Katrina's Effect on Jobless Claims vs. Sandy

Before Sandy hit I mentioned that jobless claims would be one of the more sensitive economic indicators to any disruption caused by the storm.  Checking back in, today we found out that jobless claims spiked 78,000 following the storm to 439,000.  In 2005, claims rose by 96,000 after Katrina hit and it took six weeks for claims to fall back to their previous level.

Will Sandy Affect Economic Data?

With Sandy shutting down much of the east coast today, it's likely that there will be some effect on economic data, but it's tough to say how much until the storm actually hits.  Typically initial jobless claims can be one of the more sensitive indicators to week to week variance.  In 2005 Katrina caused a clear spike in the national numbers, although I don't think anyone expects Sandy to cause anything close to the damage that Katrina did.

4 Wk Moving Average

Jobless Claims, the VIX and High Yield Spreads

Initial jobless claims were reported today and fell back towards their lowest levels for this cycle at 359k.  The chart looks similar to two other important financial charts: the VIX and High Yield credit spreads.  All three charts show a similar spike in the recession and have gradually declined since (with some bumps along the way).  Now all three are settling at low levels, but slightly higher than their lowest points reached in 2006/2007.  If the S&P 500 is going to extend its rally from here, it would seem dependent on each of these indicators breaking through to lower levels.  Can we get there?

Nonfarm Payrolls Relationship to Jobless Claims

On Friday, we'll get the monthly employment report, which is expected to show an increase of 130k non-farm payrolls.  Even though the broad employment report comes out once per month, each Thursday we get a glimpse of what the employment situation looks like from initial jobless claims.  Initial claims reports can often move the market, but how good is the initial claims data at predicting payrolls?

Below is a regression of the 4-week trailing average of initial claims against the monthly payrolls data.  The r-squared of the simple linear regression is .54--not a perfect correlation, but relatively meaningful.

Recently, the 4-week average of initial claims has risen somewhat, back to 370k.  From the regression 370k initial claims would imply somewhere around a 55k increase in non-farm payrolls.  This would be well short of estimates.