Tactical voting advice - with the benefit of hindsight

It's always easier to be right after the event. London4Europe Committee member Michael Romberg plays with the numbers for the European Parliament elections now that the results are in. Companion articles look at how useful the opinion polls had been and a comparison of 2014 and 2019.

 

Contents of this article

• Table 1 sets out the actual polling results
• Table 1a goes into more detail on the "other" parties and individuals
• Table 1b has turnout information
• Table 1c  adds the actual vote shares together into various subtotals
• Table 2: how fair was D'Hondt?
• Table 2a: what if the electoral system had been perfectly proportional?
• Table 2b: how many votes were needed for each seat?
• Table 3: what if - the three main Remain parties had formed a single platform?
• Table 4: what if - the votes for UKIP and "other" parties had gone to bigger parties?
• Table 5: what if? tactical voting advice - with the benefit of hindsight
• Table 5a: actuals and options summary table: votes and seats for the Remain Parties
• Table 5b: what if tactical voters had boosted the party closest to winning the 8th seat?
• Table 5c: what if the Remain votes had been more evenly spread?
• Table 5d: what if the supporters of the weakest party had voted for stronger parties?
• Table 5e: what if all Remain voters had swung behind the leading party - see Table 3 for results
• Table 6: what if it had been a referendum? Let's pretend, using opinion poll information

 

Conclusions

Whether D'Hondt is fair compared with a purely proportional system depends on size of the constituency, the number of credible parties, and the pattern of votes in the particular election. For 8-seat London, the system can be reasonably fair compared with a purely proportional system. Table 2 shows that in this election as usual the system benefited the largest party - this time the Liberal Democrats. By the chance of the numbers it also slightly advantaged the Brexit Party. Change, Green and Labour received the right number of seats. The Conservatives were probably disadvantaged.

The starting point is that London is a Remain city. Table 6 shows that if the election had been a referendum, the Remain parties would have won 6 seats on the back of 71% of the vote.  Table 1c shows that the Remain parties actually won four seats on 48% of the vote. I understand that many people choose the party to vote for for reasons other than Brexit/ Remain; also, some people did not know what the Labour Party's policy is or voted Labour in the hope that its policy will change. 

Counting Labour as a pro-people's vote party as the People's Vote Campaign does means that pro People's Vote parties won 6 seats and 70% of the vote (Table 1c). Let's hope that is where Labour's policy actually ends up.

About 5% of votes went on other parties. Table 4 shows that if these votes had all gone on bigger parties then - on my assumptions - it would not have made a difference to the number of seats in this election. But The Brexit Party and the Greens would have received one of their seats in an earlier allocation round than they did. 

Before the vote, a lot of effort was put into giving tactical voting advice. We do not know of course how far the actual results reflect voters principled choices and how far their decisions were affected by tactical voting considerations. The analysis in Table 5 shows that with the same number of Remain votes it would have been possible to boost the number of seats from 4 to 5 by following one of four strategies: boosting the weakest of the parties (Change UK) at the expense mainly of the Greens; eliminating the weakest party and giving all its votes and some of the Liberal Democrat votes to the Greens; everyone backing the leading party (Liberal Democrats); and spreading some of the Liberal Democrat vote between the Greens and Change UK. 

One of those strategies was promoted by RemainVoter, another by RemainUnited.

It is clear from the calculations that these results are very tight: by the time you are looking at adjusted poll numbers for later seats differences in vote shares are measured in the hundreds. That reinforces my own view about tactical voting under D'Hondt. It's too complicated because there are too many options; the quality of data in regional sub-sets of opinion polls is not good enough to allow the necessary fine tuning; and it is hard to predict how far people will take your advice. My own advice was therefore not to vote tactically but to vote strategically for the Remain party that came closest to your ideals. 

The only effective tactic, I suggest, would be to form a common platform for a campaign. That of course brings problems of its own.

 

Tables

 

TABLE 1: The actual result for the six largest parties. Votes have been turned into seats using the D’Hondt method (read more about that here).  Seats won in each round are shown in bold. Poll numbers are then adjusted for subsequent rounds of seat allocation.

