Bloomsbury Institute Ltd, R (On the Application Of) v The Office for Students

This is an appeal from the order of Cavanagh J dated 12 March 2020. By that order, the judge held that the Office for Students (“OfS”) was entitled to refuse Bloomsbury’s application to be registered with the OfS. If Bloomsbury is not registered with the OfS, its courses cannot ordinarily be “designated”, its students cannot ordinarily obtain student loans and Bloomsbury will be unlikely to survive as a higher education provider.

Bloomsbury challenged the OfS’ decision by way of an application for judicial review on a large number of grounds before Cavanagh J, who rejected them all in a judgment of great clarity and thoroughness running to 343 paragraphs. On 21 May 2020, Lewison LJ granted Bloomsbury permission to appeal on three grounds only. The issues before us have accordingly been far more focussed than those which were aired at first instance. No complaint is made of the judge’s narrative of the background, and I adopt with gratitude much of what he said.

Bloomsbury was founded in 2002. It is a private educational establishment but its profits are reinvested in the college. It has approximately 2,000 students, many of whom are from disadvantaged backgrounds: approximately 85% are mature, 66% are BAME, 16% are disabled and 90% come from families earning less than £25,000 per annum. Many of them have come to higher education from a non-traditional route. In

Bloomsbury is an “Alternative Provider” (“AP”) of education because it does not receive direct grant funding from funding councils, as do traditional universities. In order to survive financially, APs need to be “designated” to enable their students to access student loans. Bloomsbury’s courses were first designated by the Secretary of State in 2010 or 2011. Bloomsbury received some positive appraisals in 2015, 2016 and 2017, but in February 2016 and August 2018 it was issued with “improvement notices” in relation to the continuation rates of its students, that is to say the percentage who proceed from year one of their courses to year two.

Responsibility for designation was transferred from the Department for Education to the Respondent (the OfS), which was created by the Higher Education and Research

One of the most important conditions of registration laid down by the OfS (though it seems to me to be almost impossible for any institution to achieve in practice) is Condition B3, that of the institution “securing successful outcomes for all its students”. The OfS did not consider that Bloomsbury had performed sufficiently well in relation to two criteria or “indicators” relevant to Condition B3, namely continuation rates from year 1 to year 2 (“continuation rates”), and rates of progression to professional employment or post-graduate study (“progression rates”). The provisional decision letter contained an Annex explaining the OfS’ methodology. This stated that the OfS relied on three data indicators – continuation and completion rates, degree and other outcomes, and progression rates – and used “split metrics” to assess performance within each indicator for students from different demographic groups. The Annex stated that each of the three data indicators was considered by reference to “baselines” which had been determined by the OfS for that indicator, but did not say what the baselines were.

Bloomsbury made representations about the provisional decision but, on 23 May 2019, the OfS wrote to Bloomsbury refusing its application for registration. The effect of the decision was that Bloomsbury could not take on any new UK students, unless they were wholly self-funded, and, as the judge put it, if the decision stands, “in all probability, Bloomsbury will have to close”.

Section 1 of HERA establishes the OfS as the body responsible for regulating all higher education in England. Section 2 sets out its general duties. Section 3 requires the OfS to establish and maintain a register of higher education providers; the OfS must register an institution if it satisfies the initial registration conditions. Section 4 requires the OfS to notify a provider before taking a final decision declining to register it, and give the provider the opportunity to make representations.

Section 5 requires the OfS to determine and publish the initial and ongoing registration conditions. Section 5(5) provides:

“(5)

Before determining or revising the conditions, the OfS must, if it appears to it appropriate to do so, consult bodies representing the interests of English higher education providers which appear to the OfS to be concerned.”

Section 7 requires that initial registration conditions are proportionate to the risk posed by the provider. Section 23 requires the OfS to assess the quality of (and standards applied to) higher education provided by institutions which have applied for registration.

Section 75 requires the OfS to prepare, consult on and publish a “Regulatory Framework” (“RF”), providing as follows:

“(1)

The OfS must, from time to time, prepare and publish a regulatory framework.

(2)

The OfS must have regard to it when exercising its functions.

(3)

The regulatory framework is to consist of—

(a)

a statement of how it intends to perform its functions, and (b) guidance for registered higher education providers on the general ongoing registration conditions.

(4)

The statement under subsection (3)(a) must set out how the OfS intends to perform its functions in relation to a registered higher education provider in proportion to the OfS's assessment of the regulatory risk posed by the provider.

(5)

“Regulatory risk” means the risk of a breach of the provider's ongoing registration conditions.

(6)

Guidance under subsection (3)(b) must include guidance for the purpose of helping to determine whether or not behaviour complies with the general ongoing registration conditions.

(7)

The guidance may in particular specify—

(a)

descriptions of behaviour which the OfS considers compliant with, or not compliant with, a general ongoing registration condition;

(b)

factors which the OfS will take into account in determining whether or not behaviour is compliant with a general ongoing registration condition.

(8)

Before publishing a regulatory framework under this section the OfS must consult—

(a)

bodies representing the interests of English higher education providers,

(b)

bodies representing the interests of students on higher education courses provided by English higher education providers, and

(c)

such other persons as it considers appropriate.

(9)

Where a regulatory framework is published, the OfS must send a copy of it to the Secretary of State who must lay it before Parliament.”

Pursuant to its obligations under s 75, the OfS published the RF in February 2018 after consulting the bodies listed in s 75(8). The RF was approved by the OfS Board. It contains the initial registration conditions with which higher education providers must comply if they are to be registered, grouped into Categories A to E. Category B conditions concern quality, reliable standards and positive outcomes for all students. Only Condition B3 is relevant to the present appeal.

The following paragraphs of the RF describe the OfS’ general approach to assessing higher education providers:

“8.

