This study is on the effects of United States (US) Statement of Financial Accounting Standards number 133 (SFAS 133), Accounting for Derivatives Instruments and Hedging Activities, which was introduced in 2001. The first area of investigation focuses on the impact of SFAS 133 on income smoothing through the application of derivatives and discretionary accruals. The second key aspect is how SFAS 133 influences the disclosure of derivatives related information. Addressing these two issues together, provides an integrated understanding of how accounting policy can affect both risk management and risk disclosure choices.The study comprises a detailed literature review of extant empirical and analytical studies. It primarily extends the work of studies that initially looked at derivatives and discretionary accruals as income smoothing substitutes, such as Barton (2001) and Rajgopal and Pincus (2002). The theoretical framework developed during the literature review, discusses the genesis and key features of SFAS 133, and the determinants of the two key income smoothing choices of derivatives use and discretionary accruals. These include capital markets, managerial risk and corporate governance determinants. In addition, the theoretical framework outlines how SFAS 133 fair value recognition requirements can influence disclosure of related information through the footnotes. It presents the argument that the extent to which notes are complementary to recognition and measurement requirements should outweigh the extent to which they may be considered substitutes. It further describes the literature on disclosure incentives, including capital markets, proprietary cost concerns, managerial talent, managerial risk incentives based on compensation and litigation cost. This is as a precursor to the univariate and multivariate empirical testing of 1999 to 2003 data from 253 US firms (i.e. 850 firm-year observations).In conducting the empirical testing, I endeavour to address the problems of model endogeneity that could arise due to derivatives use and discretionary accruals being determined jointly. I also address the individual effects that arise due to the application of panel data. A key empirical finding is that after adopting SFAS 133, corporate managers increase discretionary accruals. However, there is no conclusive multivariate evidence of SFAS 133 reducing derivatives use, as hypothesized. An additional finding is that there is a substitution relationship between derivatives use and discretionary accruals. However, I also find that SFAS 133 adoption weakens the extent to which accruals influences derivatives use, but not the other way round. This latter finding suggests a partial substitution relationship exists after SFAS 133 and lends itself to a number of plausible explanations. These include accruals being complements rather than substitutes to derivatives use, after SFAS 133. SFAS 133 could trigger either increased earnings volatility or the greater use of speculative derivatives. Either of these could then induce the increased use of accruals in a manner that confounds the income smoothing substitution relationship.Despite the substitution relationship, the use of derivatives to smooth income is more akin to economic reality, as derivatives use also influences cash flow and fundamental economic volatility. Thus the finding that managers increase their use of accruals in general, after SFAS 133, suggests that SFAS 133 adoption, results in choices that are less beneficial to shareholders. I come to this conclusion based on the empirical evidence of Huang, Deis, Zhang and Moffit (2009). Their study shows that for income smoothing purposes, derivatives enhance shareholder value to a greater extent than the use of accruals. Further to the study of its impact on income smoothing, the empirical findings of SFAS 133 on derivatives related disclosure, build a collective picture of how different reporting practices can be influenced by accounting policy. The study finds a significant positive association between SFAS 133 and derivative information disclosure index, in all the models. This suggests that the derivatives recognition and measurement requirements have encouraged the provision of greater disclosures. The results also show a significant positive association of capital markets incentives measured as the logarithm of trading volume. However, there is no evidence of association of proprietary costs. The results further show that auditor expertise and the level of derivatives use are positively associated with observed derivatives disclosure levels. In the same vein, litigation risk and discretionary accrual levels have a negative association.In sum, this study shows that SFAS 133 adoption has potentially adverse consequences on income smoothing choices, but at the same time it has positive consequences through its encouraging disclosures that lower the information asymmetry on risk exposures.
