Carol Harlow: How not to do things with rules

For many years now, administrative lawyers have been puzzling over the relationship of rules and discretion. When is a rule a rule? Does the term properly cover ‘quasi-legislation’ or what is now generally called ‘soft law’? In his seminal study, Discretionary Justice (Urbana, Louisiana State University Press, 1969), KC Davis argued that rules of this type should be used to structure and confine administrative discretion. The consequence would be greater transparency; light would be cast on some ‘dark and windowless areas of administrative law’. The managerial style of public administration since the 1980s has taken the technique very much further. Public administration is now dominated by various forms of rule, a development hastened by the greater use of ICT by government departments. Policy statements, government papers, guidance and other information is now widely accessible on government websites and what is not openly published is often made available through requests made under the Freedom of Information Act 2000.

Generally benign, these developments have brought their own problems to which the over-burdened and generally unloved immigration service has been particularly prone. It is not so long since a previous Home Secretary declared his department ‘unfit for purpose’ and immigration case law regularly reveals scenes of great confusion, with policy changes introduced so rapidly that even officials cannot keep up (see eg R (Rashid) v Home Secretary [2005] EWCA Civ 744). The recent Supreme Court case of R (Munir and Rahman) v Home Secretary [2012] UKSC 32 casts further light ­– if this were needed – on the somewhat arbitrary way in which ministers of all political persuasions tinker with immigration policy. In 1996, the Home Office issued a Policy Document (DP 5/96) on deportation policy in cases involving children born in the UK and resident for seven or more years (the so-called 7-year concession). In practice, concession effectively glossed a later policy document contained in departmental ‘Instructions’ (the ‘Long Residence Concession’) which made no mention of the 7-year concession. In 1999, the Home Office ‘modified’ DP5/96 in a ministerial statement that significantly stated (i) that each case must be considered on its merits subject to a ‘general presumption’ that the 7-year concession would be observed; and (ii) that ‘certain factors’ should be considered as ‘relevant to reaching a judgment’ in such cases. In 2003, the Long Residence Concession was modified in a Statement of Changes to the Immigration Rules duly laid before Parliament (HC 538, 2003), which actually made no direct mention of the concession ­– never formally withdrawn, though it was at some point apparently taken off the departmental website. Suddenly, in 2008, however, the Minister for Borders and Immigration in a written parliamentary statement withdrew DP 5/96. M and R, who had hoped to benefit from the 7-year concession, challenged the withdrawal by judicial review, arguing that it was irrational and unfair and also invalid because the proper parliamentary procedure had not been followed.

Before setting out the argument in greater detail, it is necessary to feed in a second case decided by the same court (Lords Hope, Walker, Clarke, Dyson and Wilson JJSC) on the same day. R (Alvi) v Home Secretary [2012] UKSC 33 concerned the points-based system of entry for non-EEA nationals who wish to work in the United Kingdom, introduced in 2008 by a Statement of Changes in Immigration Rules duly laid before Parliament (HC 1113, 2008). So far, so good. But HC 1113 made reference both to a resident labour test as defined in ‘guidance’ published by the United Kingdom Border Agency (UKBA) and to a ‘list of skilled occupations’ to be maintained by UKBA. A, who had entered lawfully as a student but stayed on illegally, was in 2009 refused leave to remain on the ground that his job was not so listed and that his salary fell below the stipulated requirement for the policy. A challenged this decision on the ground that the list was contained in ‘Occupational Codes of Practice’ published only on the departmental website; it had never been laid, as it should have been, before Parliament.

These cases raised several tricky questions concerning the nature and quality of ‘bureaucratic rules’. The status of the Immigration Rules has always been a puzzle. They derive from the Immigration Act 1971, which for the first time imposed statutory controls on Commonwealth citizens. Section 3(2) provides that the Home Secretary

shall from time to time… lay before Parliament statements of the rules or any changes in the rules, laid down by him as to the practice to be followed in the administration of this Act for regulating the entry into and stay in the United Kingdom of persons required by this Act to have leave to enter…

In both our cases Statements of Changes had been made and laid and their validity was not in issue. The question was whether policy changes introduced through informal rule-making were valid, or if they too ought to have been laid.  Taken together the two cases raised three central questions:

  • What is the legal status of the Immigration Rules?
  • Can the Rules be changed or modified by ‘soft law’?
  • Is it open to the Home Secretary fall back on the prerogative powers?

The prerogative power

This question got a brusque reception! The only mention of prerogative in the 1971 Act was a saving clause (s 33(5)), which provided that the Act should not be taken to ‘supersede or impair’ the prerogative as applicable to aliens. Unanimously, in a single judgment delivered by Lord Dyson, the Supreme Court ruled (i) that the prerogative powers had never extended to Commonwealth citizens and (ii) that Parliament had in any event intended the 1971 Act to be the sole source of powers to regulate immigration (Munir at [23]-[26]). This simple application of the principle in De Keyser’s case (A-G v De Keyser’s Royal Hotels [1920] AC 508) will come as no surprise to constitutional lawyers, who have watched the courts in recent years ‘structure and confine’ the ultimate discretion of the royal prerogative almost out of existence. It is only in exceptional cases, involving defence and security or foreign affairs – as in the recent case of the unfortunate Chagos Islanders, expelled from the islands by an Order in Council made under the prerogative – that an argument based on prerogative powers will receive a sympathetic hearing – and even then the courts are hard to persuade (see, eg, the judgment of the High Court in R (Bancoult) v Foreign Secretary [2006] EWHC 1038, overruled by the House of Lords at [2008] UKHL 61). In our rule-based administration, such arguments will become increasingly rare.

