Self-employment in Wales during the COVID-19 Pandemic

The economic fall-out from the COVID-19 crisis continues to wreak serious damage in the labour market, and at the time of writing the full extent of this damage is some way from becoming apparent. One significant group of workers who have been very badly affected by the economic shock resulting from ‘lockdown’ are the self -employed. In previous analysis it was reported that almost 4 in 10 of the jobs created in Wales over the decade following the 2008 financial crisis were in self-employment (Henley and Lang, 2017). Furthermore, this analysis demonstrated the high degree of diversity of self-employment activity across Wales, both in spatial and sectoral dimensions. This paper looks at how this group in Wales have fared so far during the crisis and makes some suggestions on how policy may address the prospects of the self-employed.

P u blis h e r s p a g e : h t t p s://d oi.o r g/ 1 0. 1 8 5 7 3/ w er. Thi s v e r sio n is b ei n g m a d e a v ail a bl e in a c c o r d a n c e wit h p u blis h e r p olici e s. S e e h t t p://o r c a . cf. a c. u k/ p olici e s. h t ml fo r u s a g e p olici e s. Co py ri g h t a n d m o r al ri g h t s fo r p u blic a tio n s m a d e a v ail a bl e in ORCA a r e r e t ai n e d by t h e c o py ri g h t h ol d e r s .

Introduction
Due to the particular concentration of the selfemployed in 'social consumption' and face-toface service sectors, initial projections suggested that a higher proportion of the selfemployed, in comparison to employees, across the UK would be at risk of loss of job or business ). Furthermore, the female self-employed would be at particular risk. Despite an initial hiatus in late March 2020, it quickly became apparent to the UK government that a support scheme for the self-employed would need to be developed to operate in parallel with the Coronovirus Job Retention Scheme (CJRS) (furloughing). At the time of writing it has recently been announced that the Self-Employed Income Support Scheme (SEISS) would be extended, albeit in a less generous form, through to April 2021. The purpose of this paper is to provide an early assessment, drawing on the latest available official labour force statistics, as well as some other public datasets and reports data. The paper begins by providing a short overview of the extent and diversity of self-employment in Wales, in comparison to the rest of the UK. With data from the ONS Quarterly Labour Force Survey for the second quarter of 2020, the impact of the crisis on the self-employed in Wales is then examined across a range of dimensions: gender, age, sector and whether sole-traders or employers. Some significant differences are found in experience between Wales and other areas of the UK, as well as many similarities. Finally, based on this analysis, some thoughts are offered on the immediate prospects for the self-employed in Wales together with some pointers on the ways in which the design of devolved and local policy might address these prospects.

Self-employment patterns in Wales on the eve of the crisis
Figure 1 charts recent changes in the overall rates of self-employment across the UK as a whole and within the UK's constituent nations for six years to March 2020. Overall the rate of self-employment amongst those over the age of 16 for the UK as a whole has plateaued at just over 15% of the economically active since 2016, although absolute numbers have continued to grow by around 55,000. In Wales the rate of self-employment has been more or less stable over the past five years. However, in 2019/20 there were 8,100 more selfemployed than five years previously, reflecting continued upward growth in the total labour force. However, across the UK as a whole for this period self-employed numbers grew by almost half a million. It appears that selfemployment has become relatively less attractive to workers in Wales. Given the official confidence intervals on these estimates, caution is required in reading too much into these recent trends.
Sustained longer-term growth in selfemployment has been interpreted in a number of ways. On the one hand, an 'opportunity-pull' account focuses on self-employment as useful, albeit noisy, indicator of entrepreneurial activity and business start-up. While it is certainly true that a majority of the self-employed can still be classified as sole-trader business owners, a significant minority are non-business owning freelancers or sub-contractors, moving between short-term contracts or projects. Furthermore, only a minority of self-employed business owners are employers of others. So, an alternative account of self-employment growth focuses on 'necessity-push' (Henley, 2017). Traditionally self-employment has been found to be counter-cyclical, and therefore a potential alternative to unemployment (Thurik et al., 2008). More recently self-employment growth may reflect the growing use of 'flexibilization' strategies by employers and the growth of so-called 'gig working' (Henley, 2020).
Recent analysis suggests that the first of these explanations ('opportunity pull') may provide a fuller account of the development of selfemployment in the UK and across its countries and regions, as self-employment growth rates appear to be higher where local economic conditions are more vibrant, typically around London and in the south east of England (Henley, 2017). Self-employment rates have also grown more in regions with formerly higher rates of self-employment which has contributed to a long-run persistence in regional selfemployment patterns in the UK (Saridakis et al., 2020;Fotopoulos and Storey, 2017). This FFigure 1: Self-employment rates, all aged 16 or over Source: Annual Population Survey, April-March, via Nomisweb account provides some explanation for the lower rate of self-employment in Wales in comparison to England, despite the policy emphasis that Wales has placed over the past two decades on encouraging and supporting self-employed start-up activity, through initiatives such as Big Ideas Wales.
