Are New Jobless Claims Signaling A New Recession?

 |  Includes: DIA, QQQ, SPY
by: Ironman at Political Calculations

Mike Shedlock, a.k.a. Mish, raises the question of whether new jobless claims have begun to follow a "recession pattern", after noting an unexpected rise in the number of initial unemployment insurance claims filed in the week ending Saturday, 22 February 2014. Here's how he described the potential recession pattern after reviewing a chart showing the number of weekly new jobless claims since 1967:

Once claims bottom then start to rise, the above chart shows a recession usually follows. The only exception was 1993, smack in the middle of an internet boom.

It's impossible to know if claims have indeed bottomed, but a secondary pattern shows this is an area in which claims bottomed five out of the last six times. If claims bottomed again now, it would make six out of seven.

This "unexpected" event coincides with numerous other "unexpected" events, nearly all of them weaker than expected.

This is an area where we can apply basic statistical analysis to the seasonally-adjusted data to determine if the trend for new jobless claims has really taken a turn for the worse. Here, our last snapshot of the overall trend in new jobless claims, which began after 23 February 2013 and covered the period of time through the end of 2013, indicated that the number of initial claims being filed each week for unemployment insurance benefits was falling at an average rate of 1,000 per week. Our new snapshot of that trend through 22 February 2014 is below:

Click to enlarge

Since our 9 January 2014 snapshot, we find that the average pace of improvement in the number of new jobless claims filed each week since 23 February 2013 has decelerated to 231 per week. What that indicates is that the overall trend has, at the very least, flattened out since the end of last year.

However, the overall pattern of new jobless claims since the beginning of 2013 indicates that 7 out of 9 data points all fall within a range that is within one standard deviation of the mean trend line, where we would expect the data to fall approximately 68% of the time. The other two data points are on the high side of that range, but not significantly so.

So basically, we observe a slowdown, but as of yet, no real reversal of the established, year-old trend.

Digging deeper into the microtrend of data recorded since the end of 2013, in the context of Western Electric's rules for identifying potential causes for concern in similar statistical equilibrium charts, we find that there is no indication-- as of yet-- that the trend established since 23 February 2013 has begun to break down.

With that being the case, we find that the data for new jobless claims through 22 February 2014 does not-- as of yet-- indicate a recession pattern, which we would observe as a steadily rising number of claims on average from week to week.

One thing we will observe is that contrary to media reports, there appears to be very little indication of any widescale negative impact of severe winter weather in the seasonally-adjusted data on new jobless claims. And certainly nothing like we observed during the recovery from Hurricane Sandy, which had a very pronounced effect in the national data, even though its negative impact was largely concentrated in a small region of the country.

The winter storms of 2014 would therefore appear to have had nowhere near the same kind of negative economic impact, as claims to the contrary are not supported by the seasonally-adjusted new jobless claims data.