Kruger Incorporated is a decision that considers two issues in tax law. One issue concerns business taxation and the mark-to-market method of accounting. The other concerns the law of statutory interpretation. The significance of statutory interpretation law and the scant authority for mark-to-market accounting in tax reporting make an examination of this decision worthwhile, despite amendments to the Income Tax Act following the decision.
A definition of terms is in order. Mark-to-market accounting recognizes changes in a property’s value as the changes occur, regardless of whether a disposition of property has occurred. Realization accounting recognizes changes in value only upon the disposition of property. Two other terms for realization accounting are cash accounting, or cost accounting. In Kruger, the courts took a broad view of the meaning of mark-to-market accounting, essentially equating this term with the term “fair value”. Fair value is broader than mark-to-market accounting because the latter is a subset of the former. That is, fair value is mark-to-market accounting plus mark-to-model accounting. The difference between mark-to-market/model depends on the nature of the market in which the valuated property trades. Mark-to-market relates to liquid, open markets, whereas mark-to-model relates to relatively illiquid, closed markets. These definitions provide context for the discussion that follows.
In Kruger, the Tax Court of Canada and Federal Court of Appeal characterized Kruger’s method of valuing its options as mark-to-market rather than mark-to-model. In my respectful view, the latter is the more precise characterization. This characterization may have caused the courts to accept Kruger’s method as more accurate than the Minister’s realization accounting method, which led them to permit Kruger to use mark-to-market for tax purposes.
The FCA also characterized its resolution of the statutory interpretation issue as reading words into the statute, but the Court appears merely to have read down the provision. Moreover, there appears to be some inconsistency in the proper approach to statutory interpretation, and the difference between the plain and ordinary meaning of the words of a statute. Nuance and complexity permeate statutory interpretation law, and Kruger highlights the difficult issues one can encounter in this area.
Kruger Incorporated v Canada
In Kruger, the taxpayer was a pulp and paper company that engaged in foreign exchange trading. During the 1980s, the company decided to hedge its exposure to the US Dollar, so it began trading foreign currency options to mitigate the impact of currency fluctuations in the CAD/USD pair. By the 1990s, Kruger’s trading operation became a speculative operation, instead of a hedging operation.
In 1998, the relevant taxation year, Kruger suffered a trading loss, which it claimed under s.9 of the Income Tax Act. Alternatively, it claimed a tax deduction under s.10 of the Income Tax Act. The issues before the courts were whether mark-to-market accounting is a permissible method of measuring one’s taxable income under the Income Tax Act (“ITA”), thus allowing Kruger to claim the loss under s.9; and whether derivatives are inventory under s.10 of the ITA, thus allowing Kruger to claim the fair market value decline in the derivatives it held.
At the Tax Court of Canada, Justice Rip held that the plain language of s.10 refers to “any property the cost or value of which is relevant to the computation of business income”, which is very broad language. Nothing in the ITA’s language prevents taxpayers from including derivatives in inventory for tax purposes and claiming deductions for fair market value declines in the value of that inventory. That is, Kruger could claim a deduction under s.10. Under s.9, Justice Rip treated realization accounting, which was the alternative to mark-to-market accounting, as an overriding principle of tax law, holding that no provision in the ITA permitted Kruger to depart from this principle. Given Justice Rip’s reference to the Canderel framework, in effect Rip J held that realization accounting provided certainty and that it was more accurate than mark-to-market accounting on the facts. The Tax Court of Canada disallowed Kruger’s mark-to-market method and held it had to use realization accounting for tax purposes.
At the Federal Court of Appeal, after receiving submissions of a clerical error that occurred in the valuation of Kruger’s derivatives contracts, Justice Noël held that the mark-to-market method was at least as accurate as the realization method. On this issue, the Minister had not discharged its burden of proving that realization was more accurate. Thus, the FCA reversed Justice Rip’s decision under s.9, allowing Kruger to mark-to-market its derivatives losses. Justice Noël also held that the ordinary meaning of the word inventory is property held for sale. This interpretation qualified the statutory definition of “inventory” under s.248. Noël J followed the Supreme Court of Canada’s decision in Friesen v Canada, holding that the derivatives were not inventory because Kruger held them for speculative purposes instead of for sale.
