Mean reversion, or regression to the mean.
In finance, the concept of “reversion to the mean” is often used to explain how certain investment-related measures inevitably tend to return to their average over time. More technically, this is known as regression to the mean, and it is in fact a statistical phenomenon observed not only in finance but also in many aspects of everyday life.
Before looking at the markets, it is worth illustrating it with some everyday examples, because the underlying mechanism is exactly the same:
- Children’s height. Very tall parents tend to have tall children, but generally children who are somewhat shorter than they are (closer to the population average). Likewise, very short parents tend to have somewhat taller children. This is precisely the phenomenon that led Francis Galton to coin the term “regression to the mean” in the 19th century.
- Sports performance. A striker who scores in every game for a month will eventually go through a dry spell; one who was experiencing their worst run will start scoring again. The famous “magazine cover curse” — where an athlete featured on the cover after an outstanding season often performs worse the following year — is not bad luck: an exceptional season is, by definition, well above their average, and the most likely outcome is a return towards it.
- Health. Many mild symptoms reach their worst point (which is why we go to the doctor) and then improve on their own, simply because they started from an extreme. This is one reason why we sometimes attribute an improvement to a remedy when it would have happened anyway.
- Climate. An extraordinarily hot summer is rarely followed by an even more extreme one; statistically, a year closer to the historical average is more likely… allowing, of course, for the potential effects of climate change.
The same thing happens in all these cases: when we observe a value far from its average, the next observation tends to move closer to it, not because there is some magical “corrective force”, but because extreme values usually combine a structural component (talent, genetic height, the average climate of a region) with an element of chance that is unlikely to repeat in the same direction.
Applied to investing, this principle is most apparent in two measures that should be considered together: returns and volatility. Both fluctuate around a long-term average, move away from it for periods that can sometimes be lengthy, and eventually revert. Understanding this dual dynamic helps put both the euphoria of good years and the panic of market downturns into perspective.
One important clarification: this is particularly applicable to large diversified indices (S&P 500, MSCI World, etc.), and less so to individual companies. A broad index has a “mean to revert to” because it is constantly renewed: it incorporates winning companies and removes those in decline, allowing the index as a whole to maintain its long-term trend. An individual stock, by contrast, is exposed to company-specific factors — competition, technological disruption, poor management, regulatory changes — that can permanently depress its share price. There is no guarantee of mean reversion in such cases: some companies fall and never recover, while others disappear altogether.
That is why mean reversion provides a strong argument in favour of investing in the market as a whole and diversifying.
Ultimately, it is another reason to invest through the major market indices.
Mean reversion in returns
The long-term average return of the S&P 500* is around 10%-11% annualised, depending on the period considered.
But this average is almost never achieved year after year: markets go through periods clearly above the average and others below it, including negative years.
Our investment portfolios have already experienced these effects since 2017, going through periods when returns were below expectations, for example in 2018 and 2022. Conversely, recent years are a good example of the bullish extreme. Driven by technology and artificial intelligence, we have seen several consecutive years of returns well above the historical average.
When this happens, regression to the mean reminds us that it is reasonable to expect, at some point, corrections or weaker-return years that bring the average back towards its “natural” long-term level. We do not know when, but we do know that the average eventually reasserts itself.
Mean reversion in volatility
Volatility, which is simply the variability of prices, is intrinsic to equity markets. The most common way to measure it is the VIX index, which reflects the expected 30-day implied volatility of the S&P 500 and has a historical average of around 19-20 points.
Volatility reverts to its mean just like returns do, but much more abruptly: it can remain dormant for months and surge within a matter of hours. When we published the first version of this article, the VIX was close to historical lows. In July 2017, it was trading below 10, far from its average, and we wrote that, according to the principle of mean reversion, it would tend to return to its average “at some point”. What has happened since then is almost a textbook demonstration:
- March 2020: with the outbreak of the pandemic, the VIX reached a record closing high above 82. Within a few weeks, markets went from complacency to panic.
- 2022: inflation and interest rate hikes kept the VIX persistently above its average for much of the year.
- August 2024: the sudden unwinding of the yen carry trade triggered the largest intraday spike in the history of the VIX, which reached 65 points… before normalising within a matter of days.
- April 2025: the announcement of widespread tariffs in the US sent markets sharply lower and pushed the VIX to a close of 52, its highest closing level outside 2008 and 2020. In less than 100 days, it was back below 20, one of the fastest reversions ever recorded, accompanied by a strong recovery rally.
And today? The VIX is trading in the 15-19 range, meaning it is practically at its historical average. Neither the extreme complacency of 2017 nor the panic of 2020 or 2025: once again, volatility has done what statistics suggested it would do — revert to its mean.
Returns and volatility: two sides of the same coin
This is where the connection that ties everything together becomes clear. Returns and volatility do not revert independently: they tend to move in opposite directions and are closely linked. Major spikes in volatility (a high VIX) almost always coincide with sharp declines in returns, and it is precisely these moments — when returns are most depressed relative to their average — that precede the strongest recoveries. Put another way: when volatility moves significantly above its mean, returns are usually moving significantly below theirs, and both tend to revert at the same time.
That is what we saw in April 2025: the VIX at 52 and stock markets sharply down… followed by volatility reverting below 20 and a rally of more than 35% within a few months. Investors who mistook the volatility spike for a signal to sell missed the mean reversion in returns.
What does this have to do with your investment plan?
Quite a lot — and not that much. Low volatility does not mean it will spike tomorrow, and several years of excellent returns do not mean a correction is about to happen immediately. A market metric can remain out of line with its average for a long time: the VIX spent years at abnormally low levels before 2020, and the S&P 500 has recorded several years well above its recent average. This is why trying to time the market based on these signals has historically been a poor strategy.
Does this mean it is better to wait before investing? In our view, no. Waiting for volatility to fall, or for valuations to return to their average, only ensures that we take on the risk of doing nothing or the risk of inflation. And a rise in volatility does not necessarily have to be bad for a well-constructed portfolio: in fact, regular contributions made during periods of market panic have historically been among those that contribute most to long-term returns.
A diversified portfolio and mean reversion
If we build a diversified portfolio, we are also diversifying mean reversion, both in terms of returns and volatility. One asset in our portfolio may be significantly and favourably out of line with its average, but the advantage of diversification is that, when this happens, another uncorrelated asset may be in precisely the opposite situation.
This is the power of diversification, which allows risk to be reduced by combining assets optimally. And this optimisation will continue to work even as volatility and returns revert to their average levels, as sooner or later they will.
Appendix
*Why do we use the S&P 500 as a reference?
Simply because it is the index with the longest reliable history: it has existed as a 500-stock index since 1957, and reconstructed data series go back to the 1920s. The more decades of data available, the more robust the “mean” we are referring to and the more meaningful it becomes to talk about reversion towards it. An index with only ten or fifteen years of history would barely allow us to distinguish the long-term trend from the noise.
Does the same phenomenon occur in major global indices? Yes. If, instead of the S&P 500, we look at the MSCI World, which covers the stock markets of the main developed countries and has been calculated since 1969, we find the same behaviour: a long-term average return (slightly lower, at around 8%-9% annualised) around which annual returns fluctuate above and below before eventually reverting. The global average is somewhat lower than that of the S&P 500, mainly because of the exceptional strength of US technology over the past decade, but the mean-reversion mechanism is identical. In fact, the gap between the US and the rest of the world is itself a candidate to narrow over time — another example of mean reversion in progress.
The practical conclusion: mean reversion is not a peculiarity of the US market, but rather a characteristic of broad, diversified indices in general.
