Thursday, November 17, 2011

Are Nominal Wages Sticky?

Arnold Kling and Bryan Caplan are back to debating aggregate demand and supply. Kling is a skeptic. Caplan supports the "sticky wage" version monetary disequilibrium. Excess demand for money results in reductions in nominal expenditure that should result in lower prices and wages and unchanged output and employment. Prices are flexible and would fall enough to clear markets, but wages only adjust sluggishly. Falling prices and sticky wages imply rising
real wages. Employment and output fall. Easing the monetary disequilibrium would result in higher nominal expenditure, higher prices of final goods, lower real wages, and greater output and employment.

Kling's views are harder to categorize. In my view, they are inconsistent with scarcity. But whether or not his PSST (Patterns of Sustainable Specialization and Trade) makes sense as an explanation for the Great Recession, his most recent argument involved criticism of Caplan's sticky wages approach. ( Sumner also emphasizes sticky wages.) Kling generated the following table:

What Kling sees is that the rate of increase in employment cost slowed more than than the rate of increase in consumer prices. Kling thinks that if employment costs are sticky, then they would have perhaps slowed some (due to growing unemployment) but less than the supposedly flexible consumer prices.

While reasonable enough, what I "see" is that the demand for labor grew faster than the supply of labor resulting in modest growth in real labor compensation (cost to the employers,) during the first period. And then, it would seem, the demand for labor continued to rise, but only slightly more rapidly than the supply of labor during the second period, resulting in even more modest growth in real labor compensation.

But there is some missing information. What happened to employment? My explanation of the above would suggest more rapid growth in employment during the first period and then slower growth in employment in the second period. (The assumption would be that labor supply is growing at a constant rate so that it is the rate of growth of labor demand driving the faster and then slower increases in real compensation.)

Of course, starting in 2007, employment dropped. It is certainly possible that labor supply fell more than demand, and the result was a slight increase in equilibrium real labor compensation. However, a much more plausible explanation is that supply continued to grow and demand fell, and the equilibrium level of real labor compensation fell significantly. But rather than fall with equilibrium real labor compensation, the actual level of real labor compensation grew slightly.

The problem with Kling's reasoning is that during the first period, real labor compensation should have been rising, and in the second period, it should have been falling. It didn't. On the other hand, if product prices are perfectly flexible, then why didn't consumer prices fall? If they are being forced up by growing labor costs, then why didn't they rise by less than labor costs?

Here is a graph of the employment cost index:

It looks like it continued to rise until well after the recession started, and then slowed before rising again at perhaps a slower rate. It looks like it moved to a lower growth path.

What was the growth path of this index during the Great Moderation? What is its current level relative to that growth path? Unfortunately, the series only begins in 2001.

There is a series for wages that exists during all of the Great Moderation--wages of production and nonsupervisory workers. Obviously, this series doesn't include the expense of benefits and it only includes some workers. Still, what has happened to those nominal wages during the Great Moderation?

This measure of nominal wages is 1.39 percent below the trend growth path of the Great Moderation. The price level, as measured by the GDP chain-type price index, is 1.79 percent below its trend growth path. While both prices and wages have risen continuously, this approach and measure suggests that prices are slightly more flexible relative to trend than wages--but not much.

What about compensation? The wage cost index Kling reports only goes back to 2001, but aggregate labor compensation and trend for the Great Moderation are shown below:

Like other nominal measures, the amount firms spend on wages and benefits fell below the trend of the Great Moderation and remains well below trend. However, at 16.47 percent, it has fallen further than nominal GDP, which is "only" 13.87 percent below trend. The other noticeable deviation was in 1999 and 2000, when a positive the gap grew to nearly 5 percent before rapidly disappearing in the 2001 recession.

Employment has fallen during the Great Recession as shown below:

Employment is 9.53 percent below the trend of the Great Moderation. (This is level of employment from the household survey.) There is a very noticeable "boom" in employment right before the 2001 recession. Employment was nearly 3 percent above trend in the second quarter of 2000.

What has happened to compensation per employed person?

Compensation per employed person is currently 7 percent below the trend of the Great Moderation. It did drop significantly during the first quarter of 2009, at a 5% annual rate, but that was fully reversed after two quarters. A better characterization of the period is that compensation per worker quit growing during the recession, but when the economy began to recover in the third quarter of 2009, it began growing again, but remains on a lower growth path. Also note that the "boom" in employment is reflected in a boom in compensation per worker as well right before the 2001 recession.

During this period, there have been substantial changes in the hours worked. Unfortunately, the series going back to 1984 is for production and nonsupervisory workers. Still, the loss in average weekly hours for all private employees was about the same during the Great Recession (About one hour.) Government workers aren't included. (Some government workers were given furlough days. I was.) The following is the average hourly compensation calculated using the average weekly hours of production and nonsupervisory personnel.

When adjusted for the drop in hours worked, the compensation per hour did not drop in the first quarter of 2009. However, its growth certainly slowed with the onset of the Great Recession. It is currently 7.3 percent below the trend of the Great Moderation. It is not that different from the compensation per employed person. (The boom in total compensation, employment, and compensation per worker, also shows up in compensation per hour, though is lagged a bit, hitting 3 percent in the first quarter of 2000.)

These results show no evidence that firms are able (or willing) to significantly cut labor compensation per worker. However, like Kling's simpler analysis, it does suggest that the failure of final goods prices to drop significantly below the trend of the Great Moderation, in the face of a nearly 14 percent decrease in nominal GDP from trend, should not be blamed on labor costs. Nominal labor costs per employee have fallen much further from trend than the GDP chain-type price index.


  1. It’s really awesome readings such well-written articles since they help quite a lot in many ways. As a trader, I always believe in working out with proper way and strategies because only then we will be able to perform at our best. I do it rather smoothly with OctaFX and that’s thanks to their wide range of features and facilities with small spread from 0.1 pips while there is zero balance protection, swap free account and much more, it’s all awesome!