Research and analysis

The Coronavirus Job Retention Scheme final evaluation

Published 17 July 2023

A joint evaluation by HM Treasury (HMT) and HM Revenue and Customs (HMRC) of the Coronavirus Job Retention Scheme (CJRS).

Acknowledgements

HM Treasury (HMT) and HM Revenue and Customs (HMRC) would like to take this opportunity to thank all research participants who gave up their time to share their experiences to inform this final evaluation.

HMT and HMRC would also like to give thanks to our peer reviewers, Robert Joyce from the Institute for Fiscal Studies (IFS), and Alessia Montinaro with support from additional economists from the Department for Transport (DfT), who reviewed a draft of this evaluation. All decisions made about subsequent revisions to the evaluation were made by HMT and HMRC and all conclusions are our own.

Glossary

Term	Definition
CJRS	The Coronavirus Job Retention Scheme. The scheme was launched in April 2020 and aimed to protect jobs affected by the coronavirus (COVID-19) pandemic. The scheme initially offered employers the opportunity to apply for a grant to fund the wages of their employees who were on furlough, equivalent to 80% of usual wages up to £2,500 per month.
Furlough	Furlough is a temporary leave of absence from work. Employers who put employees on furlough were entitled to claim for a proportion of their wages through the CJRS.
Flexible furlough	From July 2020 the CJRS introduced flexible furlough. This provided employers with the flexibility to bring employees who were on furlough back to work part-time. Employers were able to bring employees back to work for any amount of time and any shift pattern, while still being able to claim the CJRS grant for hours not worked.
Users	Employers who were eligible for the CJRS who used the scheme.
Non-Users	Employers who were eligible for the CJRS who did not use the scheme.
Agents	People such as accountants, working either independently or for a firm, who had applied for the CJRS on behalf of an employer or provided support during the application process.
Large employer	An organisation with 250 or more employees.
Medium employer	An organisation with 50 to 249 employees.
Small employer	An organisation with 10 to 49 employees.
Micro employer	An organisation with 1 to 9 employees.
NPIs	Non-pharmaceutical interventions. These were behavioural and social interventions applied as part of the government’s response to the COVID-19 pandemic. These are measures intended to reduce COVID-19 transmission by reducing contact rates in the general population, for example national lockdown and social distancing measures.
PAYE	Pay As You Earn is the system for deducting and collecting Income Tax and National Insurance contributions from employment income.
NICs	For the purposes of this evaluation, this denotes Class 1 National Insurance contributions, paid by employers and employees. These enable individuals to qualify for certain benefits and the State Pension.
Employments	Each job reported through a PAYE scheme counts as an employment. Where someone is employed by more than one employer, their employments are counted separately. For example, an employee with jobs at 2 employers was counted twice in number of employments on CJRS data, if both jobs were on furlough.
Eligible employments	Jobs which were eligible to be claimed for via the CJRS, if the employee in the post was put on furlough. The eligible employments changed over time during the scheme was based on a list of employees included in Real Time Information (RTI) submissions.
Eligible employers	Employers who employed one or more eligible employees who met the criteria to be claimed for via the CJRS, if put on furlough.
RTI	RTI (Real Time Information) is the system used by employers to report to HMRC each time they pay their employees. Under RTI, information about PAYE, NICs and other deductions is transmitted to HMRC by the employer every time an employee is paid.
Error and fraud	Error is non-deliberate over- or under-payment, typically due to mistake, misunderstanding or misapplication of rules. Fraud is any deliberate omission, concealment, or misinterpretation of information, or the false or deceptive presentation of information or circumstances in order to gain an advantage.
Random Enquiry Programme	Taxpayers are selected at random for a post-payment compliance audit to enable HMRC to understand the scale of any incorrect claims.
Tax gap	The tax gap is the difference between the amount of tax that should, in theory, be paid to HMRC, and the amount that is actually paid.
SEISS	The Self-Employment Income Support Scheme was set up by the government to provide financial support to self-employed individuals in response to the COVID-19 pandemic.
Gross Domestic Product	A measure of the total value of goods and services in the economy.
Q1, Q2, Q3, Q4	Fiscal quarters from the start of a specified year. For example, 2020 Q1 would represent the first quarter in 2020, which is the 3 month period between the beginning of January and the end of March.
Deadweight	Desired policy outcomes that would have taken place without the intervention under consideration.
Value for money	A way of assessing the value from government spending, and the optimal use of resources to achieve the intended outcomes.
Net Present Value	The difference between the present value of cash inflows and the present value of cash outflows over a period of time.
Green Book	Guidance issued by HMT on how to appraise policies, programmes and projects. It also provides guidance on the design and use of monitoring and evaluation before, during and after implementation.
Magenta Book	Guidance issued by HMT on what to consider when designing an evaluation.
Equivalised net household income	Refers to household income that has been recalculated to take into account differences in household size and composition.

Executive summary

Introduction

The coronavirus (COVID-19) pandemic had far-reaching consequences for the UK labour market, households, businesses, and the economy more widely – some of which are still being felt. The Coronavirus Job Retention Scheme (CJRS) was a central part of the government’s response to this historic shock to the UK economy, centred on helping to stabilise the labour market by protecting jobs to ensure the economy could recover more smoothly.

Announced on 20 March 2020, 3 days before the first national lockdown, the CJRS sought to support the millions who faced disruption. A total of 1.3 million employers and 11.7 million employments were supported by the CJRS, with claims totalling £70 billion.

This final evaluation completes the government’s comprehensive evaluation of the CJRS. An interim evaluation, published in October 2022, set out initial findings on the early stages of the CJRS. The CJRS final evaluation covers the impacts across the full duration of the CJRS and also assesses the scheme’s value for money (VfM), including a consideration of deadweight and lessons learned. 

The evaluation draws on evidence from a range of sources, including internal research and analysis undertaken by HM Treasury (HMT) and HM Revenue and Customs (HMRC), as well as research commissioned by HMRC and conducted by an independent research agency, Ipsos. The evaluation has been peer reviewed by 3 external reviewers to provide independent challenge.

The scheme enabled employers to put their employees on furlough and claim a grant to cover up to 80% of their wages. This design aimed to ensure that employees could retain their job and the majority of their usual salary, even if COVID-19 disruption and restrictions meant that they could not work and that their employer could not afford to pay them. The scheme was universal in design, covering almost all employees across the UK, with only minor exceptions to prevent fraudulent abuse.

Over time, changes were made to the scheme to encourage employers to bring employees off furlough and back to work in line with the relaxation of COVID-19 restrictions. This was achieved through new flexibilities allowing employers to claim for employees who were only on furlough for part of their usual working time and increasing the amount employers would need to contribute towards their wages. These changes struck a balance between ensuring that those who required support could still access it, whilst representing value for money for the taxpayer, as well as a balance between providing support to those who needed it without preventing necessary labour reallocation. Likewise, the timing of the closure of the CJRS in September 2021 was a carefully judged decision balancing both the economic case for further support and helping employers and employees to restart their activity to allow the UK economy to recover, and providing value for money for taxpayers.

Main findings

1. The CJRS was good value for money, with a positive net benefit to society of £50 billion and a social benefit to cost ratio of around 4:1

The amount spent on the CJRS was substantial, with around £70 billion paid out. The value for money (VfM) assessment shows that this spending had significant positive impacts on jobs, incomes, businesses, tax revenue, welfare benefits spending and the macroeconomy.

The main economic benefit of the scheme was to ensure a quick and smooth economic recovery by preserving job matches and protecting jobs. This prevented employer closures and falls in employees’ income, which would also support household spending. The CJRS directly protected around 4 million jobs, and saved many businesses from closure, which contributed to lessening the economic impacts of the pandemic. Analysis of the impact on the UK labour market shows that without the CJRS, the level of economic output could have been around 1.8%^{[footnote 1]} lower in 2021 as a result of reduced employment, resulting in a long-lasting and significant scarring effect on the economy.

The unprecedented nature of the pandemic makes establishing impacts more uncertain, but the VfM assessment suggests that:

• the CJRS was a highly effective scheme providing good VfM, with a clear positive social benefit. Social Net Present Value (NPV) provides an established method for appraising the overall economic benefits of the scheme. The central estimate for the overall social NPV benefit of £50 billion, with a social benefit to cost ratio of around 4:1. These calculations are based on a thorough quantified appraisal of the scheme, using the principles of the Green Book to assess social costs and benefits
• looking just at direct impacts on the government’s finances once accounting for additional tax revenue, the central estimate for the overall exchequer NPV cost is estimated to be £25 billion. Clearly a significant amount but much lower than the gross CJRS payments
• the CJRS was an equitable scheme, benefitting a broad base of working households across the income distribution. It particularly benefitted businesses and workers in those sectors that were most affected by COVID-19, such as hospitality and those within the workforce on lower incomes. The CJRS produced an equity benefit to society of around £6 billion, included as part of the social value estimate
• however, the VfM assessment shows that 4.7% of spend (£3.3 billion) may have been paid out to employers who were entitled to claim the support but who may not have directly needed it, indicating potential deadweight in the scheme. Some deadweight is inevitable in any job retention policy, especially one developed at pace in response to an unprecedented global crisis, where there was a need to act quickly to avert job losses and put in place a scheme which was easy for employers to operate

The government’s experience from the scheme provides some valuable lessons, including on quickly delivering support and the need for improvements to data and processes. Furthermore, the policy choices throughout the scheme, while specific to the issue at hand, provide insight on how future widescale or emergency support can achieve optimal VfM.

2. The CJRS was devised and implemented at pace, reaching employers and their employees in need of support and achieving very high levels of satisfaction amongst both businesses and individuals

The significant impact of the CJRS was possible because of a highly successful process to design and implement the scheme at pace. As covered in the CJRS interim evaluation, after designing a scheme within days in March 2020, applications were opened for the CJRS grants by 20 April 2020. Employers were very satisfied with the scheme, with 9 out of 10 employers and agents pleased with the scheme’s timeliness and clarity. While the vast majority of claims were made online, 98% of customers who wanted to speak to an HMRC adviser were able to do so between March and October 2020. Employees on furlough generally also perceived the scheme positively.

3. The CJRS directly protected around 4 million jobs and saved many employers from permanent closure

As its name suggests, the CJRS was a job retention scheme. The benefit of protecting jobs and retaining employer-employee matches was that it allowed for a quicker and smoother recovery as restrictions were eased, as employers and employees were able to restart activity more promptly and efficiently than if employees had been made redundant.

Analysis estimates that around 4 million more individuals were in employment than would have otherwise been the case without the CJRS. The impact on protecting jobs was stronger towards the start of the scheme, when the impacts of the COVID-19 pandemic were at their most severe. The scheme also played an important role in protecting businesses and supporting the wider economy. Research suggests that 20% of employers who used the CJRS (around 250,000) would have closed permanently without the scheme but continued to trade throughout the pandemic or closed temporarily.

4. The CJRS played an important part in supporting the UK economy, in accelerating the labour market’s recovery from the COVID-19 pandemic and restrictions on activity, as well as reducing the long-term damage to the economy. Furthermore, the scheme supported household incomes and living standards, and therefore consumption and the macroeconomy. The gradual ending of support by September 2021 mitigated a potential spike in unemployment

Beyond the scheme’s direct benefits in protecting jobs and businesses, the CJRS also had a wider role in supporting the UK economy. Employer research shows it boosted business confidence at a time of great uncertainty. The scheme supported a large proportion of employees’ pay and as a result created additional consumer spending. Overall, the CJRS’s major macroeconomic benefit was to support economic recovery by preserving job matches, preventing falls in income and supporting household spending.

Potential negative impacts on labour market performance are explored in this final evaluation. The CJRS was found to have not restricted employees from moving around the economy during the pandemic in a way that could have harmed productivity substantially. There is also little evidence that the scheme had a role in contributing to the higher inactivity rates seen in the labour market over the last 2 years, including for employees aged 50 and over. Analysis shows that employees placed on furlough through the CJRS were no more likely to become inactive than employees not placed on furlough.

The scheme’s closure in September 2021 had very little impact on unemployment, despite the CJRS still supporting 1.14 million jobs at the time. Research suggests that employers welcomed the support remaining until September 2021, as it enabled them to be ready to reopen swiftly as non-pharmaceutical interventions (NPIs) were gradually ending. The impact of the COVID-19 pandemic and use of the CJRS varied across sectors. The significant uncertainty over the potential impacts meant that the timing of the closure of the scheme was a carefully judged decision for the government, as it had to ensure it did not unwind many of the benefits gained from the scheme, for example by risking a large rise in unemployment.

5. The scheme’s design achieved an appropriate balance between getting support to employers quickly and managing the risk of error and fraud. The scale and nature of error and fraud changed throughout the different claim periods of the CJRS as scheme design and eligibility rules changed

Given the need to support employers quickly at a period of significant uncertainty and to have a scheme that was easy for employers to operate and understand, the CJRS achieved an appropriate balance in mitigating the risk of error and fraud. The final estimate of the level of error and fraud within the CJRS lifecycle of 1 March 2020 to 30 September 2021 was found to be at the lower end of the initial planning assumptions of 5% to 10% with a most likely estimate of 5.1%, lying in the range of 3.0% to 7.8%. In monetary terms, this corresponds to a most likely estimate of £3.5 billion, with a range of £2.0 billion to £5.4 billion.

This estimate is similar to last year’s previously published most likely estimate, with a slight widening of the range of estimates. This is data-driven and not because we are less certain in our estimate; it is because a wider range of overclaims were found in the second CJRS random enquiry programme. While difficult to draw a direct comparison, for context the UK tax gap in 2021 to 2022 was estimated to be 4.8% and the rate of error and fraud in child and working tax credits has been around 5% in recent years.

From the beginning, it was expected that the CJRS would be a target for opportunistic fraud, and that some customers would make mistakes. The scheme design aimed to prevent as much error and fraud as possible before any payments were made, particularly the risks associated with organised criminal attacks. Over 22,000 suspicious CJRS claims were blocked from being paid, valued at over £114.7 million. The final estimates for the CJRS error and fraud demonstrate that HMRC’s pre-payment controls have been effective, with the amount lost to organised crime and criminal attacks being significantly lower than anticipated.

Employers claiming for employees who were working contributed the largest proportion of error and fraud across the whole lifecycle of the CJRS, however the prevalence of this particular error and fraud risk compared to other risks reduced over time. This is likely because of changes to scheme design and eligibility rules, such as varying the levels of claim amounts which allowed employers to use the scheme more flexibly. However, the introduction of these changes later in the scheme’s lifecycle did introduce more scope for error from risks other than employers claiming for employees who were working due to added complexity.

Up to March 2023, the Taxpayer Protection Taskforce (TPT), established to further expand the scope of COVID-19 compliance activity, has recovered a total of £256.1 million of overpaid CJRS grants. This is in addition to the £518.8 million recovered prior to the taskforce being established. Furthermore, there were over £1 billion in unprompted disclosures and voluntary repayments from claimants where they identified an overpayment of a CJRS grant or if they wanted to voluntarily pay the grant back, as they no longer required it.

Chapter 1: Introduction

1.1 Evaluation scope

The Coronavirus Job Retention Scheme (CJRS) formed a core element of the government’s policy response to the COVID-19 pandemic. Given its importance, HM Treasury and HM Revenue and Customs (HMRC) have undertaken a comprehensive evaluation of the scheme, whilst recognising the challenges of evaluating an unprecedented policy launched in extraordinary circumstances. This evaluation programme covers the full duration of the CJRS, and it includes a process, impact and value for money (VfM) evaluation.

The CJRS interim evaluation covered the process and early impacts of the CJRS. It found that the scheme was designed and successfully operationalised at pace within weeks of the first national lockdown starting, enabling the scheme to have a significant impact. The CJRS claims service was easy for employers to use. The methods used to communicate the CJRS and its eligibility criteria resulted in high levels of awareness and understanding of the scheme. Between the scheme opening date and October 2020, 6.6 million CJRS payments were made, of which 99% were paid within 6 working days.

The National Audit Office (NAO) published a report on the CJRS in October 2022, focusing on the management of risks, refinement of the scheme and its impact. The CJRS was found to have been “commendably implemented at pace” and that it provided “essential support to the economy. and they [HM Treasury and HMRC] achieved their primary objective of preventing millions of job losses”. However, the NAO felt there should have been earlier efforts to bear down on deadweight and error and fraud. The government agrees that it is important to understand these consequences given the scheme’s significant cost and, as planned, this final evaluation assesses the scheme’s level of deadweight and error and fraud, addressing the NAO’s recommendations.

The final evaluation builds on the evidence and findings presented in the interim evaluation and looks at the scheme in its entirety, covering impact and VfM. It covers the impact on jobs, incomes, businesses and the macroeconomy. The VfM assessment includes a consideration of deadweight and concludes with lessons learned.

1.2 Evaluation objectives and evaluation questions

The objectives of the final evaluation are to assess the extent to which the scheme achieved its intended outcomes, which are detailed in chapter 1.5.2. The evaluation also explores the scheme’s reach, considering who was affected by the CJRS, as well as how effectively the scheme was implemented, particularly through the impact of steps to mitigate levels of error and fraud in the scheme.

Additional objectives of the evaluation are to explore employer and employee perceptions of the scheme and to gather lessons that can be learned to inform future economic public policy design and implementation.

The approach for the process evaluation, impact evaluation and VfM assessment are summarised below, and the evaluation questions are set out in full in table 2.1 in the CJRS final evaluation accompanying technical information document. This final evaluation:

• presents further detail on process - looking at how effective the scheme was at reaching the intended recipients, and the estimated level of error and fraud and how this was mitigated
• expands the impact evaluation - addressing the extent to which the scheme protected jobs, supported household incomes, reduced business closures, reduced economic scarring, and the extent to which the scheme’s closure affected outcomes - the impact on the macroeconomy is also assessed
• addresses the scheme’s value for money - including its effectiveness, fairness, costs and efficiency - it also sets out what lessons can be learned to inform future policy design and implementation

1.3 Evaluation approach and evidence sources

The evaluation approach is consistent with that detailed in the interim evaluation. This final evaluation will, in particular, utilise the economic principles for appraisal and evaluation, as outlined in the Green Book.

Following the Magenta Book guidelines for evaluation, external research was commissioned and has been conducted by an independent research agency, providing objective research findings. Final findings from this research have been published alongside this evaluation. A draft of the final evaluation was reviewed by external peer reviewers who critically assessed it to provide quality assurance of the evaluation approach and assess the robustness of the findings.

The final evaluation aligns with the VfM 4 Es framework (Economy, Efficiency, Effectiveness and Equity) used by the NAO, and Green and Magenta Book guidance to inform the VfM assessment.

Evidence used to inform the final evaluation is primarily sourced from:

• recent research conducted with employers and employees
• the analysis of HMRC management information and CJRS data; and statistical and economic modelling. In some places, findings are also supplemented with evidence from external data sources to provide context or comparison, and to increase the robustness of the evaluation’s findings.

Findings from the following 3 pieces of externally commissioned research are used throughout this evaluation:

• quantitative research with employers (wave 2): a probability sample telephone survey of 4,860 CJRS Users and 1,807 Non-Users. This research explored the impact of the CJRS in areas such as the protection of jobs and the continuation of operation. The sample sizes for the research were designed to enable robust sub-group analysis by characteristics such as sector and employer size
• qualitative research with employers (wave 2): in-depth telephone or online video interviews with 40 Users. This research covered the Users’ decision-making with respect to staffing and the self-reported impact of the scheme on employers
• qualitative research with employees: 80 in-depth telephone or online video interviews with people who either are or were working for an employer that had placed at least one employee on furlough. This explored the experience of being on furlough and perceived impacts of the CJRS

Further information about the evidence sources and analysis are detailed when setting out the methods at the beginning of each chapter and in the CJRS final evaluation accompanying technical information document.

1.4 Economic context and policy in the UK

The pandemic had a sudden and severe effect on the UK and international economies, with widespread falls in economic activity resulting from the global shock. This chapter primarily focuses on the economic context and policy in the UK. A more detailed analysis of the effect of job retention schemes in other countries is included in chapter 7.3.7.

As part of its economic response to the pandemic, in March 2020 the government began to rapidly implement an unprecedented package of measures. The measures included support grants, loans, easements, rebates and the CJRS and the Self-Employment Income Support Scheme (SEISS).

Reflecting the uncertainty at the time of the scheme’s implementation, and throughout the lifespan of the CJRS, considerable falls in employment and GDP were anticipated:

• in April 2020, the Office for Budget Responsibility’s (OBR) initial illustrative COVID-19 reference scenario estimated that real GDP would fall by 35% in April to June 2020, accompanied by a sharp rise in the unemployment rate to 10% (an increase in unemployment of 2.1 million, to a total of 3.4 million)
• in the OBR’s July 2020 Fiscal Sustainability Report, the projected unemployment rate rose to 12% (to a total of 4 million)
• in June 2020, the Organisation for Economic Co-operation and Development Economic Outlook estimated UK unemployment would hit 11.7% in 2020, without a second wave of COVID-19, and could peak at 14.8% with a second wave
• in the OBR’s March 2021 Economic and Fiscal OutlookOBR’s March 2021 Economic and Fiscal Outlook, unemployment was estimated to reach 6% following the second wave, with a second peak-to-trough fall in GDP of around 5%

Figure 1.1: UK unemployment rate (percentage) (aged 16 and over, seasonally adjusted), January 2006 to October 2022 and UK GDP Index January 2019 = 100, January 2007 to January 2023

Source: GDP monthly estimate, UK - Office for National Statistics (ons.gov.uk); Unemployment rate (aged 16 and over, seasonally adjusted): percentage - Office for National Statistics (ons.gov.uk)

Figure 1.1 shows that even with measures implemented, UK GDP contracted from peak-to-trough by 27.0% between January to June 2020, with record quarterly falls in services, production, and construction output. This fall drastically surpassed the peak contraction the UK experienced from the 2008 recession of 6.1%. Yet, in July to September 2020, the UK saw a strong bounce back in GDP, with a recovery of 17.6%. The OECD November 2022 Economic Outlook shows that other G7 countries saw a peak-to-trough contraction ranging from 8% to 19%, which is smaller than the UK. However, by achieving a high rate of growth subsequently, the UK reached similar pre-pandemic levels of GDP by April to June 2021. UK and G7 GDP recovered to near pre-pandemic levels by October to December 2021.

Figure 1.1 also shows that unemployment rose from 4.0% in the 3 months to February 2020, to peak at 5.2% in November 2020, resulting in a 925,000 reduction in employees on payroll. However, the peak in unemployment was well below the unemployment rate in the aftermath of the 2008 recession, which reached over 8%, and well below the estimated unemployment rates at the onset of the pandemic. Following this peak, UK unemployment continuously fell to below pre-pandemic levels by November 2021.

Since the pandemic, there have been lower levels of participation in the labour market, which government policies have looked to increase over time. The OBR March 2023 Economic and Fiscal Outlook shows that labour market participation among 16 to 64 year-olds is down by 520,000 or –0.9% compared to its pre-pandemic March 2020 forecast. Labour market participation is explored further in chapter 7.

1.5 CJRS policy overview

1.5.1 Design

The CJRS initially paid employers a grant worth 80% of the usual wages (capped at £2,500 per month) of employees put on furlough, in addition to covering associated employee costs of employer NICs and pension contributions (up to the level of the automatic enrolment minimum), for hours not worked. Users were required to use the grant to pay employees who remained on their payroll but had stopped working while on furlough.

To limit fraud, employers could originally only claim the CJRS grant for those who were employed and on their PAYE payroll on 19 March 2020, the day before the scheme was announced. This design feature meant some newly hired employees were unable to directly benefit from the scheme. As evidenced in the CJRS interim evaluation, the scheme was wide reaching. Employers who reported a negative impact of COVID-19 (redundancies, temporary closures or decreases in sales or turnover), were most likely to receive CJRS support. Most eligible non-claimants reported that they did not need support. However, some single-employee organisations felt the scheme was not suited to them and they did not use it.

Figure 1.2: Breakdown of government and employer contributions for a salaried employee earning £28,000 a year, on furlough for all of their hours, March 2020 to September 2021

Notes: This is an example only and employer NICs varied in some circumstances. The scheme covered the costs of employer NICs, up to the level actually incurred.

