scheme has fallen steadily from
a peak of 8.9 million in May 2020 to around five million by the end of July,
while unemployment levels remained broadly stable at around 4% of the
workforce. The absence of a significant deterioration in employment up
to the end of July suggests that many furloughed employees returned to
work (Figures 7 and Figure 8 on page 32). The Office for National Statistics
(ONS) data also show that the furlough rate had fallen from around 30% of
the workforce of UK businesses in May 2020 to 9%, or around two million
employees, by mid‑September.10
9 Grants paid under CJRS and SEISS are taxable so the net cost to the Exchequer will be lower. Table 3.4 in
OBR’s July 2020 central estimate shows tax receipts of £5.2 billion for CJRS: Office for Budget Responsibility,
Fiscal sustainability report, July 2020, p58, available at: https://obr.uk/fsr/fiscal-sustainability-report-july-2020
10 Office for National Statistics, Business impact of Coronavirus (COVID-19) Survey, Waves 6 and 14, June and
October 2020, available at: www.ons.gov.uk/economy/economicoutputandproductivity/output/datasets/
businessimpactofcovid19surveybicsresults. Based on businesses that had not permanently stopped trading.
This survey source is different from HMRC’s own data on CJRS scheme take-up.
Implementing employment support schemes in response to the COVID-19 pandemic Part Two 31
Figure 7
People in employment and employments furloughed under the Coronavirus Job Retention
Scheme (CJRS), January to August 2020
Following a drop early on, the number of people in employment was stable during lockdown while the number of furloughed
jobs (employments) has since reduced from the April/May peak
Employees/employments (m)
35
30
29.0 29.0 29.0 28.5 28.4 28.4 28.3 28.3
25
20
15
10
8.7 8.7
7.5
5 6.2 5.6
0
January February March April May June July August
2020
UK Pay-As-You-Earn (PAYE) employees
Average employments furloughed under CJRS
Notes
1 The number of PAYE employees is based on experimental data from the Office for National Statistics (ONS), and HM Revenue & Customs (HMRC),
using HMRC’s Real Time Information (RTI). This shows the number of employees on employer payrolls in the UK.
2 The number of employees furloughed is based on publicly available HMRC CJRS statistics. We have calculated a monthly average of daily
CJRS data for this figure between March and July 2020. Our average figure for March 2020 is only based on figures from 23 March 2020
(the first working day after the announcement of CJRS) to 31 March 2020. The July 2020 figure of 5.6 million is an estimate based on
HMRC’s latest published data which recorded an average 5.1 million cases across July but also estimated an uplift of around 10% in the final
number for figures at 31 July.
3 There is a distinction between the number of employees outlined in the PAYE data, and the number of employments furloughed under CJRS;
an individual is counted once as a PAYE employee, but may work in two roles furloughed under CJRS.
4 Figures for the number of employees furloughed under CJRS in August 2020 were not available at the time of publication.
Source: National Audit Office analysis of Office for National Statistics’ records of HM Revenue & Customs’ Real Time Information employment data and
HM Revenue & Customs’ CJRS data
32 Part Two Implementing employment support schemes in response to the COVID-19 pandemic
Figure 8
Employments placed on furlough during the Coronavirus Job Retention Scheme (CJRS)
The number of furloughed employments peaked at 8.9 million on 8 May 2020
Employments furloughed (m)
10
8.9
9
8.4
8
7.7
7
5.9
6
5.3
5
4
3
2
1
0
1 Mar 1 Apr 1 May 1 Jun 1 Jul
2020 2020 2020 2020 2020
Recorded employments furloughed
HM Revenue & Customs estimate ofJuly employments furloughed
Notes
1 This figure shows the number of employments or jobs placed on furlough each day during CJRS. In total 9.6 million individual employments were
placed on furlough during CJRS 1, but the highest amount of employments or jobs on furlough on a single day was 8.9 million on 8 May 2020.
2 HM Revenue & Customs’ (HMRC’s) data for July 2020 are incomplete as not all claims and revisions have been submitted. HMRC estimates that
the final level of claims for 31 July will be 10% higher than its current figure. Figures for the rest of July may also be incomplete. The secondary
line in our chart inflates July’s current HMRC figures by 10% to give an indicative upper estimate of the level of claims for that month.
3 CJRS was announced on 20 March 2020 and the first claims could be made on 20 April 2020. Claims could be backdated to 1 March 2020.
Source: National Audit Office analysis of HM Revenue & Customs’ CJRS statistics, August 2020
Implementing employment support schemes in response to the COVID-19 pandemic Part Two 33
2.15 However, a 0.5 million reduction in the number of people on employer
payrolls in HMRC’s Real Time Information tax reporting system between
March and April 2020 does suggest that some people were not covered by the
support provided by these schemes at the outset. Between March and May 2020,
the number of people not in employment on Universal Credit also increased by
nearly 1.5 million. We surveyed employees to better understand their experiences
during the first months of the CJRS scheme (see Appendix Two on the limits
on the quota sampling methodology). More than half of respondents to our
questions said that their main job had been impacted by COVID-19, with more
than one‑quarter saying they had been placed on the furlough scheme and 4%
saying that they had been made redundant between March and June (Figure 9
overleaf).11 Our survey also suggests that the initial requirement that furloughed
employees should not work did help to keep costs down – around one-fifth of
respondents were not placed on the furlough scheme, despite their hours of work
or pay being reduced, presumably so they could keep working.
2.16 It is likely that the schemes will have supported some firms which would
have otherwise failed, even without the pandemic. Data from the Insolvency
Service shows that company insolvencies fell by around one-third between
April and June 2020, compared with the previous year.12 The long-term impact
of the schemes to protect jobs is unlikely to be known for some time, as firms
struggle with the ongoing impact from the COVID-19 pandemic.
HMRC is commissioning further research to better understand
take‑up of the schemes by different groups
2.17 HM Treasury considered the equality implications of the schemes as part
of early work on their development. For CJRS, it noted that employers could
discriminate between workers but concluded that furloughed workers would
retain their employment rights. It also highlighted a risk that some women moving
on to maternity leave could be affected if their reduced furloughed earnings were
used in the calculation of maternity pay.
2.18 For SEISS, HM Treasury recognised that some young people might be
disadvantaged by the requirement for self-employed people to have traded in the
2018-19 tax year. It also noted that people of Pakistani or Bangladeshi heritage
are more likely to be self-employed than the general population.
11 Based on 4,498 responses. We asked our survey questions as part of Ipsos MORI’s telephone and online
omnibus surveys. The quota sampling methodology used and the limits of the approach in terms of how
generalisable the results are to the wider population are set out in Appendix Two.
12 Some of this reduction will likely be because of reduced operational running of courts, temporary restrictions
placed on the use of statutory demands and certain winding-up petitions during this time.
Figure 9
The impact of COVID-19 on employees
Just over one-quarter of people responding to our survey said they were placed on furlough
Weighted proportion (%)
50
45
40
35
30
25
20
15
10
5
0
No impact I was put I was working, I was working I was still I was asked/ I was made My other Don’t know Refused
on furlough but the hours I my typical working, told to take redundant company
under the was expected hours, but my but the unpaid benefits were
government to work were rate of pay hours I was holiday/ reduced
scheme reduced to was reduced expected to annual (eg pensions,
between 50% work were leave (not medical,
and 99% reduced to furloughed) insurance
of the hours less than etc)
I typically 50% of
worked the hours
I typically
work
Proportion 47 27 9 8 8 5 4 4 1 1
(weighted (%))
34 Part Two Implementing employment support schemes in response to the COVID-19 pandemic
Notes
1 We commissioned Ipsos MORI to ask survey questions on our behalf as part of its telephone omnibus and an online survey. Graph based on 4,498 survey respondents who reported
being in some form of paid employment (full or part-time) on 19 March 2020. The quota sampling methodology used and the limits of the approach in terms of how generalisable the
results are to the wider population, together with the approach taken to weight results to ensure they are representative of the population is set out in Appendix Two.
2 Respondents were able to give multiple responses where, for example, their circumstances changed over time. The sum of the percentages shown is therefore greater than 100%.
3 Survey question: Thinking specifically about the period between 19th March – 30th June 2020, in your main job, which, if any, of the following happened as a result of the impact
of the coronavirus? Please say yes or no to each one as I go through them.
Source: Data from omnibus survey questions commissioned by the National Audit Office and undertaken by Ipsos MORI
Implementing employment support schemes in response to the COVID-19 pandemic Part Two 35
2.19 HMRC’s further analysis to support its communication strategy showed that
12% of self-employed workers had a Black, Asian or minority ethnic background,
with almost half based in London. HMRC developed a range of communication
and engagement plans for the schemes and worked with ethnic minority and
faith-based organisations to publicise the schemes.
