Recent advances in automation and their implications for the economy and society are central issues in global policy and academic debates (see Baldwin 2019, Klenert et al. 2020). Despite the comprehensive examination of the impact that automation has on overall employment and labour force participation (see e.g. Grigoli et al. 2020), there has been little empirical research on how automation might affect gender equality.
The impact of automation is likely to be different for men and women
Brussevich et al. (2019) construct a gender-specific routine task intensity (RTI) index, which quantifies the codifiability of tasks performed on the job. The authors find that the RTI index, on average, is 13% higher for female workers across the sample of 30 countries. This gender-RTI gap is driven by female workers performing fewer tasks requiring analytical and interpersonal skills or physical labour, and more tasks that are characterised by lack of job flexibility, little learning on the job, and greater repetitiveness. This suggests that female workers are more exposed to automation risk than male workers, on average. Here, we focus on one specific type of automation: adoption of industrial robots (i.e. robotization), which has specifically gendered effects on labour markets. After all, as opposed to many other forms of automation, robots replace workers with ‘brawn’ skills, who are more likely to be men (Rendall 2017, Ngai and Petrongolo 2017).
Figure 1.A. shows that industrial robots are mainly deployed in the automotive and transport industry (about 390 robots per 10,000 workers in 2014), although they have also begun to be used more widely in the production of plastic, chemicals, and metals, as well as food and beverages. Figure 1.A suggests that the vast majority of industrial robots are employed in industries that are part of the manufacturing sector.
To understand whether there was a change in the gender composition of the workforce over the sample period, we present the share of female workers by industry and year in Figure 1.B. Among the sectors included in our analysis, the most common sectors of employment for women in Europe are education/research/development (women accounted for 68% of all jobs in the sector in 2014), textiles (63%), and food and beverages (47%). Women are also less likely than men to be working in the automotive and transportation, metal, construction, and mining and quarrying industries. Overall, within-industry gender composition changes were minimal between 2006 and 2014.
Figure 1 Robots and women by industry
Panel A Robots per 10,000 workers by industry
Panel B Share of women in each demographic cell, by industry and year
Source: Aksoy, Özcan and Philipp (2020). Note: (M) indicates the manufacturing industry.
Figure 2 shows the gender gap in median monthly earnings in 2010 for the 20 countries included in our sample. The size of the gender pay gap varies across economies: it ranges from 4% in Romania and Bulgaria to 18% in Germany and 19% in Estonia.
Figure 2 Gender gap in median monthly earnings in 2010 by country
Source: Aksoy, Özcan and Philipp (2020). Notes: The gender gap in median monthly earnings is defined as the difference between median male earnings and median female earnings, divided by median male earnings. Earnings of part-time workers are adjusted to their full-time equivalents.
In Aksoy, Özcan and Philipp (2020), we provide the first large-scale evidence on the impact of robot adoption between 2006 and 2014 on the gender pay gap by studying 20 European countries. Specifically, we examine how changes in the number of robots per worker between survey years (henceforth, ‘robotization’) affect the gender gap in the monthly earnings of workers in manufacturing and a few other sectors that employ robots.
We find that robotization increases the gender pay gap: a 10% increase in robotization leads to a 1.8% increase in the (conditional) gender pay gap.1 Given that in many countries and industries, we have seen increases in robotization of more than 10%, this effect is sizable. To put it in perspective, the introduction of the national minimum wage led to a fall in the raw gender pay gap of about 2% (see e.g. Robinson 2002 for evidence from the UK, and Boll et al. 2015 for evidence from Germany). In addition, the effect we identified is larger than that of many family-friendly policies in European countries, where the evidence of their effectiveness for reducing the pay gap is mixed (see review in Olivetti and Petrongolo 2017).
Why does robotization affect the gender pay gap?
We explore potential underlying mechanisms and find that our results are likely to be explained by a larger increase in male earnings than female earnings, especially within medium and high-skilled occupations (through a productivity effect). Put differently, the underrepresentation of women in medium and high-skill occupations in specific industries, accompanied by robotization, exacerbates the gender pay gap, especially in countries where gender inequality was already severe. Conversely, in countries where initial gender inequality was low, robotization did not have a statistically significant effect on the gender pay gap, but increased the earnings of all workers. We also show that our results cannot be explained by changes in the gender composition of the workforce, nor by inflows or outflows from the manufacturing sector, which is in line with the findings of Freeman et al. (2020).
At a time when policymakers are putting increased efforts into tackling gender gaps in the labour market, our evidence is important. Our results suggest that governments not only need to ensure that education and vocational training systems provide people with the right skills demanded in the future, but also need to pay attention to distributional issues and increase efforts to make sure that women and men are equally equipped with the skills most relevant for future employability.
Aksoy, C G, B Özcan and J Philipp (2020), “Robots and the Gender Pay Gap in Europe”, EBRD Working Paper No. 246.
Ngai, L R and B Petrongolo (2017), “Gender gaps and the rise of the service economy”, American Economic Journal: Macroeconomics, 9(4): 1-44.
Olivetti, C and B Petrongolo (2017), “The economic consequences of family policies: lessons from a century of legislation in high-income countries”, Journal of Economic Perspectives, 31(1): 205-30.
Rendall, M (2017), “Brain versus brawn: the realization of women’s comparative advantage”, University of Zurich, Institute for Empirical Research in Economics, Working Paper 491.
Robinson, H (2002), “Wrong side of the track? The impact of the minimum wage on gender pay gaps in Britain”, Oxford Bulletin of Economics and Statistics, 64(5): 417-448.
1 Conditional Gender Pay Gap (GPG) is defined in our paper as the difference between the earnings of men and women who work within the same occupational category, industry, are of similar age, live in the same country, measured in the same year and working in similar size firms. Put differently, conditional pay gap is the pay gap after adjusting for a set of compositional factors that may account for differences between men’s and women’s earnings. Conditional GPG is more important than the unconditional (overall) pay gap, from the policy point of view, because it is related to ‘equal pay’ legislation in Europe.