Even back before computers, in one of the severest tests of STV, the 1925 Seanad election saw 19 elected in one contest using STV. An average force of 43 men worked 14 days to count the 315,000 votes and conduct the 64 counts needed to choose the 19 winners. (the count was held up by the fact that no votes were counted until they were all gathered at one central point in each district. there was no immediate poll-by-poll results, as Canada uses. It took a full week for the votes to be gathered at the central points.)
The ballot dispaying 76 names was a formidable item - but each voter was allowed to mark as few as 12 preferences and still have their ballot considered. (As far as I can tell, 12 marked preferences was the minimum allowable.) [I wish I had noted where I saw that - now I can't recall where I saw the minimum of 12 marked preferences being required!]
Low turnout 300,000 voted out of 1.3M eligible voters. Some of these stay-at-homes was due to a conscious decision - the Sinn Fein boycotted the election.
Info on this landmark election contest can be found at
and
and
https://www.jstor.org/stable/1945103 (Gosnell, An Irish Free State Senate Election)
see also
D. O'Sullivan, The Irish Free State and its Senate (London 1940, repr. 1972)
C. O'Leary, Irish Elections 1918-1977: Parties, Voters and Proportional Representation (Dublin 1979)
J. H. Humphrey, 'Irish Senate Election', Representation - the Journal of the Proportional Representation Society Vol. 15, no. 43, p. 21-30 (December 1925)
Irish Times; Iris Ofiguil; Dáil Eireann Debates; Seanad Eireann Debates (all July-September 1925)
Three parties ran
Cumann na nGaedheal
Labour Party
Farmers' Party
(Cumann na nGaedheal did not identify itself on the ballot, I believe).
(others were "ex-unionists" but they fared badly (perhaps electing none) and I don't see which of the "independents" or "others" they were.
Five were elected by the 54th Count; one in the 62nd Count.
13 in the last count, the 65th Count.
All winners but one were in winning positions in the 1st Count, so the transfers, conducted over 64 counts and which took three weeks to conduct, had very little impact.
Elected were
Toal elected in 47th Count CnaG
O'Farrell elected in 51th Count Labour
Kennedy elected in 54th Count CnaG
Cummins elected in 51th Count Labour
Hickey elected in 51th Count label uknown
Fanning elected without quota declared elected in 65th Count label uknown
Barniville elected without quota declared elected in 65th Count label uknown
O'Hanlon declared elected in 65th Count Farmers
Bellingham declared elected in 65th Count label uknown
McGuinness elected without quota declared elected in 65th Count CnaG
Dillon elected in 62nd Count Farmers
Bennett elected without quota declared elected in 65th Count CnaG
Foran declared elected in 65th Count Labour
O'Connor elected without quota declared elected in 65th Count CnaG
Lenihan elected without quota declared elected in 65th Count Farmers
Couniham elected without quota declared elected in 65th Count CnaG
Parkinson elected without quota declared elected in 65th Count CnaG
O'Mara elected without quota declared elected in 65th Count CnaG
(four who were not elected)
Bigger elected without quota declared elected in 65th Count Label Unknown
(53 less-popular candidates)
elected
CnaG = 8 elected Toal, Kennedy, McGuiness, Bennett, O'Connor Counihan Parkinson, O'Mara
Labour = 3 elected O'Farrell, Cummins Foran
Farmers' = 3 elected O'Hanlon, Dillon, Lenihan
Others = 5 elected Hickie, Fanning, Barnieville, Bellingham, Bigger
Party First-p votes Percent Seats Seats
votes of vote due won
Cumann na nGaedheal supporters 126,218 41 8 8
Labour Party 46,776 15 3 3
Farmers' Party 42,785 14 3 3
Independent* 65,230 21 4 5
Unknown affiliation* 24,69 8 1
Spoilt votes 9,466 –
Total 315,167 100
Electorate/Turnout 1,347,195 23 percent
Source: Nohlen & Stöver
*Independents and unknown affilation had no obvious and certain cross-candidate solidarity so these categories shoud not be considered to be parties where votes slide seamlessly from one candidate to another.
======
The seat counts per party are very proportional. Even the overall seats for the Independents, who each ran separately, are fairly proportional!
Hoag and Hallett (Proportional Representation (1926)) say of the 1925 Seanad election:
"the Irish Free State, after ample experience with P.R. in local elections, did not hesitate to prescribe it for senatorial elections with the whole country as one election area. The first such election took place Sept. 17, 1925. The count was of course the largest Hare (STV) count ever conducted. The number of ballots cast was 315,167. There were 76 candidates, all but 19 of whom had to be eliminated one by one. [they may not have had to be eliminated but because the seats did not fill earlier, it ended up that all but the winners had to be eliminated in this case.]
Yet the central count with an average daily force of 43 men, occupied only 14 days.
There was a preliminary sorting of ballot papers according to first choices before the ballots were assembled at the central office in Dublin.
The central count began in Dublin on Sept. 25th and ended Oct. 10.
