Initial claims last week ex-California computer glitch ~327,000

- by New Deal democrat

Computer issues in California continue to bedevil the weekly initial jobless claims reports. We can make a good estimate of what the "real" initial jobless claims have been, however, by excluding California, comparing the unadjusted average for the other 49 states this year vs. last year in the same week, and projecting this year's "real" number by assuming that the percentage of claims in the other 49 states are the same percentage of the total this year as they were last year. I used this technique last autumn in the wake of Superstorm Sandy to show that "real" initial jobless claims were actually declining slightly -- a trend that proved correct once the distortions disappeared from the data in a few weeks.

Keep in mind that the week I am adjusting in this post is last week's report, since there is a one week delay in the Department of Labor's posting of state by state numbers. So the "real" number I have estimated is comparable to last week's 310,000, not this week's 305,000.

Last year this same week total unadjusted jobless claims were 330,454 vs. 272,918 this year. Of last year's total, 61,421 were California claims vs. 40,657 this year. This gives us a 49 state comparison of 269,033 last year vs. 232,261 this year. That makes this year's 49 state number 86.3% of last year's 49 state number.

Last year, after seasonal adjustment, the report for September 15 was 379,000. Multiplying that by 0.863 gives us just over 327,000.

Last week I estimated that the "real" initial claims number for the week of September 7 was ~318,000 vs. 294,000 as reported. This means that the "real" 4 week moving average of initial claims this week is 325,250 vs. the reported 314,500.

This is still the best 4 week average in 6 years, since October 13, 2007, and it indicates that, even after we adjust for the computer glitches, the trend in initial jobless claims is still improving.