The Future of Work

Jon Cruddas Excerpts

Thursday 19th November 2020 Westminster Hall

[Jon Cruddas (Dagenham and Rainham) (Lab)]

I welcome today’s debate and the recent establishment of an all-party group on the future of work, which will be helped by the Institute for the Future of Work.

I want to emphasise a couple of points that have already been made and talk about some of the assumptions that form much of the debate about the future of work. Even before the pandemic took grip, the future of work was attracting widespread attention, reflecting a widespread belief that the robots were coming. McAfee and Brynjolfsson of the Massachusetts Institute of Technology have suggested that automation and artificial intelligence are powering a second machine age that is equivalent to the first industrial revolution. Writers such as Martin Ford have confirmed these technological shifts, and such books shape a narrative of epic technological change, often described as the fourth industrial revolution.

Numerous tabloid headlines have reported the likely displaced jobs through automation. Many of the most threatening estimates can be traced back to a single source: a 2013 article by Carl Frey and Michael Osborne that suggested that nearly half of all jobs classified by the US Bureau of Labor are vulnerable to automation. This has been used to suggest the demise of many millions of blue-collar jobs. Alongside that, in “The Future of the Professions”, Richard and Daniel Susskind suggest that technological forces will dramatically rework white-collar jobs—lawyers, consultants, accountants and health professionals. Reading these pieces makes one feel that almost no job is safe, but can we be so certain?

The approach to automation also corresponds, as has been said, with renewed interest in universal basic income. On the right, Milton Friedman, Hayek, Charles Murray and Richard Nixon have all embraced it to roll back the welfare state and replace it with an individualised transaction between state and consumer. The left-wing case tends to focus on the basic human right to a level of subsistence to shield against work poverty or job loss. Historically, Tom Paine, Bertrand Russell, JK Galbraith and Lyndon Johnson and many others have supported it. The policy is also embraced by many silicon valley titans, presumably to offset their personal responsibilities for structural unemployment. We have seen an upsurge in interest in basic income initiatives such as the Alaskan oil dividend and the UBI pilots in Finland, Scotland, Canada, Oakland, the Netherlands and New Zealand.

Much of this appears to be fuelled by the belief that the robots are coming soon. Work is ending with the wholesale replacement of humans by machines. That begs the obvious question: what do we really know about the future of work? How reliable is the data? My basic point is that technology is not destiny. Assertions of technological disruption have always been around. In the 1930s, Keynes argued that by 2030 the average working week would be 15 hours long as new methods of economising on labour exceeded its use.

[Mr Philip Hollobone in the Chair.]

We should be cautious about headlines on the future of work that often derive from a single contested source. The Frey and Osborne analysis estimated that up to half of British jobs were threatened by automation. That was famously used by the Bank of England two years later to assert that 15 million jobs were at risk. How confident should we be about such assertions, not least because wildly different estimates co-exist? For instance, McKinsey Global has suggested that only 5% of jobs are candidates for full automation. Clearly, as has been mentioned, jobs will be created by automation, not just destroyed. For example, Frey and Osborne’s projections do not consider new jobs created in health and social care, the creative industries, leisure, in sectors that require interpersonal human skills, and in the technology and telecommunications sectors. Moreover, many of these studies on the effects of automation do not contain any timelines. I read this morning about a McKinsey report based on an analysis of 800 occupations, which estimated that half of all work activities could be automated by 2055, but then said

“this could happen up to 20 years earlier or later”.

So the data is pretty unreliable.

Many of these studies also underplay patterns of labour regulation that can help or hinder automation. The classic example is one of de-automation, literally, through the resurgence of thousands of hand car washes and the disappearance of the automated alternative, driven by the exploitation of migrant labour in the deregulated British labour market. These are questions of politics, not technological destiny.

To return to the statistics briefly, the Organisation for Economic Co-operation and Development estimates that 9% of jobs are automatable, yet it also suggests that that is an overestimate, given the likely political and social constraints, redeployment and future job generation. It concludes that

“automation and digitalisation are unlikely to destroy large numbers of jobs.”

A TUC paper estimates between 10% and 30% of jobs to be at risk, yet concludes that the likelihood is that those jobs could be replaced by new occupations and professions. The evidence is at best inconclusive. The UK Government do not appear excessively worried. Their industrial strategy White Paper suggests a growing demand for high-skilled jobs and anticipates an extra 1.8 million new jobs in the next 10 years. That was before the pandemic, obviously.

In a thorough review of the literature, Phil Brown and his colleagues at Cardiff University concluded that technology is not destiny and that human decisions will determine the future of work. Their study states:

“Most studies focus on the potential for automation, without incorporating into their models economic and social factors that may stimulate or deter the replacement of workers by technology.”

In other words, politics. So the future is far from certain; it depends on the policy and political choices that we make.

In recent years, UK labour markets have seen a significant increase in atypical work, including elements of the gig economy. Prior to the pandemic there were some 5 million self-employed, 1 million workers on zero-hours contracts and 800,000 agency workers—since 2008, there have been rises of 24%, 450% and 46% respectively. Those comparatively high numbers are the product of our labour law and policy choices, resulting in work that is less regulated and protected and contributes to sluggish wage and productivity growth.

I will conclude with a few points. First, there is little consensus about future technological disruption. Secondly, the research is contested and, at best, unclear. Thirdly, it is prone to speculation and contains serious methodological flaws. This suggests that a more cautious approach is required, with an emphasis on our political choices rather than reverting to conjecture fuelled by technological determinism. There is nothing inevitable about the future of work. There are political choices about creating and rewarding good work and in upholding the dignity of labour, cruelly exposed by the pandemic. That is why it is so good that we are discussing the subject in Parliament this afternoon. I congratulate the hon. Member for East Renfrewshire (Kirsten Oswald) on securing the debate.