DEBUG: https://d3n8a8pro7vhmx.cloudfront.net/london4eu/pages/5/features/original/heart_photo.png?1501497680
DEBUG:
DEBUG: blog_post
Tactical voting advice:
21 May, 2019

Don’t. Instead vote strategically for Remain

London4Europe Committee member Michael Romberg looks at the evidence. The complexities of D’Hondt and the poor data quality due to small samples in regional subsets of opinion polls make tactical voting too difficult. Anyway, the real problem is not Remainers choosing the “wrong” Remain party, it is Remainers choosing pro-Brexit parties.

 

** updated 21 May with advice from Remain United **

  

Strategic voting: “for Europe”

We would encourage you to treat the European Parliament elections seriously in their own right. Vote “for Europe”.

We have brought together the European Group manifestos and the national party manifestos.

In my view, no vote for a party that supports Brexit can be a vote “for Europe” because the party leadership does not wish its MEPs to stay in the European Parliament and put the manifesto into effect. So a vote for the Brexit Party, Conservatives, Labour or UKIP is not a vote “for Europe” as all these parties back Brexit.

You can see what individual MEPs stand for here. That matters a little when considering which pro-Remain party is going to be best at implementing your vision for Europe. You cannot choose between candidates on a list; only the top one, two or three have a chance of being elected. But the MEP candidates do not matter at all in assessing whether the party is for Brexit or not – that will depend on events at Westminster. If the national party is for Brexit they wish their MEPs to not sit in the European Parliament.

So my key strategic advice is: choose which of the big three Remain parties suits you best and vote for them.

For those who wish to treat the elections as a referendum it is anyway the votes that count, not seats.

 

Tactical Voting: problems and caveats

This article is intended to assist those who do wish to vote tactically in order to maximise the number of seats won by the Remain parties.

London4Europe is all-party and non-party. We only care about staying in the EU.

In this article we are indifferent between the policies of the three main Remain parties: Change UK, Green, Liberal Democrat and the smaller pro-Remain parties. We are only concerned with polling results.

 

Tactical voting under the D’Hondt system is not the same as under First Past the Post

In a constituency of the size of London it is not enough just to identify the leading party and back them. Depending on the spread of votes additional votes for the leading party might not obtain another seat while those same votes for a second or third placed Remain party might lead to a seat.

So for example when you are looking at who would get the 8thseat, the leading Remain party’s total votes might have been reduced to one third say if they have won two seats and so be nowhere near winning, but the whole of the next placed Remain party’s vote might be. So who should get the tactical vote?

You also have to consider how other voters will behave, including those who intend to vote tactically.

An explanation of the D’Hondt system here, to understand it better have a look at some of the calculations using opinion polls on our latest blogs page, and read an analysis of the problems of tactical voting in an LSE blog; or a different line of argument that supports the "back the leading party" approach used by Remain United.

Tactical voting is unlikely to make much difference. When we did a counter-factual calculation and added together the votes of all the Remain parties into a single platform we found that it made no difference in half of the opinion poll calculations and increased the seats by just one in the other half, usually at the expense of Conservative or Labour rather than of Farage.  

Christina Pagel Professor  of  Operational Research at UCL looked at D'Hondt. She addresses tactical voting at the end of her blog and for a constituency the size of London advises "... just vote for the Remain Party that you like best since tactical voting is unlikely to make a difference in the final allocation of seats to Remain supporting parties in that region". She goes on to say "The D’Hondt system isn’t perfectly proportional but it’s not bad and overall party vote shares matter just as much in the post-election analysis – so don’t get too stressed out by worrying about which party to vote for!"

 

Opinion polls come with a margin of error

A rule of thumb is that headline findings from an opinion poll come with a margin of error of +/- 3 %age points. That margin of error is larger when looking at a subset of the poll, as the London numbers usually are. A good number for a poll is 2,000. By the time you are looking at just the London numbers, most polls in May were in the range 83 to 364 respondents.

