Does Size Matter? from James Sogi
Nelson Freeburg, editor of Formula Research, did some studies on asset allocation and sector models with S&P and Russell moving average crossovers that looked promising. He does some limited testing of the ideas, but unfortunately makes the error of curve-fitting to make the maximum return going forward. We know that doesn't work. He fails to test the results statistically. For these reasons I do not recommend the newsletter. He does consider maximum drawdowns and time to recovery. Too bad I didn't save my issues, as there are some thoughtful ideas to test, submitted by various money managers, including some of the more illustrious Daily Spec contributors.
Kim Zussman investigates:
The quarter ending 9/30/07 SP500 return was about +1.5% and RUT (Russel 2000 small cap) was about -3%. Usually they dance together, but this time Mrs. Small and Mr. Big pirouetted across the floor, away from each other.
Looking at index quarterly returns 3/88-9/07, what happens in the next quarter if SP500 up and RUT down?
One-Sample T: sp_1, rut_1
Test of mu = 0 vs not = 0
Variable N Mean StDev SE Mean 95% CI T P
sp_1 8 0.04919 0.08826 0.03120 (-0.02459, 0.12297) 1.58 0.159
rut_1 8 0.03954 0.11686 0.04131 (-0.05815, 0.13724) 0.96 0.370
Both up insignificantly, SP500>RUT
What about the opposite, RUT up and SP500 down? Next quarter ret:
One-Sample T: sp_2, rut_2
Test of mu = 0 vs not = 0
Variable N Mean StDev SE Mean 95% CI T P
sp_2 4 -0.04714 0.07524 0.03762 (-0.16687, 0.07258) -1.25 0.299
rut_2 4 -0.03733 0.08112 0.04056 (-0.16641, 0.09175) -0.92 0.425
Again insignificant, but this time both down, and again SP500 is the leader.
Kind of a bullish dance?