Tag Archives: #Automation

Azita Berar on ‘Automation, jobs and inequality’. GLO Policy Brief No. 1.

Azita Berar is Director Policy of the Global Labor Organization (GLO), and Senior Fellow, Graduate Institute of International and Development Studies, Geneva.

GLO Policy Brief No. 1 – Theme 3. Future of Work

Automation, jobs and inequality

by Azita Berar

There is growing apprehension about new waves of technological development in particular about the implications of increasing automation and use of robotics  displacing human labour. To date, research and policy debate have mostly focused on the estimation of potential job losses that such technologies will entail sometimes with alarmist conclusions. We argue for considering a broader frame of analysis for research and policy that takes account of wider economic and societal implications, including the distributional issues and the need for policy to re-engineer new and evolving social institutions.

What we should know

  • Recent empirical studies (see selected references below) have focused mostly on the job destruction and substitution impact in the US, other advanced countries and in a few emerging economies. Through different methodologies applied and assumptions made to project potential job losses, they reach very different conclusions about the number of jobs and/or the types of jobs that will be substituted. Their estimates vary greatly – by no less than 5 times amongst the most alarmist and the lowest band – even when focusing on the same economy and/ or sector.
  • The new job creation impact of automation within the same sector or the broader economy is usually underestimated. There are examples of  automation technology widely used today that clearly show the positive net job creation and economic expansion. The most cited example is the ATM innovation in the banking sector dispensing human tellers. There has been massive redeployment and upgrading of jobs of former tellers to offer different and higher quality banking products and services. Productivity gains have been invested in the expansion of the sector, in turn leading to new jobs and incomes within and outside the sector.
  • Another fast-evolving technological innovation is “co-bots”, where robots “co- work” along with (or assist) human workers. Used in the automotive and the health industry for example, these have contributed to reducing the drudgery and improving the quality of jobs for humans. And potentially leading to as many new jobs for humans as the use of this type of automation spreads.
  • It is also important to distinguish between the availability of technology that could  replace or displace a certain job/task – a starting premise of recent projections on automation induced job losses – and the probability that the technology will be adopted on a large scale replacing all such jobs/ tasks. There are gaps, sometimes very significant, between the two. This gap is larger in developing countries and amongst them, where as noted, there has been relatively little research. And where the diffusion and impact of automation can not be deducted from current estimates focusing on a limited set of countries and sectors.
  • An interplay of multiple factors pushing in different directions determines the pace and scale of adoption of automation innovation that goes beyond a simple comparison of relative cost among robots and human labour.
  • Understanding the decision-making process of managers and investors in the spread and large-scale adoption of automation is also key. Rarely, automation alone, in the absence of a broader firm strategy,  provides the competitive edge for increased profitability and productivity.
  • There are also wide sectoral variations in the nature of automation innovation and the dynamics of job, income and productivity gains and losses that could incur. A sector/industry focus would shed more light into specific dynamics.  Another under-researched area is the dissemination of technology within sectors and across borders including through global supply chains in highly globalized industries.
  • From a societal perspective, the key issue is how the productivity gains will be invested and/or distributed and through which mechanisms and institutions. This is highly important not from the angle of equity but for sustaining demand in the medium to long term.
  • Throughout previous waves of industrial revolutions and technological disruptions, starting with the mechanization of spinning and weaving in Britain’s textile industry the mid-18th century, there has been apprehension with regard to destruction of jobs and distribution of wealth and well-being. Then and now, there has been a sharp divide between optimists and pessimists! Looking back the reality has overtaken all projected scenarios in very different directions but it has also led to social engineering, policy innovations and the creation of new institutions to tackle wider societal implications.

Will it be different this time? Context and policy matter.

Automation in unequal and polarized environment

  • This new round of rapid automation innovations is taking place amidst a highly polarized environment.
  • Labour markets remain tight: globally 192 million people are unemployed of which over 70 million are young women and men. In many instances, the youth unemployment rate is three times higher than adult unemployment and female unemployment twice that of male unemployment. And these figures do not take into account the “discouraged workers”, those who have given up looking for a job. Quality of jobs is a concern. About 1.4 billion people are working in “vulnerable forms” of employment (own-account workers and contributing family workers) and this number is on the rise. Informality is widespread particularly in developing and emerging countries, where it reaches more than 60 per cent of total employment. There are large segments of population affected by “working poverty”, those who work but are trapped in poverty.
  • Rising inequality and polarization: in the last two decades, most regions have experienced an increase in income inequality. During the same period, there has been a decline in the share of income that goes to labour relative to capital. And this pattern has persisted while labour productivity has increased. Concerns with hollowing out of the middle class and income polarization along different groups of population galvanizes headlines regularly and feed into social unrest. These trends left unattended, automation is likely to exacerbate the inequalities and vulnerabilities and entail more apprehension, whether founded or not on evidence, and more resistance to change.

A broader policy and research agenda

  • For a more positive outlook into the future of automation and work in this environment, there is a need for a broader value-driven policy and research agenda that  engages multiple actors, spans to other geographies and provides a more integrated perspective. Beyond net job destruction and creation counts, induced by automation, distribution of jobs and skills, productivity and income gains should be the focus of research and policy.
  • Space for dialogues. There are examples of how more acceptable and equitable outcomes can be negotiated through social dialogue. The highly automated Hamburg Port is a case in point. Advanced automation has been accompanied by internal redeployment of employees to other jobs and tasks, their re-skilling and up-skilling packages, and redistribution of productivity gains through new working time arrangements and wage compensations.
  • Another innovative form of dialogue to explore, is dialogue among scientists and technology innovators with the research community in economic and social sciences.
  • Transition policies should become permanent features of public policy action and enterprise strategies, facilitating redeployment within firm, sector or economy. Use of  fiscal policies as well as private sector financing  can support active transitions in particular for those made vulnerable to automation, and who are in structural disadvantage in the labour markets.
  • Expansion and distribution of learning and re-skilling opportunities as automation and other technological innovations require different and continuously evolving  set of skills. Their accessibility to those who need them most can be powerful means to reduce inequalities.

The challenges of disruption in labour markets as a result of new waves of automation are real but more complex than a narrow focus on the potential job displacement and job creation impact. While this time too, the future may evolve in unpredictable ways, there is a need for a broader policy and research agenda and actions that accompany these transformations. In the polarized environment of today and the rapid pace of technological innovation, the time is now and the time span is limited to invest in new social engineering  and cooperative mechanisms that create more equitable and sustainable outcomes.


Arntz, Melanie; Gregory; Terry; Zierahn, Ulrich. 2016. “The risk of automation for jobs in OECD Countries: A comparative analysis”, OECD Social, Employment and Migration Working Paper No. 189. Paris, OECD Publishing.

Frey, Carl Benedikt; Osborne, Michael. 2013. “The future of employment: How susceptible are jobs to computerisation?”, Oxford, Oxford Martin Programme on Technology and Employment, a revised version of which was published in 2017 in Technological Forecasting and Social Change, Vol. 114.

ILO, Employment Policy Brief 2017. New automation technologies and job creation and destruction dynamics.

McKinsey Global Institute. 2017. A future that works: automation, employment, and productivity.

World Economic Forum. 2018. The Future of Jobs Report 2018.

World Bank. 2016. World development report 2016: Digital dividend. Washington DC, World Bank.

NOTE: The opinions expressed here are those of the author and not of the GLO, which has no institutional position.