  

	Brexit	Con	Labour	Change	Green	LD
Votes	400,257	177,964	536,810	117,635	278,957	608,725
% of votes	17.9	7.9	24.0	5.2	12.4	27.2
1	400,257	177,964	536,810	117,635	278,957	608,725
2	400,257	177,964	536,810	117,635	278,957	304,362
3	400,257	177,964	268,405	117,635	278,957	304,363
4	200,129	177,964	268,405	117,635	278,957	304,363
5	200,129	177,964	268,405	117,635	278,957	202,908
6	200,129	177,964	268,405	117,635	139,478	202,908
7	200,129	177,964	178,937	117,635	139,478	202,908
8	200,129	177,964	178,937	117,635	139,478	152,181
Total	2	0	2	0	1	3

 

Notes

Brexit = Farage’s Brexit Party

Change UK, formerly The Independent Group of MPs

LD = Liberal Democrats

 

 

TABLE 1a: Other parties and individual candidates, set out by Brexit stance

 

Name	Votes
	(absolute numbers)	%
UKIP	46,497	2.1
Animal Welfare Party	25,232	1.1
UK EU Party	18,806	0.8
Women's Equality Party	23,766	1.1
7 Climate and Ecological Emergency Independents	4,939	0.2
Independent Mike Shad	707	0
Independent Ian Sowden	254	0
Independent Andrew Venzon	731	0
TOTAL PRO-REMAIN	74,435	3.3
Independent Alan Kirkby - not allocated	401	0
TOTAL	121,333	5.4

 

 

TABLE 1b: Electorate, turnout

 

Electorate	5,478,083
Ballot papers issued	2,259,078
Ballot papers rejected	17,397
Valid votes cast	2,241,681
Turnout	41.3%

 

Source: Regional Returning Officer for London

 

TABLE 1c:  Some totals of the actual votes cast and seats won.  

 

	Votes (% of total)	Seats
Pro-Brexit	52	4
Pro-Remain	48	4
Pro-No-Deal Brexit	20	2
Pro Brexit with a deal	32	2
Pro-Remain	48	4
Anti Peoples' Vote	28	2
Pro People's Vote	70	6

 

Notes:

• Pro No-Deal Brexit: The Brexit Party & UKIP
• Pro Brexit with a deal: Conservatives, Labour
• Pro-Remain: Change UK, Green, Liberal Democrat, Other Remain
• Pro-People's Vote according to People's Vote campaign: Change UK, Green, Labour , Liberal Democrat, Women's Equality Party
• Anti Peoples's Vote: The Brexit Party, Conservatives, UKIP

 

 

TABLE 2: How fair was D'Hondt? 

TABLE 2a: What if the electoral system had been perfectly proportional?

 

	Brexit	Con	Labour	Change	Green	LD
Votes	400,257	177,964	536,810	117,635	278,957	608,725
Share of 8 seats	1.4	0.6	1.9	0.4	1.0	2.2
Rounded	1	1	2	0	1	2
Actual seats	2	0	2	0	1	3
Difference	+1	-1	0	0	0	+1

 

Note: the rounded number of seats comes to 7, not the 8 that are available. If one increased each party's unrounded seat number by the same proportion until one of them reached a number which when rounded would give them an extra seat, the Brexit Party would be the first to gain an additional seat.

 

TABLE 2b: How many votes did each party need for each seat?

 

	Brexit	Con	Labour	Change	Green	LD
Votes	400,257	177,964	536,810	117,635	278,957	608,725
Actual seats	2	0	2	0	1	3
Votes per seat	200,128	-	268,405	-	278,957	202,908

 

 

TABLE 3: Actual result but on the basis that the three Remain parties had campaigned as a single Remain Platform and received the same number of votes in total as they received as individual parties. That is an arithmetic assumption. There are arguments that a Platform would have done better than individual parties (people like to back a winner) and that it would have done worse (people wish to back their party).

