Semi-monthly 401k contributions auto-buy on a fixed payroll schedule — no choice in timing. The question: do those buy dates systematically land when the market is locally high? Each contribution date's SPY price is compared to the average of the surrounding 14 trading days (7 before + 7 after). In a fair, efficient market with random timing you'd expect ~50% above and ~50% below — and the deviations would cancel out. Do they?
Avg Deviation
—
vs window avg
Bought High
—
of contributions
Bought Low
—
of contributions
Since Jan 2025
—
avg deviation
High-Buy Avg
—
when high, how high
Low-Buy Avg
—
when low, how low
Statistical Sig.
—
z-score / p-value
SPY Price + Contribution Dates
Bought high
Bought low
Deviation per Contribution (% vs window avg)
Rolling 12-Month Average Deviation
Is the timing bias getting worse over time?
Distribution of Deviations
Shape of the spread — symmetric or skewed?
Paycheck Timing Breakdown
Mid-month vs end-of-month — one worse than the other?
All contribution dates
| Date | SPY Close | Window Avg | Deviation | Result |
|---|