Value at Risk – a new way to compare risks across funds

Value at Risk, or “VaR”, is a relatively new risk measure for the asset management industry, but is one that is rapidly becoming a benchmark for risk. The Investment Management Association (IMA) has already ruled that “sophisticated funds” (funds using sophisticated derivative strategies) must use VaR techniques, and it is likely that VaR will become a leading risk measurement in the asset management industry over the coming years.

The reason for VaR’s growing popularity is that it tells an investor what an investment’s likely downside is, provided the model’s assumptions hold true, we can say with an x% of confidence what an investment’s maximum expected loss will be over a set time period. We typically look at a 30 day time period, and use a 99% confidence interval (ie we can say with 99% confidence that it will be the case). For example, The M&G Optimal Income Fund, currently has a VaR of £1.1m, which means there is a 1% chance of the fund falling by at least 1.6% at any point in the next 30 days (based on the fund’s value of £67m).

How is VaR calculated? VaR is calculated using historical price movements and correlations of a very wide range of variables going back five to ten years (there is a greater weighting attached to recent market data). These variables are then used as inputs for models such as the “Monte Carlo” scenario set, which estimates risk by running thousands of potential scenarios. We run VaR on a daily basis, and in this way we can get an up-to-date view of how much risk we are taking, where these risks are, and whether there are any unintended risks. We can use the model to carry out stress tests by looking at specific variables, which helps us to understand what would happen to the value of a fund given, say, a 1% rise in UK interest rates. We can also look at particular historical events, and answer questions like “what would have happened to the value of a portfolio during the 1987 stock market crash?”

One output from our risk system that is particularly interesting is that the VaR of the high yield corporate bond market is around -0.6%, whereas the VaR of the UK investment grade corporate bond market is approximately -2.0%. It seems intuitively wrong to say that high yield is “lower risk” than investment grade, but if you look at what actually constitutes “risk” then it is not so surprising.

An investment’s risk is a combination of two things – duration (interest rate risk) and credit risk (essentially default risk). High yield bonds have more credit risk than investment grade bonds, but because they tend to be much shorter dated, they have much less interest rate risk. Over the past three years, corporate bond spreads have ground in steadily tighter and credit risk has therefore been remarkably stable. Meanwhile, interest rate expectations have swung fairly dramatically – in the UK, in January 2006, base rates were 4.5% and were expected to fall to 4.25%, but by the beginning of February 2007, rates were 5.25% and were expected to rise to 5.75%. Investment grade bonds have been much more volatile as a result. High yield bonds clearly aren’t always going to be “lower risk” than investment grade, just that over the next 30 days (according to the model) they are likely to be lower risk.