Developing Countries Need Robots Too

Ask any economist why some countries are poor and some countries are rich, and they will probably answer, “productivity”. Essentially, this means that people in rich countries are rich because they are able to create more wealth with less effort. But how do they do this? One of the primary ways is through better technology.

Unfortunately, instead of being recognized for its contribution to wealth, better technology is all too often demonized as a threat to employment, particularly in low-income countries without social safety nets. Intuitively, people care more about the jobs and income streams that already exist than the potential future savings from automating their jobs—a bird in hand, as they say. But a new paper by Mehmet Ugur and Arup Mitra of the University of Greenwich shows that even in very poor countries, technology is far less threatening than it may appear.

We have argued here before that robots are not taking our jobs: in the long run on a macro level productivity increases have no relationship with either the total number of people employed or with the level of unemployment. This is because when automation or other technological improvements decrease costs or increase output, there are many different ways those changes work their way through the economy. While some technologies do eliminate jobs in the short run, others create them or have no effect. Moreover, in the long run the effect is basically neutral because new productivity also creates new demand.

These “second-order” effects are very powerful on a macro level and tend to override short-term first-order effects even when those effects are relatively uncomplicated. While this has been the accepted orthodoxy in economics for some time, it hasn’t been obvious that it applies to lower-income countries in addition to developed countries. The economies of lower-income countries are often substantially different than developed-country economies, with far greater unemployment and informal employment, a much larger agricultural sector, and too little or too much regulation.

However, Ugur and Mitra conduct an extensive meta-analysis of other studies that have examined the impacts of technology on employment in low-income countries. The primary finding is that there is no easy cause-effect relationship from innovation to employment. Instead, the effect of innovation on employment depends a variety of factors in the developing country:

• type of innovation (product versus process innovation)
• skill types (skilled, unskilled and all-skill labor employment)
• level of analysis (effects at firm/farm, sector and macro levels)
• forward/backward linkages
• income levels and distribution
• international trade
• institutional quality

The authors also found that skilled labor employment typically fares better than unskilled labor, and that the manufacturing-sector employment tends to benefit more than employment in agricultural sectors.

While Ugur and Mitra caution against simplistically recommending innovation policy as employment-creation policy, they do find that in the right conditions innovation can positively impact jobs. Once again, the simple story about jobs and technology is shown to be wrong.

()