 

	Brexit	Conservative	Labour	Remain
Votes	400,257	177,964	536,810	1,005,317
% of votes	17.9	7.9	24.0	44.8
1	400,257	177,964	536,810	1,005,317
2	400,257	177,964	536,810	502,658
3	400,257	177,964	268,405	502,658
4	400,257	177,964	268,405	335,106
5	200,129	177,964	268,405	335,106
6	200,129	177,964	268,405	251,329
7	200,129	177,964	178,934	251,329
8	200,129	177,964	178,934	201,063
Total	1	0	2	5

 

 

TABLE 4: Actual result but with the votes cast for UKIP and "other" parties and individuals re-allocated. Note, the re-allocation has not been based on polling or other evidence of what their voters' second preference would be but on my assessment of the party's stance. 

 

	Brexit	Conservative	Labour	Change	Green	Liberal Democrat
Votes	446,754	177,964	536,810	136,695	332,894	610,163
% of votes	19.9	7.9	24.0	6.1	14.8	27.2
1	446,754	177,964	536,810	136,695	332,894	610,163
2	446,754	177,964	536,810	136,695	332,894	305,081
3	446,754	177,964	268,405	136,695	332,894	305,081
4	223,377	177,964	268,405	136,695	332,894	305,081
5	223,377	177,964	268,405	136,695	166,447	305,081
6	223,377	177,964	268,405	136,695	166,447	203,388
7	223,377	177,964	178,937	136,695	166,447	203,388
8	148,918	177,964	178,937	136,695	166,447	203,388
Total	2	0	2	0	1	3

 

Notes:

I reallocated the votes as following fairly arbitrarily and without consulting the organisations. Apologies for massive oversimplification of positions:

• UKIP to Brexit Party (UKIP do not approve of the Government's Brexit deal)
• Animal Welfare Party to Greens (AWP back second referendum; policies seem closest to Green)
• UKEU Party to Change UK (UKEU back immediate revocation of Article 50 notice. Some Change UK MPs have said there is no longer time to hold a referendum and that the Article 50 notice should be revoked)
• Women's Equality Party to Greens (WEP back Remain and People's Vote; have worked with Greens on EU issues)
• 7 individuals standing as Climate and Ecological Emergency Independents - to Greens. They want the EU to do stuff, so I assume they are pro-Remain
• Independent Mike Shad to Liberal Democrats (campaigns for Remain; LD as largest Remain party)
• Independent Ian Sowden to Change UK (He is pro-EU and pro-Revoke Article 50 notice)
• Independent Andrea Venzon of Volt Party to Liberal Democrats (He is pro-EU, critical of existing opponents of Brexit who just wish to form a new party or stick with "Stop Brexit" slogan; so maybe he likes the Liberal Democrats' new election slogan)
• the remaining independent, Alan Kirkby, not allocated (no information found)

 

 

TABLE 5: Tactical voting with the benefit of hindsight

TABLA 5a: Total vote of Remain Parties - Actual and options

 

Option	Change	Green	Liberal Democrat	Total
Actual - table 1 - votes	117,635	278,957	608,725	1,005,317
Actual - table 1 - seats	0	1	3	4
Boost the party closest to winning the 8th seat - Table 5b - votes	200,130	200,457	604,730	1,005,317
Boost the party closest to winning the 8th seat - Table 5b - seats	1	1	3	5
Spread the joy - Table 5c - votes	200,130	404,927	400,260	1,005,317
spread the joy - Table 5c - seats	1	2	2	5
Eliminate the weakest - Table 5d - votes	0	400,587	604,730	1,005,317
Eliminate the weakest - Table 5d - seats	0	2	3	5
Everyone backs the leader - Table 3 - votes	0	0	1,005,317	1,005,317
Everyone backs the leader - Table 3 - seats	0	0	5	5

 

Table 5b: Boost the party closest to beating the Brexit Party for the 8th seat

To obtain the 8th seat, a Remain party has to beat The Brexit Party by 1 vote, so a score of 200,130. In the last row of Table 1, Change UK is furthest behind the other Remain parties. Yet it is actually closest to beating The Brexit Party. That is because it has not yet won a seat. So the gap between the votes it has and the votes it needs represents the total number of votes needed (82,495). By contrast, because they have already won seats, the gaps for the Greens and the Liberal Democrats need to be increased by 2 and 4 respectively to show the total number of extra votes needed (121,303 and 191,795). Both the Greens and the Liberal Democrats have spare votes. So some of them could be reallocated to Change UK.   

