The scenes outside Westminster and Whitehall last May were unfamiliar to British eyes and ears. Politicians scurrying from one set of talks to another to establish who was willing to do a deal with whom and on what basis. An incumbent prime minister left hanging around as a caretaker until it was clear who was to form the next government. The nation kept waiting to find out who would take charge of the country for the next five years.
It is not the way things are supposed to happen in Britain. The nation votes.One party secures an overall majority of Commons seats. A defeated prime minister packs their bags and leaves Downing Street within hours. Days of haggling in darkened rooms to form a coalition government is meant to be an undesirable ‘continental’ practice that denies voters the right to determine for themselves who is to be entrusted with power.
The mechanism that is meant to deliver the British vision of single-party, alternating majority government is its electoral system. First past the post is said to have two key attributes that, between them, ensure that one party always has a majority of Commons seats. First, it denies both votes and seats to third parties. Second, it delivers a ‘bonus’ of seats to whoever comes first in votes. Elections are thus a choice between two big parties, one of whom is more or less bound to emerge a clear winner.
However, first past the post evidently failed to deliver in 2010. But then it did not quite have a perfect record before – nobody secured an overall majority in February 1974, either, and the country was back at the polls again after just eight months of minority government. Perhaps we have to accept that there will be the occasional accident, that, just as the 1974 experience was not repeated for another three decades, so we can confidently anticipate that, if voters reject the Alternative Vote in next May’s referendum, normal service will soon be resumed.
This, however, would be a mistake. What happened last May was not a one-off accident. Rather it was the product of long-term trends in voting behaviour that between them have done much to undermine the two crucial attributes that once ensured the system did nearly always produce a majority for either Labour or the Conservatives.
Consider, first, the squeeze that first past the post is meant to inflict on third parties. In the immediate postwar period there was no doubt about its existence. In 1951 and 1955 over 95 per cent of all votes were cast for either Labour or the Conservatives. At no election between 1950 and 1970 did more than a baker’s dozen of third party MPs manage to secure election.
But this two-party domination has long since been eroded. From 1974 onwards, typically around a quarter of all votes have been cast for third parties. In 2005 the proportion of votes cast for the two main parties fell below the 70 per cent mark for the first time. At both of the last two elections it was lower than at any time since 1922 – that is, the first election after the partition of Ireland and the first at which Labour overtook the Liberals. This increased third-party support is not just simply a reflection of the Liberal Democrats’ performance – at each of the last four elections around one in 10 votes have been cast for parties other than the Conservatives, Labour or the Liberal Democrats, well above the level at any previous election since 1922.
More importantly, third-party votes are now being translated more readily into third-party seats. Since and including 1997, a minimum of 75 third-party MPs have been elected to the House of Commons at each general election. During the previous 70 or so years that figure had not even come close to 50.
First past the post only denies third parties representation if their vote is geographically evenly spread. However, support for the various parties in Northern Ireland, together with the nationalist parties in Scotland and Wales, only arises in their respective parts of the UK. Meanwhile, although still quite evenly spread, the Liberal Democrat vote has become somewhat more concentrated geographically than before. In 1983, 26 per cent of the vote in Great Britain delivered the party just 23 seats; in 2010, 24 per cent was enough to secure 57.
The ability of first past the post to deliver a ‘bonus’ to whoever comes first also depends on the geography of party support. That bonus will only exist if there are lots of seats that are marginal between the two largest parties. That condition ensures that if there is a small swing from, say, Conservative to Labour, such that what at the previous election was a small Conservative lead in votes now becomes a small Labour one, sufficient seats change hands to turn a previous Conservative overall majority into a Labour one.
In the 1950s and 1960s around 160 seats – or just over a quarter of all the seats in Great Britain – were ones in which, in a closely fought election, the winner’s lead would be less than 10 per cent of the votes cast for Labour and the Conservatives combined. As a result a one per cent swing between the two biggest parties typically resulted in nearly three per cent of the seats changing hands.
But the number of marginal seats fell precipitously in the 1970s. By 1983 only 80 seats could be regarded as falling into that category. Thereafter, there was some sign of a limited recovery, but in May this year the number fell back once again to just 85.
The explanation for this change is relatively simple. Compared with the 1950s, Labour’s vote is now much more concentrated in the northern half of the country, while the Conservatives have become a party of the south. The divergent swing to Labour in Scotland in May was but the latest manifestation of this now long-established phenomenon. And, as the country has pulled apart into two distinct halves politically, so the number of constituencies where both Conservatives and Labour do well has fallen away.
So the mechanism that once delivered overall majorities now looks decidedly rusty. Its ability to deliver a bonus to the winner has been in question for nearly four decades, a change that until May was camouflaged from sight both because many a recent election has been won by a landslide (in votes) and because the system has (in another, separate, development) become biased to Labour. At the same time first past the post has become less effective at keeping third parties out of the House of Commons.
The degree to which, in tandem, these two developments have undermined the ability of first past the post to deliver an overall majority to one party is evident if we look at the voting leads that Labour and the Conservatives would need to win the next election if the country’s electoral geography and levels of third-party support were to remain as they were last May. Labour would need nearly a three point lead, the Conservatives nearly an eleven point one. In short, no less than a 14 point range of results would all indicate another hung parliament.
Of course, the coalition intends to introduce new constituency boundaries that will eliminate some (but not all) of the bias to Labour. However, making it a little easier for the Conservatives to win an overall majority, while making it somewhat more difficult for Labour, simply shifts the range of results that produces a hung parliament, rather than reduces it. It seems we may have to get used to the sight of politicians scurrying around after polling day after all.
This analysis seems reflective of a number of recent academic papers. As with all such papers the review process takes time enough to make the data slightly out of date. In this case this has the drmatic effect of largely ignoring the dramatic effect of the collapsed lib dem vote. This follows a pattern set by previous pacts and coalitions whereby the weaker party is disporportionately punished for deeds in power. I predict a return to an era of two party dominance unless we see major splits in the large parties.