Table 1.1: Breakdown of government and employer contributions for a salaried employee earning £28,000 a year, on furlough for all of their hours, March 2020 to September 2021

Government and employer contributions	Oct 2020; Aug 2021 to Sep 2021	Sep 2020; July 2021	Aug 2020; Nov 2020 to Jun 2021	Mar 2020 to Jul 2020	Reference monthly salary
Government grant for wages (received by employee)	£1,400.00	£1,633.33	£1,866.66	£1,866.66	N/A
Minimum employer contribution for wages (received by employee)	£466.67	£233.33	£0.00	£0.00	N/A
Government grant for employer NICs and pension	£0.00	£0.00	£0.00	£196.98	N/A
Employer NICs	£156.58	£156.58	£156.58	£0.00	£220.98
Employer pension contribution (3% of qualifying earnings)	£40.40	£40.40	£40.40	£0.00	£54.40
Reference monthly salary	N/A	N/A	N/A	N/A	£2,333.33

Note: N/A is used in table 1.1, meaning not applicable.

Figure 1.2 shows that throughout the scheme, adaptions were made to the policy to reflect the changing COVID-19 pandemic situation and associated non-pharmaceutical interventions (NPIs):

• from 1 July 2020, flexible furlough was introduced. This allowed employees to return to work for any amount of time while employers were still able to claim for the hours not worked
• from 1 August 2020, employers could no longer claim associated employee costs (NICs and pension contributions)
• from 1 September 2020, employers were required to contribute 10% of usual pay in respect of any hours not worked, followed by 20% in October 2020
• from 1 November 2020, the government contribution was increased back to 80% and subsequently extended until June 2021 in response to the increase in NPIs - accompanying a new PAYE payroll cut-off of 30 October 2020
• in summer 2021, support was tapered ahead of the scheme closure, with employer contributions increasing to 10% from July, followed by 20% from August. The scheme ended on 30 September 2021

More detail on the policy design and changes to the policy can be viewed in the interim evaluation, with a policy timeline found in the accompanying information document.

1.5.2 Policy objectives

The main objective of the CJRS was to protect jobs and therefore prevent widespread unemployment resulting from the pandemic and associated NPIs. It sought to do this by supporting employers to keep existing employees in post, therefore retaining employer-employee job matches. This would enable a prompt restart of activity when NPIs eased, meaning a smoother and quicker economic recovery.

The CJRS also aimed to reduce the risk of permanent employer closures, by supporting those who faced making decisions on pausing or scaling back their business activities due to NPIs. The scheme design also aimed to ensure employees could retain the majority of their usual salary, thus supporting a large proportion of employees’ pay and therefore household incomes.

In addition, the CJRS was designed to benefit the macroeconomy, by reducing the risk of economic scarring (a sustained loss in the productive capacity of the economy) by supporting consumption.

As NPIs were eased, the focus of the scheme shifted towards encouraging labour back into productive work. Where jobs no longer had reasonable future prospects, it aimed to encourage reallocation into roles where there was a demand for labour. This was achieved through the introduction of flexible furlough, and by increasing the cost the employer had to bear to keep the employer on their payroll.

1.5.3 Overview of the use of the scheme

Figure 1.3 shows that employments on furlough supported by the CJRS peaked during the first wave of COVID-19 at 8.9 million in May 2020. This represents a take-up rate of 30% of total eligible employments. The number of employments placed on furlough via the CJRS then declined over time, reflecting the evolving path of the COVID-19 pandemic. However, there were noticeable increases when the scheme eligibility and end-date was extended from November 2020 and again in January 2021 in response to the increase in NPIs. This resulted in a secondary peak of 5 million employments on furlough in January 2021. After January 2021 there was a steady decline in employments on furlough as the vaccine programme was rolled out, NPIs eased, and the economy reopened, with the CJRS support from the government tapered before closure.

Figure 1.3: Total number of employments on full furlough and flexible furlough, March 2020 to September 2021

Source: HMRC CJRS data

Notes: Missing information on some employments – for example incomplete or not fully processed returns – means that whether an employee has been on flexible furlough is not known in some cases. The data presented in this chart can be accessed in the interim evaluation accompanying data tables

Chapter 2: Reach of the CJRS

Responding to the unprecedented challenges in the economy caused by COVID-19, the reach and impact of the CJRS was wide ranging. At its peak between March and June 2020, 61% of eligible employers made use of the scheme and 31% of eligible employments were claimed for; this fell to 21% of eligible employers and 4% of eligible employments at the time the scheme closed.

Research shows that the CJRS was used by employers who were in greater need of support. Employers who used the scheme were much more likely to have reported that they had closed temporarily or permanently at some point, or that COVID-19 had resulted in a decrease in sales or turnover, or led them to make redundancies.

Smaller employers, and those in sectors such as accommodation and food services, were the hardest hit by the pandemic and claimed support from the CJRS for the greatest proportion of their employees.

2.1 Introduction

This chapter addresses the evaluation question:

• how effective was the scheme at reaching the intended recipients?

It explores the extent to which the CJRS provided wide-reaching support and whether it effectively reached employers most negatively affected by the pandemic, as these employers were the most likely to make redundancies, or reduce employees’ pay or hours. As a result, this would support employees most at risk of losing their jobs, the primary objective of the CJRS.

2.2 Method

This chapter presents take-up rates from the CJRS Official Statistics, alongside findings from the CJRS employer quantitative research (wave 2) and CJRS employer qualitative research (wave 2), to assess the reach of the entire scheme. ‘Users’ refers to employers who were eligible for the CJRS who used the scheme. ‘Non-Users’ refers to employers who were eligible for the CJRS who did not use the scheme.

The impact questions on turnover and redundancies cover the period between the start of the pandemic and its closure. The impact question on trading status covers the period between the start of the pandemic and when respondents were surveyed between June 2022 and November 2022. Survey questions relating to turnover were asked to those actively trading at the time of the survey (94% of Users and 95% of Non-Users).

The take-up rates presented here were produced using the CJRS and PAYE Real Time Information (RTI) data. Employer take-up rates reflect the proportion of all eligible employers who made at least one claim through the CJRS. Employment take-up rates reflect proportion of all eligible jobs that were claimed for by employers through the CJRS. The CJRS employer and employment take-up rates in this chapter are separated into the 3 eligible periods that reflect the changes made to the eligibility criteria each time the CJRS was extended.

2.3 Findings

2.3.1 Use of the scheme

The CJRS was widely used by different groups. A total of 11.7 million unique employments were on furlough for at least part of the CJRS since it began. When accounting for employees with more than one job, this amounted to 10.8 million individuals who benefited from the scheme through having their jobs supported.

The proportion of eligible employers with staff on furlough, and the proportion of employments claimed for, peaked between March and June 2020 at 61% and 31% respectively. This fell to 21% of eligible employers and 4% of eligible employments at the time the scheme closed. The CJRS employer quantitative research found that 82% of users said that at least one of their employees on furlough had returned to work, either full or part-time, before the scheme closed. There were 1.16 million employments on furlough at the time the scheme closed, across 410,000 employers.

2.3.2 Reaching employers hardest hit by COVID-19

In line with the findings in the CJRS interim evaluation, Users were much more likely than Non-Users to have reported that they had closed their business temporarily or permanently because of COVID-19 (32% compared to 5%). Users were also more likely to report a decrease in sales or turnover (72% compared to 40%), or redundancies (15% compared to 3%) due to COVID-19, as shown in figure 2.1. The impact on turnover levels was also more significant for Users than it was for Non-Users, with an average decline in turnover of 43% compared with 33%. This demonstrates that those who used the scheme were more likely to have been negatively impacted by the pandemic.

The proportion of employers reporting a decline in turnover was lower in wave 2 of the research than in wave 1, among both Users (72% compared with 83%) and Non-Users (40% compared with 48%). This suggests that the longer-term impact of the pandemic on sales or turnover was slightly less pronounced than the shorter-term impact. Around 1 in 7 Users (15%, or around 197,000 employers) made redundancies due to the impact of COVID-19, whilst the CJRS was operating. For Users who made redundancies, the average number of redundancies was 4, and the average proportion of their workforce made redundant was 38%.

In total, 4% of the total User workforce was made redundant due to COVID-19 (around 776,000 employees). Comparing these results with those from wave 1 research, the proportion of Users that made any staff redundant was the same (15%). This suggests that Users that made redundancies started to do so relatively early in the pandemic, which is discussed later in chapter 4.

With the strongest impact on turnover coinciding with employers making redundancies in that relatively early period of the COVID-19 pandemic, this suggests that the CJRS needed to provide support quickly. This is reinforced by the findings, shown in chapter 4, on how the scheme directly protected more jobs in the earlier months of the pandemic. The employer qualitative research found that Users’ main reasons for using the CJRS were for organisation survival, to protect jobs, to take precautionary action in case their outlook worsened, and to support their staff.

Figure 2.1: Negative impacts of COVID-19 on employers eligible for the CJRS, by use of the scheme

Source: CJRS employer and agent quantitative research wave 2

Base: Impact on turnover - actively trading eligible employers (6,366), actively trading Users (4,615), actively trading Non-Users (1,751). Impact on closure and redundancy - all employers (6,667), all Users (4,860), all Non-Users (1,807).

Question: Between the start of the pandemic and when the CJRS closed, what impact, if any, had COVID-19 had on the funding that your organisation received/ your organisation’s sales or turnover?; What is the current status of your organisation and for what reasons would you say your organisation stopped trading?; In total, how many staff, if any, had you made redundant between the start of the pandemic in March 2020 and until the scheme closed on 30 September 2021, due to the impact of COVID-19?

Notes: The trading status question covers the period up until the respondent was surveyed which was after the CJRS had closed. 

*Only employers actively trading at the time they were surveyed were asked whether COVID-19 had a negative impact on their turnover.

Table 2.1: Negative impacts of COVID-19 on employers eligible for the CJRS, by use of the scheme

Impact	Eligible employers	Users	Non-Users
Negative impact on sales or turnover before the CJRS closed due to COVID-19*	62%	72%	40%
Stopped trading temporarily or permanently due to COVID-19	24%	32%	5%
Made redundancies before the CJRS closed due to COVID-19	12%	15%	3%

While the survey found that Non-Users were generally less adversely affected than Users, many did still see negative impacts as shown in figure 2.1. The majority (66%) of the Non-Users who were aware of the scheme said that they did not apply because they reported not needing to. However, 21% of the Non-Users who were aware of the scheme reported they had not applied because they mistakenly did not think they were eligible. These Non-Users were more likely to be sole proprietors and from the arts, entertainment and recreation sector. They were more likely to have seen a decrease in sales or turnover than other Non-Users, however there were no significant differences with respect to redundancies or business closure. This suggests they were only somewhat more negatively affected by the pandemic than other Non-Users. Non-Users who were aware of the scheme also reported being able to get by using other forms of support or internal actions (8%) and not wanting to furlough employees (6%) as reasons for not using the CJRS.

2.3.3 Sector

Employers in all sectors reported that there were negative impacts of the pandemic on their business. Eligible employers in the accommodation and food services sector, the arts, entertainment and recreation sector and other services sector were the hardest hit. They reported the highest levels of temporary or permanent closure^{[footnote 2]}, and a decrease in sales or turnover due to COVID-19. Accommodation and food services were also significantly more likely than average to have made redundancies due to COVID-19, as shown in figure 2.2.

Figure 2.2: Negative impacts of COVID-19 on employers eligible for the CJRS, by overall and top three affected sectors

Source: CJRS employer and agent quantitative research wave 2

Base: Impact on turnover - actively trading eligible employers (6,366), actively trading sectors (all > 200). Impact on closure and redundancy - all eligible employers (6,667), sectors (all > 200)

Question: Between the start of the pandemic and when the CJRS closed, what impact, if any, had COVID-19 had on the funding that your organisation received/ your organisation’s sales or turnover?; What is the current status of your organisation?… and for what reasons would you say your organisation stopped trading?; In total, how many staff, if any, had you made redundant between the start of the pandemic in March 2020 and until the scheme closed on 30 September 2021, due to the impact of COVID-19?

Notes: The trading status question covers the period up until the respondent was surveyed which was after the CJRS had closed. 

*Only employers actively trading at the time they were surveyed were asked whether COVID-19 had a negative impact on their turnover.

Table 2.2: Negative impacts of COVID-19 on employers eligible for the CJRS, by overall and top three affected sectors

Sector	Negative impact on sales or turnover before the CJRS closed due to COVID-19	Stopped trading temporarily or permanently due to COVID-19	Made redundancies before the CJRS closed due to COVID-19
Eligible employers	62%	24%	12%
Accommodation and food service activities	81%	57%	16%
Arts, entertainment and recreation	80%	48%	14%
Other service activities	75%	42%	9%

The CJRS Official Statistics show that employers in the accommodation and food services, arts, entertainment and recreation, and other services sectors were most likely to use the scheme. As shown in figure 2.3, this was also the case as the scheme came to a close in the final eligibility period: accommodation and food services had an employer take-up rate of 64%, arts, entertainment and recreation had a take-up rate of 50%, and other service activities had a take-up rate of 42% compared to an average of 34%. This reflects how demand was slowest to return to these sectors, as was found in the employer qualitative research.

Figure 2.3: CJRS employer take-up rate for three eligible periods, by overall and top three claiming sectors

Source: HMRC CJRS and PAYE RTI data

Note: Select sectors included.

Table 2.3: CJRS employer take-up rate for three eligible periods, by overall and top three claiming sectors

Employer Size	Employer take-up rate May to September 2021	Employer take-up rate November 2020 to April 2021	Employer take-up rate March to October 2020
Other service activities	42%	72%	78%
Arts, entertainment and recreation	50%	70%	76%
Accommodation and food services	64%	77%	89%
Eligible employers	34%	48%	61%

2.3.4 Employer size

Smaller employers eligible to use the CJRS were significantly more likely to have stopped trading temporarily or permanently due to COVID-19 and to have had a negative impact on their sales or turnover. Conversely, smaller employers were much less likely to have made redundancies as a result of COVID-19 than larger employers, while the scheme was open. However, this is likely to be linked to these employer sizes having considerably fewer employees, particularly in the case of micro employers where making redundancies is likely to have a bigger impact on the ability of an employer to continue operating. The variable impact by employer sizes is shown in figure 2.4.

Figure 2.4: Negative impacts of COVID-19 on employers eligible for the CJRS, by size

Source: CJRS employer and agent quantitative research wave 2

Base: Impact on turnover - actively trading eligible employers (6366), actively trading micro (4,209), actively trading small (1,126), actively trading medium (688), actively trading large (295). Impact on closure and redundancy - all eligible employers (6,667), all micro (4,457), all small (1,142), all medium (694), all large (298).

Question: Between the start of the pandemic and when the CJRS closed, what impact, if any, had COVID-19 had on the funding that your organisation received/ your organisation’s sales or turnover?; What is the current status of your organisation?… and for what reasons would you say your organisation stopped trading?; In total, how many staff, if any, had you made redundant between the start of the pandemic in March 2020 and until the scheme closed on 30 September 2021, due to the impact of COVID-19?

Notes: The trading status question covers the period up until the respondent was surveyed which was after the CJRS had closed. 

*Only employers actively trading at the time they were surveyed were asked whether COVID-19 had a negative impact on their turnover.

Table 2.4: Negative impacts of COVID-19 on employers eligible for the CJRS, by size

Employer size	Stopped trading temporarily or permanently due to COVID-19	Negative impact on sales or turnover before the CJRS closed due to COVID-19*	Made redundancies before the CJRS closed due to COVID-19
Eligible employers	24%	62%	12%
Micro (1 to 9)	25%	63%	9%
Small (10 to 49)	20%	60%	23%
Medium (50 to 249)	13%	49%	35%
Large (250+)	10%	45%	32%

While micro employers were the most likely to be negatively impacted by the COVID-19 pandemic, as shown in figure 2.5 the employer take-up rate amongst small, medium, and large employers was consistently higher than micro employers for the first 2 eligible periods. One possible explanation could be the difficulty smaller organisations would have to continue operating with reduced staff, as mentioned above. The employer take-up rate gradually fell for all employer sizes between the 3 eligibility periods, indicating that many employers brought staff on furlough back to work gradually. Employers in the employer qualitative research reported that they brought staff back to work gradually towards the end of the scheme based on the financial health of the business.

Figure 2.5: CJRS employer take-up rate by employer size for the three eligibility periods

Source: HMRC CJRS and PAYE RTI data

Note: Employer size refers to number of employees.

Table 2.5: CJRS employer take-up rate by employer size for the three eligibility periods

Employer Size	Employer take-up rate CJRS May to September 2021	Employer take-up rate CJRS November 2020 to April 2021	Employer take-up rate CJRS March to October 2020
Large (250+)	34%	62%	75%
Medium (50 to 249)	43%	65%	79%
Small (10 to 49)	43%	67%	83%
Micro (1 to 9)	47%	45%	57%
Eligible employers	32%	48%	61%

While micro employers were less likely than other employer sizes to use the CJRS for most of its existence, a much higher proportion of employments in micro-sized employers were claimed for than other sized employers throughout the scheme. This suggests that the scheme did provide significant support to employees working for employers whose sales and trading status were most negatively impacted (see figure 2.4). The employment take-up rates across all employer sizes decreased considerably over the 3 eligibility periods as restrictions on economic activity were reduced and the economy opened up.

2.3.5 Geography

Across the UK nations, there was minimal variation in the impact of the COVID-19 pandemic on trading status, sales or turnover for Users. This was also the case when viewing the impact over the length of the scheme. However, employers in Northern Ireland (17%) and Scotland (16%) were significantly more likely to have made redundancies than those in Wales (8%). England (11%) was in line with the UK average (12%).

Employer take-up rates by nation within the UK are not explored because HMRC does not hold this data. Employment take-up was significantly above the UK average in Wales, and somewhat lower than average in Northern Ireland. Factors that influenced this may have included the difference in the NPIs, and the different types of economic support available in each nation. This is supported by the employer qualitative research that found relaxed restrictions and increased business demand to be the main factors influencing the use of the CJRS over time.

Figure 2.6: CJRS employment take-up rates by nation for the 3 eligibility periods

Source: HMRC CJRS and PAYE RTI data

Note: The employment take-up rate is based on the nation that the employee lives in and may be different from where the employer is based.

Table 2.6: CJRS employment take-up rates by nation for the 3 eligibility periods

Country/Region	Employment take-up rate CJRS November 2020 to April 2021	Employment take-up rate CJRS May to September 2021	Employment take-up rate CJRS March to October 2020
Northern Ireland	12%	7%	20%
Scotland	20%	10%	33%
England	21%	11%	34%
Wales	33%	16%	54%
Eligible employments	23%	11%	38%

Chapter 3: Error and fraud

From the beginning, it was expected that the CJRS would be a target for fraud and that customers would make mistakes. Care was taken with the design of the administrative payment system to prevent as much error and fraud as possible, whilst making payments as quickly as possible to support employers and employees.

HMRC estimates that the rate of error and fraud for the CJRS over the lifecycle of the full scheme (1 March 2020 to 30 September 2021) is between 3.0% and 7.8% with a most likely estimate of 5.1%. In monetary terms, this corresponds to a most likely estimate of £3.5 billion, with a range of £2.0 billion to £5.4 billion.

The final estimate is similar to last year’s published most likely estimate and remains at the lower end of the level expected before the scheme was implemented, of between 5 to 10%. This demonstrates HMRC’s pre-payment controls have been effective, with money lost to organised crime significantly lower than anticipated. Fraudulent claims prevented before payments were made by HMRC are not reflected in the error and fraud rate. Despite active steps to reduce the level of error and fraud, as with any scheme it is not possible to eradicate all the error and fraud risks. While difficult to draw a direct comparison, for context the UK tax gap in 2021 to 2022, was estimated to be 4.8%

HMRC continues to deal with abuse of the COVID-19 support schemes using the full range of its powers – both civil and criminal.

3.1 Introduction

This chapter addresses the evaluation questions:

• what was the estimated scale of error and fraud?
• how was it controlled through the design and implementation of the scheme, including compliance checks?

The CJRS interim evaluation published in 2022 provided a thorough account of HMRC’s compliance approach, including establishing a Taxpayer Protection Taskforce (TPT), as well as the experiences of employers who claimed for the CJRS and followed the compliance approach. Therefore, these areas will not be covered in full again in this final evaluation. Latest information on the compliance approach can be accessed through the 2022 to 2023 HMRC annual report and accounts.

3.2 Method

This chapter draws together a broad evidence base including the results of Random Enquiry Programmes, surveys and HMRC administrative data in order to assess the level of error and fraud for the CJRS.

3.3 Findings

3.3.1 Estimating levels of error and fraud

The government’s latest estimate of CJRS error and fraud from 1 March 2020 to 30 September 2021 lies in the range of 3.0% to 7.8%, with a most likely estimate of 5.1%. In monetary terms, this corresponds to a range of £2.0 billion to £5.4 billion, with a most likely estimate of £3.5 billion. This will be the government’s final estimate of CJRS error and fraud.

During the design stage, HMRC’s initial assessment of a likely error and fraud risk was 5 to10%, before post-payment compliance action to identify and recover overpayments. The final CJRS error and fraud most likely estimate of 5.1% being at the lower end of the planning assumptions demonstrates that HMRC’s pre-payment controls have been effective.

HMRC have recently completed and reviewed the results of a second Random Enquiry Programme covering the whole lifecycle of the CJRS from 1 March 2020 to 30 September 2021, selecting claims at random for a post-payment compliance audit. This, in combination with the first Random Enquiry Programme completed last year, which covered claim periods from 1 March 2020 to 31 October 2020, has given a more comprehensive understanding of the level of error and fraud. HMRC also has further insight from its ongoing compliance activity. Updated estimates are now published in the 2022 to 2023 annual report and accounts along with an updated technical report detailing methodological changes since July 2022.

Although we now have stronger evidence overall to inform our estimate, the estimate range has widened slightly since last year’s published estimate. This is data-driven and not because we are less certain in our estimate. A wider range of overclaims were found in the second CJRS Random Enquiry Programme.

The latest and final estimate shows that the scale and nature of error and fraud changed throughout the different phases of the CJRS. This was primarily due to changes in scheme design and eligibility rules, such as employers having the flexibility to bring employees who were on furlough back to work part-time. Error and fraud was found to be highest in the first phase of the scheme (March 2020 to June 2020), particularly because this phase of the scheme was not flexible, with the eligibility criteria excluding employees on furlough who did any work. The occurrence of this risk was difficult to identify and evidence. For example, employees may not have been aware if they were being claimed for and may have been reluctant to engage with HMRC given they also have a relationship with their employer. Despite it being difficult to identify, HMRC has recovered overpayments of grants where it has been able to evidence that “furloughed” employees were actually working. HMRC’s compliance activity has demonstrated that many of the overclaims in these circumstances were made in error, as well as those driven by fraudulent activity and overclaims were small in value. The scheme became more flexible from July 2020 onwards to allow employees to work some of their working hours and be on furlough for the hours not worked, reducing the error and fraud risk from individuals working whilst being claimed for.

Employers claiming for employees who were working contributes the largest proportion of error and fraud across the whole lifecycle of the CJRS. However, new insight from the second CJRS Random Enquiry Programme shows that, in the final phase of the scheme, there is a lower proportion of error and fraud attributed to the risk of employers claiming for employees who are working than the total of other risks. This is likely because of changes to scheme design and eligibility rules, such as varying levels of claim amounts which allowed employers to use the scheme more flexibly. Whilst the introduction of these changes later in the scheme’s lifecycle reduced the prevalence of employers claiming for employees who were working, it did introduce more scope for error from other error and fraud risks due to the added complexity. After incorporating the results of the second CJRS Random Enquiry Programme, we now attribute a larger proportion of the overall CJRS error and fraud estimate to error than in our previous published estimate.

The CJRS was an unprecedented support scheme. Therefore, there are limited examples to compare against. However, comparing the error and fraud estimates to HMRC’s performance as measured by the tax gap can provide some context, though this is not a perfect comparison. This is because the tax gap is estimated taking account of the downstream impacts of compliance^{[footnote 3]}, whereas the COVID-19 schemes do not include those estimates. The latest published estimate for the tax gap, for the 2021 to 2022 tax year, was 4.8% of total theoretical tax liabilities of £739.3 billion, equivalent to £35.8 billion of losses. The rate of error and fraud in the CJRS is also similar to that of child and working tax credits, which has been around 5% in recent years.

