Thursday, December 6, 2012

Notes on initial jobless claims (339,000 ex-Sandy!), the unemployment rate, and tomorrow's jobs report

- by New Deal democrat

This morning we got weekly initial claims, and tomorrow we get the monthly jobs report. I just wanted to note of few items of interest. First of all, I wanted to try to find out how much of this morning's initial claims number was still due to Sandy. To do so, I checked the BLS breakdown of initial claims by states, which gives the unadjusted state-by-state initial claims numbers. I deducted NY and NJ, the two states most hit by Sandy, and compared the number as deducted with the unadjusted number minus NY and NJ this week one year ago. Since the seasonal adjustment should be almost identical, that should give me the "real" ex-Sandy initial claims number, assuming NY and NJ would, ex-Sandy, have layoffs at a similar rate to all the other states.

The result surprised me. The unadjusted number this week was 358,541, meaning the seasonal adjustment was 1.03081, to arrive at the seasonally adjusted 370,000. One year ago the unadjusted number was 372,640, which after seasonal adjustment was 383,000.

Leaving aside NY and NJ, last year's unadjusted number was 351,454. Leaving aside NY and NJ, this year's unadjusted number was 311,269. Appling the same multiplier to this year's number as last year's, gives us the nationally seasonally adjusted number backing out Sandy's effects on NY and NJ -- and that number is 339,000 !

Secondly, Gallup's unemployment rate report has been getting play again yesterday and today. This report is new, and its monthly correlation with the Jobs Report's unemployment rate is noisy at best. For example, the official unemployment rate moved from 7.8% in September to 7.9% last month, while Gallup dove from 7.9% to 7.0%. It gives a little better reading when considered year over year. Last year the official rate and Gallup's adjusted rate were both ~8.9% for September through November. This year Gallup's rate over the same three month period averages ~7.9%. If the pattern holds true, tomorrow's official unemployment rate is likelty to also be 7.9% +/- 0.1%.

Finally, recall that the survey for the Jobs Report took place during the week of November 13, in the teeth of the worst effects from Sandy. We know from initial claims that layoffs surged. It's likely that the hiring side of the equation was affected as well. A very poor number should not be unanticipated. My dart-throw would be +50,000, recalling that the range of error is about +/-100,000 on the initial report. So if we get a report as poor as -50,000, don't be surprised. It's still likely to be Sandy.