شواهد بیشتری از مدیریت سود و رفتار های فرصت طلب با استانداردهای حسابداری مبتنی بر اصول: مورد تعهدات بازنشستگی دارایی مشروط

FASB Interpretation No. 47 (FIN 47) clarifies the diverse accounting practices for conditional asset retirement obligations (CAROs) that arose under SFAS No. 143, which is classified as a principles-based standard by the SEC. Prior research suggests that the subjectivity in SFAS No. 143 provides management with the opportunity to manage earnings and avoid the recognition of CAROs. This study examines firms that recorded adjustments for CAROs upon FIN 47 adoption. We demonstrate that effective monitoring is essential to promote adherence with principles-based standards, and that gatekeepers may not be effective when standards are ambiguous. Univariate tests and logistic regressions reveal that FIN 47 adopters have audit committees with a greater number of financial experts and are audited by BIG 4 firms. Particular firm-specific factors are also found to be associated with the adoption decision. The results also indicate that newly-reported obligations related to asbestos in firm-owned property and restoration costs for leased premises were subject to prior management discretion. This study extends existing literature on SFAS No. 143 and FIN 47 and studies examining earnings management with principles-based standards. The case of FIN 47 provides further evidence that significant opportunities for earnings management and discretion exist within a principles-based accounting environment, particularly when standards lack clarity. It also confirms the critical role of monitoring by the audit committee and external auditors to promote adherence with the substance of such standards

SFAS No. 143, Accounting for Asset Retirement Obligations, establishes standards related to the measurement and recognition of asset retirement obligations (AROs) and requires firms to recognize the fair value of such liabilities in the period incurred. The accounting for AROs requires a high degree of judgment by management and external auditors because these obligations often pertain to activities that will occur several decades in the future. After SFAS No. 143 became effective in 2003, the FASB detected that many entities were not properly accounting for “conditional” AROs (i.e., CAROs), which result when the timing and/or method of settling the obligation is dependent on a future event that may not necessarily be controllable by the entity. In response, the FASB issued Interpretation No. 47 (FIN 47), Accounting for Conditional Asset Retirement Obligations, an interpretation of FASB Statement No. 143, which became effective as of December 31, 2005 for calendar year companies. The objective of FIN 47 is to “clarify” and reiterate several aspects about the accounting for CAROs already set forth in SFAS No. 143. At the time, the FASB considered FIN 47 as an “education effort” (FASB, 2004b) and expected that adding clarity to the existing accounting guidance would assist senior management (and their auditors) implement the provisions of SFAS No. 143 as originally intended. The issuance of FIN 47 was deemed particularly important by the FASB given evidence of an unanticipated degree of subjectivity and discretion in the scope, measurement and recognition provisions of SFAS No. 143 as to CAROs. This discretion provided management with opportunities to avoid the recognition of existing obligations and associated income-decreasing accruals. The transparency provided by FIN 47 resulted in the recognition of unrecorded CAROs and cumulative-effect adjustments to income upon adoption. In prior research on the accounting for AROs, Jordan et al. (2007) posit that SFAS No. 143 provides management with the opportunity to manage earnings given the inherent subjectivity in determining the fair value of the obligations and the discount rates used to recognize annual accretion expense on the amount of the liability. Ely and Stanny (2007) examine a sample of firms that adopted FIN 47 and find that highly-levered firms were less likely to have complied with the principles in SFAS No. 143. They conclude that this provides evidence of earnings management in an attempt to avoid the violation of loan covenants. The debate and dialogue over the benefits and consequences of principles-based versus rules-based accounting standards has ensued for some time, particularly since the accounting scandals that prompted passage of The Sarbanes Oxley Act (SOX) and the increasing complexity of financial reporting in general. In a study pursuant to SOX, the SEC (2003a) expressed its preference for principles-based over rules-based standards. In general, principles-based standards are characterized by a concise statement of the accounting objective, few if any scope exceptions, an “appropriate” amount of implementation guidance, lack of bright-line tests, and consistency with a coherent conceptual framework (SEC, 2003a and FASB, 2004c). Principles-based standards also impose a greater responsibility on management and the auditors to report the underlying economic substance of transactions. Accordingly, effective gatekeepers such as vigilant