Robotisation benefits productivity and all jobs at pioneering firms
Research into robotisation and employment shows that technical innovation benefits pioneering companies over competitors, writes UNSW Business School’s Pauline Grosjean
It's time for speculation, even anxiety, about the future impact of artificial intelligence on our lives and on our jobs. While a thousand experts in the field, from researchers and entrepreneurs, have called for a pause in the development of artificial intelligence, it may be useful to look back at the lessons to be learned from another technological revolution: that of the mechanisation of jobs through the introduction of robots and other autonomous machines – a revolution that artificial intelligence could well reinforce.
Intuitively, economic theory predicts that jobs that are replaceable by robots, typically those performed by low-skilled workers performing repetitive tasks, will disappear. On the contrary, employees whose tasks are complementary to robots, typically executives and engineers responsible for design, management and marketing in robot production markets, will benefit from productivity gains and will see their salaries and job prospects improve.
The redistributive consequences of robotisation would then be clear, penalising the least qualified and benefiting the most qualified, and benefiting the owners of capital (and robots) even more. The overall impact on employment would depend on the net impact of the (negative) substitution effect on low-skilled jobs and the (positive) effect of productivity gains due to machines on the demand for qualified executives.
Increases in added value
Empirical studies confirm that the overall effect on the labour market is indeed ambiguous – both destructive and creating jobs – but for reasons and by mechanisms entirely other than those predicted by theory. In fact, the divide is not between substitutable jobs and jobs complementary to technology, but between companies that adopt technology and those that do not adopt it.
In France, as in the United States and the Netherlands, studies agree that robotisation benefits the productivity of companies that adopt these technologies and employment – for all employees’ jobs – in these companies. They are, in fact, experiencing strong increases in their added value (by 20 per cent between 2010 and 2015 in France), increasing their demand for labour (by 10.9 per cent) and gaining market share. All this is to the detriment not mainly of low-skilled employees, but of competitor companies that are late to the technology and of all the employees of these competitors, qualified or not.
The net impact on employment therefore depends on the positive effect on firms that adopt the technology, relative to the negative impact on competitors – and, therefore, who these competitors are, and where they are. A study on France covering a longer data series (1994-2015) and a more global definition of mechanisation reports the overall positive effects of mechanisation on employment, even on low-skilled employment, in sectors open to international competition. In these sectors, the technology losers are not found in France, but abroad. In sectors isolated from international competition, the overall effect on employment in France is rather neutral, and, in certain cases, negative.
Thus, if there is a pause in the development of artificial intelligence, it would undoubtedly be useful to take advantage of it to facilitate the adoption of these technologies by national firms, in order to avoid being overtaken by competitors.
Pauline Grosjean is a Professor of Economics at UNSW Business School, a Fellow of the Academy of the Social Sciences in Australia, and a Fellow of the Centre for Economic Policy Research (CEPR). Her research studies the historical and dynamic context of economic development. In particular, she focuses on how culture and institutions interact and shape long-term economic development and individual behaviour. This article was originally published in Le Monde, and was translated from French to English for republishing.