How people elect parliaments
Was the hung-parliament result of this month’s British elections affected by tactical voting?
Indeed it was – and even more so will the next election be, if it is held soon.
In the lead-up to the June 8 vote in the UK, a lot of effort was invested by ‘progressive alliance’ groups, using well-organised websites and social media promotion, to persuade supporters of the Labour, Liberal Democratic and Green parties to coalesce behind the best-placed non-Conservative candidate in key seats.
Tactical voting – a practice only needed in electoral systems based on single-member electoral divisions – is normally not an equal game between all contestants. It is an electoral strategy played by the supporters of two or more parties against one major opponent. The target of the strategy normally has a plurality – but not a majority – of support in the electorate.
Tactical voting needs a single foe, against which others unite for the purpose of keeping a potential government out of power. In England and Wales that opponent is currently the Conservative Party.
In many constituencies the Conservative candidate scores around 40% of the vote, and can win the seat even though 50% to 60% of voters oppose being ‘represented’ by a Conservative MP, if their opponents are divided among two or more alternatives.
With tactical voting, in theory a Liberal Democrat voter in one constituency will vote for the local Labour candidate because the latter has the better chance of outpolling the Conservative candidate. Meanwhile in some other electorate, a Labour supporter will vote for a Liberal Democratic candidate for the same reason.
Such tactical voting can also occur without any reciprocity between the parties, if the relevant voters can be persuaded that by doing so they are helping prevent the victory of the unwanted government.
Meanwhile up in Scotland – at least for the moment – the plurality party to be targeted by tactical voting is the Scottish National Party (SNP). There the Conservative party has tried to borrow the votes of Liberal Democrat supporters, or even Labour supporters, to oppose the SNP’s nationalist, pro-independence agenda.
Tactical voting needs to be distinguished from two other electoral realities.
Firstly, there is the situation where supporters of minor parties simply abandoning their true preferences and vote ‘against’ the major party they don’t want to form government.
For example, in the UK political observers infer that many more ‘Green voters’ exist than actually vote for Green House of Commons candidates, but that they do not cast their vote for their preferred party at elections because there is no real prospect of those candidates getting elected.
Britain’s nominal Green vote regularly registers at around 1% in elections (although it managed 3.8% in the 2015 elections). This is a far lower level of support for a green party than is seen in most other comparable democracies, the main exceptions being the other nations with election systems based on single-member districts: France, Canada and the United States.
There are less than half a dozen constituencies in the UK where Green candidates placed first or second in recent general elections for the House of Commons.
Only one Commons constituency – Brighton Pavilion – has elected a Green MP. Caroline Lucas won that seat with pluralities of 31% in 2010 and 42% in 2015, and this month won her first majority vote at 52%. Her vote growth has come primarily at the expense of Liberal Democrat vote share, and the latter party helped her out by declining to even nominate a candidate this year.
Secondly, there is always the problem that voters may simply be genuinely switching parties. If the vote for a Labour candidate goes up in a seat and the Liberal Democrat vote goes down, does that indicate tactical voting, or people attracted to Labour altogether?
There is no doubt that the British Labour Party has attracted a large growth in voter support, increasing from 9.3 million votes in 2015 to 12.8 million earlier this month. A proportion of this increase may well come from new voter turnout, especially among young voters. Such electoral shifts are not to be classed as tactical voting.
With these alternative explanations for voter behaviour in play, it is ultimately not possible to measure the extent of tactical voting in any one election, or over time.
Very detailed polling might give a good guide to the extent of tactical voting, but it is not clear that such specific polling was carried out in the UK in recent weeks. Perhaps the tactical voting campaigns will have research to reveal.
Campaign group Tactical2017 is representative of at least four tactical voting campaigns that were active in the recent UK election. (Others included Progressive Alliance, businesswoman Gina Miller’s Best for Britain, and the Guardian newspaper. They all published essentially identical recommendations, though some listed a few more target seats than others.)
Late in the recent election campaign, Tactical2017 promoted key-constituency voting recommendations in favour of the Labour candidate in 50 seats, in favour of the Liberal Democrat candidate in 25, and in one constituency they recommended that an independent be supported. (They made recommendations as well for many other, much safer constituencies.)
The Conservatives entered the election holding 38 of the 50 key seats where a tactical voting recommendation favoured the Labour candidate. At the election the Conservatives lost 19 of these seats – offset by one seat picked up in the other direction.
