Following my post on how this EP election has pulled the plug on the Greens’ perennial complaint about how the electoral system disadvantages them, a commenter made the point that, still, the electoral system needs to be reformed.
Agreed, and I take the opportunity to say how. There is still reason to change the single transferable vote (STV) system at all levels of election: local, national and European. Reasons? First, it can still result in mathematical anomalies. Not all these anomalies will necessarily result in serious consequences (as was in the 1981 general election) but they are still the source of unfair outcomes. Second, STV is unnecessarily complicated. That might sound as a purely aesthetic reason but I believe that the simpler the system the more transparent the process.
Two premises need addressing. The first — and the most controversial — is to prohibit cross-party transfers. STV’s most important count is the last one when it becomes known who are the representatives elected. This is completely at odds with every single Maltese’s view on which is the most important count: the first. The evidence is all over the place: in 1981 the Nationalists protested the fact that the technical and political outcome of the final count was different from the first and in 2009 Labour is celebrating a victory of 54 to 40 percent of the first preference votes not 4 to 2 MEPs.
This was, in my view, the greatest shortcoming of the 1994 Gonzi report on electoral reform. In its aim to retain the same way of voting but with a different method of counting the commission which drew up the report proposed a system which had two irreconcilable features: proportional representation based on first preference votes and the possibility that these votes move to other parties (thus disrupting the original proportionality) in subsequent counts.
As the suggestion to cross-voting was a prominent over at J’Accuse in the run-up to the election I will take some time to address this feature of STV. Jacques suggested to readers to give their first preference to the Greens’ Cassola, second to Labour’s Mizzi or Grech and third to Nationalist Demicoli. The reason given was that ideally try to ensure that Malta has representatives in three EP groups, in this case the Greens, the Socialists and the People’s Party.
Jacques is wrong mathematically (he also gives the wrong reasons for suggesting a higher preference to Labour candidate than a Nationalist candidate but I will not go into that). STV is not the Eurovision Song Contest where you give douze points to one candidate, ten to another and eight to another (which, incidentally, is the central feature of another preferential system, the “Borda count“). You see, it’s called single transferable vote for a reason. Jacques’ vote could not have contributed to electing more than one candidate: it would move from Cassola to Mizzi only if Cassola was eliminated from the race or if Cassola was elected and Jacques’ vote was part of the surplus (i.e. in excess and not needed by Cassola to be elected).
If Jacques concern was electing an MEP in as many groupings as possible he could have equally voted exclusively for the candidates of the Greens or Labour or the Nationalists and the mathematical chances would have been the same. Knowing from past experience and from polls that the Labour and Nationalist Party would have definitely elected at least one MEP, Jacques could have voted for Cassola and Ebejer-Arqueros only and his original aim would still have been achived. The restriction on cross-party voting would not have altered any of his ultimate intentions.
Cross-party transfers are an unnecessary complication which only stands in the way of national proportionality and serves as a source of possible anomalies (see here and here for examples with regards to casual elections, by way of mentioning just one). Prohibiting them would be one huge step ahead. But here’s a second: end even intra-party transfers. Yeah, that’s right let’s ditch the preferential system completely. Why? Because it is a complete waste of time.
Professor John C. Lane who has studied Maltese STV wrote an essay on the matter. His conclusion? That “STV vote transfers create results which, to a remarkable degree, a simpler process could also have achieved” because most candidates who are in the lead in the first count eventually are elected. The outcome of this year’s EP election adds justification to another speculation: that when candidates in the lead in the first count fail to be elected it is thanks to their rank on the ballot sheet which in Malta is determined — very unfairly — by the alphabetical ordering of their surnames.
So there you have it. Voters vote simply by putting the mark, any mark, next to the name of one candidate. Votes are counted and parties are allocated seats proportionately which are then filled by the candidates for that party who have the most votes. Counting votes would not take days and can even be made electronic (the counting, not the voting) in that the technology that recognises a mark is simpler and, therefore, cheaper than one which has to recognise the myriad ways in which voters write their “1″s and their “7″s across party lines.
Interesting analysis Fausto. Once the veil and panic of an oncoming election is removed we might be able to reason more open-mindedly.
First of all the effectiveness of the Cross-Vote was also based on a gamble – if you cannot influence with your first choice then make your second and third choices count too … if they influence different parties’ representatives all the better. The way candidates like RMTT or Marlene Mizzi hung around waiting for ANYTHING that could be inherited showed that in this case a cross-vote would at the very least send a message (as weak as it may be …). But then again the rules of the game as you very clearly point out do not allow for a clear message from the electorate – even when it comes to who you want to elect).
