Is there a best day of the week, time of the month, or month of the year to invest?
Financial Interest was founded by Damien Talks Money (413,000 YouTube subscribers) to teach the personal finance lessons he wished he'd been taught at school.
Last updated 28th August 2026
13 min read
The Weekend effect. The September effect. Sell in May then go away, but be sure to be back in time for the Santa Claus rally.
There's so much received wisdom bandied around when it comes to trying to time the stock market, it's easy to see how anyone could become paralysed by indecision.
But is there really a best day of the week, time of the month, or month of the year, to invest?
Using historical trading data and some complicated maths we'd rather not ever have to revisit, we set out to answer the question once and for all.
Financial Interest provides guidance, not advice. If you’re unsure about anything, speak with a qualified adviser. When investing, your capital is always at risk. Past performance does not guarantee future results.
The best and worst days of the week to invest
Warren Buffet once famously said that "the best time to invest was yesterday, the second best time is today".
But what if that's not actually true?
If stocks are reliably down on one particular day of the week, there would be a theoretical advantage to choosing this day for a regular investment – you could consistently buy in when prices are lower, picking up more shares for the same money, week after week, year after year. Over a few decades of compounding, even a tiny edge repeated often enough could meaningfully add up.
And there is one theory here that crops up time and time again.
The "weekend effect" is the idea that stocks tend to be down on Mondays compared to the preceding Friday. The theory goes that bad news has a habit of breaking over the weekend when markets are closed and investors have nothing to do but stew on it. By the time trading reopens on Monday morning, all that accumulated pessimism gets dumped into prices at once, dragging them down.
The hypothesis has been around for over 100 years, but was first properly documented by academic analyst Frank Cross in his paper, "The behaviour of stock prices on Fridays and Mondays", which identified this pattern across some major US equity indexes and individual stocks.
Later studies claim that the effect largely disappeared after 1975, but that hasn't stopped it from living on in market folklore.
But is the weekend effect actually a thing, or has some other day been the real one to watch all along?
We analysed just over 25 years of returns from the S&P 500 – or, 6,538 trading days in total – to find out. We calculated the daily return as the percentage change from one day's close to the previous day's close. So, if the S&P 500 closed at 5,000 on Monday and 5,050 on Tuesday, that's a daily return of +1% for Tuesday.
We also calculated how much returns typically varied on any given day, labelled below as the "typical daily swing". Technically, this is called the standard deviation, but just think of it as the market's background noise level.
And if we squint hard enough, we can see some patterns.
| Day | Mean return | % of positive days | Typical daily swing |
|---|---|---|---|
| Monday | +0.017% | 53.8% | 1.33% |
| Tuesday | +0.061% | 51.7% | 1.22% |
| Wednesday | +0.035% | 54.2% | 1.20% |
| Thursday | +0.033% | 54% | 1.23% |
| Friday | +0.008% | 53.7% | 1.13% |
All our days had an average positive return, and ended in the green between 51.7% and 54.2% of the time.
But the weekend effect is nowhere to be seen. Friday, which the theory suggests should perform reliably better than Monday, is actually our worst performer at 0.008%, while Monday is slightly better at 0.017%.
Meanwhile, Tuesday stands out marginally as the highest performer, historically producing the best average returns at 0.061%.
Monday has 116 fewer observations than most other days, because US public holidays disproportionately fall on Mondays. That means Monday's average is calculated from a smaller sample. However, we still have 1,226 days – enough that this is unlikely to meaningfully affect the results.
But there's a big clue here that, much like thinking you've seen the face of Jesus in your morning toast, we're seeing patterns that don't actually exist – at least, not in any meaningful sense.
There's actually just a tiny 0.053 percentage-point gap between our best day – Tuesday – and worst day – Friday. Meanwhile, we know that the S&P 500's return on any given weekday typically fluctuated by around 1.2 percentage points.
So in reality, returns were scattered across a huge range of individual outcomes, which may or may not have been anywhere near the average on any given day.
And to think you probably once told your maths teacher that understanding standard deviations would be useless in real life.
On top of that, although Tuesday has the highest average return, it also has the lowest percentage of days finishing positively – a clue that it's actually a handful of very good Tuesdays dragging up the average.
But, there's one final statistical test we can run just to be certain. It basically asks: if we took all 6,538 days and randomly shuffled them into five groups, how often would we end up with differences at least as large as the ones we’ve just found?
If the answer is "pretty often", then we can be pretty sure our results aren't telling us anything real or actionable.
This gives us a single number called a p-value, which runs from 0-1, with results below 0.05 considered statistically significant.
