After continuously being asked by clients what Value-at-Risk (VaR) really is, other than a fancy statistical term, I thought to give a brief overview and discuss its pros, cons, and significance.
VaR is used as a measurement of probability of loss in value for a specified confidence interval. For example, to say that a portfolio has a 95% monthly VaR of -3.5%, would mean that the portfolio has a 95% confidence interval of not loosing more than 3.5% within any given month. Three ways of calculating VaR are Parametric VaR, Historical VaR, and Monte Carlo simulation.
Parametric VaR is the most criticized since it assumes the normal distribution of the returns. The cut-off point on the left tail of the bell-shaped curve, depending on the confidence interval, will define VaR. Therefore, all we need for the VaR calculation is the mean and the standard deviation of returns. Because the returns tend to be negatively skewed with the fat tails (defined by the term leptokurtosis), the parametric VaR tends to underestimate VaR which causes the outraged criticism of the whole concept.
Historical VaR overcomes this disadvantage: it does not make any assumptions about the returns distribution because it simply uses historical returns data. When these returns are sorted in order from the lowest to the highest, the percentile defined by the confidence interval determines VaR. As the Historical VaR uses the actual returns (losses) in order to define the maximum potential loss over a specific period, it is valid as long as the current economic environment is similar to the past. Since Historical VaR is the most routinely used measurement, let’s sum up its pros and cons.
Pros:
– Conceptually easy
– Uses the actual returns (as opposed to the assumptions about the normally distributed returns)
– Gives the operational and analyzable number
– Easy to use in conjunction with the expert opinion
Cons:
– Uses historical data to predict the future, hence not as strong in a changing economic environment
– Does not determine the optimal period of returns for the most reliable outcomes
– Does not give the probability and magnitude of the abnormal loss during extreme events
To solve the problems of Historical Var and to retain its advantages, Monte Carlo simulation can be used to estimate VaR. Monte Carlo simulation generates the set of outcomes with many various assumptions behind each outcome. VaR is determined by the same principle as the percentile of the chosen level of confidence. Unlike with Historical VaR, Monte Carlo simulation implements the great variety of the outcomes and scenarios to define VaR, and, therefore, works in a wider variety of the situations.
The VaR concept allows the systematic analysis and monitoring of the risk. As any other instrument, VaR can’t be used in isolation, but rather in sync with other techniques like stress testing and scenario analysis. With the uncertainty of future trends, VaR does not pretend to replace expert opinion, but rather supplement it. Therefore, without a VaR model that will quantify the risk, the portfolio manager is left with the philosophical generalizations about the economic environment and non-quantifiable perceptions of the risk.
While some intuitive “tools” of risk assessment relying only on expert opinion might work for an individual (such as stop losses, max position limits, gross/ net exposure limits) they fail to be effective in attraction of significant investors’ capital. Nowadays, risk management is not a novelty but more of a must for asset managers, thus confirming that investors and money managers need to use a common language understood by both.
Correct interpretation and understanding of VaR assumptions and limitations, as well as proper use with other risk metrics is the new common language of risk management, giving its user an advantage in the competition for the institutional capital. As affordable risk management vendors become increasingly available, saving your money and time, VaR evolves from a misunderstood number into a valuable decision-making tool. Feel free to check out our savings calculator to see just how much you could save when outsourcing your risk management services!