Options Pricing: Marking to What?
One issue with this decision is the characterization of the valuation methods Kruger had used, which made Kruger’s method seem more accurate.
Kruger claimed that it used mark-to-market accounting, and the courts characterized it as such. This method is different from the fair value accounting method that it actually used. Though the courts used the terms synonymously, as I point out above fair value is a broader concept. Besides mark-to-market accounting, fair value encompasses valuation methods like mark-to-model accounting. Mark-to-model is different from mark-to-market.
In Kruger, there was no market in the ordinary sense, as Kruger traded the derivatives over-the-counter. These are private transactions between two trading partners. Kruger did not value its derivatives using an open market price. Kruger’s counterparties gave to Kruger the values it used to price its options. And according to Kruger’s submissions, its counterparties used a model to value Kruger’s options. That model was the Black-Scholes model, which provides a theoretical price estimate for the options. It does not establish a market price in the sense of a standardized, open price quote.
At trial, Justice Rip summarized evidence provided by Professor Klein, who was one of the expert witnesses that testified about Kruger’s valuation method:
 In Professor Klein’s view, Kruger’s mark to market valuations at the end of 1998 were not based on a market or a traded price, “they represented theoretical estimates determined by sophisticated pricing models where market values of other securities were used to determine some of the input variables.” The bank’s pricing models are “very mathematical and rely on highly idealized assumptions” with respect to how securities markets operate. Mark to market values are unreliable.
Professor Klein was the only witness to give evidence about the reliability of the models used to value Kruger’s options. In Klein’s view, Kruger’s counterparties used inaccurate models to price Kruger’s options. A market price did not exist.
The significance of using modeling for valuing a property appears when one considers the technical point that the Black-Scholes model relies on the assumption that asset returns are normally distributed. This is not true. The falsity of this assumption is crucial because a model is only as strong as its assumptions. For greater certainty, a normal distribution is one that results in a bell-curve if one plots each period’s returns. Asset returns are not normally distributed. Thus, “highly idealized assumption”, as professor Klein put it, is perhaps an inadequate description of the underpinnings of the model used by Kruger’s counterparties to value the derivative contracts.
For practical purposes, mark-to-market and fair value are often interchangeable. But it is important to be clear that they are not the same concepts because a model provides a hypothetical price the truth of which depends on its underlying assumptions, whereas a market price represents the price a taxpayer would actually realize upon the disposition of property. A model price might differ substantially from the actual market price i.e. the realizable value of the property.
The model Kruger used has questionable accuracy over the long run because one of its assumptions is false. Asset prices are not normally distributed. Characterizing Kruger’s method as mark-to-market rather than mark-to-model may have inclined the courts to find that Kruger’s model was more accurate because one associates the term “market” with greater accuracy than the term “model”. That is, if the courts had characterized Kruger’s method as a model, they may have viewed the certainty of realization as more accurate than a mere model.
The above discussion lays the groundwork for identifying an error in the Federal Court of Appeal’s decision. The distinctions between mark-to-market, mark-to-model, and fair value, are subtle. Moreover, even though strong evidence supports the proposition that the Black-Scholes model relies on a false assumption, the model’s founders won a Nobel Prize in Economics for writing the equation and laying a foundation for options pricing. In short, the complexity of this technical case caused any error that did occur. The next section of this comment discusses the particular passages of the Federal Court of Appeal’s decision that may contain an error by the Court.
Did the Federal Court of Appeal err: Inconsistent Values or Models?
On Justice Rip’s finding that realization is more accurate than mark-to-market, Justice Noël reasoned:
 The Tax Court judge went on to explain that even if mark to market accounting was authorized, he was not convinced that the values used by the appellant were reliable (Reasons, para. 116). He drew no definite conclusion on this point as evidenced by the judgment that he gave (see para. 30, above). It is nevertheless useful to comment on this point given the information that has since been brought to our attention by the appellant.
