Accelerating automation exacerbates inequality, widens wealth gap
Norton 17 Andrew Norton is director of the International Institute for Environment and Development (IIED).
[Norton, “Automation and inequality The changing world of work in the global South,” International Institute for Environment and Development, 2017, , accessed 6/28/2023] //abcchs
There is a range of mechanisms through which accelerating automation and digital innovation might affect inequality within countries. These include: boosting the advantage of wealth over labour; hollowing out of the workforce (relative declines in mid- and lower-skilled employment share); weakening of labour institutions; and erosion of the tax base (thereby weakening state capability for redistribution). As a recurring caveat — all of these effects can be influenced and reversed by policy and social action, so speculation about impacts needs to take that into account. The argument that automation will increase the advantage of capital over labour is simple and intuitive. Owners of robots will benefit more than workers they replace. This argument can also be seen as ‘Piketty amplified’, as the result is to boost returns to capital further ahead of returns to labour at the level of a whole national economy. In most industrialised countries, labour’s share of national income has been falling since the 1970s. And the process is also visible in some middle income countries, with a particularly steep drop in China.17 Automation is likely to increase this process under most scenarios, further concentrating wealth and income in fewer hands. The hollowing out of the workforce is already measurable in the manufacturing sector. Rodrik1 demonstrates that automation has been taking out more low-skilled jobs than medium- or high-skilled ones, increasing inequality. The data underpinning this important observation are quite broad-based, drawn from 40 countries (with a bias towards Europe) over the years 1995–2009, and providing a breakdown of manufacturing employment by three worker types: low skill, medium skill and high skill. Rodrik found that not only was there a steady reduction in manufacturing employment across the cohort, but also that virtually the entire reduction in employment over time comes from the low skill category. At the economy-wide level, however, this trend is not yet clear in many developed countries. The UK’s ‘Taylor Review’,47 for example, concludes that wage data in the UK labour market as a whole shows no evidence of ‘hollowing out’.