BA Buddy Group 1
BA Buddy Group 1
BA Buddy Group 1
SUBMITTED TO:
Dr. Yogesh Funde
Assistant Professor
ASMSOC
Section- TY BBA-A
Buddy Group- 1
To Calculate the Non-linear Value-at-Risk (VaR) for two derivative funds based on the price
of Crude Oil in the United States
Introduction:
The ability to quantify and estimate the possibility for losses is critical in the field of financial
risk management. This is where the notion of Value-at-Risk (VaR) enters the picture,
functioning as a vital tool for risk assessment and decision-making. VaR estimates the
potential losses that a portfolio might incur over a specified time period and at a given
confidence level. While typical VaR calculations are simple, non-linear VaR is required for
more complicated portfolios that incorporate assets with non-linear connections. Going into
the complex world of non-linear VaR in this study, investigating its computation using three
separate methods: the Delta-Gamma Method, Historical Simulation Method, and Monte
Carlo Simulation Method.
While linear VaR models are successful in quantifying risk for portfolios with linear asset
connections, they frequently fall short when dealing with more complex and diversified
assets. Options, derivatives, and other complicated financial instruments can cause non-linear
correlations in portfolios. To handle these non-linearities, non-linear VaR computation
becomes critical. Acquire a more realistic depiction of prospective losses and can respond to
market movements and uncertainty by including these intricacies.
The Purpose of This study:
In this study, two different methods for computing non-linear VaR: the Historical Simulation
Method, and the Monte Carlo Simulation Method. These methods provide various views and
levels of complexity, addressing a wide range of portfolios and risk management
requirements.
Literature Review
1. The research paper “Measuring Risk utilizing Credible Monte Carlo Value at Risk
and Credible Monte Carlo Expected Tail Loss” (Sulistianingsih, 2022) presents the
risk of the individual stocks that make up a portfolio using two unique methodologies:
Credible Monte Carlo Expected Tail Loss (CMC ETL) and Credible Monte Carlo
Value at Risk (CMC VaR). The CMC VaR combines the concepts of Credible Value
at Risk (Cr VaR) and Monte Carlo VaR (MC VaR). In the meantime, CMC ETL is
created by combining MC ETL with Credible ETL (Cr ETL). The revolutionary
strategy's capacity to evaluate the distinct risk of each asset while creating three
portfolios is supported by empirical data. Four NYSE-listed firms, two NASDAQ-
listed companies, two LSE- indexed stocks, and five LQ 45-indexed Indonesian
shares were included in the portfolios under evaluation. Kupiec Backtesting is also
used to assess the accuracy of the CMC VaR. According to the empirical results of
the study, two alternative methodologies can be used to evaluate the risk with 80%,
90%, and 95% confidence levels. The recommended alternatives can also solve the
drawbacks of VaR and ETL, which do not account for the risk associated with a
portfolio's diversified assets.
2. The research paper “Risk value at risk: Models and applications” (Engle, R. F., &
Manganelli, S., 1999) states that Value-at-Risk (VaR) is a crucial risk measure used to
quantify potential portfolio losses over a specified time frame and confidence level.
Univariate VaR focuses on single assets or portfolios, while multivariate VaR
considers multiple assets or portfolios, serving essential purposes in capital
requirements, risk management, and pricing financial products for financial
institutions.
Univariate VaR Methodologies
Univariate VaR methodologies have distinct strengths and weaknesses, including:
Historical Simulation (HS): It uses historical data to simulate future portfolio returns
and calculate VaR, but its accuracy heavily relies on the quality of historical data and
neglects return volatility.
Monte Carlo Simulation: It offers flexibility and accuracy but can be computationally
intensive.
3. The research paper "Improved techniques for using Monte Carlo in VaR estimation"
(Srinivasan, 2001) provide new techniques for improving the efficiency and accuracy
of Monte Carlo simulation for VaR estimate. Financial organisations utilise Value at
Risk, also known as VaR, as a well-known risk indicator to determine the largest
probable loss that could happen to a portfolio over a given period of time. Monte
Carlo simulation is a popular method for estimating VaR since it can simulate a large
variety of probable outcomes. One of the main tactics recommended in the study is
the use of variance reduction, which can lead to more accurate results and/or a
reduction in the number of simulations required. The authors offer numerous
strategies for reducing the variance of a Monte Carlo simulation, including antithetic
variates, control variates, quasi-Monte Carlo, and parallel processing. To improve the
efficacy and accuracy of Monte Carlo simulation for VaR estimate, additional
techniques include using a hybrid strategy, a risk factor model, and a scenario
generator. Empirical results show that the suggested solutions can significantly
improve the efficiency and accuracy of Monte Carlo simulation for VaR estimate.