audit committees and independent external auditors, whose roles were strengthened under SOX, are necessary to ensure that accounting standards are interpreted and applied in accordance with the substance and intent of the underlying principles (SEC, 2003a). Unfortunately, as demonstrated in this study, the effectiveness of gatekeepers can be significantly diminished when principles-based standards are ambiguous. In the SEC study, SFAS No. 143 was specifically classified as a principles-based standard (SEC, 2003a). Despite the expectations by many that principles-based standards will improve financial reporting compared to a rules-based system, Jamal et al. (2010, p. 143) contend that there is “virtually no empirical evidence available to support these beliefs.” They caution that standards-setters have already reduced their emphasis on verifiability as a key quality during the development of standards. Coupled with the shift toward the inherent subjectivity of fair value measurements, this “significantly impairs the ability… to limit opportunistic actions by management” and the auditor’s ability to detect them. In fact, several studies, (e.g., Nelson et al., 2002, Nelson, 2003 and Van Beest, 2009, etc.) find that significant opportunities for earnings management exist in a principles-based accounting environment. Given this background, we examine the consequences of the ambiguity in SFAS No. 143 and the subsequent issuance of FIN 47 which provided clarifying language to promote compliance with the original intent of the standard. We then examine the effect of governance (and other factors) on FIN 47 adoption in light of this clarity. We believe that the case of FIN 47 provides strong evidence that a principles-based standard with excessive ambiguity provides ample opportunity for discretionary accounting behavior for which governance mechanisms may not be effective. However, gatekeepers can be effective when such standards provide the appropriate degree of clarity to permit the exercise of professional judgment. More specifically, this study has two main objectives. The evidence of earnings management (noted above by Nelson et al., 2002 and others) and the lack of adherence to the measurement and recognition guidelines in SFAS No. 143 (Jordan et al., 2007 and Ely and Stanny, 2007) demonstrate inherent weaknesses with principles-based standards that rely heavily on the professional judgment of management and auditors. Building on existing research, the first objective of this study is to examine the characteristics of firms most impacted by the adoption of FIN 47. To accomplish this objective, we first investigate the relevant factors surrounding the issuance of FIN 47, which reveals that: (1) SFAS No. 143 contained conflicting provisions and was overly ambiguous to be effective as a principles-based standard. Essentially, it permitted excessive discretion and lacked the clarity to ensure the exercise of sound judgment by management and their auditors. It also suggests that governance mechanisms were not sufficiently effective to compensate for principles-based standards permitting excessive discretion; (2) FIN 47 provided clarifying language and reduced the level of discretion inherent in SFAS No. 143. It did not amend the scope, measurement or recognition principles as to CAROs and can also be considered as a principles-based standard. In effect, FIN 47 can be viewed as a second implementation date for SFAS No. 143; and (3) the adoption of FIN 47 permits another evaluation of the influence of gatekeepers and other factors in ensuring adherence with the underlying substance of principles-based standards. The second objective of this study is to identify the nature of the obligations that were not properly reported under SFAS No. 143. This information should be instrumental in evaluating the FASB’s success with the objectives and expected benefits of FIN 47, and provides insights into the exposures that were the subject of accounting discretion when SFAS No. 143 was issued. To accomplish these objectives, the characteristics of firms that reported a cumulative-effect adjustment upon the adoption of FIN 47 are compared against a matched sample of control firms of equivalent size and industry characteristics that did not report such adjustments. Broadly speaking, Kanagaretnam et al. (2007) suggest that effective governance mechanisms can enhance the reliability and credibility of financial information and can mitigate agency problems posed by self-serving decisions by management. We examine the influences of particular monitoring mechanisms (e.g., audit committee characteristics, audit quality, equity ownership), the incentives of senior management, and other firm-specific influences on the adoption decision. Consistent with expectations, the results show significant differences between the characteristics of FIN 47 firms and those in the matched sample. Univariate tests reveal that the audit committees of FIN 47 firms have higher representation by financial experts and meet more frequently. The results also show that FIN 47 firms are more likely to be audited by Big 4 firms. These results