2.20 For SEISS, HMRC has no data to monitor rates of take-up among different
ethnic groups. Other than age and gender, protected characteristics’ data
are not needed to calculate income tax. HMRC has also monitored take-up of
SEISS for “customers who may need extra support to claim” who it identified
from people using its ‘extra support’ service over the past 18 months. It also
monitored potential claim rates among the ‘digitally excluded’ using data
on Self Assessment ‘paper filers’.
2.21 HMRC is planning further survey work to understand customer experience
of the schemes across different claimant groups but most results from this
work will not be available until after the current schemes are closed. HMRC
also intends to survey employers eligible for CJRS who did not use the scheme.
No equivalent work is planned for those who did not claim SEISS.
2.22 HMRC has published extensive data on take-up for both schemes
(Figure 10 overleaf). This data shows that a greater proportion of younger
workers were furloughed compared with older workers, and that a greater
proportion of men were furloughed overall, but this has reduced over time and
by September there was little difference between men and women. For SEISS,
a greater proportion of potentially eligible men took up the grant than did women.
ONS data show that younger people were also far more likely to lose their jobs;
there was a 5.9% decrease in employment levels of people aged 16–24 in June
to August 2020 against the previous quarter, compared with a 0.5% decrease
in overall employment levels.
2.23 HMRC’s data also include sectoral and regional information which enables
an understanding of who has been using the schemes and in which locations.
These data show that a greater proportion of employees from smaller employers
were furloughed. Business sectors in retail, construction, accommodation and
food services, manufacturing and transportation drew the largest amounts of
financial support. There is close alignment between the sectors making greatest
use of CJRS, SEISS and the Bounce Back Loan Scheme (Figure 11 on page 37).
Post publication this page was found to contain an error which has been corrected (Please find Published Correction Slip)
36 Part Two Implementing employment support schemes in response to the COVID-19 pandemic
Figure 10
CJRS and SEISS: HM Revenue & Customs’ (HMRC’s) data on take-up of schemes, end July 2020
An uneven picture across industries, employers, individuals and geographies
Coronavirus Job Retention Scheme (CJRS) Self-Employment Income Support Scheme (SEISS)
Overall take-up
Some 1.2 million employers (61%) received a grant Some 2.6 million self-employed claimed (77% of HMRC’s
covering 9.6 million employments (32%). estimate of the potentially eligible population of 3.4 million,
and around half the self-employed population).
Employers
and sectors Employees of small- and micro-sized employers more Construction workers made 0.9 million claims for SEISS,
than twice as likely to be furloughed as those working one-third of all claims.
for large employers.
Employer take-up rates highest in Accommodation and
Food Services; Arts Entertainment and Recreation;
Construction and Manufacturing sectors.
Employees
and individuals Looking across age categories, a higher percentage of Similar claim rates across age groups (though lower
young workers under 25, were furloughed. among 16–24 year olds and people over 65).
Overall, a greater proportion of employed men were Higher percentage of potentially eligible men claimed a
furloughed than employed women. SEISS grant, compared with the proportion of eligible
women who claimed.
Around 20% of furloughed employments at the end of
July 2020 were partially furloughed under CJRS 2.
Geography
Highest furlough rates in parliamentary constituencies in Highest take-up rates in parliamentary constituencies
North West Wales, Cornwall, the Lake District, North and in Northern Ireland, the Welsh Valleys and North West
West London and West Sussex. Wales, the Glasgow City Region in Scotland and in parts
of Northern England and North and East London
Value of claims
£39.3 billion claims paid to 20 September 2020. £13.4 billion of claims paid to 20 September 2020.
Estimated value of claims to scheme end is £54.5 billion. Estimated value of claims to end of SEISS 2 is £15.2 billion.
Estimated average subsidy per job per month around Estimated average value of claim per month around £970
£1,200. (SEISS 1).
Notes
1 Taken from HMRC management information and published statistics on CJRS and SEISS. Estimates of value of claims to scheme end and
average subsidy and grant values are taken from the Office for Budget Responsibility’s (OBR’s) Fiscal Sustainability Report (July 2020).
Claims paid data is to 20 September 2020.
2 Average subsidy and grant values are based on CJRS 1 rules and SEISS 1 rules. Changes to CJRS from 1 July 2020 and amended terms for
the SEISS 2 will likely reduce these values.
3 Take-up data for employers and sectors, employees and individuals, and geography is based on CJRS 1 and SEISS 1.
Figure 11 presents sector take-up by value of claims.
Source: National Audit Office analysis of HM Revenue & Customs’ and Office for Budget Responsibility information
Implementing employment support schemes in response to the COVID-19 pandemic Part Two 37
Figure 11
Sectors accessing the most financial support from the CJRS, SEISS
and Bounce Back Loans
There is overlap between the employment support schemes and the Bounce Back loans
Rank Coronavirus Job Retention Self-Employment Income Bounce Back Loan
Scheme (CJRS) take-up Support Scheme (SEISS) market share
by value take-up by value
1 Wholesale and retail, Construction Real estate, professional
repair of motor vehicles services and
support activities
2 Accommodation and Unknown and other Wholesale and retail trade
food services
3 Manufacturing Transportation and storage Construction
Notes
1 Bounce Back Loans support small businesses with loans of up to £50,000 or a maximum of 25% of turnover.
The scheme was developed by HM Treasury, the Department for Business, Energy and Industrial Strategy,
and the British Business Bank. It launched on 4 May 2020 and was expected to last for six months.
2 Office for National Statistics Standard Industrial Classification (SIC) 2007.
Source: National Audit office analysis of HM Revenue & Customs’ data on CJRS and SEISS and National Audit Office
analysis of British Business Bank and Bank of England data from our Investigation into the Bounce Back Loan
Scheme report.
2.24 Figure 12 and 13 on pages 38 and 39 show the rates at which grants have
been claimed, by parliamentary constituency. CJRS take-up has been high in
areas popular with tourists and communities close to major south-east airports.
SEISS take-up has been highest in Northern Ireland, parts of Wales, areas
around Glasgow, parts of northern England and the West Midlands, and North
and East London. We found that areas that had periods of local restrictions since
the lifting of the national lockdown had only a marginally greater proportion of
self‑employed people making claims under the first phase of the scheme.
38 Part Two Implementing employment support schemes in response to the COVID-19 pandemic
Figure 12
Furloughing rates under the Coronavirus Job Retention Scheme (CJRS), CJRS 1
Furlough rates vary by parliamentary constituency with some of the highest in areas popular with tourists and those close
to major south-east airports
Employment furlough take-up by
Parliamentary Constituency (650)
41%–45% (7)
36%–40% (52)
31%–35% (349)
26%–30% (231)
21%–25% (11)
London
Notes
1 Take-up rates were based on an employee’s last known home address.
2 HM Revenue & Customs was unable to match the data (and so assign a location) for 104,100 furloughed employments and 201,200
eligible employments.
3 Map based on take-up data rounded to the nearest whole percent, using December 2019 parliamentary constituencies.
Source: National Audit Office analysis of HM Revenue & Customs’ data
Implementing employment support schemes in response to the COVID-19 pandemic Part Two 39
Figure 13
Self-Employment Income Support Scheme (SEISS) claims as a proportion of potentially eligible
population, SEISS 1
Take-up rates of the first SEISS grant vary by parliamentary constituency with some of the highest rates in Northern Ireland,
parts of Wales, areas around Glasgow in Scotland, parts of northern England and the West Midlands, and North and East London
SEISS claims as a proportion of the potentially eligible
population by Parliamentary Constituency (650)
81%–85% (61)
76%–80% (374)
71%–75% (200)
66%–70% (11)
61%–65% (4)
London
Notes
1 Take-up rates show the proportion of the potentially eligible population (when the scheme opened) that have applied and been paid
or are awaiting payment.
2 Take-up rates may have been higher than shown here since the potentially eligible population includes error and rejected cases,
and some individuals whose businesses have not been adversely affected by Coronavirus.
3 We have excluded 7,000 claimants and 12,000 potentially eligible individuals where addresses are listed in the Channel Islands,
Isle of Man, foreign addresses or missing and eligible Loan Charge cases who are yet to file their 2018-19 Self Assessment tax return.
4 Map based on take-up data rounded to the nearest whole percent, using December 2019 parliamentary constituencies.
Source: National Audit Office analysis of HM Revenue & Customs’ data
40 Part Two Implementing employment support schemes in response to the COVID-19 pandemic
Many people remain furloughed as the Departments’ support moves
to a new phase
2.25 Both schemes are due to end in their current form in October 2020.
CJRS will be replaced by the Job Support Scheme (JSS). For the self-employed
there is a modified SEISS scheme (see paragraphs 1.23 to 1.25 and Appendix
Four). It is not yet certain how employers will respond to the JSS. The JSS
scheme is intended to provide help for short-time working with government
providing less support and employers expected to pay more. The expanded
JSS is available for firms legally required to close. This will cover two-thirds of
the wages of furloughed employees, without requiring an employer contribution.