Overall, the work contributes significantly to the body of knowledge on Monte Carlo
simulation for VaR estimate by providing approaches that can be helpful for a range
of financial applications.
Detailed and extensive explanation of why we chose the historical method and the Monte
Carlo approach for predicting the derivative non-linear VAR (Value at Risk) margin:
HISTORICAL METHOD: The historical method is a fundamental and widely used
technique for estimating VAR. It relies on historical market data to understand how an
investment portfolio's value would have behaved in the past during different market
conditions. This approach offers several advantages:
Advantages include its intuitiveness, data-driven nature, realistic stress testing, and
transparency.
However, the historical method also has limitations, especially when dealing with non-linear
derivatives. It assumes that the future will resemble the past, which may not always hold true,
and it doesn't explicitly account for factors like volatility clustering and structural changes in
markets.
All the prerequisites and considerations for using the selected methods, which are the
historical method and the Monte Carlo approach for estimating the derivative non-linear
VAR (Value at Risk) margin.
2) Independence of Observations: we ensured that the historical data used for the
analysis had independent observations. Autocorrelation or serial correlation can affect
the accuracy of VAR estimates, so we checked for autocorrelation and made
necessary adjustments if detected.
3) Outlier Detection: Outliers in the historical data can significantly impact VAR
calculations. We employed techniques like identifying and addressing outliers, either
by excluding them or using robust statistical methods.
4) Data Quality: We confirmed the quality and accuracy of the historical data, making
sure there were no data errors or gaps that could distort the analysis.
Interpretation of data:
JAMOVI
The data collected and analysed using Jamovi, on a monthly basis for a 5-year period, has
revealed several key insights. The dataset displays a high degree of reliability, with only one
outlier identified for commodity one, and no outliers for commodity two. Furthermore, the
data sets are independent of each other, which enhances the robustness of the analysis.
Despite both commodities belonging to the same general stocks category, commodity one
exhibits a more significant impact than commodity two. Additionally, the data's normal
distribution suggests that it meets the assumptions required for statistical analysis. Given
these characteristics and the similarity in parameters between the selected stocks, this dataset
appears to be ideal for further in-depth analysis and can serve as a solid foundation for future
research or decision-making processes.
HISTORICAL METHOD:
Using the historical method with a dashboard set at a 95% confidence level, here's an
interpretation of how the data is utilized:
Dashboard Input: The dashboard is a user-friendly tool where the interpreter can input key
information:
Portfolio Value: This is where the interpreter enters the current value of the portfolio
being analyzed.
Expected Weightage: The expected weightage of each commodity in the portfolio can
also be input. This provides insights into how the portfolio is distributed among
Commodity 1 and Commodity 2.
Value at Risk (VAR) Calculation: The historical technique uses historical market data to
estimate the Value at Risk (VAR) of the portfolio at a 95% confidence level. This is done by
entering the projected weightage and portfolio value into the dashboard.
HPR: This approach relies on historical price returns (HPR) and involves the use of specific
code for computation
Risk Evaluation: The resulting VAR offers a prediction of the possible loss the portfolio
might suffer over a certain time period and confidence level. With a 95% confidence level in
this instance, there is a 5% possibility that losses will surpass the calculated VAR.
Decision Support: Decision-makers can evaluate the risk exposure of the portfolio by using
this data interpretation approach. It aids in comprehending possible negative risk and can
guide capital allocation and risk management plans.
Hence, this dashboard tool in conjunction with the historical method provides a useful and
data-driven strategy for evaluating portfolio risk. Making educated decisions about risk
management and portfolio allocation is made possible by enabling decision-makers to
comprehend the various loss scenarios.
MONTE CARLO APPROACH:
The Monte Carlo method, while valuable for obtaining optimal weightages in risk
assessment, comes with its limitations. One of its significant drawbacks is the wide range of
values it generates for scenario-based Value at Risk (VaR). This simulation technique
operates based on random variables, and the final Monte Carlo VaR is determined through
incremental percentiles of scenario-based VaR simulations. Despite its efficiency, the Monte
Carlo method falls short in terms of accuracy. It requires numerous iterations and
simulations, yet often fails to produce results that align closely with expectations. This wide
range of values can be a challenge when making critical risk management decisions, making
it necessary to consider these limitations when applying this approach.
BIBLIOGRAPHY
1.https://www.iaeng.org/IJAM/issues_v52/issue_1/IJAM_52_1_31.pdf
2. Duue, D., & Pan, J. (n.d.). An Overview of Value at Risk.
Platon, V., & Constantinescu, A. (2014). Monte Carlo Method in Risk Analysis for Investment
Projects. Procedia Economics and Finance, 15, 393–400. https://doi.org/10.1016/s2212-
5671(14)00463-8
APPENDIX