Will the trend towards increasing automation of production processes threaten the employability of graduates?
Definitions
Mechanisation: the introduction of machines or automatic devices into a process, activity, or place (Cambridge dictionary), where the process is controlled by a human operator.
Automation: the use or introduction of automatic equipment in a manufacturing or other process or facility (Cambridge dictionary), where the process is controlled by software.
Employability: the skills and abilities that allow you to be employed (Cambridge dictionary).
Graduate: a person who has a first degree from a university or college (Cambridge dictionary).
Introduction
In order to investigate the question of whether the trend towards increasing automation of production processes will threaten the employability of graduates I focus on the role of mechanisation and automation in attempting to reduce labour costs and consequently increasing the demand for high-skilled graduate workers. I examine which segments of the workforce will be most affected by automation, whether graduates are more employable than non-graduates, and whether the subject of the degree a graduate studied has any impact on employability. I also consider the increasing progression of automation and whether the threat to graduate employability will change over time.
The trend towards increasing mechanisation and automation
In order to examine the effects of mechanisation and automation on work it’s useful to look at the GDP as a measure of output per person and whether technology increases productivity. Figure 1 shows how for much of history productivity existed in a Malthusian economy (also referred to as the Malthusian Trap) of being linked to population growth (Malthus, 1798). The Malthusian Trap describes how any increase in productivity (measured by GDP) led to an increase in population, which resulted in decreasing the GDP per capita, thus sustaining a constant level of productivity. This idea is widely accepted as explaining the linear economic progress prior to the introduction of technology and social change of the industrial revolution.
Figure 1. Source: “Statistics on world population, GDP, per capita GDP, 1-2008 AD, Angus Maddison: IMF
The introduction of technology into the manufacturing processes in England and the United States of America in the eighteen century is widely accepted as being linked to the industrial revolution which took economic progress from being linear to enabling society to escape from the Malthusian trap and enabling greater output to be achieved with the same number of workers.
Contrary to the technological explanation for the onset of the industrial revolution, there is the idea that it may have been caused by social mobility (Clark, 2007). An analysis of wills showed that the wealthy had more offspring than the poor, and that this increasing upper class population disseminated it’s values across society, including education and saving for investment in capital resources (Baumol, 2002).
Although the causes of the industrial revolution are multiple and complex, and not solely limited to technology, the introduction of technological advances had a huge effect on the workers of the time. New technologies required skilled workers to install, operate and maintain the machines, and this demand for scarce skilled labour increased labour costs between 1800 and 1900 (core-econ.org), which resulted in an increase in capital investment to reduce labour costs.
Machines then, became labour-saving devices and mechanisation threatened jobs, disrupted entire sectors, and caused shifts in the production processes of every industry. But the effects were not the same or equal across all industries, sectors, and roles. Frey and Osborne (2013) illustrated in the shifting of certain production processes in the nineteenth century from artisan shop to steam powered factory increased the number of workers required but deskilled those workers through breaking the work into small, specialised sequences. Where the electrification of factories was introduced, more machinery could be utilised to automate production processes, resulting in a demand high skills in the production workers and an increase in the share non-production workers also employed (Goldin and Katz, 1998). In 1913, when Henry Ford introduced continuous-flow production the assembly lines were designed to around unskilled workers (Frey and Osbourne, 2013).
Demand for educated workers
Education and technology had to keep pace. The introduction of new technologies into the workplace resulted in demand for technologically-proficient workers to operate the new technologies. Without sufficiently educated and skilled workers the technology would fail to produce the expected productivity gains. These educated workers demand higher wages, increasing the labour costs and so driving further investment in capital and adoption of mechanisation and automation technologies. (Goldin and Katz 1995).
These examples show that the effects of mechanisation of the workplace was not as simple as machines replacing people. One interesting effect, for the purposes of this essay, is that industrial revolution technology had a profound and complex impact on productivity and employability (Baumol, 2002) through the increase in labour-saving machinery that created a demand for educated and skilled workers, and so increased labour costs fueling the introduction of further labour cost-reducing technologies.
New Growth theory, with its emphasis on knowledge creation and entrepreneurship, argues that physical assets such as capital (machinery) can only produce limited growth but that knowledge is an intellectual asset that enables increased productivity (Mankiw, Phelps & Romer. 1995) as knowledge is non-rival and non-excludable, meaning the value extracted is not restricted by the value of the asset. The endogenous model better explains how productivity can increase than the exogenous model informing Solow’s argument that productivity can increase purely through capital accumulation and technical progress (Solow. 1956), which seems to fall foul of the trap of introducing labour-saving machinery to reduce labour cost, but creating a demand for educated and skilled workers, and so increasing labour costs driving the introduction of further labour cost-reducing technologies.
