by James Phillipps on Aug 23, 2010 at 10:34
Quantitative investment strategies have come under heavy fire over the past three years and a number of investors are now heralding the death of the black box.
A number of funds from a broad swathe of asset managers fell foul of the financial crisis and then missed the subsequent rally, compounding the disappointment for many investors.
Man Group’s AHL Diversity fund, the one-time pin-up of the sector, was a case in point in 2009 when it returned -16.9%, although its performance has since picked up this year.
Several of the leading quant fund managers insist they have responded to the events of the downturn by moving to improve their transparency and strategies. In some cases, they have also integrated a degree of discretion into their investment processes.
‘Quant investors have upgraded their models, but they need to rebuild trust in the robustness of their systems,’ said Gurvinder Brar, head of European quantitative research at Macquarie Securities. ‘You could say that we are seeing the death of the black box and the dawn of the glass box.’
He stresses that quant screening processes came under huge pressure during the financial crisis as the market underwent several major inflection points, caused mainly by irrational and, at times, unexpected investor behaviour.
‘A number of quant funds have not performed because the breakdown of traditional correlations. Distressed sales and panic selling tested models to the extreme,’ Brar said. ‘A lot of investors have been disappointed. This has seen the traditional black box method come under scrutiny because they want to know what inputs are being considered. This was hitherto considered proprietary information.’
Ian Heslop, head of quantitative strategies at Old Mutual Asset Managers, agrees that fund managers have had to adapt to the ‘glass box’ model, with innovation crucial to the success of the strategy in the future.
‘Moving forward, the key criteria for the selection of quant managers should be those who have learned from the past, those whose models have evolved accordingly and those whose models stem from clear market insights, rather than just the mining of historical data,’ he said.
But just how new is this, and how helpful is it to investors?
The goalposts that quant managers use to determine how they run money have been moving about since the early godfathers of the sector, such as Ed Thorp and Jim Simons, first rose to prominence.
Heslop believes quant systems are always being refined and now factor in research beyond the traditional value and momentum formulations, sensitivity to macro conditions and cyclicality to build in greater downside protection.
The fact that fund groups are more willing to disclose their inputs has not really undermined their proprietary offering because different groups will place greater emphasis on different factors. No doubt some managers will be more willing than others to disclose how exactly these are used. This, though, only serves to ask the question of how useful it is to know this information anyway, if it is unclear how or to what extent the assorted inputs are being relied on.
Macquarie, for one, has increased the number of inputs it uses in its calculations.
‘We have advised clients to evolve their models by broadening the number of inputs and be more transparent to their client base about their investment process,’ Brar said.
‘With some groups, this will no doubt be a slow process. Some funds can have 300% turnover a month, which is obviously very hard for investors to keep track of.
‘But the outlook for quant funds is positive, with differentiated investment processes, better risk controls and new product ideas.’
No quant system can ever be perfect, though, so the funds can always expect to have their critics. Brar admits that the events of the financial crisis were so extreme that even back-testing using data from the last three decades would not have picked up so savage a pullback.
Therein lies the problem perhaps, and it serves to emphasise why quant funds should be viewed as one of a basket of strategies, rather than something relied on in isolation.
Professor John Mason, the professor of Penn State University’s management division, is more critical of the strategy, saying that quant models will always be just that: hypotheses and therefore ‘incomplete and fallible’.
‘Why is this true?’ he asked. ‘Because we, as human beings, never have complete information when building a model; we always have to work with incomplete information. This is what creates uncertainty – we don’t know what the outcomes of our decisions will be – our models are imperfect.
‘And this is the reason why quants won’t go away. The human quest is to continually build models that help people make better decisions or solve more difficult problems. It has been shown over the past 40 years that quant models can help generate millions or billions of dollars in earnings. The models were not perfect [but] quants, as we speak, are working to improve their models. The process will go on.’
Heslop is understandably less cynical, believing that the strategy has been pared back to oversold levels. He expects asset flows to increase over the next two to three years as investors regain confidence in quant funds.
If this does transpire, though, Mason warns that there is always the risk that trades will become overcrowded and that because of this, some funds will take on more leverage to improve returns.
Whatever the outcome, it looks certain that investors interested in quant funds will have to focus more on due diligence and understanding a manager’s strategy more thoroughly than has happened hitherto. As Heslop notes, wise investors will back those managers that have learned from the past.