Project Syndicate, August 19, 2013
Those following policy debates are well aware of the impasse in the stimulus-austerity debate. One side has the evidence and the other side has the central banks and parliaments, and there apparently is no way to resolve the differences.
But near-term macroeconomic policy is not the only area of economic policy in which confusion is the dominant position. There is also an overwhelming degree of confusion in discussions of the longer term future. According to otherwise serious people we have to fear both labor shortages created by an aging population and mass unemployment due to robots displacing vast amounts of labor.
It should be self-evident that these two concerns are direct opposites. It would be sort of like freezing to death when the temperature is extremely hot. One or the other can be a problem, but not both simultaneously.
The story of the aging of the population being a problem is that we will have too few workers to support a growing population of retirees. In the U.S. context, we currently have close to 2.8 workers for every retiree. This is projected to fall to just 2.0 workers for every retiree by 2030.
This would be a serious problem if we had reason to think that there would not be enough workers to both replace the retirees and also meet the additional needs that will be the result of a larger population of retirees. This is a story of a labor shortage.
A factoid that may allay concerns on this point is that the United States had a ratio of more than 5.0 workers for every retiree in the early 1960s. In spite of drop in this ratio from 5:1 to 2.8 :1 we have still seen sharp rises in living standards for both workers and retirees over this 50 year period. The key was of course productivity growth. Productivity increased by more than 170 percent over this period. The impact of productivity in raising living standards swamped the impact of falling ratio of workers to retirees in lowering living standards.
There is no reason not to expect a similar outcome going forward. Under almost any plausible scenario the increase in living standards allowed by higher productivity will swamp the impact of a falling ratio of workers to retirees. There may be distributional issues, for example if workers have to pay higher social insurance taxes but don’t see real wage gains, this would imply lower living standards. However that is an issue of distribution, it is not a problem created by the aging of the population. If the benefits of productivity growth are widely shared they will swamp the negative effects of a declining ratio of workers to retirees.
While those concerned about an aging population are worried that we will not have enough workers, those concerned about labor displacing robots are worried that we will not have enough work. In this story, we not only have robots doing most of the factory labor now performed by humans, we have robots driving cars, trucks, and buses, handling the check-out counters and inventories at retail stores, and possibly even performing medical procedures. There’s just not much need for human labor in this world.
This vision of the future may sound scary, but we will not go from here to there overnight. The way that robots will displace workers in the future is that same way that new technologies, including robots, have displaced workers in the past. Firms will bring labor saving technology into use that allows them to produce the same output with fewer workers or more output with the same number of workers.
This is exactly what productivity growth means, and it is hardly anything new. In principle this should be a positive development for the economy since it means that workers can enjoy higher wages and/or more leisure time, without firms being forced to raise prices. If robots mean that a firm can have the same output of goods and services with 10 percent few workers then workers can all get a 10 percent pay increase, or alternatively work 10 percent fewer hours at the same total pay, without firms seeing any increase in their costs.
Again, the robots can cause problems if the benefits of higher productivity growth are not evenly shared. If firms are able to avoid raising wages, and thereby get the entire benefit of productivity growth for themselves in a rising profit share, this can be bad news for workers. It can lead to a situation in which fewer workers are employed through time as demand fails to keep pace with productivity growth.
To some extent we have been seeing this picture in most wealthy countries in the years since the 2008 crash. Most workers have seen little benefit from productivity growth, with profit shares hitting post-World War II highs. This is especially clear in the data in the United States.
However, it is incredibly wrongheaded to see this story as somehow a problem of excessive productivity growth. In fact productivity growth has slowed sharply just about everywhere in the recovery compared to the pre-crash years. The problem is that in a prolonged period of high unemployment, the institutional structures in place do not facilitate real wage growth. We either need to boost employment (back to the stimulus debate) or alter the institutional framework to make it more favorable toward workers.
But the bottom line is that it’s just silly to blame the robots. The problem is an institutional structure that systemically redistributes income upward. That should be the focus of our attention.