Further detailed analysis, from the 2019 Office for National Statistics (ONS) Annual Population Survey, may provide clues as to why selfemployment rates in Wales had reached a plateau two or three years prior to the onset of the COVID-19 crisis. Table 1 documents the breakdown of self-employment by industrial sector for each of the home UK nations (with London separated from the rest of England).
Wales stands out as having a much higher proportion of the self-employed in agriculture, forestry and fishing in comparison to England and Scotland. According to ONS estimates across the UK as a whole, self-employment in the primary sector has fallen by over 17% since 2018 (ONS, 2020b). Table 1 also shows that Wales has a rather lower proportion of the selfemployed in those service sectors which are likely to provide opportunities for the more skilled. These are sectors in which across the UK self-employment numbers have been growing in recent years. For example, Wales has almost half the proportion of self-employed in the professional, scientific and technical services sector as London. The slower growth of self-employment in these service sectors in  Wales, combined with the decline of agricultural employment, makes for an industrial composition less conducive to growing self-employment. Figure 2 disaggregates the data by gender and confirms the well-recognised, although perhaps less well-understood, pattern of lower female self-employment (Saridakis et al., 2014). Two features are worth noting. The first is that the rate of female self-employment in Wales is lower than in England, although not as low as in Scotland or Northern Ireland. The second is that the plateau in self-employment in Wales is particularly a consequence of the falling rate of male self-employment. By contrast, the self-employment female rate has continued to rise. In this respect Wales is different from the other UK nations. However, although the male self-employment rate has fallen, it is important to point out that on the eve of the pandemic there were still 2000 more male self-employed and 6000 more female self-employed in Wales than five years previously.
Now rather dated analysis from the last Census in 2011 reveals further information about this significant diversity in self-employment rates across Wales (Henley and Lang, 2017).
In particular, rates of self-employment have tended to be higher in both rural Wales and in the larger cities; in the former because of economic reliance on home-based, tourism and agriculture-related activity, and in the latter because of higher numbers of graduate, professional self-employed. By contrast selfemployment rates in the south Wales valleys areas have tended to be lower, particularly for women, focused on trade-based selfemployment with higher numbers working from no fixed place of work.
Further recent overview analysis is available in ONS (2020a). Disaggregation of selfemployment within Wales to county level using ONS survey data is generally not possible, because of the scale of sampling. ONS (2020a) does provide some city region comparative analysis, revealing that the self-employment rate for the Cardiff City Region in 2019 was around 12%, whereas for the Swansea Bay City Region the rate of around 16% was well above average. In the Cardiff region the age profile of the self-employed is more heavily skewed towards younger workers than in many other UK city regions. However, these differences may reflect the spatial composition of these city regions -Cardiff City Region includes a number of south Wales valleys local authorities with traditionally very low rates of self-employment, whereas the Swansea Bay City Region extends across rural Pembrokeshire and Carmarthenshire, with the higher self-employment rates, and an older age profile, characteristic of such areas.
Henley and Reuschke (2020) undertook a projection analysis early in the UK-wide lockdown to predict where the impact on the self-employment might be most severe. This analysis focuses on an analysis of the concentration of the self-employment in sectors 'at risk' from compulsory closure of business activity and from the impact of social distancing. These sectors include non-food retailing, passenger transport, 'social consumption' activity such as restaurants and arts provision, and 'face-to-face' services such as personal-and child-care and domestic services. In general, a higher proportion of the self-employed than the employed are at risk because of the higher concentration of the selfemployed in these sectors. In this respect Wales is not an exception in comparison to other UK nations and English regions. Broad conclusions at the UK level point to the differentially worse (initial) impacts on female and older self-employed compared to corresponding groups of employees. In Wales, alongside in South West England, Scotland and Northern Ireland, the proportion of female self-employed 'at risk' is above the UK average.

