2/27 Reset, from Kim Zussman
How did the market change post the 2/27 bear-fake? One aspect is follow-up responses to up or down days. SPY day returns (w/div) were used to compare returns for days following drops and rises. Here is the result from 2/27 to the present (inclusive):
Two-sample T for after dn vs. after up
N Mean St Dev SE Mean
aft dn 15 0.00160 0.00859 0.0022 T=0.0
aft up 31 0.00160 0.00597 0.0011
There was more than twice the number of up days as down, and returns following both were positive, with means identical (after down had higher standard deviation). For the 48 days prior to 2/27 there was a somewhat different pattern:
Two-sample T for after dn pre vs. after up pre
N Mean St Dev SE Mean
aftdn pre 24 -0.00152 0.00892 0.0018 T=-1
aftup pre 24 0.00050 0.00499 0.0010
Here there were equal up and down days (close to long-term pattern), and unlike post 2/27 response to down days was also down.
Comparing all days post 2/27 and 48 pre-2/27 shows that the recent period also has a higher (NS) mean:
Two-sample T for post 2-27 vs. pre 2-27
N Mean St Dev SE Mean
post 2-27 46 0.00160 0.00683 0.0010 T=1.45
pre 2-27 48 -0.00051 0.00722 0.0010
Thus the rise since 2/27 came from:
- More up days than down
- Higher mean return of all days
- Reversal of down days rather than continuation