Electoral Reform (Senate, Commons, & Gov Gen)

[Kirkhill]

People do that now.

Yes. They do.

And they have.

And they will.

 

[Kirkhill]

Maybe I'm getting the context wrong, but at face value, I disagree with the completely. I don't want anyone doing any job that they don't want to be doing, especially running our legislature.  Sure, make the job difficult, onerous, and demanding, but I want people who are passionate about what they are doing.

No one "wants" to go to war, kill people, and see their friends killed; but I don't want my next CO to hate his job either.

Context.  Right enough.

I agree with you on your CO.  And you for that matter.

I also don't have a problem with the governors wanting to govern.  Even being passionate about wanting change.

But before they are allowed to make change, before they are allowed the funds to finance the change, I want them to face the most contrary, ornery bunch of tightwad jurors imaginable.  I don't want the Government to "enjoy the confidence of the House".  I want them to have to justify their existence to the House every day.

The business of the Government may be rendered less efficient - but when efficiency is demanded then it has the Royal Prerogative.  It can act as it sees fit.  It will be held accountable at the next election by the general public.  In the nearer term it is to be held accountable, after the fact, by the House.

But.

The House should not be contrary for the sake of being contrary.  I don't want careerists.  I don't want people that see their time in the House as a lifetime sinecure.  I want people to see it as a civic duty of limited duration.  I want them to see it in exactly the same sense as Jury duty.

 

[Brad Sallows]

>Looking inside the city of London, 32% (almost 1/3) want a conservative MP but they do not get this under FPTP single representative ridings.  A multi-member riding would give one conservative MP.

Which one?  I don't see how one of the Conservative candidates winning an MP's seat with 10-15% of the vote is an improvement over winning with 30-35%.

 

[McG]

This:

Consider that London-West, London-North-Center, and London-Fanshaw could all be lumped into one riding of "London" with three MPs.  In the last election, those three ridings elected one NDP (L-F) and two Liberals (L-NC and L-W), but as a single three member riding they would have elected one Conservative and two Liberals.

is what I mean by:

"This doesn't necessarily produce a result that is "more fair" in circumstances in which traditionally one-party ridings merge with ridings in which vote shares are more equally shared."

If a pocket of people in L-F want a NDP member for their locality, they should get one instead of another member decided in part for them by their neighbours.

I consider a system that allows representatives to be selected with non-majority pluralities, and majority governments to be formed with less than majority vote share, to be a great social stabilizer.

Yet, it was a minority of people in L-F that decided they would have an NDP representative.  Is that really more fair when what the city (inclusive of L-F) really wants is a Conservative?

Yes, I consider it "more fair" (and healthier for the stability of the country) when people in a locality can choose a representative unencumbered by the preferences of people further away.

It must be that you are being deliberately obtuse, given that the three London ridings together are all just one small "locality."  Collectively these three ridings are a fraction of the size of any adjacent riding.  Are the people of Lambton-Kent-Middlesex deprived of fairness as most of them are "encumbered by the preferences of people further away" to an extent that would be unachieved in the hypothetical three member riding that you use distance to argue against.

Looking at a graphic of London from the 2015 election, I now see that the 2 x LPC and 1 x NDP ridings seem to be surrounded by CPC ridings.  So I'm more convinced that the pocket of NDP supporters in London who can currently manage to elect a NDP representative - at least occasionally - get a fair shake from FPTP that they might never otherwise get in a multi-member riding.  And, looking at Canada-wide results in general, it is clear that "locality" matters and in a densely populated area the political flavour of the population can change dramatically.  But having lived in Metro Vancouver for years, I already knew this.  Larger multi-member ridings mean shutting some people out too much of the time.

Representation is population-based.  It doesn't have to be perfect, but those who want to have angels-on-a-pin discussions may go ahead. 

What you are feeling is called cognitive dissonance.  If "locality matters" is your underlying premise against geographically small multi-member ridings, then it does not matter what surrounds the city of London when we are talking about voting results inside the city of London.  Looking inside the city of London, 32% (almost 1/3) want a conservative MP but they do not get this under FPTP single representative ridings.  A multi-member riding would give one conservative MP.

