Hedge fund manager Samir Arora on why the lessons we have learnt from Warren Buffett are selective and wrong

The founder of Helios Capital favours diversification to capitalise on market serendipity

Samir Arora is nothing short of a rockstar in the Indian stock market. After a super successful stint as a domestic fund manager with Alliance Capital Management, he switched sides to become a hedge fund manager founding Helios Capital Management. At an investor event hosted by Outlook Business in 2010, he took a leaf from Nassim Nicholas Taleb, defending his decision to become a hedge fund manager. “I made the transition to a hedge fund manager so that you won’t have to debate if my performance was luck or skill, as people do in the case of Warren Buffett. As a hedge fund manager, you make hundreds of trades and when you come out tops after so many decisions, you would have a lesser chance of attributing it to luck!” he quipped.

It’s another matter that Buffett has made thousands of trades, over his 64-year stint as an investment legend, which largely goes underappreciated because of the disproportionate attention his long-term holdings have got, aided in good measure by him oversimplifying his own strategy.

Greatly inspired by Taleb’s Fooled by Randomness, Arora believes that stock market investing has a high component of luck, and the only way to succeed is to increase your odds of success every step of the way. A voracious reader, and an independent thinker, Arora says, the one thing that has kept him in good stead is a phrase attributed to French philosopher Rene Descartes: “De omnibus dubitandum,” which means doubt everything.

The other thing that has helped him stay on top of the game, which he does not articulate in as many words, is his strong reflexes that have helped him navigate the market turns well in time or at the very least, before it’s too late. More importantly, those reflexes saved his life as he lifted himself up from an open manhole seconds after he fell into the ditch on the streets of ‘Maximum City’.  

A quintessential Delhi Punjabi, his years of living in Singapore, traveling the world and managing global equities have made an insignificant difference to his partiality for India stocks, just as it has not diminished his love for chaat and mouth-watering sweets from Bengali market in Delhi. Combining track record of Alliance Mutual Fund’s first pure equity fund and Helios Fund into a single stream of return, Arora’s return since 1996 is approximately 25% (219x) while the Indian market is up 12.24% (16x), both in INR. Edited excerpts from conversations with Arora on his strategy: 

You have been saying that diversification is a better way to approach stocks than concentration. Concentration will surely yield better return if your calls turn out well. Please elaborate on your approach.
Let me explain the genesis of the problem. There are many investors who believe concentration is an end in itself. That it works like magic. Such investors implicitly believe that there are only a few companies that do well over each period and therefore diversification dilutes return. Meaning, every additional stock you add to the portfolio will drag down the expected return.

But the very idea that only a few companies do well over any period, both short and long periods, is a wrong starting point. The reality is that in any given period a large number of stocks outperform the index – the construct of the index itself is such that on average half the stocks will do better than index and half will do worse than index over any time period. The number of companies that do better than the index is not exactly half though because it is weighted by market cap, which distorts the movement. Still, a very large number of stocks beat the market over any period. 

To win at the game of investing, one must first understand the field. Our analysis of the field produced very interesting results. We looked at data from the past 15 years. An analysis of the performance of Top 300 stocks by market cap, the addressable set for a fund manager, for the period 2005-2019, showed that for a one-year holding period, the number of stocks that beat the market each year ranged between 93 and 169 (See: Larger the better), depending on whether the market breadth was narrow or wide. For the entire period of the study, the average number of stocks that did better than the market was 131 for a one-year holding period.

Now, these large number of outperformers broadly hold even for longer periods. Looking at the number of stocks that outperformed the market over two-year holding periods yielded a range of 116 to 161. Even over a 15-year period between 2005 and 2019, we found 105 companies outperforming the index, which is a third of the stocks. That’s a fairly large universe of outperforming stocks to play with.

What is the chance that you will end up picking the right stocks from this selection? Warren Buffett says investing is about eliminating risk, and risk is not volatility or standard deviation, which modern portfolio theory captures. When you pick large numbers of stocks, you may not be able to eliminate risk to a great degree, you eliminate volatility, not risk in the real sense.
You can never eliminate risk anyway. Your calls can go wrong. Buffett’s calls have gone wrong as well. But, there is a more interesting takeaway from our study: it is the number of companies that do really well, as against just outperforming the market over each period. For this, we considered the Top 300 companies by market cap at the beginning of each review period and calculated the performance of the 30th best stock, 60th best stock, 90th best stock, 120th best stock over each holding period – one year, two years, three years up to 15 years. Simply put, the return was calculated assuming you had perfect foresight in predicting the stocks that ranked 30th, 60th, 90th and 120th

While the index returned annualised 13.62% during the period, for a holding period of one year, on average, even the 90th stock did doubly better than the index with a 28.80% return. The average return for a two-year holding period stood at 23.46% return, somewhat lower than a one-year period but far higher than the index.