2.26 With around two million people still furloughed at September 2020, there
is a significant possibility of large-scale redundancies. Before the government’s
recent announcement of its Winter Economy Plan and the JSS,13 HM Treasury
published a summary of independent UK economy forecasts in September14
which showed unemployment ranging from 6.2% to 12.7% (average 8.3%,
around three million) by the final quarter of 2020. Analysis by the Institute for
Employment Studies – published after the government’s Winter Economy Plan
and expanded JSS – estimated around 600,000 redundancies in the final two
quarters of 2020, based on employer redundancy notifications to employees.
The latest unemployment data, published on 13 October, reported unemployment
had reached 4.5% of the labour force in the three months to August 2020, an
increase from 4.1% in the previous three months.
2.27 The Departments will need to consider how the current schemes have
operated to date, who has benefited and what this implies for the modified
schemes going forward. For example, it is not yet clear how the government’s
revised schemes will address the issue that some people will continue to be
excluded from support based on the tax information held by HMRC. The expanded
JSS scheme adds further administrative complexity with eligibility criteria based on
location and lockdown restrictions in place at particular points in time.
13 HM Treasury, Policy Paper – Winter Economy Plan 2020, CP 297, September 2020, available at: www.gov.uk/
government/publications/winter-economy-plan/winter-economy-plan
14 HM Treasury, Forecasts for the UK Economy: September, September 2020, available at: www.gov.uk/
government/statistics/forecasts-for-the-uk-economy-september-2020
Implementing employment support schemes in response to the COVID-19 pandemic Part Two 41
2.28 It will be difficult to disentangle the specific impact of the existing schemes
from other forms of government support (including the JSS and Bounce Back
loans), as well as the ongoing impact of COVID-19. Recent analysis gives more
insights into the range of effects of the schemes, finding that the average SEISS
grant came close to fully compensating recipients for the effects of the crisis.
However, furloughed employees under CJRS saw an average 13% decline in net
income and new claimants of Universal Credit a 40% decline.15 Stakeholders,
including the Treasury Select Committee, have suggested careful consideration
should be given to extending employment support on a targeted basis,
recognising that some sectors and regions may have been more acutely affected
by COVID-19 than others. The Confederation of British Industry and the Trades
Union Congress have both supported the measures announced in the Winter
Economy Plan and believe the JSS will help to support some people to stay in
work. The Resolution Foundation and Institute for Employment Studies have,
however, questioned the extent to which the original JSS will prevent significant
job losses over the coming months.
15 I Delestre et al, Income protection policy during COVID-19: evidence from bank account data, Briefing Note,
10 September 2020, available at: www.ifs.org.uk/publications/15002
42 Part Three Implementing employment support schemes in response to the COVID-19 pandemic
Part Three
Fraud and error affecting the employment
support schemes
3.1 The Coronavirus Job Retention Scheme (CJRS) and the Self-Employment
Income Support Scheme (SEISS) are both grant schemes administered by
HM Revenue & Customs (HMRC). Inevitably grant schemes are at risk of
fraud and error:
• Fraud occurs where the claimant (the employer for CJRS, the
self‑employed individual for SEISS) deliberately sets out to misrepresent
their circumstances to get money to which they are not entitled.
• Error occurs where the claimant inadvertently receives the wrong amount
due to incorrect information being provided, but with no deliberate intent
to mislead, or incorrectly calculate the award.
3.2 This part of the report considers HMRC’s:
• overall approach to compliance, including the control environment and
its tolerance of risk;
• management of the fraud and error risks;
• understanding of the scale of fraud and error; and
• plans for recovering grant overpayments.
Implementing employment support schemes in response to the COVID-19 pandemic Part Three 43
HMRC’s compliance approach
The control environment
3.3 HMRC has a long-standing approach to tackling non-compliance.
In administering the tax system and paying out tax credits, HMRC considers
“the best way to tackle non-compliance is to prevent it happening in the first
place, while cracking down on the minority who do break the rules”.
3.4 In designing the schemes, HMRC recognised that it would need to make
certain trade-offs in its normal approach of preventing as much fraud and error as
possible and the need to ensure grants reached claimants quickly. HMRC drew up
longlists of potential controls for both schemes. In total it identified 42 potential
controls for CJRS, of which 24 were implemented by the go-live date and 57 for
SEISS, of which 38 were delivered by the end of April. HMRC identified certain
controls as critical to delivery and implemented all of these. It did not implement
all possible controls because it viewed them either as unfeasible to deliver within
the time available or not sufficiently effective. This was particularly the case for
CJRS, which had to be implemented more quickly than SEISS (Figure 14 overleaf).
HMRC’s COVID-19 compliance approach prioritised developing pre‑payment
checks to prevent risk of organised criminal attacks on the schemes over other
types of risk such as opportunistic fraud, where there is a greater likelihood it can
recover overpayments retrospectively.
3.5 HMRC has maintained strong oversight of its control framework for both
schemes, which sets out in detail what all the controls do, the stage of the claims
process they operate in and the risks they were seeking to mitigate against
(Figure 15 on page 45).
44 Part Three Implementing employment support schemes in response to the COVID-19 pandemic
Figure 14
Examples of key controls for Coronavirus Job Retention Scheme (CJRS) and Self-Employment
Income Support Scheme (SEISS)
HM Revenue & Customs (HMRC) had less time to develop controls for CJRS than for SEISS
CJRS SEISS Why was control not implemented?
Promote Honesty declaration
Online fraud reporting form
Promote threat of future compliance action
Public listing of claimants N/A HMRC not convinced measure would
be effective.
Online guidance and calculator
HMRC invitation to apply Up to employer to assess need for the scheme.
Prevent Checking work status with furloughed employees N/A Insufficient time/resources.
Automated calculation of award N/A
Full verification of all input data and amounts claimed N/A Insufficient time/resources.
per employee
Payments capped at maximum grant award value
Employee/self-employed person must already be in
HMRC’s tax systems to be eligible
Agents prevented from claiming HMRC viewed CJRS risk as acceptable.
Transaction monitoring
Bank account checks
Respond Data matching to detect potential non-compliance
Voluntary disclosure window
One-to-one enquiries
means that the control was implemented. indicates that the control was not implemented.
Notes
1 The controls listed above are not exhaustive lists.
2 An online fraud reporting form was in place for both schemes. The telephone hotline was not available until August 2020 due to
home-working restrictions during the pandemic.
3 Online guidance was available for the SEISS as well as information about how the calculation was made. The calculation itself, however,
was automated by HMRC.
4 HMRC sent out communications to registered Self Assessment taxpayers notifying them that they may be eligible to apply for the SEISS.
For CJRS, no such communication was made but HMRC did contact some of the largest companies to check whether they intended to apply.
5 Furloughed employees were not able to do any work under the original rules of the CJRS. From 1 July 2020, however, they could work part-time.
6 An employee had to be registered on the employer’s Pay-As-You-Earn scheme by 19 March and included on a Real Time Information submission
made by the same date to be eligible for CJRS. Self-employed individuals had to be already registered on HMRC’s Self Assessment system to be
eligible for the SEISS.
Source: National Audit Office analysis of HM Revenue & Customs’ control frameworks
Implementing employment support schemes in response to the COVID-19 pandemic Part Three 45
Figure 15
HM Revenue & Customs’ (HMRC’s) control framework for the employment support schemes
HMRC’s control framework enabled it to effectively document risks and the controls in place to tackle them
HMRC used its control framework for Coronavirus Job Retention Scheme and Self-Employment Income Support Scheme.
Through using it, HMRC aimed to properly understand and own the scheme risks across the organisation. The framework
is influenced by the principles of the government’s Orange Book guidance, designed to support departments to better
manage risk.
A control listing sheet was used for each scheme to document controls around different phases of customer claims.
Each control had a named owner responsible for monitoring it.
Pre-claim: In-claim: Between claims: Post-claim:
For example, controls For example, controls For example, some For example,
around eligibility issues. around monetary limits. applicants needed controls to recover
to be excluded. inappropriate payments.
Control sheets are used to show progress and to ensure that the control environment is operating as expected. The sheets
were updated regularly as new controls were brought in and existing ones updated, thus ensuring HMRC had a comprehensive
overview of the overall controls regime.
Control owners reported on progress weekly. A Control Board produced a weekly update on controls that was submitted
to HMRC’s Executive and Audit Committees.
