On Adoptions of IFRS by Jurisdictions

1. Introduction

I am grateful to the editor for suggesting that I should comment here on some of the important matters raised by Song and Trimble (in press, hereafter ST) in a paper in this issue of the journal. ST point out the lack of clarity in some previous writings about adoptions of International Financial Reporting Standards (IFRS) by various jurisdictions. The examples of this given by ST (e.g., in their Section 4.3) are convincing. Future researchers can benefit from ST’s tables, which provide considerable details about the different styles of adoption chosen around the world. For 2019, ST classify each of 195 jurisdictions (hereafter referred to as “countries”),¹ into one of 10 categories of status. These are summarized in this paper in the left-hand column of Table 1. ST also usefully add the time dimension in their online table,² which covers 1990–2021.

ST sensibly point out (e.g., in their Section 5) that some of their conclusions are matters of opinion, and that their tables will need to be updated in future. I comment below on five matters which researchers might consider when making updates: the nature of ST’s statuses, the types of reporting covered by ST, how to classify countries by types of generally accepted accounting principles (GAAP), other dimensions covered by ST’s statuses (including range of firms and degree of compliance), and auditor affirmations about IFRS. I apologize to readers for a number of self-citations, but this follows from the reason that the editor kindly chose me to write this piece; and these citings are more efficient than repeating points which I have published earlier.

2. ST’s Statuses

ST’s tables try to cope with six dimensions at once: country, time, the type of GAAP, the range of firms covered, whether the GAAP is required or merely permitted, and the degree of compliance by firms. As noted above, ST classify each country into 1 of 10 statuses of IFRS adoption; they do this for 2019 (“current” table) and over time (“historical” table). The statuses try to cover four of the dimensions (all but country and time). I will suggest in Section 5 that this is too ambitious.

Before we examine ST’s scope and statuses in more detail, we should note that ST’s 10 current statuses are not quite the same as their 10 historical statuses. The left-hand column of my Table 1 shows the current statuses (from ST’s Table 2),³ and my right-hand column shows the historical statuses (from ST’s Appendix B and as used in their online table). As may be seen in Table 1, ST use 13 statuses in all, because seven of the statuses are in both columns, but each of the columns has three statuses which do not overlap. As examples, the current status “Required (NC)” has no historical equivalent, and the historical status “IFRS (Local)” has no current equivalent. This may be somewhat confusing for readers, especially as the seven statuses which do overlap are labeled differently. As examples, the current status “Required (EU)” is historically called “IFRS (EU),” and the current “Permitted” is the slightly different historical “Permit” in Appendix B.⁴ It is unclear why the labels cannot be made consistent. I will explain below how I would reduce the number of statuses and use the same list for all years.

3. Scope of Financial Reporting Covered

ST focus on a particular scope of financial reporting: the consolidated statements of listed firms. This is a sensible limitation, but ST do not fully specify it. The “consolidated” limitation is not mentioned, but the “listed” limitation eventually becomes clear (e.g., in Section 3). Perhaps, for US researchers, the limitations are too obvious to discuss. However, elsewhere (e.g., in Europe) most of the regulated financial reporting is in the form of the unconsolidated statements of unlisted firms.⁵ For that type of reporting, adoption of IFRS by countries is influenced by other factors than just those mentioned by ST in their Section 2. For example, in those many countries where tax is closely connected to unconsolidated financial reporting, it is impractical for IFRS to be adopted for such reporting (see Gee et al., 2010; Nobes, 2008; Sellhorn & Gornik-Tomaszewski, 2006).

A second point on scope is whether or not foreign firms are covered by ST. In their Table 1, Japan is shown with the status that IFRS is “Permitted,” and this is the status in the online historical table for both Japan and the United States, for the last decade or so. However, the US status of “Permitted” relates to foreign registrants, whereas the Japanese status “Permitted” encompasses domestic firms. Surely, domestic firms are the subject matter of most of ST’s adoption tables? Many countries have special rules for foreign firms, but these are not generally scored by ST. For example, ST score European Union (EU) countries as requiring the EU version of IFRS even though foreign firms listed on EU exchanges are allowed to use US GAAP. So, in the historical table, the EU country scores of requiring “IFRS (EU)” are not right, when compared to the US score. I would solve this by restricting the scoring to domestic firms, and then changing the historical U.S. score to “Local.” ST have followed this logic in their Table 1, which deals with 2019 only, and shows the US as requiring US GAAP.⁶ I make another suggestion relating to “Permitted” in Section 5 below.

