198 FOMC eves with SPY overnight data. FOMC leg net 2bps RT. Host idle-cash measured on the eve (t-1). IS<2016-01-01<=OOS.
Host idle-cash on FOMC eves
| host | eve bars | fully-flat eves | flat frac | standalone flat frac (all days) |
|---|---|---|---|---|
| eqmom | 171 | 18 | 10.5% | 8.9% |
| flagship | 169 | 26 | 15.4% | 11.8% |
Overlay P&L (FOMC net booked on host idle cash)
Annualised at the REAL event frequency (8.7 eves/yr), not 252.
| host | cash assumption | overlay events run | overlay net/event | overlay ann (host book) |
|---|---|---|---|---|
| eqmom | 100% of equity cash | 18 | +0.380% | +0.30% |
| eqmom | 10% cash buffer | 18 | +0.038% | +0.03% |
| flagship | 100% of equity cash | 26 | +0.321% | +0.37% |
| flagship | 10% cash buffer | 26 | +0.032% | +0.04% |
Ceiling: standalone sleeve (100% capital every eve)
- FOMC-only standalone: per-event +0.194% x 8.7 eves/yr = +1.69%/yr on dedicated capital, n 198.
Read
- If the hosts' flat-fraction on eves is small, the overlay can only run on a fraction of events -> the realisable ann is a fraction of the standalone +1.7%/yr ceiling. That is the correct number to compare against the M2 (dedicated account) option, not the standalone return.
- The overlay's hidden cost (order-flow coupling, reconciliation in the host account) is unaudited here and must be weighed against the small realisable P&L.
Time 0s.