 

	Brexit	Con	Labour	Change	Green	LD
Votes	400,257	177,964	536,810	200,130	200,457	604,730
% of votes	17.9	7.9	24.0	8.9	8.9	27.0
1	400,257	177,964	536,810	200,130	200,457	604,730
2	400,257	177,964	536,810	200,130	200,457	302,365
3	400,257	177,964	268,405	200,130	200,457	302,365
4	200,129	177,964	268,405	200,130	200,457	302,365
5	200,129	177,964	268,405	200,130	200,457	201,577
6	200,129	177,964	178,937	200,130	200,457	201,577
7	200,129	177,964	178,937	200,130	200,457	151,182
8	200,129	177,964	178,937	200,130	100,228	151,182
Total	1	0	2	1	1	3

 

Table 5c: Spread the joy

What if instead of the Liberal Democrats winning 3 seats they had won 2.  If the Liberals Democrats had won their second seat with the 200,130 needed to beat the Brexit party they could have given away the 82,495 votes that Change UK needed to beat the Brexit Party, and the 125,970 votes that the Greens needed to win a second seat.

 

	Brexit	Con	Labour	Change	Green	LD
Votes	400,257	177,964	536,810	200,130	404,927	400,260
% of votes	17.9	7.9	24.0	8.9	18.1	17.9
1	400,257	177,964	536,810	200,130	404,927	400,260
2	400,257	177,964	268,405	200,130	404,927	400,260
3	400,257	177,964	268,405	200,130	202,464	400,260
4	400,257	177,964	268,405	200,130	202,464	200,130
5	200,129	177,964	268,405	200,130	202,464	200,130
6	200,129	177,964	178,937	200,130	202,464	200,130
7	200,129	177,964	178,937	200,130	134,976	200,130
8	200,129	177,964	178,937	200,130	134,976	133,420
Total	1	0	2	1	2	2

 

 

TABLE 5d: eliminate the weakest 

The calculations in the lead-in to Table 5b show that it would not have been enough for Change UK voters to back the more successful Remain parties. Either everyone would have needed to come together (Table 3) or all Change-voters and a few thousand Liberal Democrats would have needed to switch to Green to give the Greens a second seat without losing the Liberal Democrat 3 seats.

 

	Brexit	Con	Labour	Change	Green	LD
Votes	400,257	177,964	536,810	0	400,587	604,730
% of votes	17.9	7.9	24.0	0	17.9	27.0
1	400,257	177,964	536,810	0	400,587	604,730
2	400,257	177,964	536,810	0	400,587	302,365
3	400,257	177,964	268,405	0	400,587	302,365
4	400,257	177,964	268,405	0	200,294	302,365
5	200,129	177,964	268,405	0	200,294	302,365
6	200,129	177,964	268,405	0	200,294	201,577
7	200,129	177,964	178,937	0	200,294	201,577
8	200,129	177,964	178,937	0	200,294	151,182
Total	1	0	2	0	2	3

 

 

TABLE 5e/ TABLE 3: Another strategy would have been for everyone to back the leading party. In the  polls in May that was clearly the Liberal Democrats. Arithmetically, that would give the same result as the common Remain platform - see Table 3.

 

 

TABLE 6: what if it had been a referendum? We turn a May 2019 London opinion poll about votes in a potential referendum into seats using the D'Hondt method (so this is not based in any way on the election result)

 

	Remain	Leave with the deal
% of opinion poll	71	29
1	71	29
2	35.5	29
3	23.7	29
4	23.7	14.5
5	17.75	14.5
6	14.2	14.5
7	14.2	9.7
8	11.8	9.7
Total	6	2

 

Source: YouGov for Mile End Institute of Queen Mary College, University of London. Fieldwork 7-10 May. Sample 1,015 - all in London

Question: "Imagine there was a new referendum on whether Britain should leave the EU under the terms of the deal [that has been agreed between the Government and the EU] or Remain a member of the EU, how would you vote?"

Excludes would not vote, don't know, did not answer (21% of unadjusted poll).

It's one poll. It comes with a margin of error (about 3 %age points). Do not take the result too seriously.

 

 

 

Articles on this page reflect the views of the author and not necessarily of London4Europe.