This was one of the most popular posts on my old blog, which surprised me a bit, but maybe the exam room holds nostalgic appeal for many of us…

I’m no doubt misremembering, but I’d almost swear that half the exercises when I was training to be an accountant related to figuring out eliminations of unrealized profits on consolidation (the other half related to income taxes). Since then, I’ve largely managed to avoid the subject. But because IFRS approaches the topic somewhat differently than old Canadian GAAP did, it may be useful to force myself to briefly step back in time.

Let’s consider items sold to a parent by a subsidiary in which a non-controlling interest exists; the items remain as inventory in the parent’s accounts at the date of consolidation. Old CICA 1600 set out the basic requirement, to eliminate the unrealized gain on these items to present them at historical cost to the consolidated entity, and provided some historical background: “There is a view that, where a non-controlling interest exists, a proportion of any intercompany transaction may be considered to have been at arm’s length, and that the proportion of the gains or losses relating to the non-controlling interest would be recognized in computing income. This view is inconsistent with the fact that the parent and its subsidiary are related and therefore not operating at arm’s length. For this reason, unrealized intercompany gains or losses should be entirely eliminated.”

It then addressed two further possibilities, that the subsidiary might either eliminate the unrealized gain entirely against the parent’s share of its income, or else that it might divide it proportionately between the parent and the non-controlling interest; CICA 1600 chose the latter view. So for example, suppose a subsidiary records a profit of $5 on items remaining in the parent’s inventory, and has net income of $100 overall; the parent has a 60% controlling interest. The $5 is eliminated in a 60/40 ratio: the parent’s share of the net income is $57 and the non-controlling interest’s share is $38. CICA 1600 pointed out this “may not be consistent with the legal position of (the non-controlling) shareholders” – in other words, from their perspective, the subsidiary’s net income, computed on its own terms, is correctly $40. But CICA 1600 was at peace with this “because consolidated financial statements are prepared primarily for the parent company shareholders and it follows that the effect on consolidated net income is based solely on the proportion of the subsidiary company’s shares held by the parent.”

IFRS 10.B86 doesn’t contain that same reference to proportionate elimination between the two interests, saying only “profits and losses resulting from intragroup transactions that are recognized in assets, such as inventory and fixed assets, are eliminated in full.” The CICA 1600 approach certainly meets the concept of eliminating the profit in full, but then so does the approach it rejected, of eliminating the unrealized gain entirely against the parent’s share of the subsidiary’s net income. According to one of the big firm texts, both approaches are commonly adopted under IFRS. So this seems to be an allowable but not a necessary difference between IFRS and old Canadian GAAP. Maybe it’s odd IFRS doesn’t specify a single approach for this, or maybe the choice reflects an inherent ambiguity about what the appropriate principle should be.

The mechanics of equity accounting, in contrast, sometimes seem to generate necessary differences. CICA 3051 said: “the elimination of an unrealized intercompany gain or loss has the same effect on net income whether the consolidation or equity method is used,” indicating then that the consolidation requirements for entirely eliminating unrealized gains and losses applied for equity accounting as well. However, IAS 28.28 says this: “Gains and losses resulting from ‘upstream’ and ‘downstream’ transactions between an entity…and its associate or joint venture are recognized in the entity’s financial statements only to the extent of unrelated investors’ interests in the associate or joint venture…The investor’s share in the associate’s or joint venture’s gains or losses resulting from these transactions is eliminated.” In terms of the CICA 1600 rationale cited above, this amounts to treating the portion of the transaction relating to unrelated parties as “arm’s length” and thus as qualifying for recognition.

In most if not all other respects, the IFRS requirements for equity accounting mirror those for consolidation – in both cases it requires uniform accounting policies for instance, and where the date of the investor’s financial statements doesn’t coincide with that of the investee, it requires adjusting in each case for significant events occurring between the two dates. But the distinction in the approach to intercompany eliminations appears rational – a transaction with a controlled subsidiary (even if not wholly-owned) might always be regarded, to some extent, as doing business with oneself, but a transaction with an entity one doesn’t control (albeit one over which some influence exists) is never just that. Maybe one could argue, for transactions with associates, that there’s no real need to mess around with eliminating the investor’s share at all, but I don’t suppose we’ll worry about that now. I mean, I have to get out of this exam room!

The opinions expressed are solely those of the author.