By Jules O’Donnell

What happens when a government decision-maker breaches a condition regulating the exercise of a statutory power? In what circumstances will an exercise of a power ­— for example, the issue of a licence, the refusal of a visa, or the seizure of property — be rendered invalid because of that breach? Alternatively, when is the breach considered a mere technicality?

In two recent cases, the High Court has revised the way it approaches these questions. Hossain v Minister for Immigration and Border Protection [2018] HCA 34, decided last year, established that administrative decisions are generally not to be invalidated on the basis of procedural errors that are immaterial. Put another way, immaterial errors are not, absent some contrary indication, jurisdictional errors. This proposition has been referred to as the ‘threshold of materiality’. Minister for Immigration and Border Protection v SZMTA [2019] HCA 3, decided in March this year, confirmed that the plaintiff must prove that the breach in question was material to the outcome in order for the decision to be set aside.

In this post, I argue that the primary appeal of the threshold of materiality is that it gives system-wide minimum content to the concept of jurisdictional error. This is significant because the Court has generally resisted developing general norms in this area, instead insisting that jurisdictional error is strictly tethered to the statutory scheme in question. A more generally applicable standard of jurisdictional error could, in some cases, provide much needed relief to courts that must otherwise resort to contestable speculation about legislative purpose.

I also highlight two remaining uncertainties surrounding the materiality threshold. First, the Court has not yet provided a clear justification for why it applies. Its doctrinal precariousness leaves it susceptible to being overwhelmed by exceptions (a number of which have already been hinted at). Second, the plaintiff’s new burden of proof may introduce complexity to judicial review proceedings. New forms of evidence may now be admissible, and courts may be called upon to consider matters that are typically reserved for merits review.

The difficulties of Project Blue Sky

Hossain and SZMTA should be read against a backdrop of an ongoing legal dilemma. Statutes creating administrative law powers rarely state the consequences of non-compliance with a condition regulating the exercise of those powers. When confronted by such an omission, courts must turn to the following passage of the High Court’s decision in Project Blue Sky (at [91]):

An act done in breach of a condition regulating the exercise of a statutory power is not necessarily invalid and of no effect. Whether it is depends upon whether there can be discerned a legislative purpose to invalidate any act that fails to comply with the condition. The existence of the purpose is ascertained by reference to the language of the statute, its subject matter and objects, and the consequences for the parties of holding void every act done in breach of the condition. Unfortunately, a finding of purpose or no purpose in this context often reflects a contestable judgment …

What this passage makes plain is that a determination of the precise consequences that flow from a breach of a statutory condition must be made by reference to the legislative purpose. In the abstract, such an approach seems sensible. Where the legislature has not spoken directly on an issue that requires resolution, the court must do its best to find an answer by adopting a purposive approach and drawing inferences from the matters that the statute does address.

The problem with this approach, however, is that it offers no baseline position to fall back on in the event that the statute creating the power provides limited guidance. The rigour of this test therefore depends on there being sufficient statutory cues from which to draw inferences about the consequences of non-compliance. A lot of faith is placed in the magic of statutory interpretation.

Compounding the complexity, the Project Blue Sky test does not necessarily produce binary results. As Edelman J observed in Hossain (at [67]), the test also requires consideration of what kind of breach of a statutory condition might invalidate a decision. The test of invalidity therefore comprises two questions:

1. Does the legislative purpose indicate that breach of this condition could render a decision invalid?
2. If so, do the purposes indicate that any breach of this condition is sufficient to render the decision invalid, or must the condition be breached to a certain degree, or in a certain manner?

This dilemma provides important context to the Court’s decision in Hossain and SZMTA. Of course, the Court in Project Blue Sky recognised at the outset that the invalidity test reflects a ‘contestable judgment’. But there is growing judicial disquiet over the difficulties of this approach. Indeed, recent appeals to the High Court on a Project Blue Sky question show the level to which the opinions of appellate judges differ on questions of legislative purpose.

The threshold of materiality

Both Hossain and SZMTA were appeals to the High Court from decisions of the Administrative Appeals Tribunal to refuse visas under the Migration Act. In both cases, the Tribunal was alleged to have breached express and implied conditions regulating the power to determine visa applications under the Act. Whether or not the alleged breaches impacted the outcome was a live issue in both cases and, at least in SZMTA, it was central to the Court’s decision to refuse relief. The facts and relevant provisions are otherwise lengthy, and for present purposes, need not be elaborated further.