The regulatory approach is designed to be principle-based because the higher education sector is complex, and the imposition of a narrow rules-based approach would risk leading to a compliance culture that stifles diversity and innovation and prevents the sector from flourishing. This regulatory framework does not therefore set out numerical performance targets, or lists of detailed requirements for providers to meet. Instead it sets out the approach that the OfS will take as it makes judgements about individual providers on the basis of data and contextual evidence.

…

13.

The OfS is committed to adopting and contributing to best regulatory practice. It will comply with the Regulators' Code, and in developing this regulatory framework the OfS has consulted widely, drawn on best practice, and sought to learn from the latest in regulatory theory.

14.

The OfS's approach to regulation puts informed student choice and institutional autonomy at its heart. It sees the dynamic of providers responding to informed student choice as the best mechanism for driving quality and improvement, and will regulate at the sector level to enable this. The OfS will regulate at provider level to ensure a baseline of protection for all students and the taxpayer. Beyond that threshold the OfS will encourage and enable autonomy, diversity and innovation…

Ensuring a minimum baseline of quality for all and promoting excellence and innovation beyond that baseline

42.

The conditions of registration for quality and standards that apply to individual providers are designed to ensure a minimum baseline of protection for all students and the taxpayer…

…

95.

The initial conditions of registration are designed to mitigate the risk that the OfS is not able to deliver its four primary regulatory objectives. The conditions are 'baseline requirements', i.e. the minimum level a provider must achieve to be registered. The conditions are expressed in terms of the outcomes that the OfS wishes to see, rather than the particular approach that a provider might take to achieve such outcomes.

Lead indicators

136.

The OfS will identify a small number of lead indicators that will provide signals of change in a provider’s circumstances or performance. Such change may signal that the OfS needs to consider whether the provider is at increased risk of a breach of one or more [of] its ongoing conditions of registration. These indicators will be based on regular flows of reliable data and information from providers and additional data sources, and will include information about outcomes for students from different backgrounds. Lead indicators are likely to include, but not be limited to, the following: ... continuation and completion rates; ... graduate employment and, in particular, progression to professional jobs and postgraduate study.

137.

The lead indicators are likely to show changes that might not, in themselves, reveal areas of weakness or concern for an individual provider, but simply flag possible increased risk, such as a rapid increase or decrease in student numbers. The OfS will not use crude ‘triggers’ or performance thresholds to monitor risk, preferring a more flexible approach that takes into account the context for an individual provider.

138.

Absolute performance against an indicator will form part of the overall context for assessing risk. For example, when monitoring continuation rates, a decrease for an individual provider could mean performance had worsened. However, levels of absolute performance need to be considered in the context of performance across the sector as a whole and might be considered to be of less concern in the wider context.

139.

The OfS will seek to ensure that the selection and specification of lead indicators allow the identification of possible increased risk before this crystallises. Indicators that provide strong signals of likely future risk (for example significant shifts during the student recruitment cycle) and data trends over time will be more useful than data that retrospectively reveals where problems have already occurred (unless those problems have not previously been identified).

140.

The OfS will ensure that its lead indicators allow it to monitor a provider’s performance for all students from all backgrounds, for example by splitting student outcome indicators for different student characteristics. The OfS will also pay particular attention to outcomes achieved for students studying at different levels and in different modes (e.g. undergraduate/postgraduate).”

The OfS’ approach to assessing performance under Condition B3 is dealt with in paragraphs 340, 350 and 352 in particular. Paragraph 340 states that in judging whether a provider is delivering successful outcomes for its students, the OfS may consider:

“a.

A range of student outcomes indicators, broken down to show outcomes for students with different characteristics that include, but are not limited to: (i) Student continuation and completion rates; (ii) Degree and other outcomes, including differential outcomes for students with different characteristics; (iii) Graduate employment and, in particular, progression to professional managerial jobs and postgraduate study…”

Paragraph 350 states that each provider must meet a minimum level of performance:

“350.

Where the provider has a track record of delivering higher education, the OfS itself will assess whether the provider is able to satisfy condition B3. The evidence used will consist of the actual performance of the provider over time rather than its performance when compared to a sector-adjusted benchmark, although the context in which the provider is operating will be taken into account. This approach is designed to ensure that a minimum level of performance is used to determine whether a provider may be registered (taking into account the context of that provider), rather than a view of the provider’s performance as compared to other providers. The

OfS will take into account the impact of a provider’s performance on students with different equality characteristics in assessing whether or not the provider meets the minimum level of performance. Where the OfS has concerns, but nevertheless decides that the provider may be registered, it may require the provider to address any issues in its access and participation plan before it is willing to approve the plan.”

Paragraph 352 states, in the context of reviewing registration conditions after initial registration, that the OfS will not set an explicit numerical target for a provider whose performance has been causing concern:

“352.

Once registered, a provider for which the risk of noncompliance with its conditions of registration for quality and standards is considered to be low will be monitored using lead indicators. These indicators will normally reflect the actual performance of the provider over time rather than its performance when compared to a sector-adjusted benchmark. However, this approach will not involve setting an explicit numerical target for, for example, continuation. An indicator is intended to signal to the OfS that further regulatory investigation may be necessary.”

The OfS published three regulatory advice documents at the same time as it published the RF. Regulatory Advice 2 (“RA2”) sets out the evidence requirements in relation to Condition B3 at paragraphs 128 to 130:

“128: For Condition B3 all providers will be assessed against the following indicators: (a) Student continuation and completion indicators; (b) Degree and other higher education outcomes, including differential outcomes for students with different characteristics; (c) Graduate employment and, in particular, progression to professional jobs and postgraduate study.

