This Seasonal is a Turkey, from Kim Zussman
Using DJIA monthly back to 11/1928, I find that following Novembers that declined worse than -3%, subsequent Decembers are up but not statistically significant:
One-Sample T: nxt dec
Test of mu = 0 vs not = 0
Variable N Mean StDev SE Mean 95% CI T P
nxt dec 15 0.01799 0.06046 0.01561 (-0.0154, 0.0514) 1.15 0.268
Furthermore, regression of Dec return vs prior November IF November was down shows no correlation between the variables:
Regression Analysis: nxt dec versus DN nov
The regression equation is
nxt dec = 0.0259 + 0.144 DN nov
Predictor Coef SE Coef T P
Constant 0.0258 0.0142 1.81 0.081
DN nov 0.1440 0.2419 0.60 0.557
S = 0.0515600 R-Sq = 1.2% R-Sq(adj) = 0.0%
When the lean on history fail
Reach for the chalice
Thy holy grail