Friday, February 5, 2010

VaR to the rescue...or not

"Mexican Deputy Finance Minister Alejandro Werner said the government will announce in coming weeks a change in risk-management regulations for pension funds to help ease the effect of market volatility on investments. 

"'Current value-at-risk regulations, which estimate how much the funds could lose in a given day, exacerbated losses by forcing funds to sell their riskiest assets during the global financial crisis and increased volatility in pension fund portfolios,' Werner said."

I find myself gravitating to stories about Value at Risk. I find the whole concept quite intriguing. Very sophisticated mathematical models are employed to project the potential loss a firm or portfolio faces. Thus, we have an ex ante or forward looking view, which in theory would be extremely helpful as the idea of being able to know what lies ahead is clearly attractive.

The problem, of course, is that this speculative view is usually far from the mark and can lead to problems. ANY ex ante estimate, on virtually any topic, is usually wrong. (As an aside, it's interesting to note that in the Old Testament, people who make such projections are subject to death!). 

The captions above come from a story which details the negative impact VaR has had on Mexican pension funds. Clearly, this isn't the intent of this measure.

When considering VaR, one must be aware of the assumptions, constraints, and risks (risk of a risk measure). Too many people treat VaR as somewhat sacrosanct, which is problematic. It isn't a panacea. No risk measure is. It may be a valuable tool but like all tools must be used properly.

p.s., we will host a webinar on VaR later this month. To learn more, please contact Patrick Fowler.

2 comments:

  1. Throw a number up in the air and everyone looking at the number would come up with his or her own interpretation. Have we ever asked if the individual reading the number actually understand what it means? Go a little further, ask what do we could use it for and observe their reaction?

    ReplyDelete
  2. Thanks for your note.
    Agreed! Yes, numbers carry a lot of weight (just read "How Genius Failed"). That's why the numbers have to be couched with cautionary language. Somewhere I've got the wording Merrill used, which was pretty descriptive and would be a good model for folks to employ. I'll post that when I can find it.

    ReplyDelete

Note: Only a member of this blog may post a comment.