The impact of lockdown on the selfemployed
The principal source of information used here about the post-lockdown experience of the selfemployed is from the Quarterly Labour Force Survey (QLFS), and in particular the second quarter of 2020 data (April to June). Data for this quarter is likely to be important because it more or less coincides with the most severe period of lockdown and business closure, between the Prime Ministerial announcement on March 23 rd and the start of the gradual relaxation of restrictions from mid-June onwards (and early July in the case of Wales). Table 2 reports the estimated aggregate change in the number of self-employed between the second quarters of 2019 and 2020 derived from the QLFS by UK nation, alongside previous annual changes for comparison. The QLFS data suggest some significant differences between the experience of the selfemployed in London in comparison to other areas of England away from the capital, and so London and the rest of England are separated in the table. As Table 2 shows across the UK, all of the growth in self-employment between 2018 and 2019 was wiped out by the estimated  However, before attributing the fall in selfemployment in Wales to the immediate impact of the pandemic, it is instructive to look at the quarterly change in self-employment between the first quarter of 2020 (January to March) and the second quarter of 2020, as shown in the middle columns of Table 2. This reveals that the experience in Wales is markedly different. Self-employment is estimated to have grown very slightly between the beginning of 2020 and lockdown; the large fall in self-employment in Wales took place before the onset of lockdown restrictions. By contrast, in the other UK nations the immediate impact of lockdown on self-employment is more readily apparent. This is particularly so in London where selfemployment is estimated to have fallen by over 13% between the first and second quarters of 2020. Table 3 provides some further breakdown of these changes, focusing on differences by gender and by age group. In London the immediate impact of the lockdown between the first and second quarters of 2020 was pronounced and felt by both male and female self-employed. In the case of females this reversed some of the growth over the previous year. In England outside London the adverse impact of lockdown affects entirely male selfemployment, although the early strong growth on female self-employment between 2018 and 2019 evaporates. In the case of Wales, the year-on-year drop between the second quarters of 2019 and 2020 for males is severe a fall of almost a quarter in the total. However, this drop occurred prior to lockdownbetween the first and second quarters of 2020 male selfemployment in Wales is stable. By contrast, female self-employment grew strongly at +14% between the first and second quarters of 2019 and 2020, more than offsetting the reduction in female self-employment over the previous year. There is still some growth between the first and second quarters of 2020 during the lockdown. Scotland follows a similar pattern to Wales, albeit with lower annual rates of change. Female self-employment in Scotland grew during lockdown. In Northern Ireland, however, the male-female pattern is reversed; women fare badly while male self-employment grows over the year between the second quarters of 2019 and 2020.
The age distribution of self-employment appears to be more stable. However, there is some evidence in Table 2 that the proportion of younger self-employed falls in all areas other than Northern Ireland. In some cases, this fall occurs over the year (Wales); in other cases, the fall is more dramatic during the lockdown (London). The age profile of the self-employed continues to shift towards older age groups year-on-year in England, Scotland and Wales, although in Wales and in London there is a slight suggestion that the impact of lockdown was to arrest this trend.
Further analysis also shows that in Wales in particular there was a sharp increase in parttime self-employment between the first and second quarters of 2020 (from 23% to 29% of total self-employment), consistent with a significant collapse of business opportunities for many self-employed. There is relatively little change in the proportion of the self-employed who are employers of others, suggesting some positive benefit from job furloughing on larger self-employed businesses.
Figure 3 also shows that the impact of the crisis varies for different forms of self-employment. The percentage drops between the first and second quarter of 2020 were generally worst in London where own account workers ('working for self') fared worst with a fall in numbers of over a quarter. In Wales, while noting that selfemployment rates had already peaked prior to the crisis, the percentage changes are less severe than in Scotland or Northern Ireland. However, they are worse than in England outside London. The quarterly fall in business owners, at 13%, was twice as severe as in England outside London. Furthermore, numbers of both business owners and own account workers fell at the same rate in Wales. The only positive note in these data are that Wales, by contrast with the other nations, did not experience a fall in numbers of subcontractors and freelancers.
Further valuable information on the impact of the crisis on the self-employed is to be found in reports by the HM Revenue and Customs (HMRC) on the take-up of the emergency Self-Employed Income Support Scheme (SEISS) administered at UK level by HM Revenue and Customs (HMRC, 2020). Although initially generous in its provision for those eligible for support, this scheme has been widely criticised because it has excluded from support a number of significant self-employed groups: those recently self-employed, those whose self-employment is part-time or who have historical profits above £50,000 per annum, and those amongst the self-employed who operate through the structure of a limited company, paying themselves through dividend rather than salary (Rouse et al., 2020;Adam et al., 2020). As official data reveal these limitations have excluded 1.6 million from the UK total of 5 million self-employed (HMRC, 2020). The 'Bounce Back Loan' (BBL) scheme for small businesses announced in April 2020 offers one route for the excluded to access funding for their business. However, whilst on favourable terms, this is a loan which requires repayment, in comparison to the SEISS which operates as grant.
