IASB member Tokar's response to academic paper on assets and liabilities published

At the IASB Research Forum 2016, an academic conference organised by the IFRS Foundation in conjunction with the academic journal Contemporary Accounting Research (CAR), International Accounting Standards Board member Mary Tokar responded to a draft research paper written and presented by Nicole L. Cade, Lisa Koonce and Kim I. Mendoza, discussing when assets and liabilities exist. Mary’s response has been published in CAR alongside the final paper by Cade et al.

Discussion of “Assets and Liabilities: When Do They Exist?”¹ 

Abstract: Cade et al. present the results of an experiment in which they test non-accountants’ natural understanding of whether items are assets or liabilities as well as the author’s summary of the Board’s current and proposed definitions of assets and liabilities. Additionally, the summaries of the current and proposed definitions are tested on a separate group of participants with accounting experience. They find that participants use a higher probability threshold in recognising assets than in recognising liabilities, but that the Board’s proposed changes to the definitions align participants’ judgments more closely with standard-setters’ intentions. I note that the experiment is not a direct test of the Board’s current or proposed definitions. I argue, however, that the paper draws attention to an important bias in natural reasoning and I suggest that standard-setters need to be aware of such biases and to consider how best to respond to them.*

I found the paper very useful – and in Conceptual Framework terms this is the highest praise possible.

The paper has an innovative design in terms of thinking about which biases – explicit or implicit — impact the application of IFRS Standards. The focus on a non-accountant population is relevant, as is the subsequent comparison of the non-accountant outcomes against outcomes from those with some accounting experience. This test design answers a recurrent concern of mine when tests (or responses to the Board’s proposals) involve people with accounting training: how outcomes might be influenced by participants’ training and experience with current requirements. I also think that the design of three scenarios — the no-definition scenario, one based on the existing definitions, and one based on the proposed definitions — is effective in isolating what factors are influencing outcomes.

The notion of natural reasoning gives a very useful insight into the environment in which IFRS Standards are used, and it is quite interesting to see how there seems to be very strong natural asymmetry in recognising assets and liabilities, suggesting inherent conservatism. This is something very familiar to those of us involved in the debate about prudence that the Board has had over the last several years. How do we express prudence? The Board’s conclusion in the revised Conceptual Framework for Financial Reporting (published in March 2018) is to say that prudence, in the context of neutrality, is caution under conditions of uncertainty rather than a deliberately asymmetrical approach. But it is very important for us to be aware of the results based on the natural reasoning of individuals to whom we are giving these neutral definitions.

The paper might be taken to imply that standard-setters think that there is “misuse” of the existing definitions. I think it would be better to say that the current definitions are not as effective as standard-setters wish in terms of driving behaviour, especially with respect to existence and separating the issue of existence from recognition and measurement. Frankly, the Board bears some of the blame for not achieving standards-level requirements that reflect fully the objectives of our elements definitions. When we started the Conceptual Framework project, the staff highlighted some instances in which current requirements in IFRS Standards may be seen as inconsistent with the wording of the elements definitions, which was a reason to revisit those definitions.

The experiments’ results suggest that the changes that the Board is making to the definitions of assets and liabilities will be useful in that people’s understanding of how to apply the definitions would be more closely, but not perfectly, aligned with standard-setters’ intentions. This is the case for both assets and liabilities.

Let me report on some testing of the proposed IASB definitions that we did ourselves with very experienced accountants. Representatives of national standard-setters from around the world met with the Board in September 2016, and we tested on them the asset and liability definitions that we proposed in the Exposure Draft. Participants were provided with about 20 scenarios, divided into groups with each group asked to analyse four or five scenarios in depth. The staff reported on this test at October 2016 meeting². The staff report noted some areas/examples for which outcomes were inconsistent or challenging to analyse using only the proposed definitions. However, these difficulties were mainly in areas not tested by Cade et al.³ Overall, while not directly comparable tests to the tests conducted by Cade et al., I think that our findings from our own testing are not inconsistent — generally, the changes proposed by the Board seem to improve realisation of the Board’s objectives.

Based on my experience from Conceptual Framework project⁴, the Cade et al. paper provides accurate insights into why the concept of probability is incorporated into both the asset and liability definitions and the recognition criteria. The 2010 version of the Conceptual Framework incorporated probability in its definitions of assets and liabilities as follows:

An asset is a resource controlled by the entity as a result of past events and from which future economic benefits are expected to flow to the entity [emphasis added] A liability is a present obligation of the entity arising from past events, the settlement of which is expected to result in an outflow from the entity of resources embodying economic benefits. [emphasis added]⁵

The Cade et al. paper explains that the notion of probability is included to convey that 100% certainty is not required before an asset or a liability is recognised. The Board’s outreach during its Conceptual Framework project indicated that many have instead interpreted the current wording to require a conclusion that an inflow (asset) or outflow (liability) of resources must be more likely than not before an asset or liability is recognised. This loss of understanding of the standard-setter’s original intent emphasises the importance of an effective basis for conclusions for our work, to help all those using IFRS Standards to understand the context in which the Board intended the standards to be applied.