3.3.2 HMRC compliance approach

From the beginning, it was expected that the CJRS would be a target for opportunistic fraud and that customers operating at pace and under pressure would make mistakes. Therefore, a range of measures to protect the scheme against organised crime, opportunistic fraud and customer error detailed below were designed into the CJRS and overall policy design, which drew on analysis and intelligence of the likely significant risks.

Promoting compliance

Compliance was promoted through educational material, behavioural nudges and good customer service, making it easier, and prompting customers to make a correct claim. This included:

• an honesty declaration in the online claims form
• an online calculator to allow claimants to calculate the size of their claim accurately
• communications on the risk of potential compliance actions
• requiring employers to write to employees to inform them they were on furlough

Preventing non-compliance

To prevent incorrect or fraudulent claims being accepted, a variety of compliance measures were built into the scheme design and the claims process itself:

• the eligibility criteria of the scheme were limited, so that claims were only accepted from employers with current records in HMRC’s PAYE system
• by building automated controls into the digital claim process, HMRC prevented more than 100,000 ineligible or mistaken claims ^{[footnote 4]}
• claims were limited to employees who were registered on the payroll by set dates (initially 19 March 2020)
• employers were required to retain the data that HMRC needed to do any later necessary checks e.g. employers were required to provide details of employees who had been on furlough and for how long
• a 72-hour risk assessment window for claims to be amended or rejected after submission

Pre-payment checks were developed, such as matching claims against lists of known and suspected organisations and devices. These checks helped to reduce the risk of organised criminal attacks. Up to the end of the scheme, over 22,000 suspicious CJRS claims were blocked from being paid valued at over £114.7 million.

Post-payment compliance

Post-payment compliance activity began once HMRC was granted powers to do so through the Finance Act in July 2020. At Budget 2021, the government announced a £100 million investment into the establishing the Taxpayer Protections Taskforce (TPT) to further expand the scope of COVID-19 compliance activity. Latest information on the compliance approach can be accessed through the 2022 to 2023 HMRC annual report and accounts.

An online repayment portal was also established to encourage claimants to make repayments, for example where they identified an overpayment of a CJRS grant or if they wanted to voluntarily pay the grant back, as it was no longer required. So far, over £1 billion has been received in CJRS unprompted repayments and disclosures.

Compliance activity is still ongoing across all COVID-19 schemes and up to 31 March 2023, HMRC has prevented the payment of or recovered the overpayment of over £1.4 billion worth of grants. HMRC has tackled a significant amount of CJRS risk. Up to the end of March 2023, the TPT has recovered a total of £256.1 million of overpaid CJRS grants. This is in addition to the £518.8 million recovered prior to the taskforce being established. The taskforce has already taken action on the riskiest claims, and with COVID-19 government support schemes closed, HMRC expect to see diminishing returns from the taskforce. Therefore, COVID-19 scheme compliance activity will now be undertaken alongside broader tax compliance activity. This is the most cost-effective way to ensure HMRC can protect and recover taxpayers’ money from those abusing the schemes.

Chapter 4: Impact of the CJRS on jobs and the wider labour market

The CJRS is estimated to have directly protected around 4 million jobs. This shows that the scheme was effective in meeting its key objective, to protect jobs during the pandemic. The benefit of protecting jobs and retaining employer-employee matches was that it allowed for a quicker and smoother recovery as restrictions were eased, as employers and employees were able to restart activity more promptly and efficiently than if employees had been made redundant. The scheme also played an important role in protecting businesses and household incomes thus supporting the wider economy.

Furthermore, a large majority of employees on furlough remained in employment after leaving the scheme; 83% of all individuals who had participated in the CJRS were employed in June 2022, of which 54% were still employed by the employer who placed them on furlough. There were only small differences in labour market outcomes across different characteristics.

The 2 main evolutions of the policy design, changes in employers’ contribution towards employment costs and the introduction of flexible furlough (allowing employees on furlough to work part-time), are shown to have encouraged employers to move individuals from furlough back into employment as COVID-19 restrictions were eased.

4.1 Introduction

This chapter addresses the following evaluation questions:

• to what extent did the scheme protect jobs?
• to what extent did employers and employees respond to scheme changes as intended?
• who was impacted by the scheme and what can be learned from their employment outcomes?
• to what extent did the winding down and closure of the scheme affect employment?

The CJRS is estimated to have directly protected 4 million jobs over its duration. It is technically very challenging to estimate this given that it means working out what would have happened without the scheme, and the options for constructing a counterfactual estimate are limited. The assessment is based on 2 estimates using different analytical approaches and reflects the uncertainty involved in estimating the scheme’s impact. The strengths and weakness of these approaches are covered in the accompanying documents to this CJRS final evaluation. Despite these uncertainties, the fact that 2 very different approaches achieve a similar outcome does underpin that this is a reasonable headline estimate of the impact of the CJRS on jobs.

The CJRS was a large government intervention and had wider effects beyond that of directly protecting jobs. The CJRS’s objectives included protecting jobs, reducing the risk of permanent employer closures, and benefiting the macroeconomy (as discussed in chapter 1.5.2). The estimates of jobs directly protected in this chapter recognise that: (i) not all jobs for which the CJRS was claimed for would have been lost in the absence of the scheme; and (ii) the CJRS could not protect all of the jobs claimed for. Eligibility for the scheme was deliberately broad, and so employers were able to place employees on furlough regardless of whether they expected to later make the employee redundant.

As discussed further in chapter 8.5.2, even if a job was not directly protected by the scheme, it is not necessarily considered as deadweight. The CJRS had secondary objectives of reducing the risk of permanent employer closure and benefiting the wider macroeconomy by reducing the risk of economic scarring. Without the scheme in place household incomes and consumption would have been significantly lower. Further positive impacts of the CJRS on employers and the wider economy, that are not captured in these estimates, are discussed in chapter 6 and chapter 7 respectively. Limitations associated with the counterfactual jobs directly protected estimate are presented in more detail within chapter 4.2.2.

The remainder of this chapter assesses the labour market outcomes of those on furlough and the associated impact of changes in the scheme’s policy design. This analysis shows that a large majority of individuals on furlough went back into employment after they had left the scheme, with most remaining with their original employer 9 months later.

Chapter 4.5 on the impact of the policy design on jobs covers changes to the employer contribution rate and the introduction of flexible furlough. These changes were introduced to encourage employers to bring staff back to work. The extent to which these 2 changes achieved that is explored.

4.2 Jobs protected estimate: matched counterfactual analysis of employment outcomes

4.2.1 Method

The first approach to estimating the number of jobs directly protected by the CJRS, compares employment outcomes for an experimental control group with an intention-to-treat group of individuals defined by their eligibility for the CJRS, using HMRC’s Real Time Information (RTI) of administrative payroll data.

The analysis looks at the outcomes for individuals starting new employments at 2 different points in time across the scheme. This analysis covers the impact of the scheme between November 2020 and April 2021. The time-periods used correspond to announced extensions and changes in eligibility for the CJRS, with the cut-off points from the scheme extensions used to determine eligible and ineligible groups for the purposes of the analysis.

Individuals in the eligible (intention-to-treat) and ineligible (control) groups are matched to account for observable characteristics that might impact employment outcomes. The difference in the employment outcomes between these 2 groups is used to estimate the impact of the scheme. These estimates are then scaled up to the total furlough population. The underlying methodology is consistent with the approach used in the CJRS interim evaluation, that being a linear probability model (LPM), with the addition of an estimate of the cumulative impact across the scheme. This is calculated by adding the jobs directly protected estimates from the March 2020 to October 2020 analysis to the November 2020 to April 2021 analysis to produce an estimate of the total number of unique jobs directly protected. This helps mitigate against double-counting jobs protected at different stages of the scheme, as discussed below in more detail.

An estimate from the final period of the scheme, May 2021 to September 2021, is not robust enough to be included in this cumulative impact estimate. There were several potential reasons for this, including low sample sizes as fewer individuals took-up the scheme and potentially changes in the behaviour of employers, both users and non-users of the scheme as the scheme wound down.

Full methodological details can be found in the CJRS final evaluation matched counterfactual analysis technical note.

4.2.2 Findings

This first method finds that the cumulative impact of the CJRS was approximately 3.7 million jobs directly protected up to April 2021. Figure 4.1 below is taken from the interim evaluation and illustrates the impact of eligibility after the first cut-off point in the analysis, March 2020, the beginning of the scheme. The difference between employment rates of the control and intention-to-treat groups is observed up to October 2020, peaking during May 2020 at 2.4 percentage points.

Figure 4.1: Employment rates for intention-to-treat and control groups from August 2019 to October 2020

Source: HMRC CJRS and PAYE RTI data

Table 4.1: Employment rates for intention-to-treat and control groups from August 2019 to October 2020

Month	Control	Intention-to-treat
August 2019	71	70
September 2019	71	71
October 2019	71	70
November 2019	71	71
December 2019	69	70
January 2020	66	66
February 2020	70	70
March 2020	99	99
April 2020	87	89
May 2020	83	86
June 2020	82	85
July 2020	82	84
August 2020	82	83
September 2020	81	81
October 2020	80	80

It is important to note that not all individuals in the intention-to-treat group were on furlough, likewise some individuals in the control group were on furlough. Approximately 18% of those in the control group in figure 4.1 were on furlough, whilst 25% of the intention-to-treat group were on furlough. This figure remains similar in the extension of the analysis illustrated in figure 4.2, where approximately 19% of the control group and 29% of the intention-to-treat group were on furlough. This proportion of individuals on the scheme, across the 2 groups, is used when scaling estimates to the total furlough population. More details of this scaling process can be found in the CJRS final evaluation matched counterfactual analysis technical note.

The scheme was initially set to close in October 2020, however it was extended to reflect the changes to NPIs. Figure 4.2 reflects the impact of eligibility between October 2020 and April 2021, not the impact of receiving the CJRS grant via the employer. As shown, the employment rates before October and November 2020 are very similar between the 2 groups, leading to the assumption that differences after the October and November 2020 cut-off can be attributed to the scheme^{[footnote 5]}. The peak difference in this period of 1.5 percentage points occurs in January 2021, after which the rates converge, suggesting a reduced impact of the scheme in protecting additional jobs in the latter months of the scheme. However, as evidenced below in chapter 6 and chapter 7, there are also wider impacts of the scheme during this period such as providing employers with certainty and supporting the economic recovery not captured within the jobs directly protected estimates.

In both figure 4.1 and figure 4.2, the employment rates between the 2 groups converge. It is difficult to draw robust conclusions as to why with the available evidence. At the beginning of the COVID-19 pandemic, the difference between the employment rates would be at its peak, given that was the period of most uncertainty and therefore a stronger likelihood of individuals losing their jobs. Therefore, this convergence could be driven by several factors, including improvements in the broader labour market and economy as restrictions eased, resulting in less unemployment in the control group, however it is not possible to determine which is the driving factor.

To estimate the impact of receiving the CJRS grant on employment rates, the figures are scaled up to the broader furlough population. This analysis shows that receiving the CJRS increased the likelihood of an employee being in any PAYE employment by approximately 16 percentage points in January 2021.

Figure 4.2: Employment rates for intention-to-treat and control groups from April 2020 to April 2021

Source: HMRC CJRS and PAYE RTI data

Table 4.2: Employment rates for intention-to-treat and control groups from April 2020 to April 2021

Month	Control	Intention-to-treat
April 2020	66	66
May 2020	65	65
June 2020	64	64
July 2020	64	64
August 2020	61	61
September 2020	66	66
October 2020 and November 2020	100	100
December 2020	97	98
January 2021	93	94
February 2021	90	90
March 2021	89	88
April 2021	83	82

Calculating the likelihood of being in PAYE employment for each month between March 2020 and April 2021 can be used to show the peak jobs protected estimate across each furlough period, as shown in figure 4.3^{[footnote 6]}. For the first period between March to October 2020, an estimate of 3.4 million jobs were directly protected, whilst between November 2020 to April 2021, an estimated 300,000 additional jobs were directly protected. These figures are derived by taking the month with the peak jobs protected figure within each period, which is May 2020 for the period between March and October 2020, and January 2021 for the period between November 2020 and April 2021.

Adding the 2 peak estimates together, illustrated in figure 4.3, produces the cumulative impact of the scheme on jobs directly protected, equal to 3.7 million. It should be noted that this figure is calculated using unique furlough employments, only within the 2 periods assessed, for example, the October / November 2020 period impact calculation is limited to using the 1.9 million new furlough employments over that period. Whilst this approach mitigates any potential issue in double counting of results presented in the CJRS interim evaluation, it likely underestimates the peak estimates because it will not capture all the jobs directly protected by the scheme. More details of this approach and the relevant calculation can be found in the CJRS final evaluation matched counterfactual analysis technical note.

This counterfactual analysis uses a LPM, which is commonly used in difference-in-differences (DiD) analysis with panel data and are generally considered robust. However, alternative methodological approaches are also available and can sometimes mitigate a bias in the estimator when estimating a binary outcome.

One of these alternatives, a logistic regression methodology, was trialled for the November 2020 to April 2021 period of the scheme. The estimates show that the scheme increased the odds of those in the intention-to-treat group being in employment relative to those in the control group. In January 2021, individuals in the intention-to-treat group had approximately 2.8 times higher odds of being in employment, relative to those in the control group. Therefore, like the LPM, the logistic regression approach suggests that the CJRS had a positive impact on employment outcomes.

However, this result highlights a notable challenge of interpretating the estimates from a logistic regression approach. The difference-in-differences coefficient is expressed as an odds ratio, which cannot easily be converted into a number of jobs protected estimate, making it difficult to compare to the jobs protected estimate from the LPM. Consequently, caution should be taken when comparing impacts across the different modelling approaches.

Even so, in terms of orders of magnitude, the logistic regression illustrates that the scheme had a positive impact on employment, which is consistent with the LPM jobs protected estimate. When assessing the impact across the period, the logistic regression shows a strong impact during December 2020 and January 2021, before a gradual month-by-month reduction through to April 2021, broadly similar to the estimate derived from the LPM.

Given the complexity of the logistic regression and difficulty in conveying both the methodology and estimated impact on the scheme in terms of jobs protected, this evaluation uses the LPM approach for the counterfactual estimate of jobs directly protected. Full details of these approaches and a discussion on the relative strengths and limitations of each methodology can be found in the matched counterfactual analysis technical note.

As noted previously, estimates of additional jobs directly protected from the final period of the scheme, May 2021 to September 2021, are not included. However, this final period would have only had a small impact on the estimate of cumulative jobs directly protected. For example, if the 16 percentage point peak jobs directly protected estimate from January 2021 were applied to unique employments on furlough in this period, it would lead to 32,000 more jobs protected. Evidence from the employer qualitative research presented in chapter 6, found that the easing of restrictions and closure of the CJRS prompted organisations to take steps to return employees on furlough to work.

Figure 4.3: Estimated peak additional jobs directly protected across the duration of the CJRS

Source: HMRC CJRS and PAYE RTI data

Table 4.3: Estimated peak additional jobs directly protected across the duration of the CJRS

CJRS Period	Number of Jobs
Mar 2020 to Oct 2020	3,400,000
Nov 2020 to Apr 2021	300,000
Total additional jobs directly protected	3,700,000

The fall in the number of additional jobs directly protected across the duration of the scheme indicates a declining impact of the CJRS from the peak in May 2020. This broadly reflects the reduced use of the scheme over its duration and is consistent with improvements in the underlying economy and relaxation in COVID-19 restrictions. It also reflects the methodology, which only captures a job protected in one period and individuals newly eligible for furlough in that period to avoid double-counting, meaning our estimates will fall as take-up falls. In reality, jobs may have been protected at multiple points during the scheme, for example during both the spring 2020 and winter lockdowns, but in this approach the job would only show as being protected in spring 2020.

Alongside this, there are several additional important impacts that are not captured in these estimates that should be considered. The cumulative jobs directly protected estimate does not account for jobs protected by preventing employers from ceasing trading, wider impacts of the scheme on other jobs or improvements in the wider labour market and economy which may have improved job prospects for the unemployed. Further limitations are discussed in the CJRS final evaluation matched counterfactual analysis technical note.

4.3 Jobs protected estimate based on research findings

4.3.1 Method

The second approach to estimating the number of jobs directly protected by the CJRS is based on self-reported research data.

The research asked employers if they would have made more redundancies if the scheme was not available and, if so, how many more. Employers were also asked if they thought they would have closed permanently without the CJRS. The analysis combines the findings from these 2 questions to identify the number of jobs protected for participating employers and then scales this up to the total population, using estimates of the total population of eligible employees.

The estimate of the number of jobs protected by the CJRS has been scaled to show the jobs protected by sector. This is expressed as a proportion of total eligible jobs for each sector. The methodology mirrors that used in the interim evaluation and is applied across the lifespan of the scheme.

Due to the self-reported nature of the questions asked to employers about the expected impact of the pandemic on their business in the absence of the CJRS, some caution should be applied to these findings.

4.3.2 Findings

The quantitative research suggests that 4.4 million jobs were directly protected by the CJRS across its lifespan. A confidence interval at the 95% level for the estimate ranges from 3.9 million to 4.8 million jobs.

Additional evidence from the qualitative research found that employers interviewed unanimously perceived the CJRS to have played a key role in keeping their business afloat during the pandemic by increasing their resilience. They reported using the CJRS to avoid making significant redundancies and retain staff with important knowledge and technical skills. Employers reported that the CJRS afforded them more time to make crucial business decisions, including balancing organisational costs with staffing costs and redundancies. The CJRS was found to have eased the weight of these decisions for employers, enabling them to delay them or remove the need to make them entirely.

Employees themselves also recognised the importance of the CJRS in protecting jobs, as found in the qualitative research with employees. They felt that they were much more likely to have been made redundant without the CJRS in place. With both employers and employees recognising the importance of the CJRS in protecting jobs, this provides further evidence of the CJRS achieving its primary policy objective.

Secondary analysis of the quantitative research findings showed that jobs were protected in all sectors, with arts, entertainment and recreation (54%), accommodation and food service activities (37%) and other service activities (39%) sectors having the highest proportions, as shown in table 4.4. These were the sectors that had the largest negative impacts of COVID-19, as shown in figure 2.2, and highest employer take-up rate, as shown in figure 2.3. In some limited instances, employers noted that they made redundancies prior to the announcement of the CJRS in anticipation of a drop in demand due to the pandemic. Some of these were newer employees, who were not on payroll at the cut-off point and hence would not have been eligible for the scheme once it opened.

Table 4.4: Proportion of jobs protected by the CJRS, by average across all and top 3 sectors

Sector	Proportion of jobs protected by the CJRS
Average across all sectors	21%
Arts, entertainment and recreation	54%
Other service activities	39%
Accommodation and food service activities	37%

Source: CJRS employer and agent quantitative research wave 2

Base: Sectors (all > 200)

Question: Which of the following best describes what would have happened to your organisation if you had not received funding from the CJRS?; To the best of your knowledge, would your organisation have made more employees redundant if the CJRS was not available, or would it have made no difference … How many more employees would your organisation have made redundant if the CJRS was not available?

4.4 Labour market outcomes and employee characteristics impacts

4.4.1 Method

The labour market outcomes of individuals on furlough are explored over the short-to-medium term after the closure of the scheme. It also explores if there was a relationship between individual characteristics and outcomes.

The analysis looks at employees who were on furlough at any point during the scheme’s lifespan (10.6 million individuals^{[footnote 7]}, from March 2020 to September 2021, and examines their employment outcomes at 2 points in time (December 2021 and June 2022). These time points were chosen so that the short- and medium-term impacts of the scheme can be assessed, whilst also accounting for natural labour market and economic changes.

Analysis of Universal Credit administrative data, for those previously on furlough was also conducted by DWP to assess the outcomes of those moving out of pay-rolled employments onto welfare support^{[footnote 8]}.

A similar approach was used to assess outcomes of employees who were on furlough in its final month (September 2021). Analysis was produced on how many were still in employment in June 2022, 9 months after the CJRS ended. These employees are compared to others that were never on furlough but were in employment in September 2021.

To understand if the scheme’s impact varied by characteristic, a logistic regression was used to calculate the probability of an employee being in employment 9 months after leaving CJRS. A range of characteristics, including gender, age and sector were controlled for.

4.4.2 Findings

The analysis suggests the scheme preserved a majority of employer-employee matches and protected a high proportion of jobs over the short-to-medium term.

Figure 4.4 shows that the majority of employees on furlough were in the same employments they were on furlough from, 3 and 9 months after the end of the scheme (December 2021 and June 2022). Across this period, there was a reduction in those still in one of their furlough employments from 61% (6.5 million) to 54% (5.7 million). However, the rate of those in other, non-furlough employments increased by 5 percentage points (from 2.6 million to 3 million). Those in other, non-furlough employments include new employments as well as those that started during or before the CJRS. Combining these 2 groups shows that 83% of employees on furlough were in any employment in June 2022. Employees not on payroll as a share of all employees previously on furlough increased by 2 percentage points in June 2022, when compared to December 2021 (from 1.6 million to 1.8 million).

To contextualise these findings, employees who were not placed on the CJRS saw similar trends. The number of all non-CJRS employees who were on payroll (PAYE-employment), fell from 83% (19.6 million) to 81% (19.1 million) from December 2021 to June 2022^{[footnote 9]}.

Figure 4.4: Employees who were on furlough, by their employment status at December 2021 and at June 2022

Source: HMRC RTI and CJRS Data

Base: All employees on furlough (10,568,650)

Table 4.5: Employees who were on furlough, by their employment status at December 2021 and at June 2022

Employment status	Jun 2022	Dec 2021
Total number on payroll for one of their furlough employments	5,665,600	6,449,000
Total number on payroll from any non-furlough employment	3,094,850	2,550,350
Total number not on payroll	1,808,200	1,569,300

Figure 4.5 displays employments previously on furlough and not on payroll in December 2021 and if they were on Universal Credit (UC) between November 2021 and January 2022^{[footnote 10]}.

Of the 1.6 million individuals who were no longer in PAYE employment and had previously been on furlough, only 8% (164,000) were on UC, not in work and in the ‘Searching for Work’ conditionality group meaning they are able and required to actively seek work unless under a temporary exemption. Whilst not all unemployed individuals are on UC or in this conditionality group, DWP’s analysis suggests that a significant number of those that were not on payroll after leaving furlough were not looking for employment and therefore not classed as unemployed.

Figure 4.5: Breakdown of the 1.6 million employments not on payroll in December 2021

Source: Labour market outcomes and DWP Universal Credit analysis Base: Employment on furlough that were not on payroll in December 2021 (1,569,000)

Table 4.6: Breakdown of the 1.6 million employments not on payroll in December 2021

Individuals previously on furlough ‘not in employment’ in December 2021:	Percentage
Universal Credit, Self-Employed or earnings off RTI	3%
Universal Credit, not working and not required to intensively seek work	7%
Universal Credit searching for work	8%
Not on Universal Credit	82%
Total furlough population ‘not in employment’ in Dec 2021	100%

There are many other reasons as to why employees on furlough may not be on payroll outside of unemployment, such as moving to self-employment, leaving the country or retiring. Analysis presented in chapter 7, which outlines the macroeconomic impacts of the scheme, shows that there were only small differences in the rate of individuals leaving the labour market when comparing the populations of those on and not on furlough. Overall, the evidence indicates that the scheme was effective at protecting employees’ jobs and keeping them in employment.

To better understand who was impacted by the scheme and if these impacts were felt evenly, the post-CJRS labour market outcomes were assessed by individual characteristics. This approach can provide insights into whether any groups in society had different employment outcomes.

Post-CJRS labour market outcomes differed by age, with younger individuals (up to age 34) more likely to remain in employment than older individuals, particularly after the age of 61. This may reflect individuals moving into retirement, alongside other barriers older individuals face in the labour market. This evidence, with a focus on age, is explored further in chapter 7.3.6.

