11 September 2019

The Technology Trap


It seems that barely a week goes by without some scare story being published in the press about how AI and automation are going to lead to mass unemployment and a world without work. The Technology Trap by Carl Benedikt Frey is a refreshing corrective to some of this doom-mongering.

Frey, an academic at the University of Oxford’s Martin School, starts by providing a thorough examination of how new technologies have influenced the power of labour and capital throughout history. He then moves on to look at how new technologies and AI may impact the future of work and society more broadly.

Most works on automation either ignore the past or, if they deal with it at all, merely focus on the Luddites, casting a with a superficial glance at the Industrial Revolution and how bad it was in the factories and the slums. However, as the historian T. S. Ashton shown, although people worked in grim conditions, things were rather more complicated than that. For many, factory work was a significant step up from a life of subsistence, squalor and back-breaking agricultural labour.

Where some authors give the impression that technological change only began in early 19th century Britain, Frey reaches further back in history for his examples. The book opens with Aristotle (‘When looms weave by themselves, man’s slavery will end’) and then moves to Ancient Rome to explain how they held back on industrialisation due to their suspicions of the private sector and the fact that their economy was reliant on slavery. We also get a fascinating snapshot of the Middle Ages and some of the technological innovations from that period such as improvements in horsepower and windmills.

Perhaps more striking is his argument that many of the key innovations of the Industrial Revolution could have been invented earlier. It was the power of guilds that hamstrung human progress. Frey draws parallels with the modern context, and expresses concerns that workers will again attempt to hold back innovation in order to protect jobs.

So, what lessons can Frey’s book teach us, and what are the implications for public policy?

First, it is a helpful reminder that fears over automation and new technologies are nothing new. They have persisted throughout history, and the concern has always been the same: new technologies will destroy jobs. Of course, new technologies do render some jobs obsolete, but they lead to new, better jobs and help to increase productivity and contribute to higher economic growth.

A related point is the distinction Frey draws between technologies which replace human labour and those which augment it. This is a subtle point which is often overlooked. It is more accurate to talk of roles being automated and not jobs. Although some of the tasks performed by humans might be at risk of being automated, this might actually free them up to focus on more enjoyable and challenging work.

Then there is the fact that automation and AI will bring huge benefits to humanity. This has always been the case throughout history and will be the same in the future. As such, we should embrace AI and automation for the many benefits it will bring. Whether it’s improving public services and freeing up civil servants to perform more worthwhile tasks, augmenting and refining health and social care, or the massive potential for improved productivity and higher economic growth, we should champion the role of AI.

What is more, we should be wary of those who try to hold back progress. Just as the guilds and the Luddites attempted to thwart technological innovation in the past, so too do trade unions and politicians today. Excessive regulation and calls for a tax on robots will stifle progress and prevent humanity from enjoying the many benefits of AI and automation.

However, as Frey reminds us, there are also downsides, especially in the short term. For example, often the benefits of increased productivity resulting from new technologies are not always immediately enjoyed by workers. This is known as an Engels Pause, and refers to the period during the industrial revolution in which profits and wealth increased dramatically, while wages stagnated for many workers. This does not always happen, but we should be prepared for the fact that the benefits of any major disruption may not be enjoyed by everybody, that inequality could increase and that many people might find themselves out of work, at least for a little while.

Concerns over inequality are often dismissed by those of us of the free market, neoliberal persuasion. Perhaps as a reaction against the fact that it seems to be the preserve of the left, our message has often been: ‘It doesn’t matter if inequality is increasing, as long as everybody is getting richer’. I do largely agree with this – the focus should be economic growth which will make us all better off in the long run, and we should not try to stifle this in order to achieve equality. However, given that any short term exacerbation of inequality could lead to the public demanding policies which hamper economic growth – such as thwarting technological innovation – we need to take this seriously.

Frey sets out some policies in his book about what to do to mitigate the negative impacts of automation. For example, he sets out the problems caused by the very restrictive planning system in a way that will be familiar to readers of CapX, but gives it an automation twist. It’s clear that planning restrictions limit supply and so increase house prices. This prevents highly skilled workers moving to the most productive areas while also creating a barrier for low skilled workers –  who may have been displaced by automation – to move to areas where they will be able to find work. For Frey, the rise of automation makes it even more imperative that planning restrictions are relaxed.

Then there is occupational licensing. As Frey points out, it’s obviously appropriate that some professions (such as doctors) require some form of formal registration, but for many others there seems to be no logical reason for doing so. Unfortunately, many countries insist on workers getting a license before they are able to perform even the most menial task. This again creates barriers to entry and has a disproportionate impact on the most vulnerable people in society and those who are most exposed to the negative short term impacts of automation.

Frey also looks at welfare. I fear that his novel idea of introducing a form of wage insurance for those who move into lower paid jobs, as well as extending means tested benefits, would risk complicating the welfare system further. I also feel that he was slightly dismissive of the idea of a universal basic income. Although he rightly pointed out the fact that it might not be politically feasible at the moment, and also challenged the flawed premise of David Graeber’s book, Bullshit Jobs, it would have been interesting if he had dwelt more on a policy which I feel has real merit.

But whatever you think of his proposals, it’s clear The Technology Trap has plenty to teach us, and should automatically be on the reading list of any serious policy maker or politician.

CapX depends on the generosity of its readers. If you value what we do, please consider making a donation.

Ben Ramanauskas is a research economist at Oxford University