Posts
![[Most Recent Quotes from www.kitco.com]](http://kitconet.com/charts/metals/gold/t24_au_en_usoz_2.gif)
![[Most Recent Quotes from www.kitco.com]](http://kitconet.com/charts/metals/silver/t24_ag_en_usoz_2.gif)
![[Most Recent USD from www.kitco.com]](http://www.weblinks247.com/indexes/idx24_usd_en_2.gif)
Job loss peaks and stock market lows
Wednesday, April 08, 2009
This is a follow-up to yesterday's piece A sucker rally or a new bull market? in which it was suggested that stock market bottoms occur around the time that job losses peak. The table below fills in some of the details behind that statement.
Remarkably, going back to 1945, stock market bottoms, on average, occur almost exactly when job losses peak, within three days as shown above. Further, if you remove the 1960-1962 outlier where stocks didn't reach their lows until two years after the job market did, stock market bottoms lead the most extreme job losses by about three months, on average.
Interestingly, as a percent of the work force, the recent three consecutive months of 0.5% nonfarm payrolls declines don't look all that severe compared to recessions prior to 1990.
4 comments:
About that 1960/1962 outlier:
Here's the result of a simple google search:
"Starting in April 1960, the recession lasted 10 months. The Dow dropped 13.9% from its December 31, 1959 height of 679.36 to its November 1, 1960 low of 585.24."
The 1962 DJIA low was a result of the Cuban Missile Crisis.
That still makes 1960 somewhat of an outlier (job losses peaked only 1 month after the recession started, that's unusual), but substantially changes the result. Taking out the 1960/62 outlier, and you get a number something like -60.
Well you'd expect the two to be highly-correlated but not necessarily on a time scale small enough to be useful. The distribution is so wide that an average value does not 'mean' much.
Thanks ndd - I figured there was something going on there but didn't have time to track it down since we've been traveling this week.
As for the usefulness of this data, with the 1960-1962 anomaly explained, the distribution becomes a lot more consistent:
+1 month: 1
-1 month: 2
-2 months: 3
-4 months: 2
-5 months: 1
-6 months: 1
A nice bell curve centered at around -3 months.
Maybe the median would have been a more appropriate metric?
Post a Comment