Friday, 30th July 1993
LORD JUSTICE MANN: I am sorry you have been kept waiting a few moments. I will ask Mr Justice Sedley to give the judgment of the Court.
MR JUSTICE SEDLEY: This is an application for judicial review of a decision of the Universities Funding Council as it then was, published on the 17th December 1992, to place the Institute of Dental Surgery, the applicant, at level 2 in its assessment of the quality of institutional research. The decision has direct implications for the level of government funding and indirect implications for the level of outside funding which the Institute can expect in the year 1993/4. Brooke J gave leave on the 3rd March 1993 to apply, with expedition, for a declaration that the respondent Council had erred in law in its assessment of the applicant Institute; certiorari to quash the decision, and mandamus to require the successor body, the Higher Education Funding Council, to reconsider the matter.
The applicant Institute is a college of the University of London. Uniquely in the United Kingdom, it is entirely dedicated to postgraduate teaching and research in dentistry. The Universities Funding Council (UFC) was established by section 131 of the Education Reform Act 1988. It was established as a body corporate consisting of fifteen members appointed by the Secretary of State, six of them being engaged and experienced in higher education. By sub section (4) the Council was made responsible for administering central funds made available to it for the support of eligible activities, these being defined by sub section (5) as including the provision of education and the undertaking of research by universities. Sub section (6) gave the Council power to make grants for the prescribed purposes to the governing bodies of universities. Sub section (7) provides:
"In exercising their functions in relation to the provision of financial support for activities eligible for funding under this section, the Council shall have regard to the desirability of not discouraging any university in respect of which grants are made under sub section (6) above from maintaining or developing its funding from other sources."
Sub section (9) places an obligation on the governing body of any university to give the Council such information as it requires for the foregoing purposes.
By section 235(1) a university is defined so as to include a college, so that the Institute of Dental Surgery acquires independent status for grant purposes.
These functions remain in place, but the body initially exercising them, the Universities Funding Council, was dissolved on the 1st April 1993 by section 63(1) of the Further and Higher Education Act 1992, and its extant property, rights and liabilities were transferred to one of the two bodies corporate established by section 62(1), the Higher Education Funding Council (hereafter HEFC) for England (the other was a similarly named council for Wales). Sections 65 and 66 of the Act of 1992 confer upon the new bodies functions broadly similar to those of their predecessor. Section 70(1) requires the Council to establish a Quality Assessment Committee to advise it on the performance of an express duty to provide for assessing the quality of education in institutions falling within its remit; but no such obligation is spelt out on assessing the quality of research. The HEFC thus has a free hand in devising proper means of arriving at its decisions on the funding of research in higher education. Although, unusually, it appears to have no express power of delegation, Mr David Pannick Q.C. for the applicant has realistically accepted that without devolving its functions to a considerable number of suitably qualified people the Council cannot function, and he has taken no point on the system which was adopted and to which we will come.
In consequence of this change in the identity of the respondent, the decisions impugned and the object of relief sought have been amended without objection. Also without objection the form of the order of mandamus sought has been expanded by amendment so as to require the respondent body not only to reconsider the matter but to give reasons for the assessment decision.
The Universities Funding Council was in its time the successor of the University Grants Committee (UGC) which in 1984 had adopted a policy of selective funding of research in order to make the best use of available funds. To implement the policy, the UGC in 1986 undertook a research assessment exercise with the object of rating institutions according to the quality of their research and of awarding grant accordingly. A similar exercise was carried out in 1989 and again in 1992.
On the 6th March 1992 the UFC issued to all interested institutions a circular letter 5/92 announcing the form to be taken by the 1992 Research Assessment Exercise. It included the following paragraphs:
"1. Circular Letter 22/91 set out preliminary proposals for the next Research Assessment Exercise and invited comments. This circular describes how the Exercise will be conducted and invites higher education institutions (HEIs) to make submissions.
2. Nearly 300 replies were received from HEIs, professional and learned bodies, subject associations, individuals and other interested parties. They have been taken into account in defining the arrangements for the 1992 Exercise.
Purpose of the Exercise
5. The purpose of the Exercise is to produce research ratings, which will be used by the new Higher Education Funding Councils (HEFCs) for England, Scotland and Wales, and in respect of institutions in Northern Ireland, by the Department of Education for Northern Ireland, in the determination of grant for research with effect from 1993/94. The White Paper in May 1991, and the subsequent letters from the Secretary of State for Education and Science to the Chairmen of the two Funding Councils, expressed the Government's wish to see increased selectivity in the allocation of research resources. The assessments made as a result of this Exercise will assist in this.
Conduct of and timetable for the Exercise
6. The 1992 Exercise will follow broadly the same approach as in 1989, with information provided by HEIs as the basis for peer review assessment of research quality by a number of specialist panels. Institutions are invited to list up to two publications and up to two other forms of public output for each member of staff whose research is to be taken into account, and these will contribute to the basis for the judgements of research quality. As in 1989, panels will also have available to them a range of data to take into account in making their judgements. However, compared to 1989, there are a number of significant changes to the Exercise, which are described below.
Coverage: research to be assessed
8. Account will be taken of the full range of research (applied, strategic and basic) as defined at Annex A. Although research covers a continuous spectrum between applied and basic, the two extremes are nevertheless distinct. In the case of engineering and science only (units of assessment 12 16, 20 23, and 27 34 in Annex B), panels will be asked to assign separate ratings for applied and for basic/strategic research. In these units of assessment institutions are therefore asked explicitly to categorise research output as either applied or basic/strategic although final judgements will be made by the relevant assessment panels. In this context we are concerned that full recognition should be given to work of direct relevance to the needs of commerce, industry, and the public and voluntary sectors.
Units of Assessment
12. The subject areas to be used as the units of assessment are listed at Annex B.............
Assessment panels
14. Each unit of assessment will be assessed by a panel with appropriate specialist membership including non academic members where appropriate. Panels, which may assess work within a group of related units of assessment, will operate on a 'core plus adjunct' model where the core panel (typically 5 8 members) is joined by a number (typically 2 3) of specialist advisers to cover specific areas as required. It will be the core panel which has responsibility for grading submissions. The names of the members of the panels and the specialist advisers will be published.
