Most seasonal edges you read about online are noise wearing a calendar as a costume. The turn-of-the-month effect is one of the few that keeps surviving serious academic scrutiny, which is exactly why it gets oversold. It is real, it is measurable, and it is far smaller and far more specific than the people selling it want you to think. Here is the plain version. Across US equity indices, a disproportionate share of the market's long-run gains has historically landed in a narrow window around the month boundary: roughly the last trading day of the month and the first three or four trading days of the next one. Strip those few days out and the rest of the month has, over long samples, contributed close to nothing. That is the claim. Now let us take it apart so you know what to do with it, and more usefully, what not to. What the turn-of-the-month effect actually is The effect describes a tendency for equity returns to concentrate in a short window that straddles the calendar boundary. The exact window depends on who is measuring. A common definition runs from the last trading day of the month (often written as day -1) through the first three trading days of the new month (+1, +2, +3). Some studies stretch it to +4, some start it a day earlier. The point is that it is defined in trading days , not calendar days, so a weekend or a holiday shifts the whole window. The original work goes back decades. Frank Russell and later researchers found that in the S&P 500, the average return over that handful of days accounted for essentially all of the index's positive drift, with the middle of the month roughly flat on average. That is a striking result, and it has held up across enough countries and enough decades to be taken seriously rather than dismissed as pure data-mining. It is not a guarantee that the market rises every turn. It is a statement about averages over hundreds of months. Why it happens Nobody has a single clean answer, but the honest list of drivers is about flows, not magic: Payroll and retirement contributions. A large slug of money hits pension and retirement accounts at month-end and gets mechanically deployed into index funds. Those funds buy regardless of price. Month-end rebalancing. Funds that target fixed asset weights trim and top up around the boundary. When equities have drifted, some of that flow lands as buying. Performance reporting. Managers who are judged on month-end marks have an incentive to hold or add to winners into the close, sometimes called window dressing. Dividend reinvestment and bonus cash. Recurring cash events cluster near the boundary and get put back to work quickly. None of these are exotic. They are calendar-driven cash flows meeting a market, and they leave a footprint. That footprint is also why the effect is strongest where passive index buying is heaviest, and weaker where flows are more discretionary. Where it shows up, and where it does not The strongest and most repeatedly documented version lives in US large-cap equity indices : the S&P 500 and the Dow. It also appears, with varying strength, in international equity indices and in a lot of developed markets. If you want the wider frame around this, the broader story of stock market seasonality covers the other calendar patterns that interact with it, and the S&P 500 seasonality picture is where the turn-of-the-month signal is cleanest. In forex it gets murkier. Month-end brings its own well-known flow event (the so-called London 4pm fix and month-end rebalancing hedges), and that can push currencies around, but it is not the same tidy long-only tilt you see in equities. If a fund needs to rebalance a large equity book, it may have to buy or sell currency to hedge, and the direction depends on how markets moved that month. So the turn-of-the-month in FX is a flow event you should be aware of rather than a directional edge you can bank on. In bonds, commodities and crypto , the evidence is thinner still. Crypto in particular has no payroll cycle feeding it, so borrowing an equity anomaly and expecting it to hold there is wishful thinking. What it is good for Treat the turn-of-the-month effect as context, not a trigger . That distinction matters. A few sensible uses: Tilting existing setups. If your system already wants to go long an index and the window is opening, that is mild tailwind. If your system wants to short into the same window, that is a reason to double-check the trade. Timing entries you were going to take anyway. Nudging a planned long entry toward the start of the window rather than the dead middle of the month is low-cost and reasonable. Avoiding fighting the flow. Knowing that mechanical buying tends to appear keeps you from being surprised by strength that has nothing to do with your thesis. It sits alongside other calendar quirks like the day-of-week effect . Individually each one is a small tilt. Stacked thoughtfully, they help you avoid taking trades at statistically poor moments. What it is bad for The failure mode is treating a small average as a reliable per-month payout. The edge is small. We are talking about a modest average return over a few days. Spreads, commissions and slippage can eat a large share of it, especially if you trade it with leverage or on a product with a wide spread. It does not happen every month. Plenty of turns are flat