SIR DAVID WILLIAMS LECTURE 2019
PRINCIPLE AND PRAGMATISM IN PUBLIC LAW
Brenda Hale

It is such a pleasure as well as a privilege to be giving thus lecture in memory of Sir David Williams – a great academic and author, a great teacher and a great University leader but above all a great person. About 45 years ago he was an external examiner at the University of Manchester and we entertained the external examiners and staff to supper in our home in Derbyshire. But David went missing. I discovered him chatting on the stairs with my two-year-old daughter who, never wanting to be left out, had crept downstairs to see what was going on. He was much more interested in entertaining her than in the rest of us. I am sure that his family here today would recognize that!

But what to talk about to honour such a man? Earlier this year I gave the Freshfields lecture in this Faculty on the subject of ‘Principle and pragmatism in private law’. This was, of course, an echo of the Hamlyn lectures given by the late, great Patrick Atiyah in 1987, on ‘Pragmatism and Theory in English Law’. He discussed the tendency of English lawyers – by which he meant practitioners and judges rather than academics – not only to be more inclined towards the pragmatic and hostile to the theoretical but positively to glory in this approach. He went on to consider the strengths and weaknesses of the pragmatic tradition. He deplored ‘a more general tendency to decide cases ad hoc, to try to settle disputes by wholly pragmatic means, without regard to the principles of law and the broader purposes which those principles must have’ (p 126). I found it easy earlier this year to consider examples where judicial pragmatism had indeed overcome principle in contract, tort and family law.

So it seemed obvious to suggest that this lecture should be entitled ‘Principle and pragmatism in public law’. But it has not turned out such an easy lecture to write. It is possible to craft pragmatic solutions to private law disputes between individuals or businesses. In private law this can be dressed up under such concepts as ‘fair, just and reasonable’ or ‘public policy’. But is there an equivalent in public law? I have picked out three topics: one which seems to me to be almost entirely pragmatic rather than principled, and that is the practice sometimes known as ‘deference’; one which seems to me to be dressed up as principled but is arguably more pragmatic, and that is the doctrine of legitimate expectation; and one which seems to me to be soundly based in principle, even though its results may sometimes seem far from pragmatic, and that is the principle of legality.



Deference

Judicial deference is the process whereby courts defer, or attach over-whelming weight, to the judgments of Parliament or the executive on certain matters. Of course, the judges don’t like the term ‘deference’. In the Pro-Life Alliance case,
Lord Hoffmann did not think that its ‘overtones of servility, or perhaps gracious concession’ were appropriate to describe what was happening. In a society based on the rule of law and the separation of powers it was necessary to decide which branch of government has the decision-making power and what those powers are. In that case, the question was whether Parliament was entitled to make party political broadcasts subject to the requirements of taste and decency to which all other broadcasts were subject. That, in his view was a perfectly proper decision for Parliament, as representative of the people, to make.

Then again, in the Lord Carlile case, Lord Sumption suggested that academic criticism of the concept ‘arises from the word, with its overtones of cringing abstention in the face of superior status’ (para 22). Assigning weight to the decision-maker’s judgment had nothing to do with deference in the ordinary sense of the term. It had two sources: the constitutional principle of the separation of powers and the ‘pragmatic view about the evidential value of certain judgments of the executive, whose force will vary according to the subject-matter’. That case was not about a decision taken by Parliament, but about a decision taken by the Home Secretary that excluding a prominent Iranian dissident from entering the country was conducive to the public good.

Deference is not the same as non-justiciability. It does not apply automatically with reference to particular subject-matters and carries no presumption that the court will have inadequate expertise. Each case is judged in its own context. It is this variability which opens up the element of pragmatism: as Professor Jowell observed, ‘there is no magic legal or other formula to identify “the discretionary area of judgement” available to the reviewed body’ .

Like Professor Jowell, I don’t think that appeals to the separation of powers or to the greater democratic legitimacy of executive decision-makers are a great deal of help in enabling us to decide in which cases to ‘defer’ and in which cases not to do so. It is, of course, correct that Parliament is accountable to the electorate and the executive is accountable to Parliament. That is true of all legislative and executive decision-making. It cannot invariably be a reason for deferring to what Parliament or the executive has decided. The whole purpose of giving legal recognition to individual human rights is to enable them to be asserted in the face of decisions made by democratically accountable actors. We have to ask ourselves very carefully which decisions we should respect and why.

