Graduate labour market update: outlook for cities and impact of AI

In his latest labour market update, Charlie Ball rounds-up the most recent economy data, and takes a look at new reports on the performance of UK cities and the impact of AI on the future of work

In October to December 2023, the estimated number of job vacancies in the UK fell by 49,000, according to data from the Office for National Statistics (ONS). However, vacancies remain above pre-COVID-19 pandemic levels.

• Annual growth in regular earnings (excluding bonuses) was 6.6%, while growth in average total earnings (including bonuses) was 6.5% in September to November 2023. Real terms growth (adjusted for inflation) for total pay rose by 1.3%, and for regular pay by 1.4% during the same period.
• The estimate of payrolled employees in the UK for December 2023 decreased by 24,000 compared to the revised November 2023 figure, reaching 30.2 million.
• The estimated total number of vacancies fell by 5.0% from the previous quarter, with arts, entertainment and recreation contracting the most, falling by 31.6%, followed by transport and storage, which fell by 23.7%.
• October to December 2023 continues the sequence of consecutive quarterly falls, decreasing for the 18th consecutive period. The industry sector showing the largest fall was wholesale and retail trade; repair of motor vehicles and motorcycles, which was down by 13,000. The current period of consecutive quarterly declines in the vacancy estimates is the longest ever recorded.
• When comparing October to December 2023 with the same time the previous year, total vacancies decreased by 226,000 (19.4%) with falls in 17 of the 18 industry sectors. The industry that decreased the most was human health and social work, where the estimated number of vacancies fell by 37,000 (17.6%).
• The total estimated number of vacancies remains 133,000 above January to March 2020 pre-COVID-19 levels, with human health and social work activities showing the largest increase, at 34,000. Five industry sectors fell below pre-coronavirus levels with the largest falls in transport and storage, and wholesale and retail trade; repair of motor vehicles and motorcycles both falling by an estimated 6,000 vacancies.

Skills Development Scotland have an overview of the December data. December 2023 saw the number of job postings down 16,100, to 34,200 compared with November 2023. This will in part reflect seasonal patterns, but postings also fell by 19% (-7,800) compared to the previous year.

The highest demand in December was for:

• Care workers and home carers
• Sales related occupations
• Cleaners and domestics
• Other registered nursing professionals
• Programmers and software development professionals.

The occupations in highest demand remain similar to last month, but the top three in the list (all non-graduate jobs) saw significant falls in vacancy numbers.

Welsh labour market data is available too. Early estimates for December 2023 indicate that the number of paid employees in Wales decreased by 700 (0.1%) over the month to 1.31 million. The adjusted experimental employment rate in Wales was 72.6%. This is unchanged on the quarter and up 0.3 percentage points on the year.

There is also new data for Northern Ireland. The number of employees receiving pay through HMRC PAYE in NI in November 2023 was 796,400, a decrease of 0.1% over the month and an increase of 1.7% over the year. Businesses reported that employee jobs increased over the quarter (0.9%) and increased over the year (2.4%) to 822,870 jobs in September 2023.

NISRA, acting on behalf of the Department for the Economy, received confirmation that 410 redundancies occurred in November 2023. Over the year December 2022 to November 2023, 2,590 redundancies were confirmed. There were 390 proposed redundancies in November 2023, taking the annual total to 4,200. Both annual totals were approximately three times more than the annual totals in the previous 12 months.

The Central Statistics Office (CSO) has also released SME data for Ireland:

• Small and medium enterprises (SMEs) (with fewer than 250 persons employed) accounted for 99.8% of all enterprises and 69.2% of persons employed in 2021. Two-fifths (41.5%) of total turnover and 34.8% of Gross Value Added (GVA) was attributed to these enterprises. In comparison, large enterprises (250+ persons employed) accounted for 0.2% of all enterprises and 30.8% of persons employed but generated 58.5% of total turnover and 65.2% of total GVA in the business economy in 2021.
• In 2021, micro enterprises (wither fewer than ten persons employed) accounted for 92.6% of all enterprises, 27.6% of persons employed, 15.3% of turnover, and 16.2% of GVA. Small enterprises (10 to 49 persons employed) accounted for 6.1% of all enterprises, 22.0% of persons employed, 12.1% of turnover, and 7.5% of GVA. Medium (50 to 249 persons employed) enterprises accounted for 1.1% of all enterprises, 19.6% of persons employed, 14.1% of turnover, and 11.1%% of GVA.
• The services sector accounted for 60.9% of all SMEs in the Irish business economy. The construction sector was the second largest contributor to the number of Irish SMEs, at 20.2% of the total. The manufacturing sector accounted for just 5.2% of all SMEs.
• The services sector accounted for the largest proportion of persons employed in SMEs in 2021 (58.4%) and contributed to 44.2% of the turnover in this size class. The lowest proportion of persons employed was in the manufacturing sector with 9.3% of persons employed while it accounted for 13.1% of  turnover in the SMEs size class. Having accounted for 19.3% of all active SMEs, the construction sector accounted for just 11.9% of persons employed in SMEs and 7.1% of turnover. The distribution sector accounted for 13.8% of all active SMEs, 20.4% of persons employed and 35.7% of turnover in the SMEs sector in 2021.

