White Paper

The digital revolution is rapidly changing the composition of the workforce across economies. In particular, a confluence of improvements in a wide range of related technological areas, including sensors, machine learning, automation, and robotics, is making technology more labor-saving and potentially less job-creating.

Our London Futures insights series focuses on the London economy and what it needs to do to maintain and reinforce its position as a leading global business hub. Our latest report in the programme focuses on the challenges and opportunities that technology presents to London. The findings indicate that a significant shift is occurring in the labour market. Jobs that do not need to be done in London, or can be fully replaced by technology, will continue to leave the city. However, the job losses will be outweighed by new jobs requiring skills that involve creativity, complex problem-solving and high technical content. The report brings together research from Oxford Martin School academics Carl Benedikt Frey and Michael A Osborne on the potential impact of automation on jobs in the UK and London over the next two decades, and a Deloitte survey of 100 London based organisations, exploring the new jobs that will be created, the skills that will be needed, and the implications for current working practices.

In this new report, Oxford Martin School academics Dr. Carl Benedikt Frey and Associate Professor Michael Osborne examine a pressing subject increasingly in the headlines: the changing nature of innovation and work, and the associated implications for the future of employment and society more widely. We are also very pleased to say that this report marks the launch of the Oxford Martin Programme on Technology and Employment, a long-term programme of research at the University of Oxford supported by Citi that will focus on many of the areas covered in this report.

We examine how susceptible jobs are to computerisation. To assess this, we begin by implementing a novel methodology to estimate the probability of computerisation for 702 detailed occupations, using a Gaussian process classifier. Based on these estimates, we examine expected impacts of future computerisation on US labour market outcomes, with the primary objective of analysing the number of jobs at risk and the relationship between an occupation’s probability of computerisation, wages and educational attainment. According to our estimates, about 47 percent of total US employment is at risk. We further provide evidence that wages and educational attainment exhibit a strong negative relationship with an occupation’s probability of computerisation.

Over the last hundred years, technological change has fundamentally transformed labour markets in most industrial nations. In recent decades, the arrival of new digital technologies has displaced a wide range of routine work, including the jobs of bookkeepers, assembly workers, cashiers and telephone operators. At the same time, the appearance of more than 1,500 new occupational titles in official classifications since the computer revolution of the 1980s, suggests that a host of new occupations have emerged.4 In the United States, around 4.4% of total employment in 2000 was in types of jobs that did not exist in 1990.5 Many new occupations are directly related to the arrival of the personal computer (PC) in the 1980s and the subsequent internet (or ‘digital’) revolution that has rolled out since the early 1990s. As computer technologies disrupt the world of work, making some occupations redundant in the process, the key challenge for governments is to help support the shift of workers into new and emerging employment opportunities. A concern is therefore that the companies leading the digital age have created relatively few new jobs compared to the giants of the early computer revolution: while IBM and Dell still employed around 430,000 and around 110,000 workers respectively in 2013, Facebook’s global headcount reached only 7,185 in that year. We turn next to examining the broader implications of this trend for the UK economy.

Canadian workers are generally more skilled than American workers in literacy, numeracy and problem solving in a technology‑rich environment (Statistics Canada 2013). However, workers who possess certain skills are not guaranteed to employ those skills in their job. Furthermore, workers may employ many other skills in the workplace (Frenette and Frank 2017). This study compares the skill requirements of actual jobs held by workers in Canada and in the United States. In total, 35 job skills are examined. The study finds that the skill level required in Canadian jobs is significantly higher than for American jobs in 30 of the 35 areas examined. The Canadian advantage is particularly large in science, technology, engineering and mathematics (STEM) and other technology‑related skill areas. In these areas, the relatively larger proportion of non‑university graduates in engineering, manufacturing and construction, and the higher numeracy proficiency in Canada account for almost all of the gap in skill level. The one group facing lower skill requirements in Canada is university graduates. In general, the jobs that they hold require lower reading comprehension, writing, social and management skills compared to their counterparts in the United States. Some implications for future research are discussed.