It is important, I think, to distinguish between the decisions of Parliament, as represented by Acts of Parliament, and the decisions of the executive. Judicial views on deference to Parliament differ widely. Lord Steyn in his famous lecture, ‘Deference: a Tangled Story’ endorses the statement of Madam Justice McLachlin in the Supreme Court of Canada on the limits of the deference principle:

‘Care must be taken not to extend the notion of deference too far…Parliament has its role: to choose the appropriate response to social problems within the limiting framework of the Constitution. But the courts also have a role: to determine, objectively and impartially, whether Parliament’s choice falls within the limiting framework of the Constitution. The courts are no more permitted to abdicate their responsibility than is Parliament. To carry judicial deference to the point of accepting Parliament’s view simply on the view that the problem is so serious and the solution difficult, would be to diminish the role of the courts in the constitutional process and to weaken the structure of rights upon which our constitution and nation is founded’.

We do not, of course, have a written Constitution, but we do have a Human Rights Act which expressly contemplates that the courts will consider the compatibility of provisions in Acts of Parliament with the Convention rights: why else do we have the duty of conforming interpretation in section 3(1) and the power to make declarations of incompatibility in section 4? Furthermore, we have to have regard to the jurisprudence of the European Court of Human Rights and other Council of Europe organs. So what part should deference play in that process? Should the courts take a different view of the respect owed to the judgment of Parliament according to whether the matter is (a) one on which Strasbourg has made its view clear, or (b) one on which Strasbourg has not yet expressed a definitive view, or (c) one which Strasbourg would regard as falling within the ‘margin of appreciation’ allowed to Member States?

The last is the most difficult case. And in the case of Nicklinson, the Supreme Court expressed a range of views. Most of you will know the facts, but here is a brief recap.

Mr Nicklinson had suffered a catastrophic stroke which left him almost completely paralysed, unable to speak or carry out any physical functions on his own except limited movements of his eyes and head, but he was not dependent on life support. He communicated through the use of an eye-blink computer. After some years he decided that enough was enough and wished to end his life. But he did not wish to inflict upon his family the pain and suffering involved in watching him starve himself to death. He wanted ‘a more humane and dignified exit from the world’. The case began as an argument that having a doctor administer a lethal injection would be justified under the common law defence of necessity or duress of circumstances. It turned into an argument about assisted suicide after it emerged that an Australian doctor had invented a machine which could be activated to deliver a lethal drug by means of the eye-blink computer. Hence, in the Supreme Court, the focus was on whether the absolute ban on assisting suicide, contained in section 2 of the Suicide Act 1961, was incompatible with Mr Nicklinson’s article 8 rights.

There was no doubt that his article 8 rights were engaged. In the cases of Haas, Koch and Gross, the European Court of Human Rights stated that:

‘an individual’s right to decide by what means and at what point his or her life will end, provided she or he is capable of freely reaching a decision on this question and acting in consequence, is one of the aspects of the right to respect for private life within the meaning of article 8 of the Convention.’

In Pretty v United Kingdom, however, the court had also taken the view that a universal ban on assisting suicide, especially when coupled with a flexible prosecution policy, was justified in order to protect vulnerable people. As a supra-national court, they could fall back upon the wide margin of appreciation accorded to Member States in an area where there is as yet no European consensus in favour of permitting assisted suicide – quite the reverse.

When the Strasbourg court accords a wide margin of appreciation to Member States, it is left to their own constitutional arrangements to decide what to do. So should the Supreme Court leave the question entirely to Parliament? In other words, is Parliament the sole arbiter of what is and is not compatible with the Convention rights in United Kingdom law? Or should the Supreme Court address the issues in proceedings brought under the Human Rights Act and reach conclusions on them, of course leaving it to Parliament to decide whether the law should be changed?

In Nicklinson, the Justices all, I think, took the view that there were occasions when the courts could decide upon the United Kingdom’s solution to an issue which Strasbourg would leave to the Member State. The House of Lords had done this with Northern Ireland’s ban on joint adoptions by unmarried couples. But that was a ban contained in delegated, not primary legislation.