On June 8 the Liberal Democrat vote declined to low single figures in almost all of these seats. But again, does this really reflect tactical voting, or is it an underlying shift in actual voter support towards Labour?
To put the question another way, are these shifting voters today really Liberal Democrat voters, who temporarily support Labour in spite of their true preference, or are they now Labour voters? (Either answer, of course, makes an assumption about the extent to which voters define themselves in terms of a partisan identity.)
In the 25 key seats targeted for a tactical vote in support of the Liberal Democrat candidate, 24 were previously held by the Conservatives. Most of these were classic ‘Conservative v Liberal Democrat’ seats, where Labour candidates typically performed poorly.
On election day the Liberal Democrats picked up 5 of these 25 seats, but one more seat was actually picked up by a Labour candidate, with voters there obviously acting in spite of the tactical vote recommendation.
Compared to the 2015 election, Labour increased its vote nationally by 3.5 million votes, and in doing so greatly increased its ratio of votes compared to the other ‘progressives’, the Liberal Democrat and Green voters. At the 2015 election Labour candidates averaged 4.7 times the combined LibDem-Green vote; at the 2017 elections this ratio more than doubled to 10.4.
(It is not the case that all Liberal Democrat voters are closer to the progressive than the conservative side of politics, but the tactical voting efforts of 2017 were targeted at those that were, so for the moment let’s leave that complication aside.)
The 25 constituencies targeted with pro-Liberal Democrat tactical vote recommendations were all, by definition, seats where the LibDem vote in 2015 exceeded the Labour vote. The Labour/LD-Green ratio in 2015 was 0.34 in these seats.
In 2017 this figure also shifted towards Labour, but only to a ratio of 0.60. So the pro-Labour shift was present, but was of a lesser magnitude in these LibDem seats than it was elsewhere. This suggests that in these seats there was some association between tactical voting considerations and a constraint of the shift towards Labour among ‘progressive’ voters.
But was this difference caused by tactical vote campaigners and recommendations, or were these seats simply those with historically lower voter support for Labour, which was not where the party’s 2017 gains took place?
Overall, all that can be said is that the tactical vote campaign recommendations were running in the same direction as the dominant trend of the 2017 election.
In the ‘pro-Labour’ target seats, the impact of tactical voting is therefore indistinguishable against the background of a strong vote swing to Labour. In the ‘pro-LibDem’ seats, the general voter swing to Labour is still present, but at a more constrained level.
Four seats were won by Labour candidates by fewer than 100 votes: Kensington, Dudley North, Newcastle-under-Lyme, and Crewe and Nantwich. The inference that Liberal Democrat and Green supporters voting tactically helped bring about these results is very strong.
While Labour was likely the main beneficiary of tactical voting in 2017, the strategy would also have helped in some of the tougher Conservative-LibDem contests, where the Labour wave was not breaking. It will also have helped the Liberal Democrats defend the small number of seats they retained.
It won’t be possible to definitely list how many – or which – seats changed hands due to ‘progressive’ tactical voting, but noting that the Conservative government fell just 8 seats short of a majority, its pretty safe to conclude that tactical voting across England and Wales, whether campaign-directed or just plain voter-instinctive, deprived the May Conservative Government of its majority and brought about the new hung parliament.
How many people voted tactically in 2017?
UK Electoral Reform Society research indicated that around a fifth of all voters were contemplating voting tactically. That suggests an army of millions of tactical voters.
The actual number of tactical voters in any election, however, will be a far smaller figure.
The great bulk of tactical-willing voters, as with all voters, are registered in politically safe constituencies, and would have had little reason to vote tactically. Sadly, these are among the millions of voters whose votes are wasted under plurality voting.
The Electoral Reform Society and others may well have correctly identified that many millions of voters understand the nature of the voting system and would be willing to vote for candidates who were not their true choice. That alone is an indictment of the effects of the plurality voting system on genuine political representation.
As discussed above, election result data alone does not allow for distinguishing tactical votes from actual support-switching votes. Nonetheless, across the 75 or so seats at the 2017 election where it made most sense to vote tactically, it is fairly plausible that as many as a few thousand votes in each constituency were deliberate tactical votes, cast against each voter’s true preference.
That adds to, at most, a few hundred thousand actual tactical votes cast in 2017. The median of probable estimates would be somewhat lower.
However limited in number these voters were, they still proved to be immensely influential. They would have contributed to the results in the many seats that changed hands on election night.
Well, if it is safe to conclude that the June 2017 House of Commons elections were influenced by tactical voting, consider what will happen at the next general election, due at some point between (say) August this year and June 2022.
