Definition of value at risk - VaR

A risk management model that calculates the largest possible loss that an institution or other investor could incur on a portfolio. Value at risk describes the probability of losing more than a given amount of assets, based on a current portfolio. It was developed (not originated) at JPMorgan in response to a demand from then chairman Dennis Weatherstone to be told at 4.15pm every day what the overall risk position of the bank was.

He was concerned the various trading desks and business units might be taking correlated bets without knowing about it. Without having time to check each trader’s position individually, he wanted a way to understand the overall position.

This has a couple of important implications: one, that it is a snapshot, a picture of what the risk level is now, based on what you hold now. The longer the time period you apply it to, the fuzzier it is likely to become.

Two, it is a lower bound for loss, not an upper bound – if your model says that at a 95 per cent VaR, the value at risk is half of your assets, you can expect to lose half or more of your assets one day in 20. It should be used to make sure you can withstand these losses, not necessarily to prevent them.  [1]

VaR is a mathematical model that purports to estimate the maximum future losses expected from a trading portfolio, with a degree of statistical confidence. VaR's calculation can be extremely technical, or it can be as simple as looking at a subjective past period and then projecting future risks from there. A big problem with VaR is that it can very easily produce very low numbers, gravely misrepresenting true exposures. Relying on the past can be treacherous: a quiet past period need not imply future quietness, historical volatility and correlation may betray you. Also, Var models can use the 'normal probability distribution', which very unrealistically assumes no chance of extreme events.

VaR played a key negative role in the 2008 credit crisis, by severely underestimating the danger from toxic mortgage products and by allowing banks to enjoy excessive levels of leverage on their trading positions. Given that toxic leverage is what sank the banks, we could argue that VaR, in the end, caused the crisis. Recognising all this, the Basel Committee for Banking Supervision, which had enthusiastically adopted VaR since 1995, has been busy at work disowning the model and tweaking the bank capital formula. Just a few weeks ago, Basel announced that it no longer wants to keep using VaR. [2]


FT Alphaville - The JPMorgan Whale's regulatory motive (June 2012)

FT Articles & Analysis

No articles are associated with this term