Skip to content

Rohitku123/VaR-Computation-and-Risk-Assessment

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

3 Commits
 
 
 
 

Repository files navigation

Portfolio Risk Optimization with Parametric Monte Carlo Simulation

Overview

This project explores portfolio risk management using Parametric Monte Carlo Simulation applied to a portfolio of three major Indian banks: SBI, PNB, and BOB.

We assess portfolio return, standard deviation, and compute Value at Risk (VaR) across multiple confidence levels and time horizons to quantify downside risk and highlight diversification benefits.

Thanks to Mehul Mehta for the insightful lecture series that laid the foundation for this simulation-based approach.


Portfolio Composition

  • Total Investment: ₹100,000
  • Weights:
    • SBI: 30%
    • PNB: 40%
    • BOB: 30%

Methodology

We used Parametric Monte Carlo Simulations grounded in the following three risk assessment methods:

1. Parametric VaR (Variance-Covariance Method)

Description:

  • Assumes returns are normally distributed
  • Uses mean and standard deviation of portfolio returns
  • Considers covariance matrix for diversification effect

Process:

  • Calculate portfolio mean and variance using asset weights and covariance matrix
  • Estimate potential losses for different confidence intervals using Z-scores

Implications:

  • Quick, reliable for portfolios with normally distributed returns
  • Highlights tail risk (extreme losses) and diversification impact

2. Monte Carlo Simulation (Parametric)

Description:

  • Simulates thousands of possible portfolio returns using random samples from a normal distribution defined by mean and SD
  • Parametric, as it assumes normality but introduces randomized variability

Process:

  • Generate random returns based on portfolio’s statistical parameters
  • Aggregate results to estimate probability distribution of losses
  • Calculate empirical VaR from simulation results

Implications:

  • Captures non-linear risk dynamics
  • More robust under uncertain future scenarios
  • Enables stress testing over multiple horizons (1-day, 10-day, 1-month)

3. Time Horizon Scaling

Description:

  • Scales 1-day VaR to longer time frames (10-day, 1-month) using the square root of time rule

Formula:

Implications:

  • Simple and intuitive for projecting risk over time
  • Assumes returns are i.i.d. (independent and identically distributed)
  • Practical for operational risk forecasting and compliance

Key Statistical Results

Metric Value
Mean Return 0.20%
Portfolio SD 0.96%

Value at Risk (VaR) Results

Time Horizon 99% CI 95% CI 90% CI
1-Day ₹-2,022.28 ₹-1,370.31 ₹-1,022.97
10-Day ₹-6,395.01 ₹-4,333.31 ₹-3,234.90
1-Month ₹-11,076.49 ₹-7,505.52 ₹-5,603.02

Takeaways

  • Risk Estimation: Parametric VaR is crucial for identifying exposure under normal market conditions.
  • Diversification Effect: Covariance analysis confirms that a balanced mix of SBI, PNB, and BOB reduces total portfolio risk.
  • Decision-Making Tool: Monte Carlo Simulations empower investors to plan for extreme loss events and optimize portfolio allocation.

Tools & Technologies

  • Python (NumPy, Pandas, Matplotlib)
  • Statistical Modeling
  • Monte Carlo Engine
  • Portfolio Theory

References

  • Lecture Series by Mehul Mehta
  • J.P. Morgan’s RiskMetrics™
  • Hull, J. (Options, Futures, and Other Derivatives)

Connect & Collaborate

Feel free to explore, fork, or contribute to this project. For discussions on portfolio risk modeling or advanced simulations:

📧 rkchs07@gmail.com]
🔗 [https://www.linkedin.com/in/rohit-kumar-38a121284/]

About

VaR calculation using Parametric and Non Parametric Approach in Excel

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors