2.4 DiscussionThe above studies exa

2.4 Discussion

The above studies examine accounting quality from the perspective of stock market investors. Many other contracting parties use accounting standards in different ways, however. For example, debt contracts normally back out the amount of goodwill from the balance sheet because goodwill does not exist during bankruptcy (Holthausen and Watts, 2001). Because they




21

must respond to the financial reporting demands from many parties in the economy, standard setters focus on a wide range of contracting parties including banks, regulators, tax bureaus, and stockholders. Value relevance tests using stock prices as a benchmark impose a narrow focus on how information is reflected into stock market investors’ expectation. With a variety of demands for financial reporting from parties other than stock market investors, value relevance tests may thus be less relevant to the goal of standard setters and the objectives of financial reporting (Holthausen and Watts, 2001). This could be one reason that results from the studies on change in accounting quality are largely mixed. Focusing on value relevance may underestimate the potential benefit of the change in accounting standards. Although a full review of the definition of the quality of financial reporting is outside of the scope of this paper, we adopt a more complete view of the usefulness of accounting to all relevant parties in the economy. This is in alignment with the TFV concept and is consistent with the FASB’s statement that accounting must be useful to all types of contracting parties to facilitate investment and credit decisions. Future studies could review the effect of IFRS on other contracting parties such as banks and regulated industries. Without study of the effect of the adoption of IFRS on different contracting parties, we will not fully understand the economic consequences of the harmonization.

The mixed results regarding IFRS adoption could also be related to methodological issues. Foremost is sample selection bias. Prior studies have necessarily included only voluntary adopters. These firms may have innate characteristics that affect their adoption decision in addition to the hypothesized economic consequences. This self-selection problem may bias either for or against finding any results. On one hand, firms whose accounting methods in their national GAAP are closer to IFRS may be more willing to adopt IFRS due to




22

lower adoption cost. These firms may be closer to IFRS firms than the non-adoption firm (i.e. control sample) even before adoption. The change in economic consequences after IFRS adoption may not be significantly large relative to the difference between adopting and non-adopting firms before adoption, resulting in a low-power test. On the other hand, firms choosing IFRS may expect a large benefit, such as facilitation of stock issuance in an international stock exchange, which may not be representative of the benefit of adoption for all firms in the economy. Studies generally address this self-selection problem using a two-stage-least-squares regression. In the first stage, a firm’s choice of accounting standards should be predicted by factors such as legal origin, cross-listing, leverage, ownership, operating cycle, international exposure, and industry affiliation. We expect that a more complete understanding of firms’ choice of accounting method and disclosure should enhance research design (Ashbaugh, 2001).

The second methodological issue is an omitted variables problem. Pricing mechanisms and the information environment, including analyst following, media coverage, and disclosure of non-financial information, differ across firms and countries (Bushman et al., 2004). Thus, comparing financial statements between two accounting standards for the same firm such as in Hung and Subramanyam (2007) is likely a better setting, as it reduces the omitted variables problem. Value relevance studies using stock price as a parsimonious measure to capture all public information in the market may have problems because stock prices may incorporate information in a different manner across countries (Morck et al., 2000). Therefore, cross-country studies should consider using fixed effects models to control for observed and unobserved country-level factors that may affect the economic consequences of accounting standards (Sun, 2007).




23

Finally, regression models in some studies that compare different accounting regimes (e.g., before and after mandatory adoption of IFRS, or comparing U.S. GAAP with IFRS) are mis-specified and thus make it difficult to compare different accounting standards. For example, some studies do not consider the non-linear relation between price and net income (e.g., Basu 1997). If accounting standards differ in likelihood of large losses or conservatism,14 treating positive and negative earnings the same in a price regression will result in a different R-square across standards. Furthermore, Ashbaugh and Olsson (2002) find that violation of clean surplus makes regressions using the Ohlson (1995) model mis-specified. One example of such a violation is asset revaluation that credits shareholders’ equity. Ashbaugh and Olsson (2002) find that more than half of the IAS sample firms trading in the Stock Exchange Automated Quotations (SEAQ) International Equity Market of London revalue their assets upward.


3. DETERMINANTS OF ACCOUNTING QUALITY AFTER IFRS ADOPTION

Figure 1 depicts a schematic framework describing determinants of accounting quality. It shows that accounting standards, legal and political systems, and incentives of financial reporting all affect accounting quality.15 Although conversion to IFRS is likely to affect financial reporting, it is only one of the determinants of overall accounting quality. Because other determinants will continue to differ across countries, it is possible that accounting quality will continue to differ across countries following IFRS adoption.