There are at least four areas where stronger international co-ordination could lead to better skills statistics: job tasks surveys, skills assessments, expert and science-based technology evaluations and online job vacancies. Each of these four approaches has its own limitations, but their combination seems able to provide useful and timely insights into the changes in skills demand driven by digitalisation (Spiezia, 2018).
Job tasks surveys are very useful to identify how job characteristics change over time and to infer the implications of these changes on the demand for skills. Very few countries, however, have established surveys of this type. The US Occupational Information Network (O*NET) is one of the best-known (https://www.onetonline.org/), and in the United Kingdom, the Employer Skills Survey provides a comprehensive picture of skills needs and training investment, including vacancies and skills shortages, employee skill gaps and the recruitment of education leavers and young people (https://www.gov.uk/government/publications/ukces-employer-skills-survey-2015-uk-report). In Germany, the BIBB/BAuA Labour Force Surveys (https://www.bibb.de/en/2815.php) provide information on the workplace as well as on the relationship between education and employment. One main reason why job tasks surveys are not common is the high cost of developing and conducting such surveys. Importantly, the measurement of workers’ skills is based on self-reporting and no formal assessment is carried out on their actual skill levels. Skills assessment surveys, therefore, function as a key complementary tool to improve understanding of skills needs.
The OECD Programme for the International Assessment of Adult Competencies (PIAAC) and the OECD Programme for International Student Assessment (PISA) are well known, cross-country skills assessment programmes. As with the job tasks surveys, PIAAC asks questions about a range of job characteristics and work skills. In addition, PIAAC tests participants through formal tests in order to assess their literacy and numeracy skills and their ability to solve problems in technology-rich environments (i.e. to use these tools to access, process, evaluate and analyse information effectively).
While PIAAC targets adults, PISA tests the skills and knowledge of 15-year-old students in science, mathematics, reading, collaborative problem solving and financial literacy. Like PIAAC, PISA not relies on the respondent’s self-assessment but also carries out formal tests of these skills.
A third, useful approach to identifying emerging skills needs is to ask experts for their assessment of what tasks, currently performed by humans, can or could be performed by digital technologies within a short time horizon. A widely cited study by Frey and Osborne (2013), which estimates that 47% of US employment is at a high risk of automation over the next several decades, is based on this approach. In 2016, the OECD asked a group of 11 computer scientists to review the test questions in PIAAC and to identify the questions that could be answered by machines today. Overall, the experts’ assessment suggests that the level of computer performance in three skill areas – literacy, numeracy and problem solving – is comparable to that of many workers. Only 13% of the workforce in OECD countries uses the three PIAAC skills on a daily basis and demonstrates a proficiency clearly exceeding the capabilities that computers are capable of reproducing (Elliot, 2017a). Based on a review of computer science research literature, Elliot (2017b) argues that IT capabilities could provide the reasoning, vision and movement skills required in most current jobs. Only in the area of language skills, does the analysis suggest that a substantial number of current jobs have skill requirements that clearly outstrip the IT capabilities demonstrated in the research literature. For this approach to become useful for skills development policies, expert and science-based assessment should be carried out systematically, considering more specific tasks and occupations, and across different countries. This is clearly one avenue where official statistics should consider greater investment.
Finally, online job vacancies have major potential as a source of information on the characteristics of job offers, job seekers and the duration of job postings. They are able to track labour market movements in real time, providing high-frequency data. Furthermore, they permit analysis of shifts in job profiles based on a large range of job requirements on skills, education and experience. Nevertheless, online job vacancies also have some shortcomings, including restricted coverage, biased samples and low international comparability, which future developments in data collection and treatment may be able to overcome.
