Stock pickers face scrapheap as investors turn to algorithms
Actively managed funds are losing out to computerised investment strategies
The “computers good, humans bad” trend in fund management is only likely to accelerate.
Is old-fashioned stock picking dying? Faith in age-old stock-picking methods is certainly waning, judging by the recent decision of BlackRock – the world’s largest asset manager – to cut back on its active management division and instead place more emphasis on computer-driven models.
“The democratisation of information has made it much harder for active management”, said BlackRock chief executive Larry Fink. Since 2009, only 11 per cent of BlackRock’s actively managed equity funds have outperformed their benchmarks, prompting Fink to “change the ecosystem” by “relying more on big data, artificial intelligence, factors and models”.
The same point was made more bluntly by BlackRock’s Mark Wiseman. “The old way of people sitting in a room picking stocks, thinking they are smarter than the next guy – that does not work any more.”
Such sentiments were once the preserve of advocates of passive investing strategies that track market indices. That they are now being uttered by a firm that operates a $275 billion active stock fund business – that’s noteworthy, and testimony to the pressures afflicting active management.
According to S&P Global data, nine out of 10 US equity funds failed to beat the market over the last year; 94.6 per cent underperformed over the last five years; 87.5 per cent underperformed over the last decade. Global fund figures are similarly uninspiring, with 81.2 per cent underperforming over the last 10 years.
Index funds have an inherent cost advantage. In the US, the average expense ratio for index funds is just 0.11 per cent, compared with 0.84 per cent for actively managed funds. Investors, increasingly cost conscious and disenchanted with active funds, are migrating en masse; according to Citigroup, there has been a global swing of almost a trillion dollars from active to passive funds over the last year.
BlackRock’s move towards a more quantitative, rules-based approach is allowing it to slash the cost of its active funds, and rival fund managers are likely to follow suit
To compete, active managers have to reduce costs and make the case that their strategies are based on hard evidence. This is where computer-based strategies come in.
BlackRock’s move towards a more quantitative, rules-based approach is allowing it to slash the cost of its active funds, and rival fund managers are likely to follow suit. “Why employ hundreds of asset managers, each selecting stocks and implementing investment strategies, when a few programmes can do it for you?” said Sudhir Nanda, a quantitative manager at US money manager T Rowe Price, in an interview with the Financial Times last year. “Humans aren’t going to be completely replaced [in fund management], but they will be mostly replaced.”
The move towards a more quantitative approach is exemplified in the enthusiasm being shown towards smart-beta strategies. Smart beta refers to exchange-traded funds (ETFs) that aim to achieve market-beating returns in a rules-based, index-like manner.
Researchers have long known that various investment styles have proven fruitful over long periods, and there is increased interest as to what specific factors have driven this outperformance. Some of the more well-known factors include value (cheap stocks beat expensive stocks), momentum (buy recent winners, sell recent losers), size (small-cap stocks have historically outperformed big companies), and low risk (less volatile stocks beat more volatile stocks).
Factor-based ETFs are custom-made indices that consist of stocks that satisfy specified criteria. In some cases, managers are constructing multifactor ETFs that combine a number of the above or other investment factors.
Although money is flowing out of traditional active funds, the smart-beta industry is thriving. Globally, more than $500 billion is invested in smart-beta ETFs. BlackRock estimates this will rise to $1 trillion by 2020, and $2.4 trillion by 2025. Continued price pressures and increased competition mean fees for such ETFs continue to fall.
The merits of adopting a rules-based, quantitative approach over a subjective, discretionary one are hard to dispute. A mountain of academic literature indicates that models generally beat humans. One major meta-analysis examined 136 studies across multiple fields – medicine, mental health, education and training, and more – comparing the performance of simple quantitative models to expert human judgment. The models either beat or equalled experts on 94 per cent of occasions; the experts’ few victories were when they had more information than the models.
One of the reasons is that models are dispassionate, whereas human judgment is undermined by all kinds of behavioural foibles such as overconfidence and loss aversion.
These self-defeating tendencies were exemplified last year in a study co-authored by quantitative money manager Victor Haghani. The study involved 61 quantitatively-trained people who were given $25 and asked to place bets for 30 minutes on a rigged coin that will come up heads 60 per cent of the time. Strictly following a rules-based approach would have generated a tenfold profit but most “managed their betting very sub-optimally”, displaying “as many behavioural and cognitive biases as you can shake a stick at”.
According to a report last month by Boston-based financial consultancy Opimas, increased automation is likely to result in 230,000 fewer people working in capital markets by 2025
The financial industry is increasingly conscious of these limitations. Behavioural finance, once a niche field, is now decidedly mainstream. References to the work of psychologist Daniel Kahneman, whose research into the cognitive quirks that bedevil human judgment won him a Nobel Prize in economics in 2002, can be found in every second analyst report these days. Indeed, Wiseman’s crack about “people sitting in a room picking stocks, thinking they are smarter than the next guy” has a decidedly Kahneman-like air about it.
Consequently, the “computers good, humans bad” trend is only likely to accelerate. According to a report last month by Boston-based financial consultancy Opimas, increased automation is likely to result in 230,000 fewer people working in capital markets by 2025.
“The asset-management industry will shrink most, with around 90,000 people being replaced by machines”, it said.
Computers, of course, are also imperfect, as evidenced by occasional flash crashes in financial markets. “Computers are the new dumb money,” said Ritholtz Wealth Management chief executive Josh Brown in 2015, following chaotic market trading on August 24th of that year that pummelled the share prices of some of the world’s biggest companies.
People knew companies like JPMorgan, Facebook and Starbucks should not have fallen up to 22 per cent in the first few minutes of trading that day, said Brown; in contrast, machines “can only do what they’ve been programmed to do. There’s no art, there’s no philosophy and there’s no common sense involved.”
There’s no denying computers can go wrong, and that calm stock pickers can profit when markets occasionally go haywire. Still, investors have clearly lost patience with paying high fees for underperforming stock-picking managers, and there is often a nagging doubt in relation to the handful of outperformers as to whether their success is the product of luck or skill.
In contrast, computer-based investment vehicles are cheap, transparent and offer the promise of a more scientific, data-driven approach.
Old-fashioned stock pickers may believe they still have a vital role to play but BlackRock’s actions indicate the writing is on the wall. “It is hard to deliver value to clients through traditional means of investing,” admitted Wiseman. “The active equity industry needs to change.”