Given the time and money invested in higher education by students, parents and governments, there is considerable interest in the economic outcomes of postsecondary graduates. Most assessments of recent graduates have focused on their short-term, early labour market results. As new entrants to the labour force, recent postsecondary graduates may be particularly vulnerable to the economic cycle. Consequently, comparisons of short-term outcomes across graduating cohorts may be highly dependent on prevailing economic conditions and may not reflect the longer-term returns on investments. This is the first study to compare the long-term labour market outcomes of two cohorts of young postsecondary graduates using linked census and tax data. Specifically, graduates who were 26 to 35 years old in 1991 were followed from 1991 to 2005 (when they were 40 to 49 years old) and compared with a similarly aged 2001 cohort, which was followed from 2001 to 2015. The results suggest that median cumulative earnings were higher among members of the more recent cohort of male and female postsecondary graduates. Increases were observed across all postsecondary levels and across most major disciplines where sample sizes were large enough to permit analysis. Also, no discipline registered a decline in cumulative earnings. Although the economic conditions faced by the 2001 cohort over the 15-year study period were generally more favourable, this cohort also registered higher earnings than the 1991 cohort during the latter portion of the period (i.e., when the 2001 cohort was faced with an economic recession). Furthermore, the initial labour market conditions upon graduation (an important determinant of career earnings) were similar for both cohorts. The improvements in long-term earnings for postsecondary graduates are important in light of the significant increase in the number of graduates over the period. However, the results also indicate that the number of years of union membership declined or remained steady across cohorts of male and female postsecondary graduates. Furthermore, while women with postsecondary qualifications registered increases in the number of years of employer-sponsored pension plan coverage, their male counterparts experienced mixed results depending on their level of postsecondary studies.

This study examines the relationship between occupational skill requirements and educational attainment (the highest level completed and the field of study). Using the 2011 National Household Survey matched to data from the Occupational Information Network (which contains information on occupational skill requirements), the study uncovers many new findings on the skill requirements of jobs held by Canadians aged 25 to 34 with different educational qualifications. First, skill‑level requirements in all areas generally increase with higher educational levels. There are three notable exceptions to this trend, however: requirements for technical operation and maintenance skills do not generally increase with more schooling; doctoral graduates require considerably less resource management skills than other university graduates; and professional‑degree holders have jobs that require lower mathematics skills than other university graduates, although this is entirely because of law graduates. Moreover, skill requirements are almost always higher among bachelor’s degree holders and college graduates of specific disciplines, compared with high school graduates. Second, skill requirements vary considerably by field of study. For example, bachelor’s degree holders in architecture, engineering, and related technologies generally work in multidimensional jobs requiring diverse, high‑level skills. In fact, these graduates rank at the top or very close to the top in eight of the nine skills examined. In contrast, bachelor’s degree holders in three fields (education; visual and performing arts, and communications technologies; and humanities) generally rank lower than other bachelor’s degree holders in most skills examined. However, the skill requirements of college graduates and bachelor’s degree holders are very different across fields of study. While skill requirements are almost always higher among bachelor’s degree holders than among college graduates from the same discipline, the relative ranking of disciplines tends to differ for each level. Finally, there are considerably more gender differences among college graduates than among bachelor’s degree holders with regard to skill requirements by field of study.

Increasing women’s participation in male-dominated trades has been identified as a means of improving the supply of skilled tradespersons in Canada, creating a more diverse workforce, and increasing women’s wages. However, little information exists about women’s decision to enter male-dominated apprenticeship programs and their subsequent labour market outcomes. This study addresses both information gaps by examining the characteristics associated with women selecting male-dominated apprenticeship programs and their labour market outcomes relative to men who selected the same types of programs. A range of outcomes are examined, including employment status, self-employment, obtaining a job related to the trade of study, hours worked per week, union membership, a series of job benefits (such as extended health care, sick leave, and retirement plan benefits) and hourly wages. Results indicate that women who were Canadian-born, who were older, or who had a father with a trades certificate were more likely than other female apprentices to choose a male-dominated program. Women who studied in male-dominated apprenticeship programs generally had poorer labour market outcomes than their male counterparts. Among apprentices who selected male-dominated programs, women were as likely as men to receive sick leave benefits in their jobs but received lower median hourly wages than men. Fewer gender differences were found among apprentices who selected female-dominated or mixed (neither male- nor female-dominated) programs. However, while the median wages of women who studied in female-dominated or mixed apprenticeship programs did not differ significantly from those of their male counterparts, these women did earn less than men at the higher end of the wage distribution (75th percentile).