 The Tax Court judge was concerned that the bank counterparties to the contracts held by the appellant at the close of its 1998 taxation year would not “necessarily” have used a model of valuation which relies on the same inputs (Reasons, para. 116). He was taken aback by evidence tendered by the respondent’s expert (Professor Klein) showing that two options written by the appellant with identical terms were ascribed by distinct bank counterparties values which were more than 20 percent apart (idem, paras. 48 and 116).
 The appellant recognizes that such a difference would shake one’s confidence. However, it submits that this discrepancy is due to a clerical error made in preparing its own expert report. The following extracts quoted from the memorandum of the appellant (references omitted) provide the explanation:
- Mr. Klein testified, with reference to his report, that identical contracts dated May 13, 1998 for a call option sold by Kruger of $10 million USD at a strike price of $1.46 to mature in one year were valued by the Bank of Nova Scotia (“BNS”) at $797,736 and by JP Morgan (“JPM”) at $612,200. Reproduced in his report was the JPM communication to Kruger of the value of that contract at 486,900 USD which, at the prevailing rate of exchange on December 31, 1998 of 1.530 CAD, yielded a Canadian dollar value of $745,200, not $612,200. The difference, therefore, between the JPM option contract relative to the BNS option contract is, therefore, not $185,536 but rather $52,536, and therefore not a difference of 26.3% but rather 6.8% …
- Mr. Klein further testified, with reference to his report, that identical contracts again dated May 13, 1998 for a put option sold by Kruger of $10 million USD to mature May 13, 1999, at a strike price of $1.40, again with both JPM and BNS as purchasers, were valued by the BNS at $8,173 and by JPM at ($113,257). The JP Morgan communication of value to Kruger sets the value of that contract at 7,900 USD which, translated to CAD at the prevailing rate at December 31, 1998, yields the amount of $12,090, and not ($113,257). …
- The remaining differences in values given to four other option contracts with identical terms by two different banks set out in Mr. Klein’s report at page 2386 are: -0.3%, 1.2%, -4.2% and 5.1% …
 The Crown does not challenge the above demonstration otherwise than by asserting that it is not supported by the evidence (Memorandum of the respondent, para. 6). However, all the figures referred to in the above quoted passage can be found in the record and when the prevailing rate of exchange is applied (1.5305 CAD; Appeal Book, Vol. 12, p. 2,387), it can readily be seen that the wrong rate was applied and that the variations are within the narrow range described.
 I do not believe that the Tax Court judge would have been troubled by the values submitted by the bank counterparties if he had been appraised of the actual figures. I should add that this no longer seems controversial as the Minister has since recognized that the appellant’s values were reliable (see para. 31, above).
The above extract references paragraph 116 of the Tax Court decision:
 There is also a difficulty with Kruger’s option contract themselves. As I view the evidence, and as Professor Klein suggests, the market prices of OTC foreign exchange option contracts are estimates, if not artificial. Kruger’s contracts are European contracts and cannot be traded during the time the contract is made until it matures, except with consent of the other contracting party. Each bank formulates its own market value based on its own model applying inputs that are not necessarily used by another bank. There is no consistency or objective, it appears to me, by the banks in fixing a common fair market value for a given day (or time of day). Professor Klein described two identical contracts as to amounts and expiry dates, among other things, having a significant difference in values on December 31, 1998. This shakes my confidence as to the other market values used by Kruger, not because Kruger was trying to do something nefarious, which it was not, but because of a probable inconsistency in values depending on the different models used by Kruger’s counterparties. I may have had more comfort in agreeing with Kruger’s valuations if all the contracts were valued using the same models, not a variety of models used by different banks. Thus even if I found that it was appropriate for Kruger to mark to market its derivatives for income tax purposes, I would not find that the inconsistent bank values used by Kruger was properly applied in calculating its losses on derivative trading in 1998.