Assessment ratings
15. The assessments made must reflect as far as possible the specific characteristics of each subject area, but at the same time the ratings applied must mean the same across all subjects. As for the 1989 Exercise the ratings will therefore be made on a common five point scale (see paragraph 11) with common definitions of the points, as given in Annex C. However, as it is expected that no research funding will be allocated in respect of any department which receives the lowest rating of the assessment scale, for funding purposes the ratings are expected to run from 0 4 instead of 1 5. More specific guidance will be given to panels than in 1989, a draft of which is given at Annex D: in addition individual panels may produce their own more specific subject related guidance.
Appeals
20. There will be no provision for appeals against individual ratings."
Annex B included clinical dentistry as a discrete unit of assessment. Annex C needs to be reproduced in full:
"Annex C
The rating scale interpretation of scale points
The following 5 point scale accompanying descriptions and notes will be used for the Exercise. The descriptions relate only to that research in a particular submission to a unit of assessment which was assessed.
.png)
Notes:
1. The concept of a 'sub area' of research activity is applicable to the work of individual researchers as well as to that of groups.
2. 'Attainable' levels of excellence refers to an absolute standard of quality in each unit of assessment, and should be independent of the conditions for research within individual departments.
3. The international criterion adopted should equate to a level of excellence that it is reasonable to expect for the unit of assessment, even though there may be no current examples of such a level whether in the UK or elsewhere. In the absence of current examples, standards in cognate research areas where international comparisons do exist, will need to be adopted. The same approach should be adopted when assessing studies with a regional basis against 'national' and 'international' standards.
4. For the Research Assessment Exercise 'national' refers to the United Kingdom of Great Britain and Northern Ireland."
Clinical dentistry was to be assessed by a panel of eight specialists under the chairmanship of Professor Roy Storer of the University of Newcastle upon Tyne. On the 15th July 1992 the panel met to establish its method of working. On the 15th and 16th October 1992 the panel met and considered submissions from sixteen institutions. Each member gave initial ratings by secret ballot to each institution except that of which he or she was a member. It is now known that the average rating received by the applicant on this exercise was 2.6, which if it had stood would have been rounded up to 3.
On the 23rd October 1992 the panel met to make its final assessments. Each member was asked to confirm or modify his or her original ratings. It was agreed that six institutions required extensive discussion, namely Bristol University, Glasgow University, the United Medical and Dental Schools, Kings College London, the London Hospital Medical College and the applicant Institute. The first the applicant knew of the outcome was the formal publication on the 17th December 1992 of all the ratings, by means of Circular 26/92. The applicant had been rated at 2. The circular contained the following further information:
"2. The ratings will be used by the Higher Education Funding Councils for England, Scotland and Wales, and in respect of institutions in Northern Ireland, by the Department of Education for Northern Ireland, in the determination of grant for research with effect from 1993/94. These bodies will issue separate documents on their methods for funding of research, and on financial allocations.
3. The Council wishes to emphasise that the assessments relate only to research quality. Teaching quality and other features of higher education institutions (HEIs) activity have not been assessed.
Conduct of the Exercise
6. Account was taken of the full range of research, as defined at Annex A. Submissions were made in 72 units of assessment; these are listed with a brief description in Annex B. All submissions were assessed against a common scale as reproduced at Annex C.
Timetable and Scale of the Exercise
8. Circular 5/92 was issued in early March 1992 and by the closing date of 30th June some 2,700 submissions had been received from 170 HEIs across the 72 units of assessment. The work of over 43,000 full time equivalent academics was included in the submissions. The submissions were assessed by 63 panels and subpanels with some 450 members and 50 assessors.
Comparability of the 1989 and 1992 Ratings
17. Although the same rating scale has been used in both the 1989 and 1992 Exercises with the same definitions, the 1992 Exercise has been carried out on a different basis and the results should not be directly compared with the earlier ratings.
In the 1989 Exercise information was requested on all academic staff within an institution, whether or not they were active in research. For the 1992 Exercise HEIs were asked to put forward for assessment only those academic staff that they judged to be active in research. Furthermore the units of assessment used in the two Exercises are not always the same."
The table of ratings in clinical dentistry, interpreted in the light of the coding explained in the body of the circular, showed that the applicant, in addition to being rated 2, had been identified as having one or more specific departments or research groups with particular excellence in research above the overall rating for the submission; the proportion of the academic staff in the department or unit submitted for assessment had been between 80 and 94%; the number of academic staff in post on the 30th June 1992 and assessed in the submission had been 28.7; and the research in developmental biology and periodontology had been assessed in part by the appropriate panel and given an overall rating, although it contained a significant proportion of research covered by another unit of assessment.
This rating caused understandable dismay in the applicant Institute. It had had a rating of 3 on the previous assessment exercise, and this change in its grading was likely to be directly reflected in a loss of some £270,000 in grant. Moreover, and it is to this that the present proceedings relate, the grade had been assigned without any reason beyond what could be inferred from the published description of the process and the interpretation of the published schedule of ratings.
On the 23rd December the Dean of the Institute, Professor Gerald Winter, wrote to Professor Graeme Davies, the chief executive of, as it then was, the UFC seeking a review of the decision or reasons for it. Neither this nor a subsequent letter brought a satisfactory response. On the 9th February 1993 Professor Davies wrote to Professor Winter:
"As undertaken in our initial telephone conversation, I have arranged to have your documentation scrutinised to establish whether or not there was any factual information which had not been available to the Assessment Panel for Clinical Dentistry.
I have confirmed that your original submission including factual information, conformed with that provided by other institutions and that all the submitted material plus the standard analyses were considered by the Assessment Panel for Clinical Dentistry.
Your letter of 21st January 1993 refers to the possibility that a higher standard was applied to the Institution, as a postgraduate institution, than to undergraduate institutions. I can confirm that this was not the case and that the guidance for assessment panels drew no distinction
between postgraduate and undergraduate institutions. On the contrary, Circular 5/92 specifically stated that the "attainable" levels of excellence referred to an absolute standard of quality in each unit of assessment, and should be independent of the conditions for research within individual departments.
