Political Reform‎ > ‎

The Electoral Setup - The Voting System

The electoral setup (the proposed Icelandic constitution does contain some electoral system changes) is potentially something of a minefield. I think it needs to be looked at though. An election asks the basic question: who do you want representing you and running the country? As is evident from the field of opinion polling, how a question is asked can have quite a big impact (and often involve quite subtle effects too) on the kind of answer received. Any voting system is, I guess, one specific way of framing this particular question.

Greater Proportionality, PR-STV and an Electoral Commission

There's a good argument for increasing proportionality, and an even better argument for a proper constitutionally-based Electoral Commission. Such a body would perhaps be of even greater importance if constituency sizes were larger, because then the purely legislative requirement to follow county boundaries as much as possible, as in the current electoral act, might not be quite as relevant or perhaps could be set aside (following county boundaries at least cuts down on the scope for Gerrymandering). An Electoral Commission would greatly limit the likelihood of a future government being tempted, or even able, to go down the Tullymandering route again (some of the points on special legal arrangements I made in the "democratic architecture" category would be relevant here). The monitoring of ethics in politics and other issues like lobbying could also be placed under its remit.

In theory, increasing proportionality is easy (just bump up the minimum allowed constituency size in the constitution from three to something larger). Personally I'd go for a seven minimum (moving back in the direction of the earlier days of the state where we had some 7 and 9 seat PR-STV constituencies). This would allow far more room for new parties to develop, plus perhaps weaken the TD-constituent bond a little, which might be no bad thing. There are inevitably counter-arguments to any proposal. There's a danger larger constituencies could lead to higher TD constituency workloads and increased constituency rivalry between TDs of the same party.

There's also still a certain case for changing from PR-STV (though it's far from being the first thing I'd place on the agenda). PR-STV is certainly an elegant system that allows voters to express their preferences in a very fine-grained and expressive way. Increasing constituency size might counter some of its problems. Having a big list of candidates, but only a limited range of parties to choose from, must accentuate the personality rather than party based aspect of PR-STV. There is a certain suspicion that PR-STV itself may be part of the problem (perhaps contributing to "amoral localism": the term coined by the late Peter Mair). It's hard to be certain though. It's an unusual system. If a dozen other parliaments used this system, then an informed judgment call could be made. As it is only Malta (with its tiny parliament of less than 70, just two main parties, an MP workload of about 1/4 of our own and quite a clientilistic politics) is about the single only other really relevant data point. That's far from enough to rule anything in or out.

A More Party-Oriented Electoral System?

Any move away from PR-STV, which at the same time maintains or increases proportionality, almost inevitably would involve a voting system that is more party-oriented in some way. Am not sure I entirely buy the argument, either, that PR-STV is little different to some forms of open list system. A difference is that list systems force a voter to commit and to pin their colours to a single party mast. That's true even for the Finnish list system where only the number of GE votes for individual candidates determines final party list electoral order (but not quite true for Swiss panachage, but then again so many other things about the Swiss system are unusual making a direct comparison difficult). PR-STV allows one to avoid a clear-cut commitment to any one party. In all electoral systems there's a pull to the centre. But, in a list system, not all the parties can fit there, which probably encourages parties to specifically target subsections of the electorate in terms of particular policies. That happens too under PR-STV to a degree, but being inoffensive (populist even) pays off in garnering down-the-line transfers, which can be crucial in getting a candidate over the line. "Harder" and less generally appealing parties do less well under PR-STV. There's an argument that the limited proportionality of our current PR-STV setup, combined with the payoffs of being generally inoffensive in this system, has partly contributed to some of the clientilist policy-lite politics we suffer from. It's a contentious point, however, and there simply isn't the data to definitely say either way.

There's also the question of whether some degree of party say in who gets elected might not be an entirely bad thing. Many relatively well-run European democracies have elements of party control in determining the list order for who gets elected for their party. There's more than one type of party control though. I much prefer the German electoral law setup where all party members in a region get to determine the party list order for that MMP regional election via secret ballot (takes control away from the party leadership). Party control in this scenario becomes more acceptable. Direct control by party headquarters would probably be anathema to the Irish public. Grass-roots party control might stand a better chance. There are various ways to introduce party involvement (semi-open lists, MMP etc.). The Sinn Féin proposal in their election manifesto (point e on p.34) to elect 2/3 of TDs by six-seater PR-STV and the remaining 1/3 by party list would be a quite natural way in our current setup to move partially away from PR-STV and introduce a limited proportion of party-based candidates. This might be more sellable to the electorate if German-style control of list order by ordinary party members via secret ballot was a constitutional requirement

Some of these changes would involve some constitutional recognition of political parties. The current constitution makes no reference whatsoever to parties (given how we use PR-STV there's no real need). I wouldn't see this as being a big problem (it'd be merely recognizing the actual reality and wouldn't likely make a big practical difference). The German basic law (article 21) achieves this in just two paragraphs (recognizing the role of parties and requiring them to conform to basic democratic norms, though the ability it grants to their constitutional court to ban political parties outright, understandable given their past, would probably be overkill here).

