Posts Tagged ‘seasonal adjustment’

‘Tis the season …

April 9, 2012

… or more precisely:  ‘tis the seasonality.

For a couple of months, we’ve been pointing out that something smelled fishy about the Fed’s employment reports.

Too much of the good news seemed to be directly tied to statistical tweaks of the the raw data called “seasonal adjustments”.

In fact, the Feds have been goosing the numbers up by more then they used to.

Well, now the Wash Post is even on the case.

The Post article — “Mild winter may have artificially inflated jobs data, economists fear“ —  suggests that we may have been underestimating the effect.

Economists are now saying that the mild winter has artificially inflated job growth.

Translation: The surge in hiring early in the year may not be as strong as it appeared.

The warm weather meant more jobs for construction workers and retail employees.

For economists, it means a statistical nightmare.

Typically, these bumps in demand are evened out through a process called seasonal adjustment.

That allows researchers to compare one month’s economic activity with the next for a more accurate picture of the nation’s health.

But this year’s weather was so abnormal that those models fell short, and economists are now scrambling to figure out how much of the growth over the past three months was simply due to a glitch in their systems.

“When the weather does not follow a normal seasonal pattern, then the seasonal adjustment cannot adjust for it.”

And that may help explain why recent data on jobs have looked rosier than actual economic growth would suggest.

Forecasts for the nation’s gross domestic product during the first quarter hover around 2 percent, a middling number at best.

Somewhere there is a disconnect, and Mother Nature is a valid scapegoat.

The labor market boost from the mild winter will eventually even itself out, though it may mean dips in job growth in coming months

Glad to see the mainstream media catching up with the Homa Files and its loyal readers …

>> Latest Posts

Cookin’ the books? … 2 numbers to watch when the BLS reports tomorrow.

April 5, 2012

Loyal readers know that I’m a bit skeptical re: the employment numbers that the BLS has been spitting out in recent months.

Two reasons: (1) Unemployment rates are diverging from the Gallup daily surveys, and (2) Seasonal adjustment factors are boosting the employment numbers.

First, the Gallup relationship …

Historically, Gallup’s mid-month unemployment rate has tracked closely to the BLS end-of-month rate.

Not so in February … Gallup reported 8% … BLS reported 8.2%.


Gallup’s mid-month rate for March was 8.9%.

Let’s see what the BLS says tomorrow.


* * * * *

The 2nd number to watch is the the seasonal adjustment adder.

For the prior10 years, the BLS has seasonally adjusted February employment numbers upward by 1.1423%.

Last month, they upped the raw numbers by 1.1688%.

That’s a big difference when floated into the unemployment rate calculation.

The prior 10 year adjustment factor for March has been .6209 %.

If the seasonal adder is higher than that tomorrow … be suspicious.


* * * * *

My bet: the unemployment rate will magically hang at 8.3% … .6% below the Gallup number.

>> Latest Posts

The unemployment numbers … digging deeper.

March 12, 2012

The Feds said last Friday that the economy added over 200,000 jobs and the unemployment rate stayed at 8.3%

I’d predicted 8.5% or higher … hmmm.

First, unemployment claims increased in each of the 4 weeks in February


That would make you think that the unemployment rate would go up, right?

Not so, using Fed math … the BLS reported that the  seasonally adjusted unemployment rate stayed at 8.3% …


But, take a peek at the raw unemployment rate … the one before the Feds adjust for seasonality.

Hmmm.  Looks like the rate has ticked up in the past couple of months … and is now around 8.7%


The secret sauce: the seasonal adjustment factors.

I guess a guy doesn’t feel unemployed if he’s unemployed in February  … seasonal unemployment is different.


Let’s look at the main data series that goes into the unemployment rate: the number of employed people.

Again, the Feds report steady improvement on a seasonally adjusted basis.


But,  when the seasonality factors are backed out, actual employment levels have been going down … consistent with the unemployment claims data.


Finally,  for fun, let’s match the seasonally adjusted unemployment rates data (which is reported by the Feds) against the raw numbers (which the Fed calculate but don’t shout out).


Pretty interesting … says we’re in a period when seasonally adjusting helps the unemployment rate appear more favorable … but when we head into Aug, Sept, Oct, Nov … seasonally adjusting makes the unemployment rate look less favorable.

My next prediction: about mid-summer, the Feds will come out with some cock-and-bull story explaining why they’re going to start report unemployment data that isn’t seasonally adjusted.

And, they’ll say with a straight face that the change in reporting methods has nothing to do with the election.

Yeah, right.

>> Latest Posts

Cookin’ the books? … About those pesky seasonal adjustments to the Fed’s employment numbers …

February 9, 2012

Earlier this week, we blogged about the “interesting” difference between Team O’s job gain claim:

The Labor Department reported that the economy gained 243,000 jobs.

But, the BLS  also reported that the economy lost 2,689,000 jobs in the month

The difference in the two numbers is in seasonal adjustment.

Here’s an interesting tidbit that I haven’t seen reported: the January seasonal adjustment factor mysteriously crept up from the factor that was used in January 2011 … with the effect of increasing the number of seasonally adjusted jobs reported.

As Gomer Pyle would say: Surprise, surprise, surprise …





Source: BLS

>> Latest Posts

%d bloggers like this: