Abstract

Derivatives recognition and hedge-accounting treatment: an empirical study of the rules prescribed by SFAS 133 and some alternatives

Read the full thesis (PDF).

Over the last decade, accountants have faced the increasingly important issue of accounting for derivative securities. SFAS 133, the recently adopted U.S. recognition standard for derivative securities, mandates that all derivatives be accounted for in financial statements at fair value. Accounting rules for hedges of forecasted transactions under SFAS 133 are not consistent with the matching principle: only the derivatives side of the hedge is marked-to-market.

Dechow et al. (1999) argue that value relevance is a necessary, but not a sufficient, condition for the financial-statement recognition of an element of information. In this thesis, it is hypothesized that, when considering whether to recognize derivatives used to hedge forecasted transactions, the matching principle is another necessary condition. Accounting for only the derivatives adds noise to the accounting numbers. SFAS 133 should produce less informative accounting numbers than those based on the matching principle.

The thesis empirical specifications are based on Barth (1994) and Ohlson (1995). Panel samples of 22 gold mining firms between 1992 and 1997 are used to test the thesis hypotheses. Since SFAS 133 does not become mandatory before 2001, "as if" SFAS 133 accounting numbers are generated using voluntary disclosure of derivative securities and no-arbitrage valuation formulas. Two sets of matched accounting numbers are considered as alternatives to SFAS 133 accounting numbers. First, "as if" marked-to-market accounting numbers, which mark-to-market both sides of the hedge, are computed from firm derivatives and reserves disclosures. Reported historical-cost accounting numbers, which keep both sides of the hedge off the balance sheet, are the second alternative.

Results indicate that the two sets of matched balance-sheet numbers show more explanatory power for common-equity market values than the SFAS 133 accounting numbers. Results also indicate that marking both sides of the hedge to market generates accounting numbers that are more informative than keeping both sides of the hedge off the balance sheet. The income statement evidence is mostly inconclusive. The transitory nature of earnings in the gold mining industry is a plausible explanation for these latter results.

The thesis evidence answers the call of standard setters for timely evidence relevant for standard setting decisions. The methodology developed to estimate the fair value of derivatives might also prove useful to CFOs and auditors who will have to comply with SFAS 133. Finally, the thesis contributes to the recognition literature, discussed in Dechow et al. (1999) by indicating that the matching principle appears to be a necessary condition for balance-sheet recognition of derivative securities used to hedge forecasted transactions.