Note
1 HMRC documented controls for manual claims using the same approach.
Source: National Audit Office analysis of HM Revenue & Customs’ control framework
46 Part Three Implementing employment support schemes in response to the COVID-19 pandemic
HMRC’s assessment of risk
3.6 HMRC conducted a detailed initial risk assessment on both schemes to
establish the key fraud and error risks. For example, HMRC was almost certain
that organised criminals would target the system for weaknesses and viewed
the likelihood of opportunistic fraud, where claimants exploit the rules of the
scheme to their own advantage, as being very likely (Figure 16). As such,
HMRC recommended limiting the schemes only to taxpayers with current
records, a move that reduced the risk of fraud considerably because applicants
had to be known to HMRC at the time the schemes were announced.
3.7 HMRC has assessed that the risk of fraud and error is greater for the CJRS
scheme than for SEISS as it was able to put in place more preventative controls
for SEISS (Figure 14). For SEISS, Self Assessment taxpayers were invited to
apply, rather than for CJRS where taxpayers could make applications themselves.
It also calculated the SEISS grant using existing tax return data. This reduced the
scope for individuals to either erroneously or deliberately misstate their incomes
to increase their claim award.
3.8 Under the CJRS scheme, employers were responsible for informing
employees about whether they were being furloughed under the government
scheme and for making payments to them. Employers retained the responsibility
for calculating claims. The scheme allowed a degree of tolerance around
the amounts employers applied for, up to a maximum of £2,500 a month per
employee. HMRC did not ask small employers to state how much was being
claimed for each employee because it concluded it was not feasible to build this
into the new system in the time available. HMRC considered validating claim
amounts for employees but rejected this because the range of variations seen
in people’s pay (for example, from promotions or performance-related pay) would
have introduced too much complexity. HMRC’s internal audit team warned before
the scheme went live that checks on amounts claimed per employee were critical
to effective and efficient compliance.
Implementing employment support schemes in response to the COVID-19 pandemic Part Three 47
Figure 16
HM Revenue & Customs’ (HMRC’s) early assessment of the key fraud and error risks
HMRC identified the same key risks for both schemes but assessed the impacts to be potentially greater for the
Coronavirus Job Retention Scheme (CJRS)
Assessment of Risk
Key risks Examples Likelihood Impact after mitigation (% of
total grant expenditure)
CJRS
Organised crime Pay-As-You-Earn schemes hijacked by organised Almost certain 2.5% to 5%
criminal gangs.
Opportunistic fraud Employer gets furloughed employee to Highly likely 2% to 3%
continue working.
Employer doesn’t pass on furlough payments
to employee.
Employer inflates claim for employee.
Internal fraud HMRC staff member colludes with employer. Low Low
Error Employer claims incorrect amount per employee or Low 1% to 2%
HMRC processes award incorrectly.
Total 5% to 10%
Self-Employment
Income Support
Scheme
Organised crime Hijacking of claimant details by organised Almost certain 0.3% to 1.5%
criminal gangs.
Opportunistic fraud Claimant makes a late filing that deliberately inflates Highly likely 0.3% to 0.7%
their profits.
Internal fraud HMRC staff member falsifies details to claim Low Low
grant money.
Error Wrong amount is awarded based on incorrect data. Low 0.2% to 0.5%
Total 1% to 2%
Notes
1 The likelihood assessments use a probability yardstick. An assessment of ‘almost certain’ indicates a probability in excess of 95%.
A ‘highly likely’ assessment has a probability of 80% to 90%.
2 The estimates above were provisional as they were based on an early assessment of the risks and carried significant levels of uncertainty.
HMRC has committed to further updates on the likely levels of fraud and error by the end of 2020.
3 The totals do not exactly match the sum of the individual components. This is because the total are calculated using Monte Carlo modelling.
This is a good-practice approach that varies key inputs and risk factors to produce a range of estimated scenarios. This means, however,
that the most likely values of total fraud and error are expected to sit within a much narrower range than is implied by simply aggregating
the most extreme values from each scenario.
Source: National Audit Office analysis of HM Revenue & Customs’ risk assessments of the schemes
48 Part Three Implementing employment support schemes in response to the COVID-19 pandemic
Tolerance of risk
3.9 In order to get money to claimants quickly HMRC accepted some factors
would contribute to significant fraud and error risks. These factors included:
• reducing the scope for preventative controls. HMRC gave itself a maximum
of three days to validate CJRS and SEISS claims so it could make payments
as quickly as possible. This is much faster than for Tax Credits and Child
Benefit, where the target for both is 22 days;
• HMRC not initially having the powers to undertake compliance work for the
grant schemes, as with tax investigations, and it could only investigate cases
where it suspected criminal intent. This delayed its ability to commence
compliance work until the 2020 Finance Act passed into law in July;
• processing large volumes of claims quickly, HMRC needed to prioritise
preventative checks for the most high-risk cases. This meant concentrating
on suspected cases of organised crime, where the potential to recover money
is normally much more difficult than for opportunistic fraud committed by
known taxpayers. HMRC’s decision to send out invitations for SEISS over a
four-day period further narrowed its window to perform preventative checks.
The invitations to apply over a four-day period resulted in four times as many
SEISS claims in the first week compared with CJRS; and
• HMRC concluding that it had limited control of the relationship between
employers and employees. It recognised there were risks that employers claim
furlough money while continuing to get their employees to work but decided
that it would need to tackle this risk in post-payment checks. We consider
HMRC’s management of furlough fraud in paragraphs 3.11 to 3.19.
How HMRC has managed risks
3.10 We looked in more detail at three key risks which HMRC viewed as
almost certain or highly likely. We considered: how HMRC assessed the risk;
the mitigation controls it put in place; and the extent to which these controls
operated effectively:
1 Employers claiming money while their furloughed employees
continue to work.
2 Self-employed individuals inflating their claims in late Self
Assessment returns.
3 Organised criminals hijacking agents’ details to submit fraudulent claims.
Implementing employment support schemes in response to the COVID-19 pandemic Part Three 49
1. Working while furloughed
3.11 Under CJRS an employer commits ‘furlough fraud’ if they claim furlough
payments for an employee but continued to require them to work (in any capacity
under the first phase of the scheme (CJRS 1), or on a full-time basis under the
second phase of the scheme (CJRS 2)). Other forms of furlough fraud include
employers falsely inflating the amount of grants they claim or failing to pass on
the full amount of furlough payments due to their employees.16
3.12 HMRC recognised that there was a high risk that employers would exploit
the initial CJRS rules and continue to get their employees to work for them
while claiming support. It informed the Chancellor it could not establish strong
preventative controls. The only specific controls in place for the risk were an
honesty declaration in the online form, whistleblowing routes, other forms
of intelligence and retrospective investigation. HMRC set up an online fraud
reporting form in addition to its telephone hotline. The telephone hotline was not
available for the duration of CJRS 1 because home-working restrictions meant
staff could not take calls in a secure environment.
3.13 HMRC ruled out further checks which could have helped to mitigate the
specific risk of furlough fraud. It considered that pre-payment checks were
impractical in the timeframe. The potential to contact furloughed employees
retrospectively to check whether they were working was considered but ruled
out on the basis it would be too resource-intensive and require resources to be
diverted away from helping employers make claims.
3.14 Employees did not know if their employer was part of the government furlough
scheme unless their employer informed them. HMRC ruled out publishing the
details of every company claiming furlough payments, which might have alerted
some employees that their employer was acting fraudulently. HMRC concluded
it would have been unrealistic to contact employees given the large numbers of
recipients of the furlough scheme. It was also concerned that naming companies
could deter too many legitimate claimants and be detrimental to the primary
policy objective to get support to those who needed it. HMRC believed it had
comprehensive and robust controls to mitigate the risk of fraud.
3.15 HMRC has announced additional controls for the Job Support Scheme (JSS)
to deter fraud and to prompt employees to disclose fraud. It intends to publish
the names of employers who have used the scheme and it will use personal tax
accounts to inform employees that their employer has claimed JSS.
16 Under SEISS, self-employed individuals were free to continue trading while claiming the grant providing that they
stated that their business had been ‘adversely affected’ by COVID-19.
50 Part Three Implementing employment support schemes in response to the COVID-19 pandemic
3.16 Up to 14 August, employees reported 8,600 potential fraud abuses through
the online form, with 8,100 of these relating to CJRS. The telephone hotline
received a further 1,600 reports, bringing total reported cases to more than
10,000.17 The Department’s experience, however, is that whistleblowing is only
likely to identify 1% to 3% of fraud cases.
3.17 HMRC has investigated around one-quarter of CJRS fraud allegations
reported online. Of those investigated, the majority of allegations relate to
employers asking employees to work while furloughed (Figure 17). In around 8%
of cases, the employer has admitted some form of error. Only one employer has
accepted that they deliberately falsified their claim. In the remaining 92% of
cases, the employer has either refused to engage with HMRC to date or denied
any wrongdoing. HMRC continues to pursue these cases. HMRC has considered
naming employers who are subsequently found to have committed fraud. Under
its compliance powers received in July 2020 it can now do so but is yet to
exercise these powers.
