There are four narratives about how increasing automation of manual and cognitive tasks will affect employment:
- Many people will no longer have paid employment, and will depend on government assistance.
- Many people will no longer have paid employment, but automation will create highly productive and equitable economies that enable people to pursue their own creative and social endeavours.
- As in the past, new technologies will create new jobs for people.
- Less a narrative than an admission that we have no idea at all what will happen this time.
Is information on employment trends helping identify which of these is more, or less, likely?
The Wall Street Journal recently highlighted that jobs in the US generally involving high levels of non-routine tasks are growing, while those involving more routine tasks are declining. They particularly noted the rapid rate of growth in non-routine cognitive (or what they call “knowledge”) workers.
- Management & professional occupations represent “non-routine cognitive” workers, or WSJ’s “Knowledge workers”
- Sales and office occupations are “routine cognitive” workers
- Service occupations are “non-routine manual” workers
- Production, transport, construction, etc occupations are considered as “routine manual” workers
This is a somewhat simplistic characterisation, since most jobs will involve routine and non-routine tasks. This is acknowledged in the Wall Street Journal, but they consider the categories generally reflect the degree to which routine tasks are associated with the work.
Josh Zumbrun, the author of the WSJ blog post, suggested that these occupation trends might be due to developments in automation replacing routine manual and cognitive tasks and jobs. But instead of fewer jobs overall, as some predict, he infers that the rapid rise in “knowledge” jobs is a result of technologies creating new types of work. That’s speculation because he doesn’t provide supporting evidence. To me the analysis isn’t a compelling example of technologies creating new jobs.
Does NZ show a similar trend?
Statistics NZ uses the Australian and New Zealand Standard Classification of Occupations, which doesn’t characterise occupations the same way as the US so a direct comparison isn’t possible. (Our “technicians and trade workers” category, for example, is split across cognitive and manual categories in the US.) And the job classification scheme here has been revised in the 1990s and early 2000’s, so a long time series of data isn’t readily available. None-the-less we can see some patterns:
The picture isn’t the same as in the US. While the number of people in “Professional” and “Managerial” jobs (part of the non-routine cognitive workers group) has grown and the number of labouring jobs has declined, the other occupations haven’t seen much change. The increase in professional and managerial workers seems to be attributable to the rise in the service sector, and the number of people in employment (an increase of about 400,000 between 2004 and 2016), rather than declining jobs in other occupations.
Job polarization – the decline of “middle-skill” occupations and the rise of high- and low skill occupations – isn’t obvious in the NZ data. That may be due to how our employment data is organised, the fact that we don’t have marked polarization, or because the characterisations of occupations is too crude.
Knowledge-intensive services on the rise
Organising employment data in a different way shows changing patterns in the types of firms and industries in New Zealand. The Ministry of Business, Innovation and Employment (MBIE) noted the rise in what they call the “knowledge-intensive services sector” in 2014. They looked at numbers of employees by type of firm rather than occupation. This illustrates job growth in the service sector, and the decline or stagnation in jobs in the goods-producing and primary sectors, a trend common elsewhere too.
Knowledge-intensive services include professional, scientific and technical services (such as science, engineering, design and consultancy services), finance and insurance firms, information, media and telecommunications, as well as several other occupations.
“What these firms have in common is provision of the specialist technical and professional expertise that underpins exporting and other activities across all sectors of the economy. Salaries and wages in this sector are 40% ($31,000) above the New Zealand average, reflecting the high qualification levels.”
But what about the future?
Whether most of these knowledge workers find meaningful employment in other areas remains to be seen, although I’d expect those with strong analytical skills to find employment elsewhere. The other critical question is whether automation enables these types of firms to provide more valuable services.
One analysis of jobs in the US banking sector indicated that automatic teller machines led to more rather than fewer bank teller jobs (at least up till 2010). This was interpreted as being due to ATMs decreasing bank operating costs, enabling them to open more branches, and to focus more on valuable non-automated services.
Collaborative robots, or cobots, working with factory workers also fits the narrative that automation can create new types of work.
However, a couple of examples isn’t sufficient to confidently predict the consequences of widespread automation of tasks.
So, I’m sticking with the “no idea what is going to happen” option for the moment. Obviously, I hope we’ll end up with narrative 2 or 3 that I mentioned at the start.
The ATM and cobot examples illustrate that it’s management decisions on value creation and how to use staff most effectively that matter when new technologies are adopted. Are you just replacing people to save costs, or are you improving what you do?
For New Zealand, discussion about automation, robots, and artificial intelligence focuses on the risks (or not) they may pose to jobs. While obviously an important factor, it can detract from the larger issue of our declining productivity [Pdf]. While near full employment is good for society, we also need to be improving our economic performance so that we are better able to address other emerging challenges.
Saving costs by replacing people with software or machines doesn’t necessarily improve the value of what we produce or provide. We need to produce more valuable things. That’s going to take more than technology. Automation without augmentation (in both economic and social senses) is a losing position. We need to think as much about what we produce as how we produce it.
Featured image: CC pixabay