Some Thoughts on Latest Initial Jobless Claims

First, we'd like to apologize for this late post as we were having technical difficulties uploading charts, thanks to Google (GOOG).  Of course, the stock is up 1% and is one of our favorite companies.  We'll let this one issue pass.

Now, let's discuss initial jobless claims, which came in significantly below expectations; 339K versus 370K.  In fact, the initial claims number was much lower than the lowest estimate of 362K out there.  The range of estimates was 362K - 375K.  The weekly report from BLS did not indicate anything abnormal in terms of certain states not reporting initial claims, etc.  However, in our opinion, the number itself was an outlier.

 

Initially, we thought that the pre-determined seasonal factor (SF) may have contributed to such a result.  Below is a chart showing the SF used by BLS for the first week of Oct. since 1980.  It is clear that BLS ups the SF every 5 - 6 years to 90 and above, then it slowly brings it back down to the low 80's or high 70's; and then that 'cycle' starts all over again.  Based on the data, the Presidential election appears not to be one of the drivers behind the 5-to-6-year increase in SF. 

 

 Source: BLS

 

We note that the higher the SF, the more favorable it is for the SA (seasonally adjusted) or headline initial claims.  Last week's SF of 96.4 was the second highest SF during the last 32 years.  The highest was 97.3, applied in 2006.  The average SF since 1980 is 86.3.  If that was applied to last week's NSA (non-seasonally adjusted) claims, the SA result would have been nearly 380K!  The median and average SF for the 'spike' years (as shown on the graph above) are 93.7 and 94.3, respectively.  With those SFs, last week's SA initial jobless claims would have come in at nearly 350K, below the consensus, but much higher than the official 339K. 

 

 

We also looked at the one week change in NSA versus SA for the first week of Oct. since 1980.  As displayed in the graph below, NSA and SA have moved pretty much in the same direction most of the time, except in 2001 and the current year.  The graph shows that while last week's NSA figure increased by nearly 26K, its corresponding SA number declined by 30K!  This supports the notion that last week's initial claims result is an anomaly. 

 

 Source: BLS

 

We are not saying that BLS employees responsible for gathering and massaging such data were on vacation.  Nor do we have any conspiracy theories to explain this.  We are simply saying that last week's initial claims number is an outlier.  We will see if we are correct or not during the next few weeks, where it is very likely that the numbers will be revised.  By the way, yes, the previous week's figure (last week of Sept.) was revised higher.  Now, that upward revision certainly wasn't 'abnormal'. 

 

The market's reaction to this news also hasn't been as much as many thought, especially after the recent downward trend that we have seen.  S&P 500 is up 0.4% at 1438.  This morning's news certainly prevented S&P 500 to go lower than its 1430 support level.  It remains below its linear regression level of 1440.  We note the index is still trading below 1447 10-day EMA.  VIX is losing some steam today, down nearly 4%. And the energy and financial sectors are doing well this morning, up 1.3% and 0.9%, respectively. Lastly, congrats to the Yankees and Rauuuuuuul!  Last night's victory was one for the ages ... in our opinion.