Explaining Numbers: Two Recent Cases

Apologies again for the long silence. Things remain rather chaotic, as we find ourselves between Canada and Cambridge, a move made all the more difficult by Brexit, and I have several writing commitments which have been filling my limited free time. These are now much closer to being fulfilled (or, at least, the deadlines are quite close…) and so intermittent blogging may resume in the next few weeks, though I do not expect to be at full throttle until at least September. I hope to be able to formulate some thoughts on theBrexit litigation, the suggestion that there is now a ‘Chevron Step 0.5’ in American administrative law, and a variety of interesting cases that have caught my eye over the past few months.*

For now, let me simply mention two interesting recent cases that highlight the need for administrative decision-makers to give reasoned explanations of their decisions. Both cases involve decision-makers who apparently plucked numbers out of thin air, much to the displeasure of reviewing courts.

In Re JR55, [2016] UKSC 22, one of the Ombudsmen for Northern Ireland — the Complaints Commissioner (see paras. 2-3) — had recommended that a general medical practitioner pay £10,000 to the widow of a patient who died, on the basis that several failings in the doctor’s care of the patient amounted to maladministration. Much of Lord Sumption’s decision for a unanimous United Kingdom Supreme Court focused on the limits on the Complaints Commissioner’s jurisdiction, which did not extend to making recommendations against private individuals such as the doctor in this case. But Lord Sumption also held that the £10,000 figure had been “plucked out of the air” and was, as a result, unlawful:

The Commissioner’s recommendations, in those cases where he is entitled to make them, are discretionary and he has more latitude in arriving at a figure than a court would have. But a monetary recommendation, like any other, must be rational, and it must be explained. The only explanation proffered is that the £10,000 should be paid “in respect of the clearly identified failings in the care provided to [the patient] and the events which consequently followed.” The report does not explain why these failings warrant a payment of £10,000 or how that figure has been arrived at. It does not say whether Mr or Mrs R suffered any loss by the failings for which the £10,000 should be treated as compensation. Some of the failings, notably the failure to take more urgent action on 6 January 2009 are found to have made no difference and others, such as the events which followed R’s death could not in the nature of things have done so. It is possible that the recommendation was intended as a solatium for injured feelings, but the report does not say so, and in the absence of explanation £10,000 seems to be an excessive amount to recommend on that basis. On the face of it, the figure has simply been plucked out of the air (at para. 30).

In Canada v. Kabul Farms Inc., 2016 FCA 143, penalties imposed by the Director of the Financial Transactions and Reports Analysis Centre of Canada for three violations of money laundering legislation were struck down by the Federal Court of Appeal. The Director has to follow a three-step methodology set out in the applicable regulations: classification of the offence as minor, serious or very serious; identification of the penalty range for these classifications (upper limits of $1,000, $100,000 and $500,000 respectively); and selection of an appropriate penalty based on criteria set out in the primary legislation and regulations. Here, the Director ultimately imposed penalties of $2,000, $3,000 and $1,000 for the three “serious” breaches in question:

He chose an amount within the [penalty] range to reflect the criterion of harm—here, $50,000, $75,000 and $25,000 in the case of the three violations. Then he reduced each amount by 20% to take into account the criterion of the respondent’s compliance history, and further adjusted that amount by 95% to reflect the criterion of encouraging the respondent to comply, not to punish the respondent. Here, appropriately, the Director considered the respondent’s ability to pay (at para. 15).

One might have thought that the proper application of a prescribed decision-making methodology in a highly discretionary statutory scheme would surely be upheld as reasonable. But, as Stratas J.A. explained: “A fact-based, discretionary decision made on the basis of proper methodology is not automatically reasonable. The reviewing court must also be satisfied that the administrative decision-maker has made an acceptable and defensible decision on the particular evidence before it” (at para. 20). Here, the figures might have been plucked from thin air (much like the Complaints Commissioner’s figure in the UK case) (at para. 28). There was no explanation of why the Director had chosen the specific amounts within the penalty range, or the 20% or 95% adjustments; indeed, there was evidence in the record pointing in favour both of leniency and harshness:

For all we know, the 20% and 95% percentages might have been plucked out of the air or adopted for reasons extraneous to the legislation. Maybe the Director did not investigate the case enough to gather the evidence necessary to support a decision. We simply cannot tell. We are left in the dark. In this case, we are a reviewing court that cannot review (at para. 32).

Counsel for the Director attempted to explain the penalties on the basis of an unpublished formula, but Stratas J.A. concluded that an unpublished formula could not legitimately be relied upon in judicial review proceedings (at paras. 36-39 and, especially, para. 47), warning in addition that the unpublished formula seemed to conflict with the legislation (at paras. 40-42) and that withholding information about the basis for calculation could be a breach of procedural fairness (at paras. 43-45). Ultimately, “jotting down a few words of explanation in the Director’s summary of calculation about why he chose the figures for the base amounts and reductions—something that would perhaps have taken a few seconds—probably would have sufficed as far as enabling this Court to review the Director’s assessment of penalties is concerned” (at para. 49).

Administrative decision-makers: beware check-box decision-making! These two cases indicate that picking a figure that may well be rational or reasonable is not enough. A decision-maker must also be able to explain the figure arrived at in terms of the evidence before it and the relevant statutory framework.

* One interesting case is Groia v. The Law Society of Upper Canada, 2016 ONCA 471. However, Mr. Groia was represented during the underlying disciplinary proceedings by Lerners LLP in Toronto while I was employed by the firm (which continues to represent him). I think it would be appropriate for me to refrain from comment on the case before it has run its course.

This content has been updated on July 17, 2016 at 05:01.