And ours comes back at 0.83, meaning we'd find differences this big 83% of the time just by randomly shuffling the data around.
The best and worst time of the month to invest
So, we've had no luck with days of the week. But is there a better time of the month to invest?
This is a more useful question to answer, as many people choose to invest on a monthly basis. And there is one theory here worth testing.
The "turn-of-the-month effect" – sometimes more excitingly referred to as the "ultimo effect" – theorises that stock prices usually increase during the last trading day of the month and the first three trading days of the following month.
There is some logic to this. The end of the month is when most people get paid, when dividends are reinvested, and when fund managers adjust their portfolios – all of which would, in theory, push prices up due to increased demand.
So does it really hold true?
We used the same S&P 500 dataset as before – 25 years, 6,538 trading days – splitting it into two groups: the four "turn-of-month" trading days, and every other day in between, looking at the returns of each individual day.
| Turn of the month | Rest of the month | |
|---|---|---|
| Mean return | +0.051% | +0.026% |
| % positive | 53% | 53.6% |
| Typical daily swing | 1.18% | 1.23% |
On the surface, the results look encouraging. Our turn-of-the-month days have roughly double the average daily return of the rest of the month: 0.051% versus 0.026%.
But yet again, the signal is tiny relative to the noise – a 0.025 percentage point difference in average returns, compared to a typical daily swing of around 1.2%.
And turn-of-month days were actually marginally less likely to finish positive than the rest of the month, at 53% versus 53.6%; another clue that the higher average isn't down to turn-of-month days being more reliably good, but a handful of unusually strong ones dragging the average up.
This time, our p-value – the number that tells us how often we'd get the same result by randomly mixing up data – is 0.52, still nowhere near the threshold we'd need to call this a real pattern.
So, all in all, we can pretty confidently say that although some individual "turn of the month" days have historically been better than other days, there's no reliable tendency for them to be better overall. At least, not with this index, and not across the past 25 years.
The best and worst months of the year to invest
Attempting to time the market based on day of the week or day of the month appears to be a fool's endeavour. But what about month of the year?
After all, there's no shortage of economic wisdom to choose from here.
The "January effect" refers to the supposed tendency for stock prices to rise in the first month of the year as everyone returns from the festive break with renewed optimism.
Meanwhile, the old adage "sell in May and go away" is based on the belief that stock market returns tend to be weaker between May and October than they are between November and April.
But if you don't pay attention to that, perhaps you should be wary come the end of summer – the "September effect" refers to the notion that stocks tend to underperform during this month, driven by the beginning of the election cycle in the US, and investors locking in profits before the final quarter of the year.
Then again, maybe give the following month a miss, too. The "October effect" theorises that this is often when markets suffer crashes or steep declines.
This one might actually be fair enough. The Wall Street Crash of 1929, Black Monday in 1987, the lowest point of the dot-com crash in 2002 and the peak of the 2008 financial crisis all occurred in October.
But the year at least ends on a theoretical high note, with the so-called "Santa rally" suggesting that markets often drift higher towards the end of the year, driven by a surge of festive optimism.
What's the pattern here? Well, aside from the distinct lack of imaginative names for various calendar quirks, you could find a reason to fear just about every month if you try hard enough.
But the big question is: do any of these phenomena exist for real?
If we take a look back at the S&P 500 over the past 25 years, it does seem like we can see a couple of patterns.
| Month | Mean return | % of positive months | Typical monthly swing |
|---|---|---|---|
| January | +0.12% | 52% | 4.26% |
| February | -0.50% | 50% | 4.38% |
| March | +1.06% | 61.5% | 4.73% |
| April | +1.56% | 69.2% | 4.75% |
| May | +0.66% | 73.1% | 3.64% |
| June | -0.05% | 61.5% | 4.05% |
| July | +1.54% | 69.2% | 3.73% |
| August | +0.16% | 61.5% | 3.59% |
| September | -1.31% | 50% | 4.95% |
| October | +1.34% | 61.5% | 5.62% |
| November | +2.12% | 76.9% | 4.39% |
| December | +0.58% | 65.4% | 3.54% |
Out of all our theories, only the September effect appears to hold true, with this month averaging the worst return of -1.31% and a 50/50 chance of ending up positive or negative.
But November – not known for being a month that draws much commentary amongst analysts – stands out as the clear winner, with a mean return of 2.12%, and producing a positive return for investors almost 77% of the time.
The others are disappointingly indistinguishable from one another – a cluster of mildly positive and negative months with no clear pattern.