or negative. The pattern is a long-run average, which means you will sit through stretches where it does nothing and feel like it broke. It is crowded. An anomaly this well-known invites front-running. Some research suggests the window has drifted earlier over time as traders try to get ahead of it, which is what you would expect once an edge is public. It is not a standalone system. On its own, buying every turn and selling four days later is thin, and thin edges are the easiest to overfit. If you optimise the exact window on past data until it looks perfect, you have almost certainly curve-fit it. How to test it before you trust it Do not take the number from a blog (including this one) as your edge. Measure it on the instrument you actually trade, with your actual costs. The workflow is simple: Define the window precisely. Pick your entry day and exit day in trading days, write it down, and do not move it after you see the results. Moving the window to flatter the backtest is the fastest way to fool yourself. Get a decent sample. One turn a month means twelve observations a year, so a few years is a small sample. You want many years to say anything honest, which ties into the wider question of how to backtest a strategy without writing code by clicking through history and logging what actually happened. Log every attempt, winners and losers. Tag each turn-of-the-month trade in your journal so you can look back and separate the pattern from your feelings about it. In TradeSave+ you can tag trades by setup and filter your stats to that tag, so the turn-of-the-month trades sit in their own bucket and you can read the real win rate, the average result and the drawdown you sat through rather than the version you remember. Judge it on expectancy, not vibes. A pattern that wins often but pays little can still lose money after costs. Look at the average result per trade net of spread, and be suspicious if the whole edge rests on two or three exceptional months. The honest takeaway The turn-of-the-month effect is one of the more durable seasonal patterns in equities, driven by real, boring cash flows rather than sentiment. That makes it worth understanding. It also makes it a poor candidate for a mechanical, high-leverage system, because the effect is small, uneven and increasingly anticipated. Use it as a tilt on trades you were going to take, size it as the minor factor it is, and let your own logged results decide how much weight it earns. A calendar pattern is only an edge once your journal, not a study, says it survives your costs.
The Turn-of-the-Month Effect Explained (and Why It Is Mostly a Stock Story)
A real seasonal pattern that clusters equity gains around the month boundary, plus what it is genuinely useful for and where it falls apart.
Most seasonal edges you read about online are noise wearing a calendar as a costume. The turn-of-the-month effect is one of the few that keeps surviving serious academic scrutiny, which is exactly why it gets oversold. It is real, it is measurable, and it is far smaller and far more specific than the people selling it want you to think. Here is the plain version. Across US equity indices, a disproportionate share of the market's long-run gains has historically landed in a narrow window around the month boundary: roughly the last trading day of the month and the first three or four trading days of the next one. Strip those few days out and the rest of the month has, over long samples, contributed close to nothing. That is the claim. Now let us take it apart so you know what to do with it, and more usefully, what not to. What the turn-of-the-month effect actually is The effect describes a tendency for equity returns to concentrate in a short window that straddles the calendar boundary. The exact window depends on who is measuring. A common definition runs from the last trading day of the month (often written as day -1) through the first three trading days of the new month (+1, +2, +3). Some studies stretch it to +4, some start it a day earlier. The point is that it is defined in trading days , not calendar days, so a weekend or a holiday shifts the whole window. The original work goes back decades. Frank Russell and later researchers found that in the S&P 500, the average return over that handful of days accounted for essentially all of the index's positive drift, with the middle of the month roughly flat on average. That is a striking result, and it has held up across enough countries and enough decades to be taken seriously rather than dismissed as pure data-mining. It is not a guarantee that the market rises every turn. It is a statement about averages over hundreds of months. Why it happens Nobody has a single clean answer, but the honest list of drivers is about flows, not magic: Payroll and retirement contributions. A large slug of money hits pension and retirement accounts at month-end and gets mechanically deployed into index funds. Those funds buy regardless of price. Month-end rebalancing. Funds that target fixed asset weights trim and top up around the boundary. When equities have drifted, some of that flow lands as buying. Performance reporting. Managers who are judged on month-end marks have an incentive to hold or add to winners into the close, sometimes called window dressing. Dividend reinvestment and bonus