The Immigration Rules (IRR)

If the power to make the IRR derives from statute, then the Immigration Act was in more ways than one ambiguous. It contained no specific grant of power to make regulations, and the Supreme Court was driven to imply one from a reference in a Schedule to ‘Instructions’. Both a power, exercisable in terms of s 3(2), and a duty to make regulations, vested in the Secretary of State, was however confirmed (Munir at [27]). Lord Dyson went on to dispose shortly of the thorny question of the legal status of the IRR: these were not, as the House of Lords had opined in MO (Nigeria) v Home Secretary [2009] ULHL 25, an executive statement of policy; they were subordinate legislation made under the authority of the 1971 Act. The IRR were in short ‘hard law’ that had to be laid before Parliament. This decisive ruling should end the uncertainty and unease registered by Sedley LJ in Pankina v Home Secretary [2010] EWCA Civ 719 at ‘rules being elevated to a status akin to law’.

Changes of policy       

The ground was now cleared for the crunch question: when was a rule a ‘statement of practice’ that needed to be laid in terms of s 3(1)? In Munir, the Court was able effectively to side-step this question since, if withdrawal of DP 5/96 was, as argued, unlawful because of failure to follow the requisite parliamentary procedure, it must follow that DP 5/96 was itself invalid for the same reason. In any event, DP5/96 was not a ‘statement of practice’ because it incorporated discretion: each case had to be considered on its merits and the factors ‘relevant to reaching a judgment’ might or might not be applied. Judged by the rule-of-thumb established by Lord Dyson, that the less the flexibility inherent in the concessionary policy, the more likely it was to fall within s 3(2), DP5/96 was merely a ‘concessionary policy statement’. Perhaps the only surprise here lies in the total absence of any mention of the doctrine of legitimate expectation.

In Alvi, it was harder to resolve the crunch question and, although there were no dissenting judgments, the Supreme Court Justices were not so tidily at one. There was general agreement that the points system did amount to a statement of practice within s 3(2) and did need to be laid. Both the ‘guidance’ and the ‘List of Skilled Occupations’ posted on the departmental website created requirements which the immigrant ‘must’ satisfy; they were in short, determinative of rights and must be laid before Parliament.

These nice distinctions (which surely fall within Humpty Dumpty’s famous observation that words mean what the speaker wants them to mean) leave much room for jurisprudential analysis. What are we to make, for example, of Lord Dyson’s conclusion in Munir that a ‘concessionary policy statement’ is not a ‘statement of practice’? Of more interest, however, is where the boundary between rules and discretion lies. This is, after all, a question that administrative lawyers have been discussing at least since KC Davis wrote, with many contributions from distinguished jurists, such as Ronald Dworkin’s ‘hole in the donut’ metaphor. It is fair to say that there is much academic support for the idea of rules and discretion as points on a spectrum or sliding-scale, though needless to say, the academic debate (summarised in ch 5 of C Harlow and R Rawlings, Law and Administration, Cambridge University Press, 2009) found no place in the two Supreme Court decisions. Indeed Lord Dyson in Alvi simply brushed aside the notion of a spectrum as advanced by Sullivan LJ in R (Joint Council for the Welfare of Immigrants) v Home Secretary [2010] EWHC 3524 (Admin) in favour of a clear binary distinction between ‘substantive’ and ‘procedural/evidential’ requirements – a distinction that he immediately rejected as not providing a ‘satisfactory basis for deciding what is and what is not a rule within the meaning of s 3(2)’. Lord Dyson’s solution was that ‘a rule is any requirement which a migrant must satisfy as a condition of being leave to enter or leave to remain’ (emphasis mine). This ‘solution’ reprises the well-known problem of distinguishing ‘must’ from ‘may’ traversed many years ago in Padfield v Ministry of Agriculture, Fisheries and Food [1968] AC 997 and rather suggests that a rule is a rule when Humpty Dumpty (a.k.a. the Home Office, UKBA or, indeed, the Supreme Court) chooses to treat it as mandatory.

What is the outcome for immigration law? It is now clear that the IRR are subordinate legislation and that any change to them involving mandatory requirements or conditions will have to be laid. This is a conclusion reached by Lord Wilson ‘without enthusiasm’; he rightly saw that it would create ‘an astonishingly prescriptive system’ (Alvi [128]). Lord Hope in the leading judgment, expressed similar concern (Alvi [65]) over the burden on Parliament, falling for the most part on the House of Lords Committee on the Scrutiny of Secondary Legislation (formerly the ‘Merits Committee’), if the process of laying were not to become a mere procedural formality. Equally, there would be a burden on the courts from the ‘rapid succession of cases’ and new opportunities afforded for challenge (Alvi [54]). Against these negative impacts, however, the dangers of allowing the Home Secretary unfettered discretion to change the rules must be weighed.

How can this rigid and inflexible system, brought about through a laudable desire for legal certainty and administrative consistency, be ameliorated or, better still, evaded? Has the Supreme Court opened an escape hatch in Munir? Provided the draftsman is careful to scatter the magic words ‘may’ liberally throughout the text, perhaps substituting ‘advice’ for ‘guidance’, statements of policy change may ’scape laying. How immigration officials will read the rules is, of course, another question altogether!

Carol Harlow is Emeritus Professor of Law at the London School of Economics

Suggested citation: C. Harlow, ‘How not to do things with rules’ UK Const. L. Blog (15 September 2012) (available at