If a decision was based on erroneous information then we would reconsider the submission. This does not appear to be so in the case of the Institute of Dental Surgery and, in the circumstances, I regret that there is no basis on which to take further action."
In response, and promptly, application was made for leave to apply for judicial review. In response to the affidavit evidence of Professor Winter affidavits have been sworn by Professor Davies and Professor Storer. From the evidence the following facts emerge:
(i) The reduction from a rating of 2.6 to a rating of 2.0 on the 16th October was the greatest reduction amongst the 16 institutions concerned and was the only case of a downgrading of in excess of 0.5.
(ii) The only other institution awarded an average score with a .6 on the ballot had it rounded up.
(iii) The way in which the 2.6 rating became 2.0 was that two members of the Panel reduced their ratings by one point, giving an average of 2.4 which, after further discussion, was rounded down to 2.0.
(iv) Of the six institutions intensively discussed, Bristol, the London Hospital, the United Medical and Dental Schools and Glasgow University had their ratings raised in consequence, and two, Kings College London and the applicant Institute did not. The four successful institutions had representatives on the Panel. No allegation of bad faith is made but it is submitted that this lends force to the suggestion that something may have gone wrong through unconscious bias.
The applicant has complained consistently of the failure to give it any reasons for this decision. In paragraph 7 of his affidavit of 1st June 1993 Professor Winter puts it this way:
"In all these circumstances, I simply cannot understand why, if there was some substantive factor which justified conferring a score of 2.0, and in particular in reducing the applicant's assessment from the score of 2.6 on the secret ballot, the respondent is unwilling to explain what it was. In academic life, we are used to giving reasons for our assessment of the quality of publications and of research work. If there were characteristics of the work done by the applicant's researchers which justified a score of 2.0, then I do not understand why those characteristics could not briefly be explained to the applicant. I should explain that the consequence to the applicant of the award of a 2.0 rather than a 3.0 is a reduction of about £270,000 in our grant from the respondent for l993/94, plus a consequent decline in private funding because of our reduction in status."
This was a response to the affidavit in reply of Professor Davies sworn on the 22nd April 1993, in paragraph 46 of which he said:
"There was no provision for an appeal against the assessment. This is because of the very nature of a peer review procedure which was formulated to ensure that the best panels were assembled to carry out the review and form a judgement. In these circumstances it was decided that the decision should be final rather than attempting to set up a further layer of appeal. This was made clear to all participating institutions at the outset........ It was also for this reason that I felt unable to respond to Professor Winter's request for reasons why the Panel graded the IDS in the way it did; to do so would undermine the whole nature of the process which had been carefully designed to ensure a fair and uniform procedure for all participating institutions. The Panel is essentially making an expert value judgement on the relative merits of the various institutions. A challenge to one result is a challenge to all."
In his further affidavit of 6th July 1993, paragraph 2, Professor Davies said:
"Professor Winter complained that the evidence filed by the respondent does not explain why the applicant was assessed at level 2.0. This is because it is not possible to explain why without undermining the whole purpose of a peer review assessment. This is not a mechanistic process but one of collective professional judgement. The grade awarded to a particular institution was not determined by a score against specific features but was a synthesis of the independent assessments of the acknowledged experts in the relevant fields. It is not possible to disaggregate the collective view without undermining the nature of the whole research assessment exercise for which professional peer review judgements are a central and essential part. This is also the case for a range of other peer review based evaluations.....
It is only possible therefore to describe the process and the safeguards built into the process according to which that judgement was made."
For the applicant, Mr David Pannick QC now submits that the respondent has acted unfairly by failing to give reasons for its Panel's decision; and that in the absence of reasons, the decision can properly be treated as irrational. For the respondent, Mr Michael Beloff QC argues that in the absence of a general duty to give reasons, no special factors call for reasons here, given in particular that the assessment is an exercise of academic judgement; and that there is nothing ostensibly, let alone demonstrably, irrational in the rating given to the applicant. These submissions are straightforward; the law, unfortunately, is not.
The principal issue is whether the court can properly require a respondent to give reasons for the kind of decision which was made here. This in turn breaks down into two further questions. One is whether either the general demands of fairness or the characteristics of a particular decision can call forth reasons when the decision in question is a collective expert evaluation of quality. The other is whether, if in principle this can happen, reasons ought to be given in the present case. In the present state of development of the law in this area, however, we do not think that they can be addressed and answered seriatim.
Beyond these questions lies Mr Pannick's alternative submission that, absent reasons, the decision under challenge is so inexplicable that it should be struck down as perverse. For reasons to which we will come, however, this submission cannot succeed if the applicant fails on the first issue of law; and if the applicant succeeds on the first issue this submission will be premature.
For his principal submission of law Mr Pannick relies on two recent decisions: that of the Court of Appeal in R v. Civil Service Appeal Board ex parte Cunningham [1992] I.C.R. 816 and that of the House of Lords in R v. Secretary of State for the Home Department ex parte Doody [1993] 3 WLR 154.
The applicant in Cunningham was unfairly dismissed from a post in the Prison Department of the Home Office and so had to seek compensation from the Civil Service Appeal Board, a body set up under prerogative powers, rather than from an industrial tribunal. The board awarded him compensation of £6,500, giving no reasons in spite of requests from the applicant. Had he been able to go to an industrial tribunal, his compensation would have been somewhere near £15,000. Otton J rejected the submission that the award was irrational in itself but declared the decision of the board and its refusal thereafter to supply reasons to be unlawful on the ground that it frustrated a legitimate expectation that civil servants would be treated no less favourably than other employees. Both parties appealed. In the leading judgment, Lord Donaldson MR quoted the already classic passage of the speech of Lord Bridge in Lloyd v McMahon [1987] AC 625, 702 3:
"My Lords, the so called rules of natural justice are not engraved on tablets of stone. To use the phrase which better expresses the underlying concept, what the requirements of fairness demand when any body, domestic, administrative or judicial, has to make a decision which will affect the rights of individuals depends on the character of the decision making body, the kind of decision it has to make and the statutory or other framework in which it operates. In particular, it is well established that when a statute has conferred on any body the power to make decisions affecting individuals, the courts will not only require the procedure prescribed by the statute to be followed, but will readily imply so much and no more to be introduced by way of additional procedural safeguards as will ensure the attainment of fairness."
