The nature of constitutional convention
Published online by Cambridge University Press: 02 January 2018
The term ‘constitutional convention’ tends to be attached indiscriminately to any regularity of conduct that is observed in the process of government. It is here argued, however, that constitutional conventions as properly understood are social rules which govern the relations between political parties or the institutions of government. Furthermore, this term is to be confined only to those social rules that are constitutional in character. Intra- party standards of conduct, in particular, have wrongly been accorded the status of constitutional convention. These considerations, together with a detailed examination of the precise status of constitutional conventions as part of the ‘morality’ of the constitution, lead to a re-assessment of the conditions under which constitutional conventions both come into existence and cease to exist.
- Research Article
- Copyright © Society of Legal Scholars 1999
1. G Marshall Constitutional Conventions: The Rules and Forms of Political Accountability (Oxford: Clarendon Press, 1984, reprinted 1993).
2. L J M Cooray Conventions, The Australian Constitution and the Future (Sydney: Legal Books, 1979).
3. A Heard Canadian Constitutional Conventions: The Marriage of Law and Politics (Toronto: Oxford University Press, 1991).
4. A V Dicey Introduction to the Study of the Law of the Constitution (London: Macmillan 10th edn, 1959) p 24.
5. Above n 4.
6. A Heard ‘Recognising the Variety among Constitutional Conventions’ (1989) 22 Can J Political Science 63. This article is substantially reproduced as chapter 7 of Heard's book, above n 3, and it is in the latter format that reference will be made to Heard's work.
7. Council of Civil Service Unions v Minister for the Civil Service (1985) AC 314.
8. T R S Allan Law, Liberty, and Justice: The Legal Foundations of British Constitutionalism (Oxford: Clarendon Press, 1993) p 242.
9. I Jennings The Law and the Constitution (London: University of London Press 5th edn. 1959), pp 134–135 (’ Mere Practice is Not Enough’). Of course, this criticism proceeds on the basis of a common meaning attached to the word ‘practice’ . It is noteworthy, nevertheless, that Jennings ignores his own strictures. He cites (at pp 119–120) the provision of the Irish Free State Constitution Act 1922, which reproduces art 2 of the Anglo-Irish Treaty of December 1921 that ‘the position of the Irish Free State in relation to the Imperial Parliament [should be that of] the law, practice and constitutional usage governing the relationship of the Crown…to the Dominion of Canada…’ (emphasis added). He then proceeds to draw the inference: ‘If the question had ever arisen, the courts would have been called upon to say what were the constitutional conventions [sic] governing the relations between Canada and the United Kingdom.’
10. Allan, it should be emphasised, does record (at p 242. n 25) the reservation expressed by Hood Phillips that, perhaps, informal arrangements governing the internal regulation of the civil service are not matters of constitutional convention. However, there is a third criticism that may be made of Allan's argument. Even if the previous practice that was in point in the GCHQ case is appropriately categorised as a constitutional convention, it must be applied by the courts as the uniquely correct standard before we may accurately speak of the legal enforcement of such a convention. For there is a fundamental difference between a court enforcing a rule because it is legally binding and (on the other hand) its choosing to create a rule the contents of which are drawn from convention. It is often the case in many diverse areas of the law that the conventionally accepted practice is simply adopted by the courts, among a range of possible solutions, as supplying the standard to be required by the law. See, for example, H Collins Justice in Dismissal (Oxford: Clarendon Press, 1992) pp 74–78.
11. Marshall Constitutional Conventions, above n 1, p 3.
12. This is the form of words used that is used in regard to the powers of the State President under s 16(3) of the Republic of South Africa's Constitution Act 1961. See also s 7(5) of the same instrument: ‘The constitutional conventions which existed immediately prior to the commencement of this Act shall not be affected by the provisions of this Act.’