The force did not work Sundays.
Only 12 men were involved in the first two days, then 44 men the next six days, and 53 for the last six days."
If people are frustrated by the long period of counting caused by use of STV, they say "it is better to wait a little for the right results than to get a wrong one in a hurry." (p. 152-153) (see also p. 249-251)
The last transfer on the 64th Count entailed the transfer of Fitzgerald's 6000+ votes. so this explains why 53 men were needed at the end.
================
another write-up on the 1925 election
I came across the largest-DM STV election up to 1990.
It was back in 1925 Ireland (Irish Senate)
no computers but 19 were elected
300,000 votes were counted across the whole country.
76 candidates.
each ballot was like 14 inches by two feet
14 days to count --
the results announced more than a month after the Sept. 17th election --
but as two leading proportionalists Hoag and Hallet wrote back in 1926 - it is better to wait for correct results than get wrong result in a hurry.
The long wait was partly due to use of the super-exact Gregory method for transfer of surplus votes. (see footnote)
but looking at the candidates in order of popularity in the first count, we see that 18 of the front runners in the first count were elected in the end.
only one came up from lower part of list to take a seat.
transfers had very little effect -- the proportionality was achieved in the first count (in what would be SNTV if it stopped there)
But the transfers, if nothing else, did concentrate the votes in such a way as to prove that the 1st-Count front runners (in total) were in fact generally and widely supported by the voters, making only one change where someone else was more widely popular.
Independent Bigger (initially the 23rd choice) replaced Farmer's Butler (19th in the 1st Count)
but Farmers' party was not excluded -- it did elect three members so Bigger's success did not shut out Farmers from rep.
three parties elected members
plus Independents were successful.
from Wiki
Results Party FP votes Percentage Seats
Cumann na nGaedheal supporters 126,218 41 8
Labour Party 46,776 15 3
Farmers' Party 42,785 14 3
Independent 65,230 21 5
Unknown affiliation 24,69 8
Total 315,167 100
Electorate/Turnout 1,347,195 23.4
Spoilt votes 9,466 –
Source: Nohlen & Stöver[34]
Non-transferable at the end 37,704
(from Wiki "1925 Seaned election"
Effective votes (used to elect someone) 260,000 85 percent*
Votes where the first-preference choice was elected 160,000 52 percent
With nine elected with full quota of 15,000 each and the other nine elected with vote tally near to quota, much more than three-quarters of the votes were used in the end. Nohle and Stover apparently felt you can compute party proportional analysis using STV!
And how proportional it is -- even Independents who ran independently got about their overall due share of seats.
first count and 64 counts of transfers
some transfers were from eliminations; and some were transfers of surplus of the first candidates to be elected.
first elimination caused transfer of only 334 votes at most.
last elimination might have caused transfer of as many as 6000 votes or more.
the number of votes that were used to actually elect someone is not recorded in sources seen but I calculate that is about 85 percent.
and of course some votes not used in the end may have seen their earlier choices elected, just without the help of the vote itself. The bulk of a particular vote might have been considered surplus, for example.
due to 14 day wait, and likely no other reason except perhaps its utter fairness -- yes some people don't like such fairness -- the 1925 election was the only country-wide contest for decades.
thereafter STV elections were broken up into separate districts (MMDs) and Senators were elected indirectly by elected members of other bodies.
But it proves the STV system can work without computers even to elect 19.
Changing to whole-vote STV, like had been brought in to elect MLAs in Winnipeg, Edmonton and Calgary just a few years before this landmark Irish election, might have allowed Ireland's whole-country experience to continue - without necessarilly any negative effect on proportionality - most P was achieved in 1st Count anyway.
(all but one winner was in winning position already in first count.)
After 1925, Ireland switched to electing the Senators indirectly by the lower house.
but at least Ireland did not simply go to FPTP to elect the lower house.
It used STV in districts of 3 to 6 or so seats.
and the Republic of Ireland's relative peace may be due to that fairness
- unlike Northern Ireland, which did go to FPTP after short STV experience in early 1920s, the PR-using Republic did not suffer the same level of violence.
This comparative history analysis is corroboration of the PR-gives-civil-peace pattern that political pundits and historians have observed.
----------
see also
Footnote: Gregory method :
This modified approach, invented by J. B. Gregory in Australia in the 19th Century and now commonly known as the [Irish] ‘senate rules’*, was adopted to give effect to certain of the provisions of the Government of Ireland Act, 1920, where provision was made for the election of certain senators from very small electorates.
It provides that, with a view to reducing the random element when ballot papers are selected for physical reallocation on distribution of a surplus, all relevant ballot papers [just last parcel - it is "Gregory method", not "Inclusive GM"] be reallocated, but at a value reduced to take account of the ratio between the number of transferable papers and the surplus.
*not to be confused with the Australian Senate rules, which are Inclusive-Gregory Method (IGM). (Farrell and McAllister, Aus. Electoral Systems, p. 60)
=============================
Comments