Four of the Justices in Nicklinson thought that the whole issue should be left to Parliament and the court should not even express a view upon what the answer was. Lord Sumption gave three reasons: first, it involved a choice between two fundamental but mutually inconsistent moral values upon which there was no consensus in society; second, Parliament had made the choice; third, the Parliamentary process was a better way of resolving issues involving controversial and complex issues of fact arising out of moral and social dilemmas (paras 230-232). This is, of course, a view which he has since elaborated very persuasively in his Reith lectures. Lord Hughes took the simple view that this was very clearly a matter for Parliament (para 267). Lord Clarke agreed that Parliament was the preferable forum and that imposing the personal opinions of professional judges ‘would lack all constitutional legitimacy’ (para 293). Lord Reed accepted that the Human Rights Act has entailed some adjustment of the respective constitutional roles of the courts, the executive and the legislature, but it did not eliminate the differences between them. It did not alter the fact that certain issues were by their nature more suitable for determination by government or Parliament than by the courts (para 296). Were these views pragmatic or principled?

Five of the Justices thought that it would not be institutionally inappropriate for the court to address the issue. Ruling it out would be an abdication of judicial responsibility. Lord Neuberger observed that it was not possible precisely to identify the boundary between the area where it is legitimate for the courts to step in and the area where it is not (para 101). But he, along with Lord Mance and Lord Wilson, thought that the time was not yet ripe to do so. This was for largely pragmatic reasons. I have great sympathy with them. The case had started on quite a different footing and so the evidence was not addressed to the issue of assisted suicide. As Lord Mance put it, this was ‘an invitation to short cut potentially sensitive and difficult issues of fact and expertise, by relying on secondary material’.

He was, of course, right about that. But my own view, shared by Lord Kerr, was that the question could be answered by reference to principle rather than evidence. It could confidently be concluded that the ban was over-broad (as the Canadians would later put it when deciding that their own ban was unconstitutional). Experience with comparable end of life decisions showed that a procedure could be devised for identifying those people who should, exceptionally, be allowed help to end their own lives. I too expressed the view that Parliament was much the preferable forum in which the issue should be decided. Indeed, in a sense, it is the only forum in which it can be decided, because only Parliament can change the law. But having reached the firm conclusion that the law was not compatible with the Convention rights, there was little to be gained and much to be lost by not making a declaration of incompatibility (para 300).

It does seem to me that section 4 of the Human Rights Act has introduced an important change to our constitutional arrangements, by expecting courts to make the same sorts of judgment as are made by the courts in countries which do have a written Constitution in which fundamental rights are entrenched, thereby creating a feedback mechanism from the courts to the sovereign legislature. We can accept, for the reasons given by Lord Sumption, that Parliament is the preferable place for such decisions to be made; but if Parliament fails to act in the face of clear incompatibility, it is our duty to say so. As Lord Neuberger pointed out, ‘difficult or unpopular decisions which need to be taken are on some occasions more easily grasped by judges than by the legislature’ (para 104). Perhaps there is an element of pragmatism on the part of Parliament?

Hence deference is, in my view, principally relevant when it comes to making judgments about the justification for executive interferences with Convention rights. We are required to decide whether an interference is ‘necessary in a democratic society’ or whether a ‘fair balance’ has been struck between the rights of the individual and the interests of society or the community at large. In many contexts these are difficult things for judges to decide. The courts may well think that political actors are better qualified to decide them – not so much because of their democratic legitimacy as because of their practical competence. They know much more about it than we can ever do.

Two examples spring to mind. In the Belmarsh case, seven out of the eight judges in the majority (and I suspect also Lord Walker, the sole dissenter) were prepared to take the government’s word for it that there was a ‘public emergency threatening the life of the nation’ which would justify derogating from the right to liberty in article 5. I put it bluntly (para 226):

‘Any sensible court, like any sensible person, recognises the limits of its expertise. Assessing the strength of a general threat to the life of the nation is, or should be, within the expertise of the government and its advisers. . . . Protecting the life of the nation is one of the first tasks of a government in a world of nation states.’