In a hypothetical world where J’accuse’s advice was followed by thousands and not just a few, having a long list of identical (block-vote) cross-votes would have sent the message that J’accuse suggested that the electorate should deliver – that they have not been conned by the PLPN shenanigans and voted by preference (with all its faults) rather than by party block vote. People wanting to leave a message instead chose to stay at home.
Now to your suggestions. I see proportionality makes a lovely reappearance. Cool. I would back your last suggestion completely. One vote, one candidate – total party votes translated into proportional representation and party representatives determined by number of votes and not by parties. Beautiful. When do we start? And what threshold for one seat? 1/6th of total votes?
P.S. Electronic votes – definitely!
P.P.S. Do you think John C. Lane is the same John Lane as the one in this thread on Running Commentary?
First, I’m a bit wary about “messages” being sent through ballot papers. You are around to explain a first preference to Cassola, a second to Mizzi and a third to Demicoli. But is there anyone to explain, for example, the “message” (if any) from considerable transfer from Lowell to Cassola?
Secondly, if there is indeed a message do not expect it to be easily “read”. If you gave your first preference to Cassola, second to Mizzi and third to Demicoli the “message” could only be “read” as far as Mizzi. When Mizzi fell out from the race Demicoli had been out for a long time and your vote went straight to the “non-transferred”. I very much doubt that at the twenty-eight count when Mizzi bowed out there was anyone around to note that you had, in fact, given a preference to Demicoli. Too little and definitely too late.
I’d rather stick to the primary aim of voting being to elect representatives. Non-transferable systems might be weaker at “sending messages” than transferable systems but as they “elect” representatives in a much more unequivocal way I’d give them my first preference any day.
Now, to electoral systems. Proportionality does not make a “re-appearance” here. It gives rise to risks of serious anomalies and is unnecessarily cumbersome but we’ve had a system of proportional representation for every election we’ve had since 1921. The discussion we’ve had since the 1980s refers to exclusively to a national proportionality as opposed to the district proportionality we have when we vote for our national Parliament. The discussion is therefore irrelevant in the context of the EP election (the subject of my post) as the country is one district and it is undesirable in the case of local election (since the choice of local government should be exclusively determined by locals).
Equally irrelevant for EP and local elections is the discussion on thresholds for mathematical and institutional reasons.
The mathematical goes like this. Take the most popular threshold in European electoral systems: 5% of the vote. As there are twenty five percents in a 100% that means that such a threshold can only possibly kick into operation in a system electing more than twenty representatives. Anything less than that and a party can cross the 5% threshold and still not elect anyone because there are others with a stronger showing and who’ll fill up the limited number of places. The discussion is not relevant here because we can only hope for six places in the EP and the Local Councils Act sets the maximum number of councillors in a locality to thirteen.
Institutionally, the arguments for thresholds is applied as a way of enhancing governability. This is completely irrelevant when it comes to EP elections and, with the automatic mechanism to elect a mayor, not very compelling in the case of local councils.
Now to your question: when do we start? When some two years ago we were discussing the adjustment introduced in this regard to the Constitution you did not appreciate my point that no matter how partial progress is, the more we move closer to national proportionality, the better.
I am well aware that changes in this regard can only come about slowly. And not only in Malta: as far as I know the only democracy to change its electoral system completely in one sweep was New Zealand in the mid-1990s. Rightly so. The way to elect representatives is at the very heart of a democratic tradition and one should only tinker and then tinker very carefully. That is what we did in 1987, in 1997 and in 2007. On each occasion it may have appeared that we were not doing much but, nonetheless, we were moving forward.
Not so quickly on electronic voting. I’m always careful to distinguish electronic voting from electronic counting as they are two completely different issues with completely different concerns. The latter is also less controversial so you can expect (or at least hope) for some progress there. As regards electronic voting I’d only just not rule it out completely. I think that despite whatever checks and re-assurances there will still be people feeling uncomfortable voting that way and, for that reason alone, I would stick to the old pen-and-paper method.
No idea about “John Lane”. Considering that the John Lane I quote seems to sign exclusively as “John C. Lane” and his interest in Malta seems to be limited to its electoral system it’s a safe bet to say that they’re not one and the same person.
[...] on 24 June 2009 at 6:24 pm I had intended to write about this for some time and a post about ditching single transferable vote (STV) and an excellent (as usual) op-ed from Times columnist Mark-Anthony Falzon provided the final push. [...]