Bad data

by Russ Roberts on August 9, 2004

in Work

The New York Times has a rather remarkable story by Edmund Andrews on the front page of the Business section of that august paper. The headline:

It’s Not Just the Jobs Lost, but the Pay in the New Ones

Before critiquing the article, I should mention that I don’t particularly credit the Bush tax cut for the recovery. If anything, Bush has raised taxes by increasing spending. He has simply pushed the bill into future years by running a deficit. But let’s read on:

The stunningly slow pace of job creation, which sank to growth of just 32,000 in July, has provided new ammunition in an intense political debate over job quality.

Stunningly slow? Curious. Job growth was rather large in May and June, even after the revisions. I’ll give Andrews the benefit of the doubt and assume he’s referring to the last three years or so. Even then, “stunningly” is rather a loaded work for a news story.

For months, Democrats have said that the long-delayed employment recovery was concentrated in low-wage jobs that paid far less than those that were lost. White House officials replied that the available data failed to settle the matter one way or the other.

The data is still inconclusive.

At this point, we should see, WARNING: What follows will be a matter of opinion and deciding who to interview.

But the weakness in job creation and the apparent weakness in high-paying jobs may be opposite sides of a coin.

Wait a minute. I thought the data were inconclusive. The author has apparently decided that the assessment of the Democrats is right. He hedges with that word “apparent” in front of weakness. But where’s the evidence that it’s right?

Companies still seem cautious, relying on temporary workers and anxious about rising health care costs associated with full-time workers. Many economists say that over the long term, the most vulnerable positions are those at the low end of the wage scale that require fewer skills and are easily replicated.

Even now, at a time when a disproportionate number of new jobs appear to be lower-paying ones, there has been growth in some high-income occupations like accounting, architecture and software.

We are now four paragraphs into the article without a single fact about its central claim.

Yet the earnings gap between the highest-paid employees and the rest of the work force is still widening, as it has over most of the last 30 years. The trend is most striking in factories, which accounted for the bulk of job losses in the last three years and tended to pay above-average wages.

Could be. Not sure how reducing factory or manufacturing employment creates the “most-striking” example of inequality. But it could contribute to it. Does it? Can I please have a fact?

In contrast to previous recoveries, when companies rehired a large proportion of laid-off workers, manufacturers have added only 91,000 jobs this year, having eliminated more than two million jobs in the previous three years.

The largely permanent decline in manufacturing employment, which has been more acute after this recession than in previous ones, spans all levels from blue-collar workers through senior management. It has coincided with a bulge in the number of jobs in low-paying fields that are comparatively easy to enter: retail sales, hotel services and clerical work.

It’s true that manufacturing employment has fallen dramatically in the last few years. I like the phrase “largely permanent.” How can you know? Especially when manufacturing added 10,000 jobs in July, an awkward fact that is not mentioned. But I’m still waiting for a fact about pay rather than employment.

The ragged pattern of the recovery has given rise to the political debate, with Senator John Kerry, the Democratic nominee, saying that new jobs pay, on average, $9,000 a year less than the jobs that were lost.

I’ve heard him say that and I’ve wondered where the number comes from. It’s a very large number. It’s one thing to say that the new jobs are not as good as the old ones. Or that they pay less. I could believe those claims. But $9,000 less? Is it true? Seems very unlikely.

White House officials disagree, saying that such calculations are based on an erroneous comparison of median wages between industries that are expanding and contracting.

Is this really where the Kerry number comes from? I wish the reporter had gone to an unbiased source to critique or understand the Kerry claim. By relying on the White House to refute the claim, the whole thing looks like just another partisan dispute.

The main error, they say, is that even low-wage industries like retailers and fast-food chains hire high-income executives and managers.

If the White House is right that the $9,000 figure comes from looking at medians from expanding and contracting industries then the White House is right. It’s not a meaningful number.

” McDonald’s has C.E.O.’s and accountants, and investment banks hire janitors,” said N. Gregory Mankiw, chairman of the president’s Council of Economic Advisers. “Simply knowing what broad categories are rising and falling doesn’t tell you anything about the jobs people are getting.”

But a growing number of analysts say the evidence increasingly suggests that the current recovery has indeed been tilted toward lower-paying jobs. Industries ranked in the bottom fifth for wages and salaries have added 477,000 jobs since January, while industries in the top fifth for wages had no increase at all, according to an analysis of Labor Department payroll data by Economy.com, an economic research firm.

“Since employment peaked, we’ve lost many more higher-paying jobs than lower-paying jobs,” said Mark Zandi, chief economist at Economy.com. “In recovery, we’ve created more lower-paying jobs than higher-paying jobs.”

Though acknowledging that the payroll data was inconclusive, Mr. Zandi said that the pattern had become firmer over the last month and that it was increasingly similar to what had been found in the Labor Department’s household survey, which categorizes work by occupation as well as industry.

We are now 2/3 of the way to the end of the article, and well into the inside continuation on page C4. We have our first real fact. Since January, 477,000 jobs have been created in the industries in the lowest-paying quintile. None in the top quintile. None? Since January, all of the new jobs are in low-paying industries? That’s hard to believe, so I called Economy.com.

They sent me the data. Their data show that since January, the lowest quintile has added 177,000 jobs and the top quintile, 135,000. Not an equal number to be sure, but not 477,000 vs. zero.

Here’s my guess as to what happened. The reporter read the following lines from the report that Economy.com sent out with the data:

Employment gains during the past year have been measurably stronger in low wage industries than in high wage industries (see table). Of the 1.3 million jobs created since last June, approximately 60% are in low-wage and only one-third in high-wage industries. Moreover, this has not changed appreciably in recent months.

What does the word “this” refer to in the last sentence? What is unchanged? The pattern is unchanged that more jobs are being created in the lower-paying industries and fewer in the higher-paying ones. But I suspect the New York Times reporter misread it to mean that the number of jobs in the high-paying industries is unchanged, ergo, zero job growth in the high-paying industries.

I have a call into Mr. Andrews. I’ll re-post if anything changes.

But here’s what’s amazing and a little bit frightening. This claim that no new jobs are being created in the highest-paying industries will become what Joel Best calls a “mutant statistic.” Whether it’s true or not, because it was in the Times, it will get quoted and cited as fact. I don’t think it is. If I’m wrong, I’ll let you know.

Be Sociable, Share!

Comments

comments

Add a Comment    Share Share    Print    Email

Previous post:

Next post: