Yesterday's Guardian reported Labour's commitment to hold a referendum on
electoral reform if it wins the next election as a new pledge.  Forgive me
if I'm wrong, but wasn't it in their 1997 manifesto?  As I recall, they
commissioned a report from Roy Jenkins, then shelved the report and reneged
on their manifesto commitment, although no one seemed to care too much at
the time.

I think the alternative system being proposed this time is the Alternative
Vote rather than the hybrid system proposed by Jenkins.  (Not that it
matters much, given Labour's chances of winning the next election.)
--
Guy Barry

In article <_V%wm.277050$_Q3.226787@newsfe20.ams2>, 
guy.barry@blueyonder.co.uk (Guy Barry) wrote:

> Yesterday's Guardian reported Labour's commitment to hold a referendum
> on electoral reform if it wins the next election as a new pledge.
> Forgive me if I'm wrong, but wasn't it in their 1997 manifesto?  As I 
> recall, they commissioned a report from Roy Jenkins, then shelved the 
> report and reneged on their manifesto commitment, although no one seemed
> to care too much at the time.
> 
> I think the alternative system being proposed this time is the
> Alternative Vote rather than the hybrid system proposed by Jenkins.  
> (Not that it matters much, given Labour's chances of winning the next
> election.)

All true. This "new" promise isn't even for a Proportional system which 
was in the 1997 promise.

AV is a snare and a delusion, which could well be less proportional than 
FPTP. Until we end the single-member constituency monopoly at least to 
some extent (and Jenkins' proposal did that to least extent yet proposed 
by any official or semi-official body) then most voters will continue to 
be unrepresented in Parliament by MPs they choose.

-- 
Cllr. Colin Rosenstiel
Council member, Electoral Reform Society
http://www.electoral-reform.org.uk/
mailto:ers@electoral-reform.org.uk

On Oct 1, 12:21 pm, rosenst...@cix.compulink.co.uk wrote:
> In article <_V%wm.277050$_Q3.226...@newsfe20.ams2>,
>
> guy.ba...@blueyonder.co.uk (Guy Barry) wrote:
> > Yesterday's Guardian reported Labour's commitment to hold a referendum
> > on electoral reform if it wins the next election as a new pledge.
> > Forgive me if I'm wrong, but wasn't it in their 1997 manifesto?  As I
> > recall, they commissioned a report from Roy Jenkins, then shelved the
> > report and reneged on their manifesto commitment, although no one seemed
> > to care too much at the time.
>
> > I think the alternative system being proposed this time is the
> > Alternative Vote rather than the hybrid system proposed by Jenkins.  
> > (Not that it matters much, given Labour's chances of winning the next
> > election.)
>
> All true. This "new" promise isn't even for a Proportional system which
> was in the 1997 promise.
>
> AV is a snare and a delusion, which could well be less proportional than
> FPTP. Until we end the single-member constituency monopoly at least to
> some extent (and Jenkins' proposal did that to least extent yet proposed
> by any official or semi-official body) then most voters will continue to
> be unrepresented in Parliament by MPs they choose.

Is that right?  I thought the point of AV was that each MP had the
support
of a majority of voters in his/her constituency, so if only a minority
of
voters go unrepresented in each constituency, then only a minority
will be
unrepresented overall.  (It may still be a very large minority of
course.)

I'm a supporter in theory of STV in multi-member constituencies, but I
can
see the drawback in terms of constituency size, especially in rural
areas.
I'm not sure if somewhere like Bath (where I live) would want to give
up
having its own MP to be lumped in with the rest of Somerset, or (even
worse)
the former county of Avon.

--
Guy Barry

rosenst...@cix.compulink.co.uk wrote:
> AV is a snare and a delusion, which could well be less proportional than
> FPTP. Until we end the single-member constituency monopoly at least to
> some extent (and Jenkins' proposal did that to least extent yet proposed
> by any official or semi-official body) then most voters will continue to
> be unrepresented in Parliament by MPs they choose.

I wholeheartedly support the LibDem and ERS view that we should use
STV. That's because I value both greater proportionality *and*
ordinality, and STV delivers both.

AV is not a proportional system, but it does provide ordinality, and
that's a good thing. To overlook the positive aspects of ordinality in
AV is to undermine some of the arguments in favour of STV.

If AV is more ordinal than FPTP and as proportional, then I think we
should support it over FPTP as a step in the right direction. I'm not
certain the evidence base is sufficient to conclude it is necessarily
*less* proportional than AV. Proportionality is heavily dictated by
the number of major parties, regional voting patterns and so on. I'm
sure there are some situations in which AV is less proportional and
others where FPTP is less proportional, but what evidence shows that
AV is *generally* less proportional to a significant degree?

It is also the case that ordinality complicates what we mean by
proportional. Calculations of proportionality that look at the first
preference vote only do not properly represent what AV or STV achieve.
If we are to defend STV against charges that it is insufficiently
proportional with reference to its ordinality, then we have to defend
AV in the same way.

So, I agree that there is little reason to trust Labour on this matter
given their track record, I agree that Labour are going to lose the
election so it hardly matters anyway, I agree with the criticisms of
AV for not being sufficiently proportional, but were there to be a
referendum on AV vs. FPTP, then I'd vote for AV.
--
Henry

"Henry Potts"  wrote in message
news:f37ed977-0aea-4d67-91c8-fc6f810113f4@h13g2000yqk.googlegroups.com...

> I wholeheartedly support the LibDem and ERS view that we should use
> STV. That's because I value both greater proportionality *and*
> ordinality, and STV delivers both.

Apologies - what does "ordinality" mean in this context?

--
Guy Barry

On 2 Oct, 09:26, "Guy Barry"  wrote:
> "Henry Potts"  wrote in message
>
> news:f37ed977-0aea-4d67-91c8-fc6f810113f4@h13g2000yqk.googlegroups.com...
>
> > I wholeheartedly support the LibDem and ERS view that we should use
> > STV. That's because I value both greater proportionality *and*
> > ordinality, and STV delivers both.
>
> Apologies - what does "ordinality" mean in this context?
>
> --
> Guy Barry

It presumably means being able to rank candidates or parties in order:
1st, 2nd, 3rd etc,, rather than FPTP just 1st, or approval voting Yes,
Yes, No etc., or range voting 100/100, 75/100, 72/100 etc.  I would
guess it does not mean that the party with most votes gets the most
seats.

There are proportionality (and ordinality in the second sense) issues
with STV: of the last seven general elections in Malta under STV, four
would have led to the party with more votes coming second had there
not been a special non-STV provision to deal with the problem; in
three of these cases the more popular party had over 50% of the first
preference vote, and in the fourth 49.3%, so this was not an issue of
cross-party transfers, but instead a multi-member constituency
rounding effect.    By comparison, in the UK to get four cases of the
party with the most votes coming second in seats would take you back
to around 1910, about 25 general elections ago.

On Oct 2, 2:16 pm, Henry  wrote:
> On 2 Oct, 09:26, "Guy Barry"  wrote:
> > Henry Potts  wrote
> > > I wholeheartedly support the LibDem and ERS view that we should use
> > > STV. That's because I value both greater proportionality *and*
> > > ordinality, and STV delivers both.
> >
> > Apologies - what does "ordinality" mean in this context?
>
> It presumably means being able to rank candidates or parties in order:
> 1st, 2nd, 3rd etc,, rather than FPTP just 1st, or approval voting Yes,
> Yes, No etc., or range voting 100/100, 75/100, 72/100 etc.  I would
> guess it does not mean that the party with most votes gets the most seatsYes, basically. Ordinality means any mechanism by which you get to
express your preferences more fully than just by voting for one
candidate or party. So, AV or STV where you rank candidates is an
ordinal system. SV (as at Mayoral elections) is ordinal (if less so)
as you can put a 1st and 2nd choice. An open list system can be
ordinal. Two-ballot systems, as used in France, where a second round
of voting is held among the top candidates if no candidate got more
than 50% -- that's an ordinal system as you can change your vote
between rounds.

> There are proportionality [...] issues
> with STV: of the last seven general elections in Malta under STV, four
> would have led to the party with more votes coming second had there
> not been a special non-STV provision to deal with the problem; in
> three of these cases the more popular party had over 50% of the first
> preference vote, and in the fourth 49.3%, so this was not an issue of
> cross-party transfers, but instead a multi-member constituency
> rounding effect.    By comparison, in the UK to get four cases of the
> party with the most votes coming second in seats would take you back
> to around 1910, about 25 general elections ago.

Malta is a small country with two very evenly matched parties. The UK
is a much bigger party with more electorally successful parties. The
problems you mention are not caused by STV; they are caused by two
evenly matched parties in a small country. FPTP wouldn't have done any
better. STV doesn't have these problems in Ireland.
--
Henry

Henry Potts wrote:

> I wholeheartedly support the LibDem and ERS view that we should use
> STV. That's because I value both greater proportionality *and*
> ordinality, and STV delivers both.

Not entirely. The STV proposed is generally for about 5 member 
constituencies that will help the Lib Dems but won't be so great for smaller 
parties like UKIP, the Greens and the BNP whose support is rarely 
concentrated even in a natural cluster. The AV+ proposal had the similar 
problem that it would largely help third and second parties rather than 
smaller ones.

> It is also the case that ordinality complicates what we mean by
> proportional. Calculations of proportionality that look at the first
> preference vote only do not properly represent what AV or STV achieve.

Well let's take a look at the Australian election results. There one of the 
key concepts is the "2 Party Preferred" vote - i.e. when all votes are 
transferred the full way does the voter prefer Labor or the Liberal/National 
Coalition? For all elections since 1983 the ballot papers have been 
transferred all the way, even when its not necessary, so both first and 
final preferences are available.

The other factors to remember about Australian federal elections are that it 
is compulsory to use *all* preferences for a vote to be valid; and that on 
polling day party activists stand outside polling stations and distribute 
"How To Vote" cards recommending an order of preferences. Although not all 
voters follow them, they are significant in determining the outcome.

The figures are as follows:

1983:

First Preferences:
Labor: 4,297,392 49.48% 75 seats
Liberals: 2,983,986 34.36% 33 seats
Nationals: 799,609 9.21% 17 seats
Australian Democrats: 437,265 5.03% 0 seats
Other: 166,611 1.92% 0 seats

2PP:
Labor: 53.23% 75 seats
Coalition: 46.77% 50 seats

(The House of Represetatives was expanded from 125 to 148 seats from the 
1984 election.)

1984:

First Preferences:
Labor: 4,120,130 47.55% 82 seats
Liberals: 2,951,556 34.06% 44 seats
Nationals: 921,151 10.63% 21 seats
Australian Democrats: 472,204 5.45% 0 seats
Country Liberal: 27,335 0.32% 1 seats
Other: 172,576 1.99% 0 seats

(The Country Liberal Party is a merger of the National and Liberal parties 
in the Northern Territory. It is sometimes listed separately and sometimes 
listed with the Liberals as their House of Representatives members always 
sit with the Liberals. Confusingly their Senators always sit with the 
Nationals.)

