Quantitative models assist in dealing with the increase in available financial information through their great capacity for systematic analysis, which facilitates the identification of investment opportunities. The quantitative approach also presents the advantage of avoiding emotional reaction in the selection of stocks so as to act in the most objective manner possible.
Two years ago, Dexia Asset Management launched its first equity funds managed on a purely quantitative basis: Dexia Quant Equities Europe and Dexia Quant Equities USA, sub-funds of the UCITS under Luxembourg Law Dexia Quant. These funds represent the outcome of many years of internal research into quantitative models to support decision-making.
In fact, within each sector, the various models select the most attractive equities based on a unique combination of factors, specific constraints in relation to risk and to turnover. Recommendations from the models are then directly implemented in portfolios through an IT optimisation process.

History of quantitative research at Dexia Asset Management
Dexia Asset Management established its quantitative analysis team in 1998. From that time, a first version of the QuaRaM (Quantitative Ranking Model) model was created and integrated into the traditional equity management process in order to assist managers and analysts in their market analysis work.
The first quantitative tool developed by this team covered the European investment universe, and some months later an American model was put in place. This first generation consisted of a single model analysing all the stocks in the investment universe, all sectors combined. Continuous improvement work by the quantitative analysis team resulted in the creation of a second generation consisting of different independent models
specific to each sector analysed.
The results obtained by these models were very convincing and contributed considerable added value to the equity investment process. The quantitative research team further increased its efforts so as to strengthen them even more. It was thus that, from 2002, the models proved to be sufficiently powerful and reliable to operate independently of a classic analysis process.
In fact, the models having proved their worth in classic equity management, the idea arose of diversifying the range of products by the launch of funds managed in a purely quantitative manner.

The development of quantitative models
In order to build a reliable decision-making tool, Dexia Asset Management’s quantitative research team developed a different model for each of the ten sectors comprising the American and European universes. This approach enabled account to be taken of specific sector features, whilst ensuring diversification of explanatory factors and modelling techniques. Each model is fed by some 40 alpha factors likely to generate an out-performance. So as to guarantee the reliability and competitiveness of each model, they are subject to constant supervision as well as regular analyses.
The models are specifically developed so as to generate stable and coherent performance over the long term. Through the analysis of different factors, they carry out a daily screening of all the stocks included in the investment universe (MSCI Europe and MSCI US). In this way, managers have up-to-date information on every single stock in their investment universe.

Simulations and real performances
It should be noted here that, before the launch of the two funds Dexia Quant Equities Europe and Dexia Quant Equities USA, the underlying models were subject to numerous simulations (back testing) with the aim of verifying the pertinence of their recommendations in relation to different market regimes. All of this research work and the tests were entirely carried out by the Dexia Asset Management quantitative research team.
Back testing simulates the implementation of a particular investment strategy according to predefined market environments. So the results from all the models were analysed in various simulations so as to verify their capacity to generate consistent performances. These simulations were carried out over a sufficiently long period to harness extremely varied market conditions (bear, bull markets and so on). The parameters used for this back testing were the same as those used in the daily management of quantitative funds (e.g. degree of tracking error, maximum turnover, etc).
These simulations demonstrated that the models developed by our teams were reliable in the long term whatever the market conditions.
The tables and charts presented above show that the real performances of the European funds1 are as good as those achieved in the simulations. The same applies to the quantitative management applied to the American universe2.

Development prospects
Dexia Asset Management’s quantitative management has demonstrated its capacity to provide a significant excess return. The management and quantitative research team works on a daily basis to update these models so that they constantly provide optimum results. Similarly, the team regularly analyses the opportunities to develop new and better-performing models or those oriented towards other investment universes. Indeed, a model for the Japanese investment universe is currently in the final stages of development.
As demonstrated by the investment process and performances of Dexia Asset Management’s quantitative funds, quantitative management offers opportunities to diversify an equity portfolio.
More fundamental traditional management and quantitative management in fact present complementary approaches, ideal to the establishment of a balanced equity portfolio capable of generating a satisfactory excess return whatever the market conditions are.
1/2 Performances net of charges based on the institutional class (management fees 0.35%)
In collaboration with
Dexia Asset Management
Contact:
Stefaan Coosemans
stefaan.coosemans@dexia-am.com
Ben Peeters
ben.peeters@dexia-am.com
Dexia Asset Management
Nederlands bijkantoor
Lichtenauerlaan 102 – 120
3062 ME Rotterdam
Gratis nummer: +00 800 589 654 95
www.dexia-am.com

Topics