Sector analysis shows that individuals in accommodation and food services (69%), arts, recreation, and entertainment (71%) and administrative and support services (71%) were less likely to return to employment post-CJRS, compared to employees in other sectors (an average of 77%). One reason could be that these sectors were heavily impacted during the pandemic, resulting in these jobs being more at risk. The analysis also shows that post-scheme outcomes were broadly similar across genders (men 71% and women 74%) and regions (average is 80% with a range of 78% to 81%), except London, which had a notably lower likelihood of employment (71%).

Although outcomes were positive for employees on furlough remaining in their jobs, there is also evidence that those previously on furlough were more likely to be out of employment after 9 months.

In September 2021 (the end of the scheme), the furlough population consisted of around 1.2 million employees. The non-furlough population was 18.6 million employees that were never on furlough and were in paid employment in September 2021^{[footnote 11]}.

The analysis shows that by June 2022, 14.0% of employees previously on furlough in September 2021 (165,000) were no longer paid by any employment, whereas for the non-furlough population, it was 6.9%. This gives an ‘excess’ separation rate, the difference between the furlough and non-furlough population, of 7.1% or 83,000 employees relative to what would have happened if the rates had been identical across the 2 groups.

Despite this, it is worth noting that the scheme’s closure resulted in very little increase in unemployment. Analysis of the closure of the scheme is outlined in more detail within chapter 7. Employers felt that the closure of the CJRS prompted them to prepare for a return to stability, which included decisions about staff who were still on furlough. Employers made decisions on staff retention based on their organisational health at the time. Staffing changes occurred where the pandemic had resulted in changes in business model and client demand after the pandemic, which varied by sector.

The labour market outcomes of those on furlough for the whole duration of the scheme, referred to as long-term furlough, were also analysed. RTI data for the population on long-term furlough found that 80% of those on full furlough for the duration of the scheme, were in PAYE employment in June 2022. This was slightly lower than the overall furlough population at 83%, as shown above. However, for those on a mix of full and flexible furlough for the duration of CJRS, 90% were in PAYE employment in June 2022.

One concern was that these individuals were on the CJRS before retiring. Limiting the analysis to those aged 55 and over on 1 March 2020, equivalent to approximately 25% of the individuals in the respective long-term furlough populations, finds that 8% of the full furlough population and 5% of the full and flexible furlough population were paid by a pension in June 2022. These populations saw 75% and 86% still paid by any PAYE employment in June 2022 respectively. This suggests that the majority of this long-term furlough population had positive labour market outcomes, leading to the conclusion that there is little evidence of negative consequences from providing furlough support to this population.

4.5 Impact of policy design on jobs

As shown in chapter 1.5 and figure 1.2, there were several changes to the design of the CJRS throughout its lifespan, including changes to the employer contribution rate and the introduction of flexible furlough. Prior to any changes, the government had covered both NICs and pensions contributions and 80% of employees’ wages for hours not worked.

A significant change to the CJRS was the introduction of flexible furlough in July 2020, which gave employers the chance to bring employees back to work for some hours, whilst still being on furlough for their remaining hours. Further changes were made through the introduction of employer contributions. From August 2020, employers could no longer claim the cost of employer NICs and pensions contributions (up to the level of the automatic-enrolment minimum) for hours not worked. From September 2020, employers were also required to contribute 10% of usual pay in respect of hours not worked, followed by 20% in October 2020, with the government contribution under the CJRS reducing to 70% and 60% respectively to maintain the wages of employees on furlough at a minimum of 80%. The government contribution rates also changed towards the end of the scheme in summer 2021, as set out in chapter 1.5.

These changes were designed to encourage employers to consider the long-term viability of jobs and where possible, move individuals from furlough back into work. Throughout the duration of the scheme, individuals on furlough were able to have a secondary employment whilst receiving the CJRS. The impact of those policy designs are explored by assessing if employers responded to them as intended and what this meant for employments on furlough.

4.5.1 Method: Employer contributions

The generosity of the scheme was reduced by the government as NPIs were eased and economic conditions improved. CJRS payments were tapered and instead a greater contribution was placed on employers to meet the minimum 80% wage support for employees. This maintained the same level of support for employees.

Econometric analysis, using linear and logistic regression, was used to assess the impacts of changes in the contribution rates. This analysed whether these changes led to more individuals being removed from furlough and if there were differences across characteristics.

4.5.2 Findings: Employer contributions

Results from analysis assessing changes in the employer contribution rates show their introduction accelerated the reduction of employments on furlough. As analysis in chapter 4.4 showed, a majority of employees remained in employment once leaving furlough, with most retained by the employer that placed them on furlough. The policy change encouraged employers to move individuals back into employments, either with their original furlough employer or a new employment.

Quantitative research with employers found that of those Users who last applied for the CJRS before 1 July 2021, 89% said they were aware of the changes to employer contributions that were made on that date. Of these, 40% said they made changes of some kind, most commonly bringing staff on furlough back to work (24%), re-opening their organisation after a period of closure (11%) or deciding to stop using the CJRS (8%). A small proportion (4%) said they closed their organisation, either temporarily or permanently, and 2% said they had made some or all of their staff on furlough redundant.

Qualitative research evidence found that increases to employer contributions were not reported as having a key role in determining use of the CJRS. However, it did emerge as part of the decision-making process for employers. This occurred when looking at balancing the increased cost of contributions and their ability to bring employees back to work. The role that increases to employer contributions played was often linked to the organisation’s financial health at the time.

The analysis found that for every one percentage point increase in the wage contribution rate, it is estimated to have had a 3% decrease in the number of employments on furlough. Whereas the reintroduction of employers having to pay National Insurance contributions (NICs) for employees on furlough is estimated to have caused a 1% reduction. However, the rate of NICs varies; some wages will be covered by the NICs secondary threshold or employment allowance^{[footnote 12]}. This may reduce the effect of this change on employments on furlough, particularly across lower earners. 

The analysis also showed that the likelihood of being removed from an employment on furlough due to changes in the employer contribution rate is similar across characteristics^{[footnote 13]}. However, there were differences in income, with employees paid less than £5,000 a year 14 percentage points more likely to be removed from furlough due to changes in the employer contribution rate, compared to those paid £80,000 or more. This suggests that lower paid employments may have been more impacted by the introduction of employer contributions. These results should be treated with caution, given the limitations of the robustness in the underlying model. More details can be found in the CJRS final evaluation accompanying technical information document.

A further uncertainty within the analysis is that it does not specify if those removed from furlough due to the introduction of employer contribution rates went into employment or non-employment.

Figure 4.6: UK redundancy rate from January 2007 to January 2023

Source: ONS redundancies time series

Note: ONS state that these are individuals who have been made redundant or have taken voluntary redundancy.

Figure 4.6 may provide an indication of this, as it shows an increase in redundancies from August to October 2020 which was larger than the increase in redundancies seen in the previous recession. As this increase coincides with the introduction of employer contributions, it may suggest that increased employer contributions are linked to redundancies during this period. There may be a reporting lag in the redundancy data shown in figure 4.6, due to the need for a consultancy period. This may suggest that the redundancy decisions occurred earlier than the introduction of employer contributions. Although there was also a slight increase in redundancies in summer 2021 when employer contributions were reintroduced, this is also subject to the same caveats.

Redundancies being larger than the previous recession may also be as a result of the broader nature, and impact, of the pandemic rather than solely due to the increase in employer contributions. Therefore, this increase in redundancies is likely to be as a result of a combination of factors, including the introduction of employer contributions, as employers began to adjust to the initial shock caused by the COVID-19 pandemic.

However, evidence outlined in chapter 4.4.2 showed that the vast majority (83%) of employees who left furlough were still in a form of employment after leaving the scheme. Therefore, this analysis suggests that introducing employer contribution rates may have had an impact on employers considering whether to continue placing their employees on furlough. This is supported by evidence from the qualitative interviews with employers, where some report that the increased employer contributions caused them to consider the balance between costs and bringing staff back to work.

4.5.3 Method: Flexible furlough

Flexible furlough was introduced in July 2020 to give employers flexibility around the number of hours that employees could be on furlough, no longer requiring them to either be completely on (not working) or off furlough (working).

The impact this policy change had on employers is covered in detail in chapter 6.3.3. Additional analysis was conducted to assess the specific impact of the policy change on jobs and how the positive impacts felt by employers would also benefit employees and the wider labour market.

To provide a full picture of how effective flexible furlough was in encouraging movements back into work and protecting viable jobs, employments that were on furlough at any point in May 2020 were analysed (approximately 9 million employments). These employments were tracked each month to October 2020 showing if they were in full furlough, flexible furlough, paid employment or not paid employment. This time period was chosen to focus on the introduction of flexible furlough.

4.5.4 Findings: Flexible furlough

Across the period of analysis, May to October 2020, an estimated 22% of the employments on furlough used flexible furlough. When looking at the pattern of use and information from the survey of employers, which is covered in more detail in chapter 6.3.3, employers used flexible furlough as a way of either easing individuals back into employment or managing their labour when dealing with fluctuations in demand for their goods and services. The analysis showed that only a small proportion of those placed on flexible furlough moved into not being paid by their employer over the period. The aim of this policy was to encourage employers to return individuals back into part-time work where viable and this analysis suggests that employers responded to the change as intended, producing further knock-on positive effects and ensuring economic activity could continue, which is detailed in chapter 7.

For these employments, it was most common to be on flexible furlough for each of the 4 months from July to October 2020, as opposed to being on full furlough, in employment or not being paid in an employment. This suggests that flexible furlough allowed employers to better manage workloads as restrictions eased, reflecting changes in broader economic activity and demand. Employers also used flexible furlough to ease employees back into work after being on furlough for several months. Only 6% of those placed on flexible furlough (1.3% of all employments on furlough) were not paid by their original furlough employment by October 2020.

Flexible furlough had a positive impact on employers, further explored in chapter 6. However, there are additional effects for individuals and their jobs. As evidenced in the employee qualitative research, some employees found that being placed on flexible furlough blurred the lines between work and personal time, as they could have been called in to work at any time. This made flexible furlough more challenging for some employees to manage than being on furlough. Overall, the evidence indicates that employers used flexible furlough to return employees to work, which could have led to positive effects on skills and development, alongside stimulating economic recovery. This is explored in more detail in chapter 7, where the scheme’s impact on the wider economy is outlined.

4.5.5 Method: Second jobs

Employees were permitted to find, or work in existing, alternative employment while on furlough and benefiting from the CJRS, unless their employment contract prevented this. Employee qualitative research explored the uptake of second jobs amongst employees on furlough and reasons for doing so. Analysis using CJRS and PAYE RTI data allowed additional employments to be identified for those on furlough.

4.5.6 Findings: Second jobs

Qualitative research with employees found that the motivations for taking on additional employment were varied. Reasons included topping-up their reduced wages, through a concern that their main employment was insecure or to keep themselves occupied. In some cases, individuals started their own business. When the CJRS closed, these employees tended to either keep both jobs or opt for the one with better pay.

Of those who did not take up additional employment, reasons included a lack of awareness that they could do so or an explicit belief that it was not allowed. Additional reasons were a lack of time, need or a lack of capability due to shielding or looking after children.

Using CJRS and PAYE RTI data to identify second jobs, over the lifetime of the scheme 1,197,800 individuals were on furlough through the CJRS and also paid in at least one other employment that they were not on furlough from in the same month. For 61% of these individuals, the additional employment started before the CJRS was introduced in March 2020, which suggests that these were existing arrangements. When looking at the pay of the job on furlough and additional employments, 45% of individuals were paid less than £15,000 in each, which may suggest that they have multiple part-time jobs.

Chapter 5: The role of the CJRS in supporting incomes

The CJRS aimed to ensure employees could retain the majority of their usual pay, thus supporting household incomes and living standards during COVID-19 and therefore consumption and the macroeconomy. Overall, the scheme provided broad support across the working-age income distribution. Middle income households benefitted the most as a percentage of pre-pandemic income.

5.1 Introduction

This chapter provides evidence of the scheme’s impact on households across the income distribution, including the extent to which employer’s topped up their employees’ wages, and addresses the evaluation question:

• to what extent did the scheme support household incomes and protect living standards?

5.2 Methods

This chapter estimates the impact of CJRS payments on the incomes of working-age households (such as where at least one person in the household is aged 16-64 and working) as a percentage of their pre-pandemic income. The CJRS was designed to support existing employees, therefore households which did not contain any individuals in employment did not benefit from the scheme. The chapter extends the analysis in the CJRS interim evaluation to later stages of the scheme, for 3 snapshot months in May 2020, September 2020, and January 2021.

The analysis presented here was carried out using HM Treasury’s tax and welfare microsimulation model. In each snapshot month, conditional probabilities are used to assign whether an individual is on furlough, flexible furlough, or loses their job. Probabilities are based on the UK Household Longitudinal Study and HMRC CJRS data, with further detail found in the CJRS final evaluation accompanying technical information document. Households which are randomly modelled to have members who become unemployed are then assumed to receive the benefit income they would have been entitled to, including the temporary £20 per week uplift to Universal Credit that was in place at the time. The outcome of this random modelling is then compared to the counterfactual scenario where, instead of being on furlough, these workers are unemployed.

The results are presented as the average impact of CJRS payments on net household income. This means that the tax that some households might pay on any CJRS income has been netted off, as well as any reduction in welfare entitlement. The resulting net change in income is then expressed as a percentage of net pre-pandemic household income. Income deciles have been calculated on equivalised net household income (before housing costs, and on a pre-pandemic basis), in order to control for changes in household size, and in turn broken down into deciles for all working-age households.

5.3 Findings

The analysis in figure 5.1 shows the impact of the CJRS in supporting pre-pandemic working-age household income. Compared to the scenario without the CJRS, in May 2020 support from the scheme was worth 10% of the pre-pandemic average net household income of all working-age households. The total support toward working-age household income by the CJRS reduced in September 2020 from the May 2020 peak but then increased slightly in January 2021. The greater total support towards incomes in January 2021 reflects the extension of the CJRS in November 2020 as NPIs were implemented and take-up of the CJRS increased.

Figure 5.1: Average net income support provided by the CJRS as a share of pre-pandemic net household income, all working-age households, by net equivalised working-age household income decile

Source: HM Treasury tax and welfare microsimulation model

Note: The findings are compared to the counterfactual scenario where, instead of being on furlough, these workers are unemployed.

Table 5.1: Average net income support provided by the CJRS as a share of pre-pandemic net household income, all working-age households, by net equivalised working-age household income decile

Deciles	May-20	Sep-20	Jan-21
Bottom Decile	3.9%	0.8%	1.9%
2nd	6.9%	1.3%	2.8%
3rd	9.3%	1.9%	3.8%
4th	10.7%	2.4%	4.7%
5th	13.2%	2.6%	5.4%
6th	13.5%	2.8%	5.4%
7th	13.9%	2.5%	5.0%
8th	12.6%	2.4%	4.4%
9th	10.6%	1.9%	3.8%
Top Decile	5.3%	1.1%	2.0%
All Households	10.1%	2.0%	3.8%

On average, the scheme offered the greatest income support (as a percentage of pre-pandemic net household income) to those working-age households in the middle-income deciles, although it provided support broadly across the distribution. In this scenario without the CJRS, as a percentage of pre-pandemic income and for each of the months modelled, the CJRS had the smallest impact on the incomes of working-age households at the very bottom and top of the income distribution.

For the lower income deciles, this is influenced by the fact that lower income households are more likely to contain either single earners, or no individuals in employment and therefore not benefitting from the CJRS. Workers in lower income households are also more likely to be self-employed, and so would not qualify for CJRS eligibility. A larger proportion of their household income is made up from other sources (such as Universal Credit) compared to households in higher income deciles. It also reflects how income received from CJRS payments will have affected benefit entitlement, for some households.

For the top income deciles, the lower support is driven by the scheme’s monthly cap (initially £2,500 and lower in later months, although, employers were required to compensate the reduction), which limited the extent to which CJRS payments compensated higher earners. Those with higher incomes would also pay a larger proportion of tax on their CJRS payments.

The household analysis above excludes the impact of any (voluntary) employer top-ups and (mandatory) employer contributions to wages for hours not worked as these were not funded by the government. Only CJRS payments themselves are considered. These top-ups and employer contributions would have provided additional support to household incomes beyond the government support analysed in figure 5.1.

Using HMRC PAYE data, the monthly earnings of employees on furlough have been compared to an estimate of their pre-furlough pay to determine whether they were receiving a pay top-up from their employer to support their incomes. This is done by comparing an employment’s pre-furlough monthly pay, which we define as the mean pay between May 2019 and January 2020, to actual furlough pay between March 2020 and September 2021. To ensure an accurate comparison, all employments with irregular or variable pay are removed. The analysis considers the impact at an individual employment level, and not at a household level.

Figure 5.2: Pay for employments on furlough compared to pre-furlough pay

Source: HMRC RTI and CJRS data

Note: This analysis only includes employments with regular pay.

Table 5.2: Pay for employments on furlough compared to pre-furlough pay

Month	Less than pre-furlough pay	Equal to pre-furlough pay	Greater than pre-furlough pay
April 2020	48%	41%	10%
July 2020	46%	42%	12%
November 2020	37%	45%	17%
January 2021	37%	45%	18%
May 2021	32%	46%	21%
September 2021	28%	48%	23%

Figure 5.2 shows that, overall, roughly half (51%) of employees on furlough earned at least 100% of pre-furlough pay in April 2020. The other 48% of individuals will have received somewhere between 80% and 95% of pre-furlough pay in April 2020, so were either not topped-up or had only a proportion of their original pay topped-up.

The percentage of employments receiving equal to pre-furlough pay is slightly higher for lower income individuals, especially for employments earning less than £15,000. Employments earning above £80,000 were slightly less likely to receive payments equal to pre-furlough pay. Despite this, they were more likely to receive a pay top-up in some form. This could be due to the payment cap as employers tried to protect their employees from drastic falls in income.

The proportion of employments on furlough receiving top-up payments gradually increased over time. By the end of the scheme, in September 2021, 48% of employments received pay equal to their pre-furlough pay. The reasons why some employments were receiving greater than their pre-furlough pay is not ascertained by this analysis, however it could be due to the sample size and types of individuals included in the analysis, which is detailed below. As noted throughout this evaluation, steps were taken to reduce the generosity of the CJRS over time through tapering the level of government contributions and increasing the amount employers had to contribute.

While using the total furlough employment population of 11.7 million was considered for this analysis, it is not possible to calculate reliable monthly pay for employments with irregular or variable income. Therefore, this analysis has a sample size of 2.3 million as employments (19% of the total number of employments placed on furlough at any point).

There are demographic differences between the sample for this analysis and the overall furlough population. While there is minimal difference between men and women, there is a smaller proportion of individuals who are under 25 years old in the sample (8% compared to 19% of the overall furlough population). This could be because a larger proportion of younger workers have irregular and variable pay and so were removed. As a result, the sample contains a larger proportion of older employees, especially for individuals aged above 35.

By sector, there is only one notable difference across the sample and total furlough population, with that being the accommodation and food services sector. This sector accounts for only 8% of the sample size, but 18% of the entire furlough population. Similar to the age breakdown, this may be due to individuals in this sector receiving variable or irregular payments. Finally, there is a larger proportion of individuals from smaller employers (PAYE scheme sizes of 1 to 4 employments) in the sample compared to the overall furlough population (32% compared to 16%).

Qualitative research suggests that employers felt that the contribution levels were suitable for dealing with the impacts of the COVID-19 pandemic. However, many employers reported affordability was their main obstacle to topping-up employees’ wages and that employers felt it did not feel fair to top-up pay for those on furlough, while their colleagues continued working. However, a higher prevalence of pay top-up towards the end of the CJRS was an indication of an improved business outlook as restrictions were eased and consumer spending increased. This coincided with increased employer contributions being reintroduced towards the end of the scheme and so may suggest that it was the right decision to increase employer contributions again, which employers reported that they felt was a reasonable development.

Additionally, employers topped-up wages, where it was affordable, for both fairness and strategic reasons. This was to support their employees out of a sense of responsibility or through demonstrating their value to the organisation. Some employers decided not to top-up due to a sense of fairness towards those in their organisation who were still working. Qualitative research with employees found that for those who were placed on furlough, their perception of the impact of the CJRS on their personal finances was affected by their financial security before the pandemic, whether or not their wages were topped up to 100% by their employer, and whether or not their pre-pandemic income fell within the £2500 per month cap, so that they received at least 80% of their usual pre-pandemic income.

Chapter 6: Impact of the CJRS on businesses

As well as protecting jobs and as a result household incomes, a further objective of the CJRS was to reduce the risk of permanent employer closures, by supporting those who faced having to scale back their business activities due to the pandemic.

The CJRS has directly supported 1.3 million businesses in total. Without the CJRS, 20% (around 250,000 when scaled up to the User population) reported they would have had to close permanently without the scheme but continued to trade throughout the pandemic or only closed temporarily. In addition, 10% reported that they would have needed to close temporarily but continued to trade throughout the pandemic.

6.1 Introduction

An important mechanism through which the CJRS intended to support the productive capacity within the macroeconomy was through protecting businesses and maintaining overall employment. The CJRS was not designed nor aimed to prevent every individual business from closing. While a key benefit of businesses is the jobs they enable, there are impacts over and above this – for example in how their scale supports more efficient and productive job creation than would otherwise be possible.

This chapter addresses the evaluation question:

• how effective was the scheme at reducing the risk of permanent business closure?
• to what extent did employers and employees respond to scheme changes as intended?

6.2 Methods

This chapter uses several approaches to assess the impact of the CJRS on businesses. The primary approach to estimate the number of businesses that would have had to permanently close without the CJRS is based on self-reported research data.

The employer quantitative research asked employers what they anticipated would have happened to the trading status of their organisation if they did not receive CJRS. It is not known if some of these employers closed permanently after they were surveyed.

Due to the self-reported nature of the questions asked to employers about the expected impact of the pandemic on their business in the absence of the CJRS, some caution should be applied to these findings. HMRC data was used to show how many PAYE schemes had paid employments in April 2020 and no paid employments in January 2022, looking at the differences between the businesses that used the CJRS and those that did not. A literature review was also conducted to help estimate the cost of rehiring and training. Further detail on the supporting analysis is in the CJRS final evaluation accompanying technical information document.

6.3 Findings

6.3.1 Preventing permanent closure of employers and reducing the downsizing of employers

The findings from the employer quantitative research shows in figure 2.1, that across the whole scheme 72% of Users had a reduction in turnover between the start of the pandemic and the closure of the CJRS, compared with 40% of Non-Users. For Users who saw a decrease in turnover, the median and mean reductions were 27% and 43% respectively, and almost a third (31%^{[footnote 14]}) of them saw their revenue fall by over half (figure 6.1). This indicates the CJRS was generally used by those who most needed it, as identified in chapter 2.

Employers participating in the qualitative research reported that the CJRS primarily helped them by preventing staffing costs that risked causing mounting debt, a reduction in their cash reserves or savings and undermining their ability to function, particularly in the short-term. This also freed up time for employers to make more prudent decisions including delaying redundancies, closing sites and scaling back their business.

Figure 6.1: Proportion of reduction in revenue for those who had reported a decrease due to COVID-19

Source: CJRS employer and agent quantitative research wave 2 (2022)

Question: by approximately how much had your funding/turnover decreased between the start of the pandemic and 30 September 2021?

Base: Users (3,169) and Non-Users (623) that were trading at the time of the survey and saw a decrease in sales or turnover.

Question: By approximately how much had your funding/turnover decreased between the start of the pandemic and 30 September 2021?

Note: Only employers actively trading at the time of the survey were asked whether COVID-19 had a negative impact on their turnover.

Table 6.1: Proportion of reduction in revenue for those who had reported a decrease due to COVID-19

Impact	Users	Non-Users
Less than 10%	6	10
10% to 25%	27	34
26% to 50%	28	32
51% to 75%	16	10
76% to 90%	8	4
More than 90%	8	2
Don’t know	8	8

As shown in figure 6.2, quantitative research evidence found that without the CJRS 20% of Users reported their businesses would have had to close permanently but continued to trade through the pandemic or only closed temporarily. This equates to around 250,000 businesses when scaled up to the User population. In addition, 10% of Users would have had to close temporarily but continued to trade throughout the pandemic. On their own, these business closures prevented could account for an estimated 2 million redundancies, which is already included in the jobs protected research estimate presented in chapter 4.

