Improving Job Outcomes in Developing Countries: The View from the 2nd IZA/World Bank/NJD Conference on Jobs and Development

This early June, for anybody interested in labour markets in developing countries, one of the most intriguing places to venture to was the World Bank building in Washington. While the World Bank is a yearlong hotspot for development economists, a summer visit became particularly appealing thanks to its hosting of the 2nd IZA/WB/NJD Conference on Jobs and Development on June 6-7, 2019. The event, co-organized by the Institut für die Zukunft der Arbeit (IZA), the Network on Jobs and Development (NJD) and the World Bank, successfully drew nearly a 100 researchers and practitioners from all over the world. Overall, barring an impromptu fire alarm, the conference proceeded in exemplary smoothness, which was in no small measure due to the excellent organization by Dominik Spitza and the conference team.

After some welcoming remarks by Gary Fields, Piotr Lewandowski and Ian Walker, the conference set off with a "policy makers panel" consisting of Prakash Loungani (IMF), Steven Ayres (DFID), Kunal Sen (UNU-WIDER) and Indhira Santos (World Bank). Although the title's format is somewhat of a misnomer - all panelists are academic economists or economic advisors - and the panel might have benefitted from the presence of actual policy makers, the discussions were nonetheless interesting. Four topics framed the discussion: (1) labour market institutions, (2) social protection, (3) female employment, and (4) agriculture. With regard to labour market institutions, Loungani drew attention to the IMF's new study on "Designing Labor Market Institutions in Emerging and Developing Economies" (not enough unemployment insurance, flexicurity and collective bargaining, too high severance pay and minimum wages). Subsequent panelists highlighted instead the politico-economic dimension of labour market institution, which often plays a key role in their effectiveness, and the institutions' impact on the divide between formal and informal labour markets, though it would be worth remembering that informal labour markets have their own institutions. On social protection, very much in line with this year's World Development Report on the changing nature of work, Santos reiterated the Bank's case for universal social protection. This implies no less than its decoupling from employment and the traditional tripartite social dialogue... it is not known what any ILO staff in the room were thinking in that moment. Concerning women in the labour market, Ayres stressed the importance of focusing not only on female employment per se but rather on better (more productive, better paid, more amenable) jobs for women and the heterogeneity of the causes behind women's oft-disadvantaged position in labour markets; however, the panel also stressed that women may have a comparative advantage relative to men in several occupations with rising future demand, including care and education. Finally, Sen emphasized that agriculture in general and smallholder agriculture in particular should be expected to remain the main driver of employment growth in Africa, leaving the issue of how to raise these smallholders' productivity (e.g. value chain integration, technology use, etc.) as key question for researchers.

The panel discussion was followed by four parallel sessions on (i) automation and the future of work, (ii) jobs and territorial development, (iii) gender and (iv) entrepreneurship. Papers included for example an exploration of international differences in routine and nonroutine work by Lewandowski, Park and Yang, an evaluation comparing the impact of grants and loans on entrepreneurship in Egypt by Osman, Crépon and El Komi, and another impact evaluation on the provision of labour market information on Pakistani parents' attitudes towards their daughters' labour force participation by Makino.

Much anticipated, the next item on the agenda was a keynote speech by Daron Acemoglu, in which he presented his most recent work on automation across the globe. In that work, using a task content framework, Acemoglu modelled the impact of automation on the demand for specific tasks and showed that automation's impact on jobs is ex ante ambiguous, especially in its distribution across countries. While automation displaces workers from the performance of old (routine) tasks and thereby makes products more affordable, it also creates new tasks - yet, it is ultimately an empirical question as to which effect dominates. In his words, even if automation might not cast a large shadow in developing countries, it presents a threat to shared prosperity by preventing tasks that could be performed in developing countries from being offshored (Apple's relocation of some of its manufacturing activities from East Asia to the U.S. is one example). As a consequence, there might be a role for policy that fosters labour-augmenting technologies rather than labour-saving technologies.

