- Patented procedure replaces the fixed correlation matrix with a dynamic process that accurately estimates the dependency between assets using copulas
- Accounts for non-linearity and asymmetry in portfolios
FinAnalytica, the leading provider of multi-asset class, predictive performance analytics solutions to investment managers, announced today that it has been granted a new patent for the process of describing the dependency structure between financial assets in estimating future portfolio risk from the U.S. Patent and Trademark Office.
The process, already commercially available on FinAnalytica’s Cognity Risk Management and Portfolio Construction platform, overcomes the deficiencies of risk management systems built on the “correlations are constant over time” premise that gave asset managers a false sense of safety through diversification prior to the destruction of portfolio values in 2008.
The patent, System and Method for System and Method for Generating Random Vectors for Estimating Portfolio Risk (USP 8,170,941), provides a framework for modelling asset returns by incorporating a flexible copula function based on a mixture between two copula functions. The first copula describes the "central" part of the dependency model. The second copula impacts specifically in the tails of the dependency model. The system allows for choosing the weight applied to each of the two copula functions to create a general copula.
Boryana Racheva-Iotova, FinAnalytica President, commented, “There are no shortcuts for modelling the behaviour and complexity of today’s markets. Our collaborative dual copulae process leverages the most rigorous commercially viable approach in forecasting the dynamic dependence between assets. It combines both parametric and empirical approaches providing significant flexibility essential for multi-asset class portfolios.”
David Merrill, FinAnalytica CEO, said, “This patent is an incredible achievement for FinAnalytica’s Chief Scientist Zari Rachev and our research team. A core foundational element of our Cognity platform is that correlations are not constant, that they become more pronounced in the tails and are sharply higher in the left tail. This newly patented dynamic dual copulae process goes beyond the known limits and flaws inherent in the classical correlation matrix approach. It provides our customers a practical, production-ready solution for modelling both what we know happens and what may happen in the future.”
This patent, the fifth awarded to FinAnalytica since 2009, has firmly established the company as the leader in providing the buy-side community with innovative solutions to accurately measure and manage tail risk and return.