There is a further problem about scope. Pownall and Wieczynska (2018) discuss the distinction between regulated stock markets and other stock markets. They report (p. 1031) that, in the first year of EU mandatory adoption of IFRS, only 44% of United Kingdom (UK) listed firms used IFRS in their consolidated statements. Much of the explanation for the low adoption was that an important stock market segment was outside the scope of the EU Regulation of 2002. ST mention this issue,⁷ but their status “Required (EU)” is defined as “IFRS (as mandated by the European Union) are required for all listed firms.” The “all listed” part of the definition is too wide, especially for a paper which is trying to correct previous imprecision. I return to the issue of how to record the scope of adoptions in Section 5 below.

4. Types of GAAP

4.1. Equivalence with IFRS: Classifying Australia, Bangladesh, Jordan, Myanmar, and China

In their Section 2.2, ST discuss countries that have tinkered with IFRS, such as Australia. ST rely on Zeff and Nobes (2010) when they state that Australia “converged the standards with IFRS between 2005 and 2007, at which point they were deemed mostly equivalent with some modifications.” However, this is a misunderstanding. Reading from the bottom of Zeff and Nobes’ Figure 1 (and consulting their p. 182), it is clear that they are saying that Australian standards are fully equivalent to IFRS, because firms complying with the Australian standards would also be complying with IFRS as issued by the International Accounting Standards Board (IASB). So, ST’s “mostly” and “with modifications” are not appropriate as descriptions of the degree of equivalence achieved in Australia. Further, full equivalence was already in place from first adoption in 2005, not just from 2007.⁸ ST use the status “Modified” or, in historical tables, “IFRS (mod.)” for Australia, which would have been fine if that status had not been discussed as compromising equivalence. It is true that Australia modifies IFRS, but not in a way that compromises equivalence.

ST also use “Modified” for Bangladesh, where a major standard⁹ has not been implemented, such that equivalence to IFRS could not be securely claimed. I would solve the scoring problem revealed in the comparison of Australia and Bangladesh by abandoning the status “Modified” and creating two new statuses. First, I suggest “IFRS-equivalent” for countries such as Australia and New Zealand. This might also be suitable for Jordan, about which ST report (in their Section 4.3) that IAS 16’s option to revalue property, plant, and equipment (PPE) is not allowed, and “hence, it is a modified adoption.” In ST’s terminology, this implies compromised equivalence, but the deletion of an option does not compromise equivalent reporting. If this were the only caveat about Jordanian firms,¹⁰ they would still be complying with IFRS even without the option to use fair value for PPE.

My second new status would be called something like “IFRS-close.” It could be used for countries in which large parts of IFRS have been implemented but where equivalence is in doubt because the IFRS content is old or has gaps. This would have applied for Bangladesh. In their Section 4.2, ST also discuss the example of Myanmar, which they say introduced in 2010 “a limited set of IFRS that were current as of that date” but “plans to again adopt IFRS in 2022–23.” ST say that it is therefore “unclear how to classify Myanmar,” but they choose “IFRS (2010),” although this is not in their list of 10 current statuses. However, to me, this incomplete set of old standards would suit my newly proposed status of “IFRS-close.”

Given the importance of China, it is worth considering this country more carefully. ST now score China as “Convergence,” and I can see why. However, it is not clear how that is different from “Modified.” Convergence would seem to imply that China started from Chinese GAAP and gradually changed its requirements towards IFRS. However, in 2006, China introduced a new set of standards based on IFRS but with some modifications, as discussed further below. This looks like “Modified” rather than “Convergence.” However, if we follow my suggestion of abandoning “Modified,” would one of my proposed new types (“IFRS-equivalent” or “IFRS-close”) be suitable? Although the Chinese standards issued in 2006 were loosely based on IFRS of that time, some IFRS options were deleted,¹¹ and there were some gaps¹² and some substantive differences.¹³ However, I have suggested above that the deletion of IFRS options does not compromise equivalence. Further, Chinese standards have now largely addressed the gaps, and the standards are being kept up to date.¹⁴ I conclude that it would now be reasonable to score Chinese standards as “IFRS-close” because, on most topics, the Chinese standards will probably lead to the same accounting numbers as under IFRS. We would then not need the status “Convergence.”

I could see an argument for retaining “Convergence” with a weaker meaning than ST’s “jurisdiction-specific IFRS” (as in their Table 1). For example, an official convergence project (the “Norwalk Agreement”) between the IASB and the Financial Accounting Standards Board (FASB) began in 2002; by 2005, it had already led to several changes to US GAAP,¹⁵ such that US GAAP was somewhat converged with IFRS. Similarly, Japan was converging with IFRS from at least 2001,¹⁶ and a formal project of convergence was announced in 2005 (IASB, 2005). However, nearly all countries have converged with IFRS to some extent, so this meaning of “Convergence” would be vague and would add little information. Thus, I support ST in not using it with this meaning. I therefore confirm my suggestion in the previous paragraph that we do not need “Convergence” at all.

