4th June, 2017

Why a robot will never pick the superstocks of the future

Experience has taught me that what matters in business is the big winners. You can suffer losers and still do well if you have just a few significant successes. I assumed that this type of return distribution — a very small proportion of one’s overall investments making all the difference — was characteristic of early-stage, venture-capital-style funding. In that sense, it mimicked the entertainment industry, where the bestsellers and smash hits are what count.

But a recent article in The Wall Street Journal shows that the same principle applies to long-term investment in the American stock market. Between 1926 and 2015, just 30 shares accounted for a remarkable one-third of the cumulative wealth generated by the whole market — from a total of 25,782 companies listed during that period.

Astonishingly, more than half of all shares lost money during their time as public companies, the WSJ reports. Not much wealth generated there. And less than 1.1% of the shares that were publicly traded created 75% of the market’s entire gains.

These statistics demonstrate that “superstocks” are what produce the true profits in the long run. They are the quoted equivalents to my personal big winners, which were private companies when I invested in them.

The research that reveals where gains have actually been made over the past 90 years, carried out by Hendrik Bessembinder of Arizona State University, appears to undermine the widely held belief in diversification and modern portfolio theory among the investment community.

It also calls into question the cult of equity, which has been followed by professional investors for more than 50 years. The experts argue that shares decisively outperform bonds and cash over time.

But Bessembinder’s research shows that the returns from 96% of American shares would have been matched by fixed-interest instruments, which generally offer more security and liquidity, and suffer from lower volatility than stocks.

Of course, getting stock selection right is very difficult indeed when such a tiny proportion of shares contributes so much to total performance. It requires investors who are truly patient and at times extremely brave.

Amazon is one of the heavy hitters that delivered a quarter of all wealth creation in the stock market during the 90 years to 2015. Yet between 1999 and 2001, the online retailer’s shares fell by 95%. Many investors probably gave up then, and having been burnt once, shunned its 650-fold appreciation over the past 16 years.

Stock picking is somewhat discredited these days, because low-cost passive fund managers argue that their tracker model delivers better value to savers by betting on an index, not individual companies.

While empirically that may appear to be correct, intuitively it feels questionable to me. If a large proportion of money is purely index driven, it distorts the market such that index components are certain to outperform — even if the underlying companies are deteriorating. Moreover, it directs savings away from tomorrow’s winners towards incumbents, disadvantaging the capital-raising efforts of industries of the future.

Economies grow thanks to new technologies and entrepreneurs, who run a fairly small number of outstanding companies funded through private capital. Half the top 20 wealth creators referred to above are in sectors such as pharmaceuticals and computers. Identifying those sorts of promising industries is not too hard. But I do not believe there is a computer program — or robotic system — that can pinpoint the great achievers of the next 10 or 20 years.

Choosing the special businesses and executives that will create enormous value, and probably large numbers of jobs, is as much a creative undertaking as a scientific one.

Rigorous analysis must include a host of variables that artificial intelligence would struggle to understand — adaptability, trust, motivation, ruthlessness and so forth. I suspect all the best investors emphasise the importance of judging management when backing companies; I am not confident that computers can do that better than humans.

Spotting a business that can grow 10 or 20-fold over a period of years is a rare art. In mature economies such as the UK, such sustained compound growth happens all too rarely.

To achieve it, a business should enjoy high returns on capital, strong cash generation, plentiful long-term expansion opportunities and a powerful franchise. And you need to buy the company at a sensible valuation. In a world awash with cash, such attractive businesses command very high prices. But if you believe the model can endure, they might be worth it.