Figure 6.2: Users reporting on what would have happened to their business without the CJRS

Source: CJRS employer quantitative research wave 2

Base: All users (4,860)

Question: What is the current status of your organisation?…Which of the following best describes what would have happened to your organisation if you had not received funding from the CJRS?

Notes: The trading status question covers the period up until the respondent was surveyed which was after the CJRS had closed. Figure 6.2 includes the category ‘Would have continued to operate’. This combines 3 individual response options.

Table 6.2: Users reporting on what would have happened to their business without the CJRS

Employers anticipated impact	Users
Would have continued to operate	51%
Would have continued to operate and are currently trading, have traded continuously	39%
Would have continued to operate but have since closed temporarily or permanently	12%
Would have temporarily closed	23%
Would have temporarily closed and have closed temporarily or permanently	13%
Would have temporarily closed but are currently trading, have traded continuously	10%
Would have permanently closed	21%
Would have permanently closed and have closed permanently	2%
Would have permanently closed but are currently trading and/or have only closed temporarily	20%

The employer research findings are reinforced by supporting analysis using VAT data which shows that a large number of VAT-registered CJRS claimants saw decreases in turnover compared to previous years.

In addition, analysing PAYE data showed that despite a considerable increase during COVID-19 in the number of PAYE schemes no longer having any paid employments, a higher proportion of Non-Users had no paid employments in January 2022 (17% for Users compared with 27% for Non-Users). Factors that may have increased the closure rate, in particular for small employers, include changes to off-payroll working for intermediaries and contractors.

Further research conducted by the Centre for Economic Policy Research analysed the impact of a range of firm-level COVID-19 support policies on productivity growth and allocation. When controlling for other variables, the study found that supported firms had a 43% lower probability of exiting the market during the first year of the pandemic. Additional counterfactual analysis also showed that the group receiving support from the policies would have been 61% more likely to exit the market if they had not received the support. This evidence further suggests that these types of policies introduced to support firms during the COVID-19 pandemic had a positive

6.3.2 Effect of the CJRS on businesses’ recovery

A key objective of the CJRS was retaining job matches and therefore enabling a smoother recovery when restrictions eased; the effect of this on the macroeconomy is explored holistically in chapter 7.

As a result of retaining job matches, costs of reopening could be reduced for employers due to lower re-hiring and training costs. Qualitative employer evidence suggests that losing these jobs would have presented the businesses with large hiring and training costs to fill vacant posts when the economy reopened. This included advertisement, consultancy costs and interview costs, as well as the loss of productivity compared to an employee who has undergone training or continued to gain experience.

Analysis suggests a saving central estimate of £9.1 billion, ranging from £8.4 billion to £9.8 billion. The central estimate was derived from estimating the costs that organisations would have incurred when rehiring employees when demand recovered, using the 4 million jobs directly protected estimate. This estimate of hiring and training costs saved does not feature in the VfM assessment in chapter 8. They are assumed to be incorporated within the main VfM estimate derived from the analysis on employment and economic output without the CJRS in place, which is explored in chapter 7 and chapter 8.

This is supported by quantitative employer research evidence, which found that 84% of employers who would have made redundancies without the CJRS, agreed that they would have had to make more employees redundant during the pandemic and rehire later. Of these, 68% agreed that the CJRS had helped them to save money on recruitment costs including 46% who agreed strongly, while 15% disagreed. The research findings also suggest that as well as the reduced costs of restarting work, the CJRS likely enabled a smoother economic recovery. Figure 6.3 suggests that without the scheme, over half (53%) would have taken longer to return to pre-COVID-19 levels, with most of this group taking at least 4 months, which as chapter 7 explores would affect the economy’s ability to return to trend output. Another 10% to 12% of businesses were not able to return to pre-COVID levels.

Figure 6.3: Anticipated time needed to return to pre-COVID-19 levels without the CJRS

Source: CJRS employer quantitative research wave 2

Base: Users that said they would have closed temporarily or operated on a smaller scale (3,054).

Question: If you had not received funding from the CJRS, how soon, if at all would your organisation have been able to return operating at normal pre-COVID-19 levels once restrictions stopped affecting your organisations?

Table 6.3: Anticipated time needed to return to pre-COVID-19 levels without the CJRS

Impact	Users
Don’t know	7%
Restrictions are still affecting the organisation	2%
Immediately- your organisation continued to operate as normal	5%
Your organisation would not have been able to return to operating at normal, pre-COVID-19 levels	10%
Four or more months after restrictions stopped affecting your organisation	28%
One to three months after restrictions stopped affecting the organisation	25%
As soon as restrictions stopped affecting the organisation	23%

6.3.3 Impacts of the changes to CJRS policy design on businesses

Changes were introduced in July 2020 to give employers flexibility around the number of hours that employees could be on furlough, no longer requiring them to either be completely on or off furlough. As evidence in chapter 4.5.4, there is strong qualitative evidence that these improvements in flexibility, such as flexible furlough, were a significant benefit to employers. Flexible furlough enabled them to manage fluctuating workloads and help employees ease into returning to work. Four in ten (40%) employers who used the CJRS also placed some of their workforce on flexible furlough. Employers reported that they found the changes a “game changer” as it gave them the ability to tailor staff hours to organisational need and bring back specific staff. However, some participants in the employee qualitative research reported that the introduction of flexible furlough did make it more difficult to plan their time outside of work, as they could be contacted to work at short notice.

This evidence suggests that the CJRS provided a significant benefit to employers, with the scheme’s adaptions over time ensuring that the CJRS met employers’ needs and could therefore have a greater impact on businesses’ chances of survival, further supporting the economy as it recovered from COVID-19. As shown in chapter 4 and evidenced in chapter 7, as a result of flexible furlough, employers were encouraged to consider the viability of their employees’ jobs in the medium and longer term, which would have enabled reallocation if that was the preferred business decision.

Figure 6.4 shows that although the average proportion of employments on flexible furlough varied by PAYE scheme size, they follow a similar trend. For all sizes, usage was lowest in July 2020, the first month that flexible furlough was available. However, usage increased in the following months, which may relate to businesses becoming more accustomed to the process of claiming for and using flexible furlough, as well as the easing of restrictions and the improved economic and pandemic outlook in summer 2020. For all PAYE schemes except micro PAYE schemes (1 to 9 employments), December 2020 was the month with the highest average proportion.

Figure 6.4: Trends in mean percentage of employments on flexible furlough at each PAYE scheme by size.

Source: HMRC CJRS data

Table 6.4: Trends in mean percentage of employments on flexible furlough at each PAYE scheme by size.

PAYE scheme size (no. of employees)	Jul 2020	Aug 2020	Sep 2020	Oct 2020	Nov 2020	Dec 2020	Jan 2021	Feb 2021	Mar 2021	Apr 2021	May 2021	Jun 2021	Jul 2021	Aug 2021	Sep 2021
Micro (1-9)	12	17	19	20	18	20	16	16	17	20	21	21	21	21	20
Small (10-49)	12	14	15	16	20	22	18	16	16	20	20	17	17	16	15
Medium (50-249)	6	8	8	9	13	15	12	11	11	13	13	11	10	9	9
Large (250+)	3	4	5	5	8	8	7	6	7	8	7	6	5	4	4
Total	12	16	18	19	18	20	16	16	16	20	21	20	20	19	19

Chapter 7: Macroeconomic impacts

The evolution of output and employment during 2020 and 2021 was highly atypical, when compared with previous recessions. Output contracted and rebounded sharply, reflecting the combination of the effects of the pandemic, changes in restrictions on economic activity and the mitigations provided by the CJRS and other policies. The CJRS scheme helped to limit the negative effect of restrictions on employment levels. Flows into unemployment were much lower than in previous recessions, despite a much larger contraction in GDP.

The maintenance of existing job matches enabled output to recover more quickly than would otherwise have occurred without the CJRS. Estimates derived from information about the speed with which unemployed individuals find employment suggest that employment could have been around 3.6% lower and output 1.8% lower in 2021 if the CJRS scheme had not been in place.

By supporting the incomes of employees on furlough, the CJRS helped to sustain household spending and hence the demand for goods and services. This effect was likely limited while restrictions were in place, but stronger when the economy reopened.

There is little evidence to suggest that the CJRS had a role in contributing to the higher inactivity rates seen in the labour market, with employees placed on furlough through the CJRS no more likely to become inactive in the first half of 2022, than employees not placed on furlough. Subsequent increases in inactivity, especially among the those aged over 50, were largely unrelated to whether individuals were on furlough or not.

7.1 Introduction

This chapter assesses the effect that the CJRS had on the economy, and the strength and speed of the recovery. The observed outcomes for output and employment during and after the operation of the scheme are the combination of underlying shocks – from the pandemic and from the restrictions on economic activity that were enacted to restrain infection rates – and the mitigations provided by the CJRS. As noted in chapter 1, the objectives of the scheme were to preserve the links between individual employers and employees, and to ensure employees could retain the majority of their usual salary, to support household incomes. The analysis in this chapter describes how these features may have affected the evolution of economic output and employment.

This chapter addresses the evaluation questions:

• to what extent did the scheme support the economic recovery?
• to what extent did the scheme reduce the risk of economic scarring – long- lasting and significant adverse effects on employment and output?

Information from the Labour Force Survey indicates that the chances of an individual being employed in 3 months’ time are typically around 97% if they are currently employed, but much lower (around 28%) if they are currently unemployed. The main explanation for this is that there is a high chance that employees will remain in employment with their current employer, while it generally takes time for those who are unemployed to find a new employment^{[footnote 15]}.

Consequently, by enabling employers to maintain job matches that might otherwise have been lost, the CJRS kept individuals in positions where they were more likely to be employed, both while on furlough and also afterwards. It helped to preserve productive capacity, thereby enabling the economy to adjust more smoothly once restrictions were fully eased, while also mitigating the risk of long-lasting scarring in the levels of employment and output.

7.2 Methods

The analysis reviews the evolution of output and employment during 2020 and 2021. It assesses how these outcomes may have been affected by the CJRS. The analysis uses the observed outcomes for the number of employments supported by the CJRS to estimate how much productive capacity might have otherwise been lost. It looks at what the evolution of vacancies reveals about the evolution of labour demand, as restrictions were imposed and eased throughout the pandemic.

The analysis derives counterfactual profiles for employment that could have occurred if the CJRS scheme had not been in place, by combining probabilities of moving between employment, unemployment and inactivity with estimates of increased flows from employment to unemployment. These are then combined with an assumption about the productivity of displaced employees to derive associated counterfactual profiles for output.

Three ‘no-CJRS’ reference scenarios have been created. These have differing assumptions about the flows into unemployment that would have occurred if the CJRS had not been in place and about the strength of labour demand from 2021 Q2 onwards. In all 3 scenarios flows into employment following the easing of restrictions on economic activity are stronger than the historical average. This is because labour demand would have recovered strongly to reinstate the productive capacity needed to meet the no longer suppressed demand for goods and services. Labour demand recovered strongly in the United States, which provides a relevant benchmark, as it made limited use of job retention programmes.

7.3 Findings

7.3.1 The evolution of economic output

The 2020 recession was distinctive in terms of its underlying causes and consequently in terms of the paths of output and employment. Specifically, output fell much more sharply than in typical recessions and rebounded more quickly, as shown in figure 7.1. This distinctive profile mainly reflected the introduction and subsequent changes to the restrictions on economic activity. All the GDP profiles in figure 7.1 were affected by multiple factors beyond the initial shocks. The recovery from the 2020 recession was curtailed in 2022 due to external factors impacting supply side constraints, such as the invasion of Ukraine and the consequent steep rise in energy prices.

Figure 7.1: GDP shortfalls in previous UK recessions

Source: Office for National Statistics: Gross Domestic Product and HM Treasury calculations

Notes: Peak quarters identified as those quarters where GDP was higher than in both the previous and the following 4 quarters.

GDP shortfalls calculated as difference from the projected level of GDP from each peak using an episode-specific growth rate. Episodic growth rate is calculated as the annualised median quarterly GDP growth rate observed in the 20 quarters up to and including each peak quarter. GDP growth rates for each episode are: 1973 Q2 2.7%; 1979 Q2 2.9%; 1990 Q2 2.5%; 2008 Q1 2.7%; 2019 Q3 2.2%

Table 7.1: GDP shortfalls in previous UK recessions

Quarters after peak quarter	1973 Q2 - gap (percentage)	1979 Q2 - gap (percentage)	1990 Q2 - gap (percentage)	2008 Q1 - gap (percentage)	2019 Q3 - gap (percentage)
0	0.0	0.0	0.0	0.0	0.0
1	-1.6	-3.0	-1.8	-1.1	-0.6
2	-2.7	-2.7	-2.9	-3.3	-3.8
3	-6.2	-4.5	-4.0	-6.2	-27.9
4	-5.4	-7.2	-5.0	-8.9	-13.1
5	-5.5	-8.1	-5.9	-9.8	-12.4
6	-7.6	-9.9	-6.4	-10.3	-14.0
7	-8.3	-10.9	-7.0	-10.6	-8.2
8	-10.6	-11.4	-8.0	-10.4	-7.1
9	-11.6	-11.0	-8.1	-9.9	-6.1
10	-11.2	-11.7	-8.2	-9.9	-6.2
11	-10.2	-12.4	-8.1	-10.4	-6.6
12	-11.0	-12.1	-8.4	-10.8	-7.3
13	-10.4	-12.6	-8.2	-11.4	-7.7
14	-9.0	-12.7	-8.3	-11.9	-8.1
15	-9.7	-11.6	-7.9	-12.6	N/A
16	-11.1	-11.7	-7.4	-12.5	N/A
17	-11.1	-11.4	-7.0	-13.2	N/A
18	-10.0	-11.4	-7.1	-12.6	N/A
19	-10.0	-11.5	-7.5	-13.4	N/A
20	-9.5	-13.2	-8.0	-13.8	N/A

7.3.2 Employment outcomes during the COVID-19 pandemic compared to previous UK recessions

Unemployment outcomes were also very different compared with previous UK recessions, as seen in figure 7.2. The unemployment rate peaked at 5.2% in the fourth quarter of 2020, 1.4 percentage points higher than a year earlier. It returned to its pre-recession rate level after 10 quarters. In contrast, it had remained 2 percentage points or more above its pre-recession level after 5 years in each of the previous 4 recessions. As with GDP, shown in figure 7.1, unemployment recovered much more quickly than in previous recessions. The unemployment peak was at an unusually low level compared to previous crises.

Figure 7.2: Unemployment profiles in UK recessions

Source: Office for National Statistics: unemployment

Note: Peak quarters identified as those quarters where GDP was higher than in both the previous and the following 4 quarters.

Table 7.2: Unemployment profiles in UK recessions

Quarters after peak quarter	1973 Q2 - percentage point change from peak quarter	1979 Q2 - percentage point change from peak quarter	1990 Q2 - percentage point change from peak quarter	2008 Q1 - percentage point change from peak quarter	2019 Q3 - percentage point change from peak quarter
0	0.0	0.0	0.0	0.0	0.0
1	-0.1	0.1	0.2	0.2	0.0
2	-0.3	0.2	0.6	0.7	0.2
3	-0.1	0.5	1.1	1.2	0.3
4	-0.1	1.0	1.8	1.9	1.1
5	0.0	1.8	2.3	2.6	1.4
6	0.0	2.7	2.6	2.6	1.1
7	0.3	3.6	2.8	2.6	0.9
8	0.6	4.3	2.9	2.8	0.5
9	1.0	4.6	3.0	2.7	0.2
10	1.3	4.9	3.5	2.6	-0.1
11	1.6	5.1	3.7	2.7	0.0
12	1.7	5.3	3.5	2.6	-0.2
13	1.8	5.5	3.3	2.7	-0.1
14	1.8	5.8	3.4	3.1	0.1
15	1.8	6.0	3.0	3.2	0.0
16	1.8	6.1	2.8	3.0	0.0
17	2.0	6.2	2.5	2.8	0.0
18	2.0	6.4	2.1	2.7	0.0
19	1.9	6.5	2.0	2.6	0.0
20	1.9	6.6	1.8	2.6	0.0

By contrast, the rise in inactivity during 2020 and 2021 was not especially atypical. It was somewhat higher than in the 1970s and 1980s, broadly the same as in 2008 to 2010, but distinctly lower than in the early 1990s, as shown in figure 7.3.

Figure 7.3: Inactivity rate profiles in UK recessions

Source: Office for National Statistics: Economic inactivity

Note: Peak quarters identified as those quarters where GDP was higher than in both the previous and the following 4 quarters.

Table 7.3: Inactivity rate profiles in UK recessions

Quarters after peak quarter	1973 Q2 - percentage point change from peak quarter	1979 Q2 - percentage point change from peak quarter	1990 Q2 - percentage point change from peak quarter	2008 Q1 - percentage point change from peak quarter	2019 Q3 - percentage point change from peak quarter
0	0.0	0.0	0.0	0.0	0.0
1	0.2	-0.1	0.0	-0.1	-0.3
2	0.2	-0.1	0.1	0.0	-0.4
3	0.2	-0.2	0.3	-0.1	0.2
4	0.1	-0.3	0.5	-0.2	0.3
5	-0.1	-0.3	0.8	0.1	0.4
6	-0.1	-0.3	1.2	0.3	0.6
7	-0.1	-0.4	1.4	0.4	0.4
8	-0.2	-0.3	1.6	0.6	0.4
9	-0.3	-0.1	1.7	0.5	0.5
10	-0.4	0.2	1.8	0.2	0.6
11	-0.3	0.4	1.7	0.5	0.6
12	-0.3	0.7	1.9	0.4	0.8
13	-0.2	1.0	1.9	0.3	0.6
14	-0.2	1.3	1.9	0.4	0.2
15	-0.1	1.6	1.9	0.2	N/A
16	-0.1	1.7	2.0	0.1	N/A
17	-0.1	1.2	2.0	-0.2	N/A
18	-0.1	0.7	2.4	-0.3	N/A
19	0.0	0.3	2.3	-0.6	N/A
20	0.1	0.0	2.2	-0.4	N/A

The 2020 recession was also highly asymmetric across different sectors of the economy, especially after June 2020. The accommodation and food and the arts and recreation sectors experienced much greater restraints on their activities. As shown in figure 7.4, this was reflected in their use of the CJRS, where employments on the scheme exceeded 75% of both these sectors’ employment in April to May 2020 and over 50% in January to February 2021, compared with ratios of 3% and 17% for all industries in these 2 periods.

Figure 7.4: The share of CJRS employments in selected sectors

Source: HMRC: Coronavirus Job Retention Scheme statistics and HMRC: Payrolled employees from PAYE RTI

Notes: ‘Wsale and retail’ refers to the wholesale and retail sector. ‘Accom and food’ refers to the accommodation and food service activities sector.

Table 7.4: The share of CJRS employments in selected sectors

Months	Total - all industries	Manufacturing	Construction	Wholesale and retail; repair of motor vehicles	Accommodation and food services	Finance and insurance	Arts, entertainment and recreation
Apr 2020 to May 2020	30.5%	36.8%	55.0%	40.5%	83.9%	6.4%	79.0%
Oct 2020	8.9%	8.8%	11.4%	8.5%	33.9%	1.9%	32.3%
Jan 2021 to Feb 2021	17.4%	14.0%	19.9%	22.7%	71.4%	3.0%	66.1%
Aug 2021 to Sep 2021	4.4%	5.7%	7.8%	4.4%	9.5%	1.3%	10.2%

Despite its high use of the CJRS, employee numbers in the accommodation and food sector fell by far more than in the rest of the economy, both in the first 6 and 12 months of the pandemic, as seen in figure 7.5. Vacancies were especially low, relative to 2019 levels, in both the accommodation and food services and also the arts and recreation sectors, shown in figure 7.6. However, in both of these industries the ratio of vacancies to employees rose well above their 2019 levels in the second half of 2021, with employment in the accommodation and food sector rising by around 340,000 over the 12 months from February 2021. This suggests that even though the CJRS preserved a high level of job matches in the accommodation and food sector, there was still a shortfall in productive capacity that needed to be rebuilt once the sector reopened fully.

Figure 7.5: Change in employee numbers on previous 6 months, in selected sectors

Source: HMRC: Payrolled employees from PAYE RTI

Notes: ‘Wsale and retail’ refers to the wholesale and retail sector. ‘Accom and food’ refers to the accommodation and food service activities sector.

Table 7.5: Change in employee numbers on previous 6 months, in selected sectors

Month	Manufacturing	Wholesale and retail; repair of motor vehicles	Accommodation and food service activities	Professional, scientific and technical	Administrative and support services	Public sector Education and health	Arts, entertainment and recreation	Other sectors
Aug 2020	-76,807	-73,650	-224,594	-39,393	-123,879	-4,302	-73,238	-134,169
Feb 2021	-35,230	-78,313	-118,191	13,133	42,395	87,819	-42,416	-7,051
Aug 2021	45,970	37,436	239,371	95,215	92,020	156,138	56,773	90,788
Feb 2022	26,217	38,511	111,637	55,306	44,863	38,133	29,720	84,836
Aug 2022	3,275	-19,220	40,108	57,894	104	97,658	22,986	94,908

Figure 7.6: Vacancies per 100 employees, in selected sectors

Source: Office for National Statistics: vacancies by industry

Notes: ‘Accom and food’ refers to the accommodation and food service activities sector. ‘All services’ is the total service sectors identified in the ONS source.

Table 7.6: Vacancies per 100 employees, in selected sectors

Time period	All vacancies	Manufacturing	Total services	Accommodation and food service activities	Information and communication	Human health and social work activities	Arts, entertainment and recreation
2019 Average	2.7	2.2	2.8	3.9	3.4	3.4	2.7
Apr 2020 to Sep 2020	1.35	1.25	1.35	0.95	1.3	2.65	0.55
Oct 2020 to Mar 2021	2.05	2	2	1.3	2.7	3.1	1.25
Apr 2021 to Jun 2021	2.8	2.7	2.9	4.9	3.5	3.6	3.7
Jul 2021 to Sep 2021	3.7	3.6	3.8	6.8	4.8	4.2	4.7
Oct 2021 to Dec 2021	4.1	4	4.2	7.9	5.2	4.8	4.7

7.3.3 Effects of the CJRS on employment and economic recovery

As evidenced throughout this evaluation, the CJRS maintained employment levels much closer to their pre-pandemic levels than would have been the case without it. As presented in figure 1.3, between April and June 2020, around 25% of the workforce was not fully utilised, with this proportion falling below 10% during 2021, when the economy began to fully reopen and many CJRS recipients were no longer using the scheme.

The analysis presented in chapter 4 indicated that around 4 million jobs could have been lost if the scheme had not been in place. This estimate has been combined with information about the probabilities of entering and leaving employment to derive counterfactual profiles, which show that employment could have been persistently lower than its observed outturn, as shown in figure 7.7.

Data from the Labour Force Survey shows that the probability of someone being employed in 3 months’ time is much higher (typically around 97%) if they are currently employed than if they are currently unemployed (typically around 28%) or if they are inactive (typically around 6%^{[footnote 16]}. This reflects the time that it takes for those who are not currently employed to find suitable employment^{[footnote 17]}. A consequence of this feature is that when there are large movements out of employment, it generally takes time for employment and unemployment to revert to their original levels^{[footnote 18]}.

Estimates of how employment might have evolved in the absence of CJRS have been generated by assuming that the flows from employment to unemployment between 2020 Q2 and 2021 Q1 would have been higher by up to 4 million (the jobs directly protected estimate reported in chapter 4), and using calibrated transition probabilities^{[footnote 19]} between employment, unemployment, and inactivity to simulate the profile of employment from 2020 Q2 onwards.

Three ‘no-CJRS’ reference scenarios have been created. These have differing assumptions about the additional inflows into unemployment between 2020 Q2 and 2021 Q1, and about the strength of the recovery in labour market demand once restrictions on economic activity had been eased in 2021.

