PARSIMONIUM
Risk Management
Innovative approach
Classical risk metrics use static volatility and correlation levels. Although they have been used for decades, they suffer from a lack of sensitiveness. Potential market krachs are evaluated with a theoretical methodology making use of those lagging indicators.
​
We are different from these multi factors and VAR models, making use of dynamic parameters that are constantly normalized :
changes in correlations
spreads of correlations
speed of change in correlations
Correlations between Bonds and Equities
How does our model work?
Unlike classical models, ours is not based on correlation levels, but on dynamic correlation spreads. This method provides a better sensitiveness of the risk changes on the markets.
​
-
Correlation spread. The spread evolves randomly respectively to moving correlations (e.g. delta 1 = 0.13, delta 2 = 0.57), indicating regular risk changes.
-
Normalization of this correlation spread. Normalizing allows to turn random values of the spread into significant mathematical values (normal law).
-
Changes in the normalized spread. The model then checks if the spread crosses significant thresholds.
Our model in practice :
Parsimonium Index
The only one way to know if our method is efficient, is to put it into practice. We use our methodology to give an arbitrage signal based on the mathematical risk between Bonds and Equities, in order to favor one or the other asset.
​
The signal changes when the normalized spread crosses the significant threshold at one standard deviation. Following this rule, our resulting strategy is depicted on the chart at right.