Which one?  I don't see how one of the Conservative candidates winning an MP's seat with 10-15% of the vote is an improvement over winning with 30-35%.

Which one?  Are we even having the same conversation, or are you just posting baffle-gab nonsense?  In a hypothetical three member riding consisting of the combined  London-West, London-North-Center, and London-Fanshaw ridings the results of the last election show 32% support for the conservatives. That same hypothetical riding would have given the Conservatives one of the three seats had it existed in the last election.

You do understand that riding boundaries are not divinely created?  They are drawn on a map by a human being.  They exist as a product of historical evolution and boundaries that shift somewhat arbitrarily in attempt to keep relative population parity between adjacent ridings.  London-West and London-North-Center are arbitrarily divided by a street, while London-West itself is bisected by a river.  I can assure you that Hyde Park and White Hills (L-W) are more closely linked to Sherwood Forest and University Heights (L-NC) than either is to Byron or Westmont (L-W).

I joined the Army in London.  The able bodied can easily walk across the whole thing in a morning, but reliable public transit gives everyone quicker access to anywhere.  I lived in one riding, worked in another riding and when to school in a third - and I was not atypical.  If you live in that city, your daily life likely takes you across riding boundaries for work, shopping, school, social activities, community events, etc.  By contrast, when I was at Gagetown, I could go weeks at a time (months if you discount trips into the training area) without leaving my riding.

So inside high density urban areas, riding boundaries are less accurate in defining communities because they are driven by population pressures, and riding boundaries are less relevant because people's daily lives (and so the communities they identify with) are much broader.  L-F no more accurately defines a community than a single riding that encompass all of L-F, L-NC and L-W.  Why is an NDP seat wanted by a minority within an arbitrary box "more fair" than giving one third of the city's seats (ie. 1 seat) to a Conservative that one third of the city wants?

Maybe instead look at all the other small multi-member riding examples and note that, with few enough exceptions that you could count on one hand, small multi-member ridings do not remove parties from cities regions.  Instead, multi-member ridings generally add a representative from one of the "big three" in a city was swept by one of the other "big three."  Instead of everybody being represented by the voice of 35-40%, the next 30-35% of the community also get a voice at the table ... and all without the addition of any new MP seats.

You could also turn your concerns to something like the riding of Elgin—Middlesex—London which, created and allowed under the current system, sees the desires of citizens in St Thomas "encumbered by the preferences of people further away" in south London.

Agreed, and that means that any consideration of multi-member ridings would have to give voters the ability to select the individual MPs (ie. no party lists).  Unfortunately, when working with FPTP results to model a hypothetical multi-member riding, the data just does not exist to replicate STV.  It also does not let you model an independent ... Brent Rathgeber scored well enough in his riding that if other ridings provided him similar support he would have been a contender in a North Edmonton multi-member riding, but data only exists for the riding he actually ran in.

In any case, I extended the model of multi-member ridings across the country (less Vancouver and Victoria which I just did not get to) to get 36 multi-member ridings (including the ones already posted).  I am sure threehundredeight.com could do a better job of this, but my model suggests use of big city multi-member ridings would have produced the following results in the 2015 election:

Liberal		Conservative		NDP		Green		BQ
163​		112​		44​		1​		11​
-21​		+13​		+7​		+0​		+1​

Attached are the ridings that I have not yet posted, if you want to see what the results would be in your neighbourhood. 

So I have applied the model to Vancouver and Victoria.  In Vancouver the model would have given two additional Conservatives, two fewer Liberals, and an equal number of NDP being more geographically dispersed.  That gives new national results as:

Liberal		Conservative		NDP		Green		BQ
161​		114​		44​		1​		11​
-23​		+15​		+7​		+0​		+1​

Victoria would go from two NDP and a Green to one each for NDP, Liberal and Green. However, I would not use the riding in the Victoria model.  Esquimalt—Saanich—Sooke and Saanich—Gulf Islands both extend well away from the urban hub.  I suspect that Victoria could support a two member riding, but the three ridings would have to be combined with Cowichan—Malahat—Langford, and then divided into a two-member city riding and two single member outlying ridings.

