Graveyard / Calendar effects / US stocks
The payday effect
Is the 16th of the month a good day to own stocks?
The idea
Paychecks land on the 15th and retirement contributions are invested the next session, so the 16th should see extra buying. Own the index for that one day.
How we tested it
| Data | SPY with the VIX as an optional filter. Long from the close of the last session on or before the 15th to the close of the next session. Variant: hold two sessions. |
|---|---|
| Period | February 1994 to October 2026 (392 months). |
| Costs | 1 bp per side, 6.5 bp per side as a stress case. |
| Holdout | In-sample and out-of-sample split; out-of-sample 78 trades. |
| Control | No random-entry control was run for this test. SPY on the 16th was compared with the all-day mean. |
Result
| Trades | 392 |
|---|---|
| Sharpe, one session | 0.34 |
| Sharpe in-sample / out-of-sample | 0.40 / 0.06 |
| Sharpe at 6.5 bp per side | 0.04 |
| Max drawdown | −14% |
| Mean return on day 16, before 2010 / from 2010 | +28.7 bp / +3.5 bp |
| All-day mean for comparison | +5.9 bp |
Why it failed
The effect decayed after the original papers, as their authors expected once employers moved to other pay schedules. From 2010 the 16th is no better than any other day, and at higher costs the strategy earns almost nothing.
What would make it worth retesting
No. Shifting the days to rescue it would be data mining.
Test id: payday-spy-event-v1. Past results of a backtest say nothing certain about the future.