More about Compensation and Productivity

In my previous post I discussed the empirical relationship between compensation growth and productivity growth in the US. One way of looking at the relationship is to look at the changes in labor’s share of total product – that is, gross compensation divided by gross output. This chart shows the evolution of labor’s share, measured 3 different ways (all computed for the business sector and expressed relative to their 1948 levels). The purple line (continued with estimates in yellow) uses NIPA data and divides “compensation of employees” by “gross value added.” The dark blue line (corresponding roughly to Lazear’s results) uses BLS data and divides “compensation” by “current dollar output.” The light blue line (corresponding roughly to Calculated Risk’s results) uses BLS data and divides “real hourly compensation” by “output per hour.”



The NIPA (purple) results are probably misleading because they ignore the decreasing role of proprietors’ income in the national accounts: labor’s share increased during the 50s and 60s because more production was shifted to the corporate sector in which labor gets separate compensation.

The BLS “product wage” (dark blue) results show a relatively steady share, although one might argue that there is a slight downward trend over time. In terms of dollars, workers are getting an output share only slightly smaller than what they got in 1948.

The BLS “consumption wage” (light blue) results show a dramatically declining share for labor. In terms of what they can buy with the money (compared to what firms can buy with their share), workers are getting a much smaller share (down by a factor of .73) than they were in 1948.

This discrepancy points to a basic flaw in the broadly phrased argument that “if productivity is rising rapidly, then standards of living for workers will also rise rapidly.” For the kind of time horizons that we’re dealing with here, the premise has to be about productivity of consumer goods, not productivity in general. Since productivity of consumer goods has in fact risen (though not as rapidly as overall productivity) and is expected to continue rising, Lazear still has a case to make, but he hasn’t quite made the case yet.