Lord Hoffmann famously dissented on this issue. But he did so, not because he did not accept that there was credible evidence of a threat of serious terrorist outrages, but on the basis that the government – and presumably the rest of us – had misunderstood what a threat to the life of the nation was. Terrorist groups did not threaten the life of the nation: terrorism does not threaten our institutions of government or our existence as a civil community (para 96).

More difficult was the Lord Carlile case, in which we were asked to take the government’s word for it that to allow a prominent Iranian dissident into England in order to meet with Parliamentarians in the Palace of Westminster would endanger our ‘fragile but imperative’ relations with Iran. We all agreed that the court was the final arbiter of whether the undoubted interference in the free speech rights of Ms Rajavi and the Parliamentarians was a proportionate means of achieving a legitimate aim. But we also agreed that on some parts of the analysis, the government was better placed than we were to make the judgment.

This was essentially a question of fact. How strong were the risks? As Lord Sumption put it in Lord Carlile:

‘How is the court to determine where the balance lies if (i) it has no means of independently assessing the seriousness of the risks or the gravity of the consequences were they to materialise, and (ii) the Secretary of State is not shown to have committed any error of principle in her own assessment of them. . . . We are not in point of law bound to accept the factual assessment of the Foreign Office about the impact on our relations with Iran of admitting Mrs Rajavi to the United Kingdom. But if we reject it we must have a proper basis for doing so. In this case, there is none. There is no challenge to the primary facts. We have absolutely no evidential basis and no expertise with which to substitute our assessment of the risks to national security, public safety and the rights of others for that of the Foreign Office.’

This was, as he himself had earlier recognised ‘no more than a pragmatic view about the evidential value of certain judgments of the executive’.

Far more difficult are those cases which turn on the justification for decisions of the executive in matters of socio-economic policy. There is much to be said for Dworkin’s dichotomy – between principle (involving moral rights against the state) and policy (involving utilitarian calculations of public good and the allocation of public resources). But there are situations where principle and policy overlap. We have had a torrid time with the benefit cap, the so-called bedroom tax and the revised benefit cap. On the one hand, these were all complaints of discrimination – indirect discrimination against women, direct discrimination against disabled people, Thlimmenos discrimination (that is, failing to treat persons in different positions differently) against lone parents. The courts have tended to claim some expertise in recognising unjustified discrimination: the House of Lords did so in the adoption case of Re G, for example. One of the principal purposes of protecting fundamental rights is to safeguard what may be vulnerable or unpopular groups or individuals from the will of the majority.

On the other hand, the courts have tended to adopt the same approach as Strasbourg in relation to measures of socio-economic policy, in particular when dealing with welfare benefits: the test is whether the measure is ‘manifestly without reasonable justification’. Even this has led to some sharp divergences of view. Denying an extra bedroom where there was a clearly demonstrable medical need for one was manifestly without reasonable justification. But depriving families with children of subsistence level benefits was not. There is obviously deference here by the majority both to institutional competence and to democratic legitimacy. But is there not also an even greater element of pragmatism in it, not only here but also in Strasbourg?

Legitimate expectation

Another area where the courts might be accused of being guided by pragmatic rather than principled considerations is legitimate expectation. The doctrine of legitimate expectation is firmly presented as a principle. Essentially, a public authority which has, by a promise or practice, conferred on a person a legitimate expectation of a procedural or substantive benefit, may not frustrate that expectation without justification. However, the underlying rationale for the concept is expressed rather broadly: the prevention of abuse of power and the promotion of good administration. A powerful exposition is that of Laws LJ in
R (Bhatt Murphy) v Independent Assessor:

‘The power of public authorities to change policy is constrained by the legal duty to be fair (and other constraints which the law imposes). A change of policy which would otherwise be legally unexceptionable may be held unfair by reason of prior action, or inaction, by the authority. If it has distinctly promised to consult those affected or potentially affected, then ordinarily it must consult (the paradigm case of procedural expectation). If it has distinctly promised to preserve existing policy for a specific person or group who would be substantially affected by the change, then ordinarily it must keep its promise (substantive expectation). If, without any promise, it has established a policy distinctly and substantially affecting a specific person or group who in the circumstances was in reason entitled to rely on its continuance and did so, then ordinarily it must consult before effecting any change (the secondary case of procedural expectation). To do otherwise, in any of these instances, would be to act so unfairly as to perpetrate an abuse of power.’

That all looks very precise. But the analysis has a repeated theme of fairness. There is plainly scope for pragmatism to play a part in the determination of what fairness requires in a particular context. When a court is testing whether there has been an abuse of power, it is not confined to a Wednesbury review, but has itself to weigh the impact of the frustration of the expectation on the individual or group against the wider public interest in failing to uphold it.

The Supreme Court has recently looked at this area in the Finucane case . The question was whether the government should be held to a promise, made in 2004, to hold a public inquiry into the death of the Belfast solicitor Patrick Finucane, who had been brutally murdered by loyalist terrorists in front of his wife and children in 1989. The requirements of good administration were again identified as underpinning the doctrine. Hence the court rejected the notion that public authorities could resile from their commitments simply because the person or group to which such promises were made were unable to demonstrate a tangible disadvantage.

However, when considering whether the frustration of this legitimate expectation was justified, the test adumbrated was one of fairness. The court accepted that it should give great weight to ‘macro-political’ issues and that the holding of a public inquiry in the circumstances was properly a matter for the Prime Minister’s political judgment. Lord Kerr said this:

‘Where political issues overtake a promise or undertaking given by government, and where contemporary considerations impel a different course, provided a bona fide decision is taken on genuine policy grounds not to adhere to the original undertaking, it will be difficult for a person who holds a legitimate expectation to enforce compliance with it.’

Professor Mark Elliott has been critical of the reliance on the concept of ‘good administration’ as the foundation for the doctrine of legitimate expectation. He describes good administration as ‘a concept whose apparent capacity to support what is, in reality, a catholic body of doctrine is attributable to nothing more than its vacuity’. He complains of

‘an overarching problem that besets this area of administrative law - namely, an unfortunate judicial tendency to seek to avoid difficult doctrinal and normative questions by sheltering behind superficially attractive but ultimately rather empty notions such as "good administration" and "fairness". Such language may be intuitively appealing, but it is incapable of doing the sort of analytical heavy-lifting that is required if the law in this area is to be placed on an intellectually cogent footing that lends itself to coherent doctrinal development.’

I am tempted to agree that this is a good example of the gap between theory and practice about which Patrick Atiyah was writing all those years ago.

Legality

So can we find an area which not only says that it is principled but actually is more principled than pragmatic? Could the doctrine of legality be such an area?

This is a rule of statutory construction. General words in an Act of Parliament will not be read so as to permit an intrusion into fundamental rights. Parliament can, of course, legislate to remove or restrict fundamental rights but it has to do so in clear and express words so that the Parliamentarians can understand what they are doing and be prepared to take political responsibility for doing it.

A recent example is the case brought by the trade union, UNISON, challenging the Order in Council imposing fees for bringing employment tribunal claims. The Supreme Court held that the general words in the Act empowering the Lord Chancellor to prescribe tribunal fees were not clear enough to authorize setting fees at such a high rate as to make it impossible, impracticable or irrational to bring a claim. As the court explained, the purposes of the Fees Order were legitimate – making resources available to the justice system, thus securing access to justice, and deterring frivolous or vexatious claims, thus increasing its efficiency. But that did not permit the Lord Chancellor to prescribe whatever fees he chose if there was a real risk that people would be effectively prevented from having access to justice. The Supreme Court had the benefit of evidence about the impact of the fees which had not been available in the lower courts. Further, as Lord Reed explained, ‘even where primary legislation authorises the imposition of an intrusion on the right of access to justice, it is presumed to be subject to an implied limitation. As it was put by Lord Bingham in Daly, the degree of intrusion must not be greater than is justified by the objectives the measure is intended to serve.’ (para 88)

The same principle was invoked in the Supreme Court’s decision in Evans, the case in which a Guardian journalist had made a request under the Freedom of Information Act 2000 for disclosure of correspondence passing between the Prince of Wales and various government departments over a few months in 2004-5. The Information Tribunal, chaired by a High Court Judge, heard evidence over several days, including the evidence of two Professors of Constitutional Law about the Conventions governing the relationship between the heir to the throne and the government. In a lengthy and carefully reasoned judgment, it concluded that the public interest in disclosure of many of the letters outweighed the various reasons given in the Act for refusing it. The Tribunal had, of course, read the letters.

