Robots and artificial intelligence have got their share of flak for killing traditional jobs. But public policy has not come up with many solutions. One policy response to the unemployment threat of automation could lie in taxing robots, suggests a new VoxEU paper.
The study, by João Guerreiro of Northwestern University and others, assumes a world with two kinds of jobs—routine and non-routine. Routine jobs are those that can be replaced by robots, and non-routine jobs need higher skills that are complemented by robots.
As automation gets cheaper, wages in routine jobs will fall and income inequality will rise, say the authors. Taxing robot purchases can then help redistribute income towards these routine workers and keep the excessive wages earned by non-routine workers in check.
Structural changes have destroyed old jobs and created new ones in the past too. But the present challenge is different since the pace of automation has been rapid. Older people may fail to adapt to these changes and lose jobs, leading to a dramatic rise in income inequality.
Since current workers already chose their skills in the past, their choice of work will not be influenced by any policy change. But the redistribution system may discourage younger workers from taking up non-routine jobs. Hence the authors suggest that these robot taxes are desirable only in the short run.
An optimal tax system should balance two objectives: incentivize young generations to acquire skills for non-routine jobs and redistribute income to support routine workers in the face of automation.
Based on a quantitative exercise, the paper suggests that it is optimal to tax robots for three decades: at 7% in the first decade, 3% in the second, and 1% in the third. Once the older generations retire, the optimal robot tax should be reduced to zero.