129: … The indicators will be constructed from existing datasets and you will not be required to submit any new data…

130: We will consider your actual performance over time rather than your performance when compared to a sector-adjusted benchmark. This is to ensure that a baseline level of performance is used to determine whether you may be registered, rather than a view of your performance as compared with other providers. We will take account of the context in which you operate, such as the type of provision you offer, when making judgments about your performance.” [emphasis added]

In May 2018, a document called the Decision-Making Guidance (which I shall refer to as “the DMG”) was prepared either by or under the supervision of Ms Susan Lapworth, the OfS’ Director of Competition and Registration. The DMG’s stated purpose was to set out the framework that the OfS would use to make registration decisions. The DMG states that it was intended for use by assessors in the OfS’ registration team, those making decisions about individual providers, and those with responsibility for the oversight and effective operation of the registration process. Paragraphs 2 and 4 state:

“2.

This is a confidential internal document intended only for those listed above. It is not intended for wider sharing, or for publication, as this would inhibit the OFS’s ability to undertake its regulatory functions effectively.

….

4.

The framework provides guidance and indicators to support assessors to reach these judgements but it is not intended to be a rigid tool that must be followed to produce a “correct” answer.”

The DMG sets out the way in which assessors will evaluate each provider, which, as far as I understand, essentially boils down to three stages.

The first is the “initial baseline analysis”. Data is collected for the indicators set out in para 340 of the RF: continuation and completion rates, degree outcomes, and progression rates. The data is broken down by level of study (such as “first degree” or “PhD”); so, for example, an assessor will be able to examine continuation or progression rates for first-degree students. The data is then broken down by study mode (full-time or part-time), enabling the OfS to examine continuation or progression rates for full-time first-degree students and part-time first-degree students. The OfS will then apply one of the “initial baselines” – a series of percentage figures set out in a table in the DMG which indicate whether the indicator data is “of concern” or “of significant concern” (or “of no concern”). For example, a provider’s continuation rate would be of “significant concern” if the proportion of its full-time, first-degree students who continue their studies from year 1 to year 2 was 75% or less.

The second stage is the “demographic group threshold analysis”. As I say below, I have found it difficult to follow how this stage works. It seems that the data is broken down by demographic group (such as white, BME, male, female, young, mature), which the DMG refers to as “split indicators”, so that the OfS can work out the number of students within each group. The OfS then works out (inter alia) the continuation and progression rates for that group to establish whether the group has at

The third stage is “context”: an assessor then takes into account “the context in which the provider operates, such as the type of provision it offers and its size”.

In relation to the first stage (“initial baseline analysis”), Cavanagh J noted Ms Lapworth’s evidence that the initial baselines were “deliberately set at generously low figures” so that the thresholds would be proportionate even when used to assess the performance of providers which had large numbers of disadvantaged students. Different baselines were set for part-time students, in some cases 15% below those for full-time students on the same type of course, but no different baselines were set for students, like those at Bloomsbury, who were on four-year courses which included a foundation year. The OfS’ explanation was that this was because the OfS did not want to tolerate poorer outcomes for foundation year students, and that data issues concerning the reporting of the outcomes of foundation year students would limit its confidence in the reliability of any baselines which it constructed.

In relation to the second stage (“demographic group threshold analysis”), Cavanagh J noted Ms Lapworth’s evidence that the demographic analysis was designed to enable the OfS to check whether an outcome of “significant concern” existed across different demographic groups.

Paragraph 3 of the OfS’ Scheme of Delegation, dated 1 April 2018, stated that the matters reserved to the OfS Board included “changes to the regulatory framework”. The other relevant paragraphs provided as follows:

“4.

… it is not practical for the Board to make every decision necessary to fulfil the OfS’s role. It therefore delegates the authority to make certain decisions to the OfS chair, the chief executive, the director for fair access and participation, other directors and Board committees, taking into account the advice of the OfS’s senior executive team or others as appropriate.

...

7.

Functions, matters, powers, authorisations, delegations, duties and responsibilities within this Scheme shall be construed in a broad and inclusive fashion and shall include the doing of anything which is calculated to facilitate or is conducive or incidental to the discharge of anything specified.

It is not practical for the chief executive to make every dayto-day operational decision necessary for the smooth running of the OfS. Operational decision-making is therefore cascaded down through directors to senior managers and others as necessary.”

Paragraph 20 confers on the Director of Competition and Registration delegated authority to grant registration where the provide satisfies the initial conditions of registration and the risk category for each condition (on a traffic light system) is no higher than yellow.

By paragraph 24, authority for making registration decisions in circumstances where the OfS’ assessment team has identified a “red risk of breach” with respect to two initial conditions of registration (as was the case for Bloomsbury’s application) is delegated to the Provider Risk Committee.

The effect of the Scheme is thus that a decision to refuse registration must be taken by the Provider Risk Committee (“PRC”), rather than by the Chief Executive or Ms Lapworth.and Registration

As I have noted, a large number of grounds of claim were raised before the judge. Those which are relevant to the present appeal were: (1) the OfS breached the Scheme of Delegation because Ms Lapworth did not have authority to set the pre-determined thresholds used in the DMG; (2) the OfS failed to consult adequately on the DMG and/or was wrong not to publish the DMG; (3) the demographic group threshold test used in the DMG was irrational.

The judge held that Ms Lapworth did have authority under the Scheme of Delegation to adopt the DMG, on the basis that the evaluative framework set out in the DMG was part of an “operational decision-making function” which was delegated to her by paragraph 13 of the Scheme of Delegation. Because the initial baselines were (in the words of the OfS’ witnesses) a means of “operationalising” the evaluative scheme set out in the RF, Ms Lapworth was entitled to formulate the DMG without approval by the Board. The baselines were matters of specific and technical detail which were suitable for being worked out by a member of the executive team rather than by a committee such as the PRC. As the handling of Bloomsbury’s case shows, final registration decisions still rested with the PRC and not Ms Lapworth: the DMG was thus not determinative of registration applications. Whether the OfS failed to consult adequately on the DMG

The judge rejected Bloomsbury’s contention that, by taking registration decisions using the DMG, the OfS adopted a very different approach to assessing Condition B3 from that which was foreshadowed in the RF consultation exercise and that, as a result, the OfS should have consulted on the DMG. The judge held that, firstly, the DMG’s use of “initial baselines” and “split indicators” to analyse performance was consistent with the approach set out in the RF. The DMG does not lay down that those metrics should be inflexible absolute values and states repeatedly that an assessor should look at the wider context in which a provider operates. Secondly, the consultation document did not suggest a “provider-specific approach”, whereby an OfS assessor would create a bespoke evaluation process for each provider. Rather, the consultation made clear that the OfS would not tolerate poor outcomes for students of any providers and that one of the RF’s primary objectives was to ensure positive outcomes for “all” students.

