Suitably Flip

Jobless Claims Much Lower Than Expected, Associated Press Coverage Approximately As Objective As Expected

This morning, the Labor Department released its weekly unemployment report, showing initial jobless claims falling 58,000 (from 404,000 to 346,000).  Math fans will be interested to note this was a one-week decline of more than 14%.

Economists were expecting a decline of less than 3%.  The actual percentage decline was the largest one-week decrease in jobless claims in nearly 3 years (since September 2005).

If this sounds like good news, it is.  Meaningful and unexpectedly good news.  (Unless you're a media outlet that delights in reporting crummy news about the American economy, that is.)

The AP Economics Desk rushed out a story originally titled "B- b- b-, but, but but..."

Jobless claims dip but don't change bigger picture of labor market struggles

Fewer people signed up for unemployment benefits last week, but not enough to obscure continuing weakness in the country's labor market.

The Labor Department reported Thursday that new applications filed for unemployment insurance fell by a seasonally adjusted 58,000 to 346,000 for the week ending July 5. A year ago, the figure was lower, at 304,000, showing a deterioration in employment conditions.

A government analyst cautioned that last week's drop did not suggest a sudden improvement in the country's overall economic health. The decline was exaggerated because of adjustment problems related to temporary shutdowns at auto plants for retooling new assembly lines. The unadjusted, or actual raw figures, showed an increase of 30,000 claims for last week.

Ah, turns out its the not-seasonally-unadjusted (the "actual") numbers that count.  I wonder why the DOL bothers with these seasonal adjustments in the first place.  By the same token, I assume the AP was careful to make the same distinction for its readers back in late January, when the seasonally-adjusted numbers showed a 23% increase in jobless claims, but the unadjusted, actual numbers showed a double-digit reduction (great news, right?).

In another bad sign for the economy, the [seasonally adjusted, non-"actual" according to us] number of workers filing new claims for jobless aid jumped by a much larger-than-expected 69,000 last week to the highest in more than two years, but the numbers were likely skewed by the timing of a public holiday.
...
The Labor Department said that initial claims for state insurance benefit totaled 375,000 in the week ending Jan. 26, the highest reading since October 2005, when claims reached 376,000. It was also the largest weekly increase since September 2005, when claims had mounted by 95,000, the Labor Department said.

Lots of stats about the significance of the seasonally adjusted number - not a peep about the unadjusted number.

And of course that's the way it should be, since the unadjusted number is a lot less meaningful (I think that might even by why they adjust it).  I just wonder why the AP only remembers this selectively.

Comments

Jobless Claims Much Lower Than Expected, Associated Press Coverage Approximately As Objective As Expected

This morning, the Labor Department released its weekly unemployment report, showing initial jobless claims falling 58,000 (from 404,000 to 346,000).  Math fans will be interested to note this was a one-week decline of more than 14%.

Economists were expecting a decline of less than 3%.  The actual percentage decline was the largest one-week decrease in jobless claims in nearly 3 years (since September 2005).

If this sounds like good news, it is.  Meaningful and unexpectedly good news.  (Unless you're a media outlet that delights in reporting crummy news about the American economy, that is.)

The AP Economics Desk rushed out a story originally titled "B- b- b-, but, but but..."

Jobless claims dip but don't change bigger picture of labor market struggles

Fewer people signed up for unemployment benefits last week, but not enough to obscure continuing weakness in the country's labor market.

The Labor Department reported Thursday that new applications filed for unemployment insurance fell by a seasonally adjusted 58,000 to 346,000 for the week ending July 5. A year ago, the figure was lower, at 304,000, showing a deterioration in employment conditions.

A government analyst cautioned that last week's drop did not suggest a sudden improvement in the country's overall economic health. The decline was exaggerated because of adjustment problems related to temporary shutdowns at auto plants for retooling new assembly lines. The unadjusted, or actual raw figures, showed an increase of 30,000 claims for last week.

Ah, turns out its the not-seasonally-unadjusted (the "actual") numbers that count.  I wonder why the DOL bothers with these seasonal adjustments in the first place.  By the same token, I assume the AP was careful to make the same distinction for its readers back in late January, when the seasonally-adjusted numbers showed a 23% increase in jobless claims, but the unadjusted, actual numbers showed a double-digit reduction (great news, right?).

In another bad sign for the economy, the [seasonally adjusted, non-"actual" according to us] number of workers filing new claims for jobless aid jumped by a much larger-than-expected 69,000 last week to the highest in more than two years, but the numbers were likely skewed by the timing of a public holiday.
...
The Labor Department said that initial claims for state insurance benefit totaled 375,000 in the week ending Jan. 26, the highest reading since October 2005, when claims reached 376,000. It was also the largest weekly increase since September 2005, when claims had mounted by 95,000, the Labor Department said.

Lots of stats about the significance of the seasonally adjusted number - not a peep about the unadjusted number.

And of course that's the way it should be, since the unadjusted number is a lot less meaningful (I think that might even by why they adjust it).  I just wonder why the AP only remembers this selectively.