Figure 17
Allegations of Coronavirus Job Retention Scheme fraud received
through online form
The majority (62%) of allegations of fraud concern employees being asked to work by their employer
while furloughed
Percentage (%)
70
60 62
50
40
30
20
15 17
10
6
0
Furloughed but Furloughed but Employer claim Other
not paid asked to work fraudulent/inflated
Note
1 Data are based off 2,205 calls made up to 7 August 2020. In some instances, multiple allegations are made in
one call, so the total number of allegations (2,453) exceeds the number of calls.
Source: HM Revenue & Customs’ analysis of post-payment calls
17 A very small number of these relate to potential fraud in the Statutory Sick Pay Rebate (SSPR) that HMRC
administers. SSPR was the reason for 15 online notifications.
Implementing employment support schemes in response to the COVID-19 pandemic Part Three 51
3.18 Our survey looked at the possible extent of employees working while on
furlough (for full details of the survey see Appendix Two). This found that 9% of
survey respondents who reported being furloughed had done some work at the
request of their employer.18 An additional 6% reported that they had chosen to
do some work despite not being asked to by their employer.19 In addition, we also
found that a further 4% of respondents to our survey who had not worked while
furloughed said that their employer had paid them less than 80% of their pay, the
minimum amount under CJRS.20
3.19 Surveys by other organisations have different results, with the
overall percentage of furloughed employees working at the request of their
employer ranging from 7% to 34% (see Figure 18 overleaf). Different sampling
methodologies, question framings and time periods surveyed are likely to
account for some of this variation. The survey commissioned by the National
Audit Office covers CJRS 1 while the other surveys covered narrower periods
in time. HMRC has not carried out its own survey of furlough fraud but it was
considering this option in September 2020, when we were finalising our fieldwork.
18 Based on 1,178 survey responses. We asked our survey questions as part of Ipsos MORI’s telephone omnibus
and online surveys. The quota sampling methodology used and the limits of the approach in terms of how
generalisable the results are to the population of furloughed workers are set out in Appendix Two.
19 Based on 1,178 survey responses.
20 Based on a sample of 503 computer-assisted telephone interviews. Due to an error in the online survey routing,
the base size for this particular question is smaller than for our other survey questions.
52 Part Three Implementing employment support schemes in response to the COVID-19 pandemic
Figure 18
Surveys looking at percentage of furloughed employees working
Furlough fraud could be in the range of 7% to 34% according to surveys
Organisation Percentage of furloughed Percentage of Period of lockdown
employees working at furloughed employees covered by survey
employer direction voluntarily working
(%) (%)
Academics at Oxford, 19 44 April 2020
Cambridge and
Zurich universities
Crossland 34 N/A March to
Employment Solicitors 11 June 2020
National Audit Office 9 6 March to June 2020
(NAO)
Resolution Foundation 7 N/A May 2020
Notes
1 The NAO survey was conducted by Ipsos MORI on our behalf. In our survey, 12% of employees who had been
furloughed responded that their employer had asked them to work, 9% had done so and 3% had not done so.
2 The survey we commissioned from Ipsos MORI combined online and telephone surveying (weighted 50:50).
While showing survey results for each method should be treated with caution because it results in smaller
response samples, we found different results across the two methods: 4% of furloughed respondents who
were contacted by telephone said that their employer had asked them to work and they had done so; 14%
of furloughed respondents who were contacted online said that their employer had asked them to work and
they had done so. These differences reflect variations between the populations from which the samples
were drawn; as well as nuances in job sector, type of employment and the mode of interview (self-completion
versus interview-led telephone interviewing).
3 Crossland Employment Solicitors, Resolution Foundation and the Academics at Oxford, Cambridge and Zurich
universities all used online panels for their surveys.
4 The Resolution Foundation’s survey, conducted by YouGov, found that 7% of employees had been asked
to work by their employer. It did not ask whether these employees complied with the request.
5 The surveys by Crossland Employment Solicitors and Resolution Foundation did not ask employees whether
they had voluntarily worked while placed on furlough.
6 Information about the polling methodologies used by other organisations can be seen in the source links below.
For further information on the survey methodology adopted by Ipsos MORI on our behalf see Appendix Two.
We did not directly validate the results of any of the other surveys listed here.
Sources: Crossland Employment Solicitors survey results are available at: www.crosslandsolicitors.com/site/media/
coronavirus-hub/furlough-fraud-coronavirus-lockdown; The Resolution Foundation survey results are available
at: www.resolutionfoundation.org/publications/the-effects-of-the-coronavirus-crisis-on-workers/; The Oxford,
Cambridge and Zurich University paper is available at: www.econ.cam.ac.uk/research-files/repec/cam/pdf/
cwpe2079.pdf
Implementing employment support schemes in response to the COVID-19 pandemic Part Three 53
2. Late Self Assessment returns
3.20 SEISS was announced less than two months after the Self Assessment
deadline meaning that many taxpayers had not yet filed their returns or still
intended to submit amendments. HMRC identified significant risks for taxpayers
to inflate the value of their returns in order to obtain larger grants under the
SEISS scheme, although this would also have increased their future tax liability.
To mitigate this risk it:
• did not allow amendments to already-filed returns to be included in
calculating the grant because it assessed the fraud risk was too high.
Any amendments to filed Self Assessment tax returns after SEISS was
announced on 26 March 2020 were disregarded in calculating an individual’s
award. An individual may have had a legitimate reason for making such an
amendment but HMRC was concerned about the potential for manipulation.
Subsequent analysis of amended returns suggests HMRC’s concerns were
well founded; taxpayers filed 7,000 amendments the day after the policy was
announced with 1,200 of these seeking to increase their trading income by
an average of £7,000; and
• permitted late returns but sought to perform additional ID verification
checks on these before receiving the SEISS claims. Between 26 March and
23 April 2020, 64,000 late returns were filed by self-employed individuals,
with 47,000 of these deemed potentially eligible for the SEISS grant.
HMRC performed ID verification checks on around 15% (7,000) of these
high-risk late returns. Of those 7,000 returns, just over 40% failed these
verification checks. Of those that passed the verification checks, 11% were
flagged for post-payment compliance checks. While HMRC detected high
levels of risk in the sample it checked, it did not carry out additional checks
on the other 85% of late returns, instead relying on its standard risk rules
applied to all SEISS claims. HMRC told us those standard risk rules were
updated to recognise the fact that SEISS claims based on late returns
should be viewed as high risk. However, this only allowed HMRC 72 hours
to decide whether to accept or reject the claim.
54 Part Three Implementing employment support schemes in response to the COVID-19 pandemic
3. Hijacked claims by organised crime
3.21 Organised criminals could exploit the schemes by stealing the identities of
legitimate taxpayers or coercing them to make fraudulent claims. HMRC identified
a risk that if agent details were stolen, fraudsters would potentially be able to
make fraudulent claims on behalf of large numbers of employers or individuals.
Agents or third parties submitted more than half (56%) of Self Assessment
returns in 2018-19. HMRC research suggests around 50% of CJRS claims
were made through agents.
3.22 Given the timescales for implementing SEISS, HMRC concluded it
could not develop a claims service for agents that would also mitigate the
risk of exploitation by organised crime. It therefore focused on developing
a system that would be easy for customers to use directly and said it would
not accept applications for SEISS from agents. It did not operate the same
restriction for CJRS.
• For SEISS, HMRC stipulated that self-employed people must make an
application themselves. It monitored activity and prevented some large-scale
attempts to attack the system. Over a single weekend in May 2020, HMRC
blocked around 87,000 claims worth £242 million due to suspicious activity.
It did not have the capacity to assess all these claims within the three-day
period and acted prudently to halt the applications. It recognised it could
have denied some legitimate claims and subsequently contacted nearly
all of these claimants to encourage them to claim themselves rather than
through agents.
• HMRC did not take the same approach on the CJRS, allowing agents to
continue to make applications on behalf of employers, even where the
employer had not explicitly consented to this. It did alert businesses through
their online business tax accounts that claims had been made in their name,
but not until August 2020. This meant businesses may not have been
aware that their details had been hijacked and used to make a claim during
the first months of the scheme. HMRC also worked proactively to shut
down a number of online scams that had been circulating that seek to trick
individuals into passing on personal details.
• HMRC has detected relatively few CJRS claims where credentials may have
been hijacked. It has rejected 224 CJRS claims to date worth £10 million.
