What skills are required in the modern economy? Economists have been grappling with this question since the explosion of information and communications technology (ICT) from the late 1980s. Over the past 30 years, the focus of the debate has shifted profoundly; with increased automation, the emphasis is now less on the broad levels of educational attainment that are correlated with the use of technology, and far more on the tasks people actually perform in the workplace. Recent research examines the nature of the skills that best complement technical change and those a worker might need to acquire in order not to be replaced by technology.

Gaining a better understanding of exactly what the most needed skills are, and how much it will cost us to equip workers with them is the subject of a special session at a conference on measurement we are co-hosting with the UK’s Economic Statistics Centre of Excellence (ESCoE).

Digital evolution and the changing demand for skills

In the early period of mass computerisation, the notion of skill biased technical progress came to the fore. It was assumed that using information technology would drive demand for more workers with ‘high level skills’. At that time high level skills were synonymous with a university education. Implementing the new technology required technical numeracy skills and critical thinking, both of which are primarily taught at universities.

As the technology matured, and economists delved more into the nature of the skills required, the focus switched to the specific tasks performed by workers. Jobs or occupations where workers performed non-routine tasks that could not be easily automated were found to be in demand, at the expense of those trained for more routine tasks, who were mostly concentrated in the middle of the skills distribution. Many occupations which had earned middle incomes before the advent of ICT, such as book keepers and typists, declined in importance. Over time, concerns about job displacement through automation and robotisation has tended to dominate the discussion of skill requirements in the digital age.

Automation, job displacement and new kinds of job

Does automation lead to lower employment? The answer is ‘not necessarily’, as set out in general terms in a recent paper by Acemoglu and Restrepo (2019). These authors suggest that automation, which involves capital replacing labour, leads to the replacement of tasks previously carried out by workers.

This process does not, however, always lead to job losses since it may raise productivity and drive demand for workers to undertake other non-routine tasks. In addition, displacement can be counterbalanced by the creation of new tasks, in areas such as web design or app development as well as more general tasks in data analysis, administration and management. These tasks might change within the same jobs as typists become administrators or lead to the creation of entirely new occupations.

Concerns about job displacement have been around since the start of the industrial revolution, but the aggregate employment rate has not shown any long run tendency to decline. While some commentators argue that the rate of displacement is higher now due to the speed of technological progress, the argument that ‘this time it is different’ is far from convincing. Nevertheless, automation does lead to lower demand for some types of workers, even if total employment does not decline.

Is technology driving wage inequality?

What is indisputable is that different types of workers have seen major changes to the level of financial reward that they can expect from their labour. There is ample evidence that wage inequality has been rising over the past few decades.

Two of the papers presented at the ESCoE conference consider impacts on earnings in more detail. Falck, Heimisch and Wiederhold (2019) investigate the use of and returns to ICT skills. They use the definition of ICT skills set out in the Organisation for Economic Co-operation and Development (OECD) administered Programme for the International Assessment of Adult Competencies (PIAAC) survey: “using digital technology, communication tools and networks to acquire and evaluate information, communicate with others and perform practical tasks”.

The authors show that the returns to ICT skills have increased, confirming the much-cited description of ICT skills by Neelie Kroes, Vice President of the European Commission, as the ‘new literacy’. Likewise Parey (2019), using data for the US, shows that technological change increases wage inequality, especially among professional workers.

Changing jobs: the need for retraining

Since digital technologies and automation lead to task and job displacement, workers whose jobs have changed or disappeared need to acquire new skills that will enable them to adapt to the new focus of their current jobs, or to change occupations.

A team at OECD have been looking into the costs of this retraining, how long it takes, and how this relates to the age of workers. The paper presented at the ESCoE conference by Andrieu, Jamet, Marcolin and Squicciarini (2019) estimates the costs of workers moving to ‘safe-haven’ jobs that are at low risk of automation. The authors show that the costs of retraining are substantial, at between one to five per cent of GDP, and are also greater for countries which have high shares of labour in manufacturing and for older workers.

Given these costs, policy discussions should consider how effective their education systems are at engaging workers with life-long learning, and who should pay these costs.  

Measurement is the key to understanding

While higher level education was once considered a useful yardstick of how well an economy was equipped to harness information technology, it is clear we need to go beyond simple measures of educational attainment if we want to gain a proper understanding of who wins and who loses from the rapid change we are seeing. Such research relies on having reliable and robust measures of skills such as the categorisation available in the PIACC survey.

Administrative or big data can also provide valuable information. For example, a recent ESCoE working paper, Djumalieva and Sleeman (2018), uses data from job adverts to generate a new skill taxonomy.

However, much more information is still required. Measurement is key to understanding, and we need more frequent survey data or access to new data sources. This richer data would allow us to develop new and broader measures of skill and link them to the new tasks required in our evolving workplaces. It is only with this information that we can understand how and where to invest to ensure that as much as possible, technological change creates more winners in our economy than losers.

 

Mary O’Mahony is Professor of Applied Economics at King’s Business School.

The Economic Statistics Centre of Excellence (ESCoE), in partnership with the UK Office for National Statistics, is holding a conference on Economic Measurement 2019 at Bush House, King’s College London, 8-10 May.

 

References and resources

Daron Acemoglu and Pascual Restrepo, (2019), “Automation and New Tasks: How Technology Displaces and Reinstates Labor”, MIT.

 

Elodie Andrieu, Stéphanie Jamet, Luca Marcolin and Mariagrazia Squicciarini (2019), “Occupational transitions: The cost of moving to a safe haven” OECD, Paris.

 

Nagui Bechichi, Stéphanie Jamet, Gustave Kenedi, Robert Grundke and Mariagrazia Squicciarini (2019), “Occupational mobility, skills and training needs”, OECD, Paris.

 

Jyldyz Djumalieva and Cath Sleeman (2018), “ An Open and Data-driven Taxonomy of Skills Extracted from Online Job Adverts”, ESCoE DP-2018-13.

 

Oliver Falck, Alexandra Heimisch, Simon Wiederhold (2019), “Returns to ICT Skills”, paper presented at the ESCoE conference Economic Measurement 2019.

 

Matthias Parey (2019), “Tasks and Technology: The Labor Market Effects of Innovation”, paper presented at the ESCoE conference Economic Measurement 2019.