End of the line for quants?

After facing pressure during the 2007 sub-prime crisis, David White reports on the future for the quantative approach to investing

Quants - investment managers that pursue a quantitative approach to equity portfolio management - faced their first real test in the market turmoil last summer.

The consensus is that they failed. Some quantitative equity funds fell more than 30% in a week. Big hitters fared worst with Goldman Sachs losing $1.5bn (€970m) in a week and Morgan Stanley losing $390m (€250m) in a day.

What went wrong? Amir Khandani and Andrew Lo of the Massachusetts Institute of Technology, looked at the TASS data and identified the main problem as the rapid unwinding of quantitative equity market-neutral portfolios.

Quants not only failed, they failed as a group. In the crisis last summer, they showed the herd behaviour - notably the similarities of their fund portfolios - which quant strategies were specifically designed to avoid.

So what lessons have quants learned? More important, will quants regain the trust they lost last summer? Some of the answers to these questions are provided the new Fabozzi-Intertek study commissioned by the Research Foundation of the CFA Institute*

The study is based on survey responses from 31 asset managers, and interviews with 12 asset managers and eight investment consultants and credit rating agencies in the US and Europe.

It attributes the deterioration in performance of many quantitatively managed funds to rising correlations, style rotation, and ‘herding' - the fact that there are now more active quantitative managers using the same data and similar models.

It notes that hedge funds are seen as the main culprits for the heavy losses of quantitative equity funds last summer.

It asks whether, given the recent poor performance of many quantitative strategies, investor demand for the strategies can be expected to hold up. And it poses the question: if quants as a group cannot outperform traditional managers, what is their future in the industry?

The Fabozzi-Intertek study asks participants to rank the factors that they think contributed to the losses incurred by quantitative funds in the July-August 2007 market turmoil. Most blame the sell-off on hedge funds covering their short positions and selling their long positions. Hedge fund managers who were highly leveraged were forced to liquidate assets for which there was no market to answer substantial margin calls from banks that had been sent running for cover by the sub-prime mortgage crisis.

Quants found themselves with positions similar to hedge fund managers who were forced to liquidate their positions. "Everyone is blaming the quants. They should be blaming the leverage," is one heartfelt response to the Fabozzi-Intertek study.

The July-August 2007 period is seen as "a dramatic departure from the normal" with a huge unwinding of positions run on quantitative models. The sums were substantial. It is estimated that there was $1.3trn invested in model-driven long-only (enhanced index and active) strategies and $800bn in hedge funds with high leverage.

The problems arose because quant managers held similar positions. As a result, there was too much money in short positions at a specific period of time, as Goldman Sachs explained to its investors.

Khandani and Lo noted a sharp rise in correlations between 1998 and 2007. The study indicates rising correlations as the most serious challenge to a quantitative approach in equity portfolio management.

Interdependence is also identified as one of the factors responsible for the spread of the contagion from the sub-prime crisis to the equity markets in July and August 2007. The crisis in liquidity - the possibility of finding buyers - began when problems began to affect the equity markets, the study suggests.

Herding - the tendency to copy what everyone else is doing - is also seen as a problem. "Everyone in the quant industry is using the same factors," one participant in the study remarks. "When you need to unwind there is no one to take the trade."

The dominance of the Fama-French model, the three-factor model devised by Eugene Fama and Kenneth French, is one possible cause of herding. This model treats the market index returns, the market cap of the stock and the book-to-price ratio as the only three factors driving returns. "Quants are all children of Fama and French," as one respondent observes.

Quants were under-performing before the sub-prime crisis, the Fabozzi-Intertek study suggests. It looked at the deterioration of the performance of quantitative equity management before the events of 2007 and asked respondents to evaluate performance trends in each market segment. Both asset managers and investment consultants agree that the deterioration of quant performance has been strongest in the US large cap sector.

Among the causes suggested for this deterioration is a greater efficiency in US large cap pricing, and an exposure to those industrial sectors that were worst hit by the sub-prime mortgage crisis. They also suggest that quants are looking for an earnings-quality component that has dispersed over time.

What can be done? The Fabozzi-Intertek study asked participants which strategies quant managers are likely to adopt in an effort to improve performance. Most focus on the need to identify new or unique factors. Quant managers must develop conceptual work that is unique - that is, work that cannot be copied by competitors.

Yet they emphasise that the primary objective of any investment manager is to hold high performing companies without paying too much for them. To that extent, all will be pursuing the same goal.

A key requirement, the survey finds, is better information. Quant managers need an ‘informational edge'. The paradox of quant management is that managers necessarily rely on the same set of data, thereby losing their informational edge. There is general agreement that quant managers need more proprietary information to be able to outperform.

Quantitative equity managers also face an uphill task persuading clients that they are not ‘herding'. Responses to the study suggest that the most significant challenge facing quant equity managers is the problem of too many market participants using similar models and the same data.

The main message from responses to the survey is that quants must update their models continuously if they are to differentiate themselves from the competition.

Fund capacity is also seen as an issue. Last year Goldman Sachs attributed the disappointing returns of its quantitatively managed market neutral funds to the fact that there was too much money in the strategy.

The market turmoil of last summer, and the knock-on effect it had on quantitative equity funds, also placed a question mark over the adequacy of risk management tools. More than two thirds (68%) of participants in the study agree that today's risk models cannot predict severe events such as those of July-August 2007, mainly because they do not take into consideration global systemic risk factors.

One manager comments that "risk management models only work under benign conditions and are useless when needed".

The authors of the Fabozzi-Intertek study question whether risk models should be expected to anticipate such events. They suggest that, by definition and by construction, a risk model measures the size of what cannot be anticipated. If the risks could be anticipated, they would not be risky. Therefore, a risk model should not include such abnormal events.

There is also an argument that risk control actually amplifies risk in situations like the market turmoil of  last summer. Did risk management models make matters worse by forcing deleveraging that was not necessary to meet margin calls?

The answer seems to be no. Most of those taking part in the Fabozzi-Intertek study do not believe that risk models amplify catastrophic events. They also remain broadly optimistic about the ability of quant strategies to build market share in equity management.

63% think that, as more firms introduce quantitative products and as exchange traded funds give the retail investor access to active quant products, quantitatively managed funds will continue to increase their market share against traditionally managed funds.

Yet they expect a slower inflow to quant funds. Quant has clearly suffered some reputational damage, and the Fabozzi-Intertek study asked whether, given the poor performance of some quant funds in 2007, traditional asset management firms that had diversified into quant management would be having second thoughts.

Slightly more than half of the study participants think that it is unlikely to have this effect. A third believes it will, but these were asset managers where quant management represented less than 5% of all equities under management.

So quantitative equity management has been given a qualified bill of health. Provided it moves towards the development of models which incorporate more unique factors and more proprietary information, it could still give traditional equity management a run for its money.

*‘Challenges in Quantitative Equity Management' by Frank Fabozzi, Yale School of Management, Sergio Focardi and Caroline Jonas, The Intertek Group.