suggest the positive influence of monitoring and demonstrate the need for effective gatekeepers to ensure adherence to the judgmental accounting issues in a principles-based standard. FIN 47 firms were more likely to have an operating loss in the year of adoption, suggesting that current operating results may have influenced management’s decision. CEOs of FIN 47 firms also have a lower proportion of the first lag of cash compensation (salary and annual bonus) to total compensation. Other firm-specific factors are also associated with the adoption decision. The logistic regression results are consistent with several of these predictions. Additional tests reveal the CAROs that were not initially recognized under SFAS No. 143. An examination of the footnote disclosures reveals that CAROs related to (1) future asbestos abatement or removal and (2) lessee obligations to restore the condition of rental property at the end of the lease were the key types of newly-reported obligations. Logistic regressions report strong associations between these obligations and audit committee expertise, audits by particular Big 4 firms, and other firm-specific attributes. This suggests that the FASB was successful in providing clarity for firms to adhere to the underlying principles in SFAS No. 143 and report unrecognized CAROs, particularly those related to asbestos and leases. This paper contributes to the existing literature on SFAS No. 143 and FIN 47. Studies by Jordan et al. (2007) and Ely and Stanny (2007) both reveal evidence of earnings management under SFAS No. 143. This paper extends those findings by examining the role of FIN 47 in reducing discretion and serving as a second implementation attempt for CAROs. We then empirically consider the critical role of governance mechanisms and other factors on adherence to these principles-based standards. It also identifies the exposures that were the subject of significant managerial discretion under SFAS No. 143. This study also extends the literature on earnings management with principles-based standards (e.g., Nelson et al., 2002, etc.) which have relied primarily on surveys and experimental studies. That stream of research is broadened by: (1) examining the causes and consequences of managements exercise of discretion using a principles-based standard, (2) revealing the FASB’s response to the implementation problems with SFAS No. 143, and (3) examining the factors that influenced compliance with the FASB’s original intent by issuing FIN 47. Additionally, accounting for asbestos, chemical-clean-ups and related commitments associated with asset retirements impact a broad array of industries and are interrelated with broader issues in the area of environmental accounting. The findings in this study may provide information useful to standards-setters in arriving at future accounting guidance in this area. The remainder of this paper is organized as follows. Section 2 provides background on the accounting issues related to SFAS No. 143 and FIN 47. It also discusses extent research on earnings management in a principles-based accounting environment and literature on AROs and CAROs. Section 3 develops the empirical predictions while Section 4 describes the research methodology. Section 5 presents the empirical results. The summary, limitations and suggestions for further research are provided in Section 6.

5.1. Descriptive statistics, univariate tests and correlations Table 4 presents descriptive statistics of the explanatory variables for (1) the total sample; (2) the 156 FIN 47 firms, and (3) the 156 non-FIN 47 firms. Univariate tests of the differences in means are also presented and indicate significant differences between the two groups. Concerning corporate governance, FIN 47 firms have on average a greater number of financial experts on their audit committees and more frequent audit committee meetings, both statistically significant at p < 0.01 and p < 0.05, respectively (one-tailed). This supports the prediction concerning the positive relationship between audit committee characteristics and FIN 47 implementation. BIG 4 firms were also more likely to audit the FIN 47 firms than smaller auditing firms (97% versus 93%, respectively), which is marginally significant at p < 0.10 (one-tailed). However, the association among the individual Big 4 firms provides more contrast. There is a positive and statistically significant difference between the two samples for audits by DELOITTE (36% versus 24%, respectively) and KPMG (26% versus 15%, respectively), both of which are statistically significant at p < 0.05 (two-tailed). Conversely, ERNST was less likely to audit FIN 47 firms (7% versus 26%, respectively), which is statistically significant at p < 0.01 (two-tailed). PWC audited substantially the same percentages of firms in both samples (27% and 28%, respectively). The results suggest that the Big 4 firm may help explain the likelihood of FIN 47 adjustments. 