If we were able to precisely pinpoint exactly what technological change was going to happen, the businesses that would be created and the ones that would fail, and the decisions each individual made about what job to do, where to live, and how to navigate change, then skills and the economy would be perfectly matched.
This of course is neither realistic nor desirable. But it is useful to have a better understanding of what skills might be needed in the future – for individuals to identify possible future career opportunities, for education and training providers to work to meet those development needs, for businesses to start thinking about how to manage existing and new staff to be the most productive, and for governments to identify areas of growth, and pursue an industrial strategy.
In addition, there will inevitably be shocks to the economy that cannot be wholly predicted. A pandemic, flood or drought, and technological leaps are all more likely in our modern economy and rapidly heating world, although when and where they might happen are less predictable.
Skills forecasts take many forms. In a project which I co-authored with a great team from Frontier Economics, we took a systematic approach to reviewing a longlist of 70 forecasts, with in-depth case study reviews of 18.
The project was commissioned by the Unit for Future Skills, part of the Department of Education (DfE), to review and assess leading and emerging methodologies for the analysis of future skills needs. The intention of our project was to review the methodologies currently used and make suggestions that would improve any future economy-wide skills forecast. Our final report is worth reading alongside the review of a potential UK-specific skills taxonomy which was published last October.
A quick history lesson
Skills forecasts have been created for the UK for some time, but by different bodies and at different levels.
You may remember the UK Commission for Employment and Skills (UKCES) which, amongst other responsibilities, was tasked with identifying the future skills needs of the UK. It did this through commissioning the “Working Futures” economy-wide forecasts. Prior to 2010, there were also a series of Sector Skills Councils that reviewed and revised the economy-wide forecast for each sector, creating a feedback loop. Similarly, Regional Development Agencies were tasked with looking at and using the regional skills forecasts, while also creating a feedback loop with UKCES. This entire infrastructure is no longer systematically funded by the government, and UKCES closed in 2016.
Since then, sectoral and regional forecasts have become more disparate. Some industries retained skills bodies, either because they pre-dated Sector Skills Councils, or because the industries were able to obtain funding from elsewhere to maintain some functions.
A range of other employer representative bodies, and sector organisations, as well as interested funders have stepped in to create sector-specific skills forecasts.
Local Enterprise Partnerships, followed in some areas by a process of devolution to Mayoral Combined Authorities, and latterly the Local Skills Improvement Plans that employer representative bodies (mostly Chambers of Commerce) were funded to develop in local areas in 2022, have all been part of the local and regional picture for creating, interpreting, and using skills forecasts. After 2016, the government continued to commission the Working Futures forecast, and our report was part of answering the question: what next?
A bundle of reports
Our project looked in-depth at a range of different forecasts: nine economy-wide and nine sector-specific. The economy-wide forecasts include Working Futures and the Nuffield-funded Skills Imperative 2035 work delivered by the Institute for Employment Research at Warwick and Cambridge Econometrics, the latter in partnership with NFER and the University of Sheffield.
We looked at forecasts on the potential impact of artificial intelligence by PwC for BEIS; on horizon scanning for labour market and skill demand by RAND Europe for DfE; and the Cedefop European approach, also delivered by IER and Cambridge Econometrics among others.
The sector-based forecasts included those from the Food and Drink Federation, Engineering UK, the Workforce Foresighting Hub, and the Construction Industry Training Board.
These reflected a range of approaches, but in general, most are quantitative approaches using existing data to create models that predict future skills demand, while reviewing the data with experts in the relevant sector or region to amend for any knowledge that is not captured by the data.
The Working Futures method works in this way, using existing data from ONS such as the Census and the Labour Force Survey, alongside data on issues like automation rates, as well as a historic understanding of the role of qualifications in different sectors. The model is informed by expert input at different points, and then published as a national projection with sectoral and regional analysis. This provides a helpful indication of the future needs for qualifications at a national level, but may be less useful at a granular level.
Other methods have been adopted in specific sectors. For example, Engineering UK’s report projects past trends into the future, using machine learning. This takes the changing requirements for jobs advertised online now and in the past few years to predict how jobs might change in future.
An increase in the demand for certain technical skills or specific job roles evidenced in this way can be combined with expert advice to project the future demand for those skills and highlight current shortages. This method would likely not reflect any changes in demand for skills around communication or working with peers, which are likely to exist on job descriptions in a similar way regardless of changing needs.
The skills challenge
These two examples highlight key challenges for skills forecasts – the first being the link between skills and qualifications, and the second being the importance of including expert views when developing forecasts.
On skills and qualifications, a model which only projects demand for certain qualification levels is not easily translated into a plan for, say, a region’s approach to upskilling adults for the jobs available in the region in the coming years. Similarly, a need for certain technical skills without a broad level attached is difficult to translate into both the supply and demand for qualifications.
If the forecasts are used as indicators, and then more work is done to investigate specific needs for an area or sector, then this need not be an issue – but it makes them difficult to use at a granular level.
Our report sets out how interpreting data for skills forecasts should always be conducted with experts in the area, whether sector or region, and ideally include both those who employ individuals, and those who train them. Involving experts in the discussions about what data to include in the model and how it might be used is already common, but it should be an essential part of the development.
Into the future
Other recommendations we give for Working Futures (or an equivalent future skills forecast) include providing forecasts for mayoral combined authorities, rather than simply for regions. The landscape here has changed significantly since the early days of UKCES and skills forecasts haven’t always provided data that can be used by the most recent local bodies.
A further recommendation for commissioners of skills forecasts, in this case DfE but of relevance to any commissioners, is to provide more guidance for users of those forecasts. In a world where misinformation can go unchallenged and trust in what is true can be low, it’s crucial to ensure that users of skills forecasts are mindful of their language in reporting on the potential implications for the economy, sectors, or local areas.
While we hope that forecasting future skills is useful to help us understand future demands for skills, and that plans can be made on the basis of forecasts in broad terms, we also know that the forecasts are only predictions of what might happen. Improving methodology can improve skills forecasts, but they will never be completely accurate predictions in all cases, so caution and moderation are advised.
Although not in scope of this report, it is foreseeable that skills forecasts and the Unit for Future Skills might come under the aegis of Skills England, as suggested by Labour and fleshed out in more detail by AoC and Work Advance’s recent report. We hope that if any future body is created, it takes account of our report on the commissioning and use of skills forecasts and its recommendations.