So while a large number of stocks display serious outperformance over each time period, the number of outperformers as well as the extent of outperformance goes down as the holding period increases (See: The winner’s game). For the 15-year period, i.e., if you picked up stocks on January 1, 2005 and held on to them until December 31, 2019 and your portfolio matched the 90th out of the 300 stock selection, you would end up with 15.23% return, still higher than the index. One reason for this decline in the number of winners is that many new stocks that got listed after 2005 (and hence not in the original selection) may have done better and been added to the index over the 15-year period.

The truth is a significant number of companies do really well every year; it’s not easy to have high confidence in them in the beginning. Typically, when you do have high confidence in a stock, more often than not, it is already ‘discovered’, which means it falls in the high-confidence, reasonable-return category. Hence, the stronger outperformance in the portfolio is unlikely to come from this category.

To put it differently, in the initial stages, when a trend is emerging or a stock is emerging, you cannot have high confidence, but you know in any given year, a third will do well. So, think of this as buying lottery tickets. The more you buy, the better the odds of winning. In any case, all you need to evaluate is if the stock will end up within the top third. Obviously, you better the odds further by doing your research well.

Let’s face it, there are some gospel truths in the stock market, like earnings performance will always be rewarded and appropriately valued in the long run. But when, what, how much are all unknown. What we have to do is reduce the odds.

Sometimes, nothing may have changed, for example, nothing has changed remarkably in the pharma sector this year, except its relative attractiveness. Other sectors have either been affected by COVID-19 or are valued out of range, so investors have latched on to pharma over the past few months. I don’t really like pharma — it does not excite me as a business. But we still have bought into a few pharma stocks because it won’t harm us. Best case, they will be top performers, worst case they might still end in the top one-third, beating the index this year.

One could argue that you put more research and try to improve your odds even more, and you have a better margin of outperformance. What’s wrong with that?
A strategy to buy stocks in only 10-odd companies will necessarily miss many stocks where prospects and current performance may look attractive but still confidence and conviction to place a 10% type bet is missing. That is why, placing smaller bets on large number of carefully selected names may make the portfolio diversified, but in essence, allows market serendipity to work in your favour. We have had many multi-baggers in our funds over the years but there is no way we thought these stocks will do as well as they did, when we first bought them.

Investors favouring concentration compound their error of concentration risk by simultaneously committing to very long-term holding periods. It’s a double-edged sword. Once you signal to the outside world that you have found this great stock where you have committed a large part of your portfolio, exit won’t be easy.

Look at Buffett. The performance of Coca Cola and American Express, two of his most favoured holdings and widely acknowledged as his best bets, is instructive: Since December 31, 1988, Coca Cola stock has given an annualised return of 11.85% per annum including dividends and American Express has given return of 10.59% per annum. Over the same period, S&P 500 is up 10.46% per annum. By holding both these stocks for so long, Buffett’s margin of outperformance has reduced significantly. On the contrary, if you compare the performance of Berkshire Hathaway versus some of these stocks over the period of these holdings, Berkshire has outperformed, which again goes to prove the excess return came from other diversified bets, not the long-term holdings.  

That does not prove the whole strategy wrong. The fact that Buffett did not time some of these exits well makes these bets look less impressive.
That’s exactly the problem. When you concentrate and you are a sizable investor, taking a ‘sell’ call is not easy. A concentrated holding signals strong commitment and endorsement of the stock. So when you hold an outsized position, you find it psychologically more difficult to walk away if things don’t work out as well. Along with your funds, you have your ego invested in the stock.

This commitment bias, for example, may have come in the way of Buffett properly analysing the PepsiCo opportunity. Not only has PepsiCo outperformed Coca Cola since 1988 clocking 12.73% per annum, it has, in fact, significantly outperformed for the past 20 years. Since December 31, 1999, Coca Cola stock is up 5.15% per annum while S&P 500 is up 6.01% per annum and PepsiCo stock is up 9.4% per annum, including dividends. If not switching to PepsiCo, even diversifying into PepsiCo would have helped Buffett.

There is another misgiving about this whole idea of making large, long-term bets – the illusion that long-term winners can be found with confidence, 10 and 20 years ahead of time. Reality is that longer-term winners normally surprise even themselves, their managements and their investors with their growth/success and can therefore not be generally identified well in advance with high degree of confidence.

