Sunday, August 9, 2009

hopeful signs for u.s. jobs

That was the headline of a front-page article about July's unemployment rate in yesterday's Wall Street Journal. Not only is this inference not derivable from the evidence, but only serves as another sad indicator that the quality of our leading financial paper has been slipping since News Corp bought it.

The "decline" of the 9.4% unemployment rate in July from the 9.5% rate in June is not statistically significant. The unemployment data is based on a survey. Like any survey there is always a chance that those surveyed may not identically represent the whole population. For example, remember all those public opinion polls before the last election? They often mentioned a margin of error like plus or minus 5 points. The unemployment data has a margin of error but due to the large number surveyed, it is only 1/10 of 1%.

Does that mean unemployment was about the same? Most of the timeit would, but not last month.

Jobs losses in July were 247,000. Just less than a quarter of a million fewer people employed is not good news.

Mathematically, how can the unemployment rate go down, even if it was not a statistically significant change, if there are a lot more people losing jobs?

The rate is merely is a percentage of the unemployed divided by the labor force. Unfortunately both of these, the numerator and the denominator, can change from month to month. To be unemployed one must not have not worked but actively sought employment. No job, not looking? Then not counted in the labor force. In July there were a lot of job loses but there were also 422,000 who dropped out of the labor force. That is, they were out of a job but not actively seeking employment.

The unemployment rate is inherently a less-than-useful gauge of changes in the unemployment situation. Funny things can happen to a ratio when both the numerator and denominator change at the same time. That is what happened in July. This kind of quirk in the unemployment rate happens infrequently, but is not a rare event.

Bottom line: When a quarter of a million jobs are lost in one month this is not good news.

Although journalists may be very good writers, and write with seeming confidence, please do not assume they know what they are writing about.

Be blessed.

NOTE: You can check out the employment/unemployment stats by going here.

No comments: