[EM] PR in student government

[James Gilmour]

> raphfrk at netscape.net > Sent: 17 April 2007 09:37
> James Gilmour jgilmour at globalnet.co.uk
> > raphfrk at netscape.net> Sent: 16 April 2007 20:08
> > > It might be easier to explain.  The real problem with  PR-STV is
the
> > > fractional transfers.  They are not very easy to explain.
> >
> > Fractional transfers are absolutely essential for STV-PR (unless you
> > accept a small element of chance).  Without the correct transfers of
> > surpluses you cannot get a proportional result.  Some students'
unions
> > in the UK use this corrupted version, but it cannot rightly be
called
> > "STV-PR" because it distorts the proportionality expressed by the
> > voters.
>                                                               
>

> It seems to me that this method is pretty close to picking random
votes
> for the surplus transfers.

I am aware of five different methods of transferring surpluses in STV-PR
and none of them could properly be described as "picking random votes".
That procedure was abandoned a long time ago.  The only rules that do
not involve fractional transfers are the rules used for elections to the
Dáil Éireann.  Those rules do involve selecting ballot papers at random,
but only after the ballot papers have been sorted into sub-parcels ready
for transfer to the next available preference.  A proportionate number
of papers is then selected at random from each sub-parcel.  Thus the
effect of random selection is minimised.  The ballot papers are
transferred at a value of one vote in all circumstances.

The other four methods of transferring surpluses (Gregory Method,
Inclusive Gregory Method, Weighted Inclusive Gregory Method and Meek)
all involve fractional transfers.  The principles are not difficult to
explain or comprehend.


> Once a candidate hits the quota, he stops
> getting any additional votes.

This is the correct procedure for transfers in accordance with the Dáil
Éireann rules and the Gregory Method (last parcel only).  But it is not
the correct procedure for transfers in accordance with the Weighted
Inclusive Gregory Method or Meek..  If applied to either of those
methods that procedure would give inconsistent results.  (The Inclusive
Gregory Method should never be used because it is fundamentally flawed
as it violates the "one person, one vote" principle.)


> In fact, random selection for surplus
> transfers might be an even easier way to explain it.

Such an explanation would be both wrong and very unhelpful.  The Gregory
Method (fractional transfers) was devised in 1880 to remove the element
of chance that is unavoidable with any method involving random selection
of ballot papers.  Most electors would now reject any voting system that
depended on random selection.  That said, random selection has survived
as acceptable in Ireland, though in the refined form described above.
When STV-PR was re-introduced into Northern Ireland in 1973, it was the
Gregory Method that was adopted.  Australia switched to the Gregory
Method long ago; they then changed to the (flawed) Inclusive Gregory
Method for their Federal Senate elections, but Western Australia has
since seen the light and is switching to the Weighted Inclusive Gregory
Method for its state elections.  In Scotland we shall use the Weighted
Inclusive Gregory Method for the Local Government elections that will be
held on 3 May.

James Gilmour