However, hijacked cases may have gone undetected. Until HMRC has
carried out enquires into cases it cannot be sure it has not failed to prevent
more attacks. HMRC continues to monitor organised criminal attacks but
has yet to assess its effectiveness in mitigating the risk.
Implementing employment support schemes in response to the COVID-19 pandemic Part Three 55
Understanding the scale of fraud and error
3.23 HMRC is unlikely to know the total level of fraud and error in each of the
schemes for some time. Its planning assumptions were based on comparisons to
Tax Credits, Universal Credit and Income Tax Self Assessment repayment fraud.
Based on these it initially estimated potential fraud and error levels of 5% to
9% in CJRS and 1% in SEISS. At the time the assessments were made, HMRC’s
internal audit function concluded that the comparators were not sufficiently
reliable to make conclusive assessments about the scale of the risk.
3.24 HMRC’s most recent analysis predicts similar rates of 5% to 10% for
CJRS and 1% to 2% for SEISS (see Figure 16). This would equate to around
£2.0 billion to £3.9 billion on the CJRS and £130 million to £270 million on
the SEISS based on spend to date. HMRC told us that while these estimates
remained assumption-based, they were informed by expert opinion and it
was continuing to update the estimates with information collected during the
schemes. HMRC believes that the CJRS was less susceptible to organised crime
than it initially forecast but also acknowledges that the greater flexibility in the
rules that allowed part-time working from 1 July increased the risk of genuine
error and more opportunistic fraud.
3.25 Survey data (see Figure 18) indicates that the true level of fraud and error
in the CJRS could be considerable. Surveys suggest that furlough fraud due
to employees working at the request of their employer could be in the range
of 7% to 34%. We also found 4% of respondents to our survey who had not
worked while furloughed said that their employer had paid them less than 80%
of their pay, the minimum amount under CJRS (see paragraph 3.18).21 Making a
precise estimate is complicated because part-time working was allowed in CJRS
2, and the work employees carried out in CJRS 1 may have been on a part-time
basis.22 HMRC will need to consider whether the spirit of the rules have been
broken and weigh this against the need to treat those who followed the rules
fairly. Of those respondents to our survey who had said that they had worked
while furloughed at the employer’s request, almost two-thirds said they did so
for less than half of their usual hours.23
21 Based on a sample of 503 computer-assisted telephone interviews. Due to an error in the online survey routing,
the base size for this particular question is smaller than for our other survey questions.
22 There were some very limited circumstances in CJRS 1 where working while furloughed was allowed, such as a
director carrying out director’s duties.
23 Based on 1,178 survey responses, of which 117 indicated that they had worked while furloughed at the
employer’s request. These are raw respondent counts rather than the weighted ones used for calculation of
proportions elsewhere in this report. See Appendix 2 for further detail.
56 Part Three Implementing employment support schemes in response to the COVID-19 pandemic
3.26 HMRC intends to produce a final statistical estimate on both schemes for
total levels of error and fraud. It aims to refine its provisional estimates again
by the end of 2020 and in spring 2021 as it undertakes more compliance work
and receives more operational intelligence. HMRC is monitoring the level of
organised criminal attack. It does not expect to produce final estimates until the
end of 2021 at the earliest. It intends to use similar methodologies to those used
in estimating the tax gap. We recently reported on HMRC’s assessment of the
tax gap, noting that it is a comprehensive assessment but with some inevitable
uncertainty attached.24 There will also be parallels with the work that HMRC does
in measuring error and fraud in Tax Credits and Child Benefit, as well as the work
the Department for Work & Pensions does in measuring fraud and error in the
benefits system.
3.27 HMRC has made some errors resulting in overpayments of grants worth
£14 million. In the first tranche of SEISS payments, 26,000 were miscalculated
due to not all tax return data being considered. Around 13,000 of these led to
overpayments worth £12 million, equivalent to less than 1% of all SEISS grants.
On CJRS, around £3 million was overpaid on 330 cases that were manually
processed by HMRC staff. There are also likely to be cases where employers
overclaimed CJRS because the HMRC online calculation tool was not accurate.
HMRC made changes to the tool in May to rectify identified problems. HMRC does
not intend to recover these overpayments despite having the legal powers to do so.
3.28 Up to 13 October, companies had paid back £278 million in grant money
claimed under the CJRS. The amounts were claimed in error or were amounts
that the company subsequently decided they did not need as they adapted to
the pandemic.
24 Comptroller and Auditor General, Tackling the tax gap, Session 2019–2021, HC 372, National Audit Office,
July 2020.
Implementing employment support schemes in response to the COVID-19 pandemic Part Three 57
Post-payment compliance
3.29 HMRC has set out how it intends to deal with potential non-compliance
post-payment. For CJRS, it has set up an approach which gives it multiple
opportunities to recover overpayments (Figure 19). HMRC has already made
three arrests in relation to suspected CJRS fraud. For SEISS, work commenced in
October 2020 targeting the highest risk areas using a similar campaign approach
to CJRS. Further SEISS risks will be considered as part of the compliance checks
on the 2020-21 Self Assessment tax returns that will be due by January 2022.
Figure 19
HM Revenue & Customs’ (HMRC’s) post-payment compliance regime for Coronavirus Job
Retention Scheme (CJRS)
HMRC’s post-payment compliance regime has yet to produce tangible results in terms of recovery of overpayments
CJRS compliance What is it? Results to date
campaign
Claims assurance checks 2,800 calls to employers during the scheme 18% of calls have identified errors resulting
to understand more about their claims. in overpayments.
20% of calls have identified errors but with
no net overpayment.
63% of calls either not answered, customer
unable/unwilling to speak or no error identified.
90-day disclosure window Amnesty period for employers to encourage 27,000 letters sent out.
to 20 October 2020 them to disclose any overclaims during the
136 notifications of overpayments received
disclosure window.
but too early to evaluate response rates.
One-to-one enquiries For remaining high-risk claimants who have 10,000 enquiries anticipated, subject to
not corrected their claims. the number of voluntary disclosures.
Protection against To protect taxpayer money where companies
deliberate insolvency look to deliberately enter insolvency.
Notes
1 Percentages for claims assurance checks calls do not sum due to rounding.
2 Total number of claims assurance checks based on data as at 7 August 2020. HMRC made the decision to discontinue these calls from
August onwards instead shifting its focus into other post-payment compliance activity.
Source: National Audit Office analysis of HM Revenue & Customs’ documentation
58 Part Three Implementing employment support schemes in response to the COVID-19 pandemic
3.30 HMRC plans to redeploy 500 full-time equivalent (FTE) staff to enable
post-payment compliance work. Its assessment in June 2020 suggested that
this work would bring in around £275 million. The work offers a positive return
on investment (we estimate around 9:1) but HMRC estimates it would come
with an opportunity cost to tax revenue of £100 million to £200 million as it
would need to redeploy staff from tax compliance. HMRC does not believe it
can recruit additional staff because there is an 18-month time-lag between
starting recruitment and getting staff fully trained to start work. In June 2020 it
considered outsourcing but concluded at that stage that there were no existing
commercial mechanisms to recruit at the speed it needed. It was also concerned
about the challenges experienced through outsourcing HMRC functions in the
past (such as using Concentrix to administer Tax Credits’ claims) and that this
would not be worthwhile given furlough arrangements may not be extended,
making this a one-off. In October 2020 HMRC informed us that with the
announcement of new employment support schemes, and informed by learning
from the operation of SEISS and CJRS, it was now planning to use private
contractors to supplement its compliance capacity where necessary.
3.31 Until it has a robust estimate of the total levels of error and fraud, HMRC
cannot reasonably know how much resource it should commit to recovery.
Its provisional assessment of £275 million is based on investigating around
30% (10,000 claims) of the highest-risk CJRS cases. If total levels of error and
fraud prove to be higher than initially forecast the work may take some time to
complete. HMRC will also need to consider the long-term impact of the pandemic
and the risks this poses to tax revenue and the ability of taxpayers to repay
grants in the short-term.
Implementing employment support schemes in response to the COVID-19 pandemic Appendix One 59
Appendix One
Our audit approach
1 This report examines the role of HM Treasury and HM Revenue & Customs
(HMRC) in implementing the Coronavirus Job Retention Scheme (CJRS) and the
Self-Employment Income Support Scheme (SEISS). The report considers how
well the schemes were implemented, recognising the need to deliver at speed in
response to the COVID-19 pandemic. Our key audit questions were:
• Did HMRC and HM Treasury manage design and delivery risks effectively
in implementing the schemes?
• Do HMRC and HM Treasury understand whether the schemes reached the
people intended?
• Are HMRC and HM Treasury managing fraud and error risks effectively
thus far?
2 This report does not consider HMRC’s other COVID-19 interventions
designed to support businesses, including the Coronavirus Job Retention Bonus
and Eat Out to Help Out, or the relationship between the schemes and wider
government support such as business loans and benefits. We have previously
reported on the Bounce Back Loans Scheme.