The central scenario assumes additional flows from employment to unemployment of 3.8 million (1.7 million in each of 2020 Q2 and 2020 Q3 following the restrictions on the economy introduced in March 2020 and 400,000 in 2021 Q1, following the renewed tightening of restrictions in January 2021)^{[footnote 20]}. The total flows into unemployment are a little lower than the estimate of 4 million jobs directly protected reported in chapter 4. This is because some individuals may have been able to transition to a new job within the same calendar quarter.

The central scenario also sets the transition probabilities into employment in 2021 Q2 from each of the categories of employment, unemployment and inactivity to values observed at the 95th percentile of the distributions of quarterly transition rates into employment in the Labour Force Survey data^{[footnote 21]}. Those transition probabilities are held at the same level until 2022 Q1, and then subsequently converge back to the average values observed between 2015 and 2019.

The lower VfM scenario assumes the additional inflows into unemployment would have been smaller (3.6 million) than in the central scenario. It also assumes that the recovery in labour demand would have been somewhat stronger after restrictions were eased, by setting higher transition probabilities into employment than in the central case^{[footnote 22]}.

The upper VfM scenario assumes the additional inflows into unemployment would have been larger (4 million). It assumes that the recovery in labour demand would have been a little weaker than in the central scenario, by setting lower transition probabilities into employment than in the central case.

All 3 scenarios assume that the flows from unemployment to employment would have been higher by 400,000 in 2021 Q2 than given by the calibrated transition probability. That was the quarter when the economy reopened, which resulted in the numbers on full furlough falling from 2.9 million to below 1 million between the end of March and the end of June, and those on flexible furlough falling from 1.3 million to 900,000 over the same period. The observed increase in labour utilisation outcomes when the CJRS was in place suggests that it is plausible that labour demand (and consequently flows into employment) would have been unusually strong in the no-CJRS scenarios in 2021 Q2^{[footnote 23]}.

Employment in the 3 scenarios in 2020 Q3 is lower by between 8.7% and 9.8% than the actual outturn, reflecting the additional flows out of employment into unemployment in 2020 Q2 and Q3 seen in figure 7.7. The scenarios diverge thereafter, with employment recovering most strongly in the lower VfM scenario. All scenarios show a marked narrowing of the employment shortfall relative to the observed outcome in 2020 Q4, reflecting a combination of high flows back from elevated levels of unemployment and the absence of an enhanced flow from employment into unemployment in that quarter. They show a further narrowing in 2021 Q2, the quarter in which restrictions on economic activity were largely eased.

The gap between employment in each scenario and observed (with-CJRS) outturns provides an estimate of the effect that the scheme may have had in limiting the scarring effect of individuals leaving employment taking time to re-enter it^{[footnote 24]}. In the central scenario this gap falls from around 3.6% of actual employment in 2021 to around 0.7% in 2022, and reduces further in subsequent years. In the lower VfM scenario, the gap is 3.2% in 2021 and 0.3% in 2022, and in the upper VfM scenario it is 3.9% in 2021 and 1% in 2022. Further information is presented in the CJRS final evaluation accompanying technical information document.

Figure 7.7: Employment in no-CJRS scenarios: percentage point difference from with-CJRS outturns and projections

Notes: The no-CJRS counterfactuals assume that up to 4 million CJRS recipients would have become unemployed if their employment had not been supported by the scheme. They use differing assumptions about labour market transition rates to track the subsequent evolution of employment, unemployment and inactivity. Outturns are projections from 2023 Q1 (January to March)

Table 7.7: Employment in no-CJRS scenarios: percentage point difference from with-CJRS outturns and projections

Quarter	Percentage point difference - lower VfM scenario	Percentage point difference - central VfM scenario	Percentage point difference - upper VfM scenario
2020 Q2	-5.1	-5.4	-5.8
2020 Q3	-8.7	-9.3	-9.8
2020 Q4	-6.2	-6.6	-7.0
2021 Q1	-6.1	-6.4	-6.7
2021 Q2	-2.9	-3.3	-3.6
2021 Q3	-2.3	-2.7	-3.0
2021 Q4	-1.4	-1.8	-2.2
2022 Q1	-0.9	-1.4	-1.7
2022 Q2	-0.3	-0.7	-1.1
2022 Q3	0.1	-0.3	-0.7
2022 Q4	0.1	-0.2	-0.6
2023 Q1	0.1	-0.1	-0.5
2023 Q2	0.1	-0.1	-0.5
2023 Q3	0.1	-0.1	-0.4
2023 Q4	0.0	0.0	-0.4
2024 Q1	0.0	0.0	-0.3
2024 Q2	0.0	0.0	-0.3
2024 Q3	0.0	0.0	-0.3
2024 Q4	0.0	0.0	-0.2
2025 Q1	0.0	0.0	-0.2
2025 Q2	-0.1	0.0	-0.2
2025 Q3	-0.1	0.0	-0.2
2025 Q4	-0.1	0.0	-0.1
2026 Q1	-0.1	0.0	-0.1
2026 Q2	-0.1	0.0	-0.1
2026 Q3	-0.1	0.0	-0.1
2026 Q4	0.0	0.0	0.0

Counterfactual profiles for output have been derived by combining the paths for employment with an estimate that the labour productivity of those who were on the CJRS scheme was, on average, 89% of the productivity of the whole workforce. HMRC RTI PAYE data indicated that CJRS take-up was more heavily skewed towards industries with lower mean pay, which is used as a proxy measure of the productivity of employees in each industry.

Employment in the no-CJRS scenarios is lower than observed employment from 2020 Q2 onwards, seen in figure 7.7. However, recipients of CJRS were employed but not producing any output (if they were fully on furlough) or reduced output (if they were on flexible furlough). As a result, the lower levels of employment in the no-CJRS scenarios would not have been associated with lower levels of output while economic restrictions were in place. Consequently, the difference in the level of output between the with-CJRS and no-CJRS scenarios is less than would be implied by the difference in employment levels between 2020 Q2 and 2021 Q1. However, because CJRS provided those who might otherwise have become unemployed with higher incomes than they would have received from Universal Credit, the level of consumer spending is likely to have been higher than if the scheme had not been in place. As a result, GDP in the central scenario could have been lower by around 0.2% between 2020 Q2 and 2021 Q1 than what actually occurred^{[footnote 25]}.

The effect of the scheme on the level of output would have been much greater in 2021 Q2, the quarter in which the economy reopened. Employments on the scheme fell by over 2 million between the end of March and the end of June 2021, as previously under-utilised capacity became operational. By contrast, in the no-CJRS scenarios, it would have taken longer to bring those who had been made unemployed back into employment, with the consequence that output would have recovered more slowly. In the central scenario, this feature accounts for a widening of the gap between observed (with-CJRS) outcomes and the no-CJRS counterfactual to 2.9% in 2021 Q2. That gap then narrows in subsequent quarters, as more of those who had become unemployed re-enter employment shown in figure 7.8.

The VfM assessment in chapter 8 calculates the Present Value (PV) of the additional post-restrictions output that may have been enabled by the CJRS.

Figure 7.8: Estimated impact of no-CJRS scenarios on GDP

Note: The counterfactual path for GDP is derived from the difference between actual and counterfactual employment shown in figure 7.7.

Table 7.8: Estimated impact of no-CJRS scenarios on GDP

Quarter	Difference from outturn or OBR projection - lower VfM scenario (percentage)	Difference from outturn or OBR projection - central VfM scenario (percentage)	Difference from outturn or OBR projection - upper VfM scenario (percentage)
2020 Q2	-0.1	-0.1	-0.2
2020 Q3	-0.2	-0.2	-0.3
2020 Q4	-0.2	-0.3	-0.4
2021 Q1	-0.2	-0.3	-0.5
2021 Q2	-2.6	-2.9	-3.2
2021 Q3	-2.0	-2.4	-2.7
2021 Q4	-1.2	-1.6	-1.9
2022 Q1	-0.8	-1.2	-1.6
2022 Q2	-0.3	-0.7	-1.0
2022 Q3	0.0	-0.3	-0.6
2022 Q4	0.0	-0.2	-0.5
2023 Q1	0.0	-0.1	-0.5
2023 Q2	0.0	-0.1	-0.4
2023 Q3	0.0	-0.1	-0.4
2023 Q4	0.0	-0.0	-0.3
2024 Q1	0.0	-0.0	-0.3
2024 Q2	0.0	-0.0	-0.3
2024 Q3	0.0	-0.0	-0.2
2024 Q4	0.0	0.0	-0.2
2025 Q1	0.0	0.0	-0.2
2025 Q2	0.0	0.01	-0.2
2025 Q3	0.0	0.0	-0.2
2025 Q4	0.0	0.0	-0.1
2026 Q1	0.0	0.0	-0.1
2026 Q2	0.0	0.0	-0.1
2026 Q3	0.0	0.0	-0.1
2026 Q4	0.0	0.0	-0.1

7.3.4 How the CJRS supported the economic recovery

As evidenced in chapter 5, the CJRS supported household incomes, enabling them to spend more than they would have done without the CJRS. The ratio of this additional spending to the fiscal cost of the scheme represents the fiscal multiplier effect of the CJRS. However, a Bank of England staff working paper shows that opportunities to spend were constrained when restrictions were in place. The household saving ratio rose sharply from 5% in 2019 to 16% in 2020, largely because consumers were unable to purchase specific goods and services. Consequently, the effect on private sector spending from higher government spending (‘the fiscal multiplier’) would have been lower, initially, than the OBR’s usual cumulative fiscal multiplier of 0.6^{[footnote 26]}.

However, the effect on consumption was likely displaced over time, occurring when restrictions were eased and the economy was recovering, instead of when the fiscal cost was incurred. There is evidence for such displacement from an empirical study of multipliers during past pandemics, including the Severe Acute Respiratory Syndrome (SARS) pandemic. That study found that multipliers increased after the immediate duration of the pandemic, and that some were twice as large in the second year of a health crisis.

As discussed in chapter 6, the CJRS helped employers to save on rehiring and training costs, by subsidising the retention of existing employees. The easing of restrictions and closure of the CJRS prompted organisations to take steps to return to work such as financial re-forecasting for staffing and operations, outreach to clients and adjusting of business models (see chapter 4 and chapter 6). This ability to position themselves for the reopening, combined with saving employers’ costs on rehiring and training employees, likely enabled employers to restart activity quickly, enabling a smoother and stronger recovery in output than would have occurred if the CJRS had not been in place.

The pace of recovery from recession is also affected by the movement of labour away from low productivity firms and activities to higher productivity ones. This is also an important mechanism for raising whole economy productivity. There was a risk that job retention schemes could inhibit such growth-enhancing reallocation^{[footnote 27]}. The introduction of employer contributions from August 2020 was designed and implemented to reduce that risk, with chapter 4 providing some evidence that it did so.

In the event, employment grew more strongly in higher productivity firms compared with lower productivity firms^{[footnote 28]}, both among those firms who used the CJRS and those firms who did not use the scheme, as shown in figure 7.9. However, the difference was lower in the CJRS group, which suggests that the CJRS may have had some, albeit small, effect on reallocation. That effect will have been limited by the increase in employer co-payments and reduced take-up of the scheme from July 2020, and by its closure in September 2021. The Organisation for Economic Co-operation and Development (OECD) published analysis on the employment dynamics across firms during COVID-19, accounting for the role of job retention schemes^{[footnote 29]}. It found that this reallocation process remained productivity-enhancing as high productivity firms exhibited higher survival rates, corroborating the evidence in this evaluation.

Figure 7.9: Employment growth rates of high and low productivity firms, by their use of the CJRS.

Source: HMRC RTI and CJRS data

Note: Employment growth is measured from March 2020 to the named month in the chart.

Table 7.9: Employment growth rates of high and low productivity firms, by their use of the CJRS.

Employment growth rate (percentage)	CJRS - Sep 2020	CJRS - Mar 2021	CJRS - Sep 2021	Non-CJRS - Sep 2020	Non-CJRS - Mar 2021	Non-CJRS - Sep 2021
Low productivty	6.7	18.7	27.1	-9.8	33.9	40.2
High productivity	7.8	21.7	31.5	-11.7	40.7	48.4

7.3.5 Impact of the CJRS closure

Before the CJRS ended, there were concerns that the closure of the scheme could have produced a rise in unemployment, as noted in the Bank of England November 2021 Monetary Policy Report. There were 1.2 million CJRS employments when the scheme closed on 30 September 2021. If most or all of the 1.2 million employments had been severed once the scheme ended, this could have led to a significant increase in one (or both) of unemployment and inactivity. However, unemployment fell by 79,000 in the 3 months after the closure of the scheme, compared with the previous quarter^{[footnote 30]}. Inactivity rose by 90,000 over the same period. This suggests that over 1 million of the 1.2 million CJRS recipients remained in employment following the closure. Additionally, as set out in chapter 6, the CJRS enabled employers to return to more normal operating capacity than there would have been without the scheme. Therefore, this would also have contributed to there being no significant increase in unemployment and inactivity when the CJRS closed.

7.3.6 CJRS and inactivity

The number of economically inactive individuals increased by 800,000 between the end of 2019 and mid-2022, largely reflecting an increase of 620,000 among those aged over 50. One way to assess how the CJRS may have affected inactivity levels is to consider the proportion of individuals who were employed at least once from January 2020 to September 2021 who were subsequently outside PAYE employment for 6 months or more by June 2022. This analysis shows that recipients of the CJRS were slightly less likely to have been non-employed (either unemployed or inactive) for more than 6 months by June 2022 than those who had been employed but had not received support from the scheme. For those aged 55 and over, the difference in the proportions of those previously on furlough and previously not on furlough was higher than for younger age groups, as shown in figure 7.10. These findings demonstrate that a greater proportion of employees that were not on furlough, at any point from January 2020 to September 2021, were non-employed than those who were on furlough at least once during the same period.

These outcomes suggest that the observed increase in inactivity among older workers is largely independent of whether they were on furlough or not. It is likely to reflect a number of other factors, such as health conditions and lifestyle choices, about the age of retirement^{[footnote 31]}. The findings also suggest that the scheme did not contribute towards an increase in the number of individuals who became non-employed for more than 6 months. The CJRS final evaluation accompanying technical information document discusses the methodology and caveats further.

Qualitative employee research found evidence that employees placed on furlough reassessed their personal circumstances, priorities and longer-term career options. This resulted in changes to working patterns and working roles, both during and post-CJRS, which aligned with some older individuals finding value in new hobbies. This impact could be attributable to the experience of being on furlough, or to their response to the COVID-19 pandemic as a whole. It was less evident for employees in the sample who were not placed on furlough.

Figure 7.10: Proportion of individuals who were non-employed for more than 6 months by June 2022

Source: HMRC RTI and CJRS data

Notes: Figure shows the proportions of each age group who had been employed in at least one month between January 2020 and September 2021, and who were subsequently non-employed for more than 6 months by June 2022. Individuals could have been employed on a full and/or part-time basis. Employment status measured by HMRC’s PAYE Real Time Information data.

Table 7.10: Proportion of individuals who were non-employed for more than 6 months by June 2022

Proportion (percentage)	0-17	18-24	25-49	50-54	55-59	60-64	65+
Previously not on furlough	15	15	9	8	14	27	39
Previously on furlough	10	10	9	8	11	23	34

7.3.7 The effect of job retention schemes in other countries

Before the introduction of the CJRS, job retention schemes had not been used in the UK. The scale of the economic shock caused by the pandemic led to many other countries offering similar schemes to support employees and businesses. The interventions that countries took in response to the COVID-19 pandemic varied around the world, but can be summarised by 3 broad types: adjusted or expanded existing short-term work schemes such as those used to mitigate job losses in the 2008 financial crisis in a number of OECD countries, new short-term work schemes, and new wage subsidy schemes, as shown in table 7.11^{[footnote 32]}. While all differing in design, they had the general aim of mitigating the fall in employment from the pandemic and subsequent economic restrictions.

International evidence on the impact and value of job retention schemes is limited, with detailed evaluations generally not available. However, it is possible to make comparisons with the design of other countries’ schemes, and detail some of the emerging outcomes of some international job retention schemes. It is worth noting that there were many varying factors between countries, for example the severity of COVID-19 impact and the level of restrictions put in place.

Effective job retention schemes in OECD countries were seen to have 3 primary qualities. They were timely, so that individuals who need support were reached as quick as possible. They were targeted, to minimise support for jobs that would have otherwise disappeared or those that did not require support. They were temporary, to ensure that these schemes did not undermine job reallocation and creation.

In Germany, the pre-existing short-term work scheme Kurzarbeit was expanded to increased accessibility and wage contributions on hours not worked. Contributions increased with the length of time on the scheme. In France, the Activité Partielle scheme was adapted with increased accessibility and contributions on hours not worked, but with similar alterations as the CJRS by adapting contributions as the pandemic evolved. Australia introduced the JobKeeper scheme, a recurring lump sum subsidy towards wages and the eligibility criteria was based on if an individual worked more than 20 hours or not. The size of this subsidy slowly fell as the pandemic went on. The European Trade Union Institute have produced a report which details the nature of job retention schemes in Europe, including take-up over time and generosity. Table 7.11 provides an overview of the various job retention schemes used in OECD countries.

An IMF working paper studied how effective the Kurzarbeit scheme was by comparing different German states. It found that the peak in unemployment of 3.9% would have increased by an additional 2 to 3 percentage points and that the contraction in domestic demand would have been 2 to 3 times larger without the scheme. Another IMF working paper on job retention schemes in the EU found that these schemes absorbed up to 80% of market income shocks, nearly twice the effect of pre-pandemic EU tax and benefit systems. Without these schemes, unemployment may have risen an additional 3 percentage points.

In the Australian Treasury’s preliminary evaluation of the first 6 months of the JobKeeper scheme, it was seen to have saved one-fifth of jobs supported and promoted the rapid recovery of consumer and business confidence. The scheme was also seen to keep the peak in unemployment at around 7% in comparison to the 14% estimate and significantly better GDP outcomes than the estimated peak contraction of 24%. A University of Melbourne paper reviewing the literature on the JobKeeper scheme also found that employment was 700,000 higher after 4 months and 812,000 higher after 12 months. These estimates corresponded to annual cost per job saved estimates of 113,000 AUD and 100,000 AUD.

An Observatoire français des conjonctures économiques (OFCE) working paper on the Activité Partielle scheme showed that it was highly effective at stabilising the peak in unemployment and keeping the fall in household incomes low at 4.7% instead of an estimated drop of 13.1%. However, the effect on the initial contraction in GDP was small at a 0.2 percentage points increase.

Table 7.11: Use of short-work schemes in OECD countries

Country	Pre-existing short-term work scheme	Increased access and coverage	Increased benefit generosity	Increased access for workers in non-standard jobs	New short-term work scheme	New wage subsidy scheme
Australia						Yes
Austria	Yes	Yes	Yes
Belgium	Yes	Yes	Yes
Canada	Yes					Yes
Chile	Yes	Yes	Yes	Yes
Czech Republic	Yes	Yes	Yes
Denmark	Yes	Yes			Yes
Estonia						Yes
Finland	Yes	Yes	Yes	Yes
France	Yes	Yes	Yes	Yes
Germany	Yes	Yes	Yes	Yes
Greece					Yes
Hungary					Yes
Iceland					Yes
Ireland	Yes					Yes
Italy	Yes	Yes		Yes
Japan	Yes	Yes	Yes	Yes
Korea	Yes	Yes	Yes
Latvia					Yes
Lithuania					Yes
Luxembourg	Yes	Yes	Yes
Netherlands	Yes					Yes
New Zealand						Yes
Norway	Yes	Yes	Yes
Poland						Yes
Portugal	Yes	Yes		Yes
Slovak Republic	Yes	Yes	Yes
Slovenia					Yes
Spain	Yes	Yes	Yes	Yes
Sweden	Yes	Yes	Yes
Switzerland	Yes	Yes		Yes
Turkey	Yes	Yes		Yes
United Kingdom					Yes
United States	Yes	Yes	Yes

Chapter 8: Value for money assessment

Value for Money (VfM) can be thought of as to what extent the scheme met its objectives whilst minimising wastage of taxpayer resources. To appraise the VfM of the CJRS, we can use 2 alternative approaches: i) the ‘social value’ - the scheme’s value to society, which takes into account its overall effect on public welfare, and ii) the ‘exchequer value’ - the scheme’s direct effect on the public finances.

The main economic benefits of the scheme were to ensure a quick economic recovery by preserving job matches, preventing falls in income and supporting household spending. The analysis in this evaluation shows that the CJRS directly protected around 4 million jobs, and prevented around 250,000 employers from closure. This contributed to reducing the economic impacts of the pandemic. It is estimated that the CJRS raised the level of economic output by around 1.8% in 2021 and 0.6% in 2022, with smaller increases in subsequent years.

Overall, the CJRS was good VfM, with significant positive impacts on jobs, incomes, businesses and the macroeconomy. Using a social value method, this results in a central positive Net Present Value (NPV) of £50 billion, indicating a benefit: cost ratio of around 4:1. Under reasonable assumptions about labour market activity, the NPV could be as high as £81 billion.

The narrower exchequer value method produces a central total NPV for the scheme of -£25 billion, much lower than the gross value of grants paid out. The scale of costs and benefits reflects how major an intervention the CJRS was in the UK economy.

The scheme is shown to be reasonably equitable, benefitting a broad base of working households. It particularly benefitted those sectors and parts of the population that were most affected by COVID-19, such as hospitality and those on lower incomes. The CJRS produced an equity benefit to society of around £6 billion, included as part of the social value estimate.

The VfM assessment shows a percentage of the spend (4.7%, totalling £3.3 billion) is estimated to have been paid out to employers who may not have needed support in hindsight. Some form of deadweight is inevitable in any scheme, especially so for the CJRS given the unprecedented nature of the pandemic. This was also the case given the need to act quickly in an uncertain economic landscape to avert widespread job losses, reduce the risk of economic harm and make the scheme easy for employers to use and understand. For the same reasons as there was some form of deadweight, the CJRS is estimated to have had a social cost of £3.5 billion in error and fraud. This level of error and fraud is at the lower end of initial planning assumptions.

8.1 Introduction and methodology

This chapter synthesises evidence from across the CJRS final evaluation to assess the value for money (VfM) of the scheme. VfM can be thought of as the extent to which the scheme met its objectives whilst being as economical as possible with the use of taxpayer resources.

When considering the VfM of the CJRS, it is important to consider the uncertain backdrop in which it was created. For example, it was difficult to know how long the COVID-19 pandemic would last, how the economy would adapt to the NPIs, and how incomes would be affected without a wage subsidy scheme. This was the case throughout the lifetime of the CJRS, not just when the scheme was introduced.

The quantified VfM study cannot account for some significant issues such as the certainty that implementing the CJRS brought, compared to the highly uncertain alternative of not introducing a scheme to support jobs and the economy.

The Green Book guidance used by the government to appraise policies provides a well-established basis for this assessment. However, the scale and nature of the CJRS was such that there are areas where the usual Green Book methodology has been adapted to more accurately reflect the scheme’s impact and value, notably the wider macroeconomic benefits of the scheme.

The CJRS VfM framework follows the ‘4Es’ approach used by the NAO, in structuring this VfM assessment. The definitions used in this framework are set out below:

• Effectiveness: the outcomes of the scheme and extent to which they achieved the scheme’s objectives. This includes the impact of protecting jobs, businesses and supporting the economy

• Equity and equality: the extent to which the scheme reached those most in need of support, and whether this was delivered in a fair and equal way. This includes the progressivity of the policy, as well as an assessment of support according to protected characteristics

• Economy and costs: the cost of, and resources used to implement, the scheme to achieve its intended outcomes. This includes the scheme’s delivery and implementation costs, as well as how much was paid out through the CJRS

• Efficiency: the relationship between the 3 factors above. The extent to which resources were well used in achieving the scheme’s intended outcomes. This includes the levels of deadweight and error and fraud

The VfM assessment quantifies the majority of the important elements, with ranges to reflect where the analysis is uncertain and challenging to assess. It highlights risks of over or under-estimation. The quantified elements are added together to create an overall range for the scheme’s VfM, in present value terms. However, there are aspects of this assessment where a qualitative approach is used, when factors are identified as relevant to the VfM of the scheme which cannot be quantified, for example due to a lack of quantitative data. These factors are still discussed and commented on, but are not included in the final quantitative assessment. This approach is fairly standard in VfM assessments and broadly follows the DWP’s social cost benefit analysis framework.

