COVID-19: Support for jobs

The age of the Self Assessment system made it more difficult for HMRC to provide financial support for the self-employed. Its tax system for the self-employed lags behind that available in other countries. The Self Assessment system was built in the 1990s, and its design and age limits the amount, timeliness and quality of the data that HMRC holds on the self-employed. HMRC had to cleanse the data in the system before it could use it, adding several weeks delay to introducing the SEISS. HMRC has an ongoing programme, Making Tax Digital, designed to upgrade its systems and permit users to provide information more frequently using digital devices—as opposed to the current annual tax return. Making Tax Digital should have been in place by 2019 but, apart from a pilot project, is now not due until 2023. By contrast, HMRC’s investment in the Real Time Information system in 2013 gave HMRC access to much more timely data about employees on the Pay-As-You-Earn system. Therefore, it was able to use this for the design of CJRS and base eligibility on more up-to-date information. Recommendation: HMRC should write to the Committee within three months to explain what it has learnt from its review of other countries’ self-employed systems and how it will apply these to its plans for delivering the Making Tax Digital programme.

The Departments have not done enough to reduce the number of people excluded from the schemes. The Departments still do not have a complete assessment of the number of people excluded from the first phase of CJRS and SEISS up to the end of October 2020, but the best estimate suggests that it may have been as many as 2.9 million workers. The extension to the schemes announced by the Chancellor in November was not accompanied by any substantial changes to the design of the schemes that would likely reduce the number of people who are excluded this time round. The more time moves on, the greater the potential number of newly self- 6 Covid-19: Support for jobs employed people that might be excluded from SEISS. The scheme continues to be based on 2018–19 tax return data submitted to HMRC by 31 January 2020—despite the fact that the 2019–20 tax year ended in April 2020 and many self-employed workers and employees who slipped through the net because of the short term nature of their employment contracts will already have submitted their 2019–20 tax returns. HMRC has also not considered whether it could use other data sources across government. Recommendation: HM Treasury and HMRC should investigate whether more data within and outside of the tax system could be used to determine eligibility for currently excluded groups and write to the committee within six weeks to explain their findings. HM Treasury and HM Revenue and Customs should liaise with departments which have a detailed knowledge of the affected sectors in order to improve access to Covid-19 related support schemes for currently excluded groups.

The Departments did not evaluate the schemes or identify which the groups they support before extending them. Both the schemes have been extended due to the prolonged impact of the pandemic, but the Departments have not yet produced evaluations of the initial CJRS and SEISS schemes. HMRC provides monthly updates on the cost of the schemes and the take-up by demographics such as age and gender. It does not, however, have data analysing take-up by protected characteristics such as race and disability. It does have some preliminary analysis that suggests 90% of furloughed workers return to their employer after furlough, but more needs to be done to track the medium- and long-term effects after furlough ends. Recent data shows the unemployment rate has now risen from 4.0% at the start of the pandemic to 4.8%, and the Office for Budget Responsibility (OBR) is projecting unemployment will rise to 7.6% in 2021. Recommendation: HMRC should, as soon as possible, develop and report monthly performance information on the schemes, such as take-up by protected groups and employment outcomes.

The Departments will not know the actual levels of fraud and error within these schemes until 2021. HMRC does not expect to have a statistical estimate of the total fraud and error levels across both schemes until the end of 2021. Whilst waiting for this estimate, there are other metrics already available that can begin to build a picture of the levels of fraud and error, such as the amounts of overpayments detected and recovered and the number of arrests and prosecutions due to criminal activity on the schemes. Levels of opportunistic fraud, where furloughed workers continued to work whilst companies claimed grants, could have occurred in between 7% to 34% of cases. HMRC also plans to publish which companies are accessing the extended CJRS and to directly notify employees when they have been furloughed, something it decided not to implement during the first phase of the scheme between March and October 2020. We welcome this suggestion that it is making some changes to try and reduce the opportunity for fraud to occur. Recommendation: HMRC should write to the Committee within three months outlining how it can utilise the information it already collects to better estimate the levels of fraud and error; and also outline what steps it intends to take to recover CJRS and SEISS grants made during the first phase of the scheme if recipients Covid-19: Support for jobs 7 made substantial profits or were not adversely affected by the pandemic. HMRC should list companies which have signed up to the furlough scheme by the end of January 2021.

