Time travelling with equities

Many of us have been advised by others (often close family and friends) to stay away from stock markets and not gamble our money away. Mostly, such advice is given in good faith, based on personal experience of such family and friends. This article aims to explore whether this is really the case - Is the stock market really a casino in disguise?

To answer this question, I am going to rely upon data - the daily closing S&P BSE Sensex values from 1986 all the way up to 2021, and some analyses on this data to understand how the market has behaved in the past:

Analysis 1: Probability of losing money across different investment horizons

Basically, this is an analysis of rolling returns over different investment timeframes. For instance, when analyzing 5-year returns, we effectively look at all 5-year periods between 1986 and 2021 (7,300+ data points) without restricting ourselves to financial year end or calendar year end returns. Hence trying to be more comprehensive and avoiding any timing bias in the analysis.

As would be expected, on a 1-day return there is a ~47% chance of losing money. On any given day, the Sensex would either go up or down, with a marginal skew to positive returns given that the general trend of the Sensex has been upwards (as is that of most other large global indices). Similarly, on a 1-year return - the probability of losing money (defined as the number of times the 1-year return is negative over the total number of 1-year returns) is 32%. Now, here is where it gets interesting - at 10 years, this probability of loss is 1%. And over any 15 year (or higher) timeframe, irrespective of how unlucky you are or how bad a market timer you are, you would have not lost money in the Sensex.

"In the short run, the market is a voting machine but in the long run it is a weighing machine." - Ben Graham

Analysis 2: Return ranges across different investing horizons

This essentially is an extension of Analysis 1, where we are looking at the probability of annualized returns across ranges in steps of 10% (and the sum of the probabilities below 0% for any investing horizon would give us the same result as Analysis 1). Additionally, apart from looking at probabilities, we also look at the average annualized returns for each of these investment horizons and the range of returns achieved. The purpose of analyzing it this way to to get some insights as to what to expect when we invest in the Sensex across these horizons.

Following are some of the takeaways:

If we invest for a period of 1 year, we should be prepared for a range of outcomes. It is equally probable to lose more than 20%, or to lose between 10% and 20% of our capital invested (refer to the table above, ~10% probability each). And it is almost equally likely to make 0% to 10% return or a 10% to 20% return (15% and 18% probability respectively). Further, while on average the Sensex has returned ~17%, it has also yielded both, as high as 250%, and as low as -50% at some points in a 1 year timeframe. So, expecting an average return of 10%-12% in a 1 year timeframe is just going to lead to disappointment.

Now consider the 5 year timeframe - things start to skew in our favor. There is a 64% chance of making at least a 20% return (36% + 28%), while only a 9% chance of losing money. The range of outcomes has reduced as compared to 1 year, with about 54% annualized returns on the upside (implying an 8.5x return in 5 years), and about -8% on the downside (reducing the investment to 0.65x). There is no instance when one has lost over 10% annualized in any 5 year investing period.

Moving on to a 10 year timeframe, the odds are even more in your favor, the average return is ~13%, and there is a 1% chance of losing money as against 66% chance of making over 10% annualized returns. And as we progress with longer timelines, the range gets narrower, the probabilities of losing money become ZERO, and we are very likely to make annualized returns between 10% and 20%.

I would like to specifically draw your attention to the 30 year timeframe. If one had invested for any 30 year duration between 1986 and 1991 (until 2016 to 2021) on any of the 1,300+ days when the markets were open, she/he would have earned somewhere between 11% (25x over 30 years) and 16% (85x). And this is without considering dividends which would have averaged at around 1.5% over and above this throughout the 30 year period.

Inflation in India has averaged at ~6.2% over the last 30 years, and the Sensex has given at least 5% excess return over the same period (and by ~6.5% if dividends are considered). So in Real terms, by investing in the Sensex over 30 years< you would be wealthier by 18x to 78x. Not bad, if you ask me!

On comparison, investments in Fixed Deposits with banks have given a return below inflation after considering the impact of taxes (which are a big drag in returns, especially in the higher tax brackets).