Almost 65% of the 216,000 self-employed in Wales were assessed as eligible to receive SEISS between the announcement of the scheme and August 2020, a little below the UK overall fraction of 67.3%. Across UK regions and devolved nations take-up rates of SEISS amongst the eligible self-employed vary littlein Wales the take up-rate was 58.6%, compared to a UK overall rate of 59.6%. Overall SEISS has supported just 38.0% of the self-employed in Wales, compared to 40.1% across the UK as a whole. The average payment size, covering a three month period, in Wales was £2,400 compared to the UK average of £2,500. Take-up rates in Wales have been highest in construction, education, finance and insurance, transportation and in other services. This is a pattern which is broadly similar to that elsewhere in the UK. In terms of proportion of the total of eligible selfemployed individuals, the construction sector dominates in Wales, accounting for 34% of claims. While this fraction is a little lower than in London and higher than in some other   Source: authors' tabulation from HMRC (2020)

Figure 5 : SEISS claims as a proportion of eligible self-employed by Wales local authority
Source: authors' tabulation from HMRC (2020) Figure 5 charts the SEISS take-up rate from those eligible for support across local authorities in Wales. The rate varies from two thirds in Blaenau Gwent to just below a half in Powys. In broad terms there has been less reliance on SEISS in rural areas, and highest rates of reliance in Cardiff, Newport and the south Wales valleys.

Implications for regional development and entrepreneurship policies
Self-employment is an imperfect indicator of entrepreneurial statusa point that is wellrecognised in the academic literature.
Interpreting levels of and rate of change in selfemployment can therefore be problematic.
Higher levels of self-employment may indicate improved business start-up activity and microbusiness survival. Self-employment is thus used, often uncritically, by researchers as a proxy indicator for entrepreneurialism. But the self-employed are not all entrepreneurs, as the literature on 'necessity self-employment' makes clear. So higher levels of selfemployment may also reflect the consequences of changes in labour market regulation and the growth of employment 'flexibilization' strategies. It is therefore hardly surprising that in a major economic crisis it is the self-employed, representing the least attached group of the workforce with low levels of organizational-specific human capital and the lowest employer separation costs, who find themselves at the top of the list if workforce reductions are required.
Policy sequencing in the UK in the early days of the COVID-19 national lockdown may not have helped. CJRS was announced simultaneously with the national lockdown and this furloughing provision was rolled out quickly by the Welsh Government, along with emergency business support provision, under devolved policy competency. The SEISS scheme only followed after a passage of time and substantial political lobbying. SEISS was administered at UK-level without explicit attention to sectoral or local/regional differences in need. Its funding was released more slowly to beneficiaries by HMRC. It is possible that a devolved roll-out might have found the means to extend the support safety net to the significant numbers of self-employed who were ineligible, as well as allowing for greater targeting of support to meet sectoral and local area needs. It should not be overlooked that schemes such as CJRS and SEISS fulfil not only a supply-side function in protecting human capital and skills investment, but also a demand-side function in supporting spending in hard-hit localities.
In any discussion, at the present time of writing, the point needs to be emphasised that it is too early to assess the full impact of the pandemic on self-employment in Wales. It would be foolish to conclude that, because selfemployment rates in Wales had already peaked prior to the onset of the crisis, the medium-term impact in Wales will be less severe. This earlier peak, two or three years prior to the current crisis, places Wales in contrast to the other UK nations. This is concerning because it suggests that the level of traction from entrepreneurship and business start-up policies in Wales was already beginning to dissipate prior to economic crisis induced by lockdown. We might speculate about the causes of this. Candidate explanations might include wider levels of uncertainty about business opportunities in Wales arising from austerity-related slowdown in growth over the past two to three years. Uncertainties may have been compounded following the UK decision to leave the European Union, particularly as Wales has been a major beneficiary of EU Structural Funds support. Nor indeed should one conclude at this stage, because female selfemployment has been less hard hit for reasons of sectoral composition that made construction vulnerable in this crisis alongside personal services, that over the medium-term women in self-employment may prove more resilient. Projections, based on the sectoral composition of 'shutdown' sectors, suggest that women would be harder hit. This has not materialised so far, but may simply indicate that women are more likely than men to 'moth-ball' selfemployed activity and withdraw temporarily from the workforce, rather than close and seek alternative full-time paid employment or Universal Credit payments.
The other big uncertainty here concerns whether reduction in the numbers of the selfemployed will be reversed once the economy starts to recover. This connects to a question of the ability of the Welsh economy to recover past levels of entrepreneurial dynamism. If damage is long-standing and reflected in future business start-up and growth levels, then the urgent attention of policy makers will be required.