Too much chopping and changing of the new schemes has created uncertainty for the UK nations, regions and businesses, regarding financial support and job security. Nations, regions and businesses, as well as their employees, need as much certainty as possible to allow them to plan ahead. Instead what they got in the autumn was a sequence of announcements, right to the very end of the end of first phase of the schemes, constantly adjusting the levels and availability of funding support for workers. The government announced on 5 November that both the CJRS and SEISS would be extended with levels of support broadly equivalent to the levels provided under the first phase of the schemes. We are concerned that businesses’ response to this initial uncertainly may have resulted in more workers being laid- off, even though eligibility for the government’s extended schemes was moved back to 23 September to cover such redundancies. Recommendation: The Departments should provide as much clarity and forewarning as possible about the employment support arrangements that will be available for UK nations, regions and businesses under conditions of national lockdown, regional lockdown and easing of restrictions for the remainder of the covid-19 pandemic. It should commit to this ahead of the Treasury minute response so employers can be clear that they can plan ahead with greater certainty.

We are concerned that HM Treasury is unable to explain how much the extended schemes are forecast to cost or what would constitute value for money. HM Treasury argues that it falls to the Office for Budget Responsibility (OBR) to produce forecast costs for the scheme extensions. However, HM Treasury is nonetheless responsible for providing assurance to Parliament of the value for money of schemes, of which costing is an important part. The OBR now forecasts that extensions to the schemes will cost the taxpayer a further £21 billion. It is also unacceptable that the Departments are unable to explain what constitutes value for money in these schemes. We accept that when the schemes were initially developed back in the spring the exceptional circumstances at that point made such analysis difficult. In the intervening period, however, we would have expected HM Treasury to have developed a more rigorous assessment of the costs and benefits of the schemes to ensure taxpayer money is spent wisely. Recommendation: HM Treasury should write to the Committee within a month to set out how it will assess value for money for the extended schemes. 8 Covid-19: Support for jobs 1 Implementing and evaluating the initial covid-19 employment support schemes

On the basis of a report by the Comptroller and Auditor General, we took evidence from HM Treasury and HM Revenue & Customs (HMRC) about the employment support schemes that they have established since the start of the covid-19 pandemic.1

HMRC used the data in its Self Assessment system to calculate how much self- employed people would receive as part of the SEISS grant award. However, this was developed in the 1990s and lags behind other countries’ systems. HMRC told us that with better quality data it may have been able to help more people and tailor the level of the grant award more to meet people’s individual circumstances. It accepted that it needed to get better quality data and that it was “definitely moving slower than the best countries around the world in resolving that”.15 HMRC conceded that poor data quality meant it had to do a fair bit of data cleansing before launching the scheme to ensure that the data was accurate. It suggested that if this hadn’t been required, it may have been able to open the scheme a couple of weeks earlier.16 The NAO found that technical issues with the Self-Assessment system also meant that HMRC had to manually enter the trading profits of around 800 self-employed people to prevent them from being excluded.17 HMRC has a programme in place, Making Tax Digital, designed to upgrade its systems and permit users to provide information more frequently using digital services. This should provide HMRC with better quality data going forward. This should have been in place by 2019, but apart from a small pilot project, is not now expected until 2023.18

HMRC’s investment in its RTI system meant that, unlike SEISS, eligibility for CJRS was based on more up-to-date data. The RTI system is used on a monthly basis by employers to submit tax information on their workforce to HMRC. HMRC said that this information was vital in enabling it to run the CJRS safely. It meant that any employee on an employer’s Pay-As-You-Earn system at 19 March 2020, the day before the first national lockdown, could potentially be furloughed.19 Similarly, with the latest national lockdown, eligibility for the CJRS extension is based on employees on employer payrolls on 30 October.20 The downside to using RTI data in CJRS is that employers could submit claims based on data they submitted after the scheme was announced, whereas the use of old self-employed taxpayer data in SEISS prevented that problem. This increased the risk that employers would deliberately misstate their positions to increase their CJRS grant awards.21 Evaluation of the initial schemes

