Thursday, July 9, 2009
Hale "Bonddad" Stewart thinks the worst is behind us, and supports this viewpoint with graphs.
I'm not convinced.
First, he makes this statement:
"(This is another reason I believe the recession is technically nearing the end; the 4-week moving average of initial unemployment claims has been dropping for several months)."
Yeah. It's dropped from about 650,000 to about 630,000. This is not exciting news, and might just be a blip. We'll know more in a few months.
Next, the non-farm payroll employment change improved from about negative 750,000 in January, to about neg. 350,000 in May. Oops - back to neg 450,000 in June. He calls the latter an outlier in the series. This is only true if the 5 month extrapolation is valid. Again - we'll see.
The Challenger Job Cut Report shows a precipitous decline in job cuts, while new jobless claims have only leveled. What gives?
The seasonally adjusted number of mass lay-offs dropped, but recently started moving up again, which Stewart shrugs off.
Retail, food services, auto, and home sales have stopped declining, as has real personal consumption expenditures. None of them have started to increase. Plus, extended unemployment benefits are starting to run out. This can only further stress our consumer-spending-driven economy.
Manufacturing and non-manufacturing indexes have been increasing rather irregularly for several months. However, they have yet to reach the downward sloping trend lines evident since 2006. (Imagine a line laying on top of the bars in Stewart's graph.)
The six month forecast of general activity has come close to its maximum possible reading, while current activity is still near zero. The last time this spread occurred was in '01-'02. Not a happy comparator.
General business conditions are still negative, and the most recent change is down.
But my biggest objection is that Stewart draws his analogies and conclusions based on data series that only go back to 1948. If the proper analogy for these times is The Great Depression, he is ignoring the most relevant historical data.