13. The same point may be made of several allied concepts, eg ‘constitutional practice’, ‘constitutional usage’ . See above n 9, and s 9(2) of the Statute of Westminster 1931, which stipulates that nothing in the statute shall be deemed to require the concurrence of the Parliament or Government of Australia in any law made at Westminster with respect to certain areas of subject matter ‘…in any case where it would have been in accordance with the constitutional practice existing before the commencement of this Act that the Parliament of the United Kingdom should make that law without such concurrence’ (emphasis added).
14. The most prominent example derives from Australia. See C J G Sampford “Recognise and Declare”: An Australian Experiment in Codifying Conventions (1987) 7 OJLS 369. Three types of difficulty may be seen to be associated with such exercises in codification. First, as a guide to the conduct that is required in particular situations as they arise, abstract statements of conventional practice tend to be unnecessary in clear cases and not especially helpful in difficult ones. Secondly, they will tend to be less accurate over time unless a conscious effort is made to revise them periodically. Thirdly, it can happen that, even when freshly drafted, they are incomplete. For example, the publication Questions of Procedure for Ministers (as to which see below) was explicit on the impropriety of gifts being received by ministers, but was utterly silent on the subject of ministers making gifts to other public figures - the situation that led to the resignation in 1993 of Mr Michael Mates as Minister of State for Defence in the wake of disclosures that he had made gifts to a prominent disgraced businessman. In circumstances such as these the view may well be taken - as it is of many a contract that has purportedly been reduced to writing - that the written document is incomplete.
15. (London: Hutchinson, 1975). See especially ch 1 (’ On Reasons for Action’).
16. Jennings, above n 9, p 136.
17. (Harvard University Press, 1969).
18. Dworkin, Law's Empire (London: Fontana, 1986), p 145.
19. (Oxford: Clarendon Press, 1961).
20. S E Finer ‘The Individual Responsibility of Ministers’ (1956) 34 Public Administration 377. The article is conveniently reproduced in G Marshall (ed) Ministerial Responsibility (Oxford: Oxford University Press, 1989) ch 16.
21. Marshall (ed) Ministerial Responsibility, above n 20, pp 125–126.
22. It is interesting to note the landmarks in the development of the two-term convention during the course of the nineteenth century. The precedent to which Americans harked back was that of George Washington's refusal to run for a third term - a refusal that was actuated by entirely personal reasons. Similar refusals by Jefferson and Jackson appear to have been motivated by considerations of constitutional principle. Jackson, in fact, was in favour of a constitutional amendment limiting the maximum possible period of office to a single term of between four and six years. A mixture of political and constitutional considerations combined to deny General Grant the Republican nomination for a third term in 1880. And in 1896 the convention appears to have been accepted by the Democratic Party as sufficient reason in itself to deny the nomination for a third term to undoubtedly their best candidate, President Cleveland.
23. See. for example, C R Munro Studies in Constitutional Law (London: Butterworths, 1987) p 51.
24. The United States was not in a state of war until December 1941.
25. Incidentally, this example provides a good illustration of the advantages of leaving the matter to constitutional convention. Legal draftsmanship is seemingly incapable of capturing concepts such as ‘imminence of war’ in contrast to a state of war itself.
26. In the unlikely colours, on that occasion, of the ‘Bull Moose’ Party.
27. As, for example, in E Barendt Broadcasting Law (Oxford: Clarendon Press, 1993) p v: ‘…broadcasting freedom in Britain only exists under weak conventions, which can easily be disregarded by confident (or arrogant) governments.’
28. According to most, though by no means all, jurisprudential theories.
29. Clearly, the passage in Barendt, above n 27, is not referring to the extent of the area over which the broadcasting conventions hold sway. It is to be understood either as dubbing all conventions as intrinsically weak or as designating the broadcasting conventions as weak members of a set, other members of which may be strong.
30. R Dworkin Taking Rights Seriously (London: Duckworth, 1977): see, especially, ch 3.
31. Above, n 30, p 54.
32. Above, n 18.
33. See, generally, D L Ellis ‘Collective Ministerial Responsibility and Collective Solidarity’ in Marshall (ed) Ministerial Responsibility (above n 20) pp 54–56.
34. The figure given by Ellis, above n 33, p 55.
35. The distinction between ‘party political’ and ‘governmental’ functions is drawn quite easily for such routine purposes as the use of official transport.
36. See below n 41 for the relevant authorities.
37. The point is often put a different way whenever any such incident becomes public knowledge. ‘Surely’, it is asked, ‘the matter is of no genuine concern to anyone other than the Minister and his family?’ The point of view that is expressed in the text, however, does not necessarily go this far. Whilst denying the governmental element (and hence the constitutional standing) of the sanction of resignation in such circumstances, it leaves open the possibility that the minister's indiscretions might be a matter of legitimate concern to his party.
38. There are a number of difficulties in maintaining the view that such misconduct does have a constitutional dimension. If, as has been argued, they do possess such a dimension since they raise the issue of fitness to hold office, individual misconduct that took place before attaining ministerial rank would be as much in point as that which occurs during the minister's time in office. Moreover, the claim that extra-marital affairs are of constitutional significance is difficult to reconcile with the career of Mr Alan Clark, who has openly admitted to several such affairs without provoking a clamour that he should resign from the several junior ministerial positions that he has held. Can this instance be explained by saying that it is the deception (rather than the betrayal) of one's spouse that forms the kernel of the constitutional convention? Or, rather, is it the embarrassment of being ‘found out’ that is the critical issue? If the latter, the cases of resignation in the wake of sexual misconduct are explained better by the need to minimise the discomfiture of the minister's political colleagues.
39. The view is expressed in R Brazier ‘The Downfall of Margaret Thatcher’ (1991) 54 MLR 47 1 at 475, 480.
40. Did convention require, further, that Ministers vote for either Hurd or Major once Mrs Thatcher was out of the race? That would seem a logical inference from the argument. Furthermore, a challenge to the incumbent Prime Minister from a fellow member of the Cabinet may spring from disagreement over style of leadership rather than policy differences. Surely, the former has nothing to do with the doctrine of collective Cabinet responsibility.
41. The only authorities in English law are Re Grant's Will Trusts (1980) 1 WLR 360 and Conservative and Unionist Central Office v Burrell (Inspector of Taxes) (1982) 1 WLR 522, which concerned the legal position of money left to the Labour Party (in the former case) and accruing to the Conservative Party (in the latter). In the case of the Conservative Party, the Court of Appeal ruled that the organisation was too loose to be characterised as an unincorporated association. See especially at 527, per Lawton LJ: ‘Agreements which confer rights and impose obligations, as membership of unincorporated associations do [sic], must be reasonably certain because they may become justiciable… No member of a local constituency association, basing his claim on contractual rights, could ask the court to protect those rights in respect of the parliamentary party's election of a leader…’ This passage, however, does not rule out the possibility that more centrally placed individuals, for example Conservative MPs, might be in a position to invoke the protection of the court in this respect. For authorities on the legal status of Canadian political parties, see Heard, above n 6, p 171, n 38.
42. Marshall Constitutional Conventions, above n 1, p 4.
43. Set out in Marshall (ed) Ministerial Responsibility, above n 20, ch 8. The summary of conclusions and recommendations is contained in Appendix 2 of the report: Cmnd 6386 (1976).
44. Para 1: see Marshall (ed) Ministerial Responsibility, above n 20, p 68.
45. Para 2: see Marshall (ed) Ministerial Responsibility, above n 20, p 68.
46. By ‘alegal’ we mean neither legal nor illegal, but simply having no reference to law or legal procedures.
47. Subject, of course, to their satisfying the general criteria of constitutional conventions.
48. Lewis, above n 17, ch 3.
49. Ibid, p 84.
50. Ibid, p 88.
51. A case in point is the present procedure for the appointment of archbishops and diocesan bishops in the Church of England. The new procedure was ‘promulgated’ by the then Prime Minister, Mr Callaghan, in a written reply to a question placed by the Leader of the Opposition, Mrs Thatcher. The text of the statement is at 912 HC Debs Written Answers, cols 612–614 (8 June 1976).