12No differences were noted between FIN 47 and non-FIN 47 firms for the variables BLOCK, INSIDER% and CFO_CPA. However, the FIN 47 sample has a higher percentage of firms reporting a loss from continuing operations (LOSS) for the year than non-Fin 47 firms (23% versus 13%, respectively) which is statistically significant at p < 0.05 (one-tailed), suggesting that firm performance is associated with the adoption decision. CEOs of FIN 47 firms also had a lower proportion of lagged cash to total compensation (CASH_COMP) than the matched sample which is statistically significant at p < 0.05 (two-tailed), suggesting their longer time horizon and willingness to absorb additional adjustments in the current year. The variable RENT% is positive and marginally significant at p < 0.10 (one-tailed), providing some evidence that FIN 47 firms have a higher proportion of operating leases than non-FIN 47 firms. 13 The correlation matrix in Table 5 reveals low to moderate correlations among the independent variables. The highest correlation is between REG and SFAS143 (0.52). A review of the Variance Inflation Factor for each variable reveals no significant multicollinearity.5.2. Logistic regression results Table 6 presents the coefficient estimates of regressing FIN 47 adoption against the explanatory variables. Model 1 employs a single indicator variable representing the existence of a BIG 4 auditor, while the Expanded Model uses an indicator variable for each of the Big 4 firms.As expected, the corporate governance variable AC_EXPERT is positive and significant at p < 0.01 (one-tailed) in both models, confirming that greater financial expertise of audit committee members and the related monitoring benefits are associated with firms making the FIN 47 adoption decision. This is consistent with the univariate results and is perhaps due to their deeper understanding and involvement with complex accounting issues, coupled with the clarity provided by FIN 47. However, audit committee activity (AC_MEET) is not significant. The coefficient for BIG4 in Model 1 is positive and significant at p < 0.05 (one-tailed), confirming the expectation that the perceived higher quality of the largest auditing firms is likely to be associated with firms reporting FIN 47 adjustments. 14 The Expanded Model shows the impact relating to the individual Big 4 firms. Consistent with the univariate tests, audits of FIN 47 firms were more likely to be performed by DELOITTE and KPMG, which are both statistically significant at p < 0.05 (two-tailed). The coefficient for ERNST is not statistically significant, whereas PWC is positive and marginally significant at p < 0.10 (two-tailed). As noted previously, these results suggest that particular characteristics of the individual auditing firms may be an explanatory factor in their clients’ decision to adopt FIN 47. Also consistent with the univariate results, the coefficient for LOSS is positive and statistically significant at p < 0.05 in Model 1 and p < 0.01 (one tailed) in the Expanded Model. This suggests that firms experiencing poor operating results may be more willing to report the cumulative effect adjustment arising from FIN 47 adoption. The coefficient for lagged CASH_COMP is negative and significant at p < 0.05 (two-tailed) in both models. The coefficients of the ownership structure variables BLOCK and INSIDER% are not statistically significant, suggesting that concentration of equity holdings by blockholders and management had no influence on the FIN 47 adoption decision.15 The coefficient for SFAS143 is positive as expected and marginally significant at p < 0.10 (one-tailed) in both models providing some evidence that prior experience with the accounting for AROs was a factor in the adoption decision. The coefficients of the remaining firm-specific variables CFO_CPA, LEV, REG, 16 LN_PPE and RENT% are not statistically significant and therefore had no influence on the FIN 47 decision. Both models are statistically significant at p < 0.01. 5.3. Additional insights into the types of CAROs recognized under FIN 47 An additional objective of this paper is to examine the types of CAROs reported as cumulative-effect adjustments upon the adoption of FIN 47 and to identify significant relationships with the independent variables. This provides additional insights into areas of complexity under SFAS No. 143 and where possible opportunistic decisions were made by management. It also reveals contentious areas that were successfully clarified by the FASB in FIN 47 and extends the findings by Jordan et al. (2007) and Ely and Stanny (2007) in this area. We reviewed the footnote disclosures for each of the 156 FIN 47 firms and found that the primary causes of the cumulative effect adjustments were related to obligations for asbestos (49 firms) and the future restoration of leased premises (46 firms). The remaining 61 firms reported a variety of causes or were generic in their descriptions. Using firms in the Asbestos and Lease classifications as binary dependent variables (and the respective control firms), two logistic regressions were run using a total sample size of 98 firms for the Asbestos model and 92 firms for the Lease model. 17 Directional predictions for the independent variables are the same as those in Table 6, and variables for the individual Big 4 firms are used. Table 7 displays the test results.