Value at Risk - Not a Perfect Measurement Tool

by: Susan Mangiero, CFA

In case you missed it, Joe Nocera wrote a user-friendly and quite interesting article about Value at Risk and the role of mathematical models in dealing with uncertainty. In "Risk Mismanagement" (January 2, 2009), this New York Times business columnist intimates that financial professionals may have relied too much on a single number, no matter how often it is updated.

This blogger wholeheartedly agrees with the premise that risk, in its totality, cannot be succintly captured with one number, one metric, one tool, one time period and so on. While Value at Risk, a statistical measure of likely loss during a given business time interval such as a trading day, can be helpful as ONE gauge of financial exposure, it can also be of limited use if underlying assumptions do not accurately capture "typical" price behavior for a particular financial instrument. Expressed as a single number, VaR is often measured in dollars (Euros, etc) and is deemed by some as an easy way to compare trading risk across banks/companies/portfolios. It can be calculated in several ways and is provided as part of company SEC filings and bank regulatory reports.

Alas, reality intervenes.

I won't repeat what Mr. Nocera so eloquently wrote in his lengthy piece but will add the following:

  • Numbers can guide but not replace solid decision-making by rational human beings.
  • No single number can measure the many types of risks that investors and traders face.
  • Even a collection of numbers can lead to bad decisions if not supplanted with common sense and a key understanding of qualitative and quantitative portfolio risk drivers, both in isolation and in conjunction with one another. We need look no further than market outcomes in 2008 to know that bad news can beget bad news (i.e. the notion of contagion and compounding of risk factors).
  • Longer term investors such as pensions, endowments and foundations should acknowledge that Value at Risk is designed more as a tool for short-term (trading) decisions.

An important area not addressed by the article is the nature and extent of due diligence being conducted by institutional investors and their consultants before allocating monies to various asset managers. I've yet to see too many institutional investors publish a comprehensive Risk Management Policy that is separate and distinct from an Investment Policy Statement (assuming that even an IPS exists). Besides Value at Risk, Sharpe Ratio, Standard Deviation, Correlation Coefficient(s) and maybe an Information Ratio and Tracking Error(s), how much risk management analysis is being conducted before institutions say "here's my money?" The current state of pension disclosure reporting leaves us mainly in the dark about how plan sponsors drill money managers on the topic of risk identification, measurement and management. Unless you're a mushroom, too little sunlight is not necessarily a good thing.

Note: Check out for a broad array of papers on the topic of Value at Risk. I've also written a book entitled Risk Management for Pensions, Endowments and Foundations (John Wiley & Sons). While some of the statistics are dated, it includes some good case studies and checklists. (I am in the process of updating the book.)