Estimating Risk for Extreme Value Distribution and its Implication in Passive Asset Management
Abstract/Description
As precise evaluation of risk is fundamental to survival and progress of financial institutions as well as individual investors, we present in our paper well known facts of the term ‘heavy tails’ that are extreme outcomes in real financial world which may occur more frequently than predicted. As, these heavy tails represent down-side risk for investment firms, we use Value at Risk (VaR) as a risk measuring tool to manage exposure in an asset portfolio based on passive style of investing.
Passive style of investing allows fund analyst to track index values for managing asset portfolio on daily basis without applying vast trading strategies. In order to limit our research on extreme values and events that may incur in an index-based portfolio, we choose our portfolio of assets on basis of widely formed KSE-100 index stocks, so that frequent adjustments are not required. We, then employ Extreme Value Theory (EVT) method and utilize its heavy tail statistical distributions such as Gumbell, Frechet and Weibull to highlight unlikely extreme values from our historical data of KSE-100 index that have occurred in the past and similarities may also lie in future as uncertainty prevails in economy and financial markets. Further, after studying our financial data of stock index with Extreme value distribution characteristics, we estimate risk behavior using Semi-parametric model of VaR in our distribution tail index.
Our research methodology comprises of quantitative research and we use secondary and experimental data (randomly generated) to employ and test in our model parameters.
By large, this research allows us to understand critical issues in relation to risk exposure of any portfolio. Detailed study of down-side risk characteristics with historical data leads us to conclude that we can minimize wealth loses for investors if we implement sound methodology of risk estimation.
As we focus on extreme events and distributions, which by definition is uncommon in normal conditions, EVT demand larger sample sizes (we use 5,000 days of data to identify extreme observations).
Keywords
Value-at-Risk, heavy tails, extreme value theory, risk exposure, asset portfolio, passive style of investment, index-tracking
Track
Finance
Session Number/Theme
3A: Finance
Session Chair
Dr. Mohsin Zahid Khawaja ; Dr. Falik Shear
Start Date/Time
30-5-2024 5:00 PM
End Date/Time
30-5-2024 6:00 PM
Location
MCS – 3 AMAN CED Building
Recommended Citation
Ahmed, M. (2024). Estimating Risk for Extreme Value Distribution and its Implication in Passive Asset Management. 3rd IBA SBS International Conference 2024. Retrieved from https://ir.iba.edu.pk/sbsic/2024/program/33
COinS
Estimating Risk for Extreme Value Distribution and its Implication in Passive Asset Management
MCS – 3 AMAN CED Building
As precise evaluation of risk is fundamental to survival and progress of financial institutions as well as individual investors, we present in our paper well known facts of the term ‘heavy tails’ that are extreme outcomes in real financial world which may occur more frequently than predicted. As, these heavy tails represent down-side risk for investment firms, we use Value at Risk (VaR) as a risk measuring tool to manage exposure in an asset portfolio based on passive style of investing.
Passive style of investing allows fund analyst to track index values for managing asset portfolio on daily basis without applying vast trading strategies. In order to limit our research on extreme values and events that may incur in an index-based portfolio, we choose our portfolio of assets on basis of widely formed KSE-100 index stocks, so that frequent adjustments are not required. We, then employ Extreme Value Theory (EVT) method and utilize its heavy tail statistical distributions such as Gumbell, Frechet and Weibull to highlight unlikely extreme values from our historical data of KSE-100 index that have occurred in the past and similarities may also lie in future as uncertainty prevails in economy and financial markets. Further, after studying our financial data of stock index with Extreme value distribution characteristics, we estimate risk behavior using Semi-parametric model of VaR in our distribution tail index.
Our research methodology comprises of quantitative research and we use secondary and experimental data (randomly generated) to employ and test in our model parameters.
By large, this research allows us to understand critical issues in relation to risk exposure of any portfolio. Detailed study of down-side risk characteristics with historical data leads us to conclude that we can minimize wealth loses for investors if we implement sound methodology of risk estimation.
As we focus on extreme events and distributions, which by definition is uncommon in normal conditions, EVT demand larger sample sizes (we use 5,000 days of data to identify extreme observations).
Comments
This research paper is a working paper with results I am still compiling. Due to time limitations, it's in progress and I am expected to complete this paper in a week time.