The keynote was again followed by a set of four parallel sessions, this time on (i) India's employment challenges, (ii) jobs outcomes in Latin America, (iii) job quality and better jobs, and (iv) migration. The migration session was particularly interesting: based on work in post-Soviet Central Asia with Abdulloev and Epstein, Gang raised the question whether migration might lower education investment by raising the opportunity costs of education for young adults, Khamis, Conover and Pearlman analyzed the reversal in Mexican emigration to the U.S. on local labour markets in Mexico, and Kirdar presented his study (with Aksu and Erzan) on the impact of mass migration of Syrians on the labour market in Turkey. The highlight of that day, of course, wasn't any keynote or session but rather the informal get-together of several dozen persons passionate about labour market research in developing country, which was pleasantly hosted at Kellari Tavern, and left onself with a few new acquaintances and many good memories.

The next day witnessed a veritable marathon of presentation sessions, only briefly interrupted by another keynote speech. Imagine it is 08:30, you are jet-lagged, didn't have coffee yet and have to optimize between learning about (i) unemployment and job search, (ii) formality and informality, (iii) gender II, and (iv) social protection systems. Of course, since your preferences are well-behaved - read: display diminishing marginal utility - and transportation costs between sessions are negligible (you have to switch floors...), you will end up attending a bit of every session. This is how I ended up listening to Paolo Falco about an impact evaluation providing job search assistance or a transport stipend to jobseekers in Addis Ababa (work with Abebe, Cariat, Franklin and Quinn), to Mariana Viollaz about the effect of extending internet access on the labour market outcomes of Jordanian women (work with Hernan Winkler), to Sarra Ben Yahmed about the potential of trade openness to exacerbate gender wage discrimination, and to Diana Kruger on the impact of longer school schedules on maternal employment in Chile (work with Matias Berthelon, Catalina Lauer and Carlos Zamora).

That day's first round of sessions was followed by an excellent keynote by Ricardo Hausmann that showcased his work on development, growth and jobs diagnostics. After introducing the concept of growth diagnostics - essentially a systematic search for the bottleneck constraint on an economy's development process - and its translation to the realm of labour markets (jobs diagnostic), Hausmann highlighted that labour market policy may not be the answer to labour market problems. He then presented his own concept of development, or - as he calls it - "the Scrabble theory of economic development". The key idea behind this framework is that the level and change in a country's development are determined by the range and relative complexity of the products that this country can produce, which in turn is determined by the required skills and resources. For instance, a country that only excels at natural resource exports can be thought of as having only a few, relatively common letters, whereas a well-diversified economy with high value added (e.g. Austria) has many different letters, including some common ones but also many rare ones. In this theory, the name of the game of development is to improve the range of "letters" at its disposal, be it by improving the quality of its education or - much neglected - by "importing" foreign professionals who can address skills bottlenecks and from whom local professionals can learn.

After lunch, one had to choose between parallel sessions on (i) growth and labour markets in low-income countries, (ii) education and training, (iii) skills and (iv) fragility, conflict and violence. Honorable mention (by virtue of my attendance...) goes out to Alicia Marguerie and Patrick Premand's work on the impact of a livelihood intervention (a combi package of capital, training and social cohesion) on vulnerable groups in C?te d'Ivoire, the impact evaluation of a vocational training programme in Bangladesh by Abu Shonchoy and his co-authors Fujii and Raihan and a quasi-experimental study of the Colombian apprenticeship programme following some political screw-tigthening. Finally, the final run of parallel sessions presented the exhausted participant with a choice between (i) measuring and modeling jobs' impacts, (ii) firms and labour demand, (iii) income and employment, and (iv) wage policy. Attending the first session, I enjoyed an explanation by José Romero on how jobs-linked externalities change the calculus for the allocation of grants oriented towards job creation (work with Robalino and Walker), an analysis of the behavioural biases affecting Ethiopian jobseekers by Christian Meyer, the dissection of the employment effects of the South African youth wage subsidy by Amina Ebrahim (with Leibbrandt and Ranchhod), and the exposition of automation's threat to Chinese employment by Osea Giuntella (with Wang).

Even though the conference lasted only two days, I trust that I am not alone in returning to my work with a treasure trove of new ideas and insights as well as many new email addresses scribbled into my notebook (somehow researchers don't have business cards...). Here is to hoping that there will already be a sequel in 2020.