Alpha Forecasting
At Clermont Alpha, we continuously rethink every part of the investment process with the ultimate goal of extracting as much information from all data / signals available.
Visualization techniques, robust statistical methods, adaptive econometrics and proprietary signals leads to alpha forecast with higher predictive ability.
What's New:
More Educational Content coming soon
More educational content to come. Sujects include double taxation of bonds and corporate governance.
April 19, 2012: Death of a great investment counselor, Carl H. Otto
Updated: April 20, 2012
Alpha Forecasting
Investment skill and forecasting ability is critical in delivering alpha. In the “Investment Skill Measurement” section, I show the importance of measuring that skill. How can one improve its forecasting ability if it is not measured? The Information Content (IC) is this measure of skill. It can be used by any investment managers, whether traditional or quantitative. AT the stock level, alpha is directly related to a manager’s IC:
Where σr stands for the cross sectional dispersion of stocks’ returns, and S stands for our standardized forecast
In order to deliver greater alpha, one needs to increase his IC. IC can be improved by anyone of:
- Adding new information, new signals which correlate to future return, yet add information not contained in other signals
- Replacing current information/signals by better one.
- Better statistical data preparation (this is often neglected)
- Better forecasting techniques
- Better forecast combination techniques
- Better valuation models
THUS Better information management leads to more alpha.
Typical information/signals used in forecasting returns include:
- Valuation models (most important)
- Growth factors
- Momentum/Technical signals
- Insider/Management signals
- Earnings Revision
- Earnings surprise
- Earnings manipulation:
- Accruals
- Earnings Quality measures
- Other
In each signal category, sophisticated models can be developed to improve its forecasting ability (i.e. increase its IC). At Clermont Alpha, we have developed such sophisticated models in most signal categories and other models/signals in categories not listed above. Our edge in Alpha forecasting comes from better information management, including:
- Superior data preparation (pre-processing)
- The use of robust statistical techniques
- Several proprietary high IC signals/models
- The use of both static and adaptive forecasting econometric techniques
- Superior universe selection
For more information on these and other techniques and methods used to increase the overall model’s IC, contact us.