312 FOMC dates. Overnight prev close -> open, net 2bps RT, IS<2016<=OOS.
A. Cross-asset FOMC overnight
| asset | n | mean | t | IS | OOS | placebo p |
|---|---|---|---|---|---|---|
| SPY | 197 | +0.213% | +4.34 | +0.236% | +0.183% | 0.0002 |
| EFA | 197 | +0.189% | +2.53 | +0.163% | +0.223% | 0.0054 |
| VWO | 188 | +0.230% | +2.32 | +0.232% | +0.227% | 0.0156 |
| VUG | 196 | +0.271% | +5.20 | +0.275% | +0.267% | 0.0006 |
| MTUM | 110 | +0.265% | +4.76 | +0.207% | +0.281% | 0.0042 |
| HYG | 171 | +0.148% | +4.65 | +0.212% | +0.084% | 0.0002 |
| EEM | 197 | +0.216% | +2.52 | +0.184% | +0.257% | 0.0190 |
| VGK | 188 | +0.283% | +3.20 | +0.320% | +0.238% | 0.0002 |
| VPL | 188 | +0.263% | +3.20 | +0.310% | +0.206% | 0.0006 |
Risk-on basket (SPY+EFA+VWO+HYG): mean +0.200% (t +3.62), IS +0.216%, OOS +0.179%, placebo p 0.0008.
B. Beta deciles on FOMC eves (~500-name panel)
| leg | n | mean | t | IS | OOS |
|---|---|---|---|---|---|
| top-decile beta | 172 | +0.383% | +4.01 | +0.422% | +0.344% |
| bottom-decile beta | 172 | +0.065% | +1.98 | +0.078% | +0.053% |
| top - bottom | 172 | +0.318% | +4.02 | +0.344% | +0.291% |
C. Multi-night (SPY, overnight into t-k)
| night into | n | mean | t | IS | OOS |
|---|---|---|---|---|---|
| t-0 | 197 | +0.213% | +4.34 | +0.236% | +0.183% |
| t-1 | 197 | -0.033% | -0.59 | -0.136% | +0.099% |
| t-2 | 197 | -0.020% | -0.43 | +0.002% | -0.048% |
| t-3 | 197 | -0.007% | -0.12 | -0.093% | +0.104% |
Reading
- A: if the risk-on basket / high-beta ETFs beat SPY OOS, the event exposure can be diversified or amplified.
- B: top-decile beta >> bottom => the drift is a beta effect; size via high-beta names instead of SPY.
- C: t-0 is the eve; a positive t-1/t-2 means the drift extends back.
Time 9s.