3 Our audit approach is summarised in Figure 20 overleaf.
60 Appendix One Implementing employment support schemes in response to the COVID-19 pandemic
Figure 20
Our audit approach
The objective of
government To provide support for businesses affected by the COVID-19 pandemic to enable them to retain
employees and to provide support to self-employed individuals whose businesses were adversely affected
by the pandemic.
How this will
be achieved HM Revenue & Customs (HMRC) and HM Treasury together designed two grant schemes: the Coronavirus
Job Retention Scheme (CJRS) to support businesses; and, for self-employed individuals, the Self-Employment
Income Support Scheme (SEISS).
Our study
We examined how HMRC and HM Treasury designed and delivered these two support schemes, recognising
the need to deliver at speed in response to the COVID-19 pandemic.
Our evaluative
criteria Did HMRC and HM Treasury Do HMRC and HM Treasury Have HMRC and HM Treasury
manage design and understand whether the managed fraud and error risk
delivery risks effectively in schemes reached the effectively thus far?
implementing the schemes? people intended?
Our evidence
Interviews with key HMRC and Interviews with key HMRC and Review of key documents,
(see Appendix Two
HM Treasury staff. HM Treasury staff. including control frameworks.
for details)
Document review, including Document review, including Review and analysis of any
ministerial submissions on ministerial submissions. emerging evidence on fraud
scheme design. and error levels.
Interviews with
International comparisons relevant stakeholders. Interviews with key HMRC and
using data from other HM Treasury staff.
Analysis of our survey data.
audit institutions.
Analysis of our survey data.
Our conclusions
HM Treasury and HMRC met their objective to rapidly implement the schemes and the Departments should
be commended for making these available ahead of schedule. The schemes were relatively straightforward to
apply for, and payments quickly reached those who applied. Indications are that this has helped to protect jobs
in the short-term and the numbers of people moving from furlough arrangements back to work are encouraging.
However, many other people have lost earnings and have not been able to access support. The long-term
impact of the schemes will also depend on wider financial support and the ongoing impact of COVID-19.
A key value-for-money test for these schemes will be how far the Departments can mitigate fraud and error.
The pace at which the schemes were designed and implemented meant the Departments had to accept a
greater risk than normal. As such, there are likely to be considerable amounts of fraud and error, particularly
on the furlough scheme. Limiting applications to existing taxpayers should have reduced the fraud risk, but
HMRC could have done more to make clear to employees whether their employer was part of the furlough
scheme. In future, the Departments should do more while employment support schemes are running to
protect employees and counter acts of fraud. The Departments will need to ensure sufficient resources are
committed to recover money where it is cost-effective to do so.
Implementing employment support schemes in response to the COVID-19 pandemic Appendix Two 61
Appendix Two
Our evidence base
1 Our conclusions on the two employment support schemes were reached
based on our analysis of evidence collected between June and October 2020.
Our overall audit approach is outlined in Appendix One.
Using our back catalogue
2 In designing and carrying out our work we took account of previous National
Audit Office (NAO) analysis examining the Departments’ responses to financial
and humanitarian crises using the learning from this work to inform our evaluative
criteria. We also deployed our framework for assessing the Departments’
approaches to managing fraud and error, an audit method that builds on
significant past experience of government interventions.
Departmental meetings and information
3 We met with key officials from both HM Treasury and HM Revenue &
Customs (HMRC) involved in the design and delivery of the schemes. Specifically:
• the senior responsible owners for both the Coronavirus Job Retention
Scheme (CJRS) and the Self-Employment Income Support Scheme (SEISS),
and the key staff members supporting them;
• staff from HMRC’s Customer Compliance Directorate who led on the
compliance regime for both;
• HMRC’s internal audit team, who offered initial advice to their Departmental
colleagues on the appropriateness of the schemes’ control environments;
• staff from HMRC’s Knowledge, Analysis and Innovation Team to discuss the
monitoring and evaluation of the schemes; and
• staff from HMRC’s Behavioural Insights and Trials Team.
4 We reviewed relevant documents, including: options papers, ministerial
submissions and strategy documents, accounting officer advice, controls
frameworks, risk and decision logs, scheme overviews and customer journey maps,
lessons-learned documents, performance dashboards and internal audit papers.
62 Appendix Two Implementing employment support schemes in response to the COVID-19 pandemic
Stakeholders and experts
5 We reviewed key external reports and data sources including those
published by the Office for National Statistics (ONS), the Organisation
for Economic Co-operation and Development (OECD), and the Treasury
Select Committee.
6 We engaged with experts from the Institute for Government, National
Institute for Economic and Social Research, and the Resolution Foundation.
We also discussed with the authors the results of findings from a survey
conducted by Dr Abigail Adams-Prassl and colleagues from Oxford, Cambridge
and Zurich universities. We corresponded with Crossland Employment Solicitors
in relation to their survey work. For the surveys featured in Figure 18 in the
report, we asked: who conducted the survey; what questions were asked
about furloughed working; and over what time period the survey was carried
out. We have not audited the underlying data.
Survey research
7 To gain feedback from employees furloughed as part of CJRS, we
commissioned Ipsos MORI to carry out primary survey research on our behalf.
Ipsos MORI interviewed a sample of 8,445 adults aged 18+ in the UK using a
quota sampling method. Research was conducted using its telephone omnibus
and online interviews. For both methodologies, quotas were set on age, gender,
working status, education level, ethnicity and government office region to
reflect the general population. The quotas were set based on the 2020 mid‑year
population estimates for UK adults 18+ published by the ONS. Quota sampling
techniques provide accurate information about the responding sample, but
findings derived from this approach cannot generally be used to make inferences
about a wider population in the same way as having taken an entirely random
sample of respondents from a defined population, where each potential
participant has a known probability of being selected.
8 A total of 4,367 interviews were conducted via the telephone (CATI)
omnibus between 14 August and 2 October 2020. The CATI omnibus is a
nationally representative sample of adults (18+) using standard landline Random
Digit Dialling (RDD), mobile RDD and targeted mobile samples. Interviewing took
place over a four-week period, from a Friday to the following Wednesday and
interviewing was conducted between 9am and 9pm on weekdays and 10am
to 7pm on weekends.
Implementing employment support schemes in response to the COVID-19 pandemic Appendix Two 63
9 A total of 4,078 interviews were conducted online between 26 August and
8 September 2020. For the online survey, a nationally representative sample
of adults aged 18+ were interviewed online. The research was conducted by
contacting respondents through Ipsos Interactive Services. The panel consists
of individuals who have volunteered to take part in market research surveys.
The respondent panel is continuously maintained and monitored on response
rates, attrition, data quality and respondent engagement, as well as on
demographic composition to ensure gender, age, regional and socio‑economic
representation. Panellists receive appropriate incentives to participate in
surveys. Email invitations were sent out to the selected respondents, with a
unique survey link for each respondent. Respondents participating in the study
were actively monitored during fieldwork across a number of criteria, such as
response time, multiple participation, and so forth, and this process was repeated
at the end of fieldwork. Of the sample, 47 cases were removed as a result of
noticeable primacy effect, speeding or odd answer profiles and are not included
in the totals above.
10 Telephone and online data were weighted separately to the known offline
population proportions for age within gender, social grade within gender, region
within gender, working status within gender, ethnicity and education using rim
weighting. The weight targets were based on the 2020 mid-year population
estimates for UK adults aged 18+ published by ONS. An additional second stage
rim weight was applied to both data sets to yield an even blend within the final
combined weighted data.
11 We observed some variation in response between online and telephone
survey methods on the key question “Still thinking specifically about the period
between 19 March – 30 June 2020, did your employer of your main job ask you
to work while you were furloughed, or not?” The figure of 9% of furloughed
employees who worked following a request from their employer is an aggregated
result; separate figures for respondents interviewed by telephone and for those
undertaking the survey online were 4% and 14% respectively.
International comparison
12 We also engaged with our national audit counterparts in other countries to
better understand the different ways in which their respective governments were
implementing short-time working or wage subsidy schemes in response to the
pandemic. See Appendix Three for more details.
64 Appendix Two Implementing employment support schemes in response to the COVID-19 pandemic
Limitations
13 In reaching our independent views, we are aware of the following limitations
to our review of HMRC’s employment schemes and the value for money
conclusion we draw:
• We conducted our audit at a relatively early stage and the full impact of the
schemes is not yet clear.
• We conducted all our feedback remotely and, in the context of the
pandemic, could not supplement our fieldwork with site visits.
• Survey research draws on quota sampling techniques which cannot
generally be used to make inferences about a wider population (see above).