Does automation affect all jobs equally?
If we equate the knowledge that Mankiw, Phelps and Romer refer to with skills and abilities of being employable (as per our definition above) we can consider how the automation of work affects workers of different skill levels. Frey and Osbourne found 47% of US employment are “at risk should these technologies materialise”.
- Routine manual work (e.g. assembly line worker), and routine non-manual work (e.g. book-keeper) required low and middle education levels respectively, and both would suffer a decrease in demand as automation substitutes these workers.
- Non-routine manual work (e.g. janitor) is likely to see no change in demand as automation does not perform non-routine tasks.
- Non-routine non-manual work (e.g. lawyer) is likely to experience an increase in demand with automation being a strong complementary (Frey and Osbourne, 2013).
Equating education level with skill level, as the Department of Education does (Graduate Labour Market Statistics 2017), we can take from Frey and Osbourne’s work that high-skilled workers performing non-routine, non-manual work are least susceptible to being replaced by automation (Michaels, 2010), and those jobs involving any kind of routine work are likely to be substituted with automation technologies. These low and middle skilled workers that are substituted by automation or had their wages reduced through computerisation will move to low-skilled service occupations (Autor and Dorm 2013).
Frey and Osborne’s (2013) and Autor and Dorm’s (2013) suggestion that this changing demand is ‘squeezing the middle’ of the employment market with jobs that require “cognitive and manual tasks that can be accomplished by following explicit rules” (Autor, Levy and Murnane. 2013) being substituted by automation leaves what Goos and Mannings (2003) call ‘lousy jobs’ and ‘lovely jobs’. ‘Lovely jobs’ are those that require creative thinking and the ability to confront novel situations successfully. Automation will complement these workers in “performing non-routine problem solving and complex communications tasks” but is unlikely to replace them (Goos and Mannings. 2003).
Employability of graduates
The Graduate Labour Market Statistics 2017 report by the Department of Education showed that in 2017 for the UK graduates and postgraduates shared a similar employment rate of 87% whilst non-graduates had an employment rate of 71%. The report showed that of the 87% of graduates in employment, 66% were in high skilled jobs whilst only 22% of non-graduates were in high skilled jobs. For the purposes of this report, high skilled can be defined as “a role where the tasks typically require knowledge and skills gained through higher education”, and Autor, Levy and Murnane’s (2003) definition of non-routine work.
Looking more closely at the young population (21 to 30 year olds) which are more likely to be impacted by the automation of their role in the coming decades, the report shows that 58% of graduates were in skilled roles compared to 18% of non-graduates. Comparing the overall to young we see that non-graduates have an 18% decrease in skilled roles and whilst graduates have a 12% decrease. There could be a number of reasons for the difference between the overall working age population percentage in skilled roles compared to the young population, but one possible impact for the future. is that the percentages in the young population will move towards the overall percentages over time as they learn more creative skills relevant to non-routine work whilst in their roles, which would suggest that graduates are less threatened by the automation of work than non-graduates. Another possibility is that the percentage of the young population in skilled roles (both graduates and non-graduates) will become the trend into the future. Although there is less of a difference for graduates than non-graduates, both may experience a decline in the percentage in skilled roles in the future, possible due to the effects of automation on routine work.
Automation effects on employability depends on timeline
We can accept that being a graduate makes a person more employable, and more likely to be employed in a high-skill job, and that the subject of study has less impact on these than having a degree. However, the question of whether this is sufficient to protect graduates from the threat of automation depends very much on the timeline one considers. The list of things computers ‘can’t do’ is rapidly becoming shorter and shorter (Bakhshi, Frey & Osbourne, 2015) as “developments in Machine Learning and Mobile Robotics, associated with the rise of big data, which allows computers to substitute for labour across a wide range of non–routine tasks – both manual and cognitive. As McCormack and d’Inverno (2014) put it, “We now know how to build machines that can ‘learn’ and change their behaviour through search, optimisation, analysis or interaction, allowing them to discover new knowledge or create artefacts which exceed that of their human designers in specific contexts”” (Bakhshi, Frey & Osbourne, 2015). This view of the future of the effects of automation on work does not need to distinguish between types of work (routine or non-routine) or skill level of the workers (graduates or not), it simply implies that automation will replace all jobs eventually.
As far back as 1933 Keynes predicted technological unemployment “due to our discovery of means of economising the use of labour outrunning the pace at which we can find new uses for labour” (Keynes, 1933, p. 3, in Frey & Osbourne, 2013) so as automation technologies improve they will undoubtedly squeeze not just the middle of the labour market but the high and low too.
