The Weekend and Day-of-the-Week Effect

If you averaged US stock market returns by day of the week through most of the 20th century, Monday stood out — its average return was not just lower than other days, it was reliably negative, while Friday's average return was the highest of the week. That gap was large and persistent enough that a trader who simply bought stocks Friday afternoon and sold Monday morning, doing nothing else, would have beaten a buy-and-hold strategy over long stretches of the sample. This is the weekend effect, one of the earliest documented calendar anomalies in finance, and studying why it existed — and why it largely stopped working — teaches a general lesson about anomaly discovery.

What caused it — and the competing theories

No single explanation is fully settled, which is itself instructive. One theory ties it to settlement conventions: in the era of physical trade settlement, buying on Friday meant paying later than buying any other day of the week (because settlement periods spanned the weekend), effectively giving Friday buyers a few extra free days of financing — a real, if small, incentive to prefer buying on Fridays and pushing prices up, and to prefer selling on Mondays. A second theory points to corporate behavior: companies historically released bad news disproportionately after Friday's close, when trading volume and media attention were lowest, so the negative news hit prices over the weekend and applied downward pressure at Monday's open. A third, more behavioral, theory suggests retail investors do more of their trading — and their selling, driven by weekend reflection on their portfolios — on Mondays, pushing prices down independent of any informational cause.

Worked example

Suppose historical data shows the average Friday-to-Monday return on a broad index was -0.10% over a multi-decade sample, versus an average two-day return of roughly +0.03% for any other adjacent pair of trading days. A strategy that goes short the index at Friday's close and covers at Monday's open, sized at $1 million notional, earning that average -0.10% edge, would net roughly $1,000 per weekend before costs — small per trade, but compounding across fifty-two weekends a year into a return stream uncorrelated with holding stocks through the week. In practice, once this pattern became well documented in academic papers during the 1980s, and settlement conventions moved to shorter cycles removing the financing-timing incentive, the gap between Monday and other days narrowed substantially in the following decades' data, consistent with the anomaly being traded away once it was known and its structural cause partly removed.

What this means in practice

The weekend effect is taught less because it's tradeable today — it largely isn't, at meaningful size, after costs — and more because it's a clean case study in how researchers distinguish a real anomaly from noise: multiple independent studies, across different markets and time periods, found the same pattern, which is the bar a calendar effect needs to clear before it's treated as more than a statistical fluke.