4.2. Parsimony in types of GAAP

The American Accounting Association (AAA, 1977, p. 77) set out criteria for good classification. These have been questioned (e.g., by Roberts, 1995), but we should bear them in mind. According to the AAA, one of the criteria is that there should be sufficient subsets to cover the whole field. I propose that we refine this to: “as few as possible subsets to pick up all substantial differences between objects, and yet to cover the whole field.” Empirical researchers might find ST’s 13 statuses too many. Indeed, some might ideally prefer a 0/1 variable. So, let us try to remove some of the statuses.

I suggest that ST’s distinction between “IAS” and “IFRS” (see the right-hand column of my Table 1) is not useful. In ST’s historical table, various countries change status from “IAS” in 2000 to “IFRS” in 2001, but nothing substantive changed between those years in the nature or content of the standards. It is true that at its first meeting, in 2001, the IASB adopted all the extant IASs which had been issued by the International Accounting Standards Committee (IASC). However, it was not until 2003 that the IASB issued its first IFRS, with mandatory effect from 2004; and even now most of the individual standards that are part of IFRS are still those with the prefix “IAS.” I therefore suggest that it would be better to conflate “IAS” with “IFRS” throughout ST’s entire period.

Also, ST’s status “In progress” (see Table 1) is perhaps not useful. For any particular year, it is presumably more useful to know the status of IFRS at that moment. Any change in status over time is picked up in ST’s historical table. I further suggest that we do not need the status that IFRS is “Prohibited” because that is effectively the same as “Local” being required. The latter is a better label because it is unlikely that IFRS is actually prohibited (as supplementary reporting) in countries where local GAAP is required.¹⁷

Finally, we can remove ST’s historical status “IFRS (SME).” Out of 195 countries, ST only use this for the Maldives, and then only for 2005–2009. This is, in itself, odd because IFRS for SMEs was not issued until 2009, so the Maldives cannot have been requiring it from 2005.¹⁸ I suggest that we can use my “IFRS-close” to cover any cases where the GAAP is IFRS for SMEs, because the accounting numbers produced might be the same as under IFRS for many topics, though not for a few topics.¹⁹

In terms of types of GAAP, the above discussion leads to my Table 2, containing five possible GAAPs throughout ST’s period. I will explain in Section 5 how I would deal with the other statuses of ST which have not yet been discussed. In Section 6, I suggest how to convert ST’s statuses into these types of GAAP.

4.3. The United Kingdom from 2020

The UK left the EU on January 31, 2020, after ST had written their paper. Given that UK companies comprise the largest country sample in many empirical studies of IFRS, it is important to discuss how this “Brexit” should affect the classification of the UK from 2020 in tables such as those prepared by ST or as amended by my proposals. Initially, there was no effect because EU rules were maintained during a transition period.²⁰ However, the UK has set up a “UK Accounting Standards Endorsement Board” to replace the EU’s endorsement procedure, and this will lead to “IFRS as adopted in the UK,” which will be overseen by a new regulator (The Audit, Reporting, and Governance Authority).

In terms of my Table 2, it is unclear yet which type of GAAP will be required for UK listed firms after 2020.²¹ It is possible that the UK will fall into line with future EU decisions about IFRS. The first major test case will be considerations about the endorsement in the EU and in the UK of the controversial IFRS 17, Insurance Contracts. It is also possible that the UK will move nearer to, or further from, “IFRS as issued by the IASB.” It will therefore be necessary to re-assess the UK’s position, though (in terms of my Table 2) it will presumably be IFRS-IASB, IFRS-EU, or IFRS-equivalent. It would be cumbersome to invent a special label solely for the UK.

5. More Dimensions: Range of Firms, Required/Permitted, and Compliant/Not

Up to now, I have been discussing three of ST’s dimensions: country, year, and the type of GAAP. However, as noted in Section 2, ST also try to score three other dimensions: the range of firms covered, whether the GAAP is required or merely permitted, and the degree of compliance by firms. ST try to deal with these three dimensions with their statuses, which are already dealing with the type of GAAP. For the range of firms, ST want to record, for example, whether IFRS is required for all firms, or for financial institutions only, or for all firms except financial institutions. They do this by scoring, for example, “Required (some firms)” in their current table or “IFRS (FI)” in the historical table. However, what if “IFRS (EU)” were required, but only for financial institutions? That is, sometimes the content of ST’s brackets relates to the type of GAAP but sometimes it relates to the range of firms affected. The further dimension of required/permitted is also very apparent in Table 1, e.g., in the “Permitted” and “Prohibited” labels.

