Robots are going to take our jobs, right? Well, maybe. The available evidence provides conflicting results:
- In the US, one new robot reduces employment by 5.6 workers and local wages by about 0.5% according to estimates in Daron Acemoglu & Pascual Restrepo’s much cited 2017 paper.
- In Germany, more robots decrease employment in the manufacturing sector but increase employment in the services sector, suggesting that the spread of robots creates both losers and winners in local labour markets, according to Wolfgang Dauth and his colleagues in a 2018 paper.
So, what’s been happening in Japan, the most robot-dense country in the world?
Hot off the press is a paper from Adachi, Kawaguchi and Saito that finds that the employment-to-population ratio (that is workers per number of local residents) decreased in labour market areas highly exposed to robots in Japan, similar to Acemoglu and Restrepo’s findings for the US. But the decrease in Japan was not caused by a reduction in employment. Overall employment increased, mostly in services-sector jobs (like in Germany) but there was also an increase in the population in the local labour-market area.
The employment-to-population ratio decreased not because robots led to a net decrease in jobs, but because the local population increased by a larger amount. This would resolve the apparently inconsistent US and German findings.
A plausible mechanism
Adachi, Kawaguichi and Saito suggest that robots increase productivity, which flows into higher wages and profits – generating larger local purchasing power. The increased disposable incomes create demand for local firms to produce locally consumed services. There may be a further local multiplier effect where the increased economic activity and extra tax revenue collected funds local amenities, such as public schools and hospitals (see Moretti, 2010). These amenities make a place more attractive to the non-working population. Population growth may be a downstream (third, fourth or fifth order) effect of the robots coming to town.
So maybe the finding that more robots in local Japanese labour markets decrease the employment-to-population ratio isn’t that surprising (or negative) after all. It looks like robots can help fund nicer and more attractive places to live.
Don’t generalise from robots to other automation technologies
These findings are about industrial robots, which are easy to measure. They go into well-defined workplaces, usually in manufacturing. The results may not generalise to other automation technologies. This point is made by Acemoglu and Restrepo who also looked at the impact of IT more generally. They found that IT investment, unlike robot investment in the US, did not reduce local employment.
We can’t generalise about who benefits from local effects
We should be cautious about assuming that local tax and public spending effects found in other places have similar outcomes here. Local government in New Zealand has a smaller scope of responsibilities than local governments in most other countries. This is in part because many of the functions undertaken at the local level in other countries, such as hospitals and schools, are funded centrally in New Zealand and provided through Crown Entities.
Local policies can undermine this positive story
Chang-Tai Hsieh and Enrico Moretti recently reported that local government authorities can dampen the effect of high productivity in their cities by restricting growth in housing supply and making housing unaffordable. Such policies restrict worker access to more abundant job opportunities, including high-productivity work. High relative house prices can also lead to the non-working population leaving a high-productivity city. More on this in a later post!
My main take-out: Robots coming to town have more complex effects on employment and community wellbeing than we might at first expect.
Judy Kavagnah is an inquiry director with the Productivity Commission.
- Fukuoka is the economic centre of the Kyushu region with a high rate of firm start-ups, many service sector firms, and a high-tech manufacturing sector.
- Sally Davenport attended the ISPIM innovation conference in Fukuoka, Japan in December 2018.
Daisuke Adachi, Daiji Kawaguchi & Yukiko U. Saito (2019). Robot, Employment, and Population: Evidence from Articulated Robot in Japan’s Local Labor Markets, July 16, 2019. http://www.esri.go.jp/en/data/esri2019_presenter_paper1.pdf
Daron Acemoglu & Pascual Restrepo (2017). Robots and Jobs: Evidence from US Labor Markets. March. NBER Working Paper No. 23285.
Wolfgang Dauth et al. (2017). ‘German Robots – The Impact of Industrial Robots on Workers’. London, Centre for Economic Policy Research. https://cepr.org/active/publications/discussion_papers/dp.php?dpno=12306
Enrico Moretti (2010). “Local Multipliers,” American Economic Review, 100 (2), 373–377.
Chang-Tai Hsieh & Enrico Moretti. (2019). “Housing Constraints and Spatial Misallocation,” American Economic Journal: Macroeconomics, 11 (2): 1-39. DOI: 10.1257/mac.20170388