Abstract
Much of the discussion of the impact of automation on employment has focused on developed countries. Y et, for developing countries, a key concern is the prospect of "reshoring" or "nearshoring"-the opposite of offshoring- in which production particularly of labor-intensive manufactures shifts from developing back towards developed countries. These shifts in the global division of labor are enabled by automation in such critical industries as apparel and electronics, that have provided developing countries with strategic entry points into global markets and employ large numbers of workers. For example, the more readily and cheaply that apparel, sewing, and electronics assembly can be automated, the less readily can developing countries retain their competitive advantage based on lower labor costs. For lead firms in global supply chains, reshoring would provide the considerable advantages of lower transport costs, as well as shorter lag times between design, production, and final sales, enabling more just-in-time production. While there is not at present an overall trend towards reshoring, recent empirical studies find evidence that the increased use of robotics and other automation technologies in developed countries is associated with reshoring.1 The concern for developing countries is heightened when one considers the figures for risk of potential automation in different occupations via a method developed by Carl Benedikt Frey and Michael Osborne and applied in theirs and other widely-cited studies.2 The purported risk of potential automation by computer-controlled equipment in the next ten years or so is 99% for hand sewers, 89% for sewing machine operators, and 95% for electrical and electronic equipment assemblers. These figures are high because the occupations are classified as being comprised largely of routine work that is judged to be more amenable to automation. Since developing countries have much higher shares of such occupations than developed countries, the method inevitably yields the result that the country-wide share of jobs at high risk (70% or greater) of potential automation is considerably higher in developing than developed countries.