Applying that test, Lord Donaldson held that, in addition to the ground founded on by Otton J, the applicant succeeded
"upon the broader ground that fairness requires a tribunal such as the Board to give sufficient reasons for its decision to enable the parties to know the issues to which it addressed its mind and that it acted lawfully."
McCowan LJ and Leggatt LJ did not consider that a case of legitimate expectation was made out, but upheld the first instance decision on grounds which, although broader than those relied on by Otton J, were not those of the Master of the Rolls. McCowan LJ said at 831:
"To this day neither he, nor for that matter this court, has any idea why the Board recommended that he receive so little.
As Mr Pannick says, it cries out for some explanation from the board. As I would put it, not only is justice not seen to have been done but there is no way, in the absence of reasons from the board, in which it can be judged whether in fact it has been done. I find that a thoroughly unsatisfactory situation, in which this court should hold, if it can properly do so, that the board ought to give reasons for its recommendation.
In reaching a conclusion as to the propriety of Otton J's order, I am influenced by the following factors. 1. There is no appeal from the board's determination of the amount of compensation.
2. In making that determination the board is carrying out a judicial function. 3. The board is susceptible to judicial review. 4. The procedure provided for by the code, that is to say the provision of a recommendation without reasons, is insufficient to achieve justice. 4. There is no statute which requires the courts to tolerate that unfairness. 5. The giving of short reasons would not frustrate the apparent purpose of the code. 7. It is not a case where the giving of reasons would be harmful to the public interest. These considerations drive me to the view that this is a case where the board should have given reasons and I would, therefore, dismiss the appeal."
Leggatt LJ at 832 said:
"But it seems obvious that for the same reason of fairness that an applicant is entitled to know the case he has to meet, so should he be entitled to know the reasons for an award of compensation, so that in the event of error he may be equipped to apply to the court for judicial review. For it is only by judicial review that the board's award can be challenged."
At 834 he said:
"Without an obligation to give reasons the board's procedures cannot be checked let alone challenged;
and without reasons neither the person dismissed nor the court can tell whether to apply for or to grant judicial review.
In relation to many, if not most, administrative decisions it may well be undesirable, for one reason or another, to give reasons. But there are not here, as in certain contexts there are, any valid grounds for adhering to the general rule that there is no duty to give reasons. On the contrary, there are here particular grounds for departing from the general rule. The applicant has a legitimate grievance, because it looks as though his compensation is less than it should be, and yet he has not been told the basis of the assessment."
Later he said:
"The cardinal principles of natural justice are that no one shall be judge in his own cause and that everyone is entitled to a hearing. But the subject matter of the decision or the circumstances of the adjudication may necessitate more than that. An award of compensation by the board concerns the applicant's means of livelihood for the period to which the award relates. The board's determination binds the Home Office, and also the applicant subject to his right to challenge it by applying for judicial review. But that right is nugatory unless the award is so aberrant as to compel the inference that it must have been wrong, or unless the board explains how the figure was arrived at, so as to enable the applicant to tell whether the award can be successfully impugned.
Those two grounds for requiring reasons elide here."
He concluded:
"In default of explanation the applicant's award was so far below what, by analogy with the award of an industrial tribunal, he was entitled to expect as in my judgment to compel the inference that the assessment was irrational, if not perverse. Because there was no general duty to give reasons, the absence of reasons does not by itself entitle the court to hold that the award was not supportable. But the unexplained meagreness of the award does compel that inference. As Lord Keith of Kinkel said in R v. Secretary of State for Trade and Industry, ex parte Lonrho Plc [1989] 1 W.L.R. 525, 540A:
'The only significance of the absence of reasons is that if all other known facts and circumstances appear to point overwhelmingly in favour of a different decision, the decision maker, who has given no reasons, cannot complain if the court draws the inference that he had no rational reason for his decision'.
In my judgment the duty to act fairly in this case extends to an obligation to give reasons. Nothing more onerous is demanded of the board than a concise statement of the means by which they arrived at the figure awarded. Albeit for reasons which go wider than those relied on by the judge, I too agree that the appeal should be dismissed and cross appeal allowed."
In Doody the issue was very different: the Home Secretary appealed to the House of Lords against a decision of the Court of Appeal that prisoners mandatorily sentenced to life imprisonment for murder were entitled to be told by the Home Secretary what period or periods had been recommended by the judiciary as necessary for the purposes of retribution and deterrence and to be given an opportunity to make representations to him before he determined what the period should be. The Court of Appeal, however, had rejected the further submission that the Home Secretary should give reasons for his decision, at least when departing from the judicial view of the appropriate period to be served. On the Home Secretary's appeal to the House of Lords, the decision of the Court of Appeal in favour of an opportunity to receive information and make representations was upheld, and on the prisoners' cross appeal their Lordships further declared that:
"The Secretary of State is obliged to give reasons for departing from the period recommended by the judiciary as the period which [the prisoner] should serve for the purposes of retribution and deterrence."
Giving the single reasoned speech, in which all their Lordships concurred, Lord Mustill said at 169:
".....the respondents acknowledge that it is not enough for them to persuade the court that some procedure other than the one adopted by the decision maker would be better or more fair. Rather, they must show that the procedure is actually unfair. The court must constantly bear in mind that it is to the decision maker not the court that Parliament has entrusted not only the making of the decision but also the choice as to how the decision is made......
Accordingly, I prefer to begin by looking at the question in the round, and enquiring what requirements of fairness, germane to the present appeal attach to the Home Secretary's fixing the penal element. As general background to this task, I find in the more recent cases on judicial review a perceptible trend towards an insistence on greater openness, or if one prefers the contemporary jargon 'transparency', in the making of administrative decisions."
