The “no substantial difference” test for judicial review remedies – a snapshot of the trends

Mustaqim Iqbal (University of Oxford) and Lee Marsons (University of Essex)

Reforming judicial review remedies is back on the legislative agenda. The last time similar proposals were suggested, the result was section 84 of the Criminal Justice and Courts Act 2015 (CJCA), which inserted section 31(2A)-(2C) and (3C)-(3F) into the Senior Courts Act 1981. This provided that a judge should refuse relief “if it appears…to be highly likely that the outcome for the applicant would not have been substantially different” if the unlawful action had not occurred and also provided that a judge should refuse permission to bring a judicial review in the same circumstances. We call these provisions the “no substantial difference” (NSD) test.

Particularly, in this blog we wish to provide a snapshot of the courts’ approach as of 2021, updating and extending Rebecca Williams’ examination of the provisions in 2017. Notably, we identify some core trends in the courts’ understanding of this duty and identify what factors the courts consistently appear to take into account when making NSD assessments. These include: the availability of written evidence expressly foreshadowing the inevitability of the decision; how “close” the decision was; the number of options available to the decision-maker; and the seriousness of the legal error.

The politicised nature of the duty

Immediately after enactment, judges were already aware of the politicised nature of the NSD test – namely, how is a judge to make a guess about whether a decision-maker would have reached the same conclusion, had they not committed the legal error? For the earliest example, see comments made by Blake J in R. (Logan) v Havering London Borough Council [2015] EWHC 3193 (Admin) (at [59]).This sentiment has not been lost and, if anything, courts are now even more aware of the political guesswork required in discharging the statutory duty. The Court of Appeal provided the most in-depth discussion of this issue at [273] in Plan B Earth v Secretary of State for Transport [2020] EWCA Civ 214:

“In particular, courts should still be cautious about straying, even subconsciously, into the forbidden territory of assessing the merits of a public decision under challenge by way of judicial review … Courts should also not lose sight of their fundamental function, which is to maintain the rule of law.“

Earlier this year, in Pearce v Secretary of State for Business, Energy and Industrial Strategy [2021] EWHC 326 (Admin), we saw the practical impact of a court’s awareness of the politicised nature of the duty. Responding to an argument from the Secretary of State to withhold a quashing order because a development consent order for a windfarm would have been granted irrespective of legal errors, Holgate J at [156] said:

“[T]he court does not have any notion as to what the evaluation of cumulative impacts by the Defendant would have been if he had considered the matter … It would be impermissible for the court to make findings on that issue for itself. To do that would involve entering forbidden territory.“

The judges are acutely aware of the considerable element of conjecture involved in the “evaluation of the counter-factual world in which the identified unlawful conduct … is assumed not to have occurred” (Sales LJ at [89] in R. (PCSU) v Minister for the Cabinet Office [2017] EWHC 1787 (Admin)). This trend has not waned and has become a standard motif relied on by the judges.

High threshold

Initially, the language of “highly likely” seemed to lower the threshold for the refusal of relief, as compared to the prior Simplex “inevitability” test – something can be highly likely without being inevitable, after all. This line was repeated by Kate Grange QC at [52(ii)] in R. (Cava Bien) v Milton Keynes Council [2021] EWHC 3003 (Admin), and in Glatter v NHS Herts Valleys Clinical Commissioning Group [2021] EWHC 12 (Admin), Kerr J said at [98] that the standard lay:

somewhere between the civil standard (the balance of probabilities) and the criminal standard (beyond reasonable doubt)”

However, some recent authority suggests that the “highly likely” threshold is not materially different to the Simplex standard. Even in Cava Bien, it was noted that the standard still “sets a high hurdle” ([52(ii)], despite the comment about the threshold being lowered.

