SEC Market-Risk Disclosures: Enhancing Comparability

In 1997, the U.S. SEC mandated through Financial Reporting Release No. 48 (FRR48) the disclosure of forward-looking market risk information. Because of the recency of the required risk disclosures, little has been written about them or about how analysts might use them. FRR48 allows three disclosure formats—sensitivity measures, value at risk, and a tabular format. The issue is that variations among the disclosure formats and the discretion allowed about assumptions underlying sensitivity and VAR measures may impair analysts' use of the disclosures. We demonstrate how sensitivity and VAR measures can be derived from the tabular format. Our methodology allows financial analysts to derive risk measures based on consistent assumptions among companies. Tabular data provide a common denominator by which companies may be compared and provide a means of overcoming the limitations of sensitivity and VAR measures. In 1997, the U.S. SEC issued Financial Reporting Release No. 48 (FRR48) requiring companies to disclose qualitative and quantitative market-risk information. The release defines market risk as the risk of loss arising from adverse changes in market rates and prices, such as interest rates, foreign currency exchange rates, commodity prices, and other relevant market rates. Companies may disclose market risk using one of three presentation formats—sensitivities, value at risk (VAR), or a tabular format.We describe the FRR48 disclosure requirements and discuss obstacles analysts may encounter in using them. We demonstrate how comparable and reliable sensitivity and VAR measures can be derived from the tabular format.Sensitivities are measures of the potential losses arising from hypothetical changes in market rates. FRR48 permits companies to choose any reasonably possible near-term rate change to use in calculating sensitivity. They may measure loss in terms of fair value, cash flow, or income. Companies may report the loss they consider most informative if all three are material.Value at risk is the largest loss a company could experience from its market-risk-sensitive instruments in a given holding period under normal market conditions with a specified probability (typically no greater than 5 percent). Companies have leeway in selecting VAR models and assumptions. As with sensitivity, FRR48 permits companies to measure VAR in terms of cash flows, earnings, or fair value. Companies must report backtesting results—the number of times during the year that actual daily losses exceeded the reported VAR or the range of actual daily losses over the year.In the tabular format, risk is not explicitly quantified. Rather, a table lists the attributes of market-sensitive assets and liabilities (including instrument type, book value, maturity, average rate, strike price, and fair value) according to exercise date or time-to-maturity. The tabular format is intended to provide sufficient information to determine the cash flows from the company's financial instruments—presumably, to allow analysts to compute summary risk measures.FRR48 does not require risk disclosures for commodity positions or derivative contracts that settle in “other than cash,” but companies may choose to include such positions in their reported sensitivities or VARs. Consequently, what instruments or positions are covered by management-reported risk measures is not always obvious.FRR48 requires a point estimate for the sensitivity or VAR measure. Analysts, however, typically consider a broad range of outcomes. Analysts cannot easily convert sensitivity disclosures to reflect different hypothetical rate changes. For example, if a company provides a 10 percent sensitivity when analysts believe that a 30 percent change is likely, tripling the number will provide a biased estimate because expected losses are not always linear in rate changes—especially for companies with significant option positions. Few companies report multiple loss sensitivities or VARs, and even fewer companies report gains.Analysts cannot directly compare disclosures in the same format because companies select their own hypothetical rate changes, time horizons, outcome likelihoods, and modeling techniques. Analysts cannot accurately assess a company's models and assumptions because disclosures are cryptic. Furthermore, companies are permitted to define loss in terms of fair values, net income, or cash flow, each of which measures a fundamentally different type of risk.Even more difficult—in fact, inappropriate—is comparing FRR48 disclosures among formats. Sensitivities and VARs are inherently dissimilar risk constructs; directly comparing them will result in erroneous risk assessments. Therefore, analysts need a way to calculate sensitivities and VARs from FRR48 tabular data and publicly available market rates. We demonstrate such a method. Our methodology allows analysts to generate a range (or distribution) of risk measures that reflects consistent assumptions and models.We provide an example of the conversion methodology that uses actual FRR48 disclosures. Although interest rate risk is the focus of the example (because it is the most common form of market risk), the methodology is equally applicable to foreign currency exposure, equity-price risk, and commodity risk. We also discuss applying the method to a number of more complex issues, including derivative positions.