Reflexion Weeks, from Kim Zussman
This last week the SP was up about 3.8%, following the prior week which was down about -3%. Using SPY, looked for consecutive calendar weeks with the first week down more than -2% ("DN wk"), and the next week up more than 2% ("UP wk"); then checked the return for the third week ("NXT wk"):
One-Sample T: DN wk, UP wk, NXT wk
Test of mu = 0 vs not = 0
Variable N Mean StDev SE Mean 95% CI T
DN wk 37 -0.0462 0.0337 0.0055 (-0.0574, -0.0349) -8.34
UP wk 37 0.0481 0.0260 0.0042 ( 0.0395, 0.0568) 11.26
NXT wk 37 -0.0021 0.0276 0.0045 (-0.0112, 0.0070) -0.47
NXT week was down slightly, N.S.
Bivariate regression suggested the size of DN wk was an important predictor (smaller DN wk resulting in higher NXT wk). Here are the results if -0.05 < Dn wk < -0.02 :
One-Sample T: DN wk_1, UP wk_1, NXT wk_1
Test of mu = 0 vs not = 0
Variable N Mean StDev SE Mean 95% CI T
DN wk_1 25 -0.0296 0.0082 0.0016 (-0.0330, -0.0262) -17.95
UP wk_1 25 0.0397 0.0166 0.0033 ( 0.0328, 0.0466) 11.94
NXT wk_1 25 0.0005 0.0251 0.0050 (-0.0097, 0.0109) 0.12
Now NXT weeks were about flat, N.S.