Justice Noël’s reasoning that Justice Rip’s lack of confidence arose because inconsistent values attributed to the derivatives contracts is one possible interpretation of the Tax Court decision. Justice Rip wrote, “even if I found that it was appropriate for Kruger to mark to market its derivatives for income tax purposes, I would not find that the inconsistent bank values used by Kruger was properly applied in calculating its losses on derivative trading in 1998.”
However, Justice Rip stated, prior to the above sentence, “This shakes my confidence as to the other market values used by Kruger … because of a probable inconsistency in values depending on the different models used by Kruger’s counterparties. I may have had more comfort in agreeing with Kruger’s valuations if all the contracts were valued using the same models, not a variety of models used by different banks.”
That latter part enables one to interpret paragraph 116 of the Tax Court decision in one of two ways. First, the inconsistent values themselves troubled Justice Rip. Second, the inconsistent modeling troubled Justice Rip. These are not the same. The values are the outputs of the models, while the models are equations that use inputs to produce outputs.
If one accounts for the proposition above that the Black-Scholes model relies on the false assumption that asset returns are normally distributed, it is reasonable to suppose the models themselves, rather than their outputs, troubled Justice Rip. But the falsity of the model’s assumption is not clear on the evidence, nor in the courts’ reasoning. Regardless, Justice Rip’s reliance on the realization principle, and his view of the evidence of Professor Klein, suggest that the inconsistency in models was more troubling than the inconsistency in value. Consider the word “may” in the sentence, “I may have had more comfort in agreeing with Kruger’s valuations if all the contracts were valued using the same models”. The word “may” suggests that Justice Rip disagreed that Kruger’s method was more accurate than the Minister’s was even if identical contracts had identical values. It suggests that Justice Rip’s primary concern was the use of a different model by each one of Kruger’s counterparties.
This alternative interpretation of paragraph 116 suggests the FCA might have erred in substituting its view of the facts on the issue of whether mark-to-market accounting was as accurate as realization accounting. If Justice Rip’s problem was the inconsistent values assigned to identical contracts, then this substitution was likely appropriate. It prevented a re-examination of the facts at trial, thereby conserving scarce judicial resources. But if the problem was inconsistent modeling, then the substitution should not have occurred, because nothing disclosed at the FCA undermined Justice Rip’s findings in this respect. Justice Rip heard the evidence about how Kruger priced its options, and made a finding of fact about the accuracy of this method and the permutations thereof. Rip J held that realization provided the more accurate picture. Nothing at appeal affected the evidence that supported this particular finding.
Whether Inventory is Property Held for Sale
The second issue in Kruger was the interpretive issue under s.10 of the Income Tax Act (“ITA”), whether options were “inventory”. This issue also highlights the importance of characterization.
This issue invokes the fundamental statement of the law of statutory interpretation, the modern principle. In Canada Trustco Mortgage, the Supreme Court of Canada held:
10 It has been long established as a matter of statutory interpretation that “the words of an Act are to be read in their entire context and in their grammatical and ordinary sense harmoniously with the scheme of the Act, the object of the Act, and the intention of Parliament”: see 65302 British Columbia Ltd. v. Canada, 1999 CanLII 639 (SCC),  3 S.C.R. 804, at para. 50. The interpretation of a statutory provision must be made according to a textual, contextual and purposive analysis to find a meaning that is harmonious with the Act as a whole. When the words of a provision are precise and unequivocal, the ordinary meaning of the words play a dominant role in the interpretive process. On the other hand, where the words can support more than one reasonable meaning, the ordinary meaning of the words plays a lesser role. The relative effects of ordinary meaning, context and purpose on the interpretive process may vary, but in all cases the court must seek to read the provisions of an Act as a harmonious whole.
Courts have interpreted this legal principle in at least two ways. At times, courts have interpreted the statement as referring to the plain text of the statute. Other times, courts distinguish “plain meaning” from “ordinary meaning”. The former references a literal textual approach. The latter references a contextual approach. Indeed, the Supreme Court in Canada Trustco Mortgage might have conflated these two difficult concepts. The Court’s statement appears to equate “plain meaning” and “ordinary meaning” in the sentence, “When the words of a provision are precise and unequivocal, the ordinary meaning of the words plays a dominant role in the interpretive process.” This statement suggests that the ordinary meaning of a word is capable of only one interpretation. This interpretation is exactly what the plain meaning is.