The judge held, thirdly, that the DMG did not ignore the demographic make-up of a provider’s student body. The initial baselines were formulated generously and the OfS could also take into account the number of disadvantaged students at a provider at the third (“context”) stage of its analysis. Finally, although the initial baselines were not set out in the consultation document, this did not render the consultation process defective. The purpose of s 75(8) and s 5(5) of HERA was “not to require consultation on every granular detail of the assessment process that was carried out by the OfS”, but to consult “in broad terms about the Regulatory Framework and… about the Conditions that should be adopted”. It was therefore not necessary for the consultation document to say where the initial baselines for “significant concern” would be set. Consultees had sufficient information to make intelligent representations about the OfS’ approach to evaluation without such material.

The judge held that the OfS was not obliged to publish the DMG. The judge acknowledged that this ground was predicated on similar arguments to the ground on consultation. He considered that because the DMG did not depart from (and was not inconsistent with) the RF, the OfS was under no obligation to publish it. The RF said that the OfS would evaluate providers using data and contextual evidence; the DMG provided for this. Although the RF does not set out initial baselines, the DMG’s use of these numerical performance targets did not mean the OfS had to publish the DMG, since the targets were only the first stage of the process and were not absolute.

The judge also held that Bloomsbury was given adequate information to make intelligent representations in the provisional decision letter of 29 January 2019. The Annex to that letter set out information about the methodology used by the OfS to make its evaluation of Bloomsbury’s performance under Condition B3.

The judge held that the demographic group threshold analysis was neither irrational, arbitrary, nor contrary to the statutory purpose. He held that it operates as a sensecheck: “if a high proportion of a provider’s students fall into demographic groups which achieve outcomes of “significant concern”, this is a sign that there is a problem across the board at the provider”. The judge noted that, in any case, “failing the 75% demographic test does not automatically mean that the provider has failed to comply with Condition B3”.

Lewison LJ granted Bloomsbury permission to appeal on three grounds. Under Ground 1 (“delegation”), Bloomsbury submitted that Cavanagh J was wrong to find that Ms Lapworth had lawful authority under the OfS’ Scheme of Delegation to adopt the DMG. Under Ground 2 (“publication and consultation”), Bloomsbury submitted that the judge was wrong to hold that the OfS acted lawfully in neither publishing nor consulting on the DMG. Under Ground 3 (“irrationality”), Bloomsbury submitted that the judge erred in holding that the “demographic group threshold” analysis set out in the DMG was not irrational.

For the Appellant Jessica Simor QC relied on the case of Queen Mary University London v Higher Education Funding Council for England [2008] EWHC 1472 (Admin) in which Burnett J held that if a public body has a scheme for delegating its statutory decision-making, the decision must be taken by the person to whom the decision-making function is delegated. She submitted that if the OfS had wanted to delegate the policy decisions in the DMG to Ms Lapworth, it could and should have done so expressly.

Ms Simor submitted that the OfS could not rely on paragraph 13 of the Scheme to authorise Ms Lapworth’s decisions since that paragraph only authorised “day-to-day operational decisions for the smooth running of the OfS”. It was not designed to authorise the delegation of the formulation of the “ground rules” for evaluating higher education providers’ compliance with Condition B3. Ms Simor submitted that the phrase “operational decisions” could not cover decisions which determine the circumstances in which an applicant can meet a condition of registration and that such an interpretation would run contrary to the rest of the Scheme, which carefully prescribes the circumstances in which important decisions can be delegated below Board level. Ms Simor submitted that the OfS’ unlawful delegation was potentially material because, had the Board itself considered the approach set out in the DMG, the DMG might have been formulated differently.

For the Respondent Monica Carss-Frisk QC invited us to afford the OfS a degree of latitude in interpreting the Scheme of Delegation given that it was designed to serve as an internal guidance document. She cited R (Springhall) v Richmond upon Thames

Ms Carss-Frisk submitted that, even if we decided to interpret the policy objectively, we should do so flexibly and not subject the Scheme to the level of analysis which might be applied to a statute or contract. She submitted that paragraph 13 did authorise Ms Lapworth to formulate the evaluative process in the DMG. She relied on a witness statement of Mr Coleman (deputy chair of the Board and chair of the PRC)

Ms Carss-Frisk further submitted that the DMG was a guidance document to assist OfS officers and did not have to be followed rigidly. The DMG did no more than “operationalise” the broad policy decisions made in the RF and did not materially add to or vary it. While the DMG did set out baselines which were not contained in the RF, they were not determinative of a provider’s application and the ultimate registration decision was reserved to the PRC. Ms Carss-Frisk submitted that in any event, the Board would not have reached a different conclusion if it had not delegated the drafting of the DMG to Ms Lapworth or had been asked for its approval of the document.

Ms Simor relied on the case of R (Lumba) v Secretary of State for the Home Department [2012] 1 AC 245 for the propositions that any guidance which sets out how decision-makers should exercise their discretionary powers must be published; that decision-makers are obliged to follow published policies absent good reasons to the contrary; and that individuals have the right to know what a policy is so that they can make representations in relation to it. She drew our attention to paragraph 141 of R (Justice for Health Ltd) v Secretary of State for Health [2016] EWHC 2338 (Admin), where Green J set out the reasons why such guidance documents must not be secret: so that individuals can regulate and foresee the consequences of their conduct and make representations about how any discretionary power should be exercised, and so as to ensure that decision-making is accountable and not carried out pursuant to unlawful guidance.

Ms Simor submitted that the OfS was obliged by the common law, as well as by provisions in HERA and the Regulators’ Code, to publish any part of the DMG which set out how the OfS’ discretion would be exercised. She argued that Cavanagh J was wrong to hold that the duty arises only when an unpublished policy is inconsistent with a published policy (as was the case in Lumba). Narrowing the duty in the way in which Cavanagh J did would be unworkable because it would enable a decisionmaker such as the OfS to determine for itself, without external scrutiny, whether the unpublished policy is inconsistent with the published one.

Ms Simor submitted that a statutory duty to publish arises from ss 5 and 75.

Ms Simor submitted that because of the OfS’ failure to publish the DMG, Bloomsbury did not know that numerical thresholds had been set, the level at which those thresholds were set, that a “demographic group threshold” would be applied, or

Ms Carss-Frisk submitted that Lumba imposes no duty to publish the DMG. In her submission, Lumba is authority for the proposition that a policy should be published if it will inform discretionary decisions in respect of which the potential object of those decisions has a right to make representations. The right to know the content of the policy is therefore parasitic on the right to make representations; there is no general obligation on a public body to publish every policy which it applies. Where an internal guidance document does not depart from a published policy, and the published policy gives sufficient information to enable representations to be made, there is no duty to publish the internal guidance. Since the DMG did not set out a different approach to evaluating registration applications from the RF (which was published), the DMG did not have to be published.

Ms Carss-Frisk further submitted that the OfS was under no duty to publish the DMG since Bloomsbury was given sufficient information to enable it to submit meaningful representations, consistently with the OfS’ duty under s 4 of HERA or at common law. The RF and RA 2 contained information which made providers aware that continuation and progression rates would be material to their application. Providers did not need to know the precise baselines in order to make meaningful representations, given that the baselines were not determinative of providers’ applications and, in any case, related to their historical performance. The OfS’ provisional decision of 29 January 2019 gave Bloomsbury further information about the OfS’ evaluative approach. Ms Carss-Frisk rejected any suggestion that this approach contravened the requirements of transparent regulation and submitted that there were good reasons for not publishing the DMG, such as not wanting providers to “game the system” by discontinuing lower performing courses for tactical reasons.

Ms Simor submitted that the OfS was obliged to consult on the policy decisions which were taken in relation to how Condition B3 would be assessed. The OfS’ obligation to consult on the way in which compliance with Condition B3 would be assessed arose from HERA, the Regulatory Code and the common law. Sections 5 and 75 of HERA required the OfS to consult in respect of the RF, the guidance on the ongoing conditions of registration, and the initial and ongoing conditions of registration. Paragraphs 2.1, 3.2 and 3.3 of the Regulators’ Code required the OfS to engage with the organisations it regulates to develop regulatory policy and their approach to risk assessment. At common law, cases the courts have held that regulators such as the OfS may be obliged to communicate the methodology which they intend to use when evaluating the organisations that they regulate. Ms Simor submitted that the requirements of fairness are particularly exacting when the proposal in question involves the deprivation of an existing benefit: Bloomsbury was registered under the previous system, and had been so for many years.

Ms Simor submitted that the OfS was obliged to provide consultees with sufficient information about how the OfS would assess compliance with Condition B3 to enable them to make informed representations which could influence that approach. The OfS failed to discharge this obligation by not consulting on four particular policy choices: the use of the two-stage numerical test to determine compliance; the level at which the numerical thresholds would be set; the differences in thresholds which would be applied to part-time (rather than full-time) and “other undergraduate” (rather than first-degree) courses, in particular whether there should be a lower threshold for courses with a foundation year which had lower average sectoral results; and the factors which would be considered as “context”, for example whether demographic factors should be taken into account after the application of any threshold.

Ms Carss-Frisk submitted that the extent to which a public body must consult is determined by the statutory context, that in general a consultation needs to do no more than tell consultees enough to enable them to respond intelligently, and that a high threshold exists for establishing that a consultation process was procedurally unfair. She submitted that the OfS discharged its duty by carrying out a single public consultation from October to December 2017, which covered both the proposed RF (as required by s 75 of HERA) and the initial and ongoing registration conditions (as required by s 5). That consultation proposed using “baseline requirements” and evidence of providers’ actual performance including continuation and progression rates.

Ms Carss-Frisk submitted that the OfS was not obliged to consult on the contents of the DMG, or the precise baselines which it included, for several reasons. First, there was no such obligation under s 5(5), which required broad consultation on the design of the registration conditions, not the detail of how the OfS assesses compliance with the initial conditions. Secondly, there was no such obligation under s 75(8): that subsection required broad consultation on the RF and on the OfS’ general approach to registration conditions, not on what she described as the “detailed internal steps” which OfS assessors would take to reach an initial review on whether a condition was satisfied. Thirdly, the DMG did not need to be consulted upon because it did not depart from the detailed policy choices which were set out in the RF and included in the consultation. Fourthly, for the same reasons, the Regulators’ Code did not impose on the OfS any obligation to consult on the DMG.

Ms Simor submitted that the “demographic group threshold analysis” stage of evaluating registration applications is irrational for several reasons. First, it yields perverse results. As an example, if Provider A and Provider B both pass the numerical threshold test (and have similar levels of attainment), but Provider B has a concentration of students who fall within a demographic group that disproportionately lowers its attainment level, Provider A may pass the demographic analysis threshold when Provider B will not. Thus, for the purposes of the demographic group threshold, what matters is how low attainment is spread through the various demographic characteristics, not what the failure amounts to. In Ms Simor’s submission, the demographic group analysis obscures the role of other factors in student outcomes, such as their impecuniosity, level of previous education or age, and incentivises providers to take fewer disadvantaged students, because the more disadvantaged students a provider has, the more likely it is to fail.

Ms Carss-Frisk submitted that the hurdle for challenging the rationality of a policy decision of a specialist regulator such as the OfS is extremely high. She submitted that there were good reasons for using the demographic threshold analysis: as Cavanagh J found, it enabled OfS assessors to bring together all the rates into a single figure, providing a “sense check” and an overview of the provider’s performance. Ms CarssFrisk rejected the argument that the analysis yielded perverse results. She submitted that even if the demographic group threshold was irrational, it was immaterial to Bloomsbury’s registration application: the OfS considered the wider context in which the provider operated, including the broad demographic issues which Bloomsbury identified.

It is common ground that the DMG was written by Ms Lapworth (no doubt with the assistance of others) and that she was responsible for its contents. Indeed, she emphasises this in her own evidence: at paragraph 62 of her first witness statement she says that the PRC “agreed with the proposed approach to assessing B3, including that the baselines appeared appropriate” but adds that “in reaching this view the PRC was clear that it was not the decision-maker for determining the baselines”. The DMG was “considered and noted” by the PRC but was never approved by that committee, still less by the Board. The question is therefore whether its contents, so far as relevant for present purposes, can be described as an “operational decision” within paragraph 13 of the Scheme of Delegation.

It is well established that in general a public authority can choose to whom to delegate its delegable functions. The Court should intervene only if the choice is irrational or otherwise beyond the authority’s powers: see e.g. R (Chief Constable of the West Midlands Police) v Birmingham Justices [2002] EWHC 1087 (Admin), per Sedley LJ at [16]. Thus, as Ms Simor correctly pointed out, the issue under Ground 1 is not whether the Board could have approved the DMG, nor whether they could have approved a Scheme of Delegation which expressly stated that all baselines and other criteria used for assessing compliance with Condition B3 were to be determined by the Director of Competition and Registration alone. The question is whether they did so by the reference in paragraph 13 to “operational” decisions being delegated.

I accept that when the Court is interpreting a document such as the Scheme of Delegation, it should not subject the wording to the kind of fine analysis which might be applied to a statute or a contract: Tesco Stores Ltd v Dundee City Council [2012] PTSR 983 per Lord Reed. But it must still be interpreted objectively. We were referred to the well-known observations on the interpretation of policy statements of Lord Steyn in Re McFarland [2004] 1 WLR 1289 at [24] (in a partly dissenting speech which is nevertheless accepted on this point to be good law) that, while they need not be construed as though they were legislation, and it “seems sensible that a broad and wholly untechnical approach should prevail”, nevertheless:-

“…what is involved is still an interpretative process conducted by a court, which must necessarily be approached objectively and without speculation about what a particular minister may have had in mind.”

The aspect of the DMG which is critical in the present case is the way in which Ms Lapworth chose to make allowance for demographic factors, in particular the presence in the student population at some providers such as Bloomsbury of significant numbers of students disadvantaged by socio-economic factors or by having left school without taking A-levels. She did so by:

Ms Carss-Frisk submits that there is no “magic distinction” or bright line between a policy decision and an operational decision. I agree that there may be cases where it is difficult to say on which side of the line the decision falls. But I consider that the baseline decisions I have just summarised fall well on the policy side of the line and could not by any stretch of the imagination be described as merely operational. Such an interpretation of paragraph 13 of the Scheme of Delegation may not be from what Lord Reed in Tesco (alluding to the famous speech of Lord Atkin in Liversidge v Anderson [1942] AC 206) called the “world of Humpty Dumpty”, but it is not far off it. Ms Simor pointed out that if the OfS’ argument on this point is correct it would have been within the Scheme of Delegation for Ms Lapworth to increase all the baseline figures without seeking anyone’s approval. Indeed, since paragraph 13 allows operational decisions to be cascaded down from level to level, the setting of baselines could on the same logic have been delegated by Ms Lapworth to the office junior. I cannot accept that this could be described as an operational decision.

The OfS placed reliance on the evidence of Mr Coleman (deputy chair of the Board and chair of the PRC) that the Board normally only meets six times a year; that its focus has tended to be on high-level decisions and strategic oversight; and that when the Board agreed the RF it was well understood that the more detailed decisions which would be needed to “operationalise” the framework would be devolved to others. No doubt this is true. But I do not think it of any assistance in determining whether the decision not to make special allowance for courses beginning with a foundation year was merely operational. It is not significant whether Mr Coleman or any other individual thinks now, or thought then, that the Scheme of Delegation permitted the formulation of the DMG to be entrusted to Ms Lapworth alone. On an objective interpretation of paragraph 13, it did not permit the baseline decisions to be delegated to Ms Lapworth. The claim therefore succeeds on Ground 1.

Although the failure to comply with the Scheme of Delegation is in itself fatal to the validity of the decision under scrutiny, I should record my view that it should not be assumed - and was not asserted in evidence - that the Board would have simply rubber-stamped the DMG drafted by Ms Lapworth if it had been presented to them for approval, especially if the omission of special provision for foundation year courses had been highlighted. As the Appellant’s skeleton argument suggests, the Board might have rejected the use of numerical thresholds to create a presumption of noncompliance; or it might have decided that the logic of applying lower numerical thresholds for part-time courses, which Ms Lapworth decided was appropriate, applied a fortiori to foundation year courses.

Even if the DMG had been approved by the Board or expressly delegated by them in all respects to Ms Lapworth, Bloomsbury would still have had available to them their challenge based on failure to publish and consult. That brings me to Ground 2. Both sides treated publication and consultation as separate issues, although to a large extent in the present appeal the two are intertwined.

It is an ironic feature of this case that the baselines for the various indicators under Condition B3, which Ms Simor repeatedly called the “secret guidance”, are no longer secret. They were published in detail on the OfS website on 30 October 2019 as part of an OfS report on its registration process and outcomes. This was, of course, some months after Bloomsbury had been refused registration. The OfS’ case remains that it was under no obligation to make the baselines public.

As both sides accepted, the relevant law as to the publication of the policies of a public body is authoritatively stated in R (Lumba) v Secretary of State for the Home Department [2012] 1 AC 245, where Lord Dyson JSC said:

“35.

The individual has a basic public law right to have his or her case considered under whatever policy the executive sees fit to adopt provided that the adopted policy is a lawful exercise of the discretion conferred by the statute: see In re Findlay [1985] AC 318, 338E. There is a correlative right to know what that currently existing policy is, so that the individual can make relevant representations in relation to it.

…

38.

The precise extent of how much detail of a policy is required to be disclosed was the subject of some debate before us. It is not practicable to attempt an exhaustive definition. It is common ground that there is no obligation to publish drafts when a policy is evolving and that there might be compelling reasons not to publish some policies, for example, where national security issues are in play. Nor is it necessary to publish details which are irrelevant to the substance of decisions made pursuant to the policy. What must, however, be published is that which a person who is affected by the operation of the policy needs to know in order to make informed and meaningful representations to the decision-maker before a decision is made.” [emphasis added]

We were referred to numerous cases at various levels of court about the duty to consult. There is no need for me to conduct what Laws LJ used to call an “anxious parade of learning” on this topic. In the leading case of R(Moseley) v Haringey LBC [2014] 1 WLR 3947 Lord Reed JSC said at [35] that, while there is no general common law duty to consult persons who may be affected by a measure before it is adopted, “a duty of consultation will however exist in circumstances where there is a legitimate expectation of such consultation, usually arising from an interest which is held sufficient to found such an expectation, or from some promise or practice of consultation”. As to what the common law duty involves, Lord Wilson JSC at [25] gave his approval to what public lawyers have long called the “Sedley criteria”, that is to say the four requirements suggested by Stephen Sedley QC (as he then was) and accepted by Hodgson J in R v Brent LBC ex p Gunning (1985) 84 LGR 168, namely that: (1) consultation must be at a time when proposals are still at a formative stage; (2) the proposer must give sufficient reasons for any proposal to permit of intelligent consideration and response; (3) adequate time must be given for consideration and response; and (4) the product of consultation must be conscientiously taken into account in finalising any statutory proposals. “It is hard”, said Lord Wilson, “to see how any of his four suggested requirements could be rejected or indeed improved.”

Ms Carss-Frisk cited, and placed great emphasis on, paragraph 90(v) of the judgment of Hickinbottom LJ in R (Help Refugees Ltd) v Secretary of State for the Home Department [2018] 4 WLR 168:

“The courts will not lightly find that a consultation process is unfair. Unless there is a statutory specification as to the matters that are to be consulted upon, it is for the public body charged with performing the consultation to determine how it is to be carried out, including the manner and extent of the consultation, subject only to review by the court on conventional judicial review grounds. Therefore, for a consultation to be found to be unlawful, “clear unfairness must be shown” (Royal Brompton and Harefield NHS Foundation Trust [2012] EWCA Civ 472 at [13]; or, as Sullivan LJ said in R (Baird) v Environment Agency [2011] EWHC 939 (Admin) at [51], a conclusion by the Court that a consultation process has been so unfair as to be unlawful is likely to be based on a factual finding that something has gone “clearly and radically wrong”.”

It is important to read this passage, especially the citation from Baird, carefully and in context. The Help Refugees case itself concerned s 67 of the Immigration Act 2016 (better known as the “Dubs amendment”), which imposed on the Home Secretary a duty to make arrangements to relocate some unaccompanied refugee children who sought to enter the UK and provided in subsection (2) that “the number of children to be resettled shall be determined by the Government in consultation with local authorities”. The judicial review claim involved an all-out attack on the adequacy of that statutory consultation with local authorities. Similarly in Baird Sullivan LJ (sitting as a single judge of the Administrative Court) had been considering a claim that part of a non-statutory consultation about a local flood and erosion risk management strategy had been so “fatally flawed” that it should be started again. He said this:

“In R(on the application of Greenpeace Limited) v the Secretary of State for Trade and Industry [2007] ELR 29, it was submitted on behalf of the defendant that the court should interfere with a consultation process "only if something has gone clearly and radically wrong." The claimant had submitted that there was no support for this proposition in the authorities. In paragraphs 62 and 63 of my judgment, I said: "This difference between the parties is one of semantics rather than substance. A consultation exercise which is flawed in one or even in a number of respects is not necessarily so procedurally unfair as to be unlawful. With the benefit of hindsight it will always invariably be possible to suggest ways in which a consultation exercise might have been improved upon. That is most emphatically not the test. It must also be recognised that a decision maker will usually have a broad discretion as to how a consultation exercise should be carried out. This applies with particular force to a consultation with the whole of the adult population of the United Kingdom. The defendant had a very broad discretion as to how best to carry out such a far-reaching consultation exercise. In reality, a conclusion that a consultation exercise was unlawful on the ground of unfairness would be based upon a finding by the court not merely that something was wrong but that something went "clearly and radically" wrong."

Sullivan LJ went on to refer to observations about his judgment in Greenpeace made by Ouseley J in Devon County Council and Norfolk County Council v Secretary of

Sullivan LJ continued in Baird by saying: “I respectfully agree with that observation. The test is whether the process was so unfair as to be unlawful. In Greenpeace, I was not seeking to put forward a different test, but merely indicating that in reality a conclusion that a consultation process has been so unfair as to be unlawful is likely to be based on a factual finding that something has gone clearly and radically wrong.”

These remarks, endorsed by this court in Help Refugees, appear to me to say two things about the phrase “a factual finding that something has gone clearly and radically wrong”. Firstly, it is not an additional hurdle to be jumped: the test remains whether the process was so unfair as to be unlawful. Secondly, it is directed principally at cases where a consultation exercise is under attack on such a broad front that it is alleged that the whole process is unlawful. It does not assist much, if at all, in cases where the allegation is that a claimant or class of claimants likely to be particularly affected by the operation of the policy was not given sufficient information which they needed to know in order to make informed and meaningful representations to the decision-maker before a decision was made.

Since the DMG was not published (either in draft or in the form in which Ms

I agree with Cavanagh J when he said at paragraph 143 of his judgment that the OfS was not required by ss 75(8) and 5(5) of HERA or the Regulatory Code to consult on

“every granular detail of the assessment process”; and I would add that it was not required to do so at common law either. But I consider that the foundation year course policy decision cannot properly be described as a matter of granular detail; and that the failure even to publish it, still less to consult on it when it was at the formative stage, constituted clear unfairness in the treatment of Bloomsbury.

It is not an answer to this complaint to say that Bloomsbury were sent the “minded to refuse” letter of 29 January 2019 giving the OfS’ provisional decision, and had the opportunity at that stage to make representations before the decision was made final. They were by then in a far weaker position than they would have been if allowed to make representations at the policy formulation stage. Moreover, the provisional decision letter states that the OfS had considered contextual issues and taken them into account when reaching its decision.

The complaint of procedural unfairness cannot be described as theoretical. The foundation year course policy issue is one on which different views can reasonably be held. The Claimant’s evidence includes a witness statement of Professor Sir David Melville, a very distinguished figure in the world of higher education, who is highly critical of the OfS’ decision not to make special allowance for students on foundation year courses. He says:

“In evaluating the performance of a provider’s first degrees, the Foundation Year is a relevant factor that cannot be ignored.

The OFS’ own report has recognised the lower continuation rates for students on a foundation year. At the same time the OFS has recognised the foundation year’s positive contribution to the promotion of equality of opportunity in connection with access to and participation in higher education. …”

It is not for a court to say whether special provision ought to have been made; and I have not overlooked the fact that Cavanagh J dismissed a challenge based on the public sector equality duty. But this is an issue on which Bloomsbury and other providers in a similar position ought to have been consulted at the policy formulation stage; and also an issue whose resolution should have been approved by the Board unless (which, as discussed above under Ground 1, was not the case) they had clearly entrusted it to someone else.

Ground 2 accordingly succeeds.

I found the Respondent’s explanation of the second stage of the demographic threshold analysis very hard to follow; and I believe that I was not alone in that respect. But in the light of the above conclusions it is unnecessary to say anything more about Ground 3.

For these reasons, as well as those in the judgment of Males LJ with which I agree, I concurred in the decision of this court, communicated to the parties and published on 31 July 2020, to allow the appeal on both Grounds 1 and 2 and quash the OfS’ decision of 23 May 2019 refusing Bloomsbury’s application for registration. I would invite further submissions on consequential orders by 3 September 2020.

Ensuring access to high quality education for students from disadvantaged backgrounds is one of the top priorities for providers of higher education. It follows that, for a regulator such as the OfS, the way in which it takes account of demographic factors in assessing the performance of such providers, and in particular of providers with a high proportion of students from disadvantaged groups, is an important policy choice.

It is apparent from the witness statement of Ms Susan Lapworth, the OfS’s Director of Competition and Registration who was responsible for establishing the OfS’s approach to assessing compliance with Condition B3, that she decided on the use of “baselines” as the means by which account would be taken of demographic factors. Having made that decision, she then set what she regarded as “somewhat generous baselines” which she thought “would take adequate account of demographic differences”. There were, therefore, two relevant decisions, the first being to use baselines as the means of taking these factors into account, and the second being her approach to the levels at which they would be set.

It is true that the Decision-Making Guidance provides for a third stage, after ascertaining (inter alia) continuation rates and progression rates for each of a provider’s courses and undertaking the “demographic group threshold analysis”. That third stage requires the assessor to take into account “the context in which the provider operates, such as the type of provision it offers and its size”. It is apparent, however, that the intention of the Guidance is that such factors should primarily be taken into account at the first two stages, with only limited scope thereafter to consider arguments based on “context”. This was in fact how the Guidance was applied in practice, as appears from the minutes of the Provider Risk Committee meeting on 29^th April 2019 at which the decision to refuse Bloomsbury’s registration was made. At that meeting the Committee determined that “limited weight” should be given to arguments based on the characteristics of a provider’s student population or to arguments that some students withdrew from their courses for reasons outside the provider’s control because allowance was already made for such factors in the baselines which had been set.

Bean LJ has emphasised Ms Lapworth’s decision to make no distinction in the continuation rates required to be achieved between four year degree courses which include a foundation year (typically undertaken by mature students and those without A-levels) and three year courses which do not. I agree that this is a particularly striking example of a policy decision, not least in view of the economic and other circumstances likely to affect students taking four year courses which have nothing to do with the quality of education provided and over which a provider has no control. In my judgment, however, the decision to use baselines more generally as the means by which demographic factors would be taken into account in assessing a provider’s performance, leaving only limited scope for consideration of the context of a particular provider, was itself an important policy decision.

Such a decision cannot be regarded as an “operational” matter within the meaning of the Scheme of Delegation. It was a decision which, pursuant to that Scheme, was required to be made by the Board.

Once it is appreciated that the Guidance (or at any rate this aspect of it) embodied an important policy decision as to the way in which account would be taken of demographic factors, it becomes obvious in my judgment that this was a matter on which the OfS was required to consult, and which, once the decision was made, it was required to publish.

For these reasons, in addition to those given by Bean LJ with which I agree, I would allow this appeal.

I agree with both judgments.