Political Reform‎ > ‎

The Electoral Setup - Other Issues and Complications

Rights, the "democratic architecture", and certain aspects of direct democracy are all categories that can be considered relatively independently of other constitutional changes. That's not really possible, or probably even wise, for the electoral system. And there's more to the electoral setup than just the voting system.

Proportionality and Stable Government Formation

Increased proportionality isn't inevitably a good thing if certain measures aren't put in place to counteract potential instability in government formation. I've a lot of sympathy for efforts by portions of the Italian electorate to strike proportionality out of their electoral system. Italy has a limited form of citizens initiative: the abrogative referendum. It allows voters to strike down laws or aspects of laws, but not to make any positive proposals. In the absence of an ability to introduce stabilizing measures I suppose it makes sense for voters to try to reduce proportionality. An abrogative referendum succeeded in the past in striking some proportionality out of the system. Relatively recently, an abrogative referendum came within a narrow whisker of striking out most of the remaining proportionality.

One common stabilizing measure, in the context of a list system, is the presence of a threshold; this typically blunts proportionality by not allocating any representation to a party that fails to meet some set share of the vote (perhaps 3%). With PR-STV, the combination of a German-style constructive vote of no confidence mechanism (i.e. a vote of no confidence is only allowed if the name of an alternative Taoiseach is attached who becomes the new Taoiseach if the motion succeeds) and semi-fixed (perhaps 5 year) terms (modeled after the Swedish Riksdag setup rather than the more inflexible and artificial German approach) would help ensure stable governments even with the injection of extra proportionality.

From page 86 onwards of the 1996 constitution review group report gives a suitable wording for a constructive vote of no confidence in the context of our own constitution. In Sweden, ordinary parliamentary elections are always held exactly every four years. However, the Prime Minister also has the option of dissolving parliament and calling midterm extraordinary elections. But a parliament formed in this way can then only last up until the next scheduled date for ordinary elections. This incentivizes parliaments to last their full term but is flexible enough to also recognize that sometimes early parliamentary dissolution is a necessary evil.

Such measures in conjunction with larger PR-STV constituency sizes and an Electoral Commission would roughly form the type of overall electoral package that I'd feel would most likely garner support in a referendum. I suspect the Irish people are quite comfortable with PR-STV. Some increased proportionality might counteract some of its downsides (and I feel there are some downsides).

The Voting System and Parliamentary Alternatives

There are also probably certain aspects of the legislature-executive relationship that might need to be considered at the same time as a look at the electoral system (some elements cannot be neatly disentangled). Iceland proposes to bring in parliamentary alternatives. An MP would list an alternative when standing for election (or perhaps a list of names like our own European elections). If the MP later becomes a minister or speaker of parliament (or resigns) he has to give up his seat and his alternative takes his place and vote (temporarily anyway). That's a reasonable proposition in a list system (strengthens legislature-executive separation). With PR-STV and strong dynastic tendencies here, I'd feel the exact same system wouldn't be such a good choice here. One could instead simply take the next candidate in the PR-STV list in the constituency (the last to be eliminated) as the replacement. But that might not be a government candidate. That would strongly discourage a government from picking ministers from constituencies where the next-in-line wasn't a government candidate (cutting down on the ministerial talent pool).

Or one could require that the next government party candidate in line in the constituency would be the alternative (some constitutional recognition of parties would seem necessary for such mechanisms). Choosing the next government candidate in line does also have downsides. It is inevitably promoting a ministerial constituency rival to TD status. That's not likely to discourage ministerial pork from finding its way into the constituency as a minister keeps a very sharp eye on his own rival. My preferred solution so far takes partial inspiration from a previous post by Daniel Sullivan on politicalreform.ie. My variant would be to keep track of the elimination counts of all candidates in a GE. If a candidate has 2000 votes when he is eliminated this is recorded (along with his party affiliation in the GE). My parliamentary alternative scheme would simply rank all failed GE candidates from the same party according to this count (using their party affiliation at the time of the election, ignoring whether or not they were even still affiliated to the party). If a TD is made minister the next candidate in the list becomes his replacement. This scheme rightly favours candidates from smaller magnitude constituencies and also constituencies that are somewhat more under-represented than average. Constituencies with a higher turnout would also be favoured. It can cope with unequal constituency sizes fairly as far as I can see. It cuts the local connection (a quirk is that some constituencies might gain an extra TD, but this is more likely for smaller magnitude constituencies which don't have particularly proportionate representation anyway). Probably less than half of ministers would end up with promoted constituency rivals, ameliorating some of the extra temptation to ministerial pork. Plus there's no direct party control involved and it doesn't bias who might be chosen as a minister (there's no control over who'll be the alternative so they might as well choose whoever it was they wanted in the first place as minister anyway). 

Parallel and Top-up Voting Systems

Plus there's no real reason why all this (and very much copying the Daniel Sullivan type approach from the link above) couldn't be extended to the entire voting system. If, like the Sinn Féin GE proposal, some portion of TDs, 1/3 in the Sinn Féin manifesto, were to be elected by party list, then voters, as well as the usual PR-STV ranking, could be asked to tick a box next to a single party. Ultimately parties would be allocated a number of seats proportional to their support in this portion of the vote. The above system (using elimination counts) could be used to decide which candidates ultimately got the party seat allocation.