I suggest that we have to recognize these other dimensions separately from the type of GAAP. We would then not need ST’s specific statuses “Required (most firms)” and “Required (some firms),” which are anyway not in the list of their 10 historical statuses. We could also remove “Permitted” because that is not a type of GAAP. Each status would then be a type of GAAP, such as “IFRS-IASB” or “IFRS-EU,” and its scope of application could be recorded separately in brackets. For example, some EU countries could be recorded as “IFRS-EU (firms listed on EU-regulated markets).”²² I deal with “Permitted” further at the end of this section.

As mentioned above, ST attempt to cover yet a sixth dimension: do firms comply with the requirements? However, this dimension only appears in the “current” statuses and only in one of them: “Required (NC).” This relates to non-compliance with IFRS as issued by the IASB, and is recorded for three countries in ST’s Table 1. However, I doubt that reliable surveys have confirmed that there is full compliance in all the other “Required” countries. Furthermore, are we sure that there is full compliance in countries requiring “Modified” or “Convergence” versions of IFRS? In my view, it is too ambitious to try to add this dimension, so I would delete the status “Required (NC).”

This returns me to the AAA’s criteria for classification. In addition to exhausting the universe of objects, as discussed in Section 4.2 above, the AAA (1977) suggested that subsets should be mutually exclusive. However, it would be possible for the EU version of IFRS to be imposed on most firms in a country but for the firms to be poor at complying. If we followed ST’s types of statuses, that single country would need three statuses at the same time.²³ My proposals above would lead us to a single GAAP type for each country, with a single modifier for range of required application in each country.

This discussion has not yet addressed how to deal with ST’s “Permitted” status. I have proposed removing this for the United States (in the historical table) by dealing with domestic firms only. For the small number of remaining countries to which “Permitted” is applied in ST’s tables, it would be necessary to record two or more types of GAAP (e.g., IFRS-IASB and Local), each with the modifier “(permitted).” As part of this logic, I assume that (as in Japan and Switzerland) it is always IFRS as issued by the IASB, rather than any other version of IFRS, which is permitted.

6. Conclusion on Statuses

All in all, my proposals in Sections 4 and 5 would reduce ST’s 13 statuses (which are shown in my Table 1) to five types of GAAP, as in my Table 2. If researchers would like to convert ST’s tables towards my proposed basis, they could use the right-hand column of Table 2. This would still necessitate resolving ST’s statuses 6, 7, and 8: Modified, Convergence, and In progress. However, this is a limited task because, of ST’s 195 countries in their Table 1, only 13 were scored as “Modified,” four as “Convergence,” and one as “In progress.” I have already suggested solutions for several of these countries in Section 4.1.

Nevertheless, the discussion of China illustrated the problem which will sometimes be met when assigning a type of GAAP to a country, and keeping it up to date. Many researchers will find it impracticable to try to assess (and to keep re-assessing) the closeness of financial reporting practices in a country to those that would occur under IFRS as issued by the IASB. It will be more practicable to assess the standard-setting process in the countries concerned. For example, the stance of the Chinese standard-setter seems now to be that new content of IFRS should be closely and rapidly absorbed into the national standards, whereas the FASB does not take that approach. This comparison would justify scores of “IFRS-close” for China but “Local” for the USA.

For empirical research, the appropriate use of ST’s tables depends on a researcher’s objective. In particular, is the ultimate aim to examine how countries behave or how firms behave? If researchers wanted a 0/1 variable for some country-related research questions, they could consider combining my types I–III as versions of IFRS. For other country questions, a wider or narrower selection of types would be suitable, perhaps comparing the results — for sensitivity analysis — with those achieved by other combinations.

Other IFRS researchers use firm-level data, and ST anticipate²⁴ that their tables could be used in this context. The modifications proposed here might assist. For example, researchers might wish to restrict their data to firms which have mandatorily adopted IFRS, or they might wish to contrast mandatory adopters with voluntary adopters. Researchers would then be able to use tables adjusted to my proposed country scores, such as IFRS-close (banks), IFRS-EU (listed on regulated markets), or IFRS-IASB (permitted for all listed). More generally, researchers who use a firm’s GAAP as an independent variable might find these country scores useful as part of assembling or checking such data.

