اندازه گیری همگرایی استانداردهای ملی حسابداری با استانداردهای بین المللی گزارشگری مالی

This paper analyses three methods for measuring the success achieved in effecting convergence between any two sets of accounting standards. We begin by reviewing a measurement method based on the concept of Euclidean distances. We then propose two better measures (involving Jaccard's coefficients and Spearman's coefficients) to assess the progress of National Accounting Standards setting bodies in converging their standards with International Financial Reporting Standards [IFRS]. For illustrative purposes, we measure the convergence of National Accounting Standards in Portugal with International Accounting Standards [IAS] and IFRS over the period 1977–2003.

This paper explores sophisticated measures for assessing and comparing the success achieved in converging National Accounting Standards with internationally-prescribed sets of accounting standards (such as those comprising IFRS). 1 The development of sophisticated measures of convergence is important for two principal reasons. First, not all countries have committed to adopting IFRS. For example, Iceland, Japan and Saudi Arabia are reported by the International Forum on Accountancy Development [IFAD] (2003) to have not yet expressed an intention to converge with IFRS. Some other countries (e.g. New Zealand) have opted to converge with IFRS over a longer time period (by 2007). Indeed, IFAD (2004) has reported also that as at December 3, 2004, IFRS were ‘not permitted for domestic listed companies’ in 36 countries (including Argentina, Brazil, Canada, Chile, Fiji, India, Indonesia, Mexico, Philippines Taiwan, Tunisia, United States and Vietnam). Consequently, for such countries, there are likely to be information benefits in measuring and monitoring the extent to which National Accounting Standards approximate IFRS. There are also likely to be benefits for capital markets and other users of financial statements in helping to assess the quality and comparability of published accounting data in those countries. Measurement of convergence is important also because, in some EU countries that have adopted IFRS in 2005, the application of IFRS does not extend to all entities, but is confined to listed companies. The accounting standards that are to apply to non-listed companies are being debated in such countries, and an important issue is whether IFRS will affect the accounts of non-listed companies. Street and Larson (2005, p.1) conclude that ‘most EU members do not plan to converge national GAAP with IFRS, thereby highlighting … concerns regarding the emergence of a “two-standard” system in the EU’. The main barriers to convergence identified by the survey are the link between financial accounting standards and tax accounting; and disagreements about the complicated nature of certain IFRS, especially those associated with ‘fair value’ accounting (see Street & Larson, 2005, p.23). It seems to be taken for granted that IFRS are good for non-listed companies and that they should supersede National Accounting Standards. Perhaps this is because the official discourse of international accounting standardization is made by ‘the audit industry and its agents’ in such a way that ‘users tend to be represented rhetorically rather than physically’ (Hopwood, 1994, p.243). Why has there been such a sudden rush to converge national GAAP with IFRS, even for non-listed companies? Is financial accounting ‘solely a functional reflection of the internationalization of financial markets, or are other factors at stake?’ (Hopwood, 2000, p.763). How does accounting, as a technology and a social practice, serve to structure various institutional fields affected by globalization? Why is it that accounting technologies and accountants help to propagate organizational agendas, policies and purposes and, in doing so, amplify certain voices but do not ‘amplify others, yet these others deserve to be heard’ (Graham & Neu, 2003, p.467). In 2003, the Portuguese Accounting Standards Board (Comissão de Normalização Contabilística) [ Comissão de Normalização Contabilística (CNC), 2003] proposed a dual accounting model for Portugal. 2 This model required individual and consolidated accounts of listed companies to be prepared using IFRS. However, other entities have the discretion to use either Portuguese Accounting Standards (issued by the CNC) or IFRS. Thus, a dual regime of accounting standards would prevail: one for listed companies and the other for non-listed companies. 3 In these circumstances, a measure of convergence will be invaluable in assessing the extent to which the accounting methods allowable for non-listed entities converge with the methods permitted under IFRS. Obstacles to convergence in Portugal include the tax driven nature of national accounting requirements, the complicated nature of certain IFRS and the legal basis of Portuguese GAAP. Non-listed and local companies, audited by local auditors, feel more comfortable using National Accounting Standards because these standards are more in tune with their code-law, economic, and social backgrounds. It is important to examine and measure formal harmonisation carefully because of the increasing influence of accounting regulations on accounting practice. The first of three measures of convergence we analyse was outlined by Garrido, León, and Zorio, 2002. It was intended to assist in evaluating progress in converging any two sets of accounting standards, and is based on the concept of Euclidean distances. But we argue that measures based on Euclidean distances have serious shortcomings. Accordingly, we propose more robust measures involving Jaccard's [association] coefficients and Spearman's [correlation] coefficients because they provide a much stronger foundation for evaluation and seem likely to benefit a wide range of users of published financial statements. For investors, they provide a compact yardstick for comparison of the quality and content of financial statements published by companies in a variety of countries and settings. They provide enhanced measures of confidence and reliability by showing the similarity and dissimilarity between the allowable accounting methods that underpin those reports, and the variety of accounting ordained in IFRS. For regulators, the measures we advocate can be disaggregated to reveal the areas of accounting (such as valuation or revenue recognition) that are most dissimilar from IFRS and for which reform should be a priority. We illustrate the strengths and weaknesses of the three measurement methods by calculating the performance of Portuguese Accounting Standards setters in achieving convergence with IAS and IFRS between 1977 and 2003. Our choice of Portugal is opportunistic. It is motivated by our in-depth knowledge of Portuguese Accounting Standards-setting issues, and by assertions that Portuguese Accounting Standards are converging rapidly with international standards (see Jarne, 1997). We begin by briefly reviewing prior literature on the measurement of formal and material harmonisation; and delineate several recent phases in the evolution of IFRS and of National Accounting Standards in Portugal. Thereafter, we review the measure of convergence (based on Euclidean distances) that was outlined by Garrido et al. (2002), and highlight its shortcomings. We then propose Jaccard's association coefficients, supplemented by Spearman's correlation coefficients as better measures of the convergence of National Accounting Standards with IFRS.

