I like statistics that back up what I already know. The ones that let you say, “See, I told you we were in a recession,” or “Only 12 year-old kids are on MySpace.”

Last week the Bureau of Labor Statistics released figures showing changes in productivity and labor costs during the first quarter of 2009. The numbers indicate that productivity — output per hour worked — goes up a little bit when you lay off a few people, and then drops dramatically when you lay off a lot of people:

  • Business sector – hours worked down 8.8 percent, productivity increased 1.1 percent.
  • Nonfarm business sector – hours worked down 9 percent, productivity increased by 0.8 percent.
  • Nondurable manufacturing sector – hours worked down 13 percent, productivity decreased by 0.1 percent.
  • Durable manufacturing sector – hours worked down 23.4 percent, productivity down 10 percent.

That sounds about right: As smaller-scale layoffs take effect, people are forced to take on extra work, or simply work harder because they are scared they will be singled out for a permanent holiday. Or in other cases, long-standing inefficiencies are inadvertently addressed through the layoffs.

I recently ran into an ex-colleague who had survived a 12 percent reduction in his company’s workforce. “We lost a lot of good people, but we also cut the fat and as a company and we’re better for it,” he commented. I was part of that 12 percent reduction, and I couldn’t help but agree with him — the inefficiencies and overstaffing at that particular company were obvious.

Now before anyone accuses me of writing a piece on the positives of layoffs, let me get back on track.

The part of these statistics that seemed totally counterintuitive was the change in hourly compensation during the same period. In this time of economic turmoil, the Bureau of Labor Statistics was reporting a 15.7 percent increase in hourly compensation during the first quarter of 2009. I took this to mean that people were getting big pay rises, something that is obviously not happening in the real world.

It would be nice to think that companies are compensating employees for the extra work left by retrenched ex-colleagues, but how is that possible when many companies have applied wage freezes, cut hours and reduced benefits?

As it turns out, there is a perfectly good explanation for this erroneous statistic, but it took me a while to find it. After contacting two analysts, two manufacturing industry experts and one economist, I was still no closer to an answer. So I decided to go straight to the source.

Jane at the Bureau of Labor Statistics cheerfully greeted me with a serving of technical mumbo-jumbo about the accuracy of their report — something about a possible lag in the updating of benchmarks and sources of compensation data coming from outside agencies. Jane was going to get a copy of the report and her colleague who was looking into the discrepancies — and she would be right back …

She didn’t come back.

I immediately got suspicious, sensing a cover-up in the Major Sector Productivity group of the U.S. Department of Labor. Was this my chance for a Pulitzer? Was Jane my Deep Throat? Would I have to fly to D.C. and meet her in a shady parking lot to get my answers?

Not really.

I finally managed to speak with one of the fine young economists at the bureau, Shawn Sprague, who dug down into the statistics and gave me the answer I was looking for. The reality was that compensation in the durable manufacturing sector dropped 11.4 percent in the first three months of 2009. This follows from decreases of 3.7 and 5.8 percent in the final two quarters of 2008.

Nondurable manufacturing compensation dropped 5.5 percent for the same period, a slight improvement on the 5.6 percent drop at the end of 2008.

I won’t bore you by trying to explain the statistical definitions that resulted in me misinterpreting the report — but suffice to say that these numbers are meaningless without an understanding of how they are derived and how they apply to a particular industry.

After speaking with Pat McGibbon at the Association for Manufacturing Technology, it turns out that this drop in compensation doesn’t necessarily indicate that workers are getting pay cuts.

“In manufacturing, people don’t do 40 hour weeks — they do 50 or 60 hour weeks. When output drops, the first thing to go is overtime, which is usually paid at time-and-a-half. We’ve seen compensation drops in the 11 to 15 percent range before.”

At the end of the day, “Statistics are like bikinis. What they reveal is suggestive, but what they conceal is vital.”
- Aaron Levenstein

Have you lost overtime hours? E-mail me at