The Act contains a power, in section 53, for Ministers to override the decision of the Tribunal and to veto disclosure if ‘on reasonable grounds’ they consider it not to be in the public interest. Rather than exercising the right to appeal against the Tribunal’s decision, the Attorney-General exercised the override power. Mr Evans then brought judicial review proceedings to challenge the decision of the Attorney-General. He failed in the Divisional Court but succeeded in the Court of Appeal and, by a majority of five to two, in the Supreme Court. None of the judges hearing the judicial review proceedings had seen the letters. They were simply ruling on the lawfulness of the Attorney-General’s decision.

Lord Neuberger, with whom Lord Kerr and Lord Reed agreed, said this:

‘A statutory provision which entitles a member of the executive...to overrule a decision of the judiciary merely because he does not agree with it...would cut across two constitutional principles which are also fundamental components of the rule of law. First... a decision of a court is binding as between the parties and cannot be ignored or set aside by anyone... Secondly, . . . decisions and actions of the executive are...reviewable by the court at the suit of an interested citizen. Section 53, as interpreted by the Attorney General’s argument in this case, flouts the first principle and stands the second principle on its head’ (paras 51, 52).

In his recent Reith lectures, Lord Sumption has characterized this reasoning as saying that the government’s power to override the judicial decisions of the Information Tribunal is such a bad idea that Parliament cannot have meant it.

I think that this is rather different from the reasoning of Lord Mance. He stated that any test must be ‘context-specific’, ‘in the sense that it must depend upon the particular legislation...and upon the basis on which the Attorney-General was departing from the decision’. It was clear that the Attorney-General had to show that he had reasonable grounds for refusing disclosure, which was a higher test than mere rationality. In considering what is reasonable one must consider the factual investigation by the Tribunal and the extent to which the Attorney-General can replicate that; effectively, if a Tribunal is better equipped to make a decision, then a Minister or the Attorney would need solidly reasoned grounds for issuing a certificate. I agreed with Lord Mance. While our approach might be seen as a pragmatic compromise, I prefer to think of it as a principled interpretation of what the section actually says, in the light of the principle of legality, which mandates a minimal intrusion upon fundamental rights.

Lord Sumption agrees with Lord Hughes, the dissenting Justice who said that ‘the rule of law is not the same as the rule that courts must always prevail no matter what the statute says’ (para 152). I agree entirely with that.

The point I am making is that in cases such as these, the courts have been prepared to construe Acts of Parliament in the light of the principle of legality without a hint of deference or pragmatism, indeed some might say quite the reverse. In the first substantive case to be heard by the Supreme Court, Ahmed v Her Majesty’s Treasury, the court held that the very general words in the United Nations Act 1946, allowing the making by Orders in Council of such provision as appeared necessary or expedient to carry out the decisions of the Security Council, did not permit the Treasury to make Orders in Council permitting it to freeze the assets of people blacklisted by the Security Council without any sort of due process. This was a classic application of the principle of legality. Furthermore, the court refused the Treasury’s application to suspend its order. Although it did have power to suspend the effect of any order, suspending an order declaring something to be ultra vires and quashing it did not alter the fact that it was void and of no effect. The court should not lend itself to something which might obfuscate the effect of its judgment. Not a hint of pragmatism there.

Conclusion?

I have convinced myself that, after all, there is a great deal of pragmatism in public law. Pragmatism – ‘a pragmatic view about the evidential value of certain judgments of the executive’ – is a more convincing rationale for deference than democratic legitimacy. Essentially pragmatic concepts of good administration and fairness are used to explain the workings-out of the doctrine of legitimate expectation. And I expect that any competent academic public lawyer could even find pragmatic considerations creeping into our application of the principle of legality. The trouble is that real judges have to make judgments in real cases involving real people and it is unrealistic to expect complete doctrinal coherence of them – though I suspect that many people do.