« Paul Krugman: A Tale of Two Moralities | Main | Fed Watch: A Mixed Bag of Data »

Friday, January 14, 2011

Capacity Utilization and Unemployment

According to the latest data released by the Fed, both capacity utilization and industrial production increased in December:

Industrial production increased 0.8 percent in December after having risen 0.3 percent in November. ... At 94.9 percent of its 2007 average, total industrial production in December was 5.9 percent above its level of a year earlier. The capacity utilization rate for total industry rose to 76.0 percent, a rate 4.6 percentage points below its average from 1972 to 2009.

I had graphed the relationship between unemployment and capacity utilization in the past, and there was a fairly close relationship between the two series, so I was about to write that this is good news for unemployment. But when I updated the graph with the latest data, the relationship appears to have broken down to some degree:


Why has the relationship changed, i.e. why are GDP and industrial production recovering faster than unemployment, and why has this relationship changed over time? I wrote this in response to that question for a "Room for Debate" at the NY Times, but it got bumped due to the events in Arizona and I don't know if it will run:

The outlook for the economy has improved a bit recently, but we still have a lot of ground to make up – millions of workers who lost jobs during the recession are still in need of employment – and it will take quite of bit of time to close the gap.  Thus, I am still expecting a slow recovery for output and an even slower recovery for employment.

Why will the recovery of employment take even longer than the recovery of output? A combination of factors is at work.  First, firms do not want to make a commitment to hiring new workers until they are sure the recovery is solid, and uncertainty about the strength of the recovery near turning points leads firms to delay in hiring new workers.

Second, during a downturn it's natural to reorganize production.  As firms lay workers off, they reassign tasks to the workers who remain. Then, as things improve they install labor saving equipment in an attempt to cut costs.  This reassignment of tasks and the replacement of labor with software, robots, and other machinery lead to a delay in the recovery of employment.

The third reason for a delay is that firms do not want to let their highest productivity workers, or workers that require costly training, go in a recession even if there's not enough work for them to do.  Since these firms will not hire new workers until this excess capacity is used up, this also delays the time until new workers are hired.

Fourth, when there is a considerable amount of structural change – leading to large numbers of workers who must be retrained and/or relocated as they move out of industries such as housing and finance – labor markets will have difficulty recovering.

I hope I am wrong, but I believe these factors will interact to produce an extended period of unemployment. Historically, financial meltdowns of the type we experienced are difficult to recover from and this creates considerable uncertainty. Thus, the first factor listed above is particularly strong. The second factor is likely to be strong as well since firms will take advantage of expanded opportunities created by technological advances to improve productivity. Furthermore, this recession has induced a large amount of structural change in addition to the usual cyclical problems making the fourth factor much stronger than in a typical recession.

What about the longer run? Will the troubles for labor end when the recession is over? Unfortunately, the answer is no.  Labor markets have experienced tumultuous change in recent decades due to globalization and technological change, and these forces will still be there after the recession ends.

[Alternative explanations are welcome. Also posted at MoneyWatch.]

    Posted by on Friday, January 14, 2011 at 09:36 AM Permalink  Comments (32)


    Feed You can follow this conversation by subscribing to the comment feed for this post.