7. Auditors’ Affirmations

In their Section 5, ST refer to Nobes and Zeff’s (2008) finding that, in 2005, auditors generally did not affirm compliance with IFRS as issued by the IASB, even when such compliance was probably being achieved. This still applies generally in the EU, with the continuing exception of EU firms that are listed on US exchanges. Later, Zeff and Nobes (2010, p. 182) explained that Australia had rectified this problem by requiring dual audit reports (on Australian standards and on IFRS).²⁵ However, dual reports are no longer universal practice in Australia, after a further change to auditing standards.²⁶ Later still, Nobes and Zeff (2016, p. 287) lamented the fact that South Korean auditors from 2011 were giving affirmations restricted to “Korean IFRS” even though compliance with IFRS was probably being achieved. Despite all this, audit reports in many other recently adopting countries affirm compliance with “IFRS as issued by the IASB” or just with IFRS (e.g., in Brazil, Canada, and Russia).

Incidentally, ST conclude that “the reluctance of auditors to confirm IFRS adoption negates its effectiveness.” However, it might not be the auditors who are reluctant. They would perhaps be pleased to be asked to do more paid work. The reluctance can instead be blamed on those national regulators who miss the point of having international standards by not requiring auditors’ affirmations about IFRS. Of course, in such countries, firms themselves could ask for dual reports, and some do.

8. Conclusions

ST should be congratulated for addressing in detail how to classify IFRS adoptions for a wide range of countries over time. They convincingly demonstrate previous imprecision on this, and that “adoption exists on a wide and complex spectrum” (ST’s Section 6). I have commented on five issues: the nature of ST’s statuses, the scope of ST’s paper, the classifying system for GAAPs, other dimensions (e.g., firm compliance), and auditor affirmations.

ST restrict their scope to consolidated statements. This is perfectly reasonable, but they do not explain that they have done this or note that it excludes most of the financial reporting in many countries. Further, ST’s scope is said to be “listed” firms, but this is too wide for ST’s actual scoring because, to take the example of the many EU countries, the requirement to use IFRS does not apply to listings on some types of stock market. Also, although ST score most countries on the basis of requirements for domestic firms only, this is not always the case (e.g., for the United States in the historical table). I have suggested that the tables should consistently be restricted to domestic firms, and that this should be acknowledged.

ST use a long list of “adoption statuses.” At first sight, there are 10 statuses, but really, there are two partially overlapping sets of 10, comprising 13 statuses in all (see my Table 1). I recommend using fewer statuses and being consistent over time. ST’s statuses try to cover four dimensions: type of GAAP, the range of firms covered, whether the GAAP is required or merely permitted, and the degree of compliance by firms. In all, ST have six dimensions because the tables cover many countries and many years.

One dimension covered by ST’s statuses is the type of GAAP. For this dimension, I would abandon ST’s “Modified” IFRS status because it covers cases that are importantly different from each other. Instead, I would create two new categories: (a) “IFRS-equivalent” for countries such as Australia and New Zealand, where changes to IFRS have not jeopardized compliance by firms with IFRS as issued by the IASB; and (b) “IFRS-close” for countries with less-than-full implementation of IFRS, as was the case for Bangladesh and Myanmar. As part of this, the status “Convergence” can be dropped, and the status “IFRS (SME)” can be subsumed within “IFRS-close.” We could also delete the “Prohibited” status because it is redundant, since it means the same as “Local.” I have included a discussion about the position of the UK from 2020, given that UK companies often comprise the largest country component of samples of IFRS reporters.

The remaining dimensions which are somewhat covered by ST’s statuses are the range of firms covered, whether the GAAP are required or merely permitted, and the degree of compliance by firms. For example, three of the current statuses are “Required (EU),” “Required (most firms),” and “Required (NC),” but versions of all three could be true at the same time for a country which required the EU version of IFRS for most firms, which were not very compliant. I have suggested that attempting to include the non-compliance dimension is too ambitious because we do not have data for most countries.

I propose that the classification should primarily deal with types of GAAP. This involves abandoning the “Permitted” status because it is not a type of GAAP. Overall, my changes would reduce ST’s 13 statuses to five types of GAAP, with scope of application picked up as another dimension. I have suggested how researchers could translate ST’s statuses to mine. I have also suggested how 0/1 variables could be created by empirical researchers. As ST point out, adoptions are in constant flux, so future researchers will need to check and update ST’s scores. While they are doing that, they might wish to consider the modifications proposed above.

On auditor affirmations, I have updated ST’s point by noting that many countries do now require affirmations relating to IFRS. I also suggest that it is regulators, not auditors, who should drive improvements on this.

Acknowledgments

The author is grateful to the editor for commissioning this paper. He thanks the following for comments on an earlier draft: Christian Stadler (Royal Holloway, London), Ann Tarca (IASB) and Stephen Zeff (Rice University, Texas).

Notes