The VfM assessment uses a no-CJRS counterfactual scenario, as stated throughout this evaluation. To consider the full opportunity costs of the scheme, an approach to take could be assessing alternative scheme choices. For example, this could be comparing against a more targeted CJRS or a different policy altogether. However, there is difficulty in producing robust data and appraising alternative policies. Instead, this evaluation includes external evidence on the policies adopted by other countries during the COVID-19 pandemic in chapter 7.3.7, to help address these elements.

Two alternative valuation methods are presented both for transparency, and to provide a full picture of the CJRS’ VfM:

• the first is the social value method. This is the primary methodology set out in the Green Book for appraising government policies and shows the overall effect on economic welfare (including individuals, businesses, the Exchequer and wider society)
• the second method is to calculate the exchequer value of the scheme. This is a less holistic assessment of the scheme’s value, but shows the direct impact of the scheme on the public finances

A major difference between the social value and exchequer value methods is how they treat the transfer of resources from the government to individuals and organisations. Following the guidance set out in section 6.3 of the Green Book, the CJRS payments are treated as a transfer of resources between groups but do not create or destroy resources and so do not make society as a whole better or worse off. Therefore, when determining the social value of the CJRS, these scheme transfer payments are not treated as full costs, but they are fully taken into account when determining the exchequer impacts and exchequer value of the CJRS. Consequently, only 20% of the cost is counted on a social value basis when following DWP’s social cost benefit analysis framework. This is financed by either borrowing, taxation or reduced spending elsewhere.

Similarly, given its focus on public finances, the exchequer value basis accounts for the additional tax revenue and changes to benefit spending resulting from additional GDP associated with the scheme. More specifically, the exchequer value is calculated as 33.5% of the total social benefit, reflecting the current UK tax to GDP ratio.

Table 8.1 below sets out the scheme’s overall costs and benefits on both a social and exchequer method, for each of the 4 ‘Es’ (effectiveness, equity, economy and efficiency). Following this, the chapter then explains how the costs and benefits of each of these 4 Es is derived, for both methods.

There are several instances where this VfM assessment deviates from the Green Book:

• Effectiveness: to quantify the effectiveness of the CJRS, a macroeconomic approach has been adopted, as this more holistically captures the effect on the UK economy, rather than a ‘bottom up’ approach that could be used by calculating the value attributable to individual impacts of the scheme – the impact on jobs, businesses and incomes. This approach limits the risks of double counting benefits and better reflects the longer-term benefits beyond when COVID-19 restrictions and the CJRS ended
• Efficiency: error and fraud is treated as a social cost. The major driver in the difference between the headline exchequer and social valuation is the fact that using a social value method, CJRS subsidies are treated as transfers between taxpayers. However, error and fraud is separated out to reflect that it is counted as a social cost. Using an exchequer method, error and fraud (and also deadweight costs) are contained within the main economy costs

As shown in table 8.1 the main benefit of the scheme comes from effectiveness where the macroeconomic impact covered in chapter 7 is counted. The benefits from equity are only quantified through improvements to inequality and are not actualised as any benefits on tax revenue. Although economy represents the largest cost for both methods, there are some significant savings that bring it down from the £69 billion gross cost of grants. Efficiency is only considered on the social side, as it is already counted within the overall economy cost, reflecting the exchequer cost of the scheme.

Table 8.1: Quantified CJRS value for money assessment

Type	Social value, £ billions	Exchequer value, £ billions
Effectiveness	58	19
Equity	6	N/A
Economy and costs	-11	-45
Efficiency	-4	-7 (included within economy)
Total benefits	64	19
Total costs	-14	-45
NPV	50	-25

Notes: N/A is used in table 8.1, meaning not applicable given that those elements of VfM do not apply or are already counted within the exchequer calculation. Figures rounded to the nearest £ billion, and therefore may not sum. Costs are denoted as negative values. Table 8.1 benefits and costs totals differ to those in table 8.6 and table 8.7, due to presenting on an ‘E’ basis rather than a benefits and costs basis. The NPV remains the same across all tables.

8.2 Effectiveness

8.2.1 The value of the CJRS preserving economic output

As set out above in chapter 8.1, a macroeconomic approach is adopted to estimate the majority of the CJRS’s benefits. This is broadly the same when using both an exchequer and social valuation, with small deviations explained in further detail where relevant.

Chapter 7 presented estimates of the effect that the CJRS had in enabling economic output to recover more rapidly than it would otherwise have done. Those estimates indicate that the CJRS raised the level of economic output in the central scenario by around 1.8% in 2021 and 0.6% higher in 2022, with smaller increases in subsequent years. The PV of that additional output can be calculated by valuing those increases at fiscal year 2020-21 prices, and by discounting them at the rate of 3.5% per annum, as set out in the Green Book. This gives an estimate of around £58 billion in the central scenario described in chapter 7.3.3, and shown in table 8.2.

The present values in the lower and upper VfM scenarios reported in table 8.2 provide the range of the value to society of additional economic output in the overall quantified VfM assessment, as shown in table 8.1.

Table 8.2: Present value of additional output based on labour market conditions

Year	Lower VfM scenario - £ billion of additional output at fiscal year 2020/21 prices	Central VfM scenario - £ billion of additional output at fiscal year 2020/21 prices	Upper VfM scenario - £ billion of additional output at fiscal year 2020/21 prices
2020	2	3	5
2021	34	41	46
2022	5	13	21
2023	0	2	9
2024	0	0	5
2025	0	0	3
2026	0	0	2
Present Value Total	41	58	91

Source: HM Treasury estimates

Note: Chapter 7.3.3 describes the derivation of the no-CJRS scenarios.

8.2.2 Effectiveness of the CJRS in protecting jobs, businesses, incomes and spending throughout the economy

Although output has been used to value and monetise the majority of the CJRS’s effectiveness, the effectiveness criterion can also be assessed in terms of its impact on jobs, businesses and incomes.

The CJRS supported employers to retain labour by subsidising one of their largest costs (wages) whilst economic activity was suppressed, and revenues weakened. When restrictions eased, employers could bring back employees without the delay and cost of hiring. By subsidising wages and preserving jobs via employers, the CJRS also supported household incomes, meaning households could continue to consume from open sectors. This helped unemployment to remain low in these sectors. The CJRS directly protected around 4 million jobs that would otherwise have been lost and it supported the creation of new jobs for employees to reallocate into.

The scheme also supported incomes and thus consumption and jobs in the wider economy. It is plausible that being placed on furlough prevented individuals from undertaking productive work, and that some individuals would have returned to work at a faster pace without the CJRS – although this was unlikely, certainly during the earlier stages of the pandemic. However, as evidenced in chapter 7, analysis found limited evidence of the CJRS restricting this productivity enhancing reallocation of labour.

The CJRS also directly supported employers. The pandemic resulted in a significant shock to supply chains and employers’ activity and balance sheets. Without the provision of support, many employers may have permanently closed, leading to redundancies, a loss of income for many employees and their households, and thus reduced prospects for a rapid economic recovery once restrictions had eased. As reported in chapter 6, research suggests that close to 250,000 employers would have permanently ceased trading over the lifetime of the scheme. However, the CJRS may have kept potentially inefficient businesses afloat and reduced the expansion of new, more productive employers.

The value of preventing business failures is not specifically incorporated into the VfM assessment, as these are largely incorporated into the estimates of increased direct economic activity from protecting jobs and therefore its potential impact on economic output. Additionally, as reported in chapter 6, the CJRS enabled businesses to save on their rehiring and new training costs, given that the jobs of their existing employees were protected. This will have indirectly protected some businesses from closure. These hiring and training costs saved are not included in this VfM assessment, as they are assumed to be incorporated within the economic output estimate.

As well as directly protecting jobs and businesses, the scheme also had an indirect and wider effect on the economy. As shown in chapter 7, the CJRS’s major macroeconomic benefit was to ensure a quick economic recovery by preserving job matches, preventing falls in income and supporting household spending. This ensured the loss to output was significantly less compared to a scenario with no scheme and created an additional boost to the economy.

Through preserving support to incomes, the spending resulting from the CJRS support had positive macroeconomic effects, as outlined in chapter 7. These effects are difficult to assess and there is wide evidence in non-pandemic times of different types of multipliers associated with different types of spending. In the November 2020 EFO, the OBR used a 15% fiscal multiplier in the first year of the CJRS, meaning that every £1 spent, resulted in a £0.15 increase in output. This is significantly lower than a normal multiplier of 60% because supply was constrained by the NPIs put in place. However, as restrictions eased, households were able to begin resuming consumption at an increased level due to pent up demand. Therefore, the multiplier effect would tend back to its more normal rate across the period, and may have been larger than normal as a result of pent up demand. Whilst these impacts have not been monetised for this VfM assessment, they provide further evidence on how the CJRS is likely to have supported the economy.

The CJRS may have also produced health and wellbeing benefits to individuals on furlough and the wider population. Without the support of the scheme, employers or individuals may have been more inclined to put themselves or others at greater risk of catching COVID-19 by travelling to work or not self-isolating when advised to do so to keep earning. The scheme may also have produced wider wellbeing benefits by providing stability and limiting increases in household debt and payment arrears.

As highlighted above, the central total present social value estimate is £58 billion, with a range of £41 billion to £91 billion in additional output. An adjustment is then made to account for the CJRS’s preservation of economic output on an exchequer basis. This is estimated by the tax revenue which is generated from the additional economic activity. The exchequer benefit is calculated as 33.5% of the of the social benefit, which is broadly in line with the current UK tax to GDP ratio. Therefore, the exchequer value of the additional output from protecting the economy from economic harm has a central estimate of £19.4 billion with a range of £13.7 billion to £30.5 billion.

Overall, the evidence shows that the CJRS was highly effective, across multiple dimensions of the economy, although its effectiveness decreased over time. It had a significant social benefit in particular, and this does not account for the fact there are likely to be significant unquantifiable benefits.

8.3 Equity and equality

Equity and preventing increased inequality were not explicit goals of the CJRS, but as set out in chapter 8.1 it can still be considered as part of the scheme’s VfM assessment. While not explicitly captured in this assessment the CJRS was found to have supported the employers most affected by the impacts of the COVID-19 pandemic, as well as the perceptions of both employers and employees generally finding the scheme to be fair. Overall, this provides further evidence of the equitable and equality impacts of the CJRS.

The CJRS aimed to support all employments at risk due to the pandemic. While it was at employers’ discretion who to furlough, the policy design included a cap on individual support implemented within the scheme’s rules. It is therefore an important aspect of the VfM assessment to demonstrate that these measures were effective in supporting the population fairly, targeting support to those in higher need.

This support creates a benefit to society. However, these equity effects do not produce a direct exchequer benefit and are a social benefit only.

8.3.1 How the CJRS supported income equality

The CJRS directly contributed to increased household incomes for those who were on furlough, and therefore helped maintain living standards during the pandemic. Due to its broad eligibility criteria, the scheme overall supported a significant proportion of working households, especially during the first lockdown, although the limits on claim amounts capped the support received by those with the highest earnings. The CJRS had a broad effect across the household income distribution, particularly benefitting middle income households. The scheme was targeted at employees, therefore workless households did not benefit. However, ongoing welfare support and other temporary support such as the Universal Credit uplift was available. CJRS payments were capped at £2,500 per month. At an individual level, the scheme was progressive, with internal analysis of RTI data finding evidence that employments on furlough had lower average incomes compared to the total population. This analysis includes all employments for individuals that were eligible for the CJRS between March and October 2020. The mean wage of an employment on furlough in 2020 to 2021 was £16,602 while for the total population it was £23,648.

The Green Book values benefits to lower incomes more highly than benefits to higher incomes. To quantify the benefits of lower inequality it is necessary to estimate the marginal utility of income (the value of an extra pound of income to an individual which increases as income levels decrease). This means that the bigger the difference in income of CJRS recipients relative to average income, the greater value the scheme had from the perspective of reducing inequality. This benefit is then multiplied by the net spend of the scheme. This means not accounting for the component that is returned to the exchequer through Income Tax, National Insurance contributions (NICs) and savings to Universal Credit, had the scheme not existed. This analysis is detailed in the CJRS final evaluation accompanying technical information document.

This analysis produces a total equity benefit to society of £6.0 billion to £6.2 billion. As set out, these equity effects do not produce a direct exchequer benefit.

8.3.2 Equalities impacts of the CJRS

Although not quantified in the overall VfM assessment, the impact of the CJRS on protected characteristics is reviewed as recommended under the 2010 Equalities Act. The protected characteristics HMRC CJRS data contains are age and sex, both of which have been examined extensively in official statistics published on the scheme. While the CJRS was always intended to support those in employment, other support was put in place at the start of the pandemic for other groups including the self-employed and benefit recipients.

As evidenced in chapter 2, the CJRS had benefits across the regions and nations of UK, with the CJRS supporting businesses from the impact of the COVID-19 pandemic on sales, turnover, operating status and prospect of making job losses. Men and women benefitted relatively equally. The scheme tended to benefit younger employees at the start, but over time older individuals formed the largest group of beneficiaries.

The analysis below in figure 8.1 and figure 8.2 focuses on the additional protected characteristics of disability and ethnicity. The evidence suggests that there are no statistically significant differences, as shown by the uncertainty ranges in the charts below, in usage of the CJRS by different ethnic groups or by disabled individuals relative to the total population, with the exception of ‘Other Ethnic Groups’.

Figure 8.1: Usage of the CJRS by different ethnic groups

Table 8.3: Usage of the CJRS by different ethnic groups

Ethnic group	Usage Rate (percentage)	Uncertainty range - upper limit	Uncertainty range - lower limit
White	16.51	17.64	15.39
Asian/ Asian British	13.45	17.97	8.94
Black/ African/ Caribbean/ Black British	12.96	19.01	6.91
Other ethnic group	7.98	14.59	1.37
Total	16.13	17.18	15.07

Source: Family Resources Survey, 2021 to 2022

Note: Mixed or multiple ethnicities are not included in this analysis due to their small sample size in the data.

Figure 8.2: Usage of the CJRS by disability status

Table 8.4: Usage of the CJRS by disability status

Source: Family Resources Survey, 2021 to 2022

Disability status	Usage Rate (percentage)	Uncertainty range - upper limit	Uncertainty range - lower limit
Disabled	18.43	21.45	15.41
Not disabled	15.77	16.90	14.64
Total	16.13	17.18	15.07

8.4 Economy and costs

The economy element of measuring VfM considers the cost of the scheme, going beyond the headline exchequer cost reported in terms of the gross spend on the scheme. The Green Book emphasises that economic appraisals should be performed relative to the ‘Business as Usual’ (BAU) option, and so should account for how the Exchequer would have been impacted had the CJRS not been made available. Therefore, the cost of the scheme here reflects potential savings from additional Universal Credit benefit payments having been avoided as a result of protecting jobs.

For a comprehensive CJRS VfM assessment, other costs have been considered including the cost to HMRC of delivering the scheme (for example the cost of additional staffing or IT infrastructure), the cost of increased government debt and the need to service this, and the additional tax revenue from increased economic activity supported by the pandemic.

8.4.1 Cost of CJRS grants (net and gross tax)

The CJRS was a significant intervention in the labour market. Around £69 billion was paid out in grants through the CJRS, which is reflected as the main gross cost in the exchequer basis method. In their March 2022 EFO, the OBR also published an estimate of CJRS grants that excluded the £15.4 billion in Income Tax and NICs due on these payments, resulting in a net exchequer cost of £54 billion, which is set out in figure 8.3 below.

However, the social cost is more relevant from an economic and VfM perspective. As CJRS payments are transfers from the Exchequer to wider society, there is directly no cost to society of these grants being paid (aside from implementation costs and error and fraud) – this is the standard Green Book treatment for taxes and subsidies. Alternatively, there is a cost on society of this exchequer spending, which is discussed in more detail in chapter 8.4.4.

Figure 8.3: Monthly gross and net cost of the CJRS

Source: OBR March 2022 EFO

Accruals basis, £ billions	Private sector CJRS payments	Net CJRS cost
Mar-20	2.2	1.7
Apr-20	10.1	7.8
May-20	10.2	7.9
Jun-20	8.1	6.3
Jul-20	5.6	4.4
Aug-20	3.4	2.7
Sep-20	2.0	1.6
Oct-20	1.5	1.1
Nov-20	3.1	2.4
Dec-20	2.9	2.2
Jan-21	4.0	3.1
Feb-21	3.9	3.1
Mar-21	3.6	2.8
Apr-21	2.7	2.1
May-21	1.9	1.5
Jun-21	1.4	1.1
Jul-21	1.1	0.9
Aug-21	0.8	0.6
Sep-21	0.7	0.5

8.4.2 Administrative costs of the CJRS

The administrative cost of the CJRS is estimated to be £85 million (£74.1 million in 2020 to 2021 and £11.4 million in 2021 to 2022 respectively). This covers staffing and IT costs, capital spending and other expenditure required to deliver the scheme, including post-scheme costs. As a proportion of the grants paid through the CJRS, this represents 0.1%. The opportunity cost, meaning the cost of delivering the scheme that could have been spent on alternative government policies or programmes, is defined as a social cost as well as an exchequer cost.

8.4.3 Compliance recoveries and Universal Credit savings

As reported in chapter 3, through compliance activities by the Taxpayer Protection Taskforce (TPT), they recovered £256.1 million in total from overpaid CJRS grants across 2021 to 2022 and 2022 to 2023. This is in addition to the £518.8 million recovered prior to the taskforce being established. Furthermore, there were over £1 billion in unprompted disclosures and voluntary repayments from claimants where they identified an overpayment of a CJRS grant or if they wanted to voluntarily pay the grant back, as they no longer required it.

Had the CJRS not existed, there would have been higher unemployment and lower hours worked for some households on lower incomes. Thus, there would have likely been a bigger than observed increase in Universal Credit (UC) claims. By preventing individuals from losing their jobs, this in turn saved government benefit expenditure via UC claims, which can be offset against the exchequer cost of the scheme. Given this is a transfer of resources, as per Green Book guidance, it is only counted as a benefit of the scheme within the final VfM quantitative calculation on an exchequer basis and does not feature in the social value basis.

UC claimant amounts are based on household income levels, and the circumstances of that household; not all households will also be eligible for UC. Analysis produced by DWP using the Policy Simulation model, found that across 2020 to 2021 and 2021 to 2022, accounting for varying take-up rates and average UC payments of around £620 a month, the CJRS saved around £8.2 billion with a range of £6.3 billion to £9.8 billion in extra UC savings, as a result of the millions of jobs directly protected by the CJRS. This analysis is detailed further in the CJRS final evaluation accompanying technical information document.

As well as the direct exchequer saving of reduced UC benefit expenditure, the government will also have benefited from reduced administrative costs as a result of not having to process as many UC claims due to individuals receiving support from the CJRS instead. However, an estimate of this operational cost is not available, though it is expected to be small relative to the overall costs and benefits of the scheme, and so this is not included in the VfM assessment.

8.4.4 Cost of increased national debt and raising taxation

Established guidance set out in section 5.4 of the Green Book does not recommend accounting for the increase in government debt and the cost of servicing this as government spending is normally set out and accounted for at fiscal events, with individual policy decisions made within a fixed fiscal envelope. However, in March 2020, the CJRS created an unprecedented increase in government borrowing that was not initially part of the government’s spending plans. As borrowing creates an additional cost to the Exchequer in the form of debt interest, borrowing costs between March and the next fiscal event in July 2020 are accounted for, for completeness. Beyond this, it is assumed that in line with the standard Green Book guidance, the CJRS was funded through the usual fiscal event cycle.

Additional borrowing can lead to a higher cost of servicing government debt, as well as a future reduction in other public spending, or an increase in tax to reduce the debt. Increased government debt and tax rises simply represents a transfer between present and future taxpayers and so is not a social cost. When accounting for this on an exchequer basis, the OBR March 2020 Economic and Fiscal Outlook supplementary expenditure fiscal tables suggested that in 2020 to 2021, increasing national debt by £5 billion would result in an extra debt servicing cost of £0.1 billion. As shown in chapter 8.4.1, the CJRS cost around £54 billion, after accounting for tax and NICs, and is therefore assumed that it would result in a £1 billion cost of servicing debt. This is a simplified (and conservative) estimate factored into the exchequer costs, given the difficulty of estimating debt-servicing costs for a specific in-year period.

8.4.5 Social cost of exchequer finance

For the social value calculation, there is also a cost to society of raising additional funds. All increases in government spending are financed by either borrowing, taxation or reduced spending elsewhere. Theory suggests a reduction in economic efficiency will arise from the transfer of resources from the private to public sector. These costs are important to consider given the macroeconomic impact of the CJRS and its broader objective to support the wider economy. This variable is sometimes referred to as the Social Cost of Exchequer Finance. DWP’s social cost benefit analysis framework suggests this is set at 20%, therefore for every £1 raised, the economy is supressed by 20 pence. Applying this factor to the net exchequer cost of the scheme is estimated to increase social costs by £10.7 billion.

8.5 Efficiency

Efficiency is concerned with whether resources were well used to reach the intended population. There are 2 main components accounted for in this VfM assessment:

• error and fraud: non-compliant claims do not contribute to the scheme achieving its objectives yet increase the cost to the exchequer
• deadweight: can be thought of as the proportion of spend allocated to delivering desired outcomes that would have happened even in the absence of the policy intervention. In the circumstances of an unprecedented pandemic, what would have happened in the absence of the intervention is a difficult concept to measure

All schemes are likely to have some form of error and fraud and deadweight, and so it is important to consider how well the CJRS minimised them. The CJRS was designed to prevent as much error and fraud as possible, whilst making payments as quickly as possible to support employers and employees.

A wide range of options were considered to target the CJRS and minimise deadweight. This led to changes to the eligibility criteria over time, including the tapering of government support. Other targeting options were not used where these would have meant unfairly excluding customers genuinely in need, creating additional avenues for fraud, or where they would not have been possible to deliver as emergency support, for instance because of limitations in data availability or challenges in defining populations.

On the exchequer method, error and fraud and deadweight are already included within the CJRS payments, and so are not additional exchequer costs. Conceptually, both error and fraud and deadweight are transfers and so would not ordinarily be included as social costs as per the Green Book. However, grants that were claimed in error or fraudulently may not have been used for socially beneficial activities and could have negative consequences. Therefore, an estimate of error and fraud is incorporated into the VfM assessment as a social cost of the scheme.

Deadweight payments represent spend that may not have contributed to the objective of the CJRS or gone to those who needed the support, but do not necessarily have a social cost either. Deadweight payments means that the scheme was not perfectly optimised, as they increase the exchequer cost of the scheme and add to the social cost of exchequer finance without necessarily creating benefits, but they do not add to the social costs directly. Therefore, an estimate of deadweight is not incorporated into the VfM assessment as a social cost of the scheme.

Although not monetised within this VfM assessment, an additional factor of efficiency is the timeliness of a policy intervention, especially at a period of significant uncertainty like the COVID-19 pandemic. As evidenced in this evaluation, 6.6 million CJRS payments were made between the scheme’s opening date and October 2020, of which 99% were paid within 6 working days. As well as this, employers were very satisfied with the scheme, with 9 out of 10 employers and agents satisfied with the scheme’s timeliness and clarity.

8.5.1 Error and fraud

As reported in chapter 3, the scheme was subject to some erroneous claims and targeted by fraudsters. HMRC put in place safeguards to prevent this and continues to recover incorrectly claimed grants through compliance activity as shown in the HMRC annual report and accounts. The final most likely estimate of CJRS error and fraud is 5.1% with a range of 3.0% to 7.8%. In monetary terms, this equates to a CJRS error and fraud most likely estimate of around £3.5 billion with a range of £2.0 billion to £5.4 billion. While difficult to draw a direct comparison, for context the UK tax gap in 2021 to 2022 was estimated to be 4.8%.