2PP:
Labor: 51.77% 82 seats
Coalition: 48.23% 66 seats

1987:

First Preferences:
Labor: 4,222,431 45.76% 86 seats
Liberals: 3,175,262 34.41% 43 seats
Nationals: 1,060,976 11.50% 19 seats
Australian Democrats: 554,017 6.00% 0 seats
Country Liberal: 21,668 0.23% 0 seats
Other: 189,975 2.06% 0 seats

2PP:
Labor: 50.83% 86 seats
Liberals & Nationals: 49.17% 62 seats

This election took place at a time of political chaos on the conservative 
side. The Queensland branch of the National Party launched the bizarre "Joh 
for Canberra" campaign to elect Queensland premier Sir Joh Bjelke-Petersen 
as Prime Minister, in the process forcing the splitting of the Coalition. In 
the election there were more seats contested by both the Liberals and 
Nationals than normal, plus a number contested by "Joh's Nationals" on a 
separate ticket. As if this isn't complicated enough, the National Party has 
a very decentralised structure with the result that the state parties are 
often a law unto themselves even in federal elections. So it's unclear to 
what extent votes for the Nationals especially were and weren't for the 
Liberals or Coalition.

1990:

First Preferences:
Labor: 3,904,138 39.44% 78 seats
Liberals: 3,468,570 35.04% 55 seats
Australian Democrats: 1,114,216 11.26% 0 seats
Nationals: 833,557 8.42% 14 seats
Independents: 252,116 2.55% 1 seat
Other: 327,077 3.30% 0 seats

2PP:
Labor: 49.90% 78 seats
Coalition: 50.10% 69 seats

1993:

First Preferences:
Labor: 4,751,390 44.92% 80 seats
Liberals: 3,923,786 37.10% 49 seats
Nationals: 758,036 7.17% 16 seats
Australian Democrats: 397,060 3.75% 0 seats
Greens: 196,702 1.86% 0 seats
Independents: 328,084 3.10% 2 seats
Other: 221,721 2.10% 0 seats

2PP:
Labor: 51.44% 80 seats
Coalition: 48.56% 65 seats

1996:

First Preferences:
Labor: 4,217,765 38.75% 49 seats
Liberals: 4,210,689 38.75% 75 seats
Nationals: 893,170 8.21% 18 seats
Australian Democrats: 735,848 6.76% 0 seats
Greens: 314,594 2.89% 0 seats
Australians Against Further Immigration: 73,023 0.67% 0 seats
Call to Australia: 42,683 0.39% 0 seats
Natural Law Party: 41,573 0.38% 0 seats
Country Liberal: 38,302 0.35% 1 seat
Independents: 246,986 2.27% 0 seats
Other: 69,219 0.35% 0 seats

2PP:
Coalition: 53.63% 94 seats
Labor: 46.37% 49 seats

Note: The Liberals disendorsed their candidate in Oxley, Pauline Hanson. 
However it was too late to nominate a replacement and the ballot papers were 
printed listing her as a Liberal candidate. She won and is listed here as an 
independent.

This election has some of the big distortions. Labor polled the highest 
number of first preferences (albeit only because the Coalition parties are 
listed separately) yet the Liberals won 50% more seats. Labor had been in 
power since 1983 and was trying to win a sixth term so was probably transfer 
weak.

1998:

First Preferences:
Labor: 4,454,306 40.10% 67 seats
Liberals: 3,800,721 34.21% 64 seats
One Nation: 936,621 8.43% 0 seats
Nationals: 588,088 5.29% 16 seats
Australian Democrats: 569,935 5.13% 0 seats
Greens: 238,035 2.14% 0 seats
Independents: 212,522 1.91% 1 seat
Other: 308,835 2.78% 0 seats

2PP:
Coalition: 49.02% 80 seats
Labor: 50.98% 67 seats

2001:

First Preferences:
Labor: 4,341,420 37.84% 65 seats
Liberals: 4,291,032 37.40% 69 seats
Nationals: 643,926 5.61% 13 seats
Australian Democrats: 620,225 5.41% 0 seats
Greens: 569,074 4.96% 0 seats
One Nation: 498,032 4.34% 0 seats
Independents: 332,669 2.90% 3 seats
Other: 177,696 1.55% 0 seats

