Canadian Parliamentary Review

Current Issue
Canadian Region CPA
Upcoming Issue
Editorial and Stylistic Guidelines

HomeContact UsFranηais

Government Defeats in the House of Commons: The British Experience
Philip Norton

At the time this article was published Philip Norton was Reader in Politics at Hull University in Great Britain. This is a revised version of his submission to the Special Committee on Reform of the House of Commons in May 1985.

The belief that a government defeat necessarily entails resignation or a dissolution has influenced significantly parliamentary behaviour. It has served to reinforce party cohesion in the division lobbies. Government backbenchers who disagree with a particular government policy but who have nonetheless voted for it have on occasion justified their voting behaviour on the grounds that they "do not wish to see the other side in office". A number of Conservative backbenchers were still proffering this explanation in the early 1970s, even after a number of government defeats, none of which had propelled the government into resigning or requesting a dissolution. It is, then, clearly a largely ingrained and powerfully held perception of the government's required response, one which a number of observers may have been inclined to elevate to the status of a convention.

It is, however, a perception which rests upon no continuous basis of practice or upon any authoritative original source and one which is belied by the experience of both the nineteenth and twentieth centuries. In the twentieth century alone, the British Government has suffered more than one hundred defeats in the division lobbies in the House of Commons. In response to only three of these defeats (two in 1924, one in 1979) did the government resign or request a dissolution. Indeed, on most of the remaining occasions, the question of confidence was not seriously considered to have been raised. When the governments of the l970s suffered a spate of defeats in the division lobbies (sixty-five in the seven-year period between April 1972 and April 1979) a number of commentators were of the opinion that: there had been a change in constitutional practice. They claimed the government of the day was deviating from past practice in continuing in office, accepting most of the defeats imposed upon it, and declining to resign or request a dissolution unless defeated on an express vote of confidence. In fact, clear precedents can be found for the government's response to each defeat. Far from there being a deviation from past practice, there was a clear continuation of it.

What, then, is the constitutional position of the government in the event of a defeat in the division lobbies of the House of Commons? An analysis of the defeats of the past century suggests that there are essentially three types of defeat, each of which invites a different response from government.1

Defeats on Votes of Confidence

A government defeated on a vote of confidence is expected to resign or seek a dissolution of Parliament. The precedent was established in 1841 and has been maintained since. What, though, constitutes a vote of confidence? Three distinct types of vote fall within this rubric:

(1) Explicitly worded votes of confidence. These state in express terms that the House has or has not confidence in Her Majesty's Government. The carrying of a motion of no confidence against the Government is the strongest and least ambiguous of votes involving the question of confidence. The passage of such a motion puts it beyond doubt that the government has lost the confidence of the House there are no other issues to cloud the picture and that a visit by the Prime Minister to Buckingham Palace is required, there to request a dissolution of Parliament or to tender the resignation of the Government.

(2) Motions made votes of confidence by the declaration of the government. The government, usually at the instigation of the Prime Minister, may declare that if defeated on a particular motion before the House, even though not an explicitly-worded one of confidence, it will resign or seek a dissolution. It may do so because the issue in question is so central to its policy that there would be little point in continuing in office if defeated upon it; or it may do so simply as a political expedient in order to maximise its voting support. It may even do so, in the face of likely defeat, because it has lost the political will to continue. Though defeats in such instances are not quite as straightforward as those taking place on explicit votes of confidence the picture is complicated by the issue which ostensibly is covered by the motion the position of the government remains the same. Having lost on a motion which by government declaration is one of confidence, the Prime Minister must make the trip to the Palace.

(3) Implicit votes of confidence. Traditionally, certain matters have been deemed to involve confidence, even though not declared to do so by the prior statement of government. Falling within this category primarily is the granting of Supply. Failure to grant Supply is regarded as the established means by which the House can demonstrate its lack of confidence in the ministry. However, it is a category of vote which has largely fallen into desuetude. Given the number of divisions which occur annually on items of Supply, and the fact that the Second Reading debates on Consolidated Fund Bills ceased long ago to be substantive debates on Supply, the opportunity for the House to debate and then refuse Supply in one single division hardly exists; it might be possible on Second Reading of the Finance and Appropriation Bills, but even that is not certain. A defeat on a specific Estimate would not in itself constitute a vote of no confidence; Prime Minister Balfour refused in 1905 to treat such a defeat as a vote of censure. For the House to demonstrate its lack of confidence in the government by denying it Supply would necessitate now a series of defeats. Such action would seem perverse given that the desired outcome could be achieved by one explicit vote of no confidence. Only in the unlikely (and unprecedented) event of a government seeking to ignore a defeat on a explicit motion of confidence would it be likely that the House would resort to denying Supply through a series of negative votes.

