/){kind=link}
At the beginning of the twentieth century, Scott’s Porage Oats were marketed as being produced entirely through an automated process: “untouched by human hand”. Towards the end of the century, a new variety of asset management could lay claim to such a slogan – “quants” or quantitative managers fed mathematical models of financial markets into computers, which were then supposed to select stocks with none of the biases and fallibility of human stock selectors.
Investors were easily convinced of this concept, perhaps seduced by impressive technical descriptions of models, perhaps dubious about the possibility of traditional asset managers beating the market. Large amounts of money poured into quant funds, both long-only and hedge funds.
“Assets managed quantitatively grew phenomenally in the late 1990s and early 2000s,” says Martin Knowles, head of equity manager research at investment consultant Towers Watson.
More recently, quant funds have struggled as their performance, by and large, failed to satisfy investors through the credit crisis and afterwards. Fund research company Morningstar looked at the US quant sector recently, finding a group of 65 quant funds were on average in the bottom quartile of all US equity funds over three years to July 2010. This includes funds from Bridgeway, Goldman Sachs, JPMorgan and Vanguard. Axa Rosenberg is shortly to liquidate four Laudus funds following a controversy over the handling of an error in one of its models. The controversy saw founder Barr Rosenberg and head of research Thomas Mead leave the company.
“Quant has got a bad press because of performance issues,” admits Eric Sorensen, president and chief executive of PanAgora, a US manager with a strong quant tradition. “When a lot of people come in and do the same thing, there are capacity constraints.”
It is difficult to get a precise handle on exact numbers flowing in or out of quant funds because the definition of such products is not clear-cut. Most fundamental managers use some quantitative techniques to filter their universe or provide suggestions, so the line between the two has become blurred.
“My impression is there have been asset outflows,” says Mr Knowles. “In one way, outflows are a good thing, because then you have less money chasing the same factors.” Nevertheless, this advantage will be minimal, he predicts, because the money leaving the sector is not material by comparison with the money that flowed in when quant was in fashion.
Investor attitudes to quant managers are not uniform, even after the struggles of the past few years. “Some clients have an appetite for the quant approach – it’s systematic, it’s easy to explain what the investment story is,” says Mr Knowles. “Some would regard it the opposite way, as a black box.”
Because many quant models identified and attempted to exploit similar factors in the market, such as momentum and value, the more money chasing these factors, the less outperformance any individual fund could achieve. They have also struggled in recent market conditions, explains Mr Knowles, because they frequently do best when the market is taking a clear direction.
“Inflexion points”, at which that direction changes, can be harder for a model to adapt to than a human, he explains. “Over the past year or so, the market’s been oscillating between worry and belief that everything’s going to be okay.
“That’s effectively been a whole series of inflexion points, which is generally not good for momentum strategies,” he says.
This is not to say quant managers are sitting back and waiting for things to go their way again. Instead, many of them are re-examining their models, either because they are concerned they were not correct in the first place, or because they believe markets have changed sufficiently that new models, or at least adjusted versions, are necessary.
“The problem is that finding unique factors becomes ever more difficult as more and more people are looking for them,” says Mr Knowles, “and there’s an increased risk of identifying spurious relationships.”
Although it may be true figures cannot lie, it is very easy for them to mislead, suggesting trends or correlations that do not actually exist.
Other quants have looked at improving their trading techniques to squeeze extra performance from their models, while others are examining questions of timing in switching to eke out returns.
A reconsideration of the role of risk in portfolio management is another avenue for exploration by the mathematically minded. The fundamental indices that re-weight traditional market capitalisation weighted indices according to company metrics such as dividends have been in vogue for some time, but are often criticised for resulting simply in value investment strategies. Now alternative methods of weighting indices are being developed, such as the “anti-benchmark” approach of French quant Tobam, which systematically weights index components based on their correlation to each other to maximise diversification.
Since this approach is intended to minimise risk, as does the “minimum variance” approach, its success should be measured largely by the lower volatility it produces, but it has outperformed in most of the big markets Tobam invests in.
Like many investment innovations, quant managers’ core ideas have become mainstream – now they find themselves at the edge of thinking about what exactly investors want from their portfolios and how that might be produced.