The Commons’ attempts at indicative votes on Brexit have failed: no option secured more support than opposition, leading many to propose using preferential voting instead. However, this carries its own risks in the form of strategic voting and a perverse winner, writes David Klemperer. He explains why Majority Judgment may be better in similar future situations.
On 27 March and 1 April, MPs took part in a series of ‘indicative votes’ on Brexit, designed to enable the House of Commons to make clear its preferred outcome. On both occasions however, none of the Brexit options voted upon succeeded in securing more votes in favour than against. In a crucial sense then, these votes were a failure: although they provided MPs and the public valuable information on MPs preferences, the process failed in its aim of determining a clear ‘Will of the House’.
A central reason for this failure was strategic voting. Under the system used, ‘Approval Voting’, MPs were expected to vote in favour of any option they could accept, signalling their ‘approval’. But in practice, a significant chunk of MPs voted strategically, refusing to vote approval of compromise options, and instead abstained on or opposed all options except their ideal outcomes. In doing so, they each hoped to secure the sole success of their own most preferred outcome, but in practice collectively secured the failure of every possible outcome.
In response to this situation, a number of politicians and journalists have advocated a different voting system be used, specifically ‘preferential voting’, arguing that this would improve the process. Kenneth Clarke MP had in fact submitted an amendment to use preferential voting back in February, and on 1 April Layla Moran MP explicitly linked her decision to only vote for her ideal outcomes to the lack of preferential voting, stating: ‘I cannot … rank [compromise options] equally with the outcomes I believe to be truly in the national interest. I wish we had preferential voting.’
So, what is preferential voting? A preferential voting system is one in which voters rank the options presented to them in order of preference – putting a 1 by their first choice, a 2 by their second choice etc., for as many choices as they choose to express a preference for. Although there are different ways in which these preferential votes can be counted, the preferential system we hear advocated on this matter is the ‘Alternative Vote’.
Under this system, if no option received over 50% of first preferences, the option that received the lowest number of first preferences from MPs would be eliminated, and the ballots of the MPs who gave that option their first preference would be redistributed to their second preference. This process of elimination and redistribution of ballots to their next remaining preference would continue until one option had accumulated over 50% of ballots, at which point it would be declared the winner. This system is also known as ‘Instant Runoff Voting’, because it essentially uses the preferences expressed on the ballot papers to conduct a series of runoff votes, until one option has a majority.
Although this system would have the merit of producing one ‘winner’ amongst the options, it comes with its own problems. Firstly, it completely fails to eliminate or even reduce the incentives for strategic voting by MPs, who might wish to give a higher preference to options they believe capable of beating their least favourite options in later runoffs, rather than voting honestly. Secondly, and perhaps more importantly, the options with broadest (but shallowest) support would risk being eliminated early on, leaving the later runoffs occurring between options with narrow (but intense) support, and ultimately producing a ‘winner’ strongly disliked by a majority of MPs.
How then can we allow MPs to express more complex preferences, while reducing their incentives to vote strategically, and being sure to avoid a perverse outcome? A possible solution is a voting system called ‘Majority Judgment’, devised in 2007 by the French mathematicians Michel Balinski and Rida Laraki. Under this system, rather than simply accepting or rejecting options, or ranking them in preference order, MPs would assign each option one of a pre-determined number of grades, for instance ‘excellent’, ‘good’, ‘acceptable’, ‘poor’, or ‘unacceptable’. Crucially, each MP would be able to assign the same grade to multiple options.
This allows for the expression of complex preferences: a pro-Remain MP could for instance give the ‘excellent’ grade to both the options to revoke Article 50 and to hold a referendum, the ‘good’ grade to Common Market 2.0, the ‘acceptable’ grade to a Customs Union, and the ‘unacceptable’ grade to all the No Deal options. By contrast, a pro-Hard Brexit MP could give the ‘good’ grade to the No Deal and EFTA options, the ‘poor’ grade to the Customs Union and Common Market 2.0, and the ‘unacceptable’ grade to the revoke and referendum options.
How could these votes be counted? For each option, the median grade assigned to it would be determined, and the option with the best median grade would be deemed to have won. In the event of a tie, the ballots giving the median grade would be progressively removed until one option has the highest median grade.
Majority Judgment would therefore provide us with detailed information about the distribution of preferences within Parliament, and would be able to produce one winner, in a way that rewarded options capable of attracting broad support or acceptance. This system would also present MPs with fewer incentives to vote strategically, as in most cases the most effective strategy is to vote honestly.
Perhaps then, before MPs risk carrying out another series of inconclusive votes, or blunder into the strategic nightmare of Instant Runoff Voting, they should consider making the Will of the House clear using Majority Judgment.
_______________
David Klemperer is Research Fellow at The Constitution Society. He writes here in a personal capacity.
Two questions:
1. What if the preferences of the UK government are currently the OPPOSITE of the people who elected them?
2. What if the UK has a government-electing voting-system that will USUALLY deliver the situation as described in question 1?
In such a case it wouldn’t matter which (indicative) voting system was used to make a collective governmental decision, as the result would be the opposite of what the electorate wanted anyway.
It’s also worth keeping in mind that the UK’s current voting system IS the type that’s described in point 2 (the other commenter Clay Shentrup can tell you all about that!).
This is why we’re currently stuck in ‘Brexit Limbo’ – because the referendum result said ‘Leave’ while the General Election result said ‘Remain’.
It’s yet another sign that the UK must change its voting system to either Approval Voting or STAR Voting (just web-search those terms to learn about them).
The only way that that will happen is by convincing the various political parties (mainly the big two) to use Approval and STAR Voting in their internal voting processes.
SO LET’S STOP FAFFING AROUND WITH THE SYMPTOMS OF OUR BROKEN VOTING SYSTEM AND FOCUS ON THE ROOT CAUSES INSTEAD!!
This was technically Score Voting on a scale of -1 to +1 (Yes = +1, Abstain = 0, No = -1). Approval Voting is Score Voting on a 0-1 scale. Like a FPTP ballot but where you can vote for as many options as you wish.
Majority Judgement (Score Voting using media instead of sum or average) is a highly problematic system, as explained here by Warren Smith, the Princeton math PhD who’s work formed the basis for the book “Gaming the Vote”.
http://scorevoting.net/MedianVrange
You can see Bayesian regret values for a host of different voting systems here.
http://scorevoting.net/BayRegsFig.html
Note that Score Voting and its simplified form of Approval Voting are generally dominant over the other commonly discussed alternatives.