The total number of online job adverts on 26 January 2024 was 18% lower when compared with the level seen in the equivalent period of 2023:

• The total number of online job adverts on 26 January 2024 increased by 4% when compared with 22 January 2024. This is 18% below the level seen in the equivalent period in 2023.
• Between 22 and 26 of January 2024, the total number of online job adverts increased in 11 of the 12 UK countries and English regions. However, when compared with the equivalent period in 2023, all 12 UK countries and English regions decreased, with the largest decrease of 26% seen in Scotland.
• At present, education is the only industry with a higher vacancy level than this time last year.

29% of trading businesses reported turnover had decreased in December 2023 when compared with November 2023, up six percentage points; while 14% reported their turnover was higher.

47% of trading businesses reported they were not considering raising prices in February 2024, broadly stable with January 2024; of those businesses that are considering raising prices, 23% reported labour costs as their main reason for doing so.

The level of hybrid working in the UK remained stable in January, and 10% of businesses with ten or more employees reduced their headcount in December while 12% increased it. Accommodation and food service businesses were most likely to reduce headcount, and health employers most likely to increase it. In February, 16% of employers with more than nine employees expect to increase headcount (health and social care again most likely) and 9% expect to decrease it (transportation and storage employers most likely to expect to reduce their workforce).

The Department for Education has released a new tranche of data on FE and apprenticeships. Of the 1,071,900 adult (19+) government-funded further education and skills learners participating so far in 2023/24:

• Females account for 58.3% (624,500).
• Higher level (level 4 or above) participation increased by 3.7%, to 211,800 from 204,220 in 2022/23.
• Learners recorded as having a learning difficulty/disability (LLDD) account for 19.2% of the cohort (198,750) an increase of 7.9% from 184,140 in 2022/23.
• The cumulative number of enrolments on Free Courses for Job qualifications, reported between April 2021 and October 2023, stands at 61,380, with 12,170 of these reported in the 2023/24 academic year so far.
• Adult essential skills including digital skills participation reported was 290,880 - a decrease of 4.8% since 2022/23.
• Female learners participating in Community learning was 74.7% (97,120) compared to 75.3% (89,760) in 2022/23 while males increased to 25.3% (32,890) compared to 24.7% (29,418).
• From August 2023 Community learning types have been replaced by seven new purpose types. Of the 130,010 learners, 32,070 are participating in Improving essential skills including English ESOL Maths and Digital.

There's lots of data available at these dashboards.

The Recruitment and Employment Confederation's (REC) January Jobs Outlook reports that businesses' mood about their own prospects continues to improve.  The net balance of employers' confidence in making investment and hiring decisions was +4 in October-December. Employer sentiment on the wider economy remains strongly negative (-43, down 4 points).

Employers' sentiment towards short-term (next three months) permanent hiring was strong this October to December 2023 quarter (at net: +16) - especially among medium-size businesses. And medium-term (four to 12 months) permanent hiring intentions also remained strong in this quarter (at net: +23). On the temp side, the balance remained positive (net: +2) for the short term, although medium-term planned demand for agency workers turned marginally negative (net: -1).

The Centre for Cities have released their annual Cities Outlook. It's one of my favourite labour market publications of the year, and this year's is as good as ever. This edition compares cities' performance since 2010 against the previous period of 1998 to 2010:

• The 63 largest cities and towns in the UK account for 9% of land but 63% of output and 72% of high-skilled jobs.
• All but two of Britain's 62 largest cities and towns had more jobs too. London led in relative terms. Its total number of jobs grew by almost 30%, adding a further 1.4 million positions to the capital's economy - the equivalent of adding the whole of Wales to the UK jobs market. It was joined by Luton, Reading, Cambridge and Milton Keynes in the top five. And half of cities had jobs growth that was higher in this period than the national average between 1998 and 2010.
• In contrast, Aberdeen and Worthing were the only two places that had fewer jobs. Both these cities are heavily reliant on a single industry - oil and gas for Aberdeen, and aerospace for Worthing. Around 30% of jobs in Aberdeen's export base - the part of the economy that trades with other areas of the UK and the world - are directly related to oil and gas. And unlike the 2000s, this sector struggled in the 2010s. The result is that the city has lost an estimated 9,000 jobs in areas related to oil and gas. This is likely to have impacted the amount of money spent in the city: the number of retail jobs fell by almost 30% compared with 6% nationally. And house prices have been affected too - averages were 3.4% higher in 2022 than in 2010 (compared with a national increase of 50%) and were 15% down on their 2014 peak.
• At the national level, gross disposable household income grew in real terms by 0.7% on an annual average basis, or £1,800 over the period. This compared with a yearly average growth rate of 1.4% between 1998 and 2010, an increase of £3,300 (in 2022 prices).
• Despite its poor productivity growth, London was the only city that saw its disposable income per head rise faster than the national increase between 1998 and 2010. On average, incomes of Londoners were almost £4,300 higher than in 2010 (in 2022 prices). Meanwhile, Burnley and Aberdeen had lower incomes than in 2010.

