Monolith Research
In plain terms
Some calendar effects — turn-of-month and seasonal patterns — recur in equity returns. This checks whether they survive honest testing.
How it works
Calendar features are turned into cross-sectional tilts and held only over the windows the pattern predicts.
What it’s tested against
Out-of-sample with stressed costs and a placebo on shuffled calendars, to separate a real seasonal effect from data-mined coincidence.
Data
CRSP point-in-time panel.
Researched — calendar effects held to the bar.