 

[Brad Sallows]

You seem to assume there would only be one CPC candidate running in a multi-member riding.  Why wouldn't each of the major parties run three candidates in a three-member riding?  Certainly the LPC would have to run at least two candidates to claim the two seats you suggest they could capture.

 

[McG]

You seem to assume there would only be one CPC candidate running in a multi-member riding. 

No.  I have shown the math for each party running one candidate for each seat in the hypothetical multi-member ridings.

 

[Brad Sallows]

You've shown some vote counts, but how do you justify the assumption that all of the votes for each party are distributed among candidates with perfect efficiency?

If 58,917 votes were cast for CPC candidates in the three distinct London ridings, why would you assume those voters would support one particular CPC candidate in great enough numbers to capture a seat in the equivalent multi-member riding?  Why would the 58,917 votes not be distributed among three CPC candidates, such that each of them falls short of the candidates who won the election?

 

[McG]

You've shown some vote counts, but how do you justify the assumption that all of the votes for each party are distributed among candidates with perfect efficiency?

I did not make that assumption.

I have previously described some limitations that come from using real world FPTP results to create a model for multi-member ridings, and the assumptions made can be found there.  While the limitations prevent modelling a number of factors (relative popularity of candidates within a party, popular independents, popular candidates in disliked parties, and disliked candidates in popular parties), I think the results are a good approximation of what a real world run would get us.

 

[Brad Sallows]

Strictly speaking, I agree the assumption of "perfect efficiency" of distribution of votes was not made; I exaggerated the case.  Nevertheless, discard the absolutism of "perfect" and there is still an implied assumption that the votes will somehow be distributed more efficiently, sufficiently to  lift one of the (in the example of London) CPC candidates above the count obtained by the NDP member.  That assumption applies in all cases: the distribution of votes for each member must actually be different for the outcomes to be different.

The assumption is unsound and unreasonable.  What is the basis for making it?  The premise can't be assumed. It must be proven.  There is no reason to believe the model is a useful approximation, because there is no way of knowing whether the distribution of votes would be nearly the same, more uneven but favourable in just the right way to benefit a party on the basis of vote count, or so wildly different as to result in a completely unforeseen outcome; the model is too weak to draw any conclusion.  At this point, it is just a bunch of hand-waving:
1) The vote count is such-and-such.
2) ???
3) The outcome should be thus (but not this, or that, or anything else).

What is left, then, is a solution which has not been shown to solve the suggested problem (that there is an unfair imbalance of votes which needs to be resolved with slightly larger boundaries, but not any larger), and still no demonstration that a problem exists which needs to be solved.  The measurement of party vote share over a region larger than a single riding is no more relevant than the distribution within a province, or within the country as a whole.

If a riding votes 90% to elect a CPC MP, then they have elected one CPC MP.  The fact that they did so with a large surplus of votes is irrelevant.  No rational basis has been advanced for extending the boundaries of ridings some arbitrary distance, and no further.  If there is some objective criterion for making some ridings larger but not others, explain it.

 

[jmt18325]

Is it not possible to create a ballot where you rank party choices, and then within that rank your choice of candidates if you so choose?

 

[Jed]

Is it not possible to create a ballot where you rank party choices, and then within that rank your choice of candidates if you so choose?

Sure it is possible. But what is the point of doing this?  No one needs this complicated process. The KISS principle for voting is the only efficient and effective choice.

 

[McG]

... there is still an implied assumption that the votes will somehow be distributed more efficiently, sufficiently to  lift one of the (in the example of London) CPC candidates above the count obtained by the NDP member.  That assumption applies in all cases: the distribution of votes for each member must actually be different for the outcomes to be different.

No there is not "an implied assumption."  The only undeclared assumption is being made by you.  You have assumed that SNTV is the only voting option for a multi-member riding.