In Wales, perhaps more so than in other regions of the UK, the COVID-19 crisis is compounded by some structural shifts in the sectoral and spatial composition of the selfemployed workforce, and in particular the steady decline in numbers of older selfemployed in rural areas engaged in agricultural activity or activity closely related to the rural economy. On the other hand, levels of selfemployment in Wales appear more reliant on lower skilled occupations, for example in construction trades, but also more on older, more experienced self-employed. This may have protected Wales thus far, compared in particular to London, in that these selfemployed are longer established and experienced, with a relatively higher proportions of sole-traders and business owners, rather than subcontractors and freelancers. These groups have, for example in the case of construction sector workers, been better able to access short-term support through the SEISS scheme. Wales, in common with other English regions away from London and the South East, has seen less growth in numbers of young 'gig-economy' workers, who appear to have been particularly hard hit during the crisis. On the other hand, if remote working is to become a more prominent and permanent feature of the UK employment landscape (Felstead and Reuschke, 2020), then lower levels of potential graduate, knowledge economy business founders may turn out to be a long-term structural weakness for Wales. This is a weakness which informs the broader context of Wales' productivity and innovation gap both within and beyond the UK.
Alongside productivity, well-being is also important, as recognised in the Wales Wellbeing of Future Generations Act of 2015 which now frames the policy landscape of Wales. The self-employed have generally reported higher overall levels of job satisfaction than the employed (Georgellis and Yusuf, 2016). One driver of this, particularly for women, might be the better opportunities offered by selfemployment to balance work with other home commitments (Reuschke, 2019). However, the isolation of long-term remote working, combined paradoxically with a sense of never being 'disconnected', may lead to significant loss of wellbeing (Daniel et al., 2018). This needs to be kept in mind if future policy is directed to support the development of knowledge economy-based self-employment as a career option for graduate workers.
The economic consequences of the COVID-19 crisis are likely to be more severe in particular areas of Wales, such as the south Wales valleys, where levels of self-employment and microbusiness activity were already low before the pandemic. The policy challenge in terms of support for entrepreneurship has therefore become now more severe. At present, policy to support the self-employed, either directly or indirectly, entails no sector-specific or locationspecific dimension. For example, as seen in Table 1, a greater proportion of the selfemployed in Wales, in comparison to England outside London, work in the arts, entertainment and recreation sector. This sector has been particularly badly hit by venue closure and the requirements of social distancing, with the consequence that business re-opening has lagged all other sectors. It is a sector which in the Welsh context provides wider public and cultural value beyond its role in supporting businesses and as an employer. For the selfemployed in this sector there may be a strong case for longer term policy attention and support. The self-employed in tourism-related activity may also have been hit more severely in Wales because the re-opening of this sector was slower than in England, and it may have been differentially affected by the greater use of travel restrictions within and into Wales as a tool to limit coronavirus spread. The sector is proportionately more significant to the economy in Wales than in some other parts of the UK and may therefore merit particular medium-term support.
The immediate impact of the lockdown on the earnings of the self-employed was severe. Reuschke et al. (2020), analysing data from the Understanding Society household panel survey, find that over the first month of lockdown at UK level just under half of all selfemployed men and just over half of selfemployed women experienced a fall in earnings of at least 50%. These falls are very severe, and particularly so in comparison to corresponding groups of waged or salaried employees, illustrating the differential impact of the support schemes (CJRS and SEISS) for the respective groups. There was only limited recovery in earnings for the self-employed after two months, as the economy began to reopen.
While the survey is not large enough to provide a reliable regional breakdown there is little reason to anticipate that the self-employed in Wales fared any better. At the time of writing, available evidence suggests that consumer spending and high street footfall is still subject to considerable uncertainty and volatility, made worse by rising unemployment levels and by reimposition of lockdown restrictions across the whole of Wales.
Weakness in consumer spending will continue to have serious impacts on the self-employed, particularly for those in business sectors which provide 'final' goods and services to households, and especially where social distancing restrictions are problematic or infeasible.
On a more positive note there is survey evidence that a minority of the self-employed experienced increased demand for their services during lockdown, and an abundance of anecdotal evidence that the more resilient microbusinesses have been able to undertake pivoting of business models alongside new-tobusiness innovation. This does not take place in a vacuuminnovative activity is supported by an appropriate balance of 'hard' and 'soft' skills, access to knowledge, networking social capital and appropriate infrastructure. Each of these are legitimate wider concerns for policy design and support in the months and years ahead.