This week, we continue a special series of blogs on the future of economic development, work, and wages in developing countries.
Last week we presented the three headlines from our new paper.
Specifically, we looked at the rise of a global ‘robot reserve army’ and its likely profound effects on developing countries. We made three points. First, that automation is not just a rich country issue. The developing world is both affected by automation trends in high-income countries and is itself catching up in terms of automation. Second, that automation is not only about technology. The current debate focuses too much on technological capabilities, and not enough on the economic, political, legal, and social factors that will profoundly shape the way automation affects employment. Third, that policy makers need to pay more attention to stagnating wages than unemployment. In contrast to a widespread narrative of ‘technological unemployment’ (© John Maynard Keynes), we argued that a more likely impact in the short-to-medium term at least is slow real-wage growth in low- and medium-skilled jobs as workers face competition from automation.
In this blog we take a look at the question of employment generation and what actually is automation. In the following blogs we’ll look at what drives automation; assess whether 1.8bn jobs in the developing world will really be lost by automation; consider the public policy responses to automation, and finally conclude with a wrap up blog.
Employment growth in crisis?
Employment generation is crucial to spreading the benefits of economic growth broadly and to reducing global poverty. And yet, emerging economies face a contemporary challenge to traditional pathways to employment generation: automation, digitalisation, and labour-saving technologies.
A broad range of international agencies have recently flagged issues relating to the future of employment, and the consequences of automation and deindustrialization in a number of global reports. Employment prospects have also come into sharp focus because of the contested experiences of “premature deindustrialization” (see Felipe here and here, Palma as well as Rodrik) and weakening employment elasticities of growth. The World Bank estimates 1.8 billion jobs or two-thirds of the current labour force of developing countries are estimated to be susceptible to automation from today’s technological standpoint. Cumulative advances in industrial automation and labour-saving technologies could further exacerbate this trend. Or will they?
How real is the widespread concern? Heintz examines employment growth and the productivity growth rate in 35 countries between 1961 and 2008, and finds that increases in the productivity growth rate slow down the rate of employment growth, and that this pattern is getting stronger over time. In the 1960s, a one percentage point increase in the growth rate of productivity reduced employment growth by just 0.07 percentage points. However, in the 2000s, that same one percentage point increase in the growth rate of productivity reduced employment growth by a substantial 0.54 percentage point.
Several possible explanations are as follows: (i) it could be that increases in productivity over time are reducing the employment elasticity of growth; (ii) it could be that the proportion of wage labour is increasing; or (iii) it could be that increases in real wages, employers’ social contributions, or strengthening labour institutions are raising unit labour costs and dampening employment creation, though this is ambiguous in empirical studies. A meta-review of 150 studies of labour institutions covering minimum wages, employment protection regulation, unions and collective bargaining, and mandated benefits with an emphasis on studies in developing countries, found that in most cases, effects are either modest or work in both directions in terms of productivity.
Automation and employment
Global automation trends are a candidate for explaining reduced employment growth in manufacturing and agriculture. Overall, the World Bank estimates that “the share of occupations that could experience significant automation is actually higher in developing countries than in more advanced ones, where many of these jobs have already disappeared”. However, they note that the impact will be moderated by wage growth and the speed of technology adoption. There are numerous estimates of job displacement and much in the way of gray literature.
The developing world is both affected by automation trends in high-income countries (HICs) and is itself catching up in terms of automation. Indicative of this, the International Federation of Robotics (IFR) reports that Asia is currently the “strongest growth market” in a “significant rise in demand for industrial robots worldwide”. A double-digit growth trend includes not only China, Korea, and Japan but also emerging economies in South East Asia. The IFR estimates that by 2019, more than 250,000 units of multipurpose industrial robots will be installed in Asia on a yearly basis, with the main industries driving demand in robots being the automotive, electrical/electronics, metal, and machinery, as well as the rubber and plastics industries.
This only captures the more easily measurable demand for robotics hardware and does not take account of the widespread use of software in the context of economic production. In some domains of automation, emerging economies are, in fact, ahead of many OECD countries – a phenomenon known as ‘leapfrogging’. Take, for instance, the opening of Beijing’s first driverless subway line in 2017 or the popularity of the mobile phone-based financing platform M-Pesa in Kenya illustrate.
The digitalization and automation of economies raises the question of what lessons the developing world can draw from extant evidence. “Late developers” are facing the digital revolution considerably earlier and under different conditions than today’s advanced economies. There is thus an increasing worry, as Frey et al. argue, that “increased automation in low-wage countries, which have traditionally attracted manufacturing firms, could see them lose their cost advantage and potentially lose their ability of achieving rapid economic growth by shifting workers to factory jobs” which today’s high-income countries used to have.
What do we mean by automation?
We have so far talked about automation without defining it – and indeed the concept is more difficult to define than might seem at first glance. Throughout history, humans have used tools to save time and effort when completing laborious tasks and thanks to innovation, such tools have gradually increased in sophistication. Today, the spectrum of what economists call “physical capital” ranges from simple manual tools to intelligent machines. One could argue that a “robot” is simply a highly advanced version of a tool which requires minimal (manual) human input for completing a task, although currently all machines still require considerable human intervention in their design, production, installation, and maintenance.
The potential of artifical intelligence (AI) is to move machines beyond human oversight, at least in everyday operation. An intelligent machine performs a set of complex tasks autonomously and may be capable of adapting to new and changing circumstances, i.e. “learning”. Workhorse animals could be considered a biological equivalent of complex machines and have been used in transportation and agriculture since at least the agricultural revolution in 10,000 BC. Contemporary automation often tends to be associated with physical hardware such as industrial robots, but also includes software which plays a critical role in service automation (see the work of Willcocks & Lacity). The wider process of structural economic change toward an automated economy has been referred to not only as a digital transformation but as the “fourth industrial revolution”.
To sum up: current technological change is set to heighten tensions between employment growth and productivity growth in both developed and developing countries. While a consensus is emerging that labour-saving innovations are impacting developing and developed economies and while there is a mushrooming of guesswork on the magnitude of its labour market effects, it is, paradoxically, less clear precisely what ‘automation’ (derived from the Greek words for ‘acting of one’s own will’) or ‘robots’ (derived from the Czech word for ‘servants’) actually is.
Are these radically novel techno-social phenomena or simply the latest of many iterations of the Industrial Revolution which has been unfolding for two and a half centuries? Depending on this, our understanding of the implications of technological change for economic development would either (i) draw on tried-and-tested theories of political economy or (ii) be in need of radically new approaches.
We’ll continue the series of blogs with some discussion next of the drivers of automation which aren’t only technological.
Lukas Schlogl is a Research Associate with the ESRC GPID Research Network at King’s College London. He works on structural change, digital transformation, and political behaviour in developing countries.
Andy Sumner is a Reader in International Development in the Department of International Development, King’s College London. He is Director of the ESRC Global Poverty & Inequality Dynamics (GPID) Research Network.