52. As, for example, in the resolutions passed at the various Colonial Conferences and Imperial Conferences of the late nineteenth and early twentieth centuries: eg the resolution of the Imperial Conference of 1930 that Dominion ministers, rather than their British counterparts, would thenceforth advise the Crown on appointments to the post of Governor General.
53. None of the ways that were proposed succeeded in finding favour with the conference. In fact, it was on the problems of how to proceed on the Home Rule Bill that the conference eventually foundered. See J Jaconelli ‘The Parliament Bill 1910-1911: The Mechanics of Constitutional Protection’ (1991) 10 Parliamentary History 277 at 285–286.
54. Any such convention that might have emerged would have been described more aptly as ‘meta-constitutional’, in so far as it would regulate the mode of proceeding on measures, such as the Government of Ireland Bill, which are themselves constitutional in nature. The same description could also be given to the convention that the Committee stages of any ‘constitutional bill’ are to be taken on the floor of the House of Commons.
55. Heard, above n 6, p 11.
56. To emphasise once again the point that was developed in section 5 above, the agreement must be, in any event, ‘constitutional’ in nature. Compare the sort of agreement on specific policies that is exemplified by the ‘Lib-Lab’ pact of the late 1970s. See D Steel A House Divided: The Lib-Lab Pact and the Future of British Politics (London: Weidenfeld and Nicholson, 1980) pp 36–37, for the text of the first draft of the pact.
57. Analogous issues have arisen in regard to another device of private law, the trust. Certain governmental arrangements come to be described loosely in terms of trusteeship. It was used, for example, to characterise British policy towards the colonies between the two World Wars. However, such attempts as have been made to extrapolate ‘a true trust’ enforceable by the courts - as opposed to ‘a trust “in the higher sense”, or a governmental obligation’ - have not been successful. See Tito v Waddell (No 2) (1977) Ch 106 (from which this terminology is drawn (at p 222) and Guerin v The Queen (1983) 143 DLR (3d) 416.
58. R Brazier ‘The Non-Legal Constitution: Thoughts on Convention, Practice and Principle’ (1992) 43 NILQ 262 at 268.
59. Dicey, above n 4, p 24.
60. H L A Hart Law, Liberty and Morality (Oxford: Oxford University Press, 1963) p 20.
61. Marshall Constitutional Conventions, above n 1, pp 11–12.
62. Ibid, p 12.
63. Indeed, to appraise social rules in the light of critical morality (as that term is used by Hart) is, in some respects, a futile exercise. Consider a typical social rule, that men are to remove their hats in church. One may sensibly ask whether people should show respect on entering a place of religious worship. The form that a show of respect will take, however, will be specific to the culture in question.
64. Lewis, above n 17, and accompanying text.
65. Above n 8 and accompanying text.
66. Dworkin, above n 30, p 53.
67. P P Craig Public Law and Democracy in the United Kingdom and the United States of America (Oxford: Clarendon Press, 1990) pp 210–211, furnishes a good example of the way in which certain writers are prone to assume that the breach of (or change in) constitutional convention is, in itself, objectionable from a critical moral viewpoint. He cites a passage in R Holme and M Elliott (eds) 1688-1988: Time fora New Constitution (London: Macmillan, 1988), in which the writer denounces the abolition of the Greater London Council and the metropolitan local authorities as exhibiting scant regard for the conventions governing the distribution of power between central and local government. Even if the abolition was in breach of constitutional convention, it does not follow that the new position is constitutionally improper. The latter conclusion, as Craig says, ‘requires a normative analysis, not just simple assertion’.
68. Cabinet Office, July 1997.
69. Ibid, para 24.
70. Ibid, para 86.
71. Ibid, para 126.
72. Ibid, para 62