We calculate that there will be a job creation need of 305 million jobs between 2020 and 2030, given trends in population growth; changes in the populations age structure, the labour forces gender composition and in age- and gender- specific labour force participation rates; and the desire to reach specific target unemployment rates. The job creation requirements are split unequally among country income groups, such that most of these jobs will be needed in low- to middle-income countries. Our estimates indicate that the job creation needs due to automation by 2030 are substantial, but not insurmountable. According to our preferred specifica- tion of the medium adoption of industrial robots, these will replace 37.9 million workers in 2030 in a high-displacement scenario and 12.2 million workers in a low-displacement scenario. Manufacturing workers seem to be the most vulner- able and Asia is projected to be the region in which robots will substitute the largest number of jobs. Despite burgeoning automation, our calculated job dis- placement due to automation by 2030 is dwarfed by the job creation needs to accommodate projected demographic trends, labour force participation changes and target unemployment rates. A noteworthy relationship exists between the job creation needs due to demo- graphic change and the projected job displacement due to automation: the physically demanding, routine, low-skilled, entry-level jobs required to meet the demand of youth working-age population growth in lower-income countries are the most susceptible to automation and automation-driven reshoring. This pre- sents a challenge for lower-income countries if the growing youth workforce is not well educated and skilled, or performs predominantly routine tasks. From a policy perspective, investing in high-quality education for children who are not yet in the labour force will be key in enabling them to cope with competi- tion not only with their peers in lower-income countries but also with robots in higher-income countries. For higher-income countries with an ageing labour force, the challenge will be to keep older workers healthy and in productive work. Automation could plausibly be beneficial in these countries, as the demand for an older labour force generally consists of less physically demanding and higher-skilled jobs. In these countries, key areas in addressing the challenge will be providing high- quality health care, developing legislation that allows people to work for longer and incentivizes firms to hire older workers, and designing projects to foster collaboration between robots and older workers. For example, robots can be helpful for older workers with hip problems who have to lift heavy objects. Moreover, investments in lifelong learning programmes are also essential to keep this group in the labour force. Lastly, it is important to stress that job creation results from a multifaceted interplay among supply-and-demand factors. Economic policies can play a role in facilitating the process of job creation by providing supportive legislation, by ensuring that workers have the required skills in a fast-changing technological environment, and by promoting healthy ageing in the workforce. Our contri- bution highlights some considerable expected shifts on the labour supply and demand sides that could be helpful to take into consideration when enacting policies that influence labour markets. As far as the potential for further research is concerned, a number of areas are particularly promising and call for a deeper analysis. First, the potential displacement of workers in low-income countries with a young population age structure and the potential boon presented by automation for countries with ageing populations might lead to further divergence in living standards between low- and high-income countries. Thus, investigating policy measures that can help prevent a rise in global inequality will be important. Second, the inconclusive results in the literature on the differential effects of automation on women and men call for a more thorough analysis of the gender-specific impact of automation. This analysis needs to account for the changing nature of auto- mation (away from substituting mainly routine, low-skill-intensive tasks towards substituting more non-routine, high-skill-intensive tasks). Third, analysing the second-order effects of global migration would be interesting given that migra- tion could change labour force participation, for example, when female labour force participation in the countries to which migrants move is higher than in the countries from which migrants originate. Finally, while modelling job creation as a partly endogenous response to characteristics of labour supply is beyond the scope of this article, research in this domain would be useful. This might include research focusing on the nature of job creation in relation to the char- acteristics of employees and the technologies available, specifically models that endogenize factors such as wages and human capital characteristics.

If you were a leader of one of the developing countries and believed the findings of this study, what will you do? Support your answer with references.