Brad Delong writes
Give us another four weeks and we will know where the labor market is…
The clear concern is that enormity of the seasonal adjustment in the new Claims data makes the standard adjustment tricky to interpret. This is only excererbated by the fact that the Global Financial Crisis hit just as winter new claims were peaking, making it hard for the computer to parse apart how much of that huge spike in the middle was the GFC and how much was the fact that winter is just a worse time than it would otherwise estimate.
There are a few tricks to get around this just with your eyes though. Obviously you can go peak-to-peak and trough-to-trough. Though if you think the dynamics really have changed or if the situation is changing rapidly between peaks and troughs that can be hard to eyeball.
You can also look at the following.
This is the year-over-year change in unadjusted new claims plotted as a bar graph.
The trick is that when the area above the line is balanced by area below the line you are back to where you were before the GFC.
This is like taking the derivative of Initial Claims as function of time and then reintegrating to give back the original function. However, because its year-over-year it destroys any seasonal information, which is good for us because most of that is orthogonal to what we want to know and only serves to confuse our eyes.