Managing risk is difficult. It takes time and continual effort, and even then no risk management system can be perfect. How do you price the risk of a trade becoming so large it warps the very thing it is trying to trade? Or the risk that your losing trade becomes known to the outside world and they line up to make you pay extra to get out of it? There are just so many inputs and possibilities, that it sure would be nice if the whole process were simplified – say, by boiling all the risk possibilities down to a single number perfectly represents a firm’s potential losses over any one day, week, month, etc.
Enter Value at Risk (VaR), a metric that tries to express how likely it is for a firm to lose a given amount of money over a particular period of time. For example, a daily 5% VaR of $10 million means that there is a 95% chance that the firm’s losses on any given day will be less than $10 million. Sounds nice, but unfortunately, like most things that sound too good to be true, it probably is.
There have been hundreds, if not thousands of articles, blog posts, books, etc. calling out VaR as useless… Some were written before the 2008 financial crisis, and many more after. The number of critics that have been railing against VaR for years is staggering:
- “Against Value at Risk” – Nassim Taleb’s site Fooled by Randomness
- “Killing VaR” – Financial Times blog Alphaville
- “Risk Mismanagement” – New York Times
- “The Number That Killed Us: A Story of Modern Banking, Flawed Mathematics, and a Big Financial Crisis” – by Pablo Triana
- “Value At Risk Underestimates Extreme Scenarios” – Seeking Alpha
- “How To Get Permission To Put $2 Billion Of Your Employer’s Money At Risk” – Business Insider
- “JPMorgan Loss Bomb Confirms That It’s Time to Kill VaR” – Naked Capitalism
So when we heard that JPMorgan, the world’s largest and generally most respected bank, declared a staggering $2 billion loss, we were floored to find out how they were measuring risk:
JPMorgan also changed how it calculates so-called value at risk, or VAR, a measure of how much the company estimates it could lose on securities on 95 percent of days. The company restated its VAR for the first quarter, previously disclosed at $67 million, at $129 million. The bank used a new model for calculating its trading risk in the first quarter that Dimon said was “inadequate.” It is reverting to the old model.
A few things about this strike us. First of all, you’re still using VaR? Where were you the day they taught “lessons from the 2008 financial crisis?” Second, if JPMorgan, the supposedly most risk-centric and well-run bank out there, is doing things this dumb, what are the other banks doing? And lastly… they’re reverting to the old model? Most critics of VaR aren’t just saying it’s poorly executed, but that it’s a terrible idea in the first place. So JPM’s plan is to try it again with the older model? Fool us once, shame on you. Fool us twice…