However, as Professors Siebrasse and Sullivan note, the ordinary meaning is not a literal, a-contextual approach that presumes text can have only one meaning. The plain meaning is a literal, purely textual approach. The ordinary meaning technique acknowledges the possibility of many meanings. It looks to the meaning that jumps out at the particular reader of the text. This is a contextual approach because it presumes text can have many meanings. In other words, the textual approach in Canada is a contextual approach.
In Kruger, the trial and appeal decisions engaged with this subtle distinction. Justice Rip at trial interpreted the words of s.10 and s.248 literally. Rip J found that there was no qualifying language in the definition of inventory. Any property the cost or value of which is relevant to the computation of business income is inventory. The interpretive form of this argument is persuasive given the language of the ITA.
One issue with this interpretation emerges only if we look to the textual argument’s corollary: What is not capital property is thus inventory. But there are several classes of property that are not inventory and not capital property, such as windfalls, depreciable property, Canadian resource property, and even mark-to-market property held by financial institutions. The literal interpretation implies something that is not true.
On the other hand, Justice Noël interpreted the definition of inventory according to its ordinary meaning i.e. that inventory is property held for sale. This approach followed the Supreme Court’s analysis in Friesen. Note, one will not find the words “held for sale” in the s.248 definition of “inventory”, nor in s.10(1), which is the operative provision here. Interestingly, they appear in only two places in the ITA: s.10(4) and s.10(5).
Justice Noël’s analysis was succinct, relying on the Friesen decision. One could view the interpretation that inventory is property held for sale as reading words into the statute, a sort of judicial legislating. Justice Noël characterized the remedy as such. One could also view it as reading down the statute, giving meaning to the very word “inventory”.
Reading down a provision of a statute is within a court’s jurisdiction, whereas reading in words generally is outside a court’s jurisdiction unless it is interpreting the constitution. Both remedies add words to the statute, but reading down is a form of justifiable paraphrase that restates the text of the statute in a more comprehensible manner. More precisely, reading down qualifies or restricts the language of the statute, and reading in expands the language.
As Professor Sullivan puts it,
7.13 At first glance, reading down and reading in may seem to be symmetrical techniques and remedies, two sides of the same coin. However, the courts are right to distinguish them and to be much more cautious in using the reading in technique or remedy. As an interpretation technique, reading down merely makes explicit what the court finds to be implicit in the legislative text. It is impossible for drafters to spell out every qualification or limitation that might appropriately apply in a given set of circumstances. Otherwise, provisions would go on for pages. Modern legislation is drafted in general terms, effectively delegating to official interpreters the work of adapting the language to particular facts and reading down its scope when there is a good reason to do so.
7.14 Reading in is different. It does not purport to operate within the scope of the legislative text, but rather to expand that scope to matters that are neither explicit nor implicit in the legislation. It expands legislation to matters that cannot come within any plausible understanding of the wording adopted by the legislature.
The effect and legitimacy of reading down a provision are evident from the analysis in Manrell v Canada, which the courts cited in Kruger. In that case, the Federal Court of Appeal interpreted the definition of “property” in s.248, and held that the words “a right of any kind whatever” are so broad one might interpret them as having an infinite meaning (Para. 50). So, the court interpreted the words according to their ordinary meaning, to qualify the breadth of the ITA’s s.248 definition of “property”.
Similarly, the words “any property the cost or value of which is relevant to the computation of business income” are words with a very broad scope. Indeed, cash held on hand is property the cost or value of which is relevant to the computation. Thus, the fair market value decline in a cash balance that is attributable to inflation ought to be a deductible expense under s.10. But it is not. Something restricts the language of s.248 and s.10.
The ordinary meaning of the word “inventory” is the only way to restrict the language. This interpretive technique excludes cash balances even though cash is property the cost or value of which is relevant to the computation of business income. One might view Justice Noël as having read down the definition of inventory to exclude property held for speculative purposes, such as Kruger’s derivatives.