After dealing with arguments based on the statutory framework, Lord Mustill went on at 170:
"One further argument for the Secretary of State must be mentioned, namely that since the prisoner already knows all the circumstances of his offence, in the light of which the trial judge made his recommendation on the penal element, he can deduce without the need for any more information both the factual basis of the Secretary of State's decision, and the intellectual reasons why the penal element was fixed at a particular term of years. Although some thing akin to this argument has found favour in other cases, I am quite unable to accept it here. The prisoner does indeed know what primary materials were before the court, but he does not know what the judge and the Home Secretary made of them, nor does he know what other materials, not brought out at the trial, may have formed an element in the decision. That the choice of the penal element is not self evident appears quite clearly from the number of occasions on which the Home Secretary's appraisal differs from that of the judges. Either there is something in the material before the Home Secretary which was not known to the judges, or the Home Secretary approaches his task in a way which is different from that adopted by the judiciary when passing sentence. In either event, the missing factor is hidden from view, and the prisoner can do no more than guess what it might be."
Having upheld the decision of the Court of Appeal on the prisoner's right to make informed representations, Lord Mustill went on at 172:
"I accept without hesitation, and mention it only to avoid misunderstanding, that the law does not at present recognise a general duty to give reasons for an administrative decision. Nevertheless, it is equally beyond question that such a duty may in appropriate circumstances be implied, and I agree with the analyses by the Court of Appeal in R v. Civil Service Appeal Board, ex parte Cunningham [1991] 4 All E.R. 310 of the factors which will often be material to such an implication.
Turning to the present dispute I doubt the wisdom of discussing the problem in the contemporary vocabulary of "prisoner's rights", given that as a result of his own act the position of the prisoner is so forcibly distanced from that of the ordinary citizen, nor is it very helpful to say that the Home Secretary should out of simple humanity provide reasons for the prisoner, since any society which operates a penal system is bound to treat some of its citizens in a way which would, in the general, be thought inhumane. I prefer simply to assert that within the inevitable constraints imposed by the statutory framework, the general shape of the administrative regime which ministers have lawfully built around it, and the imperatives of the public interest, the Secretary of State ought to implement the scheme as fairly as he can. The giving of reasons may be inconvenient, but I can see no ground at all why it should be against the public interest: indeed, rather the reverse. This being so, I would ask simply: Is refusal to give reasons fair? I would answer without hesitation that it is not. As soon as the jury returns its verdict the offender knows that he will be locked up for a very long time. For just how long immediately becomes the most important thing in the prisoner's life. When looking at statistics it is easy to fall into the way of thinking that there is not really very much difference between one extremely long sentence and another: and there may not be, in percentage terms. But the percentage reflects a difference of a year or years: a long time for anybody, and longer still for a prisoner. Where a defendant is convicted of, say, several armed robberies he knows that he faces a stiff sentence: he can be advised by reference to a public tariff of the range of sentences he must expect; he hears counsel address the judge on the relationship between his offences and the tariff; he will often hear the judge give an indication during exchanges with counsel of how his mind is working; and when sentence is pronounced he will always be told the reasons for it. So also when a discretionary life sentence is imposed, coupled with an order under section 34. Contrast this with the position of the prisoner sentenced for murder. He never sees the Home Secretary; he has no dialogue with him: he cannot fathom how his mind is working. There is no true tariff, or at least no tariff exposed to public view which might give the prisoner an idea of what to expect. The announcement of his first review date arrives out of thin air, wholly without explanation. The distant oracle has spoken, and that is that.
My Lords, I am not aware that there still exists anywhere else in the penal system a procedure remotely resembling this. The beginnings of an explanation for its unique character might perhaps be found if the executive had still been putting into practice the theory that the tariff sentence for murder is confinement for life, subject only to a wholly discretionary release on licence: although even in such a case I doubt whether in the modern climate of administrative law such an entirely secret process could be justified. As I hope to have shown, however, this is no longer the practice, and can hardly be sustained any longer as the theory. I therefore simply ask, is it fair that the mandatory life prisoner should be wholly deprived of the information which all other prisoners receive as a matter of course. I am clearly of the opinion that it is not.
My Lords, I can moreover arrive at the same conclusion by a different and more familiar route, of which ex parte Cunningham [1991] 4 All E.R. 310 provides a recent example. It is not, as I understand it, questioned that the decision of the Home Secretary on the penal element is susceptible to judicial review. To mount an effective attack on the decision, given no more material than the facts of the offence and the length of the penal element, the prisoner has virtually no means of ascertaining whether this is an instance where the decision making process has gone astray. I think it important that there should be an effective means of detecting the kind of error which would entitle the court to intervene, and in practice I regard it as necessary for this purpose that the reasoning of the Home Secretary should be disclosed."
We readily accept Mr Beloff's submission that Lord Mustill was not holding, in the final part of this passage, that reasons are called for wherever it is desired to know whether grounds for challenge exist; for to do so would be to create just such a general duty as Lord Mustill at the start of the passage was careful to exclude. Rather he was holding that in the situation of near total ignorance and impotence in which the prisoner found himself about something as vital to him as his prospects of liberty, such a duty arose. It follows nonetheless from Lord Mustill's reasoning that the "more familiar route" exemplified by Cunningham may be broader than the Cunningham situation alone and capable of embracing other situations in which "it is important that there should be an effective means of detecting the kind of error which would entitle the court to intervene".
This being so, it seems both desirable and practical to test by a common standard both the fairness of not telling a person the reasons for a decision affecting him and the desirability of exposing any grounds of legal challenge. There are, moreover, reasons of principle for a unitary test. As the judgments in Cunningham show, one aspect of unfairness may be precisely the inability to know whether an error of law or of process has occurred. But since the latter is not a freestanding ground for requiring reasons (for if it were, it would apply universally), it can only be on grounds of fairness that it will arise; so that the need to know whether there has been an error of law or of process is rightly seen not as an alternative to the demands of fairness but as an aspect of them.