In this regard, the recent Court of Appeal judgment in R. (Hudson) v Royal Borough of Windsor and Maidenhead [2021] EWCA Civ 592 is helpful. The applicant argued that the test was not the “highly likely” standard but that used in relation to EU obligations – the inevitability test, identified in R. (Champion) v North Norfolk District Council [2015] UKSC 52. Coulson LJ stressed that while the two tests appear different, they likely operate in the same manner in practice ([74]). He said at [79]-[80]:

[W]hilst I can see that the two tests are potentially different … I consider that the difference may have become overstated. … once an analysis has been undertaken … the precise formulation of the test may not matter: in the real world, it would be a very unusual case in practice in which the court’s consideration of the seriousness of the breach and any prejudice caused was sufficient to establish that it was highly likely that the outcome would not have been substantially different, but insufficient to establish that the result would have been the same. That may well explain why the authorities demonstrate that the hypothetical difference in the formulation of the test has not given rise to a real issue in practice.”

“A change of words but not necessarily of substance” appears to be the message.

Taking the duty seriously

Since the early caution that followed Logan, what has become clear is that the duty is not one that courts are shirking. Whether rejecting or accepting that the test has been met, judges do neither in a cursory manner; the initial reticence in applying the test, identified by Williams, has been displaced by greater confidence in balancing various considerations. Concerns about granting “non-remedies”, like that seen in Logan [(58)] and R. (Hawke) v Secretary of State for Justice [2015] EWHC 3599 (Admin) [(66)], have largely disappeared.

Behind this increasing confidence may be a better appreciation of the objectives of the provision. In Gathercole v Suffolk CC [2020] EWCA Civ 1179, Coulson LJ noted at [38] that the NSD test is:

designed to ensure that, even if there has been some flaw in the decision-making process which might render the decision unlawful, where the other circumstances mean that quashing the decision would be a waste of time and public money (because, even when adjustment was made for the error, it is highly likely that the same decision would be reached), the decision must not be quashed … the provision is designed to ensure that the judicial review process remains flexible and realistic.”

Therefore, the main impact of the NSD test has not necessarily been to lower the standard for the refusal of relief, but rather to transform the previous exception into the norm. The test may be more or less the same but that test is now being applied as a matter of course. No longer is the withholding of a remedy left solely to the discretion of the court; it is now a statutory requirement that is being entertained virtually whenever argued by counsel for the defendant, and can be considered in every application for judicial review. There are similarities here with the debate related to suspended and prospective-only quashing orders, which according to Public Law Project’s research can already be made but which the government seeks to make the norm.

Factors taken into consideration

As often in public law, much hinges “on the particular facts of the case before the court” (Plan B Earth, [273]). In EFW Group Ltd v Secretary of State for Business, Energy And Industrial Strategy [2021] EWHC 2697 (Admin), Dove J at [74] confirmed that NSD analysis “will vary from case to case and no single analysis will be capable of resolving all of the many scenarios which courts will have to address”.  Nevertheless, patterns have slowly emerged in the factors that judges consider when assessing the NSD test.

1. Written evidence

The most persuasive evidence favourable to public bodies arguing that relief should be refused is written evidence signalling that, irrespective of the decision-maker’s errors, no substantially different outcome would have been reached. This might be planning officer reports, written correspondence or expert evidence.

There are many recent examples. Most recently, relief was granted to the claimant retail company in Cava Bien partly because pre-decision email exchanges confirmed that the legal error in question was “the central focus” ([55]) motivating the defendant council’s decision to exclude the claimant from a grant to support businesses during the Covid-19 pandemic. In R. (Advearse) v Dorset Council [2020] EWHC 807 (Admin), relief was denied despite legal errors in the Planning Officer’s Report ([31]), mainly because an Inspector’s Report on the area’s Local Plan – a separate document – made the overwhelming public benefits of a development of the type proposed abundantly clear. In R. (Blundell) v Secretary of State for Work and Pensions [2021] EWHC 608 (Admin), relief was similarly denied because an ex post facto equality impact assessment made it clear that the Secretary of State would have still applied the same policy of deducting a fixed rate from a person’s Universal Credit allowance to pay off criminal fines ([123]-[134]), despite the failure at the time to discharge the public sector equality duty under section 149 of the Equality Act 2010.