The extension of both schemes was announced in November 2020. At that stage the initial schemes had not been formally evaluated. The Departments told us that they were undertaking informal evaluations to help them tailor their communications to those the schemes were aimed at. They explained that they would be undertaking full evaluations of all of the schemes, and that these would be made public, but that full evaluations were 14 C&AG’s Report, para 1.11 15 Q 27 16 Q 27 17 C&AG’s Report, para 2.11 18 Q 75 19 Q 10, C&AG’s Report, figure 5 20 HMRC, Extension of the Coronavirus Job Retention Scheme, available at: www.gov.uk/government/publications/ extension-to-the-coronavirus-job-retention-scheme/extension-of-the-coronavirus-job-retention-scheme 21 C&AG’s Report, para 3.7 Covid-19: Support for jobs 11 not due until “later next year”. HMRC expected that such evaluations would consider the value for money of the schemes; how individuals with protected characteristics were covered, and the ability of the schemes to meet the needs of different communities.22 The NAO found that, to date, HMRC’s evaluation of the schemes had largely focused on lessons learned around project management as its sought to refine its ongoing delivery. On the wider issue of whether the schemes will have prevented mass unemployment, HM Treasury told us it was too soon to tell, but that the government felt it was the right thing to do to extend the schemes.23

We asked whether the Departments had undertaken any evaluation on the regional differences in take-up or the schemes and take-up by groups with protected characteristics.24 HMRC currently reports monthly on the cost of the schemes and provides analysis on take-up by different demographics. The data for October 2020 showed that CJRS had cost £41.4 billion and the two SEISS grants had cost £13.7 billion.25 Take up for CJRS was largest amongst the wholesale and retail, and repair of motor vehicles, sectors, whilst in SEISS it was largest in the construction sector. It is also possible to analyse take-up by age and gender, as well as by parliamentary constituency for both schemes.26 HMRC does not, however, monitor take-up of the schemes by different ethnic groups as it is not required to collect that data as part of its regular tax administration. HM Treasury commented that it looked at Office for National Statistics (ONS) data to consider the potential effects of the schemes on unemployment, including regional unemployment.27 ONS’ unemployment data for July to September 2020 showed that the unemployment rate was 4.8%, up from 4.0% in March 2020.28 HMRC told us it had done some preliminary research about what happened to furloughed workers once they come off furlough, which suggested that around 90% of them return to their previous job.29 The medium to long-term projections on unemployment, however, are less encouraging; the Office for Budget Responsibility (OBR) forecast, released on 25 November 2020, suggests that unemployment might peak at 7.6%, or 2.6 million people, in 2021.30

HMRC’s initial planning assumptions suggested that the level of fraud and error would be 5% to 10% within CJRS and 1% to 2% within SEISS. We asked whether it expected the level of fraud and error to change under the new scheme. HMRC told us that it had originally estimated that around half of the fraud and error in the CJRS (2.5% to 5%) would potentially be due to fraud committed by organised criminal gangs, but it had subsequently revised that down to 0.6% due to the controls it had put in place. There were fewer controls in place, however, to tackle the risk of furlough fraud where, for example, employers furloughed workers but then continued to get them to work, against the rules 22 Qq 58–59 23 Q 61, C&AG’s Report para 1.17 24 Qq 58–59 25 HMRC, HMRC coronavirus COVID-19 statistics, available at: www.gov.uk/government/collections/hmrc- coronavirus-covid-19-statistics 26 C&AG’s Report, figures 11, 12 and 13; HMRC, Coronavirus Job Retention Scheme statistics: October 2020, available at: www.gov.uk/government/publications/coronavirus-job-retention-scheme-statistics-october-2020/ coronavirus-job-retention-scheme-statistics-october-2020 27 Q 58, C&AG’s Report, para 2.20 28 Office for National Statistics, Unemployment rate (aged 16 and over, seasonally adjusted), released 10 November

As many as 2.9 million people may have been excluded from the first versions of CJRS and SEISS. The NAO found that people were excluded either because of policy design choices or due to constraints in the tax system. An estimated 1.1 million people were excluded from CJRS because HMRC did not have sufficient data to verify claims, whilst on SEISS, 0.2 million missed out because HMRC did not hold enough data about their self-employment. A further 1.6 million were estimated to have been ruled out of SEISS due to not meeting the scheme’s criteria; for example, because they received less than half their income from self-employment.32 We asked the Departments what issues existed that meant people were unable to receive support and how they were working to resolve them. HMRC confirmed that the eligibility criteria for the extended schemes remained broadly the same as the original scheme. It acknowledged that the structure of these schemes meant that those with more casual working patterns were likely to have found it more difficult to qualify for the schemes than those with more regular working hours.33 It said that the schemes had been designed to help as many people as possible but that it simply wasn’t possible to help everyone.34

