Applications Of Risk Measures

The risk measures used by an organization will depend on the types of risks it is exposed to, the regulations that govern it, and whether the organization uses leverage. For each type of organization, differences among firms will result in differences in the risk measures used. In what follows, we focus on similarities of risk measures used among organizations of the same type and typical differences between the risk measures used by different organizations.

Banks typically use sensitivity measures (duration of held to maturity securities and foreign exchange risk exposure), scenario analysis and stress testing (for their full balance sheets), leverage risk measures, and VaR (especially for trading securities). Banks also estimate risk from asset liability mismatches, estimate VaR for economic capital, and disaggregate risk by both geographic location and business unit type.

Traditional (long only) asset managers typically focus on relative risk measures unless their goal is an absolute return target. Typical risk measures used include the size of positions, sensitivity measures of interest rate and market risk, historical and hypothetical scenario analysis, and options risk. A risk measure more specific to asset management is active share: the difference between the weight of a security in the portfolio and its weight in the benchmark.

Ex post tracking error (backward looking) and ex ante tracking error (forward looking) measures provide different information. Ex post tracking error is a measure of a portfolio’s tracking error relative to a benchmark portfolio over a lookback period. Ex post tracking error is used for performance attribution and to assess manager skill over prior periods.

Traditional asset managers mostly use ex ante tracking error for risk estimation, which focuses on the potential underperformance of the current (rather than a historical) portfolio. Managers with an absolute return target may use VaR instead.

For hedge funds, the risk measures used depend, to some extent, on the strategy employed. For hedge funds in general, the risk measures used include sensitivity analysis, leverage measures, scenario analysis, and stress tests. Funds with both long and short positions will estimate risk measures for long positions and short positions, as well as for the overall portfolio (gross exposure).

Hedge funds that use VaR focus on VaR measures of less than 10% for short periods.

Hedge funds with significantly non normal returns distributions use a risk measure referred to as maximum drawdown: the largest decrease in value over prior periods of a specific length. As we have noted, sensitivity measures based on standard deviation or beta may be misleading for large changes in risk factors when returns are non normal.

Defined benefit pension funds calculate the difference between the present value of their assets (often market values) and the present value of their estimated future liabilities (payments to retirees and heirs). A risk measure used by pension funds is surplus at risk, a VaR for plan assets minus liabilities. A negative surplus must be made up by the firm if higher than expected asset returns do not reduce it significantly over time.

The term glide path refers to a multi year plan for adjusting pension fund contributions to reverse a significant overfunded or underfunded status. To reduce surplus uncertainly, a pension fund may match its assets to its liabilities. A related risk measure is an estimate of the hedged exposure and unhedged (returns generating) exposure of the fund.

Insurance companies are often subject to significant regulation of their products and their investment portfolios (reserves). Property and casualty insurers sell auto, home, boat, liability, and health insurance. The insurance risks of a P&C company are not highly correlated with the market risk of their investment portfolios.

Insurance risks are reduced by purchasing reinsurance (from another insurance company) and by geographical diversification. Life insurers primarily sell life insurance policies and annuities, some of which make payments until the death of the annuity owner.

P&C insurers use sensitivities of their exposures to market risk factors in their investment portfolios for risk management. Premium income is expected to cover the cost of insurance claims in a typical year, with the investment portfolio available to cover extraordinary claim losses, such as those in a year with a natural disaster.

P&C insurers use VaR and capital at risk as measures of their risk exposure in their investment accounts. They also use scenario analysis, often combining portfolio risk factors and insurance risk factors in a scenario. Regulations may require specific amounts of reserves (based on policies issued), and regulators discount the values of riskier assets held as reserves in determining their adequacy.

The insurance risk of life insurers is more highly correlated with the market risk exposures of their investment portfolios than it is for P&C insurers. Because annuities pay over relatively long periods into the future, the present values of these liabilities are quite sensitive to the discount factors used, although they have significant mortality risk factors as well. (The longer a life annuity pays, the larger the current liability.)

For this reason, life insurers estimate the sensitivities to market risk factors for both their investment portfolios and their annuity liabilities.

Because life insurers are able to somewhat match the market risk of their portfolio assets to their liabilities, they must consider the risk of the remaining mismatch between assets and liabilities. Life insurers also use scenario analysis that includes both nonmarket (insurance) risk factors and market risk factors.