Please explain these jobless claims numbers.

[Just thinking]

OK, the numbers came out yesterday for jobless claims for the week before Thanksgiving --- 407,000 (seasonally adjusted). A huge drop from the previous week, 34,000 less. But wait --- the non-adjusted data shows that there were in fact 52,500 more people filing this week than the previous week. Now, I understand the need for seasonal adjustment and all, but if we look at how both sets of numbers went the previous week we see that 409,500 actual claims got adjusted to 441,000, yet this week 462,000 gets adjusted down to 407,000? Just how much of a seasonal change can occur in one week? I don't recall anything major occurring in the previous two weeks, or even minor for that matter. Now, in all fairness, the same seems to have occurred 1 year ago, but it only shows for the same week as for our previous week --- I'd like to examine both previous weeks a year ago to see if the same reversal of adjustment occurred then.

 

[Puppycow]

For whatever reason, probably because, for example, the last harvests of the season are typically over by such-and-such week, certain seasonal kinds of jobs probably end the week before and then lots of people apply for jobless benefits in the same week every year. Some factor like this may explain it. But I don't know the specifics.

I assume that the seasonal adjustments are based on many decades of data and that certain patterns tend to occur at specific times of year on a regular basis. That's why one week the seasonal adjustment is up, and the next week it is down.

The adjustment itself should be roughly the same each year, and they should all add up to zero by the end of the year. All the positive adjustments have to be offset by equal negative adjustments (cumulatively of course) in other weeks and vice versa. Another way of saying this is that the total of seasonally adjusted data for the full year should be equal to the total of non-seasonally adjusted data for the full year.

 

[drkitten]

OK, the numbers came out yesterday for jobless claims for the week before Thanksgiving --- 407,000 (seasonally adjusted). A huge drop from the previous week, 34,000 less. But wait --- the non-adjusted data shows that there were in fact 52,500 more people filing this week than the previous week. Now, I understand the need for seasonal adjustment and all, but if we look at how both sets of numbers went the previous week we see that 409,500 actual claims got adjusted to 441,000, yet this week 462,000 gets adjusted down to 407,000? Just how much of a seasonal change can occur in one week? I don't recall anything major occurring in the previous two weeks, or even minor for that matter.

Holiday season's kicking off. A lot of retailers hire in the weeks before T'giving so that they can train staff and have a full slate for black Friday, which means a lot of hiring doesn't happen just before and just after training. Unemployment gets kind of wiggy in late November through early January.

 

[Just thinking]

Thanks pup and kitty ... I did manage to get a look at previous years, and yes, there seems to be somewhat irregular claims numbers around this time (up, then down, then up) in only a handful of weeks. But this is somewhat recent, as in many years past (+20), this sort of thing didn't seem to happen. Weird.