The Unemployment Rate and Compensation Growth

Last week I took a look at the way that higher labor productivity has not translated into higher worker compensation, particularly during the 1980s and 2000s. This is at odds with classical labor market theory, which suggests that as workers become more productive, their increasing value to firms should cause their wages to be bid higher so that their compensation rises accordingly.

There are a number of possible explanations for the divergence between productivity and compensation, and for how this may play into the broader phenomenon of stagnant wages for average workers. Part of the explanation is that an increasing share of worker compensation takes the form of benefits rather than wages and salaries. As shown in the chart below, fully one-fourth of worker compensation in 2010 took the form of benefits. (Source: BEA personal income data.)


This upward trend has been driven almost entirely by the rise of health care costs in the US, and the corresponding rise in health insurance premiums. Note that the one dip in the series in the late 1990s was due to the widespread implementation of HMOs - but they clearly proved to provide a one-time gain rather than a permanent increase in health insurance efficiency. So part of the reason that workers' paychecks have not been rising is directly attributable to the rise in health care costs in the US.

But that's not the whole story, and doesn't address the question of slowly growing total compensation (as opposed to stagnant wages). There are, I think, reasonable arguments to be made about social and political factors, such as the decline in the power of unions. Along similar lines, Mike Konczal recently wondered to what degree this could be due to the Fed's consistent and explicit desire to prevent wage increases.

And then there's plain old supply and demand as a possible explanation. What did the 1980s and 2000s have in common from a macroeconomic point of view? One answer is this: multi-year long periods of high or rising unemployment rates.

The chart below shows, in blue, the seven-year moving average of the portion of increased labor productivity that were paid to workers in the form of higher compensation. During the 1960s and 70s, for example, workers typically received around 80% of gains in labor productivity over any given seven year period. Then during the 1980s that portion fell to about 40%. Meanwhile, the series in red is the seven-year moving average of the unemployment rate.


To make it a little easier to interpret, I've color coded the 60 years shown in the chart by shading the periods when workers were losing their share of productivity growth red, while the periods when workers were increasing their share of productivity gains are shaded in green. This helps to make it quite clear that "green" times - i.e. times when workers seem to be enjoying more of the gains in productivity - were periods when unemployment was falling. "Red" times (I guess it actually looks more pink than red in this chart) are clearly associated with periods when the unemployment rate was stagnant or rising.

One implication of this is clear: the high unemployment rate in the US right now, which is expected to decline only slowly over the next several years, is likely to mean that it will be a long time before worker compensation begins to rise as rapidly as worker productivity. Put another way, the overall level of high unemployment right now not only has the obviously enormous personal implications for those who are unemployed -- it also is likely to seriously affect the compensation of workers who have never lost their jobs, for years and years to come.