The voting system drives our politicians mad

After the first big Brexit Postponement, the Euro elections and all that excitement, the day job in No10 became somewhat dull. There are limits to what an administration that is definitely, unavoidably, finished can do (for another blog). Life was dull, unlike the polls, and so with some extra hours on my hands I decided to build a Swingometer.

The ostensible pretext was to showcase what you can do on Excel. The honest reason was to translate those barmy voting intentions figures into seats, because the system is well and truly insane. 

This is not just the standard LibDem whinge about the unfair voting system, though it is that too (in 2017, a percentage point of national vote share won the Tories 7.5 seats, Labour 6.5 seats, and the LibDems just 1.5).  If anyone should complain, it is the Brexit Party – on most of my simulations, their incredible performance earns them … Thurrock-all. 

No, it is about the crazy volatility the system induces, which in the hands of risk-seeking politicians is bound to encourage similarly crazy, risk-seeking behaviour.  Which I will try to show with Charts. 

First, a few geeky words on how it works.  Back in the days of Peter Snow, there was only need for one swing – Labour-Tory.  As this moved, red would shift to blue or vice versa. Now it is far harder. As far as I can see, to reflect what is currently going on, you need:

  • The loss of Tory voters to Brexit Party
  • The loss of Labour voters to Brexit Party
  • Ditto both to the LibDems
  • Ditto both to the Greens and
  • A swing from Labour to Tory 

Fortunately Excel is good at this, and you can take every seat in the country, apply it uniformly, and presto! There you have an aggregate result.  Here is a snapshot of the thing:

Screenshot the model

Of course, it is a model, and hence a simplification. Swings will not be uniform around the country, voter shifts will be way more dynamic, and so on – but it gives you a base case. And if you know techniques such as Data Tables, you can quickly eyeball a whole range of outcomes and see just how batshit the system is.  

For example, here is a Leaver’s nightmare: what happens if we assume the Brexit Party takes 2 Tory votes for every Labour vote, and the LibDems are held at 14%: 

Leavers Nightmare

Even if there is a substantial (5%) swing from Labour to Conservative, a Brexit Party scoring 10% will see the Conservatives short of a governing majority.  It is possible for Farage’s outfit to hit 20%, take hardly any seats, but see loads of Conservative Leavers replaced by Labour MPs, and the odd LibDem. 

Now what about the LibDems doing well? Let’s see what happens if the Brexit Party is crushed down to 5%, but the LibDems gain from the Brexity Mess the others have made: 

LD Scenarios

(correction: x axis should say LibDem vote share)

Again, even with a reasonably good Labour to Tory swing going on, the Tories go well short of a governing majority.  And see above what happens if there is a substantial LibDem surge – a sudden explosion of seats when they cross into the mid 20s.  Those plum Surrey seats all begin to fall. 

Finally here is what happens if you assume the spirit of mid 2019 can be recaptured in a general election, the LibDems retain some 20%, and the Brexit Party can look upwards to 25%

Flight from the Labservatives

A really mad chart, no? Brexit votes begin by taking seats from the Conservatives and handing them to Labour and the LibDems. Then as BXP rises further the Tory losses continue and the Labour gains plateau. This time, at around 20% the Brexit Party explodes into the House of Commons, effectively taking over from a Conservative Party already gravely weakened by the loss of voters to the LibDems.  

At around 23% each for Brexit Party and Labour (and 19 for the Conservatives), the Tories are reduced to a rump of 40 seats and Brexit Party takes Uxbridge; two or three more percentage points, and they walk into Downing Street themselves. 

News stories around the middle of the year talked about a Conservative Party terrified of being wiped out. It was not an exaggeration. Pious types like me sneer at their favouring party over country. I will continue to. No Deal is a horrible failure. But the beating this voting system delivers them in the face of a combined Remainer and Brexit-hardcase surge makes it perfectly understandable. 

Of course, you can tell the polls from June were bad. What is worse is what the model shows about the glitches in the system, such as:

  • LibDems do better when the Brexit Party is strong. This is because BXP votes lower the threshold for winning any seat off the Conservatives and Labour. The same applies, vice versa. 
  • Labour benefit from many of these scenarios, even as their vote share plummets from 2017’s levels.  In plenty of cases, Labour emerges as the largest party, even when the combined Tory-BXP share is closing in on 50%. In fact …
  • The major trick is hoping your main opposition is split. That is what makes the major difference for any one party – the fragmentation of opponents. 
  • The leads the Conservatives have recently shown could return them to power.  I plugged in 31 – 25 – 15 – 15 and saw the Conservatives with 342 seats. 
  • But if you start at a four-way split, any party is 3-5 percentage points from Nirvanna (or hell).  I could show scenarios where the Conservatives get back into power with an increased majority scoring just  27% in the polls. 

All of these points, taken together, add up to a mad voting system.  Shifts within the margin of error make gigantic, Britain-altering differences to the next House of Commons. Incentives are incredibly dodgy: the Remain/Soft Brexit parties actually benefit from the ongoing strength of the No Deal Brexit Party, the Conservatives sort-of like the LibDem vote being riled up, and so on.  And there are very good reasons for the Farage-istes to remain in existence. The prize is not just Brexit any more: some of the results from mid June might actually pitch them into government. 

This is not a system to encourage rational behaviour, but gambling on a resurrection. It explains why Labour hangs on grimly – they may just pull it off.  Likewise Johnson, likewise Farage. We are in the hands of gamblers. 


Published by freethinkingeconomist

I'm a mid 40s, former special adviser (Downing Street 2017-19, BIS from 2010-14), former FT leader writer and Lex Columnist, former financial dealer (?) at IG, student of economic history, PPE like the rest of them, etc. This blog has large gaps for obvious reasons. The name is dumb - the CentreForum think tank blog was called Freethink, I adapted that, we are stuck now.

9 thoughts on “The voting system drives our politicians mad

  1. Beautiful thanks. Might grouping LibDem targeted Cons seats and Lab/Cons battlegrounds, then applying some squeeze factors, show important differences? What do you think is the likely final % Brexit vote given an Oct/Nov election?

  2. The Betfair site has the chance of 0-9 MPs for the Brexit Party at around 70%, or 4/11. Looking at your figures this looks like generous odds, unless some Brexit seats could be taken with a squeeze.

  3. Have you modelled what would happen if the LibDems and the Greens formed an alliance for the next election as they did in the last byelection? Would expect, I think, an increase in those voting Green (and by proxy LibDem) if they thought it was going to be more than a protest vote and with the increasing worry about climate change?

    1. I need to have a go at that.

      Problem is, everything I am modelling is seemingly aimed at boosting non-Conservative voting results, which would suggest a particular blind spot …

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: