From January through the end of August 2017, Initial Weekly Jobless claims enjoyed a remarkably consistent run, landing in a range between 261,000 and 232,000 each week of this year. That’s a total range of slightly more than 11% for a number reporting weekly that aggregates weekly unemployment insurance filings across a workforce numbering over 170 million.
That consistency ended the instant Hurricane Harvey washed ashore.We learned last week that jobs in South Texas can be added to the list of things the storm ravaged havoc on.
While most of the talk surrounding Harvey’s impact on markets has centered on fuel, which makes sense especially given Houston’s centrality to the US energy landscape, last week’s initial weekly claims release reminded us of the impact a catastrophic weather event can have on other facets of the economy, like jobs.
In the week ending September 2, the advance figure for seasonally adjusted initial claims was 298,000, an increase of 62,000 from the previous week's unrevised level of 236,000. This is the highest level for initial claims since April 18, 2015 when it was 298,000. The 4-week moving average was 250,250, an increase of 13,500 from the previous week's unrevised average of 236,750.
In total, Texas reported an uptick of over 50k new unemployment filings over the previous week. Based on analysis of the national numbers, that 52k uptick from Texas is almost entirely responsible for the sharp increase last week.
The key question this week for economic event traders is, what will the impact of Harvey be in the second reporting week after the storm?
What History Tells Us
Historically, we can use both Katrina in 2005 and Sandy in 2012 as a basis for a forecast of initial weekly claims as the report progresses from the first week after the storm to the second week.
In the first week after the storm, Katrina produced a 29% increase in weekly claims, as the number jumped from 326,000 the week before the storm to 422,00. Sandy produced a 22% increase as the number went from 365k to 446k. Last week, we saw a 26% increase from the week before to the week after.
In the second week, Katrina produced a 30% increase from the week before the storm, and Sandy produced an 11% increase from the week before the storm.
While this is a limited data set to be sure, and storm damage can be challenging to forecast these two storms provide us a rough second week range of between 50-100% of the week one impact.
Applied to this year, that provides us a range of net new claims landing between 274,000 and 298,000.
Outlook for this Week
Modeled mathematically, an econometric model tuned to this week’s data and variables provides an outlook of 289k.
It is reasonable to assume in reviewing the previous storm data that the magnitude of week one is predictive of how much fallout will be seen in week two. The week one data tells us Katrina was more destructive to jobs than Sandy, and consequently the week two impact was far more severe.
The week one damage from Harvey is conveniently between the two. If we likewise place the week two impact roughly in the middle of the range of week two impacts of the previous storms, that provides a number in the 280k range.
We see significant risk for a number to land beyond that range. This is attributable to a potential marginal uptick in claims from the country as a whole as part of a natural drift up from the numbers in the 230k’s we had seen recently, which are near all time lows. Hence, the baseline case for this week is 289k initial claims with elevated risk to the upside for a larger number.