A complicating factor for this view is that the ITA’s definition of “inventory” under s.248 suggests that the ITA’s definition is the sole, exclusive meaning of the word as it applies in the statute. This interpretation of the law of statutory interpretation is consistent with Parliament’s subsequent amendments, which left intact the definition of inventory under s.248 by introducing s.10(15). That section excludes derivative forward agreements from the application of s.10. On this view, one should interpret the definition of “inventory” literally. Yet, this view disregards the Supreme Court’s explicit statement that the words of a statute are always to be read in their ordinary and grammatical sense i.e. according to their ordinary meaning, which is a contextual reading.
Ultimately, the case is decided and Parliament has acted, so it is difficult to tell how courts will interpret “inventory” going forward. It is also difficult to tell how much precedential value Justice Noël’s reasoning has. What is clear, however, is that the reasoning is consistent with the Canadian approach to statutory interpretation and textual analysis.
The point here is that it is important to distinguish between ordinary meaning and plain meaning, and reading in and reading down. Whether one should read the words according to their ordinary or plain meaning is a difficult question, as highlighted by the Supreme Court’s statements above and the different way in which courts have approached textual analysis. In the trial decision, Justice Rip read the words plainly, disregarding the ordinary meaning of the text, in contrast to Justice Noël’s approach. As for reading in and reading down, the two are subtly distinct. At appeal, Justice Noël characterized the remedy as reading words into the statute, in my respectful view, potentially mischaracterizing the decision as an unjustifiable judicial amendment, rather than a justifiable judicial paraphrase. Simply changing the word “in” to “down” could have made Justice Noël’s decision more palatable.
I hope this note highlights the effect of the characterization of an issue, as well as the effect of subtle differences in the meaning of different words or terms. Mark-to-market, mark-to-model, fair value, and fair market value are complex accounting terms that are different from each other, as are reading in and reading down. Sometimes, one uses them interchangeably. But it is crucial to be precise when using technical terms, as conflation or confusion can arise, leading to adverse outcomes for a client’s case.
Below is a brief introduction to Parliament’s amendments and proposals following Kruger.
Parliament’s Response to Kruger
Parliament acted swiftly following the Kruger decision. In addition to enacting s.10(15), Parliament introduced other provisions and several proposals to amend the ITA.
Subsection 10(15) excludes derivative forward agreements from the operation of s.10, by deeming such agreements to not be inventory. Parliament also restricted losses claimed on derivatives agreements, by amending s.18 of the ITA. Paragraph 18(1)(x) precludes any deduction caused by a reduction in value of a derivative property if (a) the method used to value the property is the lower of cost or fair market value, (b) the property is described in s.10(15), and (c) the property is not disposed of. Parliament’s reasoning was that the application of s.10 to derivatives could destabilize the tax base.
The 2017 Budget proposed an elective regime for “eligible derivatives” held on income account, which will require taxpayers to elect whether they will use the mark-to-market or realization methods to report their taxable income. Once a taxpayer elects, the taxpayer can change their method only with the consent of the Minister. A derivative agreement is eligible if the taxpayer has produced audited financial statements that are prepared according to Generally Accepted Accounting Principles or the derivative agreement has readily ascertainable fair market value, and if the agreement is not capital property.
Parliament made other proposals, including a proposal to exclude eligible derivatives from the definition of eligible property under s.85 and s.97. The proposals are complicated and will need to be analyzed more to understand the effects.
 See Benoit Mandelbrot, The Misbehaviour of Financial Markets: A Fractal View of Financial Turbulence. See also JPM Asset Management, Non-Normality of Financial Returns: A Framework for Asset-Allocation Decision Making.
 See Ruth Sullivan, Sullivan on the Construction of Statutes. See especially Norman Siebrasse, “The Essential Elements Doctrine in Patent Infringement: Free World and Whirlpool in Light of Kirin-Amgen” at p. 9.
 Ruth Sullivan, Sullivan on the Construction of Statutes at §7.10ff.
 Sullivan supra, at §7.13-.14.