Measuring RisksSome companies focus

Measuring Risks
Some companies focus mostly on tail risk—the low-prob- ability, large-loss outcomes. As a result, when they measure the risk of changes in the present value of cash flows, they use a measure like VaR at a probability level that corre- sponds to a default threshold. Some of these companies also complement their VaR estimates with stress tests in which they investigate the impact on firm value of rare events (such as the crisis period of August and September 1998 that followed Russia’s default on some of its debt).
Though VaR is widely used, it is important to under- stand its limitations and to complement its use with other risk measures. Perhaps the main problem is that while VaR measures the loss that is expected to be exceeded with a specified probability, it says nothing about the expected size of the loss in the event that VaR is exceeded. Some have argued that companies should instead focus on the expected loss if VaR is exceeded. But focusing on this risk measure, which is often called conditional VaR, instead of focusing on VaR has little economic justification in the context of firm- wide risk management. Setting the company’s capital at a level equal to the conditional VaR would provide the firm with a lower probability of default than the targeted level, leading to an excessively conservative capital structure.
But a more important reason for companies to look beyond a VaR measure estimated at the probability level corresponding to a default threshold is that ERM adds value by optimizing the probability and expected costs of financial distress. It is therefore critical for companies to make sure that the equity capital set based on a VaR estimate leads to the targeted optimal probability of financial distress. Such an effort requires a broader understanding of the distribu- tion of firm value than is provided by a VaR estimate for a given probability of default. Further, since different levels of financial distress have different costs, a company can take these different costs into account and focus on the probabil- ity distribution of different levels of financial distress.
To compound the problem, when a company has a high rating target, the estimation of VaR becomes more of an art as the estimated VaR corresponds to an extremely low probability level. To see this, consider a company that has determined that an A rating is optimal. Since the proba- bility of default for an A-rated company is only 0.08% over a one-year period, to estimate its optimal amount of capital the firm must therefore estimate the loss in value that is exceeded with a probability of 0.08%. The problem, however, is that few A-rated companies have any experience of losses that come anywhere near that level. And without any historical experience of such losses, it is difficult for management to estimate the VaR at that probability level and then evaluate the result.