Some observations on one of the core prohibitions of IFRS for first-time adopters

One of the first things you learn about the mechanics of retrospectively adopting IFRS for the first time is probably the exception that relates to estimates: “an entity’s estimates in accordance with IFRSs at the date of transition to IFRSs shall be consistent with estimates made for the same date in accordance with previous GAAP (after adjustments to reflect any difference in accounting policies), unless there is objective evidence that those estimates were in error.” In other words: a first-time adopter doesn’t apply hindsight to derive better estimates than it reported originally. The broad point of this can easily be grasped: adopting IFRS shouldn’t provide an opportunity to rewrite an entity’s financial reporting history as it originally unfolded. If, for example, it necessarily took several annual accounting periods for the ultimate amount of a legal settlement to become clear, then adopting IFRS shouldn’t be a reason for retrospectively reporting the item as if the outcome had been known from the outset.

This prohibition is perhaps most valuable when, as happened in Canada, a large part of the financial reporting population makes the transition to IFRS at the same time, and users only want the disclosures about the comparative figures to explain the incremental effects of the new accounting model on the numbers they’re already familiar with, rather than to confuse things with a different substantive analysis of the facts that gave rise to those numbers. I’m not sure the prohibition is always as important in other situations though. For example, for various reasons, an entity reporting for the first time under IFRS might provide a reconciliation to a set of previous-GAAP numbers that no current investors have really seen or have any reason to care about. In that situation, you could argue (and I’m only writing a little think-piece here) that an investor assessing the entity for the first time might be better served by statements that clean up the messiness of past estimation processes; nothing about the history of an entity’s accounting estimates as it originally unfolded carries any predictive value regarding the entity’s future direction or prospects. Still, even if we all agreed on that (and I don’t suppose we would), we couldn’t expect IFRS 1 to set out different expectations for different kinds of adoption scenarios I guess.

This isn’t the only place in IFRS where you run across the concept. For example, the most recent Annual Improvements to IFRSs clarified some aspects of the IFRS 7 disclosures about continuing involvement in a transferred financial asset. In determining the basis of application for the amendments, the IASB noted that for an entity to apply them to any period it presents that begins before the period in which it applies the amendments might “..require an entity to determine the fair value as at the end of the period for a servicing asset or servicing liability, which the entity might not have previously determined. It might be impracticable for an entity to determine the fair value of such a servicing asset or servicing liability without using hindsight.” The same passage refers later as well to “the risk of hindsight being applied.” Try as I might, I can’t see why it would matter in this case if hindsight was applied though. As I understand the IASB’s reasoning, because hindsight might result in an entity reporting a better estimate of fair value for those preceding periods than it would have computed at the time, the Board allows it not to report an estimate at all…which surely for purposes of ongoing decision-making can only possibly constitute worse information than whatever the hindsight-affected number might have been? You might say this elevates the avoidance of hindsight into something of a fetish.

I wondered what the IASB said originally about this aspect of IFRS 1, but I couldn’t track down the original exposure drafts and other material. This is how the basis for conclusions summarized the issue though:

• An entity will have made estimates under previous GAAP at the date of transition to IFRSs. Events between that date and the reporting date for the entity’s first IFRS financial statements might suggest a need to change those estimates. Some of those events might qualify as adjusting events under IAS 10 Events after the Balance Sheet Date. However, if the entity made those estimates on a basis consistent with IFRSs, the Board concluded that it would be more helpful to users—and more consistent with IAS 8—to recognize the revision of those estimates as income or expense in the period when the entity made the revision, rather than in preparing the opening IFRS balance sheet (paragraphs 31–34 of the IFRS).

Perhaps one should resist trying to over-interpret the reference to “more helpful to users,” but it suggests that the decision was partly pragmatic rather than solely principled, and that at some point there might have been some discussion of taking the opposing view. And as I argued at the top, I’m not sure how opening the door to hindsight would always be less helpful to users. Basically, it doesn’t seem to me that hindsight is such a bad quality in life generally, if it’s used rigorously and non-self-servingly, for the purpose of more clearly understanding the past, so as to do better in the future. Some applications of hindsight in financial reporting would surely amount to no more than that…

The opinions expressed are solely those of the author