This approach places on an even footing the multiple grounds on which the giving of reasons may in any one case be requisite. The giving of reasons may among other things concentrate the decision maker's mind on the right questions; demonstrate to the recipient that this is so; show that the issues have been conscientiously addressed and how the result has been reached; or alternatively alert the recipient to a justiciable flaw in the process. On the other side of the argument, it may place an undue burden on decision makers; demand an appearance of unanimity where there is diversity; call for the articulation of sometimes inexpressible value judgments; and offer an invitation to the captious to comb the reasons for previously unsuspected grounds of challenge. It is the relationship of these and other material considerations to the nature of the particular decision which will determine whether or not fairness demands reasons.
In the light of such factors each case will come to rest between two poles, or possibly at one of them: the decision which cries out for reasons, and the decision for which reasons are entirely inapposite. Somewhere between the two poles comes the dividing line separating those cases in which the balance of factors calls for reasons from those where it does not. At present there is no sure indication of where the division comes. Asked to give an example of the kind of decision in which in the light of his submissions fairness will not require reasons to be given, Mr Pannick was unable or unwilling, at least without further reflection, to commit himself. No doubt the common law will develop, as the common law does, case by case. It is not entirely satisfactory that this should be so, not least because experience suggests that in the absence of a prior principle irreconcilable or inconsistent decisions will emerge. But from the tenor of the decisions principles will come, and if the common law's pragmatism has a virtue it is that these principles are likely to be robust. At present, however, this court cannot go beyond the proposition that, there being no general obligation to give reasons, there will be decisions for which fairness does not demand reasons. It follows that in appraising each case, the present included, too catholic an approach will amount to generalising what is still a particular obligation though we are not prepared to accept Mr Beloff's contention that it is any longer an exceptional one.
It may be useful here to touch on an unresolved but potentially important question highlighted by the relief which Mr Pannick seeks in the form of mandamus to give reasons: is the giving of reasons, in a case where it is "the justice of the common law" which requires it, a freestanding duty enforceable by mandamus or simply, as Mr Beloff contends, a form of relief where independent grounds for it, such as irrelevant factors or irrationality, are established? If it is the latter, then (pace the limited remedies in RSC Order 53) relief should consist of the remission of the matter for reasons to be given, and if adequate reasons are then not given the court will be entitled to infer that there were none and that the decision was therefore irrational: see Lonrho v Secretary of State for Trade and Industry [1989] 1 WLR 525, 539 per Lord Keith. Neither Cunningham nor Doody gives an explicit answer. In the former case, however, a declaration was made and upheld on appeal that the refusal to give reasons was unlawful and ultra vires, strongly suggesting the existence of an independent legal obligation. In the latter case the House of Lords declared that the Home Secretary is obliged to give reasons for departing from the period recommended by the judiciary. If in such a case the maximum remedy was simple remission for the giving of reasons with the threat of quashing if none were given, mere non compliance would frustrate the court's order without redress. The prisoner would lose, by its quashing, even the first review date set by the Home Secretary. This too powerfully suggests that the obligation to give reasons, where it is established, is an independent and enforceable legal obligation and hence a ground of nullity where it is violated. Such an outcome would have a satisfactory symmetry with the ordinary consequence of non compliance with a statutory requirement to give reasons. In both cases the discretion as to remedy would of course remain.
That an apparently inexplicable decision is not, however, a general requisite for leave or for relief is now clear from the decision of the House of Lords in Doody, where the length of the actual tariff periods set by the Home Secretary when matched against the facts of appellants' crimes formed no part of the grounds for requiring reasons in the interests of fairness. It follows that an apparently inexplicable decision may be a sufficient but is not a necessary condition for requiring reasons; it may equally be fair to require them on other grounds. It is arguable that since the decision in Doody the role of the inexplicable decision is to be regarded as evidential rather than legal, bearing principally on the discretionary decisions whether to grant leave and whether to grant relief by pointing to the need for reasons in the particular case. But we prefer the view that in the present state of the law there are two classes of case now emerging: those cases, such as Doody, where the nature of the process itself calls in fairness for reasons to be given; and those, such as Cunningham, where (in the majority view) it is something peculiar to the decision which in fairness calls for reasons to be given. This does not mean that differing tests of fairness are to be applied; only that, as always, the requirements of fairness will vary with the process to which they are being applied. In this context we unhesitatingly reject Mr Beloff's submission that the judicial character of the Civil Service Appeal Board and the quasi judicial function of the Home Secretary in relation to life sentence prisoners distinguish the cases requiring reasons from cases of purely administrative decisions such as the present one. In the modern state the decisions of administrative bodies can have a more immediate and profound impact on people's lives than the decisions of courts, and public law has since Ridge v Baldwin [1964] AC 40 been alive to that fact. While the judicial character of a function may elevate the practical requirements of fairness above what they would otherwise be for example by requiring contentious evidence to be given and tested orally what makes it "judicial" in this sense is principally the nature of the issue it has to determine, not the formal status of the deciding body.
The first limb of Mr Pannick's submission is accordingly that the decision of the respondent was of a kind for which fairness requires that reasons be given. His written contention is that this will be the case "when the relevant decision has important consequences for the individual or body concerned, especially if the absence of reasons makes it very difficult for the applicant and the court to know whether the respondent has acted by reference to irrelevant factors, and especially if there is no justification for withholding reasons". In our view this formula will not do. The absence of reasons always makes it difficult to know whether there has been an error of approach. The question of justification for withholding reasons logically comes after the establishment of a prima facie duty to give them. Neither can therefore add to the principal ground advanced, which is of such width that it would make a duty to give reasons a universal rule to which the only exception would be cases of no importance to anybody. There are certainly good arguments of public law and of public administration in favour of such a rule, but it is axiomatically not, or not yet, part of our law. It remains to be seen what the "continuing momentum in administrative law towards openness of decision making" (per Lord Mustill in Doody at 174A) will bring.
To his written criteria Mr Pannick has added, in argument, the absence of an appeal. This may well be crucial in justifying recourse to the court, but it cannot logically render unfair a silence which is otherwise not unfair.
In this situation it is not in our view possible to formulate a different or more precise rule or test in substitution for Mr Pannick's. Instead it is necessary to turn to the factors on which Mr Pannick relies and to see which pole they tend towards. They are these:
(a) The decision is very important to the future of the applicant Institute. Both directly and indirectly it affects the Institute's funding in very large sums.