Defeats on Items Central to Government Policy

If defeated on an item regarded as central to its policy, but one not made an issue of confidence by prior declaration, the government may either seek an explicit vote of confidence from the House or resign or request a dissolution. The effect of deciding on the latter course (resigning or requesting a dissolution) would be to make the division retrospectively one of confidence. It was a course followed on occasion in the nineteenth century but not one which has found favour with governments in the twentieth. Not surprisingly, not wishing to jeopardise their tenure of office, Prime Ministers have preferred to ask the House subsequently to declare its confidence in the government rather than assuming that such confidence has been denied in the vote in question. In March 1976, for example, the Labour Government was defeated on its Expenditure White Paper. Though one member of the Cabinet raised the possibility of seeking a dissolution, the Cabinet opted instead to seek a vote of confidence from the House. Such a vote was forthcoming.

A number of divisions in each Parliament and one or two automatically each session may be deemed to be on items central to government policy; notably the Second Reading of major Bills promised in the Government's election manifesto and, sessionally, the Loyal Address and (as indicated above) the Second Reading of the Finance Bill. Defeat on a Second or Third Reading per se would not automatically fall within this category. The two defeats in this century on Second Readings (in 1924 and 1977) and the one on a Third Reading (also in 1977) were not treated as falling within it, either by government or opposition. To qualify for inclusion the motion must be recognized as central to government policy by the government itself or, failing that, by the Opposition. The number of divisions that qualify, be it relative to the total number of divisions or in absolute terms, is a small one.

Defeats on Items Not Central to Government Policy

Divisions which take place on items not at the heart of government policy are the most numerous in a Parliament. A Second Reading defeat on an important Bill may fall within the second category of defeat above; defeat in one or more of the multiple divisions which take place during Committee and Report stages usually would not. Defeats in this category do not raise questions beyond the confines of the issue on which they have taken place. The government's defeat in 1972, for example, on an amendment to the Local Government Bill – on the issue of which local councils should have the power to incinerate refuse – can hardly be said to have raised the question of confidence. Indeed, it was a defeat little noticed by either MPs themselves or by the media. In response to such defeats all that the government need do is decide whether to accept the defeat or to seek its de facto reversal at a later stage.

Of the government defeats in the division lobbies in this century, all but a handful have fallen in this category. They have included (most frequently) defeats on amendments to Bills, including Finance Bills, as well as on prayers to annul orders, Ways and Means Resolutions, on motions for Mr. Speaker to leave the Chair, on adjournment motions, on various procedural motions, and on a number of declaratory motions, including Supply motions. None has been considered to strike at the heart of government policy but a number have not been insignificant either in political or economic terms. They have, for example, included a change in the standard rate of income tax, a devaluation of the European Green Pound, and the loss (following defeat on a guillotine motion) of the Labour Government's main measure of constitutional reform the Scotland and Wales Bill – in 1977. In the majority of cases the government of the day has accepted the defeats. (All defeats on Finance Bills, for instance, have been accepted). In only a minority of cases has the government sought to obtain later reversal. Even then it has not always been successful: the House refused to reverse the defeat imposing the so-called "40% threshold" requirement in the Scotland Bill in 1978, instead sustaining the provision by a vote of 298 votes to 243.2 On only one occasion has the government sought to reverse a defeat by making the vote for reversal one of confidence. That was in 1944 on an amendment to the Education Bill, when Prime Minister Winston Churchill appears to have overreacted by demanding an "adequate" majority. It was a move that attracted adverse publicity and has not been resorted to by any subsequent government.

Given the nature of most divisions in the Commons, and the political significance attached to them by the government, defeats in the division lobbies will normally fail into this category Defeats involving questions of confidence in the Government are very much the exception, not the rule. It may be that if government defeats failing within this third category nevertheless occur persistently, the government may consider its ability to govern is impaired. But that is a matter for political, rather than constitutional, judgement. As long as the government retains the confidence of the House it may, at its discretion, remain in office until the statutory five-year maximum for the life of the Parliament reached.

Who Decides?