On the question of natural reasoning, I think a good metaphor for what the paper is telling us would be to think about a sailboat. A sailboat has to contend with the wind and with the currents underwater. You can think about what we often try to do with standard-setting as trying to direct or manipulate how the sails are set to cope with the winds. But the paper reminds us that we are also operating in water full of currents, and as standard-setters we have to think about that – about the effects of natural reasoning — when we ask people to apply our definitions, whether we are writing a conceptual framework or writing accounting standards.

One of the other things that I think the experiment suggests is that words do matter — as we see from the difference in conclusions from the test subjects between those using a definition based on the current definition and those using a definition based on the new definition.

The authors look at the definition of an asset and the definition of a liability and, in the instances in which they give definitions to the experiment participants, ask them to apply these definitions. But it should be remembered that, in most cases, a conceptual framework does not flow directly to the decisions that are being made by preparers of financial statements because standards apply directly, and the Conceptual Framework does not override them. IFRS Standards now are sufficiently complete for virtually all transactions and events to be within the scope of a directly applicable standard, even if not addressed directly. If the only tool required for a complete, high quality financial reporting framework capable of being applied consistently were a conceptual framework, we could fold our tents and go home once we completed the 2018 version of the Conceptual Framework and we would not need individual IFRS Standards. So, a conceptual framework is a necessary but not sufficient part of IFRS Standards. But I agree with the authors’ views about why the Conceptual Framework is important.

One thought that I took away from reflecting on the paper is: how useful is information if we are trying to impose it on someone with a natural framework that it is not aligned with the framework used to prepare financial statements? Is information that is not aligned with natural reasoning useful to users of financial statements?

This is a question of how well people understand financial statements. The paper challenges symmetrical thresholds for asset and liability recognition from this point of view, but what are some of the other instances in which a misalignment might arise? One may be best estimates versus expected values. Take the example of a lawsuit in which, if you win, the cost is zero; if you lose, then the cost is 100. If you have a 50% probability of losing, then you have an expected value measurement of 50, which is neither possible outcome. Is that useful information? Another example is uncertain tax positions. The guidance in US generally accepted accounting principles on uncertain tax positions requires recognising the greatest amount that has a greater than 50% chance of being paid. This is logical once you work through it and understand both the objective and the calculation, but it is very difficult to get people to understand both the calculation and what the recognised amount represents.

So, what does the Board have to do? Do we change how we approach things? Maybe we should challenge ourselves to do that, but I have some different suggestions.

Words matter, but they may not be sufficient. The identification of the tension between natural reasoning and the application of the definitions says to me as a standard-setter that we have to reinforce the Conceptual Framework with well-written standards that articulate our logic. It means that we have to think about examples to say, “Here is how we expect the requirements in the standards to be applied.” We also have to think about supplemental educational materials and what I have sometimes called “speechifying”, reminding people in informal ways, like speeches, of how we have balanced costs and benefits. Those costs may include tensions with natural reasoning.

Sometimes changing words in a standard encourages re-examination of past practices, and again this may be something that the Board should be more proactive in trying to prompt. An example of this kind of change management is a change we made in December 2014 to IAS 1 Presentation of Financial Statements.⁶ We were addressing concerns about disclosure shortfalls: in some cases too much boilerplate, in other cases too little entity-specific information. One of the relatively minor changes we made was to introduce a discussion of understandability as a counterbalance to the notion of completeness. It is only a two-sentence paragraph (IAS1.30A), but the hope is that it will prompt people to challenge decisions that they have taken in the past about, for example, the materiality of certain disclosures. These are all things that we have to think about from a behavioural perspective.

The Cade et al. paper is not a direct test of the Board’s current or proposed definitions in its Conceptual Framework for the elements of financial statements. For example, the definition of an asset in the 2010 version of the Conceptual Framework is: “A resource controlled by the entity as a result of past events and from which future economic benefits are expected to flow to the entity”. For the purposes of the experiment, this is paraphrased as: “An item is an asset if it provides the company with a probable future economic benefit”. I am not dismissing the study because of that; I just want to point out that it is not a direct test of the 2010 or 2018 Conceptual Framework’s wording.

To sum up, these are some of the conclusions that I take away as a standard-setter from this paper.