2. How “close” the decision was

Courts will also evaluate how close a decision-maker was to reaching a different outcome. In Gathercole, relief was denied because the decision taken – to grant planning permission for a new primary school – was a near inevitability. Despite the unlawful failure of the officer’s report to consider the impact of overhead aircraft noise on children with protected characteristics, there was no site for a school in the area that would not have subjected children to this noise, and the school had already been designed to mitigate the sensory impact ([39]-[43]).

On the other hand, in R. (Buckley) v Bath and North East Somerset Council [2018] EWHC 1551 (Admin), the argument from the defendant council was rejected, because outline planning permission was granted for a controversial development on a small majority vote of 5-4. Thus, the decision was far from obvious, and all that would have been needed was for one committee member to have been swayed by the unlawfully omitted information, relating to loss of homes for the elderly and disabled ([41]-[43]).

3. Number of options open to the decision-maker

R. (EG) v Parole Board [2020] EWHC 1457 (Admin) was a case concerning an applicant alleging an unlawful breach of his Article 5(4) ECHR right to timely review of his continued detention. May J rejected the defendant’s argument that even if the applicant had been dealt with promptly, it was inevitable that he could not have had an effective hearing due to adjustments that the applicant’s learning difficulties required.

May J granted relief partly because there were at least three different points in the applicant’s review process when the defendant could have intervened to make reasonable adjustments. Considering the multiple avenues open to the Parole Board, it was far from “highly likely” that a hearing would inevitably have been unfair.

4. Seriousness of the legal error

An interesting corollary of the Court of Appeal’s comments in Hudson – discussed above – is that the seriousness of the legal error may be a relevant factor when applying the NSD test. When assessing whether relief should be withheld for a procedural defect in the Environmental Impact Assessment process, Lord Carnwath in Champion at [58] said, in relation to EU law, that the court should “take account of the seriousness of the defect invoked and the extent to which it has deprived the public concerned of the guarantees designed to allow access to information and participation.”

Dove J applied the same approach with regard to EU Law in Canterbury City Council v Secretary of State for Housing, Communities and Local Government [2019] EWHC 1211 (Admin) (at [84]). In Hudson at [79], Coulson LJ said (emphasis added):

When considering whether it is highly likely that the outcome would not have been substantially different (or whether the outcome would not have been different, or even whether the outcome would have been the same), what matters in each case is the seriousness of the failure or breach …”

Thus, it would appear that if withholding relief would result in a particularly egregious breach of human rights, civil liberties, public law principles, or the purposes behind the legislative scheme, the facts may favour the granting of relief.

The discussion in Hudson helps to explain the trend, identified by James Maurici QC and Admas Habteslasie ([54]-[57]), of courts not being persuaded by defendants’ NSD arguments in cases involving unlawful detention and infringements of ECHR rights. From this, we can infer that, despite increased awareness of the normative underpinnings of the NSD test, courts are still eager to fulfil their function of upholding the rule of law.

Conclusion

It is clear that judges are now more familiar with the NSD test than when it was introduced in 2015. In particular, there are a range of factors that are consistently invoked in a court’s consideration of the test and these include: the availability of written evidence expressly foreshadowing the inevitability of the decision; how “close” the decision was; the number of options available to the decision-maker; and the seriousness of the legal error. The most interesting aspect of the NSD test, however, is how, while it may not have materially affected the practical test for refusal of relief, it has turned what was previously left to judicial discretion into a statutory norm. What was a rarity is now the rule. There are lessons here in relation to the government’s current reform agenda, where, though suspended and prospective-only quashing orders may or may not be permissible at common law, there is an attempt to change what is rare into the presumptive remedy.