Who determines into which category a defeat will fall? The determination rests primarily but not exclusively with the government. It may, though nowadays rarely does, seek an explicit vote of confidence from the House. It may determine that a motion to be divided upon involves fundamentally its ability to continue in office or that, for political reasons, it should constitute a vote of confidence. It may determine that a defeat on a particular issue, while not having been declared previously to have been a vote of confidence, nevertheless raises a question of confidence, and hence allow the division to fall in the second category of defeat identified above. Or it may decide – and usually does – that the issue involved is not one so central to its policy goals that it raises a question of confidence and so contents itself with deciding whether or not to accept the defeat. Most defeats, as already noted, are relegated to this third category.

In determining into which category a defeat falls, a government will be influenced by a number of factors. These will encompass both the nature of the issue involved and the circumstances in which the defeat has taken place. Defeat on a minor amendment to a routine Departmental Bill will be a straightforward candidate for relegation to the third category of defeat. Defeats which are the product of opposition parties combining against a minority government, which are the product of some deliberate Opposition ploy (opposition members leaving the precincts, hiding in nearby houses and then rushing back when the division bells sound), or which are the result of confusion or accident (a number of members stuck in a lift, for example) will affect the government's assessment as to whether defeat on an important issue under such circumstances raises a question of confidence. In 1964, when it looked as if the Government might be defeated on an amendment to the Address because thirteen Labour MPs were fog-bound at a Scottish airport, the Prime Minister made it clear that if the absence of the MPs resulted in such a defeat it would fall in the second category of defeat rather than being construed as a defeat on an implicit vote of confidence. Only in the event of a defeat on an explicitly-worded vote of confidence can it be said that the circumstances of the defeat are of no consequence: the government is obliged to resign or request a dissolution.

Two problems may be identified with the exercise of this power of determining into which category a defeat falls. One concerns relegating certain defeats to the third category of defeat; the other concerns elevating them to the first category. It is the latter problem which has aroused the greater controversy and generated calls for reform.

A government may relegate to the third category a defeat on an issue which opponents consider to be central to the government's policy and hence appropriate for inclusion in the second category, necessitating a vote of confidence. That governments should err on the side of putting such defeats into the third category is not surprising. Such action avoids embarrassing publicity and minimises the potential damage to its reputation. Should the Opposition nonetheless be convinced that the issue involved raises a question of confidence it has the option of tabling a motion of no confidence in the government; in effect, to raise to the second category a defeat which the government prefers to consign to the third. By convention, the government makes time available to debate an opposition censure motion. Hence, this particular power of determination is not exclusive to the government.

What is exclusive to the government, and at the heart of the controversy surrounding the issue of confidence, is the power of government to declare any vote to be one of confidence. For the government to declare the Second Reading of major Bills, at the heart of its parliamentary programme, to be issues of confidence – as Edward Heath did with the European Communities Bill in 1972 – may be considered a legitimate use of the power. More controversial is the power to declare minor amendments to entail a question of confidence. The capacity to abuse this power was well illustrated by Harold Laski. "Under the Balfour administration of 1900 it actually led to a question of whether pillar boxes should be green being made a matter of confidence; and the unfortunate critic of the executive (who was actually one of its supporters) found himself compelled to deny his own colour preferences. No one is ever so completely right as to deserve the power to attach to itself the penalties contingently involved in a vote of confidence."3

It is this capacity to declare a vote, any vote, to be one of confidence that has led to various proposals for reform. Most have been aimed at removing the power from the hands of government. One proposal is that the power to determine what constitutes a 'hanging matter', a confidence vote, should rest with the Speaker.4 Another is for explicit votes of confidence alone to be construed as confidence motions.5 A third proposal, politically and constitutionally the most radical, is for fixed term Parliaments, with no Provision for dissolution except in extraordinary circumstances. The problem with the first two proposals is that each would entail giving potentially contentious political powers to individuals who occupy positions of political neutrality. The Speaker would be placed in an invidious position if called upon to determine what was or was not a motion of confidence. The second proposal would likely entail the Sovereign being drawn into the political fray, given that the monarch alone could deny a Prime Minister's resignation or request for a dissolution. Inherent in the third proposal (fixed term Parliaments) is the risk of subjecting the country to a period of weak or even no government. Its implementation would also be fraught with difficulty, constituting a dramatic departure from constitutional and political norms and practice. It would, in any event, constitute the use of a hammer to crack a nut.