This is, as always, a fascinating publication, and there's a nice data tool here to explore the many stats the Centre have furnished us with.

The Resolution Foundation have a labour market outlook for the coming quarter:

The UK is unique among G7 countries in having lower employment than pre-pandemic. While employment in the UK remained resilient during the height of the pandemic (in large part thanks to the Job Retention Scheme), our recovery has lagged other countries. The Foundation suggests that our high level of workers with health issues might be a cause.

Places with pre-existing health problems, high levels of reported disabilities and low graduate shares have seen particularly steep increases in inactivity due to ill health. Places with already-high rates of sickness-related inactivity like Merseyside (+1.6 percentage points), Tees Valley and Durham (+1.5 percentage points) and West Wales (+1.5 percentage points) have experienced twice the national increase. In contrast, Inner London East (-0.4 percentage points) and West (-0.3 percentage points) have actually seen falls in the share of working-age people who are inactive due to ill health.

AI and the future of work

There are two heavyweight reports on the future of work to cover this time around. The first is from the International Monetary Fund (IMF):

• Almost 40% of global employment is exposed to AI, rising to about 60% of jobs in advanced economies like the UK. Half may be negatively affected by AI, while the rest could benefit from enhanced productivity through AI integration.
• More educated workers are better prepared to move from jobs at risk of displacement; older workers may be more vulnerable to the AI-driven transformation. In the UK, most graduate level employment is exposed to AI, but in most cases will have a positive impact (in the IMF's terminology, have 'higher complementarity'). The exceptions are in some professional roles, and in some technician occupations, and largely in younger and/or less experienced workers, suggesting some impact of AI on roles for career starters and new hires. The strongest negative impacts, however, will be amongst clerical workers in the IMF's analysis.
• That said, the IMF does express concern for older workers in the UK displaced by AI; the likelihood of reemployment of older workers exposed to AI even in high skilled jobs to a job of similar skills level is estimated by the IMF as being well below 50%.
• Because more educated, and better paid workers in the UK are both more exposed to AI but also more likely to benefit from AI, it is felt that it is likely to increase labour inequality, with earnings boosts being linked to existing incomes - in other words, better paid professionals are particularly likely to see their wages rise still further, whilst the less qualified are less likely to see their jobs affected but will also struggle to realise benefits from AI adoption.

The other report is from Microsoft. Quite a bit of this is spent explaining to the reader about how great Microsoft CoPilot is, but there is lots of interesting information in there.

• Mostly early studies have found that new or low-skilled workers benefit the most from Large Language Models such as ChatGPT or Microsoft CoPilot.
• Information search as well as content production (manually typing, writing  code, designing images) is enhanced by AI, so general information work may shift to integrating and critically analysing retrieved information.
• With more generated text available, the skills of research, conceptualization, planning, prompting and editing may take on more  importance as LLMs do the first round of production.
• Skills not directly related to content production, such as leading, dealing with critical social situations, navigating interpersonal issues, and demonstrating emotional intelligence, may all be more valued in the workplace.
• Development of prompts is difficult and nuanced. Even seemingly minor changes, including capitalisation and spacing, as well as the ordering of prompt elements can result in dramatically different LLM outputs. The same prompt can result in significantly different performance across model families, even with models of similar parameter size, And nobody really understands why. There is little theoretical understanding for why any particular technique is suited to any particular task.
• AI can also make tasks worse and less productive if used ineffectively. Overreliance on AI leads to poorer performance than either the human or the AI acting alone.

There's a lot more in there and it's a neat reference to the state of current thinking and research in the field.

Meanwhile, Indeed believes UK labour market demand has largely returned to normal after the disruption caused by the pandemic, with one notable exception - they saw no meaningful reduction in hybrid working taking place in 2023.

The Indeed Job Postings Index, which includes more recent data through late January, is now slightly below pre-pandemic levels - as of 26 January, total UK job postings on Indeed were 4% below their 1 February, 2020 level. Postings for social science jobs are more than double their pre-pandemic baseline. By contrast, tech categories are among the weakest performers, with job postings for software developers down 46% from the baseline.

Notifications of potential redundancies have remained modest in recent months despite some layoff announcements in sectors including tech, professional services and construction.

And finally, Prospects Luminate have published a series of pieces on the shape of the 2024 graduate labour market from Stephen Isherwood of the Institute for Student Employers, Rebecca Fielding of Gradconsult, and also from me.

Written by

Charlie Ball

Head of labour market intelligence

Jisc