Mean Reversion: Universal Truth Or Human Delusion?

The basic definition of ‘mean reversion’ is that asset prices and historical returns will revert to the long-run mean or average level of the entire dataset. The concept hopes to identify abnormal activity that will revert to a normal pattern in the future. In other words, ‘mean reversion’ suggests that cumulative observations matter more than the most recent one.

Many of you would have come across the following statements related to ‘mean reversion’ while investing in equities:

• Booking profits (versus booking losses): “Since I have to sell part of my equity portfolio, let me sell stock A where I am making a profit rather than selling Stock B where I will have to book a loss.”
• Investing at 52-week high!: “I want to add two stocks to my portfolio—Stock A and Stock B. Stock A is trading close to its 52-week high while Stock B is trading close to its 52-week low. Hence, I should buy Stock B now and I should wait for Stock A to fall before I decide to buy it.”
• Investing at rich P/E multiples: “Current P/E multiple of Stock A is 30% higher than its average P/E multiple for the last 10 years. Hence, Stock A is expensive.”

Often, the concept of mean reversion fails to hold true for these three statements. Why?

For example, if the city of Mumbai receives rainfall on a day in the month of January, we do not see shoppers on the high-street buying several umbrellas and raincoats. This is because the concept of mean reversion holds true and shoppers do not expect unseasonal rainfall to continue over the next few days. Our cumulative observations over many years matter more than the most recent observation.

Here is another example. A father has two children—a son and a daughter. The daughter always ranks amongst the top 10% of all students in her school (like a stock trading at its 52-week high). The son always ranks amongst the bottom 10% of all students in his school (like a stock’s 52-week low). The mean (average) performance between the two children is around the fifth decile. While planning for his children’s future, this father does not expect ‘mean reversion’ to result in his daughter’s performance to drop down to the fifth decile in the future. Why? Because the father understands the factors that have driven the daughter’s strong performance and hence if these driving factors sustain into the future, then her strong performance will also continue in the future. In other words, the father understands that the real ‘mean’ applicable for his daughter is the historical track record of top decile performance, and not the fifth decile which is the average of the two children.

For instance, in the chart below, an incorrect understanding of the underlying fundamentals (i.e. the ‘mean’) leads to incorrect decision-making by the investor.

This error can only be avoided by a correct understanding of ‘mean’—as shown in the chart below.

Let us now go back to the three examples from equities investing we had highlighted above:

• Booking profits (versus booking losses): In most cases, this decision should be based on the fundamental strengths of underlying stocks. It is prudent to exit from stocks with weak fundamentals no matter whether that implies booking profits or losses for an investor.
• Investing at 52-week high: Companies that consistently compound their earnings at a healthy rate will, more often than not, trade at prices close to their ‘52-week highs’. ‘Mean reversion using 52-week highs and lows’ has no fundamental logic. So, for example, selling Nestle India Ltd. because it is trading at close to its 52-week high does not make any sense (unless you believe infants will stop drinking Nestle’s formula and/or you believe Nestle is about the lose its hammerlock on the infant formula market). Similarly, buying Larsen & Toubro Ltd. just because it is trading close to its 52-week low makes no sense unless you have researched the company and convinced yourself that L&T has built a strong franchise in, say, civil construction and engineering.
• Investing at rich P/E multiples: As highlighted on Jan. 4, 2020, P/E multiples higher than 30x can be cheap and P/E multiples lower than 10x can be expensive. An investor looking to buy and hold stocks over a long time period, needs to be able to differentiate between businesses which can deliver longevity of consistently healthy free cashflows and those that run a high degree of uncertainty in their fundamental prospects.

Investment Implications

Although the concept of mean reversion sounds obvious, clarity of thought around the definition of ‘mean’ is critical, especially in times when the external environment, reported earnings, and share price movements are volatile. In our Consistent Compounders Portfolio, we use the concept of mean reversion to periodically re-balance client portfolios. Our view of the underlying fundamentals of stocks (i.e. the ‘mean’) defines our model portfolio weights and if relative share price movements cause substantial deviations in a client’s stock allocations versus the model portfolio weights, we revert these allocations to the ‘mean’. For instance, over the last two months, we re-balanced most of our existing clients’ portfolios by buying more of stocks like Bajaj Finance whose share price had fallen by almost 60% between Feb 2020 and May 2020. By purchasing more Bajaj Finance over the past couple of months, we retained our model portfolio weight in this stock.

Note: Nestle and Bajaj Finance are part of most Marcellus’ portfolios.

Saurabh Mukherjea, Rakshit Ranjan, and Deven Kulkarni are CIO, Fund Manager, and Analyst respectively at Marcellus Investment Managers.

The views expressed here are those of the authors and do not necessarily represent the views of BloombergQuint or its editorial team.