In Hossain, the majority — consisting of Kiefel CJ, Gageler and Keane JJ — reminded us that labelling an error ‘jurisdictional’ is, and has always been, a judgment about the gravity of an error. There need not be a clear-cut philosophical difference between errors that are, and are not, jurisdictional. Rather, the distinction is functional; the point of making it is to determine whether an error is of such gravity that it invalidates a subsequent decision. Critically, the only way to determine whether a breach meets the requisite order of magnitude is to construe the statute.

The majority went on to recognise that this directive to ‘construe the statute’ requires more than simply reading the bare words of the provisions. A number of presumptions supplement consideration of the text, such as presumptions about procedural fairness, reasonable decision-making and the like, which the legislature is taken to accept as conditioning the exercise of public power (unless it has expressly provided otherwise).

Nestled in this list of familiar presumptions, the threshold of materiality rears its head. It is stated as follows (at [29]):

Ordinarily, a statute which impliedly requires that [a] condition … be observed in the course of a decision-making process is not to be interpreted as denying legal force and effect to every decision that might be made in breach of [that] condition. The statute is ordinarily to be interpreted as incorporating a threshold of materiality in the event of non-compliance

The majority noted that the threshold of materiality may be set ‘higher or lower’ depending on the particular statutory regime, but that — ordinarily — the threshold will not be met if compliance would have made no difference to the decision ultimately made.

A year later, in SZMTA, Bell, Gageler and Keane JJ confirmed the Hossain principle and reiterated that ‘materiality is … essential to the existence of jurisdictional error’ (at [45]). Their Honours also stated that ‘a breach is material to a decision only if compliance could realistically have resulted in a different decision’ (at [45]).

The majority in SZMTA also ruled that the question of materiality is an ordinary question of fact that is established through inferences drawn from evidence. The plaintiff bears the onus of proving this fact. In determining whether compliance could have made a difference, a court is to consider ‘what can be expected to occur in the course of the regular administration of the Act’ (at [47]).

Telling people where they stand

Nettle and Gordon JJ, in SZMTA, objected to the idea that materiality is an essential precondition to jurisdictional error. Their Honours said that judicial review plaintiffs are entitled to ‘know where they stand’ (at [88]) — that is, to expect that decisions will be valid and enforceable according to the statute, not according to some margin of error described as ‘materiality’. The answer to the question of whether jurisdictional error has occurred is always to be ‘found in the statute’ (at [83]).

There are two problems with this critique. First, it does not directly address the status of the threshold of materiality as a legislative presumption. No one doubts that judicial review plaintiffs are entitled to have decisions made according to ‘the statute’. But the majority’s point is that the threshold of materiality is presumptively part of the statute — it is an application of ordinary principles that stand unless the legislature displaces them. In order to dismiss the threshold on these terms, the minority needed to explain why — unlike all other common law presumptions — this presumption rests on an indefensible hypothesis about what the legislature intended.

Second, the critique amounts to an unsatisfactory repetition of the need to apply Project Blue Sky — that is, to find the contents of jurisdictional error ‘in the statute’. Apparently, this is necessary in order to tell people ‘where they stand’. But if this attack is levelled at the threshold of materiality, it would apply with even greater force against the Project Blue Sky test. Do the minority judges really expect the Project Blue Sky test to tell people where they stand with respect to jurisdictional error? At most the test tells people that their legal rights stand on contestable inferences about a matter that is not even addressed in the statute.

In fairness, the majority judgment could have been clearer about the genesis of the presumption. Common law presumptions are generally supported by some kind of ‘working hypothesis’ which explains why the legislature can presumptively be taken to mean something until it indicates the contrary.

In Hossain, the majority’s hypothesis may be interpreted as follows:

1. Jurisdictional error has long been recognised as an assessment about whether the administrative error was serious enough to warrant judicial intervention (at [25]);
2. Ordinarily, an error would not be considered serious — by any measure — if it had no tangible impact on the process in which the error occurred. Otherwise the concept of seriousness would have little work to do;
3. The legislature is presumed, absent any indication to the contrary (at [30]), to intend for ordinary notions of seriousness to apply with respect to the consequences of a procedural error (at [29]).

The majority does not expressly set out proposition (2) — rather, this step is implicit in the shift from (1) to (3). But proposition (3) may also be a contestable interpretation. I derive (3) from the fact that the word ‘ordinarily’ is used three times when describing the threshold of materiality, implying that the application of the threshold springs from a common sense notion of what constitutes ‘seriousness’.

In any case, there is an assumption at work in the majority’s reasoning that a judgement about the gravity or seriousness of an error is coextensive with a judgement about whether the error caused something different to occur (particularly something unfavourable to an applicant). A lack of explanation for this assumption somewhat undermines its persuasiveness.

