Happiness and Certainty, from Kim Zussman
Assuming the majority hold stocks long, happiness is associated with a large rise, and unhappiness with a large decline. Independently of happiness, uncertainty can be thought of as volatility; for example, intra-day range.
SPY 93-present was checked for daily return/happiness as well as range/uncertainty, defined as:
(H-L) / {(H+L)/2}
Uncertain days were defined as those with range >3% (V). Happy days were up >2% (U), and unhappy ones down more than -2% (D).
With these assumptions let's see whether uncertainty presages uncertainty or unhappiness. First, compare returns "r" after high-uncertainty days which were either happy or unhappy:
t-Test: Two-Sample Assuming Unequal Variances
VUr VDr
Mean 0.0007 0.0045
Variance 0.0004 0.0011
Observations 70 99
Hypothesized Mean Difference 0
df 165.0000
t Stat -0.9313
P(T<=t) one-tail 0.1765
Uncertain days which were either happy or unhappy were both followed by up days on average, though the difference was NS (due to higher variance after down days).
Next see what happens to uncertainty after uncertain days which were either happy or unhappy:
t-Test: Two-Sample Assuming Unequal Variances
VUv VDv
Mean 0.0303 0.0421
Variance 0.0003 0.0007
Observations 70.0 99.0
Hypothesized Mean Difference 0.0
df 166.0000
t Stat -3.6144
P(T<=t) one-tail 0.0002
After uncertain days, uncertainty after happy days (up) was significantly lower than after unhappy days (down).
When uncertainty is high, market participants become more uncertain when they are unhappy.