We were concerned that the extension to SEISS could leave more self-employed people without support than the initial scheme. HMRC based eligibility for the initial SEISS on tax return data up to 2018–19 and estimated, as part of initial planning at the start of lockdown, that this meant around 0.2 million newly self-employed people were unable to claim the grant. HM Treasury told us that it considered alternative arrangements, but that basing the grant awards on previously submitted data helped prevent the risk of people manipulating their returns to receive higher payments. HMRC confirmed that the extension to SEISS would continue to operate on the basis of 2018–19 returns despite HMRC accepting that the 2019–20 tax year ended in April 2020 and returns were starting to come in ahead of the 31 January 2021 deadline. The NAO noted in its report that the number of self -employed people excluded from the initial SEISS could have been greater than 0.2 million if lockdown occurred further from the annual tax return deadline. The SEISS extension occurred nine months after the 2018–19 deadline rather than the two months for the initial scheme. HMRC told us that if it had more up-to-date data on those who were the self-employed it may have been able to operate the scheme differently. Self- employed tax reporting on a quarterly basis to HMRC is due to come in as part of its Making Tax Digital project, but other than the pilot scheme, is not due until 2023.35

Two further groups that were largely excluded from support were freelancers and owner-managers of companies. Freelancers generally have short-term contracts with employers and as a result many might not have been on a company’s PAYE system at the cut-off point for furlough. HM Treasury said that was aware of the issues, but that the lack of data held by HMRC and the fact that the tax system wasn’t designed to regularly capture information on such people made it particularly difficult to bring freelancers into the scheme. Some 0.4 million freelancers may have missed out on support from the 32 C&AG’s Report, para 14 33 Qq 73, 79 34 Public Accounts Committee, Oral Evidence: HM Revenue & Customs 2019–20 Standard Report, HC 690, Q 13 35 Qq 73–75, C&AG’s Report paras 2.9–2.10 14 Covid-19: Support for jobs first schemes.36 We also asked about owner-managers of companies who may have been excluded. HMRC explained that owner-managers qualify for the furlough scheme to the extent that they pay themselves through their PAYE system. It acknowledged that many owner-managers opted to pay themselves a small amount in this way, and rely predominantly on dividend income. HMRC maintained, however, that the schemes were not aimed at supporting investors for the loss of dividend income and that it did not hold sufficient information from tax return data to be able to distinguish between what is investment income and what is income in lieu of salary. We asked HMRC whether other sources of government data could be used to verify the eligibility for the schemes from those groups currently excluded; for example, information held by Companies House might distinguish between different types of dividend income. It said that this was something it had looked at closely over the years from a tax perspective but had not found a way of distinguishing between the different types of dividend income.37 Changes to the schemes created uncertainty

The initial CJRS scheme was due to end on 31 October 2020. In September, the government announced that it would be replaced by a Job Support Scheme (JSS) that would top up the wages of workers working at least one-third of their normal hours, with the employer also contributing.38 HM Treasury told us that this scheme was developed at a time when the “hope was that the economy would be largely open and recovering, but with particular temporary restrictions in particular areas”.39 The JSS was amended by government on two subsequent occasions, with an announcement on 9 October stating that where businesses were legally forced to close due to lockdown restrictions the government would pay two-thirds of wage costs, with the employer not obliged to top that up.40 A second modification on 22 October decreased the percentage of hours that an employee would be expected to work for businesses still open and increased the government contribution to topping up wages.41 However, by 5 November the government had moved all of England back into a full lockdown and subsequently announced that CJRS would be extended on broadly equivalent terms to the way the scheme had been operating back in August.42 Government also announced that SEISS would be extended to the end of January 2021 at the rate of 80% of a self-employed person’s trading profits. This was a more generous offering than previous versions of the SEISS extension that had been announced during September and October.43 We were concerned that constant changes to the levels of financial support available across the different schemes created job instability with businesses and employees,44 and also impacted on the future funding settlement for the devolved administrations.45 36 Q 49, C&AG’s Report, figure 6 37 Qq 82–83 38 C&AG’s Report, para 1.23 39 Q 55 40 C&AG’s Report, para 1.24 41 HM Treasury, Plan for Jobs: Chancellor increases financial support for businesses and workers. Available at: www.gov.uk/government/news/plan-for-

