Voting theory is a big subject, with a mathematical foundation, (e.g. see http://www.nuff.ox.ac.uk/Politics/papers/2002/w12/NewTheory.pdf ) but we don't need to go into all that.   We know the basis for our election counting will be the single transferable vote.  Here is a good account of that:

In the nineteenth century, Thomas Hare in Britain and Carl Andru in Denmark independently invented the core principles of the system. STV uses multi-member districts, with voters ranking candidates in order of preference on the ballot paper in the same manner as the Alternative Vote (see Alternative Vote). In most cases this preference marking is optional, and voters are not required to rank-order all candidates; if they wish they can mark only one. After the total number of first-preference votes are counted, the count then begins by establishing the "quota" of votes required for the election of a single candidate. The quota is calculated by the simple formula:

 


votes
Quota = _________ + 1
seats + 1

 

 

The first stage of the count is to ascertain the total number of first-preference votes for each candidate. Any candidate who has more first preferences than the quota is immediately elected. If no-one has achieved the quota, the candidate with the lowest number of first preferences is eliminated, with his or her second preferences being redistributed to the candidates left in the race. At the same time, the surplus votes of elected candidates (i.e., those votes above the quota) are redistributed according to the second preferences on the ballot papers. For fairness, all the candidate's ballot papers are redistributed, but each at a fractional percentage of one vote, so that the total redistributed vote equals the candidate's surplus (except in the Republic of Ireland, which uses a weighted sample). If a candidate had 100 votes, for example, and their surplus was ten votes, then each ballot paper would be redistributed at the value of 1/10th of a vote. This process continues until all seats for the constituency are filled.

Put even more generally and with less exactness, candidates are either "elected" or "excluded" to reduce the remaining field, and people who voted for those elected and excluded candidates get a extra chance to influence the final outcome.

How this basis is adapted to the IBM constraints is not recorded publicly anywhere (except here!) but Electoral Reform Services have been helpful in providing details.

Essentially the process described above is carried out but candidates cannot be elected where that would break the constraints.  However, candidates who are unelectable solely for that reason are not "excluded" in the sense given above, and the secondary choices of their supporters do not come into play.

So, possibly, after the point when three employees have been elected the process will continue until there is just one retiree candidate left, but the further employees will not be excluded (unless they come bottom of the poll at some iteration), they will just be unelectable.

The fairness of this is arguable, but life is not fair and no voting process can meet all the ideals of a voting system.  It is left as an exercise to the reader to work out voting tactics for the particular case.

 Index of election pages


Home