top of page
Tom Monto

Many exaggerate the effect of STV vote transfers and their complexity

Updated: Oct 31, 2021

As Meslin mentioned in a recent TVO zoom call broadcast on ranked votes, it is not necessary to make voters mark back-up preferences under STV (or in the single-winner version, Instant-runoff voting).


if it has no back-up preference, a vote is more likely to be wasted but that would be the voter's choice.


And even when a vote bears many back-up preferences, it may still be wasted by becoming exhausted.


It is not necessary but if a voter does mark back-up preferences their vote is more likely to be used.


Even with a lack of or low number of marked back-up preferences, STV still has higher rate of effective votes and lower rate of wasted votes than single-winner FPTP. The selection of the exact mathematical method used to formulate transfers - whether the random "Andrae" or "Hare" method, the Gregory, the Weighted Inclusive Gregory, etc. - often has no importance because in most cases transfers (however done) have no effect on the order established in the first count. This is due to any or all of these reasons -- the transfers (potential or actual) in the vote counting merely reinforce the existing order, or transfers are not actually done (due to legitimate reasons outlined below) or even when done do not use math.


Scotland uses Weighted Inclusive Gregory in its STV elections. The final result is often the same as the result in the First Count before any transfers are done.

In the STV used in the Scottish Local Government elections of 2007, only 35 (18%) of the three “front runners” in the 190 three-member wards were not elected and only 20 (12%) of the four “front runners” in the 163 four-member wards were not elected.


Alberta provincial elections that used STV (eight, from 1924 to 1956) elected 90 MLAs in Edmonton and Calgary. Of the front runners in the first count, all but 14 in Edmonton and seven in Calgary were elected in the end.


Manitoba provincial elections that used STV (nine, from 1920 to 1954) elected 94 MLAs in Winnipeg. Of the front runners in the first count, all but 12 were elected in the end.


Thus, in Alberta and Manitoba provincial elections -- and Scottish local elections in 2007, back-up preferences and transfers usually made no change to those who would have been elected under SNTV anyway.


I would say that the number of seats in a district is more important to fair representation and to having a low number of wasted votes than back-up preferences and transfers.


Single-voting in multi-seats districts means mixed balanced representation, which is not produced by IRV or FPTP.


If waste of votes is a concern, the 80 to 90 percent of effective votes produced by STV is something far above the guaranteed 50 plus 1 in IRV.


Taken another way, in FPTP up to 83 percent of the voters in a district may be wasted.

in IRV, as much as one less than 50 percent may be wasted.

in STV, usually no more than 10 to 20 percent are wasted, depending on number of number of seats in a district.


STV does this despite many back-up preferences not being considered. Vote transfers are not actually that important.

Making voters rank each candidate when many/most back-up preferences will not be considered anyway seems a waste of energy to me.


concentrating on how vote transfers are done seems a waste of energy to me also, when the result of STV elections is entirely or mostly the same as the result without any transfers at all.


In all STV elections, most of the front runners in the first count are elected in the end after transfers. Sometimes all the front runners are elected in the end. So the result under SNTV would be the same, or almost the same, as under STV. SNTV can be thought of as multiple plurality but where each voter casts just one vote.

(The effect of transfers is often merely to prove that indeed all or most of the front runners did have the right to be elected.)


Back-up preferences and transfers are meaningless under many circumstances.


A back-up preference marked for a candidate already elected or eliminated is un-usable and ignored. It is put on the exhausted pile. The votes marked for an already-elected candidate are where exhausted votes overlap with votes cast for elected candidates. I suggest use of two piles: "exhausted but satisfied" and "exhausted and not used". The first of these piles would tell you how many voters did not have their vote used but did see a favoured candidate elected.


Back-up preferences marked on all votes cast for those who win in the end are never considered, except to determine where the surplus votes above the quota (usually a small percentage of he elected candidate's votes) should go to. Other than that, back-up preferences are considered when the candidate marked as the first preference is seen to be un-electable and is eliminated. Votes placed on candidates who are electable do not have their back-up preferences consulted (except to transfer the surplus).


Surpluses are not transferred when they would have no effect, either because they are too few (compared to the distance between any other pair of candidates) or because the last seats have been filled or are about to be.


Even in Australia STV elections people do not need to rank all the candidates but can avoid the full-ranking rule by choosing a party slate and giving no preferences to candidates of any other party (at least that is my understanding).


Canadian practice was to allow voters to mark as many or as few preferences as they wanted, either along party slates or crossing party lines. full liberty. Canadian STV elections did show more exhausted votes than Australian ones, although some do show up in Australian elections (at least that is my understanding). and still Canadian STV elections produced balanced crops of reps and high rates of effective votes.


Because votes may run out of marked back-up preferences while still being eligible for transfer, the exhausted pile grows.


Even when surpluses are to be transferred, no math is required when the number of the surplus votes is greater than the number that can be transferred. To explain, if a vote does not have a usable back-up preference, it cannot be transferred. Sometimes the number to be transferred (the surplus) is greater than the number of transferable votes. in those cases no math is used, simple reference to the next usable back-up preference on the few transferable votes is all that is required. No math! So the exact mathematical method used to formulate Many people may not know this but it is true.


Back-up preferences marked on votes cast for the last surviving unsuccessful candidate are usually not considered. This is because the candidate's elimination drops the number of remaining candidates to the number of remaining seats and ends the vote count.


So in STV, back-up preferences are not considered when the vote was cast for a winner (except sometimes to transfer surplus), often when the vote was cast for the longest-surviving unsuccessful candidate, or when the back-up preferences are marked for candidate(s) who have been elected or eliminated when the back-up preference comes up.


So even when votes mark back-up preferences, they are often ignored.


But despite that, single voting in multi-seat districts ensures that most of the voters -- as much as 90 percent -- see their vote go to help elect someone.


Note too that in some IRV elections, someone receives majority of votes on the first count so no transfers at all are done.


If voters and even pundits have an exaggerated idea of the importance of the ranking of candidates past the first preference,

the question then becomes:

do you give in to that by adopting a system that uses as few transfers as possible (such as MMP) or do you give people the other side of the story?


Either is fine as long as you can live with the result.


MMP's use of top-up seats to overcome un-balance is elegant in its own way although using party lists is off-putting to some.


Top-up seats are usually used to address imbalance created by single-winner FPTP district elections.


Top-up seats could also address unbalance resulting from SNTV elections.


Under a SNTV/MMP system, people cast an X and multi-member districts create a semblance of balance (sometimes as much as STV or MMP produces) which is further fine-tuned if necessary by top-up seats.

==================================

4 views

Recent Posts

See All

Comments


bottom of page