Sunday, April 06, 2008

Exaggeration by the New York Times on the employment report

Although nobody wants to lose a job, an editorial in the New York Times entitled "Unemployment Rising" exaggerates the current employment situation by labeling the employment report on Friday as "awful" and referred to a loss of 80,000 payroll jobs as "a severe contraction." That is a completely invalid interpretation of the data. A loss of 80,000 payroll jobs is a modest decline, not a "severe" decline. The Times goes on to improperly remark that "The job contraction last month was on top of large losses in January and February." In fact, the losses in January and February were actually modest as well: 76,000 each. A "large" loss for payroll employment would be something on the order of 300,000 or more. We are not even close to "large" losses.

Given that there were 137.926 million payroll jobs in February, a decline of 80,000 jobs amounts to a mere 0.06% decline. Now if we were seeing a 0.3% decline in a single month, then I would start to be comfortable labeling the decline as "awful", but a 0.3% decline would amount to a loss of 413,778, and we are very far from seeing such a decline in a single month. In fact, the total decline over three months is only 232,000.

Oh, and incidentally, that 137.926 million is the number after seasonal adjustment, which is a notoriously subjective and error-prone process. The actual nonfarm payroll employment level before seasonal adjustment was 136.441 million in February and 137,015 million in March, which means that there was a gain of 574,000 payroll jobs in March. Seasonal adjustments are needed to smooth the data, but be alert to situations where they end up masking the real data.

The unemployment rate is not calculated from the payroll data but rather from household surveys. The household survey showed a decline of employment from February to March of 24,000 and a rise in unemployment of 434,000. Given a civilian labor force of 153.784 million, that works out to an unemployment rate of 5.1%, up from 4.8% in March. But those are seasonally adjusted numbers. Before seasonal adjustments, employment was 144.550 million in February and 145.108 million in march, for a net gain of 558,000. Before seasonal adjustments, unemployment was 7.953 million in February and 8.027 million in March, a gain of only 74,000. Before seasonal adjustments, the unemployment rate was essentially unchanged in March compared to February (5.2%.) Before seasonal adjustments, the civilian labor force was 152.503 million in February and 153.135 million in March, a gain of 632,000. In other words, the net loss of jobs was due to the entrance of more people into the labor force. The New York Times didn't disclose any of these facts.

We may in fact be at the edge of a recession, but you will have to look somewhere else other than this employment report to find evidence that we are actually in the middle of a recession rather than merely in a soft patch.

The Times does point out that "Not since the 1950s has job growth contracted for three months straight without recession." I do not know whether that is specifically true,but I would urge two cautions: 1) do not confuse correlation with causation, and 2) be careful when the trend in seasonally adjusted data diverges from the trend in the raw data. Besides, if it happened in the 1950s, why is the Times arguing that it could not happen again? I would also remind readers that in 2002 the Times participated in the "Rush to War" with Iraq, so let us try to step back, calm down, and be more deliberate to avoid a "Rush to Recession" that could do more harm to the economy than good.

Shame on the New York Times for such shoddy financial journalism. I am sure they can do better. OTOH, this is an election year, so the editorial staff is ever-alert for even the slightest excuse to paint the economy in a bad light.

-- Jack Krupansky


Post a Comment

Subscribe to Post Comments [Atom]

Links to this post:

Create a Link

<< Home