Is it not possible to create a ballot where you rank party choices, and then within that rank your choice of candidates if you so choose?

It is possible:

...but I would rather any ranked ballot just rank the individual candidates while identifying which party each one represents.

 

[Brad Sallows]

Whether you depend on the voters making a single SNTV choice each or an alternate mechanism, there is still an "assumption that the votes will somehow be distributed more efficiently".  What alternate mechanism(s) would you propose to channel the outcome?

>Is it not possible to create a ballot where you rank party choices, and then within that rank your choice of candidates if you so choose?

A ranking system gives "anybody but" voters a mathematical advantage over "single preference" voters.  A system which yields imbalanced strategic voting power should be discarded.  (There must be no imbalances, because safeguarding the output legitimacy - essentially, the sense of a level playing field - is of paramount importance).  That is one of the strong advantages of FPTP: every voter makes exactly one choice, with the same information, with no possible way of gaming a path through a decision tree.

 

[Jed]

Whether you depend on the voters making a single SNTV choice each or an alternate mechanism, there is still an "assumption that the votes will somehow be distributed more efficiently".  What alternate mechanism(s) would you propose to channel the outcome?

>Is it not possible to create a ballot where you rank party choices, and then within that rank your choice of candidates if you so choose?

A ranking system gives "anybody but" voters a mathematical advantage over "single preference" voters.  A system which yields imbalanced strategic voting power should be discarded.  (There must be no imbalances, because safeguarding the output legitimacy - essentially, the sense of a level playing field - is of paramount importance). That is one of the strong advantages of FPTP: every voter makes exactly one choice, with the same information, with no possible way of gaming a path through a decision tree.

This is why we should leave it alone and stick with the status quo.

 

[jmt18325]

Sure it is possible. But what is the point of doing this? 

Proportional representation.  Many would argue that our current structure doesn't allow for people to have their voices heard in the way that they wanted them to be heard.

 

[McG]

Whether you depend on the voters making a single SNTV choice each or an alternate mechanism, there is still an "assumption that the votes will somehow be distributed more efficiently". 

No there is not.

What alternate mechanism(s) would you propose to channel the outcome?

There is Single Transferable Vote or Schulze Transferable Vote.

A ranking system gives "anybody but" voters a mathematical advantage over "single preference" voters. 

The federal 2015 election shows that, informed by polls, the "anybody but" camp will just as easily organize a mathematical advantage in FPTP.

That is one of the strong advantages of FPTP: every voter makes exactly one choice, with the same information, with no possible way of gaming a path through a decision tree. 

I can game the FPTP system by running an irrelevant candidate with a platform similar to the contender(s) that I oppose, and in strong contrast to the contender(s) that I support.  My irrelevant candidate stands no chance to win, but despite that my irrelevant candidate will draw votes away from those competitors that I want to loose while not affecting my preferred candidate.

 

[Jed]

No there is not.
There is Single Transferable Vote or Schulze Transferable Vote.
The federal 2015 election shows that, informed by polls, the "anybody but" camp will just as easily organize a mathematical advantage in FPTP.
I can game the FPTP system by running an irrelevant candidate with a platform similar to the contender(s) that I oppose, and in strong contrast to the contender(s) that I support.  My irrelevant candidate stands no chance to win, but despite that my irrelevant candidate will draw votes away from those competitors that I want to loose while not affecting my preferred candidate.

I have seen a couple of elections were that process did not work for gaming the system.  Ask Roy Romanow how he was beaten by Jo Z. the giant killer.

 

[McG]

I have seen a couple of elections were that process did not work for gaming the system. 

From 1997 to 2004, that split gave the Liberals a series of majority governments.  It does work.

 

[Jed]

From 1997 to 2004, that split gave the Liberals a series of majority governments.  It does work.

It worked because Quebec and Ontario could game the system not because FPTP.

 

[jmt18325]

It worked because Quebec and Ontario could game the system not because FPTP.

Game the system?  How were they gaming the system, exactly?  Ontario is actually still, to this day, shortchanged for seats.