A quantitative model can cover a far larger number of shares than a traditional analyst and can do so more quickly and accurately.
Internal consistency
Shares can be compared within a rigorous and homogeneous framework which may not be humanly possible - especially if a team of fundamental analysts is involved.
A Diversified Source of Alpha
To enhance reliability, we reduce the dependence on any particular factor by diversifying across more than 15 factors at any one time in each country-specific model.
Transparency
With a well constructed quant model the reasons for a forecast can be precisely attributed. This contrasts with the “warm fuzzy feeling” that may explain a human’s recommendation.
Objectivity
A good quantitative approach exploits behavioural biases rather than being subject to them.
Repeatability
The effectiveness of the model over a particular period of time is not subject to the “mood” or time-varying enthusiasm of the fund manager.
A unique alpha diversifier
Given that traditional fundamental analysis is the most prominent investment philosophy applied and that our proprietary models are unique - they represent a different source of alpha from the norm. Furthermore, due to the style-adaptive native of the investment approach adopted it can perform during those periods when a particular fundamental style is out of favour.
(b) Competitive advantage over other quantitative models:
Research depth and quality
A substantial, award-winning and ongoing body of robust quantitative research, aspects of which have been published in peer-reviewed academic journals. We also have a close strategic relationship with FutureAlpha.com who is an international specialist in the construction of quantitative stock selection models.
Track record
A compelling ‘real money’ track record of sustained outperformance since January 2000. It is admitted that some other models may also have good prima facie track records but it is emphasized that ours is not merely one of the ‘back testing’ variety.
An advanced algorithm that allows for time-varying rewards to the factors concerned
A rolling estimation history involves a trade-off between sample-size (as more history is used) versus dynamic adjustment (as a smaller history is used to estimate factor payoffs). We have evolved a unique factor weighting methodology that uses the full available history of multivariate factor pay-offs but optimises on the prominence given to more recent observations. Each factor is also modelled individually rather than a single ‘learning rate’ being ascribed to the entire model.
Transparency and qualitative overview
The transparent nature of our models is conducive to a qualitative overview to check data correctness or if something is going on that the model is not aware of e.g. corporate actions or data errors . In a minority of cases we may override the model by not investing in a share forecasted to do well but we never take active positions opposite to what the model is predicting. ‘Pure’ quants approaches are liable to a number of easily remediable weaknesses that should not be neglected.
A Cutting Edge Approach to Portfolio Construction
Usually a two-step approach is taken: (i) factor weightings are optimised in order to attain predictive accuracy in the context of an alpha forecasting model and, thereafter, (ii) portfolio weightings are optimised in conjunction with a risk model as part of the portfolio construction process. We have developed advanced original technology that allows these two steps to be blended into one which considerably enhances the effectiveness of our portfolios.
Focus
We are exclusively an equity portfolio manager and all of our funds are managed off a single investment methodology that underlies all of our models. We are focused specialists and pioneers in the development and application of multifactor stock selection models and do not accept any other mandates.
We outsource all non-core functions in order to ensure an undiluted focus on our investment process.