Bakhshi, Frey & Osbourne’s statement that “there is nothing inevitable about the impact of automation on jobs and skills” may be historically true as they state their examples of looking back over the past few centuries, and their conclusion that the creative industries are likely to be the least disrupted by automation in the future also rings true in light of all of the thinking we have looked at, but as they say, only 24% of jobs in the UK have a high probability of being creative, which suggests a large percentage of jobs can and will be replaced by automation technologies.
Conclusion
In answering whether the trend towards increasing automation of production processes will threaten the employability of graduates we looked at the possible historic causes of the mechanisation of the industrial revolution era, and how mechanisation and automation had profound and complex impacts on jobs, workforce distribution, and employability since the industrial revolution and continues to today, including creating the trap of introducing technology to reduce labour costs whilst increasing the need for high-skilled workers and so increasing labour costs. We saw that the automation of work will have the greatest impact on middle and low skilled workers with high-skilled knowledge workers suffer least threat to employment. This is partly due to the need for high-skilled workers to operate new technologies and partly due to high-skilled workers being more likely to work in creative industries that are difficult to automate. We found that more graduates are in high-skilled jobs either through education and/or opportunity to learn at work, and can therefore conclude that over the next few decades graduates are least likely to suffer threats to their employability from automation.
However, over a longer time span, as computers improve their learning capabilities and become more able to tackle novel situations successfully, it’s my opinion that automation will threaten the employability of graduates. How far automation goes in changing the employability and nature of work for graduates and in fact all employees is a factor of how far we choose to look into the future. Autor, Levy and Murnane’s study looked back at a less than forty year time span but the history of mechanisation and automation goes back hundreds of years, and the future of automation has an unknown time span, making it impossible to predict how automation will affect work, jobs and employability in the future.
I can imagine a shift in how organisations invest in automation in the coming decades as they realise that digital transformation cannot be achieved by capital investment in technology alone and move to investing more in knowledge creation and turning those intellectual assets into a competitive advantage in line with the New Growth theory that will allow automation to escape the current trap of increasing automation to reduce labour costs increasing the demand for skilled workers which increases the labour costs. This change of approach in investment, along the rapidly advancing progress of Artificial Intelligence will allow automation to eventually replace all workers and completely reshape society.
References
Malthus, T. (1798). An Essay on the Principle of Population.
Maddison, A (). Statistics on world population, GDP, per capita GDP, 1-2008. International Monetary Fund.
Clark, G. (2007). A Farewell to Alms: A Brief Economic History of the World. STU – Student edition ed., Princeton University Press, 2007.
Baumol, W. J. (2002). The Free-market Innovation Machine: Analyzing the Growth Miracle of Capitalism. Princeton University Press
Frey, C.B. and Osborne, M.A. (2013). The future of employment: How susceptible are jobs to computerisation? Oxford Martin School. University of Oxford.
Goldin, C. & Katz, L. (1995) The Decline of Non-Competing Groups: Changes in the Premium to Education, 1890 to 1940. NBER, Cambridge, Massachusetts.
Mankiw, N. G., Phelps, E. S. and Romer, P. M. (1995). The Growth of Nations. Brookings Papers on Economic Activity, Vol. 1995, No. 1, 25th Anniversary Issue (1995), pp. 275-326. Published by: Brookings Institution Press.
Solow, R.M. (1956). A Contribution to the Theory of Economic Growth, The Quarterly Journal of Economics, Volume 70, Issue 1, February 1956, Pages 65–94.
Autor, D. H., & Dorn, D. (2013). The Growth of Low-Skill Service Jobs and the Polarization of the US Labor Market. American Economic Review 2013, 103(5): 1553–1597.
Goos M., Manning A. (2003) McJobs and MacJobs: The Growing Polarisation of Jobs in the UK. In: Dickens R., Gregg P., Wadsworth J. (eds) The Labour Market Under New Labour. Palgrave Macmillan, London.
Autor, D. H., Levy, F., & Murname, R. J. (2013). The skill content of recent technological change: An empirical exploration.
Graduate Labour Market Statistics 2017. Department for Education.
Michaels, G. (2010). The shrinking middle: how new technologies are polarising the labour market. Centerpiece.
Abel, J. R. & Deitz R. (2014). Agglomeration and Job Matching among College Graduates. Federal Reserve Bank of New York.
Bakhshi, H., Frey, C.B., & Osborne, M., (2015). Creativity Vs. Robots: The Creative Economy And The Future Of Employment. Nesta.