[EM] Voting mathematics in Irish system (STV)

[James Gilmour]

David GLAUDE drew attention to:
> http://www.math.nus.edu.sg/aslaksen/papers/voting.pdf

This is an interesting paper that deals with some of the peculiarities of the Dáil Éireann
implementation of STV-PR.  Many of the peculiarities arise from the use of random selection of
ballot papers when surpluses are transferred, but these are not the main concern.

The paper contains two small mistakes.
Candidates names in Northern Ireland Assembly and District Council elections are listed in
alphabetical order of family name, not by party.
To be elected, a candidate's vote need not exceed the quota;  the Dáil Éireann rules say "equal to
or greater than the quota", as is usual in STV-PR.

The main problem with this paper, however, is that the authors do not appear to appreciate that the
use of the 'last parcel transfer' procedure is not just a convenient simplification.  There is a
fundamental difference in the philosophy of representation between those who advocate the simple
'Gregory method' of dealing with consequential surpluses and those who advocate the more inclusive
approaches such as the 'weighted inclusive Gregory method'.

The authors present the issue only as one of convenient simplification and as a "problem" to be
fixed by adopting some more sophisticated arithmetic.  This is how the issue was also raised in the
Local Government and Transport Committee of the Scottish Parliament during discussion of the Bill
that will introduce STV-PR for future local government (council) elections in Scotland.  But there
is much more to it than that.  For a discussion of the issues involved (in the context of the Bill
before the Scottish Parliament) see this downloadable (PDF) Briefing Note:
 http://homepages.phonecoop.coop/James.Gilmour/STVTransferRules.pdf

James Gilmour