We noted that a large number of business, particularly within more restricted areas, had not been able to operate normally for many months. While support for businesses is expected, it is essential that government provides this money as quickly as possible.46 The Job Retention Bonus was supposed to pay businesses £1,000 for each furloughed worker brought back to work and continuously employed until the end of January 2021.47 We asked HM Treasury how much money it had set aside for the scheme and whether it had been cancelled or deferred as reports appeared to be mixed. HM Treasury told us that this scheme had been deferred as it would not make sense to pay the grants at the same time as extending the furlough scheme and that instead it was something that the Chancellor will come back to next year.48 HM Treasury noted that the Chancellor felt that the broader and more generous safety net provided by extending the existing CJRS was required—rather than implementing the JSS—in order to help keep people in work.49

Employees who were let go after 23 September 2020 can be rehired and furloughed again with CJRS extended, but if they were let go before that date then it would not be possible to furlough them. HM Treasury told us this was similar to the approach it took when the CJRS scheme was first announced in March 2020, as eligibility for that initial phase of the scheme was backdated to 28 February to include those made redundant just before the scheme was announced.50 We asked the Departments whether they were concerned there might be some perverse incentives for regional areas to stay in higher lockdown restrictions if it meant that greater funding from the employment support schemes was available as a result. HM Treasury told us that this was not something it was directly involved in, but that the sense it had was that most areas were keen to avoid the restrictions on normal life and economic activity that went with higher tiers owing to the potential damage to jobs, livelihoods and wellbeing.51 Costing the schemes and assessing value for money

We asked the Departments how much they expected the extended furlough scheme and the SEISS would cost the taxpayer on top of the £55 billion spent so far. HM Treasury told us that it was not responsible for forecasting the expected cost of the scheme, which would be published by the Office for Budget responsibility in a few weeks. We were concerned that neither Department was able to provide any details on how much the extension to the schemes would cost. HM Treasury said the final costs would depend on take-up levels which in turn depended on labour market forecasts. It told us that it had a broad range of estimates, so it had a sense of what the maximum cost of the extended schemes might be and what range the costs were expected to be within, but we unable to provide the figures during our evidence session.52 We were concerned that, even if the OBR does the detailed number crunching, HM Treasury should at the very least be providing the Chancellor with some ballpark costings before implementing any government policies.53 46 Qq 97–98 47 HM Treasury, Further details of the Job Retention Bonus announced, available at: www.gov.uk/government/ news/further-details-of-the-job-retention-bonus-announced 48 Q 101 49 Q 38 50 Q 51 51 Q 99 52 Qq 15–17 53 Q 34 16 Covid-19: Support for jobs

HM Treasury asserted that this was a similar approach to that it had taken when introducing the original schemes in March.54 It told us that, at the height of the initial schemes in the spring, they cost around £10 billion per month, but that this had reduced “quite significantly” over the period to October. It confirmed that it expected the cost of the extended schemes to be lower as a result of differences in the restrictions in place, but did not have the data available to be able to say by how much.55

The total cost of the schemes is now estimated to be £76 billion, with OBR estimating that the extensions to the schemes will add an additional £21 billion to the total. We asked the Departments what calculations they had made of the value for money provided by the schemes. HM Treasury told us that the introduction of the schemes in March did not lend itself to traditional cost-benefit analysis because “it was not a marginal change in policy” but “a major structural change” almost without parallel in recent history. It told us that at the point at which ministers had to decide whether to intervene in the economy the cost of doing so could not be known with any certainty. It explained that it had no way of predicting in March what the impact of the schemes would be, in part because it did not know how long the lockdown would last or what the future costs of the pandemic would be.56

The condensed timetable for introducing the schemes meant that a lot of the standard documentation that would accompany such a major policy initiative—business cases, options appraisal and detailed cost-benefit analysis—wasn’t undertaken back in the spring.57 HM Treasury asserted that the potential economic costs and human cost of large-scale unemployment if the schemes had not been introduced meant that they were necessary even if it was not possible to quantify this.58 We were nonetheless concerned that while there was a speedy intervention with the schemes initially it was important not to lose sight of the need to undertake value for money calculations on behalf of the taxpayer.59 54 Q 16 55 Qq 18–21 56 Q 36 57 C&AG’s Report, para 1.8 58 Q 36 59 Q 39 Covid-19: Support for jobs 17