As with many published anomalies, the classical Turn-of-the-Month (TOM) effect, originally defined as the period from the last trading day of the month to three days after, appears to have largely disappeared in the past decade....

There are ways to stack the deck in your favour without really *relying* on the effect.

For example, say it's a few days before the end of the month, and you're planning on buying some stock shortly. Better to do it right away, not wait a week. If you're thinking of selling something, you might want to wait till after the end of month.

I have noticed that the effect has (apparently) disappeared from easy to trade indexes, but has (apparently) persisted more strongly in certain particular stocks. As if some people still do end-of-month buys automatically from their pay packets causing the effect, but traders haven't bothered to look at individual names.

It's an interesting effect in that, assuming undiscerning month-end automatic buying is the true underlying cause, there is no really obvious way for it to be arbitraged away. If someone tries to front-run it, the optimal buy date just moves a hair earlier until the incremental return is balanced on not being worth it to push further...then sits at that date indefinitely. A testable hypothesis: the optimal number of days before end of month to do your buying is a function of prevailing interest and liquidity conditions. (and, for the deeper past, trading costs)

Jim