• The absence of a clear estimate of the nature and extent of fraud and error
meant we were unable to definitively conclude on how well risks have been
managed. HMRC expects to produce a more certain estimate, but not before
the end of 2021, and we will need to consider that as part of any future work.
Implementing employment support schemes in response to the COVID-19 pandemic Appendix Three 65
Appendix Three
International comparisons
66 Appendix Three Implementing employment support schemes in response to the COVID-19 pandemic
Figure 21
International Comparison
Most countries we reviewed had pre-existing schemes
Country Short-time Wage subsidy Maximum monthly Estimated total cost Initial period
work scheme? scheme? benefit1 to date given of support2
UK Yes No £2,500 £39.3 billion 1 March to
(New) (20 September 2020) 31 May 2020
Australia No Yes £1,791 £33.1 billion 30 March to
(New) (6 October 2020) 27 September 2020
Canada Yes Yes £2,144 £24.0 billion 15 March to
(Pre-existing) (New) (4 October 2020) 6 June 2020
Denmark Yes No £3,637 £1.5 billion 9 March to
(Pre-existing (12 October 2020) 9 June 2020
and new)
France Yes No £6,249 No data3 1 March to
(Pre-existing) 31 December 2020
Germany Yes No £3,385 £27.0 billion4 1 March to
(Pre-existing) 31 December 2020
Ireland5 Yes Yes £1,603 £2.6 billion 26 March to
(Pre-existing) (New) (8 October 2020) 17 June 2020
Netherlands5 Yes Yes £8,603 £17.9 billion 1 March to
(Pre-existing) (New) (6 October 2020) 31 May 2020
Sweden Yes No £3,827 £2.5 billion 16 March to
(Pre-existing) (14 October 2020) 31 December 2020
USA6 Yes Yes Employee Retention ERC: No data ERC: 12 March to
(Pre-existing) (New) Credit (ERC): 31 December 2020
PPP: £402.0 billion
£3,831 (one-off)
(8 August 2020) PPP: 3 April to
Paycheck Protection 30 June 2020
Program (PPP):
£6,385 (benefit for
four weeks, maximum
eight weeks)
Notes
1 All maximum monthly benefits and estimated total costs have been converted to pound sterling based on the exchange rates as at
14 October 2020. Monthly benefit figures have been converted to the per-month equivalent where necessary.
2 Some countries have subsequently extended their support schemes beyond the initial period. For example, the UK Coronavirus Job Retention
Scheme was initially due to run only to 31 May but was subsequently extended to 31 October 2020.
3 Total spend figures have not been confirmed for France.
4 The German Supreme Audit Institution told us in August that the Federal Employment Agency estimated €30 billion in spending would be
attributable to increasing numbers of workers using the Kurzarbeit short-time work scheme throughout 2020.
5 In Ireland and the Netherlands, existing short-time work schemes were replaced by new wage subsidy schemes.
6 The ERC and PPP provide a benefit to employers with the intention of mitigating layoffs. The PPP scheme provides a loan which may convert into
a grant if employee retention conditions are met. The CARES Act 2020 also provided funding for states to establish short-time work programmes,
however the US Government Accountability Office told us that at 30 June 2020 no federal funding had been obligated for spending.
Sources: National Audit Office analysis of data from the Supreme Audit Institutions contacted; Institute for Government report (2020): Coronavirus and
unemployment: a five-nation comparison; Organisation for Economic Co-operation and Development (OECD) report (2020): Job retention schemes
during the COVID-19 lockdown and beyond; Governmental websites.
Implementing employment support schemes in response to the COVID-19 pandemic Appendix Three 67
Appendix Four
Evolution of HM Revenue & Customs’ employment
support schemes
Figure 22
Comparison of Coronavirus Job Retention Scheme (CJRS) and Job Support Scheme (JSS)
Government support has changed over time
Employee pay
Government pays
CJRS 1 June 80
reduction
20
Employee pay
Employer pays Government pays
CJRS 2 October 60
reduction
20
20
Employee must work at least Government pays Employee pay
Employer pays
JSS 33 22 reduction
22
22
Government pays Employee pay reduction
Expanded JSS 67 33
0 10 20 30 40 50 60 70 80 90 100
Percentage (%)
Notes
1 From August 2020, the employer has paid National Insurance Contributions and minimum pension contributions. Employers can top up
employee salaries if they want. For all schemes, the government has capped the maximum level of support it provides.
2 For CJRS 2 and JSS, government support is based on the hours not worked by the employee. For CJRS 2, the level of support reduced
over time to 60% in October, the final month of that scheme.
3 Percentages may not sum to 100 due to rounding.
4 The Expanded JSS is for businesses legally required to close as a result of lockdown restrictions.
Source: National Audit Office analysis of government information
68 Appendix Four Implementing employment support schemes in response to the COVID-19 pandemic
Figure 23
Comparison of Self-Employment Income Support Scheme (SEISS)
and SEISS Extension
Government support has changed over time
Self-Employment Income Support
80
Scheme 1st Grant (SEISS 1)
Self-Employment Income Support
70
Scheme 2nd Grant (SEISS 2)
Self-Employment Income Support
Scheme Extension (SEISS Extension) 20
from November
0 10 20 30 40 50 60 70 80 90 100
Percentage of average monthly trading profits (%)
Note
1 The maximum level of government support is also capped.
Source: National Audit Office analysis of government information
CORRECTION SLIP
Title: Implementing employment support schemes in response to the
COVID-19 pandemic
Session: 2019–2021
HC 862
ISBN: 978-1-78604-339-9
Ordered by the House of Commons to be printed on 21 October 2020
Correction One:
Paragraph 15 (page 9) of the report was produced in error, the second sentence
referred to a ‘high proportion’, it should have been 12%.
The paragraph currently reads:
15 The Departments considered the equality implications of their design
decisions. The Departments carried out equality impact assessments for both
schemes. They identified that a high proportion of Black, Asian and minority
ethnic people were self-employed and undertook additional work to raise
awareness of SEISS with stakeholder groups. HMRC’s monitoring data provide
information on the age and gender profiles of people covered by the schemes but
do not report on other protected characteristics, such as ethnicity, as these data
are not necessary for the administration of taxes. The data show that a greater
proportion of younger workers were furloughed. A greater proportion of men were
furloughed initially, but this reduced over time and by September there was little
difference between men and women. In August 2020 HMRC began tendering
for survey and qualitative research to gain additional feedback on the schemes
(paragraphs 2.17 to 2.22 and Figure 10).
The paragraph should read:
15 The Departments considered the equality implications of their design
decisions. The Departments carried out equality impact assessments for both
schemes. They identified that 12% of self-employed workers had a Black,
Asian or minority ethnic background and undertook work to raise awareness of
the SEISS scheme with stakeholder groups. HMRC’s monitoring data provide
information on the age and gender profiles of people covered by the schemes but
do not report on other protected characteristics, such as ethnicity, as these data
are not necessary for the administration of taxes. The data show that a greater
proportion of younger workers were furloughed. A greater proportion of men were
furloughed initially, but this reduced over time and by September there was little
difference between men and women. In August 2020 HMRC began tendering
for survey and qualitative research to gain additional feedback on the schemes
(paragraphs 2.17 to 2.22 and Figure 10).
BACK
Correction Two:
Paragraph 2.19 (page 35) of the report was produced in error, ‘one-third’ should
have been 12%.
The paragraph currently reads:
2.19 HMRC’s further analysis to support its communication strategy showed
that one-third of self-employed workers had a Black, Asian or minority ethnic
background, with almost half based in London. HMRC developed a range of
communication and engagement plans for the schemes and worked with ethnic
minority and faith-based organisations to publicise the schemes.
The paragraph should read:
2.19 HMRC’s further analysis to support its communication strategy showed that
12% of self-employed workers had a Black, Asian or minority ethnic background,
with almost half based in London. HMRC developed a range of communication
and engagement plans for the schemes and worked with ethnic minority and
faith-based organisations to publicise the schemes.
BACK
Date of correction: 12 October 2020
This report has been printed on Pro Digital
Silk and contains material sourced from
responsibly managed and sustainable
forests certified in accordance with the
FSC (Forest Stewardship Council).
The wood pulp is totally recyclable and
acid-free. Our printers also have full ISO 14001
environmental accreditation, which ensures
that they have effective procedures in place to
manage waste and practices that may affect
the environment.
£10.00
ISBN 978-1-78604-339-9
Design and Production by NAO External Relations
DP Ref: 009194-001 9 781786 043399
You have reached the end of this document
File and source
- File
- implementing-employment-support-schemes-in-response-to-the-covid-19-pa.pdf
- Size
- 741,828 bytes
- SHA-256
- 4a9438925b8a546076b03d85b775c1c990e921773b8c30ef6961ac884b8530d9
- Original
- www.nao.org.uk