(b) It is a decision to lower the applicant's grading from that which it previously enjoyed.
(c) In the absence of reasons there is a real risk that the respondent may have been influenced by irrelevant factors. This is fully argued out by Mr Pannick by reference to the detailed history of the readjusted rating, but for reasons which we will come to it is not appropriate to set them out here.
(d) The applicants are frustrated by their inability to know whether the respondent acted on irrelevant factors.
(e) There is no justification for the failure to give reasons, and certainly not that given by Professor Davies.
We will consider them in reverse order.
We agree with practically all of Mr Pannick's criticisms of the respondent's objections to giving reasons. It is not suggested for a moment that the panel collectively and its members individually did not have, and did not have to have, reasons for each grading they made. It seems to us disingenuous to talk, as Prof Davies repeatedly does, about the impossibility of "disaggregating the collective view", when that view was an aggregate view in the first place. Schiemann J's reference in R v Poole Borough Council ex parte Beebee [1991] 2 PLR 27, 1 to "the theoretical difficulties of establishing the reasoning of a corporate body which acts by resolution" does not bear directly on what in our judgment is more correctly viewed as a collegiate than a corporate decision. To suggest that the panel's conclusion cannot be explained without "undermining the whole purpose of a peer review assessment" seems to us, with respect, to devalue the assessment process. It would take a great deal more than this to persuade us that experienced and distinguished academics, whether individually or collectively, cannot assign reasons for their own judgment. It is unlikely that a similar reticence affects them either individually at examiners' meetings or collectively when writing joint reports. But Mr Pannick's contention assumes what is yet to be proved, for if otherwise there is no obligation to give reasons, the question whether the respondent is justified in withholding them does not arise. The casuistic grounds advanced by the respondent for not giving reasons, while they do not help the respondent, do not help the applicant.
Inability to know whether the rating has been arrived at in reliance on irrelevant factors is universal in unreasoned decisions, except where such factors are revealed by chance.
It is a ground for a general duty to give reasons, but for that very reason it cannot found a particular duty.
The oddities of the decision, on what is known to the applicant, meet a further logical difficulty. In considering whether the nature of the material decision and decision making process require the giving of reasons, the particular outcome does not matter. As illustrations of the risk of a concealed flaw such oddities will form part of the broad argument in favour of a general duty, but no more. It is under the alternative approach the argument from an aberrant outcome that these factors may matter, and to this we will come.
Mr Pannick has not dwelt on the submission that the lowering of the grade contributed to the need for reasons. Not only was this a product and not a part of the questioned process; the whole statutory exercise was predicated on the proposition that no institution had any right to have its previous grade preserved until displaced: every point had to be freshly earned. Mr Pannick was wise to pass over it.
This leaves the importance of the decision to the applicant's future. We do not doubt a word of Professor Winter's evidence about this. Not only does a reduced rating directly affect the quantum of government funding; it reduces the standing and competitiveness of the institute in seeking outside funding. The morale and the future of an institution such as the applicant are shaken by such a decision. We do not accept Mr Beloff's suggestion that an institutional applicant is fundamentally different from an individual in this context. In many contexts the difference will be real and material, but for what is a collegiate body of individuals collaborating in one area of science, we would be loth to diminish its expectations of the protection of the law below that which individuals can expect.
The chief benchmark of significance which we have at present in this setting is the Doody case. There the applicant knew the evidence on which he had been convicted but little else, while a considerable body of highly relevant matter had accumulated in the hands of the decision maker and was going to affect many years of his liberty. If the Home Secretary were then to depart from the judicial view of tariff, it is not easy to think of a stronger case for the disclosure of reasons not merely to the applicant but to all mandatory life sentence prisoners, to each of whom result of the case will necessarily apply. Equally here the argument, it seems to us, must be good for all applicants, not just disappointed ones, if they want to know why they have been rated as they have been. One would like to be able to hold that for all such applicants, disappointed or not, the importance of the decision alone was enough. But to do so would generalise the duty to give reasons to a point to which this court, at least, cannot go.
We must therefore look also at the other indicia: the openness of the procedure, widely canvassed in advance and published in circular form; the voluntary submission of self selected examples of work; the judgment of academic peers. These, it seems to us, shift the process substantially away from the pole represented by Doody, not on mere grounds of dissimilarity (there will be many dissimilar cases in which reasons are nevertheless now required) but because the nature of the exercise was that it was open in all but its critical phase, and its critical phase was one in which, as Prof Davies deposes, "the grade awarded to a particular institution was not determined by a score against specific features". We shall return to this, which we find remarkable, but it is a fact and not one which Mr Pannick has been able to assault on legal grounds. In the result, the combination of openness in the run up with the prescriptively oracular character of the critical decision makes the respondent's allocation of grades inapt, in our judgment, for the giving of reasons, notwithstanding the undoubted importance of the outcome to the institutions concerned.
From this case specific conclusion, it is possible to generalise to a certain extent. The only mystery left in the process is precisely why the final grade of 2 was arrived at. As Mr Pannick points out, the evidence is replete with answers to the question how it was arrived at, but not why. The question "why", in isolation as it can now be seen to be, is a question of academic judgment. We would hold that where what is sought to be impugned is on the evidence no more than an informed exercise of academic judgment, fairness alone will not require reasons to be given. This is not to say for a moment that academic decisions are beyond challenge. A mark, for example, awarded at an examiners' meeting where irrelevant and damaging personal factors have been allowed to enter into the evaluation of a candidate's written paper is something more than an informed exercise of academic judgment. Where evidence shows that something extraneous has entered into the process of academic judgment, one of two results may follow depending on the nature of the fault: either the decision will fall without more, or the court may require reasons to be given, so that the decision can either be seen to be sound or can be seen or (absent reasons) be inferred to be flawed. But purely academic judgments, in our view, will as a rule not be in the class of case, exemplified (though by no means exhausted) by Doody, where the nature and impact of the decision itself call for reasons as a routine aspect of procedural fairness. They will be in the Cunningham class, where some trigger factor is required to show that, in the circumstances of the particular decision, fairness calls for reasons to be given.