A number of developing exceptions add to this doctrinal precariousness. The threshold of materiality may not apply where:

1. The statute indicates that a lower or higher threshold should be set (see above);
2. A decision maker is required to make a decision by reference to one specified criterion, but addresses themselves to the wrong criterion in circumstances where choosing that wrong criterion made no difference (Nettle J in Hossain at [40]);
3. There is a need to protect the dignity of individuals engaging in administrative processes, which requires invalidating a decision even if the breach was inconsequential (Edelman J in Hossain at [72]).

If the Court keeps discovering instances in which a jurisdictional error appears to occur absent materiality, it may need to reconsider whether the gravity of an error should be presumptively assessed by reference to the impact it has on the decision-making.

A heavy burden?

In SZMTA, Nettle and Gordon JJ make two sharp objections to the fact that the threshold of materiality shifts the burden of proof onto the plaintiff.

First, their Honours stated (or rather repeated) that the shifted burden of proof throws into doubt the ‘fundamental principle that a statutory power is to be exercised under … the terms of the statute’. The terms of the statute are not to be subjected to ‘some level of materiality that is identified after the decision is made’ (at [93]).

Second, their Honours added that the shift of the burden of proof fundamentally alters the nature of judicial review. This is because (at [95]):

Instead of a court concluding that an act or omission constitutes an error going to jurisdiction … it would become a form of merits review where jurisdictional error is found only if the breach is material to the applicant for review because it has denied that applicant the possibility of a successful outcome.

The first part of the criticism — about a power needing to be exercised according to the terms of the statute — contains the same oversight discussed earlier. The fact that the plaintiff has to establish materiality as a matter of fact after the decision is entirely appropriate if materiality is accepted as a statutory precondition of jurisdictional error. The question is thus whether the presumption is appropriate, which is not examined in detail.

The second part warns that judicial review may turn into a form of merits review. What does this mean? There are two points to consider here.

First, a plaintiff now has to put forward evidence that goes to materiality in a judicial review trial. If materiality is a fact in issue, then evidence that supports an inference about materiality becomes relevant — and therefore admissible — in the proceeding.

The admissible evidence in SZMTA was not drastically different from what might be considered in a trial before materiality was a fact in issue. But other cases may be different. Plaintiffs may now attempt — for example — to tender testimonial and documentary evidence about how they intended to conduct themselves if the breach did not occur.

Second, and perhaps more significantly, the court must draw inferences from the above evidence about how the breach impacted the outcome of the proceeding. The question ‘what would have happened otherwise’ could require the judge to consider a range of non-legal matters that influence a decision, such as government policies, regular practices, or even character assessments. If evidence has been put before the judge (but not the original decision-maker) about how a plaintiff would have argued, the judge may need to consider how convincing that evidence would have been and how it affects the merits of the case.

It is too early to say for certain whether this fear is well-founded. I am inclined to think that it is overstated. The threshold of materiality has been set quite low — reserved for those cases where the breach is clearly inconsequential, and could have had no plausible influence on the outcome.

Recall that all the plaintiff has to show is that they lost the possibility of a successful outcome; they do not actually have to establish that the outcome would have been different. In SZMTA, the majority suggested that what needs to be shown is that compliance could ‘realistically’ make a difference. As soon as doubt is introduced as to the difference that compliance could have made, there is a good argument that a judge would not need to go any further to recreate a hypothetical merits hearing. Doubt about the impact of the error tips the balance of probabilities in favour of the view that compliance could — realistically — have made a difference.

Conclusion

The threshold of materiality attributes to the legislature a view about when an error is too minor to warrant judicial intervention. Some, including two members of the High Court, will think that this oversteps the judicial role to strictly defer to legislation when deciding whether an error is jurisdictional.

In my view, an approach which gives some minimum content to jurisdictional error is better than one by which courts treat the ‘seriousness’ of a procedural error as something that is reinvented every time parliament passes an act. A commitment to legislative purpose as the only cue to the boundaries of jurisdictional error appears restrained and deferential, but in practice it requires a lot of judicial conjecture, as the Project Blue Sky experience demonstrates. A baseline common law presumption clearly declared by the Court is far more transparent. Whether the threshold of materiality is the best common law presumption to apply in the circumstances remains to be seen.

AGLC4 Citation: Jules O’Donnell, ‘A Threshold of Materiality for Judicial Review: Common Sense or Injustice? Hossain and SZMTA‘ on Opinions on High (11 June 2019) <https://blogs.unimelb.edu.au/opinionsonhigh/2019/06/11/odonnell-hossain-szmta/>.

Jules O’Donnell is a recent JD graduate of Melbourne Law School with a Bachelor’s degree in Literature. He is interested in administrative, constitutional and criminal law. All views are his own.