top of page
Tom Monto

Purpose of STV vote transfers

Updated: Nov 19, 2023

if voter doesn't like ranked voting, he or she doesn't have to mark back-up preferences under the optional-preferential STV system used in Ireland, Malta and historically in Canada -- if you do not or did not want to mark back-up preferences you do/did not have to. (You just mark your first preference just like in FPTP., except use the number 1 not an X)


Even at that, STV works better than FPTP - in STV you have choice of 15 or more candidates to choose from and choice even within each party slate.


Most votes go to first preferences anyway and most candidates in winning position in the first count go on to be elected, so back-up preferences and vote transfers make little difference.


But what transfers do is either confirm the fairness of the first-count positions or make small changes so that more voters' choices are actually used to elect someone.

Any change in candidates in winning position arises from electing someone more generally acceptable or if the same candidates remain in winning positions, the vote transfers shift votes so more votes go to the people already in winning positions.


Those two different possible outcomes are what drives transfers -

- Which will it be? --

Will an initially-less-popular candidate pass someone initially in winning position

Or not?

Single voting in MMD is the backbone of STV and does not depend on vote transfers.


STV vote transfers do change candidates' vote tallies, and vote tallies at the end are different from vote tallies in first count.


Whether candidates' first-count vote tallies or candidates' votes at the end are used to collate and compose parties' due share of seats, the result is broadly proportional in either case. (I have not seen a case where that is not true - if DM is more than 3.)


I say this because in every election the candidates in winning positions in the first count are of different parties (more than one party is represented) so mixed representation is already established in the first count.


Advantage of MMD and fair voting (STV or district list PR or SNTV) is that each voter has an elected member whose sentiment he or she shares, or someone very similar, actually representing the district where the voter lives.


No SMD can do that. but of course no one is defending FPTP.


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

looking at transfers in Alberta's use of STV (which is fairly typical),

we see:

- the complicated transfers that some people want to avoid were actually few in number -- only two in each election in each city (Edmonton and Calgary) on average. - the few times when those kind of transfers would have been done in Albta use of STV, they were done without fractions by transferring whole votes of just the surplus.

yes I do say --the times when surpus votes are transferred -- tht is when fractional transfers are done in other system andwhen they were done using whole votes in Alberta's experience -- there was little effect on results - This is not reflection of the transfers or the method of transfer but due to the relative few number of times that surpus votes were transferred. in STV transfers are done when: - a candidate is eliminated due to being un-electable - no math is used here except just to move ballots from one candidate to another or when a candidate is elected with surplus votes and then only if - the surplus is of enough voters to make a difference - there are still un-filled seats, and there are two or more candidates stil in running more than the number of unfilled seats. if all three conditions apply, a transfer is done an it can use somewhat complicated math. In Alberta's experience, these conditions all applied only twice each election in each city. and then this was done with whole votes in Alberta, not the fractions as done under Gregory method.elesewhere.


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


The same was noted in the first STV election in a North American city - Ashtabula (Ohio), 1915.

A Proportional Representation League leaflet of 1915 pointed out that less than one percent of valid votes were transferred as surplus votes.

So transfers of surplus votes, in Ashtabula's 1915 elections and many other STV elections, involve relatively few votes so are not likely to have great impact and therefore a super-precise method of such transfers seems superfluous when the amount of transfers are so slight. (the transfers are important so should not be avoided altogether but they are so marginal that excessive attention should not be paid to them.)


Transfers of votes arising from elimination of least-popular candidates on the other hand involve perhaps 30 percent of votes. These transfers are important and also should not be avoided. Here there is no need to worry about how to conduct the transfer - each vote is looked at - the vote is then simply moved to the next usable preference marked on the vote if any.


Half of the valid votes typically are never transferred at all.


Hoag and Hallett's book Proportional Representation (1926) had something to say about surplus transfers.

By 1926 the whole-vote method of transfer was in use in Ireland and Malta.

The Gregory method was in use in Calgary city elections (as of 1917).

other variants of Gregory (WIG, etc.) may or may not have been invented by then.


Hoag and Hallett:

the book speaks of three methods of transferring surplus votes:

random method - three methods

- a candidate who achieves quota does not receive more votes and any additional incoming votes are moved on to next marked preference. Used in Ashtabula, Kalamazoo, Sacramento and Cleveland (and today (2023) in Cambridge Mass.)

- a transfer is composed by taking a nearly identical amount randomly from each precinct.

- the random method used in Cincinnati where each vote is numbered at time of counting and when it is sorted by first preference. if one fourth of votes are to be transferred, then each fourth vote by number is transferred. (Hoag and Hallett say this is more guaranteed random than the previous two methods and is therefore recommended.)


"Exact" method [whole-vote method] -- math reduction is used, based on next usable preference, so transfers and quota that remain with the candidate are proportional. Used in Boulder, West Hartford [and in the British-Irish-Canadian STV systems]

(within whole-vote method. two variants exist -- those that consider all the candidate's votes and those that base transfer on just last incoming parcel of votes.)


Fractional method (Gregory method) - each vote is transferred at fractional rate based on relationship between surplus and total votes held by successful candidate. used in South Africa Senate, Irish Senate, Tasmanian Assembly. [and in Calgary and perhaps other Canadian cities.]


These methods may be used for all transfers or in combination as per different types of quota.


although random methods do produce some chance, Hoag and Hallett wrote:

to "add to the rules of STV considerable complications merely for the sake of eliminating a further infinitesimal trifle from elements of chance that even in the aggregate are known to be very small seemed of doubtful wisdom to the fully-informed and practical men who drafted the standard STV rules of Great Britian, Ireland, and Canada, and to those who drafted the still simpler rules of the U.S. cities. ...

We regard the complications of the South African and Irish Senate rules as not worthwhile for ordinary public elections.

The American count in its "Exact" form [rules of which were described on pages 345-6] we regard as quite as satisfactory, for practical purposes, as the slightly-more--complicated British-Irish-Canadian count.

We recommend for public elections the simple count [form of random method] used in Cincinnati."


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

British-Irish-Canadian STV

Hoag and Hallett provided election rules for the British-Irish-Canadian STV on pages 356 to 375.

British-Irish-Canadian STV used whole-vote transfer method for all surplus vote transfers. (Calgary and perhaps other Canadaino STV systems did not use the usual "British-Irish-Canadian STV" method.)

Surplus votes are transferred in the whole vote method as described on page 361.

The method is different for candidates who have quota on first count from those who get quota later.

If it is later, you would just look at the last parcel of votes that came to the candidate to establish proportions.



U.S. rules used combination of whole-vote transfers and simple cut-off according to Hoag and Hallett.

The simpler U.S. rules had it that if the candidate attained quota in first count, the whole-vote method was used for transfers.

and if the candidate attained quota by way of incoming vote transfers, the candidate would simply stop receiving votes when he or she attained quota. Election officials are to count votes as they arrive so that the cutoff can be done at proper time. (p. 347)

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


uniform quota see p. 348 footnote


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

0 views

Recent Posts

See All

Comments


bottom of page