Title: Technological Unemployment in a Stock-Adjustment Model

Abstract: Innovation can cause overproduction of consumer durable goods relative to the desired stock. Unemployment can result, until new industries emerge to absorb labor.

Think of the auto industry in the 1920s. Given the price of cars, suppose that the desired stock of cars is one car per household. A car needs to be replaced every three years. If we were in a steady state, then annual production would equal one car for every three households.

However, productivity rapidly improves, reaching an annual rate of one car for every two households with the same labor input. This reduces the price of cars, which raises real incomes and increases the desired stock of cars just a bit. However, it increases purchases of cars by even more, because in the short run people are willing to hold an excess stock of cars, given their high real incomes and the low price of cars.

At some point, though, when people have accumulated too large an excess of cars (there are so many new cars out there that replacement demand is low), they cut back on car purchases, even though they are cheap. Since the car industry already has the capacity to build cars faster than they need to be replaced in a steady state, layoffs take place. Some of the layoffs are temporary (reflecting unusually low demand for replacement vehicles), but most of them are permanent. In fact, assuming that productivity continues to increase, all of the layoffs are permanent.

Full employment will eventually be restored by the expansion of another industry (say, televisions). But for now, the laid off workers have no income, and so they replace their cars at an even slower rate, and they reduce their desired stock of cars.

Beware that this story sounds tighter than it is. I have not done anything about keeping track of the higher real income that comes from the higher productivity. Somebody is richer now than before. Perhaps it is just the factory owners. Or it could also be the fortunate workers who kept their jobs, with higher real wages because of the lower cost of living.

The people with higher income have to do something with it that doesn’t create employment for the laid-off auto workers. Otherwise, we don’t have an interesting story. Keynesians would have the winners save too much. But I do not think that is necessary. The income gainers could spend their bounty on string quartets (nod to William Baumol) or other goods and services that a former auto worker cannot help to produce. The result will be a big increase in the relative price of string quartets.

Note that Ed Leamer makes a convincing case that cyclical movements in GDP are mostly about housing and consumer durables. This makes stock-adjustment models very important, in my opinion.

I also think that this story means that the measured loss of output in a recession overstates the loss in well-being. At the peak of a cycle, the auto industry is producing more cars than people want, and at the bottom of the cycle people are keeping their older cars going a little longer. The value of the flow of services from cars at the peak is not as high as the government statisticians want to impute to it, and the value of the flow of services from cars at the bottom is not as low as the government statisticians want to impute to it.