Market statistics
Day-of-week and turn-of-the-month effects, measured across three eras
Monday was the worst day of the week for the first two thirds of this record and has been an ordinary one since 1993. That decay is the finding — an anomaly cut into eras tells you something an anomaly averaged over a century cannot.
24,781 sessions · data through 28 Aug 2026
1928–2026
- Worst weekday, whole record
- Monday
- -0.059% per session
- Monday since 1993
- +0.041%
- 55.0% of Mondays positive
- Best weekday, whole record
- Wednesday
- +0.075% per session
- Turn-of-month premium
- +0.127%
- against +0.009% on other sessions
| Weekday | Average | Positive | Cumulative | 1928-1959 | 1960-1992 | 1993-present |
|---|---|---|---|---|---|---|
| Monday4,766 | -0.059% | 49.4% | -96% | -0.113% | -0.104% | +0.041% |
| Tuesday5,035 | +0.047% | 51.5% | +644% | +0.038% | +0.035% | +0.066% |
| Wednesday5,052 | +0.075% | 54.3% | +3,045% | +0.079% | +0.097% | +0.051% |
| Thursday4,975 | +0.039% | 52.6% | +413% | +0.066% | +0.035% | +0.018% |
| Friday4,943 | +0.054% | 54.4% | +959% | +0.066% | +0.071% | +0.026% |
| Position | Average | Median | Positive | Sessions |
|---|---|---|---|---|
| 5 sessions before month end | -0.071% | -0.013% | 48.1% | 1,184 |
| 4 sessions before month end | +0.012% | +0.029% | 51.7% | 1,184 |
| 3 sessions before month end | +0.009% | +0.021% | 51.3% | 1,184 |
| 2 sessions before month end | +0.097% | +0.088% | 55.2% | 1,184 |
| Last session of the month | +0.085% | +0.095% | 54.6% | 1,184 |
| 1st session of the month | +0.142% | +0.172% | 58.2% | 1,184 |
| 2nd session of the month | +0.154% | +0.136% | 58.5% | 1,184 |
| 3rd session of the month | +0.124% | +0.108% | 55.4% | 1,184 |
| 4th session of the month | +0.052% | +0.038% | 51.5% | 1,184 |
| 5th session of the month | -0.018% | +0.008% | 50.3% | 1,184 |
| 6th session of the month | +0.021% | +0.065% | 52.2% | 1,184 |
| 7th session of the month | +0.016% | +0.030% | 51.5% | 1,184 |
| 8th session of the month | +0.017% | +0.049% | 52.3% | 1,184 |
| 9th session of the month | -0.014% | +0.023% | 51.1% | 1,184 |
| 10th session of the month | +0.004% | +0.026% | 51.2% | 1,184 |
| 11th session of the month | +0.086% | +0.108% | 56.1% | 1,184 |
| 12th session of the month | -0.003% | +0.033% | 51.2% | 1,184 |
| Period | Turn-of-month session | Every other session | Gap | Positive, in window |
|---|---|---|---|---|
| Whole record | +0.127% | +0.009% | +0.118% | 56.7% |
| 1928-1959 | +0.194% | -0.014% | +0.208% | 59.2% |
| 1960-1992 | +0.120% | +0.006% | +0.114% | 56.6% |
| 1993-present | +0.068% | +0.034% | +0.034% | 54.5% |
Price returns from daily closing levels. Index price returns exclude dividends. Full provenance, method and a citation line are in Sources and method below.
What happens to an anomaly after it is published
Monday's average session across the whole record is -0.059% — the only negative weekday in the table, and the finding that the academic literature named the weekend effect around 1980. Split the same sessions into eras and it stops being one fact: negative in 1928-1959, negative in 1960-1992, and +0.041% in 1993-present. The effect that got published is not the effect that is there now.
The turn of the month has held up better and it is also fading. Across the whole record, the last session of a month plus the first three of the next averaged +0.127% a session against +0.009% for every other session — roughly 14 times as much per session. In the most recent era the same comparison is +0.068% against +0.034%: still positive, much less dramatic.
Both patterns have plausible mechanisms — settlement and news timing around weekends, salary and pension flows at month end — and neither mechanism is strong enough to survive being traded. What the era columns show is the ordinary life cycle of a published anomaly: documented, arbitraged, and left as a residue too small to cover the spread.
Read the sizes before reading the pattern. The largest gap on this page is a few hundredths of a percent per session. That compounds into something visible over decades, which is why the cumulative column is here, but it is well inside transaction costs for anyone trading it directly — and the strongest evidence on the page is evidence that trading it is what removed it.
Questions people ask about this
- Is the Monday effect still real?
- Not in the last three decades of this record. Monday's average return is negative across the full history and negative in both of the earlier eras, and it is positive in the era since 1993. The effect was documented in the academic literature around 1980; what the era split shows is the pattern that literature described weakening after it became common knowledge.
- What is the turn-of-the-month effect?
- The tendency for the last session of a month and the first few of the next to carry a disproportionate share of the month's return. It is one of the more robust calendar patterns and it is visible in every era here, but the gap between those sessions and the rest of the month has narrowed substantially since 1993.
- Can I trade these?
- The averages here are hundredths of a percent per session, which is inside the cost of trading them for most people, and the pattern that was strongest is the one that has faded most. The page is a description of what the record contains, not a strategy — and the strongest evidence on it is evidence that acting on published calendar effects is what removes them.
Sources and method
- Data
- Financial Modeling Prep — Daily adjusted closing levels and quotes, retrieved through Plutux's own data service.
- S&P Dow Jones Indices — Publisher and methodology owner of the S&P 500 index itself.
- How it was calculated
- Every close-to-close change is assigned a weekday and a position within its own month, counted both from the first session of the month and from the last. The turn-of-the-month window is the final session of a month plus the first three of the next, compared with every other session on a per-session basis. The era boundaries are 1960, roughly when the anomaly literature began, and 1993, the arrival of the first US index ETF; they are a choice and are stated so a reader can disagree with them. Price returns, dividends excluded.
- How often it changes
- Regenerated from the full daily history about once a year; the date it runs through is at the top of the page.
- Citing this page
Free to quote — please link rather than copy the table.
Plutux. "Day-of-week and turn-of-the-month effects, measured across three eras." Data through 28 Aug 2026. https://plutux.ai/resources/tools/stock-market-calendar-effects
Historical figures for information only — not investment advice, and not a forecast.
Related tools
- Monthly seasonalityWhat every calendar month has actually done across ninety-eight years — average, median, and how often it finished up.
- Streaks and base ratesThe odds of an up day, week, month, quarter and year in one table — plus the longest winning and losing runs on record.
- Overnight vs intradayEvery session split into the gap and the trading day — and the reason the index and the fund give opposite answers.
Plutux is not an investment adviser. Market data and AI-generated analysis are for information and education only, not investment advice. Disclaimer