2PP:
Coalition: 51.03% 82 seats
Labor: 48.97% 65 seats

2004:

First Preferences:
Liberals: 4,781,313 40.81% 75 seats
Labor: 4,409,117 37.64% 60 seats
Greens: 841,734 7.19% 0 seats
Nationals: 690,275 5.89% 12 seats
Independents: 292,036 2.49% 3 seats
Family First: 235,315 2.01% 0 seats
Australian Democrats: 144,832 1.24% 0 seats
One Nation: 139,956 1.19% 0 seats
Other: 180,554 1.54% 0 seats

2PP:
Coalition: 52.74% 87 seats
Labor: 47.26% 60 seats

2007:

First Preferences:
Labor: 5,388,184 43.38% 83 seats
Liberals: 4,546,600 36.60% 55 seats
Greens: 967,789 7.79% 0 seats
Nationals: 682,424 5.49% 10 seats
Australian Democrats: 89,813 0.72% 0 seats
Family First: 246,798 1.99% 0 seats
Independents: 276,370 2.23% 0 seats
Other: 222,004 1.79% 0 seats

2PP:
Labor: 52.70% 83 seats
Coalition: 47.30% 65 seats

Tim Roll-Pickering wrote:

> Independents: 276,370 2.23% 0 seats

Sorry that should be 2 seats.

Henry Potts wrote:
> On Oct 2, 2:16 pm, Henry  wrote:
>> There are proportionality [...] issues
>> with STV: of the last seven general elections in Malta under STV, four
>> would have led to the party with more votes coming second had there
>> not been a special non-STV provision to deal with the problem; in
>> three of these cases the more popular party had over 50% of the first
>> preference vote, and in the fourth 49.3%, so this was not an issue of
>> cross-party transfers, but instead a multi-member constituency
>> rounding effect.    By comparison, in the UK to get four cases of the
>> party with the most votes coming second in seats would take you back
>> to around 1910, about 25 general elections ago.
> 
> Malta is a small country with two very evenly matched parties. The UK
> is a much bigger party with more electorally successful parties. The
> problems you mention are not caused by STV; they are caused by two
> evenly matched parties in a small country. FPTP wouldn't have done any
> better. STV doesn't have these problems in Ireland.

The size of the country doesn't really matter.  What causes this in 
Malta is the constituency size of five seats, plus as you say two evenly 
matched dominant parties.  Deciding who came first in Ireland has in the 
past been less of an issue as the same party has won the most votes and 
seats since 1932.

There are other issues, such as differential turnout. Take the last 
three Northern Ireland Assembly STV elections as examples, where turnout 
is often lower in unionist areas: in 1999 the Ulster Unionists won more 
seats than the SDLP despite fewer first preferences; in 2003 the Ulster 
Unionists v. Sinn Fein; and in 2007 the Ulster Unionists v. SDLP again. 
  This matters because it can affect the number of executive members. 
That being said, it is again a constituency issue rather than just STV: 
in a party-list election in 1993 the DUP won more seats than the SDLP 
despite having fewer votes.

Henry  wrote: [...]
> There are other issues, such as differential turnout. Take the last
> three Northern Ireland Assembly STV elections as examples, where turnout
> is often lower in unionist areas: in 1999 the Ulster Unionists won more
> seats than the SDLP despite fewer first preferences; in 2003 the Ulster
> Unionists v. Sinn Fein; and in 2007 the Ulster Unionists v. SDLP again.
>   This matters because it can affect the number of executive members.
> That being said, it is again a constituency issue rather than just STV:
> in a party-list election in 1993 the DUP won more seats than the SDLP
> despite having fewer votes.

Differential turnout is an issue for any election with constituencies.
That said, the Unionist vote is also more fractured between a larger
number of candidates, so this is an example of ordinality confusing
proportionality calculations. The UUP pick up more transfers despite
getting fewer first preferences.
--
Henry