Q&A on: Robotics and reshoring - Employment implications for developing countries

This new publication explores the implications of increasing automation through case studies across four different sectors in developing countries. The volume examines how robotics and reshoring can affect both the quantity and quality of jobs, focusing on workers at the shop-floor level. This Question and Answer gives a brief overview of the motivations for producing the volume and its findings.

Article | 13 August 2020
Q: What is the main focus of this new publication, and what motivated the ILO to publish it?

A: Much of the discussion of the impact of automation on employment has focused on developed countries. Yet, for developing countries a key concern is the prospect of “reshoring” or “nearshoring” in which production shifts from developing countries back towards developed countries. These shifts would be enabled by automation in the labour-intensive sectors that have provided developing countries with strategic entry points into global markets and continue to employ a large number of workers, often disproportionately women.

The COVID-19 crisis has lent renewed urgency to the discussions on such restructuring of global supply chains. Relatively absent in these discussions, however, is the sense of how automation plays out on the shop-floor. This has motivated the industry case studies in this volume, addressing the implications of the increased use of robots and ITC-enabled automation in the apparel and electronics industries as well as in retail warehousing and business process outsourcing.

Q: What are some of the key findings emerging in this publication?

A: Based on the case studies, automation may not present an immediate threat to jobs on a large scale in developing and emerging countries, but reshoring and near-shoring concerns should not be overlooked.

The research finds that there are several bottlenecks to adopting new technologies at the shop floor level. There are still substantial limitations to the use of robots in activities which, although repetitive and thus seemingly easy to automate in theory, require (among others) flexibility and dexterity that machines are not yet fully able to replicate. Much concern about automation is based on labour displacement, but as the case studies highlight, technology can also affect jobs through complementarity effects, where robots work in collaboration with humans, and market expansion effects. Moving away from net job creation and destruction estimates, it is essential to assess how jobs in these sectors would be affected at the task level and the implications for job quality.

Overall, on the basis of the analysis in the four case studies, it is fair to say that largescale replacement of humans by machines in the sectors studied as well as the possibility of large-scale reshoring is unlikely in the near future. The actual impact on employment would depend on the ability to overcome the technological bottlenecks that exist, and the extent to which products are for niche markets or mass markets, as well as economic factors such as relative costs. Nevertheless, much of what actually happens and how fast these changes take place will depend on policies and social dialogue at the sectoral, national and global levels.

Q: According to the case studies, how could automation affect the quality of jobs in these industries?

A: It is important to note that the impact of greater automation on workers is not restricted to job displacement but also affects working conditions and compensation. While some studies argue that robots could improve working conditions by completing dangerous or less desirable work, the case can also be made that workers may be pressured to increase their pace when they work in collaboration with robots. Greater use of automation technologies in assembly lines could also reduce the number of workers or their working hours, with potentially negative consequences on wages. This is even more crucial given that many of the countries with large electronics assembly industries are not signatories to international conventions on freedom of association and collective bargaining.

Q: How will reshoring affect workers in LDCs?

A: For most developing countries – except in some middle-income countries – the risk of job loss would not only be from automation in the country, but also, and more immediately, from automation in advanced countries.

Recently, debates on reshoring have been gaining space. Studies suggest that a developing country with low capital stock and low wages attracts offshored labour-intensive activity. As the country develops and capital is accumulated, wages rise, and offshoring appears less attractive. While this is temporarily countered by a decline in rental cost of capital, at some point in time, the increased wages will outweigh the decline in the rate of capital rental, leading to reshoring.
Research finds that effects are particularly negative for low-skilled machine operators and technicians in manufacturing, and high-skilled service workers in managerial and professional occupations.

Q: What types of bottlenecks could be anticipated when adopting and deploying new technologies in these industries?

A: One key bottleneck is the superiority of human capabilities over that of machines. Sewing continues to be predominately done by the familiar process of workers manipulating pieces of fabric by hand through stand-alone sewing machines. Wages tend to be low in the industry, creating less incentive to automate, but a fundamental impediment is technical. This results from the pliability of fabrics, pieces of which need to be accurately aligned before they are sewn, something the human hand and eye can readily accommodate but which poses daunting challenges for automation.

In the electronics industry, production involves distinct processes in the manufacture of components and the assembly of final products, each with unique technological needs and bottlenecks. The assembly of electronic products necessitates handling small and fragile parts, put together in compact, tightly-packed products. In addition, rapid technological progress has led to short product life-cycles. These bottlenecks create the need for adaptable and reusable robots fitted with force and vision sensors to improve the handling of miniaturized parts, allowing costs to be amortized over longer periods of time.

The greatest technological bottleneck of warehouse automation is the challenge of integrating multiple automated systems. For example, both the Amazon and STO Express systems require workers to manually pick up and scan items before transferring them to a conveyer belt. This challenge arises due to the variety of different shapes and sizes of packages; while a human hand can easily adjust to grip objects of different shapes, robotic “hands” do not have the same level of dexterity. An institutional bottleneck that has significantly slowed the implementation of robotic warehouse systems by other retail firms is the tendency for large retailers to acquire robotics companies, effectively removing their technology from the market. In other words, there are market barriers created by large retail firms.

The bottleneck for complete automation in the service industry appears more related to consumer preferences and operational costs than technological shortcomings. With their low labour costs relative to developed nations, call centres in countries such as India and the Philippines are likely to still remain competitive.

The technicalities of the bottlenecks differ in each case, but come down to the lack of quickly adaptable technology in the cases of the electronics, warehouse and clothing sectors and consumer preference in the service sector. Nevertheless, removing the technological bottlenecks and changing the attitudes of consumers towards automation is just part of the equation. What actually happens and how fast these changes take place will depend on policies and social dialogue at the sectoral, national and global levels.