On whether to move away from PR-STV, the Finnish list system probably makes a good comparison. This is a list system with very little party control. A voter's choice is far more limited however. A voter can merely pick a single party and then a single individual within that party list. Seats are allocated proportionately to parties according to their overall vote (in potentially very big constituencies, up to 25 or so seat constituencies have been used) to the party candidates who get the most personal votes. A voter is forced into choosing a single party and, even within that party list, a single candidate. There's a complete absence of the fine-grained expressiveness of PR-STV but none of the ambiguity either. I suppose an intermediate system, lying somewhere between our own PR-STV and the Finnish list system, would be to force voters to first opt for a single party and only then use PR-STV as usual for the candidates within that party list. This would certainly allow far bigger constituencies (can't see why 15-20 seat constituencies wouldn't easily then be possible). One would merely need to rank in order the perhaps 20 or so candidates on the single party list one opted for. The Finnish system merely goes a step further in that even the internal candidate choice within a party is made starkly all-or-nothing (a voter merely ticks a box beside one single candidate). Would the Irish electorate buy a move in this direction (maybe via other methods: open-list variants or some open-list variant of MMP)? Probably not! 

But, if they did actually buy into a more party-oriented system, then I suppose the next question that naturally arises would be whether parties themselves should have some element of say into which of their candidates gets elected. I suppose the somewhat peculiar and rather forced Australian Senate "above the line" variant of PR-STV would be one way of shoehorning some party control onto PR-STV. In the party list version of PR-STV above, that might include the extra option of ticking a single box to opt for the party's official PR-STV ordering for the candidates, which could be listed right there on the ballot paper, rather than having to individually rank the party candidates oneself. Or one of plethora of other electoral systems could be chosen to similar effect (semi-open lists, some suitable version of MMP etc.). Have strong doubts, though, as to whether the Irish public would ever buy this, even with an accompanying constitutional insistence that party grass-roots determine the candidate ordering. There are pluses and minuses to all such arrangements. Greater party control might blunt rivalry within a constituency between candidates of the same party. In an ideal world such TDs might then spend more time at national level legislating and not at parish-pump work. There's indeed some evidence that TDs who don't have a rival of the same party in their constituency do spend less time at constituency work. But, of course, an alternative possibility is that this time might instead be merely spent (in a non PR-STV setup) sucking up to party members. And TD accountability to voters becomes more indirect being partially shielded behind the party ticket. A counter to that is that Irish party organizations often resemble a loose web of local fiefdoms with volunteers motivated more by personal than overall party loyalty (think of Bertie Ahern's "Dromcondra Mafia"). Is that really such a good state of affairs? Party control is not a straightforward topic!

Non-Geographic Constituencies?

A different strategy entirely would be to keep PR-STV itself but depart radically from the usual geographical constituency organization. I do find some ideas for non-geographical PR-STV constituencies rather attractive, e.g. see a previous article on the politicalreform.ie blog or a past article on the thejournal.ie website. Daniel Sullivan also had an interesting exploration of this and related electoral ideas up on his blog page. Voters could be randomly assigned to constituencies on reaching voting age, or perhaps, in an administratively easier fashion, according to some non-geographic personal attribute (like final digit of year of birth, or month of birth as suggested by the thejournal.ie article above). Probably several steps too far into the exotic though. And would be tricky enough to make manageable for general elections, if not impossible. It could perhaps form a very viable alternative, though, for electing a modest-sized second chamber on a non-geographic basis (a more natural alternative to extending the rather artificial university and vocational panel system to a more universal franchise).

Just Stick with PR-STV?

A sensible reform strategy might be to stick initially purely with changes to the PR-STV setup and more general issues such as an Electoral Commission, initially just making the argument for greater proportionality, and only then later revisiting non PR-STV options in further referendums, maybe with a PR-STV (the irony! ;) ) preferendum amongst various other electoral options, with a final referendum on the winning choice. If an initial referendum on increasing PR-STV proportionality failed, then there would probably be no further point in trying anything else anyway. And, of course, there's the perennial dilemma of why would larger parties even vote in the first place to increase proportionality in a system that's already skewed in their favour?!

The electoral system chosen for a country doesn't seem to make quite as big a difference to the overall system as might be expected. Maybe that's an argument in Ireland for leaving well enough alone and sticking with what we have (and perhaps merely just increasing PR-STV proportionality). But one can just as easily turn that argument on its head. For all PR-STV's elegance, it's still possible it might be part of the problem. We could change to some boring middle-of-the-road list system, with an intermediate balance of voter and party control over list orderings, as is common across much of Europe, including in many quite well-governed countries. This might well not solve much (if PR-STV really isn't a problem), but with suitable care and having a mind to the stability of government formation (see the next page on the electoral setup), this likely wouldn't cause too many problems either. And we'd no longer be an outlier with a rare and unusual parliamentary voting system. And I suppose we'd no longer be able to blame PR-STV (justified or not).

But, in summary, immediately plunging headlong straight into an exploration of the more subtle pros and cons of PR-STV probably wouldn't be a good idea in a constitutional overhaul. I personally buy the argument for proportionality. Have more mixed feelings, though, on arguments for moving the electoral system in a more party-oriented direction. Non-geographic forms of PR-STV may become increasingly feasible as communication technology continues to improve, but this approach is probably a step too far for the moment (for the large scale of general elections anyway). There seem to be few definite answers in respect of voting systems and plenty of ifs, buts and maybes. PR-STV is nonetheless one of the really distinguishing features, for good or ill, of the Irish system. There are no easy answers.