Steve Jobs had more than 80% of his wealth in Walt Disney (which he got for selling Pixar to Disney) and not Apple when he died; Bill Gates has not been the biggest individual investor in Microsoft for a long time. Even Jobs and Gates did not know that their companies would become so big and successful and how the future would pan out. And we have investors (in many cases, retail investors) confidently announcing how they have found stocks that they will hold for the next 10 to 12 years.

The market is replete with examples that demonstrate this. Even in India, Sunil Bharti Mittal has gone on record about how his forecast for the telecom sector in the early years was way lower than how it turned out. For a good part, Vodafone was seen as the best managed company in telecom, and the joint venture between Birla, AT&T and Tata, which later became Idea was the best horse to bet on. No one could have dreamt of who would rule telecom in India finally.

Even better example is Eicher Motors, which most fund managers bought into in 2009 as a cash bargain and a bet on its partnership with Volvo for trucks. The Royal Enfield opportunity was nowhere on the radar but ultimately it was that obscure motorcycle that made it a multi-bagger stock. There are tonnes of other examples.

What's the right number when you diversify? Is it a number you derive by limiting the position size?
Yes, absolutely. Since the total has to be 100%, position size will influence number of positions. We feel a portfolio should have two kinds of stocks — “High confidence in reasonable return" and "Reasonable confidence in high return".

Unlike many of the younger and less-experienced fund managers, we do not have a category of "High confidence in high return" and therefore limit the size of the initial position to around 7.5% for the highest weight.

If your fund size were 10X or 100x, would this number still work?
In that case, the number of stocks would increase to 50-60. I don't think that would reduce return in any way as long as one could get a large part of the portfolio in the top one-third. Remember that Peter Lynch used to have hundreds of positions and still massively outperformed the index over his tenure.

I also have data to show that Berkshire's return has been negatively hurt by Buffett’s top positions in the last 20 odd years and his performance has been helped by the tail rather than by the top positions. Even this year, where his top position (Apple) has done so well, losses in his other big positions (JP Morgan, Kraft, Wells Fargo, Bank of America and Airline stocks) have neutralised the gain from Apple. So overall, the concentrated portfolio has not done as well as the market.

What will be the effect of additional churn on the overall return, say over 15 years, which is your maximum period?
Assuming that the portfolio will churn in two vis-a-vis every five years for the long term (buy and hold) investors, the tax difference will approximate to around 0.5% per annum with current Indian tax rates. In practice, the churn for the whole portfolio is more like three years. We have seen that this can be easily compensated by better pre-tax return of a diversified portfolio. 

Are there any rules you follow for diversification? How do you pick your outperforming set?
It is important and easy to differentiate between bad and good investments, but it is not so easy (or even useful) to try and differentiate between two good investments only for the sake of concentrating on one. It is in this context that diversification is really useful for no one knows anything about anything beyond a point, and knowing that a large number of companies do well in each period, there is no need to make a forced choice between two good stocks.

Our strategy has been to build a portfolio of about 30 to 35 stocks bought with a 1- to 3-year initial view. Sweet spot for an investing horizon is one to three years for one can visualise most factors such as industry trends, disruption, company strengths, management strategy, government policies, market preferences, external environment and so on more easily over this horizon. 

If a company continues to do well, there is absolutely no reason to not hold it for another one to three-year horizon and so on. I have held HDFC Bank for the past two decades. I have stayed with the stock for long, but I still do not call myself a ‘long-term investor’ in the stock. It has delivered consistently, and we have held on. But if you think of yourself as a long-term investor, on the contrary, you are committed to holding on to the stock even during periods of long-term underperformance, like Buffett has done in the case of Coca Cola. That is not rewarding.

Is there any aspect of Buffett’s style that appeals to you?
There is an irony to what is propagated as Buffett’s philosophy. While the idea of buying and holding fewer bets has come to be accepted as the cornerstone of Buffett-style investing, Buffett’s own portfolio has displayed far higher level of churn than one would expect. According to a research paper by John Hughes, Jing Liu and Mingshan Zhang, Berkshire Hathaway’s quarterly filings from 1980 to 2006 comprising 2,140 quarterly stock observations found Buffett’s median holding period to be a year, with approximately 20% of stocks held for more than two years. Approximately 30% of stocks were sold within six months. It’s trial and error, at the end of the day, but as I said, you have to do it intelligently. Buffett has done all kinds of trades, from leveraged trades to arbitrages and derivatives and that is how he has been as successful as he is. Unfortunately, the lessons we have drawn from him are selective and wrong.

What's you view on the market right now? Is the current rally sustainable? What pockets are you bullish on?
We normally invest in three big picture themes: financials, private sector; consumer, broadly defined; and exporters – technology, pharmaceuticals and specialty chemicals. In recent months, we have added to the new emerging technology behemoth and two companies expecting to benefit from China +1 strategy.