Is there then such a trigger factor here? The second limb of Mr Pannick's submission is that the applicant Institute has been confronted with a decision which, on the evidence, is inexplicable: the Institute's excellence is widely acknowledged and attested; its original rating of 2.6 would have qualified for rounding up to a 3; and the reduction to 2.4 and hence to a rating of 2 followed reconsideration in circumstances which, at the lowest, can be regarded as unsatisfactory. Mr Beloff responds, and we agree with him, that neither intrinsically nor on the evidence is there a sufficient basis on which this court can hold the eventual rating to be so aberrant as in itself to call for an explanation. We lack precisely the expertise which would permits us to judge whether it is extraordinary or not. It may be a misfortune for the applicant that the court, which in Cunningham could readily evaluate the contrast between what the Board awarded and what an industrial tribunal would have awarded, cannot begin to evaluate the comparative worth of research in clinical dentistry; but it is a fact of life. The applicant's previous grading, the volume and frequency of citation of its research and the high level of peer reviewed outside funding which it has attracted, to all of which Mr Pannick points, may well demonstrate that the applicant has been unfortunate in the grading it has received, but such a misfortune can well occur within the four corners of a lawfully conducted evaluation.
Mr Pannick also points to the fact, undoubtedly an uncomfortable one for the respondent, that of the six institutions reconsidered in detail, the four which were uprated had members on the panel, albeit these members very properly left when their own institutions were discussed, while the two which were downrated did not. He does not found upon this as a discrete head of challenge. Rather and wisely he relies on it as converting into a probability the possibility that the oddity of the decision betrays a flawed approach. But without the major premise of patent oddity the minor premise of unconscious bias cannot advance his argument.
The inexorable conclusion is that neither the nature of the material decision nor its eventual content was such as to oblige the respondent to furnish reasons for it.
However, we have pointed out that, irrespective of any obligation to give reasons, it goes without saying that any decision making body, above all one with the powers and duties of the respondent, must have reasons for every decision it takes, and we have given our grounds for rejecting Mr Beloff's fallback argument that in matters of academic judgment such reasons are somehow incapable of being formulated or articulated. Each panel member will have had reasons for his or her own initial rating, and in discussion each will have been in a position to advance those reasons and to modify them in the light of the reasoning of others. This case, however, has had to proceed on the given fact that the respondent, a modern body charged with important public functions and recognising, as the respondent does, the need for uniformity and consistency of approach, did not work by agreeing the criteria by which it would judge the point on the scale to which each institution was to be allocated, so that each panel member might before and after discussion make an evaluation within an agreed range on each criterion. Somehow, according to its minutes, the Clinical Dentistry Panel on the l5th l6th October agreed that at its coming meeting on the 23rd October "all ratings will be reviewed to ensure that the criteria on which assessment had been based had been applied consistently to all institutions across the board; and to enable the panel to make any modifications they felt necessary to the ratings recorded", when on the evidence no such criteria of assessment existed: the only fixed points were the verbal descriptions of the level represented by each scale point in Annex C, but not the criteria by which the allocation of institutes to points on this scale was to be carried out. Indeed we have cited above Professor Davies' unequivocal testimony that "the grade awarded to a specific institution was not determined by a score against specific features".
Industrial tribunals and courts have come to appreciate that such a procedure enables evaluation to be carried out on a basis of parity both among panel members and among candidates, providing a uniform foundation for differential evaluations and making the eventual collective evaluation both explicable and defensible, so that if as may happen the law calls for reasons to be given, they exist and do not have to be called into being. Parenthetically, such a procedure also enables candidates, successful as well as unsuccessful, to be told why and in what areas they have done well or badly information which may be as important to them as their actual rating. It is of course not for the court to advise an independent body on how to arrange and conduct its procedures. But it is necessary for public decision making bodies to appreciate that there are already some circumstances (e.g. where unlawful race or sex discrimination is alleged), and more may well come, in which their legal position may depend on their ability to account intelligibly for their decisions by explaining not simply how but why they have reached them. This much, we think, bears a practical relationship to the movement of the law towards open or "transparent" decision making to which Lord Mustill refers in Doody.
In summary, then
1. There is no general duty to give reasons for a decision, but there are classes of case where there is such a duty.
2. One such class is where the subject matter is an interest so highly regarded by the law for example personal liberty that fairness requires that reasons, at least for particular decisions, be given as of right.
3. (a) Another such class is where the decision appears aberrant. Here fairness may require reasons so that the recipient may know whether the aberration is in the legal sense real (and so challengeable) or apparent.
(b) It follows that this class does not include decisions which are themselves challengeable by reference only to the reasons for them. A pure exercise of academic judgment is such a decision.
(c) Procedurally, the grant of leave in such cases will depend upon prima facie evidence that something has gone wrong. The respondent may then seek to demonstrate that it is not so and that the decision is an unalloyed exercise of an intrinsically unchallengeable judgment. If the respondent succeeds, the application fails. If the respondent fails, relief may take the form of an order of mandamus to give reasons, or (if a justiciable flaw has been established) other appropriate relief.
But just as it is outwith this court's powers to judge degrees of excellence in clinical dentistry research, or for that matter the wisdom of a body's administrative arrangements, so it is not open to this court to require the communication of reasons, even where such reasons must necessarily exist, in the current absence of a legal basis for the requirement. We would accordingly dismiss this application.
LORD JUSTICE MANN: Mr Richards?
MR RICHARDS: My Lord, I ask for my costs.
MR PANNICK: My Lords, I would suggest to your Lordships that the appropriate order in the circumstances is no order as to costs, given the importance of the issues that your Lordships have dealt with.
LORD JUSTICE MANN: Do you wish to say anything about that?
MR RICHARDS: My Lord, I have clear instructions that the financial memorandum that my clients have with the Government requires me to apply for my costs. At the end of the day it is a matter of discretion for you. In my submission, this is a case where costs should follow in the normal course of events. (Learned Judges confer)
LORD JUSTICE MANN: The application will be dismissed with costs. We are very grateful to counsel for the argument in a case which has raised a question of some general importance.
© Crown Copyright