The results for Euclidian distances can only be assessed in dynamic terms (considering more than one stage of time), since this measure yields absolute values that are not easily interpretable. Consequently, the value 38.96, by itself, does not mean anything special: it must be compared with the results of the previous stage (56.86). Only by comparing the evolution of the distances over time can we conclude about the convergence of National Accounting Standards with IFRS. In contrast, the results produced by Jaccard's coefficients can be interpreted in dynamic terms (increasing results over time denote formal harmonisation advances), and in static terms also. Accordingly, the result of this measure regarding phase NC to IC, means that the level of convergence between Portuguese and international accounting standards achieved in the period 1989–2003 is 59% for the 43 accounting issues analysed (and without considering the strength of methods). The calculation of Spearman's correlation coefficient reinforced these results and provided further evidence of the progress achieved in converging Portuguese accounting standards with IFRS. In the first phase there is a negative correlation between PAS and IFRS. In the last two phases the correlation coefficient is positive and increasing. However, despite the positive trends in convergence, the benefits of Spearman's correlation coefficient help us to conclude that the correlation between Portuguese accounting standards and IFRS only became statistically significant in the last (NC) stage (rs = 0. 387; p (0) < 0.01). The convergence of Portugal with IFRS started in 1991, when the Portuguese Accounting Standards Board (CNC) began issuing accounting standards (DCs) very similar to international accounting standards. When Portugal could not receive any support from the EU to regulate new accounting issues, international accounting standards were adapted by Portugal using DCs. The Portuguese system is slowly losing its French influence and is becoming similar to IFRS. Based on our sample, and considering the strength of the accounting methods, Portuguese standards were found to exhibit a 50% similarity with IFRS. A major question now is will the convergence process continue? After requiring EU listed companies to use IFRS (Regulation [EC] No 1606/2002), and after the needs of capital markets users have been satisfied, the convergence process is now being required to all companies in the EU. Clearly, as Hopwood (1994) has stated, the process of institutionalizing international accounting is not well understood and the interest in international accounting is not well known. It seems that the ‘apostles of internationalization’ (Cooper, Neu, & Lehman, 2003) want to erase local standards in the interest of harmonization, and in support of deregulation, privatization and globalization. More research is needed to understand if the process of convergence should continue in the EU. It seems an opportune time to develop understanding of ’the process of institutional change, the pacing of policy agendas and the influence structures around supranational accounting policy-making’ (Hopwood, 1994, p.246). Further, it is tempting to explore whether there is any concordance between the convergence similarity measures for Portugal and those obtained for other [so-called] ‘Latin countries’: that is, those classified by Hofstede (1987) into a ‘pyramid and people’ cluster (for example, Brazil, Italy, Spain, Belgium, Greece, Mexico, Japan and Korea), and characterised by large power distances and high uncertainty avoidance. Additionally, it would seem timely and beneficial to further explore the idea (introduced in footnote 1) that conventional measures of ‘convergence success’ might be regarded, contrarily, as measures of a ‘failure of cosmopolitanism’.