Important points to note: While the entire analysis is based on historical Sensex data, please do realise that it is just that - 1) Historical and 2) Sensex data. We can never be certain that these will be the odds or the ranges of return which we will encounter when we invest in the future (in fact, we can be certain that they will be different). Nor is the Sensex representative of all types of stock investing in the Indian or Global markets. It is just a basket of stocks of the 30 largest companies by market capitalisation in India and has been used as a proxy for illustration. The data set is of 35 years only, and gleaning insights about 30 years is not without risks.

Hence, we conduct a similar analysis of the monthly rolling returns of S&P 500 in the US market, which has a much larger history (officially started in 1926, though with a small number of constituents which grew to 500 somewhere in 1957). Here is what it looks like:

Data Source: multpl.com

As you will notice, the overall picture is not very different even in the US market. The 1 year returns are pretty much as random as the 1 year Sensex returns. And as the time horizon stretches, the variability comes down, with the probability of making a loss diminishing to zero eventually, again like the Sensex.

Do note that this is the price index, and does not consider the dividends yield, which has averaged around 3.75% during this timeframe considered in the data, though has fallen to ~1.8% over the last 20 years - given the larger dividend yield, a total return analysis would have given a more favorable return profile, with almost zero probability of making a loss even in a 20 year time frame.

The timeframe we are talking about here includes a number of boom and bust phases of market cycles, with the notable down periods including - The Great Depression (1929), The Dot-Com Bust (2000), The Great Recession (2008) and the Covid-19 Pandemic (2020), along with the corresponding boom periods. So, in that respect, this would be much more comprehensive as compared to our Sensex data.

Can we conclude that you would never make losses in a large enough timeframe of ~30 years? Unfortunately not. There is at least one black swan which negates that conclusion - Japan.

No analysis is complete without searching for disconfirming evidence, to overcome our confirmation bias!

A simple glance at the chart above, without any sophisticated analysis, would tell is that anyone who invested in the Nikkei in 1989, hasn't yet recovered his/her money 32 years later in 2021. Further, the Lost Decades of Japan ensured that an investment made in 1995 also wouldn't have made a profit 20 years later in 2021.

So, what's the point of all of this? And isn't it ridiculous to talk about 30 year timeframes?

Here is my take on it:

• Expecting an average rate of return in 1 year is like asking a blindfolded monkey to throw a dart and expecting it to land near the bullseye.

• Your horizon for expecting any reasonable returns from equity investing cannot be less than 5 years

• There is an overwhelmingly high chance of making a profit (and likely, a reasonable quantum of it) by staying invested for long periods of time.

• The next time your friends and relatives tell you stories about how they lost money in equities, do as them what their timeframe was, and whether they would have been better off by staying invested (also, how did they go about selecting stocks).

• Equities are probably the best protection against inflation over the long run. The most simplistic form of investing in an index has yielded phenomenal real returns over the long run.

• If you are 30, and plan to save for retirement, your timeframe is 30+ years; of you are 50 and plan to save for retirement, your timeframe is still 30+ years!

• While the takeaway is simple - invest for long timeframes - it is not easy. Every 30 year timeframe has 30 1-year timeframes embedded, with many of those 1 year periods yielding huge losses; people telling you to get out of the markets repeatedly along the way; your emotions creating an urge to sell when you panic; and the list could go on. Endurance is much more important than intellectual prowess or analytical rigor!

• Lastly, you can never be certain about making great returns - as was the case in Japan, so have a plan-B. A couple of simple ways of mitigating this risk is to diversify into other asset classes, and invest amounts across time (dollar cost averaging or SIP).

"History doesn't repeat itself, but it often rhymes" - Mark Twain

So, are the stock markets comparable to casinos? I'll leave that for you to decide. 🙂

I hope you liked this article, and were able to take some insights home from this analysis. Do let me know your thoughts on this topic through the comments, I look forward to read them. If you liked this post, please share it with your friends and family, and do subscribe to get access to these articles directly in your inbox as soon as they are posted.

PS: In case you would like to have a look at the Sensex data I have downloaded for this analysis, and look at / modify the analysis on excel - you can download it by clicking the link below.

Sensex_Historical_Data.xlsx

Download XLSX • 2.20MB