If there are two routes to getting elected (say PR-STV and a list), a further important design choice that arises is whether to allow candidates to run in only one or in both electoral routes simultaneously, and whether the list acts in a top-up compensatory relationship to the first list or runs completely in parallel; the resulting dynamics can be quite different depending on these choices. Parallel voting systems can allow candidates high in a party list a relatively easy ride. Is that a good thing? This has its upsides. Ministers might have to be less worried about re-election where the list is to some degree party controlled, leading to less pork perhaps, but maybe also leading to less direct accountability to the voters (and more to party "selectorates" even if they consist of the party grass-roots).  This also likely tends to foster party cohesion, but what kind of party cohesion? And again is this necessarily a good thing? The MMP (mixed-member PR) voting system has an element of this, but if a big party does particularly well in constituency contests it's possible it may be left with no allocation of list seats whatsoever (even though if a version of MMP was used with a greater proportion of list seats than constituency seats, then there would always be some party "safe" seats).

Proportionality and Presidential Systems

One good way of dispersing power would be to opt for a presidential type system. Donal O'Brolcháin has written much in favour of having a Rialtas-Dáil model with a far greater separation of powers. Michael Williams also wrote quite well in favour of such a model in an article in the Spring 2010 edition of "Studies", the Irish Studies magazine produced by the Jesuits. Can't say I often buy this, but my local magazine shop actually stocks it, and this article (and another by Eoin O'Malley) in this issue looked quite interesting.

It has occurred to me though that there is perhaps a certain issue with how easily such a model might sit with greatly increased electoral proportionality. The Israelis previously experimented in the 90s (not entirely successfully) with such a model in conjunction with an almost maximally proportionate electoral system. There seems to be an inherent tension between giving stronger powers to the executive when more weakly counterbalanced by the more splintered legislature engendered by a more proportionate electoral system. The Israeli directly-elected PM was relatively weak (a 2/3 vote not only overrode his decisions as in the US, with its almost non-existent electoral proportionality, but could remove him entirely, plus he was far more constrained in his freedom to choose a cabinet). But would it really have been practical to give him the powers of a US president given the somewhat unstable nature of the Israeli parliament? Though there isn't huge proportionality in our own current electoral system, so an appropriately tailored directly-elected Taoiseach model might work fine. There'd probably be even scope for a modest extra injection of proportionality (maybe the minimum constituency size could be increased to five). There's quite a range of variations on the basic model amongst the various US states so no doubt some appropriate calibration of legislative versus executive power could be found to match the electoral system. Though I can't say I fully understand the issue or am fully au fait with the Israeli experience.

Dáil Size

There's a lot involved in the electoral setup category. Dáil size would be another important aspect of the electoral setup. According to the polsci cube root "rule", a Dáil size of around 166 is about bang on for our population (would not count the Seanad in this, it's almost never used as a source for cabinet ministers or much else). It seems that countries tend to have parliament sizes approximately clustered around the cube root of their population size. It's not a simple linear relationship. Otherwise one could take some very small countries like Iceland (with about 5000 people per MP) and Malta (around 6000 per MP) and argue on this basis that Ireland should have a Dáil size of 750-900! Or, conversely, one could take a very large country like the US (around 580,000 per member of Congress and Senate) and argue that Ireland should have a Dáil size of 8 ! Obviously, the relationship between population size and parliament isn't a simple linear one. But, even on purely linear terms, our current representation levels would be fairly comparable to many other similar-sized small countries, e.g. Sweden and Norway and the other Scandinavian countries. There's always an exception though! New Zealand has a particularly low representation level for a country of this size. Its parliament has only 120 members. Linearly scaling to Ireland would give a Dáil of just 124 TDs.

In terms of the more sophisticated cube-root approach, the UK would be heavily over-represented in terms of this "rule"; extrapolating their level of representation purely from the House of Commons, and not even accounting for the House of Lords, using this rule would give an equivalent of a Dáil of around 250 (the precise final figure depending on how exactly one does the extrapolation). On the other hand, the US is very much under-represented according to this rule (extrapolating their level of representation to here would still give about 130 however). However their cabinet is not chosen from Congress or the Senate.

Slashing Dáil size would likely be very popular here. The New Zealanders perhaps do manage to get away with a much smaller parliament. Could we? I've my doubts. There are many examples of FPTP (first-past-the-post) and list systems, so I suppose the relationship between parliament and population size may be better understood in these cases. While there are not that many examples of MMP systems like New Zealand's, it still is really after all just a FPTP setup topped up by a list. Perhaps Ireland with its unusual PR-STV system should err on the higher side of caution (plus almost non-existent meaningful local government in Ireland means TDs have also taken on extra duties that might ordinarily be performed by local councilors).

So I'd be against any drastic reduction in Dáil size here unless it was accompanied by a requirement (not just an option) that several ministers be chosen from outside the Dáil (and ideally also proper local government). Or if there's a meaningful second house, then I suppose it wouldn't really matter how representation is split between the two chambers, as long as the second chamber is also used as a ministerial talent pool (with perhaps hard minimum requirements for the number of external or Seanad ministers). Dáil size, having an adequate ministerial talent pool and the procedures for ministerial appointment are all strongly bound up together and cannot really be considered in isolation from the electoral system.

That probably touches on most aspects of the electoral system and some factors needing be borne in mind if altering the electoral system (though it wouldn't surprise me if there are several more that haven't even occurred to me yet).