Adrian Kavanagh, 15th October 2010
When electoral boundaries were being redrawn for City abd County Council electoral areas in 2008, the terms of reference set for the two electoral area committees were changed so that, apart from exceptional cases, the minimum number of seats per electoral area would now be four, and not three as previously the case, with the maximum number of seats remaining at seven. The electoral area committee responsible for Dublin and the cities responded to this by increasing the number of 6-seat electoral areas, while the number of 4-seat constituencies were increased in the case of electoral areas in the rest of the country. See the reports for Dublin and the Cities and for the rest of Ireland. But does increasing distirct magnitude have an impact on electoral behaviour. In this piece, I’ll try to draw some conclusions in relation to this by comparing different constituency sizes in terms of factors such as party support, turnout levels and gender.
In all, there were 42 different electoral areas drawn up by the Dublin and Cities committee (46 electoral areas in the previous report), with no 3-seaters (8 in the previous report), 15 4-seaters (15), 16 5-seaters (16), 9 6-seaters (5) and 2 7-seaters (2). The committee responsible for the rest of the state created 140 different electoral areas (146 electoral areas in the previous report), with 3 3-seaters (24 in the previous report), 50 4-seaters (34), 35 5-seaters (37), 25 6-seaters (27) and 27 7-seaters (24).
When we look at support patterns across the different sized constituencies, we note a propensity for Fianna Fail and Fine Gael to do better in the smaller-sized constituencies. Fianna Fail won an average of 26.1% of the vote in 3/4-seat electoral areas, dropping down to averages of 25.6% for 5-seaters and 24.7% for 6/7-seat electoral areas. Similarly, Fine Gael won an average of 33.7% of the vote in 3/4-seat electoral areas, dropping down to averages of 29.9% for 5-seaters and 32.9% for 6/7-seat electoral areas. Labour also polled better in the smaller constituencies, winning an average of 16.2% of the vote in 3/4-seat electoral areas, dropping down to averages of 16.3% for 5-seaters and 12.6% for 6/7-seat electoral areas. There was no relationship between constituency size and Green support (2.2% average for 3/4 seat electoral areas, 2.3% for 6/7 seat electoral areas), but support increased in line with constituency size in the case of Sinn Fein (6.2% average for 3/4 seat electoral areas, 7.3% for 6/7 seat electoral areas) and the other smaller parties/independents (15.7% average for 3/4 seat electoral areas, 20.2% for 6/7 seat electoral areas). Turnout propensity did not seem to vary significantly with constituency size however; turnout levels stood at an average of 57.7% for 3/4-seat electoral areas, 58.1% for 5-seat electoral areas and 57.5% for 6/7-seat electoral areas.
The really interesting trends that do emerge relate to the linkages between female candidacy/support levels and constituency size for these elections. The larger number of cases (N) for local elections means that such an analysis can pose more interesting findings than one would get with a similar analysis of general election constituencies, with 182 different electoral areas/local election constituencies in the 2009 contest as opposed to 43 general election constituencies used for the last, and indeed to be used for the next, general electoral contests. In a previous posting, it was suggested that increasing constituency sizes had created an opportunity space that allowed for more female candidates and an increase in the number of votes won by females, as evidenced in the example of the three Dublin Inner City electoral areas. But was this an exceptional cases, or one that proves the rule?
In the local elections, female candidates contested 155 of the 182 different electoral areas, meaning that 27 constituencies (14.8%) did not have a female candidate. 18 of these constituencies were 4-seat constituencies (28% of all 4-seat electoral areas), 4 were 5-seat electoral areas (8%), 3 were 6-seaters (9%), and 2 were 7-seaters (7%) – the significantly lower female participation levels associated with the smaller, 4-seat, constituencies being particularly interesting here. In terms of the number of female candidates run per constituency, there is also a notable difference according to district magnitude; 20.2% of all candidates in 6/7 seaters were female against a level of just 17.9% for 3/4 seaters; when looking solely at Fianna Fail and Fine Gael candidates this difference proves even starker – just 14.8% of FF/FG candidates were female in 3/4 seaters while 19.1% of FF/FG candidates were female in 6/7 seaters. It is interesting to note that the percentage share of the vote won by female candidates did not deviate in relation to constituency size to such a dramatic extent; female candidates won an average of 16.2% of the vote in 4-seat electoral areas, 16.8% in 5-seaters, 18.3% in 6-seaters and 16.4% in 7-seaters.
Irish Politics Forum 
Is Ireland the only STV jurisdiction in the world that allows a mixture of odd and even district magnitudes for the same council or legislative house?
Given the importance of uniformity – if not all the same DM, then at least either all odd (Malta, Tasmania, ACT, Victoria) or all even (Aust Senate, Western Australia) – and the need to avoid “Tullymandering” or the appearance of it, there is a case for fudging (slightly) the allocation formula. Rather than, say, a 6-seater district having between 5.5 and 6.5 times the average in population, the cutoffs could be adjusted so that only (say) 4.8 to 5.2 gets 4 seats, 5.2 to 5.8 gets 5 seats, 5.8 to 6.2 gets 6 seats, and so on. This would weight the result overall more towards odd numbers.
(It would even be possible in theory to make the rule “round off to the nearest odd number” – or, if allocating using highest averages, jumping among successive odd numbers only – but this would be undesirable in practice as it would mean 2 seats, not 1, would ride upon a slight increase in population. It is probably unavoidable for a district with exactly 6 times the quotient to have 6 seats, but it is not obviously wrong for a district with, say, 5.7 quotients to have 5 seats or 6.3 quotients to have 7 – both being within a 10% variation).
I assume that to adopt this in Ireland for the Dail would require a Const amendment by referendum, but could it be introduced for local electoral districts by ordinary legislation?
Fascinating analysis. Interestingly, research shows that constituencies should have at least 7 representatives to be considered ‘women-friendly’ and is related to the fact that small district magnitudes tend to favour incumbents and larger political parties, as we all know. Since women tend to be less likely to be incumbents, and are often more likely to run as candidates for smaller parties, it creates a vicious circle inhibiting an adequate increase in female candidacy levels.
As Adrian pointed out, local elections in Ireland provide a fruitful point of study because we have slightly more diversity in constituency size than we do in general elections. Yet a look at comparable figures for Dáil elections is also interesting for the patterns that emerge. Looking at the two largest parties, Fine Gael are more likely conform to the expectations of the academic literature (Prof Yvonne Galligan has also noted this dynamic in her work). In the 2002 election, Fianna Fáil ran female candidates in 31.2% of 3-seaters (5/16), 16.6% of 4-seaters (2/12) and 42.9% of 5-seaters (6/14). Fine Gael ran women in 25% (4/16) of 3-seaters, 25% of 4-seaters (3/12) and remarkably in 57.1% of 5-seaters (8/14) (22% higher that their average). A similar pattern occurred in 2007. FF ran women in 33.3% of 3-seaters (6/18), 23% of 4-seaters (3/12) and 41.6% of 5-seaters. The respective figures for FG – they ran females in 11.1% of 3-seaters (2/11), 38.4% of 4-seaters (5/13) and, interestingly, 66.6% of 5-seaters (8/12)(32% higher than the average!).
All in all, it seems that the larger the constituency size, the more space there is created for women to run for FG. The pattern is slightly reversed for FF who are more likely to run women in 3-seat constituencies than those with 4 seats, but who are still considerably more likely to run women in 5-seaters. Would an increase in district magnitudes lead to more women candidates? A look at the figures for both general and local elections would suggest this to be the case.