8.5.2 Deadweight

Deadweight can be considered in different ways in VfM assessments. It can be simply thought of as desirable outcomes that would have taken place even in the absence of the policy intervention. The degree of deadweight can often be revealed by considering what the (BAU) outcome would have been. In the circumstance of an unprecedented pandemic, BAU is difficult to measure. The importance of comprehensively assessing the extent of deadweight was highlighted by the NAO in their 2022 report on the employment support schemes and the departments have committed to assessing deadweight in this evaluation. To assess the deadweight of the CJRS it is important to consider the scheme’s objectives, as set out in chapter 1.5.2.

Introduction to deadweight

The primary objective of the CJRS was to protect jobs and prevent widespread unemployment as a result of the pandemic. It had secondary objectives of reducing the risk of permanent employer closure and benefiting the wider macroeconomy by reducing the risk of economic scarring. Considering potential deadweight of the CJRS is concerned with whether the scheme could have achieved these same objectives but at a lower cost.

Given a secondary objective of the CJRS was to provide macroeconomic support, even CJRS payments claimed by employers who may not have required them to retain their employees, will have contributed to the success of the scheme by helping to stimulate additional economic activity. Therefore, the discussion below may arguably over-estimate deadweight by failing to account for the wider benefits of the scheme. Without the CJRS, the economy may have entered a negative downward spiral where the initial impact of COVID-19 created further business closures and redundancies, as suggested through the macroeconomic impact evidence presented in chapter 7.

It is important to note that nearly every policy intervention will contain some element of deadweight. At the time the CJRS was introduced there was uncertainty on the path of the COVID-19 pandemic and how the economy would be impacted, both initially and over a longer time horizon. As acknowledged throughout this evaluation, to account for this uncertainty the CJRS needed to have broad coverage and be practical in terms of how the data HMRC collects for tax purposes could have been used in devising and implementing the scheme. Restricting grant amounts or eligibility (for example on characteristics like sector) and therefore drawing the eligibility rules too tightly risked not supporting businesses who needed the grants, running the risk of deterring valid claims.

Additional complexity may have provided employers with less certainty and delayed payments, or risked putting off eligible employers in need of support from claiming, potentially resulting in further economic harm. However, the scheme did evolve over time to reflect changing economic circumstances, for instance with the introduction of flexible furlough and the increases in employer contributions.

Other targeting options were not used where these would have meant unfairly excluding customers genuinely in need, creating additional avenues for fraud, or where they would not have been possible to deliver as emergency support. Targeting at specific sectors most affected by COVID-19 was explored in order to improve VfM and reduce deadweight. However, a tapering approach was used instead due to limitations in turnover data, and the challenges in defining a sector and how impacts are not solely defined by sector given the inter-related nature of the economy.

Possible deadweight components

An overly simplistic estimate of deadweight would be the total amount of CJRS grants paid out to employers for individuals who did not have their job directly protected, based on the around 4 million jobs directly protected estimate presented in chapter 4. However, this interpretation of deadweight is incomplete as it implies that the remaining 7.7 million employments did not require any form of support at all. Whilst some individuals on furlough may not have required the scheme to protect their jobs, these employees may have faced reduced hours or pay, thus affecting their income, as a result of businesses having to scale back operations further if the scheme was not in place, as evidenced in chapter 5 and chapter 6.

Furthermore, the CJRS provided businesses with a safety net and certainty at a time of considerable unpredictability (see chapter 6 and chapter 7). Finally, as evidenced later when presenting the deadweight estimate section and referenced throughout chapter 2, the majority of these employments were in firms impacted by COVID-19. Therefore, whilst the employment may not have been protected by the scheme, they were still in need of support as the business was being negatively impacted. This explains why it is incorrect to assume that all the individuals on furlough who did not have their job directly protected amounts to deadweight, as having a job protected is not the only benefit to using the CJRS.

A further possible component of deadweight would be identifying those who used the scheme, but who did not have viable long-term job prospects. Jobs supported that ended shortly after the scheme closed might not be seen as needing support. However, analysis assessing the employment outcomes of those who used the CJRS indicates that there is little evidence of this as a significant risk. Analysis of the impact of the CJRS on jobs and the labour market set out in chapter 4.4 shows that outcomes were largely positive for furlough users, meaning there is limited evidence of negative consequences from providing furlough support long-term. This interpretation of deadweight does not therefore appear to be a particular concern in the context of the CJRS.

Potential deadweight may also be considered through assessing redundancies made while the scheme was in operation. As shown in figure 4.6, redundancies peaked at 14.5 per 1,000 employees in October 2020, potentially following the introduction of employer contributions. This may suggest that when employers needed to contribute towards the costs of placing their employees on furlough, some employees were made redundant. Although the scheme changes were likely a factor, it is also likely that there is a reporting lag from earlier in the pandemic. In addition, as shown in chapter 4.4.2, 83% of employees who left furlough were still in a form of employment after leaving the scheme. Furthermore, the CJRS objective of protecting jobs was achieved by providing certainty to businesses who are best placed to decide whether to continue placing employees on furlough, return employees to active work or make them redundant.

Deadweight calculation

This evaluation considers deadweight through examining if employers receiving CJRS grants were in need of financial support and in hardship. Whilst deadweight is difficult to measure, the employer research set out in chapter 2, shows that users of the scheme were materially impacted by the pandemic. Quantitative employer research covering March to October 2020, evidenced in the CJRS interim evaluation, found that 15% of employers claiming the CJRS experienced no reduction in turnover. Applying this 15% to the amount of CJRS spending to the end of October 2020 results in an estimate of around £6.5 billion.​ However, as noted by the NAO, this does not account for a large majority of these employers stating that they would have made redundancies or closed without the scheme​, and so is a comparatively narrow way to define those who may not have needed the CJRS. Therefore, a preferred definition is to account for firms who would not have made redundancies or closed permanently without the CJRS, and simultaneously saw their turnover stay about the same or increase.

This methodology has been extended and adjusted to incorporate new quantitative research evidence covering the full length of the scheme (March 2020 to September 2021). This builds on previous estimates, including in the NAO report, which covered the March to October 2020 period. Employers who were not actively trading at the time of the survey were not asked about the impact of COVID-19 pandemic on their turnover. For these employers, a question which asked them why they were not actively trading was used instead. Those who were not trading for reasons unrelated to the pandemic, and who satisfied the criteria relating to permanent closure and additional redundancies, have been included in the deadweight estimate. ​

The measure of deadweight used is based on cases in the survey where firms would not have made redundancies or closed permanently without the CJRS, and simultaneously either did not see their turnover decrease or were not trading at the time of the survey for reasons unrelated to the pandemic. This measure represents claims for around 555,000 jobs and equates to around 4.7% of employments that were claimed for through the full length of the CJRS. A confidence interval at the 95% level for the estimate ranges from 469,393 to 641,185 jobs, which is 4% to 5.5% of employments. The central estimate of 4.7% of employments is equivalent to £3.3 billion (95% confidence interval ranges from £2.8 billion to £3.8 billion) if representative of all jobs claimed for, indicating that some jobs may not have needed supporting.

Figure 8.4: Proportion of Users satisfying individual deadweight criteria

Source: CJRS employer and agent quantitative research wave 2

Base: Users answering all questions (4,545), actively trading Users (4,615)

Question: Which of the following best describes what would have happened to your organisation if you had not received funding from the CJRS?;The CJRS was available to employers between 1 March 2020 and 30 September 2021. During this time, to the best of your knowledge, would you organisation have made more employees redundant if the CJRS was not available, or would it have made no difference?; Between the start of the pandemic and when the CJRS closed (up to 30 September 2021), what impact, if any, had COVID-19 had on your organisation’s sales or turnover? Had it?; What is the current status of your organisation?; For what reasons would you say your organisation stopped trading?

Notes: Findings exclude “don’t know” and “not sure” responses.

*Employers actively trading at the time they were surveyed were asked whether COVID-19 had a negative impact on their turnover

Table 8.5: Proportion of Users satisfying individual deadweight criteria

Impact	Users (percentage)
Would not have closed permanently if they had not received the CJRS	75
Would not have made additional redundancies if they had not received the CJRS	38
Turnover increased or stayed the same due to COVID-19*	23
Not actively trading at the time of the survey for reasons unrelated to COVID-19	7

Figure 8.4 shows the proportion of Users who satisfied each of the deadweight criteria that were used to produce the final deadweight estimate (around 4.7% of employments that were claimed for through the full length of the CJRS may not have needed support). This chart looks at each of these factors in isolation, for example 75% of Users said they would not have closed permanently if they had not received the CJRS. However, 23% of these employers would have closed temporarily and many would have made additional redundancies or saw a reduction in turnover, which indicates that they were still impacted by the pandemic and therefore in need of support. This highlights how assessing one possible business outcome in isolation masks the various impacts felt by employers as a result of the pandemic, reinforcing the need to take a combination of outcomes in assessing deadweight.

Further considerations

As noted above, this estimate still does not account for the additional and indirect macroeconomic boost provided by the CJRS support and so is likely to be an over-estimate of the deadweight of the CJRS.

Whilst error and fraud and deadweight are considered as 2 separate components of inefficiency in the VfM assessment, there may be some overlap between them. Employers who claimed for the CJRS in error or claimed fraudulently may also be employers who did not need the support provided by the CJRS, as they were not affected by the impacts caused by the pandemic. However, it is not possible to robustly calculate the amount of overlap, given the independent and distinctly separate methodologies used.

8.6 Overall value for money assessment

As illustrated in figure 8.5, table 8.6 and table 8.7, the CJRS is shown to have been good value for money and had a substantial net social benefit overall, despite the high costs to the Exchequer. Overall, it provided a total central social benefit of £67 billion with a range of £50 billion to £100 billion derived from the value of additional economic output, equity benefits through reducing income inequality and NICs and Income Tax due on grant payments. The CJRS is seen to have a total central social cost of £17 billion with a range of £16 billion to £19 billion from the gross cost of grants, the cost of implementing the scheme, error and fraud, and the deadweight measure on inefficient payments. This gives a high central net social benefit of £50 billion with a range of £34 billion to £81 billion in present value terms. This indicates a benefit: cost ratio of around 4:1 and that the CJRS was a successful and important intervention to support the UK labour market during COVID-19.

This assessment also sets out the exchequer valuation that shows the scheme had a lower effect on the public finances than the ‘headline’ £70 billion of grants and other costs, with a net central exchequer cost of £25 billion with a range of £13 billion to £33 billion. This is a conservative estimate, reflecting the stronger labour market that the macroeconomic analysis assumes. It is possible however that with labour market conditions closer to average, benefits were significantly higher on both an exchequer and social value basis.

Figure 8.5: CJRS overall value for money assessment, central estimate using social method

Notes: The shorthand [low] is to denote a low figure which is non-zero given the rounding to the nearest £ billion. Costs are denoted as negative values.

Table 8.6: CJRS overall value for money assessment using social method, broken down by financial years with lower, central and upper estimates

Type	Benefits and costs	Financial years	Lower value, £ billions	Central value, £ billions	Upper value, £ billions
Effectiveness	Preservation of economic output	2020/21 to 2025/26	41.0	58.0	91.0
Equity and equality	Equity benefits	2020/21 to 2021/22	6.0	6.2	6.2
Economy	Gross cost of grants	2020/21	-12.1	-12.1	-12.1
Economy	Gross cost of grants	2021/22	-1.7	-1.7	-1.7
Economy	Tax and NICs	2020/21	2.7	2.7	2.7
Economy	Tax and NICs	2021/22	0.4	0.4	0.4
Economy	Cost of implementation	2020/21	-[low]	-[low]	-[low]
Economy	Cost of implementation	2021/22	-[low]	-[low]	-[low]
Efficiency	Error and fraud	2020/21 to 2021/22	-2.0	-3.5	-5.4
All Es	Total Benefits PV	2020/21	50.1	67.3	100.3
All Es	Total Cost PV	2020/21	-15.9	-17.4	-19.3
All Es	NPV	2020/21	34.2	49.9	81.0

Notes: The shorthand [low] is to denote a low figure which is non-zero given the rounding to one decimal place. All values except the equity benefits have been deflated and then discounted to be present values for 2020 to 2021. Costs are denoted as negative values.

Table 8.7: CJRS overall value for money assessment using exchequer method, broken down by financial years with lower, central and upper estimates

Type	Benefits and costs	Financial years	Lower value £ billions	Central value £ billions	Upper value £ billions
Effectiveness	Preservation of economic output	2020/21 to 2025/26	13.7	19.4	30.5
Economy	Compliance recoveries	2020/21	0.5	0.5	0.5
Economy	Compliance recoveries	2021/22	0.2	0.2	0.2
Economy	Compliance recoveries	2022/23	0.1	0.1	0.1
Economy	Voluntary repayments	2021/22	0.9	0.9	0.9
Economy	Voluntary repayments	2022/23	0.1	0.1	0.1
Economy	UC savings	2020/21	6.0	7.8	9.1
Economy	UC savings	2021/22	0.3	0.4	0.7
Economy	Gross cost of grants	2020/21	-60.6	-60.6	-60.6
Economy	Gross cost of grants	2021/22	-8.4	-8.4	-8.4
Economy	Tax and NICs	2020/21	13.5	13.5	13.5
Economy	Tax and NICs	2021/22	1.9	1.9	1.9
Economy	Cost of implementation	2020/21	-[low]	-[low]	-[low]
Economy	Cost of implementation	2021/22	-[low]	-[low]	-[low]
Economy	Cost of servicing debt	2020/21 to 2021/22	-1.0	-1.0	-1.0
Efficiency	Error and fraud	2020/21 to 2021/22	-2.0 (included within Economy)	-3.5 (included within Economy)	-5.4 (included within Economy)
Efficiency	Deadweight	2020/21 to 2021/22	-2.8 (included within Economy)	-3.3 (included within Economy)	-3.8 (included within Economy)
All Es	Total Benefits PV	2020/21	37.3	44.8	57.4
All Es	Total Cost PV	2020/21	-70.0	-70.0	-70.0
All Es	NPV	2020/21	-32.8	-25.2	-12.6

Notes: The shorthand [low] is to denote a low figure which is non-zero given the rounding to one decimal place. All values except the equity benefits have been deflated and then discounted to be present values for 2020 to 2021. Costs are denoted as negative values.

Chapter 9: Final evaluation conclusions

This evaluation has shown that the CJRS was an important policy in mitigating the negative impacts of the pandemic on the UK labour market. Indeed, such was the scheme’s breadth and scale, it proved to be a major benefit to the wider economy and have a high central social value for money (VfM) of £50 billion using conservative assumptions about the impact on the economy.

Throughout the scheme’s existence, a range of policy choices were considered regularly to maximise VfM, including means to target support, incentivise a return to work, and when to unwind support. This evaluation has shown the government’s decisions across the scheme achieved a good return for the taxpayer. The government has also learned valuable lessons from the experience of the CJRS, including:

1.processes for how it manages future crises: as evidenced in the interim evaluation, the successful delivery of the CJRS at pace was enabled by focussing on a simple and deliverable design initially to address the high public uncertainty and severe risk to jobs, delivering work in parallel, and then building and adding complexity over time. Additionally, the close working between departments (HMRC and HMT) from the start, on both policy and delivery, was a key aspect in ensuring a deliverable policy to be formulated, publicly announced, and rolled out at pace

2.policy design: as discussed in chapter 4, the introduction of flexible furlough and tapering of government contributions had positive impacts, helping to incentivise employers to scale up activity as NPIs evolved and ended. Such design choices should be considered in the development of future employment support schemes. In addition, CJRS payments were made in advance which had a range of benefits. However, this could have led to some error in claims, due to changes in business’ plans and activities. In future, both payment in advance and payment in arrears should be considered, to best meet policy objectives

3.policy targeting: while targeting at specific sectors most affected by COVID-19 was explored in order to improve VfM, a tapering approach was used instead. Reasons included limitations in turnover data, and the challenges in defining a sector and how impacts are not solely defined by sector given the inter-related nature of the economy. Furthermore, while possible to explore sectors and regions to view where impacts are being felt, it is more difficult to then determine within that who is directly in need of support. In general, fast and broad-based support does risk supporting individuals who do not require support, while more targeted supported generally is slower, is more complex with lower take-up, and means some may not have got the support they need

As discussed in chapter 4, the impact of the CJRS was greatest in its early stages, demonstrating increasing benefits from targeting in later stages of the scheme. More consistent and up-to-date data would have enabled greater flexibility, which would be useful for the future targeting of labour market policies. The government has recently published a response to its consultation on options for improving the range of data HMRC collects, uses and responsibly shares across government. In designing and targeting future policies, consideration should be given to data available at the time

4.management of error and fraud: chapter 3 demonstrates that policy and operational design choices successfully achieved an appropriate balance between getting support to employers quickly, and managing the risk of error and fraud. The digital repayment service worked well, and future schemes should consider providing a similar service from the outset where appropriate

5.supporting customers: the interim evaluation highlighted high levels of customer satisfaction with the CJRS processes, and showed the importance of customer contact channels in supporting customers to understand a new scheme at a time of uncertainty. In particular, the claims process was designed to be simple, and resulted in high levels of customer satisfaction with the application process. The process evaluation also highlighted some feedback and insights gained while the CJRS was operating. These led to improvements in the claims process and guidance, as the scheme progressed

The NAO recommended that the government should consider the trade-off between error and fraud and deadweight more explicitly in its policy design.

In the short term, the scheme was created and announced in days, prior to the first nationwide lockdown, to reassure businesses struggling with cashflow issues that support would be available. The CJRS was delivered successfully within a month, achieving a reasonable balance against the risk of error and fraud. This enabled the scheme to have a major impact in preventing a significant spike in unemployment and business failures in spring 2020. It also protected the incomes of many households who would otherwise have had to rely on benefits. Evidence presented throughout this evaluation suggests that the CJRS was targeted to support those most in need.

In the medium term, as the economy reopened for business, the CJRS and subsequent policy design changes during spring and summer 2020 enabled a quick economic recovery. Future adjustments to react to the evolving severity of the pandemic over autumn and winter 2020 to 2021 meant that the scheme continued to be effective in protecting jobs and businesses. The timing of the scheme’s closure was a carefully judged decision, based on an uncertain outlook. However, the gradual winding down of support was viewed favourably by employers and timed to coincide with the ending of restrictions. The CJRS has also had a sustained benefit on the UK economy, reducing scarring and minimising the economic harm caused by the COVID-19 pandemic, with economic output higher than it would have been without the scheme. Additionally, the observed increase in inactivity, especially among older workers, is largely independent of whether they were placed on furlough or not. The increase in inactivity is likely to reflect a number of other factors, such as health conditions, lifestyle choices about the age of retirement and the impacts of the COVID-19 pandemic more generally. Overall, this evaluation demonstrates that the CJRS achieved its ambitious objectives with considerable net benefits to society.

---

---

1. GDP increased at an average annual rate of 2.1% between 2014 and 2019, so this is equivalent to around 10 months of pre-COVID-19 GDP growth. ↩

2. Between the start of the CJRS and when employers were surveyed between June 2022 and November 2022. ↩

3. Downstream yield refers to tax receipts generated by activities undertaken by HMRC after non-compliance has occurred. ↩

4. This figure includes claims that were made across all of the COVID-19 schemes. ↩

5. October 2020 and November 2020 have been merged as the reference month used in this analysis. This is due to the timing of the announced scheme extension at the end of October and the impact this had on the data used to construct the counterfactual. ↩

6. Each furlough period coincides with an announced extension of the scheme, which is exploited in the counterfactual methodology to derive the estimates presented. ↩

7. The 10.6 million referenced in this analysis is not comparable to the CJRS official statistics as some individuals are excluded from the labour market outcomes analysis. ↩

8. Individuals who were on Universal Credit in at least one assessment period ending between November 2021 to January 2022 were included. ↩

9. It should be noted that the furlough and non-furlough population have different employment characteristics and demographics. Therefore, this comparison is only used to contextualise the furlough population findings. ↩

10. Individuals are counted as being on Universal Credit during this period if they had at least one assessment period ending between 1 November 2021 and 31 January 2022 in which their claim had not closed. UC assessment periods last one month and individuals typically receive their payment 7 days after this. Individuals are not excluded from these figures if their UC award was zero but their claim had not yet closed. The number counted as not working and not required to intensively seek work is a count of those who were not working and not in the ‘Searching for Work’ conditionality group in any month during the period. ↩

11. It should be noted that the furlough and non-furlough population have different characteristics. Therefore, this comparison is only used to contextualise the furlough population findings. ↩

12. The NICs secondary threshold was £732 per month in 2020 to 2021, Rates and allowances: National Insurance Contributions (see table 1.2). ↩

13. The analysis assessed characteristics such as age, gender, and income. A full breakdown of the characteristics assessed can be found in the accompanying technical information document. ↩

14. This proportion does not sum to the values shown in figure 6.1, due to the rounding of percentages. ↩

15. The chances of remaining in employment are also strengthened by employees having the option to look for alternative employment while remaining in their existing one. ↩

16. Source: ONS Labour Force Survey flows estimates. The “typical” estimates are median values observed between 2002 and 2022. ↩

17. This is discussed in the analysis of labour retention costs saved in the CJRS final evaluation accompanying technical information document. ↩

18. In addition, the probability of unemployed people becoming employed tends to decline as the time spent in unemployment increases. ↩

19. The transition probabilities indicate the proportions of people who move between each of 3 labour market states (employed, unemployed or inactive) in a given calendar quarter to each of the 3 states in the following quarter. ↩

20. There is no extraordinary change to the employment to unemployment flow in 2020 Q4, reflecting the easing of restrictions in the second half of that year. ↩

21. The Labour Force Survey data has 85 quarterly observations for each transition rate, from 2001 Q4 to 2022 Q4. The decision to use the 95th percentile was informed by the strength of the US recovery in labour demand. ↩

22. In the lower value for money scenario the transition probabilities into employment in 2021 Q2 from each of employment, unemployment and inactivity to values observed at the 98th percentile of the distributions of quarterly transition rates into employment in the Labour Force Survey data; and in the upper value for money scenario to values at the 93rd percentile of the distribution. ↩

23. Outcomes for transition rates between employment and non-employment in the United States in 2020 and 2021 provide a benchmark for how such rates could have evolved if the CJRS had not been in place. The USA made limited use of job retention schemes. Hiring rates out of non-employment were around 20% higher when restrictions were eased there than is typically the case. That supports the assumption that flows into employment would have been unusually strong when the economy reopened in 2021 Q2. ↩

24. Figure 7.7 shows that it takes time for employment levels in no-CJRS scenarios to close the gap with employment in the with-CJRS observed outcomes. By contrast, figure 4.1 illustrates the impact of eligibility for the CJRS, by comparing the difference between employment rates of similar individuals within the control and intention-to-treatment groups in the with-CJRS scenario. ↩

25. These calculations assume that the hours and pay of those remaining in employment would not have been different if the CJRS had not been in place. ↩

26. The OBR fiscal multiplier ready reckoner of 0.6 applied to CJRS spending of £70 billion implies support to private sector spending in the region of £42 billion, equivalent to 2% of annual GDP. Much of this would have boosted spending once the supply constraints, that were prevalent while restrictions on economic activity were in place, had eased. The impact of additional spending on labour demand would have contributed to increased flows from unemployment to employment that is a feature of the 3 no-CJRS scenarios. ↩

27. There were other ways that inhibiting the re-allocation of labour could have been sub-optimal. For example: if it inhibited job moves that would have enabled people to acquire new skills; or moves that would have allowed people to find a new job that better reflected changed personal circumstances, such as preferred number of working hours or location. ↩

28. Productivity is proxied by each firm’s average wage in the 2018 to 2019 tax year. High and low productivity firms are those with productivity 1 standard deviation above or below the mean firm-level productivity. ↩

29. The COVID-19-shock and productivity enhancing reallocation in Australia: real-time evidence from single touch payroll and COVID-19, productivity and reallocation: timely evidence from 3 OECD countries. ↩

30. This is consistent with the evidence presented in chapter 4 that indicates that the employment outcomes of those on furlough following the end of CJRS were generally positive. ↩

31. For further information on drivers of inactivity see Box 3.4 in Spring Budget 2023 document or section 7 in the ONS Employment in the UK, July 2023 publication. ↩

32. Other countries, notably the United States, made limited use of job retention or job subsidy schemes. The US supported household incomes through payments to households that were not tied to their remaining with their existing employer. ↩