Those who set out to organize tactical vote campaigns in 2017 had a reasonable amount of information to go on. They were able to identify around 75 seats out of the 650 to target, and they could probably estimate that they needed to reach and persuade between 100,00 and 200,000 willing voters with their online, social media and community campaigning.
Those setting out to organize an anti-Conservative tactical vote campaign for the next election are in a vastly more convenient position.
To start with, they know that they only need to capture just nine currently Conservative-held seats to reach the point where a Labour government could be formed with Liberal Democrat and SNP support (with the votes of minor parties possibly also helpful).
(For Labour, higher targets include winning 25 Conservative-held seats to require only SNP support to form government, and winning 60 seats to be able to govern alone.)
Future tactical voting campaigners also know exactly which seats to target, and with what vigour.
Starting with 2017 figures, the most marginal seat in the Commons is Southampton Itchen, won by Conservative MP Royston Smith by a mere 31 votes.
So for Labour to win that seat – all other things being equal – they need only persuade 32 of the 2017 Liberal Democrat or Green voters to back the next Labour candidate.
There were 1,421 such Liberal Democrat votes and 725 Green votes at the recent elections. With an organized campaign, tactical vote campaign organisers should have little trouble in individually signing up the required number of voters.
Even easier, statistically, could be defeating the Conservative MP for Richmond Park, Zac Goldsmith. Goldsmith won the seat on June 8 by 45 votes, and his next opponent will clearly be the Liberal Democrat candidate. So only 46 tactical Labour voters registered in the constituency are needed. There were at least 5,773 Labour voters earlier this month. There would be far more than 46 active Labour party members alone within the constituency.
And so it goes. There are in fact 36 constituencies where, based on 2017 election results, tactical voting alone would be sufficient – in the absence of any other vote-altering events – for the seat to change hands.
The nine seat changes needed for the parliamentary numbers to bring in an anti-Conservative minority Labour government can be won by persuading no more than 24% of 2017-election 3rd-party supporters to deliberately change their vote.
The number of new tactical voters needed to win nine seats starts will be as low as around 3,200 votes (depending on whether ‘closest’ seats are defined by relative or absolute numbers).
Just over 20,000 votes found in just the right seats would be sufficient to take 25 seats from the Conservatives.
All this is without actually persuading any additional June 8 Conservative voters to fully switch their vote away from the current government.
For the Conservatives, the strategic problem is that they cannot fight back in the same way. Tactical voting is only a device that the divided opponents can play. The large party against which others are colluding can’t make use of the same tactics.
There might be some slight prospect for tactical voting against the SNP in Scotland to yield similar outcomes for the Conservatives, but not at nearly the same magnitude.
The Conservative party probably reached close to the high water mark of its possible Scottish prospects at this month’s election. In any case, well-placed Scottish Conservatives cannot coordinate tactically with similar Scottish Labour or Liberal Democrat candidates, because any wins of SNP seats by those parties are either neutral or negative outcomes for the Conservative government.
Needless to say, the campaign headquarters of the Conservative, Labour and Liberal Democrat parties worked all this out in the days following the election. The tactical voting campaigns can see it too. The situation places the balance of British politics on a knife-edge if a new election comes to pass any time soon, should the May government be unable to govern effectively.
How this pans out at the next election all depends on how long the current government lasts. Five years hence, things may not be the same. By then, the result numbers discussed above will have dated, the electoral register will have grown and slowly shifted, and innumerable political events will have intervened.
But if we learn nothing else, the figures above illustrate just how close the 2017 election proved to be.
Finally, it’s worth pausing to reflect on how extraordinary this situation is in terms of British voters having an ideally equal say in choosing the government of Britain.
The 2017 election, the results landed in hung parliament territory.
Out of 46 million registered voters – 32 million of which did vote – had as few as 5,000 ‘progressive’ votes been better aligned to the best-placed non-Conservative candidates in each constituency, the nation would now have a minority Labour government.
Conversely, had as few as 134 votes actually fallen for the Conservatives in just four of the seats narrowly won by Labour, the May government would have eked out a nominal (but not very workable) 322-seat parliamentary majority.
The tiny populations of voters in question are made enormously powerful by such a voting system. If only they can locate them, such voters can be targeted with surgical precision by election campaign strategists, while millions of British voters in safe seats look on, entirely powerless.
Such close elections prove just how unequal voting for parliaments using plurality-single-member divisions can be.