Value at Risk (VaR) An Important Measure of Risk and Easy to Understand Too

Risk Management is older than written history.historical data used to construct the VaR estimate
However, it wasn't until the mathematical advances incontains useful information to forecast the loss
the 18th and 19th centuries when actuarial sciencedistribution. Value at Risk aggregates all of the risks in
and civil engineering where flourishing did proactivea portfolio into a single number ideal for reporting to
risk management evolve. Financial risk came later asmanagement or in shareholder reports.
an invention of the second half of the 20th century.One of the tools for assessing VaR is Monte Carlo
Before, gold acted as a hedge against interest ratesimulation. Like most financial models some
or foreign exchange risk but short-term interest rateassumptions are made first such as the distribution of
changes or extreme currency fluctuations madechanges in asset prices are normally distributed. Data
financial risk management difficult. In the early 197osis then collected to estimate the parameters of the
when the gold standard was dropped and stagflationdistribution (ie standard deviation). Monte Carlo then
was the word of the day financial risk managementprovides several sets of possible future outcomes
had become more important than ever. But how?given the changes in asset prices. The portfolio is
Fortunately, Modern portfolio theory was introducedthen revalued with each outcome leaving you a set
by Harry Markowitz back in 1952 with his paperof portfolio valuations relating to the set of possible
"Portfolio Selection," and a new breed of quantitativeoutcomes of the asset price change. Then from this
mathematicians were finding their way to Wall Street.distribution of data you take say a 95 or 99
With the rise of "quants" and the fall of markets onpercentile (the 1% part of 1% 1 day VaR) loss as
Black Monday October 19, 1987 had finally pushedVaR.
risk management to the front of every boardroomOne of the common questions with Monte Carlo is
agenda. Value at Risk was born from the crosshow many simulated asset price changes should be
breeding of sound knowledge of operational risk andused. It really depends on the complexity of the
a blend of modern portfolio principles.portfolio. For non-linear instruments like options more
A recent risk management study bysimulations would be prudent than linear asset
Pricewaterhousecoopers found that the 3 mostportfolios. There are ways to improve accuracy in
important objectives of risk management wereMonte Carlo models such as antithetic variance
identifying new and emerging risks, measuring andreduction allowing for less simulations for a given
monitoring risk and being able to communicate theseportfolio.
risks to the executive team. Value at Risk fulfills theHere is a simple example of the VaR concept with
last 2 criteria quite well.our Monte Carlo data in hand. Assume that our
VaR as it's also known is a market risk measure fordistribution of possible one day changes yields a 2%
extreme or highly unlikely outcomes. Denominated inprobability that a portfolio loss will exceed $10,000, a
units of currency it is a statistical measure of the1% probability the loss will be between $8,000 and
potential loss that a portfolio would be subject too$10,000 and a 2% probable loss of between $6,000
generally with a 1 day time horizon (you may haveand $8,000. There is then a 5% probability that the
heard of 1% 1 day VaR). It can be expected toloss will exceed $6,000. If we use this 5% probability
close out or hedge a portfolio within 1 day. Indeed(5% 1 day VaR) as the criteria to define a loss due
there are instances when liquidating a portfolio mayto normal market conditions then $6,000 is the Value
take several days such as with credit products. Inat Risk.
this case, one can use an average time horizon for aGiven the ability for risk managers to easily report a
multi- horizon portfolio.single number to the board or to her regulators it's
Of course there are assumptions with VaR as wellno wonder that VaR has remained an important
such as the portfolio remaining constant over themeasure of risk.
time horizon. A VaR model also assumes that the