What is Value at Risk (VaR)? A quick google search will return multiple approaches to computing this metric. But broadly and generally defined, Value at Risk(VaR) is a statistical technique used to estimate the dollar value amount at risk of loss over some time period or future time horizon. Usually VaR is measured using a statistical confidence level of either 95% or 99%. What does this mean in layman’s terms? Value at Risk helps us compute the potential loss and uses statistics to estimate the answer to this question: ‘What is the minimum that I will likely lose on this asset over the next trading day?’. When you apply a statistical threshold of 95% or 99%, VaR quantifies the tail of the distribution of returns and allows you to say, 'Roughly 1% (for 99% statistical confidence) of the time, I can expect to lose at least the VaR amount over the next trading day.' Banks use VaR to calculate and control risk exposures in respect to business line risk limits set both internally by risk management desks and by regulators.
Why is VaR useful to traders and risk officers? Value at Risk is useful because it helps quantify an estimated loss. Still struggling with the concept? It may be useful to go through some examples. If Bitcoin has a 1 Day VaR metric of -6% at the 95% confidence level; this is implying is that there is roughly a 5% statistical chance of a MINIMUM loss of 6% occurring tomorrow. Five percent (5%) of the time translates into once every twenty (20) days. If Bitcoin has a 1 Day VaR metric of -6% at the 99% confidence level, this implies that there is a statistical 1% chance of a loss in a magnitude of at least 6% occurring over the next day. The VaR calculation uses the term ‘minimum’ loss because there is ALWAYS the possibility that a loss will exceed VaR estimate; and unfortunately, it is a known limitation that VaR does not help us in estimating losses that occur past the statistical threshold.
Value at Risk is useful because it helps quantify an estimated loss
Value at Risk metrics can also be used to estimate a loss over a time horizon that is longer than one day. It would not be unusual to monitor a weekly (5 day) VaR exposure or 10 day time period. The 5 day value generally corresponds to the equity market’s trading week. If the 10 Day VaR metric for Bitcoin at the 99% confidence level is -9%, this implies there is a 1% chance of a loss of at least -9% occurring over the next 10 trading days. VaR can be used to help an investor or trader manage his overall risk exposure or profile; but it is only an estimate, and never a certainty.