Though the problem identified is a real one, it is not perhaps as serious as it is made out to be. It is rare for governments to elevate divisions to motions of confidence. Various constraints operate to ensure such restricted use. As Churchill discovered in 1944, making divisions on amendments to Bills votes of confidence attracts ridicule from the media as well as from political opponents. It may constitute the exercise of a constitutional power, but it is an exercise that suggests political weakness. It is used most often to reduce the incidence of cross-voting by dissident government backbenchers. Its immediate effect is to ensure that the government wins the vote. Its medium term effect is to generate resentment among those dissident backbenchers, an effect that most governments would prefer to avoid especially if morale on the backbenches is not high. Furthermore, it is a power that cannot be used on a consistent and regular basis without inviting political opprobrium from press and opponents and even one's own supporters. Edward Heath could make the Second Reading of the European Communities Bill a vote of confidence: he could not easily have done so with the subsequent eighty divisions that took place on the remaining stages of the Bill. Finally, there is always the danger that the government might actually lose the vote. Even if declaring a vote to be one of confidence reduces to nil the number of cross-votes by dissident backbenchers (though even that is not certain), the government has no control over London traffic, the operation of' elevators, or fog at airports. Mr. Heath had a majority of eight for the Second Reading of the European Community Bill. Fifteen of his own backbenchers voted with the opposition. Had nine government supporters been held up in traffic, not only would the Bill have been lost but Mr. Heath would have had to make the unwelcome trip to the Palace. Fear of losing a vote even if made one of confidence appears to have served as a clear constraints upon the Labour Government of James Callaghan, not least in relation to its devolution legislation.


The essential problem concerning confidence is not the government's power to make a vote one of confidence. Rather, it is a problem of perception: the perception on the part of Members of Parliament that each vote is, in effect, implicitly one of confidence. It is not and realisation of this by members, and indeed by government, can have, and has had beneficial effects both for backbenchers on both sides and the government.

Recognition by Members that a government is constitutionally bound to resign or request a dissolution only in the event of losing a vote of confidence rids them of a significant and largely self-imposed constraint upon their parliamentary behaviour. Government backbenchers can vote against their own side on a substantive issue with the intention of defeating the government, knowing that by so doing they do not necessarily raise any wider constitutional issues. The backbenchers get their way; the government remains in office. Recognition of that fact also imparts to government a degree of stability. It knows its position is secure unless the vote is one of confidence. It can thus bear the burden of the occasional defeat without serious questions being raised as to whether or not it retains the confidence of the House.

However, recognition of what may be termed the constitutional reality by itself is insufficient to generate a change in parliamentary 6ehaviour. As noted earlier, the myth that every vote was implicitly one of confidence served to reinforce, not to cause, party cohesion. Recognition of the constitutional reality must be conjoined with the political will to impose a defeat on government. Members may recognize that they can impose a defeat upon the government without jeopardising its continuation in office; such a recognition does not imply that they will do so.

The British House of Commons in the 1970s witnessed both the realisation of the constitutional position of the government in the event of defeat and the growth of the political will to impose defeats by government backbenchers. The Heath government of 1970-74 suffered six defeats in the division lobbies. Defeat was not necessarily planned by dissident Conservative backbenchers; but that was the effect of their action. In response to none of the defeats did the government contemplate resigning or requesting a dissolution. Gradually, Members realised that the constitutional – as well as the political – constraints presumed previously to operate upon them were largely the imaginative constructs of government and the whip; they had no basis in constitutional practice or political fact. (The whips, as one Member observed, "were not so much whips as feather dusters"). Increasingly, Members began to take a degree of voting independence for granted and a number of independent-minded government backbenchers – notably George Cunningham on the Labour side – were prepared to engineer defeats on those issues on which they disagreed with government. In so doing, they incurred no hideous retribution; and the government was able to continue in office. This greater independence on the part of Members was continued in subsequent Parliaments and also made possible significant structural reform.

Recognition of the constitutional position of government in the event of defeat has thus been a partial product of, and has contributed to, the behavioural and attitudinal change of British Members of Parliament in recent years. Whether a similar recognition can be achieved in other Commonwealth Parliaments without a concomitant behavioural and attitudinal change is for others to determine.


1. These three categories I identified in some detail in an article in Public Law in 1978 though they were identified in broad terms as early as 1936 in a speech in the House of Commons by Prime Minister Stanley Baldwin.

2. House of Commons, Debates 944, cols. 597602.

3. H. J. Laski, A Grammar of Politics (1925), p. 349.

4 Ibid.

5. This proposal is advanced, in effect, by Nevil Johnson in In Search of the Constitution (1978).



Canadian Parliamentary Review Cover
Vol 8 no 4

Last Updated: 2020-03-03