Increasingly sophisticated risk measurement tools have evolved to help financial institutions measure market risk (value-at-risk measurement tools), credit risk (expected and unexpected loss measurement tools), and insurance risk (dynamic financial analysis tools). There also have been advances in using these evolving risk metrics to help guide executive management in their strategic decision-making. The framework through which this is accomplished typically has two parts:
1. Risk is related to the amount of capital the firm requires to achieve a sufficient level of protection against adverse circumstances.
2. Risk is used to adjust the returns from business activities to determine whether activities are value adding or value destroying.
The first part should reflect a debtholder's perspective on risk (Is there sufficient capital to cover "worst case" risks?). The second part should reflect a shareholder's perspective on risk (Are we getting a sufficient return for the systematic risk being taken?).
The debtholder and shareholder views of risk differ but are related. Actions that tend to increase risk for debtholders also tend to increase risk for shareholders. However, in important ways, the views diverge. For example, debtholders value risk diversification at every. level, while shareholders do not value diversification they can replicate on their own. In short, these two constituencies can draw on the same underlying risk measurement framework, but will view the resulting risk measures through different lenses.
A Conceptual Framework for Attributing Capital
The first step in attributing capital is developing a theoretical framework for relating risk to the amount of capital a financial institution needs to hold. Many financial institutions have developed a framework based on Merton's model of default(1). This model effectively assumes the following:
* Shareholders own the right to default on debtholders and will do so if the value of the firm's equity (or "net assets") drops to zero.
* Debtholders charge shareholders for default risk by demanding a spread over the risk-free rate on the funds they provide.
* The probability of default is a function of the firm's net asset value distribution and its current net asset value
The last element of the above approach can be inverted. That is, a theoretically robust estimate of required capital can be determined if the net asset value distribution is known and a probability of default (or solvency standard) is selected.
Most financial institutions have developed methods for relating risk to capital which, on the surface, look similar to the framework described above. However, there are several fairly common flaws:
Standard deviation, rather than confidence interval. Many proposed risk measurement frameworks rely exclusively on a set "multiple of standard deviation" as the solvency standard. While this would work if all risks were normal, it introduces serious distortions for non-normal risks, such as credit risk and insurance risk. For example, a 3 standard deviation loss in foreign exchange trading could have the same probability as a 6 standard deviation loss in commercial lending. The solvency standard for linking risk to capital should instead be set as a common confidence internal on the value distribution.
Accounting earnings, rather than value. Risk is often modeled as volatility of accounting earnings. Hence, the distribution of asset value is approximated by a distribution of accounting earnings. This can seriously misstate the true economic risk to the firm. For example, the balance sheet interest rate risk in many U.S. thrifts during the 1980s, if measured by [Delta]NII, would have looked manageable, but if measured on [Delta]NPV bais, would have looked catastrophic. The latter view was correct but was obscured by the one-year accounting …