Explanations for The Jobless Recovery

Two of the world’s leading economists publicly discussed possible reasons for the jobless recovery in the past week. In a speech on November 6, Ben Bernanke, Governor of the Federal Reserve Board, articulated the argument that the main culprit is high productivity growth in the US. On the other hand, Alan Krueger, one of the world’s leading labor economists and a professor at Princeton University, threw cold water on this explanation in a NYTimes column on November 13. Who’s right? They both make great arguments (you should read them in full), but if you keep reading, I’ll suggest a new possibility that neither of them addressed.

Let’s consider a few alternative explanations for the jobless recovery:

1. Maybe global political and economic uncertainty is keeping firms from hiring. The problem with this explanation? As Bernanke said:

The observations that workweeks have not risen and temporary employment have risen only modestly… is a bit of evidence against the uncertainty explanation. If firms needed more labor services but were reluctant to commit to new hiring, one would expect to see them lengthening workweeks and hiring temporary workers in large numbers, measures that increase labor input but are also easier to reverse.

2. Maybe high productivity growth in the US is allowing output to rise without increasing employment. This is Bernanke’s preferred explanation:

In real terms manufacturing production in the United States has risen rapidly over the past fifty years. The recent recession has affected that trend only modestly. For example, although as of September 2003 U.S. manufacturing output was about 6 percent below its mid-2000 peak, it was also about equal to the level reached in 1999.

If manufacturing output has not declined in the United States, then what explains the sharp reductions in U.S. manufacturing employment? The answer is a stellar record of productivity growth. Over the years, new technologies, processes, and products have permitted manufacturing firms to produce ever-increasing output with ever fewer workers.

This observation brings me to my fifth and final possible explanation of the jobless recovery, which is the remarkable increase in labor productivity we have seen in recent years, not only in manufacturing but in the economy as a whole… Although other explanations for the jobless recovery – overstaffing in the boom, benefits costs, uncertainty, and structural change–have played a role, in my view the productivity explanation is, quantitatively, probably the most important.

However, Krueger disagrees with this explanation:

On examination, rapid productivity growth is unlikely to account for the dismal job picture in the United States over the last two and a half years.

Implicitly, this view argues that for some reason there are limits to how fast the gross domestic product (that is, output of goods and services) can grow, so, by definition, faster labor productivity growth results in slower job growth… There is no reason the gross domestic product could not have grown faster once productivity accelerated. Monetary and fiscal policy have not restrained growth.

Second, in the United States greater job growth tends to accompany faster productivity growth, over either a quarter or a year, as well as over longer periods. [It is normal for] productivity growth to surge at the beginning of a recovery – [only] the job losses are unusual this time. Furthermore, during the second-longest jobless recovery on record, which occurred during the previous Bush administration, productivity growth was lower than it is now, so accelerating productivity is not the only potential cause of a jobless recovery.

I have to admit that I find Krueger’s counterargument persuasive. I agree that there has to be more to the story than just the high rate of productivity growth. Such as? Citing Larry Katz, Krueger proposes numbers 3 and 4 on our list:

3. “Maybe recoveries that follow longer booms have weaker job growth initially because companies postponed restructuring during the boom. Therefore, more time is needed for companies to reorganize work, which spills over into the recovery phase.” Krueger goes on to cast doubt on this explanation, as well, however:

Indeed, the recessions in 1991 and 2001, notable for extended jobless recoveries afterward, both followed long booms. But in the eight earlier postwar recessions, longer booms were typically followed by shorter jobless recoveries, not longer ones. For example, after the recession following the 1960’s boom – the second longest, after the Clinton boom – job growth resumed immediately.

Again, I agree with Krueger. I don’t think it has anything to do with the length of the boom that preceded the recession. Krueger’s next proposed explanation is:

4. Maybe jobs haven’t been created because the tax cuts were directed more at savings and consumption, rather than job creation. The problem with this idea? Even if they weren’t explicitly directed at job creation, the tax cuts still stimulated the economy, and caused GDP to grow rapidly. So we’re still left wondering why a growing GDP isn’t accompanied by growing employment. Whether you subsidize the firm or the consumer through a tax cut shouldn’t matter – either way more economic activity is generated. Also, counterexamples abound. For example, the tax cuts of the early 1980s were also primarily in the form of reduced income taxes, yet led to a massive boom in employment.

So what’s my own theory? I think that there’s a crucial clue contained in #3, which I will use to propose my own explanation…

5. Maybe recoveries from milder recessions take longer to generate jobs. The two mildest recessions in postwar US history have been the two Bush recessions. And those are the two that took the longest to start generating new jobs. Why? I think it’s a variant on what Krueger said in #3: in a mild recession, there’s relatively little restructuring, and relatively little firing, so the economy doesn’t have as deep a hole to climb out of. Sharper recessions necessitate drastic actions by firms to stay alive – and those that survive are therefore strong and healthy from the very beginning of the recovery. But a mild recession doesn’t do as much to weed out the weaker firms, so you have more firms at the beginning of the recovery that still can’t afford to hire more workers.

Note that I’m not at all advocating sharp recessions over mild ones – mild recessions are still definitely preferable, in my book. But we may have to recognize that if the pain isn’t as severe, the bliss of job creation won’t be as great, either.

Kash