Risk and value at risk

Increasingly complicated tools known as financial derivatives have been introduced in recent times to manage the market risk arising from floating exchange rates. The rapid development of the derivatives markets has in turn introduced new risks into the business of finance - witness the highly-publicised trading losses at Metallgesellschaft and Procter and Gamble. A principal method for measuring and reporting market risk in the portfolios of banks and their clients is 'value at risk' (VaR). Fred Stambaugh explains the concept of 'value at risk' and describes three principal approaches to calculating it - correlation matrix, historical simulation and Monte Carlo simulation; they are alternatives, not competitors. As well as setting out their uses, he considers those situations that go beyond 'value at risk', i.e. dire events that lie beyond the confidence level of VaR. Techniques for portfolio stress testing are discussed.