Fast internet and jobs

Intuitively we assume that landing a cable in a location creates jobs and growth. There’s the inductive reasoning: if it didn’t, no one would pay to lay the cables. But for many individual workers today, the arrival of technology in their workplace is bad news.

In many parts of the world parts of the world technological change, such as the arrival of fast internet, has been (in the jargon used by economists) “skills biased”. This means that it creates work for educated people with the right experience of qualifications, but puts others out of work, by automating their jobs (the poorest jobs remain: for example, it’s still cheaper to employ a cleaner than make a robot to do the same job). Therefore, the arrival of fast internet may “hollow out” the jobs market, with high-skilled people unaffected, and medium-skilled people shunted down into unskilled or low-skilled jobs, as theirs disappear or are outsourced. This observed effect one of the explanations for low average income growth among the middle classes in the 21st century.

In the 21st century, technology has also literally landed in Africa, whose emerging middle class might be vulnerable to the same effect. Jonas Hjort and Jonas Poulsen, two US-based academics, have been analysing the impact of technology in a paper called “The Arrival of Fast internet and Employment in Africa” (NBER Working Paper no. w23582). And (spoiler alert), it’s remarkable news.

As the authors point out, “In 2000, Africa as a whole had less international internet bandwidth than the country of Luxembourg. By 2013, 13% of all Africans used the internet.” This is still far less than the international average, of course, but during this time 10 submarine cables landed on the coast of Africa. This meant that the authors could measure the impact on the economic situation of individuals who were and were not connected to the landing-point cities via terrestrial internet cables in 12 African countries. In effect, they measured the impact on a combined population of 500 million people.

They did this partly by looking at employment surveys in each country, compiled before and after the cables landed, to find out who was working, and what type of jobs they had. The results are startling: in newly “connected” cities, the employment rate rise that could not be accounted for by any other cause was between 4.2% and 10%. With this amount of data, the analysis could rule out confounding factors such as local growth in jobs before the cables landed (that is, ruling out the effect from the likelihood that cables land close to places with the best economic performance already), or a formalisation of Africa’s large grey economy, which changes the data on the number of jobs without increasing employment. This impact on jobs is far greater than almost anything that aid or national governments could have achieved. “Building fast internet infrastructure may have some of the great potential to create employment among any feasible policy options in Africa,” the authors speculate.

Unlike in developed economies, the “hollowing out” effect, pushing semi-skilled workers into unskilled jobs, hasn’t happened. After cables landed and fast internet became available, the analysis shows that fewer workers were employed in unskilled jobs, for example in agriculture. This means the increase in skilled and semi-skilled employment has been even bigger than the headline figure for employment growth. Average wealth and average incomes also increased.

As you might expect, more businesses were created in these areas, they exported more, and workers with secondary and university education did best form technological change. But even the population whose education finished after primary school were also more likely to hold a job, and middle-educated workers, who finished school but hadn’t gone into further education before getting a job, did better too. In South Africa for example, middle-educated workers in the affected areas had the highest jobs growth.

This isn’t what has happened in the developed world, but in the economies that Hjort and Poulsen measured, the impact of technology isn’t only skills-biased, it is skills-creating. The authors argue two trends from the data: companies with better internet access provided more training to their employees, lifting their skills using technology rather than replacing their jobs, and children in connected areas stay in school longer, and get a better education. 

The lazy assumption that information technology is always good for everyone is being challenged by incomes and productivity data in many developed economies. But for 12 countries in Africa, it has improved the working lives of literally millions of people.