If you consider Basel II, there are two ways of calculating Market Risks VAR:
• Historical Simulation Approach
• Model Building Approach
What makes them different?
Historical Simulation approach is most frequently used by organisations. As the name suggests, we consider daily changes in past/historical values to compute the likelihood of the variations in values of current portfolio between given time frame. The other advanced version of this model places more emphasis on recent observations. The key assumption in historical simulation is that the set of possible future outcomes is fully represented by what occurred in a definite historical time frame/window.
On the other side, model-building approach involves assumptions about the joint probability distributions of the returns on the market variables. This model is also known as variance-covariance approach.
This is more apt for portfolios which has short as well as long positions in their bucket. This consists of commodities, bonds, equities, etc. in the portfolio. Here, the mean and standard deviation are computed from the distribution of the underlying assets returns and the correlation between them.
Daily returns on the investments are normally assumed to be multivariate normal which can be the models biggest drawback. Hence, model-building approach makes it easy to calculate Var.
Model Building approach assumes two things:
• The daily change in the value of a portfolio is linearly related to the daily returns from market variables
• The returns from the market variables are normally distributed
Shortcomings of Historical Simulations
Over reliance on past data can fail to serve the purpose as markets change every moment. The momentum can be gradual or sudden, but does not remain static.
Large number of factors like Technology, regulatory changes, economic conditions, seasonal patterns, etc. influence market and in such scenarios manager who are using historical simulation can face unfavorable situation.
Shortcomings of Model Building Approach
Also this approach is much more complex to use when a portfolio comprises of nonlinear products such as options. It is also a grim task to relax the assumption that returns are normal without a significant increase in totaling time.
When to use? Model building vs. Historical simulation.
Depending on the situation, appropriate model should be adopted by the organisation. While both of them have pros and cons, it is important to list down the objectives of risk model before adopting either of them.
Model building approach producer quicker results and can be used in conjunction with volatility and other correlation procedures.
The advantage of the historical simulation approach is that the joint probability distribution of the market variables is determined by historical data. This approach may not be very complicated however, it is little slow for computation. However, the methodology used in historical simulation is in line the risk factor and does not involve any estimation of variances or covariance’s which are statistical parameters.
One should use historical simulation model only when they have data on all risk factors over a justified historical period if they want the model to depict strong representation of the outcome in future.
To know more about model building join Imarticus Learning’s Financial Modeling Certification Courses, which will help you understanding opportunities in the Investment Banking, Private Equity, Budgeting and Financial Control space.
Picking stocks is hard work and there’s a reason why Mutual fund managers get paid so much to do it. So if you don’t have the time, we suggest starting out with an index fund like Franklin India Index or HDFC Index Fund – Sensex. An index fund is a mutual fund that invests in a predefined stocks of an index in a percentage allocation that resembles the index. Your portfolio could be a mix of different index funds, NSE Small caps, BSE Sensex and maybe even an international index fund.