MADELEINE BRAND, host:
The government has released its monthly employment statistics. They show that the unemployment rate for August dropped from 4.8 to 4.7 percent, which means absolutely nothing, according to Gene Epstein. He's the economics editor at Barron's and a critic of how we media people report government statistics.
Mr. GENE EPSTEIN (Barron's Magazine): Hi. How are you doing?
BRAND: Fine, thank you. So on the face of it, a .1 percentage point decline in joblessness. Looks like good news. So what's wrong with that?
Mr. EPSTEIN: Well, you really surprised me when you said that it means absolutely nothing. I thank you for acknowledging my little bit of wisdom in that regard. I should emphasize, though, that since we want to tell people out there that something is going on, indeed something is going on. We have an unemployment rate that's in the high four percent region. That's basically been fluctuating in terms of number but not in terms of reality - between 4.6 percent and 4.8 percent, and now 4.7 percent - and those are basically the same numbers.
Those movements don't have enough statistical significance to have any meaning. The range of error on the unemployment rate is at least 2/10 of a percent. Therefore we don't know whether any change took place, and that's why the Bureau of Labor and Statistics will report - just as it did this morning - the unemployment rate was essentially unchanged.
BRAND: I see.
Mr. EPSTEIN: However, essentially unchanged is pretty good. We're talking about an unemployment rate that a year ago was almost a full percentage point higher, and that's what's significant.
BRAND: Okay. So percentage-wise, what does it take to make the statistic significant?
Mr. EPSTEIN: Its change on a month-to-month basis has got to be at least 2/10 of a percentage point. But the range of error on any number is much greater than what you would just attribute to pure probability. It's much greater because it's all these - a lot of survey error in this number. When you tell people that this number is based upon a survey of just 60,000 households nationwide, where those 60,000 households are supposedly speaking for a labor force of nearly 150 million people, and when you knock on the door, the person who answers often isn't the individual in that household who either has a job or looking for a job, then you start giving the average person the impression that we must have a clue about what's really going on with the unemployment rate. In fact, of course, these numbers are really just very out of focus, badly out of focus snapshots of what's going on.
BRAND: So Gene, are there statistics that are useful with employment and unemployment news?
Mr. EPSTEIN: The last six snapshots shows that it's been averaging about 4.7 percent. That's almost a percentage point lower than it was about a year ago.
BRAND: But you're saying that the very methodology used to take these surveys is flawed, so how can you trust any of the numbers?
Mr. EPSTEIN: Well, flawed is the point, but not fatally flawed. In other words, flawed to the point where any single one-month number is a very badly out of focused snapshot. That doesn't mean that with a little adroitness, with a little industry, we can't lay these out of focused snapshots end to end and find that we want to know were the subjects in the snapshot is smiling or frowning, or merely crying. Well, sometimes you can tell from out of focus snapshots.
But if we get carried away and if we start putting sizzle into every one-month figure, then we're going to just confuse people much more than we teach them about what's going on with the labor market.
BRAND: Gene Epstein is the author of the book, Econospinning: How to Read Between the Lines When the Media Manipulate the Numbers.
Thank you, Gene Epstein, for joining us.
Mr. EPSTEIN: Sure. Thank you.
BRAND: Stay with us on DAY TO DAY from NPR News.
NPR transcripts are created on a rush deadline by Verb8tm, Inc., an NPR contractor, and produced using a proprietary transcription process developed with NPR. This text may not be in its final form and may be updated or revised in the future. Accuracy and availability may vary. The authoritative record of NPR’s programming is the audio record.