Where Did All The Productivity Gains Go?

By now the whole wonky set is familiar with the stylized fact that for thirty years after World War II, productivity gains fed into higher earnings for the median worker. In the past thirty years, that hasn’t been the case. But people rarely dive deep into the question of where the “missing” gains went. A great report that Larry Mishel wrote for the Economic Policy Institute last week does the work and the answer is pretty interesting. Rather than a single cause, there are three separate sources of divergence. One is structural shift in which labor claims a smaller share of national economic output and capital claims a larger share. A second is a shift within the labor share toward greater inequality such that the mean wage rises faster than the median wage. And a third is a very interesting phenomenon that has to do with the existence of different inflation indexes.

When we translate nominal incomes into “real” incomes, we normally deflate with either the Consumer Price Index or the Personal Consumption Expenditure Deflator. These are similar indexes that both attempt to measure the price of goods and services that consumers buy. But the economy as a whole does produce some stuff that isn’t intended for consumption—capital goods and exports. So when we calcuate Real GDP we deflate with a GDP Deflator that includes all those things. A productivity measure is focused on all kinds of output, while a real compensation measure is focused on the stuff people buy. So a divergence in the price series can create a wedge between productivity and compensation. In theory, this kind of thing ought to even out in the long term but what Mishel finds is that it in the post-1973 period it hasn’t evened out.

Here’s a summary table of all three factors:

You should probably think of the growing capital share as largely reflecting the growth in the effective labor pool. Lots of people moved to the United States from Latin America during this period, lots of Asian countries became viable production platforms for tradeable goods, and lots of women entered the labor force. The timing makes the rise of China in the aughts look like the big player here. The growing inequality within the wage distribution is things like the declining clout of labor unions and other formal and informal labor market institutions, plus the rise of finance. The income of rich bankers is labor compensation in the relevant sense for this exercise. The price wedge is more mysterious, but seems to relate to a mix of oil price dynamics and the explosive productivity growth in the production of computers.

A somewhat related point also illustrated by the chart is that median compensation has risen faster than median wages because the amount that employers spend providing health insurance has gone up. To the extent that you feel like we’re getting value for our extra spending on health care services, this picture looks brighter than if you think this has been pure cost inflation with no value.