So do we have a clear answer? Invest throughout September when prices are down, November: profit?
Maybe, maybe not.
The difference between our best and worst months is just 3.43 percentage points – again, outweighed by the typical monthly 4-5% swing in either direction.
But running our extra statistical test again gives us a p-value of 0.21 – close enough to warrant further investigation.
So what if we're just not looking hard enough?
Let's instead look at what happens if we use global market data from the MSCI World Index and go as far back as we reliably can – to 1987 – covering 39 years, or 468 months.
By adding more data, we'd expect any pattern truly driven by investor behaviour to hold up, while any pattern occurring by chance would become less pronounced or disappear.
| Month | Mean return | % of positive months | Typical monthly swing |
|---|---|---|---|
| January | +0.62% | 61.5% | 4.03% |
| February | +1.06% | 64.1% | 4.13% |
| March | +1.02% | 69.2% | 3.63% |
| April | +1.23% | 56.4% | 3.57% |
| May | +0.95% | 61.5% | 3.19% |
| June | +0.32% | 56.4% | 4.17% |
| July | +1.26% | 64.1% | 4.16% |
| August | -0.05% | 59% | 4.74% |
| September | -0.76% | 53.8% | 4.69% |
| October | +0.91% | 64.1% | 5.33% |
| November | +1.47% | 59% | 4.39% |
| December | +1.35% | 69.2% | 3.52% |
And disappointingly, all we see are the same patterns as before, only weaker.
September remains our worst month, but significantly less bad than it was before, averaging -0.73% instead of -1.31%. The percentage of months that produced a positive return has improved slightly, too, from 50% to 53.8%.
And although November is still our strongest performer, this finding has also tempered significantly – down to 1.47% from 2.12%, with the percentage of positive months dropping dramatically from 76.9% to 59%.
This time, the difference between our strongest and weakest month is just 2.2 percentage points – yet again, far less than the typical swing. And our p-value is 0.47, meaning our results have become even more consistent with what we'd expect to see by chance.
October stands out as the most volatile month in both datasets, with typical swings of 5.62% and 5.33%. These are the fingerprints of a handful of historically catastrophic months, dragging the figure up. Remove them, and October looks reassuringly boring.
But if we've seen the same results twice now, doesn't that mean they must be real?
Well, they are real in the sense that they're really there in the historical data. But being real is different from being meaningful or actionable.
Think about it like this: if you flipped a coin once on every weekday for 30 years, some days would end up with a slightly higher average of heads than others just by chance. And a coin flip is a simple 50/50 outcome. The stock market has far more variation than that, which makes it even more likely to throw up patterns that look meaningful but aren't.
Or, as Daniel Kahneman puts it in his book, Thinking, Fast and Slow: "Random processes produce many sequences that convince people that the process is not random at all."
And there's something else worth bearing in mind, too.
Even if our results were 100% consistent and reliable, trying to capitalise on them would be near impossible. To avoid any loss or exploit any gain, you’d have to buy and sell, with perfect timing, meaning trading fees and bid-offer spreads would almost certainly dwarf any tiny edge you were chasing in the first place.
What does all this mean for investors?
Right about now, you might be thinking: Well thanks. You've told me precisely nothing.
Not so fast. Just because we can probably conclude there is no "best time" to invest, that doesn't mean there are no useful takeaways.
For a start, it's actually quite reassuring to know that nobody really knows what the stock market is going to do on any given day, month, or year.
Secondly, the vast majority of days, weeks and months we looked at had a positive average return. That means that the kind of short-term dips we're talking about trying to avoid are just bumps along a road that has reliably climbed over every multi-decade stretch.
And trying to avoid those bumps isn't just notoriously difficult and costly, it's often counterproductive. Missing just the 10 best trading days in a 20-year period can reduce total returns by more than 50%, and the best and worst trading days tend to happen close together.
In fact, our previous research has shown that even the unluckiest investors in history, who put their money in the market just before major crashes, would have seen returns that substantially outperformed cash just by staying invested.
The same principle also explains why lump-sum investing has historically beaten drip-feeding money into the market. Because markets tend to rise over time, getting your money into the market sooner has generally produced better results than waiting.
Which, let's face it, is also a hell of a lot less complicated.
Ultimately, our analysis shows that there's probably no way to reliably time the stock market based on day of the week, time of the month, or month of the year. Anyone telling you otherwise is simply selling you a dream… or a trading course.
Financial Interest provides guidance, not advice. If you’re unsure about anything, speak with a qualified adviser. When investing, your capital is always at risk. Past performance does not guarantee future results.