Only three recent polls had large numbers of replies in London: YouGov 7-10 May (1,015), YouGov 12-16 May (863) and YouGov 8-17 May (1,111), which all gave similar results. But the results may reflect biases inherent in YouGov’s methodology.

Our tracker brings the polls together and the trends should be more accurate than individual polls. We will publish our final tracker on Tuesday 21 May evening in our latest blogs section

 

Other sources of tactical voting advice

I suggest that you look at a range of sources of tactical voting advice, including Remain Voter and Gina Miller’s Remain United, though I can offer no assessment of their competence or credentials. You will be able to look at how they reach their conclusions to see whether they have a sophisticated understanding of tactical voting under D’Hondt and can address concerns on data quality. Neither treats Labour as pro-EU.

Here is a 14 minute interview between Mike Galsworthy of Scientists for Europe and Cheryl Hung, the lead programmer of Remain Voter. It gets interesting from about 4 minutes where she starts talking about what they are trying to do and really interesting from 7 minutes when she makes clear that she recognises the complexity of tactical voting under D’Hondt. She is a software engineer and not a statistician and so she did not address data quality as an issue. That is what will undermine the usefulness of any analysis.

Remain Voter issued their recommendation on Sunday 19 May. For London they recommended voting for Change UK. They might update that on Tuesday if polls change.

INFacts say that they are not making recommendations but setting out the opportunities and risks (that if a Remainer moves from a Remain party they damage that party's chances) in each region. For London they see the Liberal Democrats and the Greens as the best bet.

Edwin Hayward is a pro-Remain computer scientist who has obtained a Twitter following. You will need to decide how much of a quality recommendation that is. Colleagues at UKPEN are drawing attention to his election analysis. For London, he recommends Liberal Democrat or Green.

Remain United have recommended voting for the Liberal Democrats as the leading Remain party in London.  Their survey uses an MRP analysis which they hope improves data quality. MRP takes the answers to the survey, relates them to the demographic characteristics of the respondents, and then applies the results to the demographics of a constituency.

 

Do not vote for the micro parties

There are a number of micro-parties and independents standing, some of which are pro-EU. The three large YouGov polls showed that they had total support of 5% in London. That is not enough for any of them to win a seat. No opinion poll has suggested that any of them could.

Nor will anyone bother to count the votes of the micro-parties and independents when looking at the total Remain/ Leave vote.

But when parties are scrabbling for 8thplace differences can be small. So those votes that went on “other” parties might be enough to make a difference. So a vote for an “other” party will not affect anything but could have produced another seat for a Remain party.

 

What the opinion polls tell us

The Liberal Democrats are the leading Remain party in London. They are also in the lead nationally. The poll lead looks to be large enough to be statistically valid. They are in the running for one or two seats.

Support for the Greens is running at much the same level as for Change UK. They are each in the running for one or zero seat. The difference between them in the London polls is too small to allow a statistically valid conclusion to be drawn. In the three large London polls, the Greens have led Change UK. In the national polls the Greens tend to lead Change UK. But the difference between the two parties is not enough to be certain, given the margin of polling error.

I have done no analysis of my own of which party is likely to win an extra seat if it received additional votes or what the risk is if taking votes away from another Remain party.

 

Conclusion

Low data quality in the regional sub-set of national polls, the complexities of D’Hondt and the need to predict voter behaviour in the light of tactical voting advice mean that tactical voting is unlikely to work in London.

Anyway, the problem is not Remainers choosing the “wrong” Remain party; it is Remainers choosing a pro-Brexit party.

So, vote strategically: if you are a Remainer and prioritising that then choose the one of the three Remain parties that comes closest to your vision of Europe – and hope for the best.

 

 

 

The London4Europe blogs page is edited by Nick Hopkinson, Vice-Chair. Articles on this page reflect the views of the author, not necessarily of London4Europe.