top of page

Tsallis Relative Entropy

In an earlier post (Entropy), we discussed Information Entropy and how it can be used to define risk in financial markets. We also briefly touched upon Tsallis Entropy which unlike Shannon Entropy, defines entropy for systems with long range interations. Since the performance of financial markets is the result of many interacting agents, an appropriate entropy used to define risk is the Tsallis Entropy.

Relative Entropy is the risk associated with an equity, compared to a standard index such as S&P 500 , Nasdaq ---.  This is important in portfolio management, where one is interested in constructing and managing portfolios that beat market returns in the long run (> 5-6 yrs).  

Tsallis Relative Entropy(TRE) shows how much the risk (defined by Tsallis entropy) of an individual equity differs from that of a standard index. Our back studies show that over a period of 6 years or longer, portfolios constructed/managed using  'Tsallis Relative Entropy' as the risk measure, outperform both the markets and those managed using the risk measure 'beta', usually used by investment managers.

 

For definitions and details of Tsallis Entropy and Tsallis Relative Entropy , see publications belowAlso see A New Risk Measure for details.

 

Financial Portfolios based on Tsallis relative entropy as the Risk Measure, Sandhya Devi, 2019,  Journal of  Statistical Mechanics  093207


       https://doi.org/10.1088/1742-5468/ab3bc5
    
         This article can also be found in  
         https://arxiv.org/abs/1901.04945v3

 

bottom of page