cash. Recurring cash events cluster near the boundary and get put back to work quickly. None of these are exotic. They are calendar-driven cash flows meeting a market, and they leave a footprint. That footprint is also why the effect is strongest where passive index buying is heaviest, and weaker where flows are more discretionary. Where it shows up, and where it does not The strongest and most repeatedly documented version lives in US large-cap equity indices : the S&P 500 and the Dow. It also appears, with varying strength, in international equity indices and in a lot of developed markets. If you want the wider frame around this, the broader story of stock market seasonality covers the other calendar patterns that interact with it, and the S&P 500 seasonality picture is where the turn-of-the-month signal is cleanest. In forex it gets murkier. Month-end brings its own well-known flow event (the so-called London 4pm fix and month-end rebalancing hedges), and that can push currencies around, but it is not the same tidy long-only tilt you see in equities. If a fund needs to rebalance a large equity book, it may have to buy or sell currency to hedge, and the direction depends on how markets moved that month. So the turn-of-the-month in FX is a flow event you should be aware of rather than a directional edge you can bank on. In bonds, commodities and crypto , the evidence is thinner still. Crypto in particular has no payroll cycle feeding it, so borrowing an equity anomaly and expecting it to hold there is wishful thinking. What it is good for Treat the turn-of-the-month effect as context, not a trigger . That distinction matters. A few sensible uses: Tilting existing setups. If your system already wants to go long an index and the window is opening, that is mild tailwind. If your system wants to short into the same window, that is a reason to double-check the trade. Timing entries you were going to take anyway. Nudging a planned long entry toward the start of the window rather than the dead middle of the month is low-cost and reasonable. Avoiding fighting the flow. Knowing that mechanical buying tends to appear keeps you from being surprised by strength that has nothing to do with your thesis. It sits alongside other calendar quirks like the day-of-week effect . Individually each one is a small tilt. Stacked thoughtfully, they help you avoid taking trades at statistically poor moments. What it is bad for The failure mode is treating a small average as a reliable per-month payout. The edge is small. We are talking about a modest average return over a few days. Spreads, commissions and slippage can eat a large share of it, especially if you trade it with leverage or on a product with a wide spread. It does not happen every month. Plenty of turns are flat or negative. The pattern is a long-run average, which means you will sit through stretches where it does nothing and feel like it broke. It is crowded. An anomaly this well-known invites front-running. Some research suggests the window has drifted earlier over time as traders try to get ahead of it, which is what you would expect once an edge is public. It is not a standalone system. On its own, buying every turn and selling four days later is thin, and thin edges are the easiest to overfit. If you optimise the exact window on past data until it looks perfect, you have almost certainly curve-fit it. How to test it before you trust it Do not take the number from a blog (including this one) as your edge. Measure it on the instrument you actually trade, with your actual costs. The workflow is simple: Define the window precisely. Pick your entry day and exit day in trading days, write it down, and do not move it after you see the results. Moving the window to flatter the backtest is the fastest way to fool yourself. Get a decent sample. One turn a month means twelve observations a year, so a few years is a small sample. You want many years to say anything honest, which ties into the wider question of how to backtest a strategy without writing code by clicking through history and logging what actually happened. Log every attempt, winners and losers. Tag each turn-of-the-month trade in your journal so you can look back and separate the pattern from your feelings about it. In TradeSave+ you can tag trades by setup and filter your stats to that tag, so the turn-of-the-month trades sit in their own bucket and you can read the real win rate, the average result and the drawdown you sat through rather than the version you remember. Judge it on expectancy, not vibes. A pattern that wins often but pays little can still lose money after costs. Look at the average result per trade net of spread, and be suspicious if the whole edge rests on two or three exceptional months. The honest takeaway The turn-of-the-month effect is one of the more durable seasonal patterns in equities, driven by real, boring cash flows rather than sentiment. That makes it worth understanding. It also makes it a poor candidate for a mechanical, high-leverage system, because the effect is small, uneven and increasingly anticipated. Use it as a tilt on trades you were going to take, size it as the minor factor it is, and let your own logged results decide how much weight it earns. A calendar pattern is only an edge once your journal, not a study, says it survives your costs.