May 13 – BC Referendum, RIP STV

Posted May 13, 2009 by Sacha Peter
Category: Analysis, Referendum

The BC Referendum was soundly defeated by the people that voted. The quick analysis is that people overwhelmingly were turned off by both the large electoral districts that would be created by such a system, and also the complexity of the voting system.

This referendum, in conjunction with Ontario’s previous referendum on electoral reform, virtually guarantees that electoral reform will not be discussed again for a generation – First Past the Post is here to stay.

My advice for any budding political parties would be to look at the FPTP system and realize that it is a system that rewards regional concentration and punishes diffuse support. Just examine what the Bloc Quebecois has been able to do in this country with 10% of the popular vote, but consistently pulling in about 16% of the seats elected across Canada. The Green party could get 10% of the popular vote nationally but it will translate into zero seats unless they can start to regionally concentrate.

Top 10 voters of First Past the Post

Top 10 voters of STV

Comments Off

Prediction on Referendum on Electoral Reform

Posted May 11, 2009 by Sacha Peter
Category: Predictions, Referendum

In order for the referendum to “pass”, the government set a 60% threshold for approving a change to another electoral system. The other criterion was 50% of the ridings must have 50% support for the referendum.

The 60% requirement on a binary question makes it very difficult for the referendum to pass – for every 2 people voting for status quo, it requires 3 STV voters to offset.

The way that the ballot was worded (from a yes/no question in 2005 to a much more wordy ballot in 2009), combined with the general lack of public education on the nuances on the electoral system, meant that the pro-STV forces had an impossible task at hand.

The referendum was likely to fail on that basis alone – if you have enough people voting on an issue they know nothing about, it will normalize to 50/50. If you have 100 people in a room, and 60% of them don’t know of an issue (which is generally what proportion do not have a good grasp of electoral reform issues), you then require the remaining 40 people in the room to vote 75% in favour of STV in order to have the room reach the 60/40 threshold.

For those of you that understood the previous paragraph, it becomes mathematically obvious that the referendum was never going to be passing. In fact, it was a statistical miracle it ever came to 58/42 as it did in 2005 – most of this was due to the positive wording on the ballot. Specifically the change of the words from the 2005 ballot of “as recommended by the Citizen’s Assembly” to “as proposed by the Citizen’s Assembly” is crucial.

The pro-STV campaign made some serious mistakes. The first mistake was that their main slogan on signage everywhere was “Power up your vote”, which might work well for a video game, but the bulk of the voting populace will be older than the “video game generation”. The second mistake they made was spending too much time on the mechanics of how STV worked (which is an unwinnable proposition). A better alternative would have been to tap into the general detachment the public has with the government (“Don’t like Gordon Campbell or Carole James? Choose STV!”), and to link together the concept of STV and better government.

Conversely, the anti-STV campaign was very successful on their messaging – the single message that “STV will cause larger ridings, which will reduce your local representation” carried. For rural ridings, the argument of “all of your representatives can be elected from a single city” is incredibly powerful, easily understood, and will crush the STV campaign. It doesn’t even matter if it isn’t true (which it is not!), but counter-arguments will get lost in the noise.

In terms of measuring the relative effectiveness of both campaigns, my guess is that Bill Tieleman and David Schreck beat the pro-STV campaign.

Prediction on 2009 referendum on electoral reform:

First Past the Post 54%
BC-STV 46%

STV ridings that vote above 50% for STV: 20/85 (24%). My guess is that Vancouver Island would vote for STV, and dense urban ridings. The interior of the province will reject STV convincingly.

I’ll consider this post to be the obituary on any serious electoral reform in the province. It is too bad considering that STV is the best of alternative systems. For the record, I consider first past the post a good system – just that STV allows voters much more precision in voting. I voted for BC-STV.

Inevitably, if people want “proportional” results in elections with the First Past the Post system, they will need to consider increasing the numbers of members in the legislature to a higher level. The more electoral districts there are, the more proportional the FPTP results will be.

10 Comments

What if we had STV last time?

Posted May 3, 2009 by Sacha Peter
Category: Referendum

Last Wednesday, Christina Myers of the Burnaby Now called me and asked me for some questions and comments about STV and how it would affect the newly proposed Burnaby-New Westminster STV electoral district. Some of the conversation can be found in the opinion page, What if we had STV last time?

But it’s a question that comes with plenty of assumptions, warns Sacha Peter, who maintains a website called B.C. Election 2009: The Race for Victoria. Peter, who is a parks commissioner in Cultus Lake and has worked on federal Conservative campaigns in the past, says politics is something he’s just personally interested in. He crunched the numbers to come up with a prediction about what the provincial results might have looked like riding by riding for his site.

It’s an exercise in curiousity, more than one of determining fact, he says.

“You have to look at those new riding boundaries and take each area and add up the popular vote, and you end up with approximately what it might look like,” he said. “But let me give a big disclaimer. If the way we vote changes, then those numbers change too. It’s not entirely appropriate to do that sort of analysis. The results that come out of it, it’s really almost impossible to say because you’re making assumptions about how people would vote.”

For example, with the Green Party – which he suggests might have ended up with three representatives – it’s difficult to say how voters would rank their preferences.

“With respect to third parties, it’s very mathematically easy to say that would happen – but whether that would actually happen or not is a very good question.”

Peter says the interesting question is not which party has more or fewer candidates, but which candidates within each party get the votes. If voters have a choice between several candidates from each party, would they vote for candidates in one party only or show a preference for candidates in different parties?

A similar situation would be in US Presidential elections – the case study of getting rid of the electoral college. If the President of the United States were elected based on an at-large country-wide popular vote, the nature of the election campaign would be drastically different.

Implementing STV would also change the nature of campaigning. It would not be better or worse, but it would be different.

The last paragraph alludes to the most interesting thought experiment of STV: Which individual candidates would get elected under STV had it taken place in 2005? Specifically in the Burnaby-New Westminster area we had the following candidates, ranked from most votes to least votes:

BC Liberal:
Richard Lee (10,421) (Elected in 2005)
Harry Bloy (10,054) (Elected in 2005)
Joyce Murray (9,645)
Patty Sahota (9,599)
John Nuraney (8,754) (Elected in 2005)

NDP:
Chuck Puchmayr (13,226) (Elected in 2005)
Pietro Calendino (10,365)
Raj Chouhan (10,337) (Elected in 2005)
Bart Healey (9,682)
Gabriel Yiu (8,355)

Green:
Robert Broughton (2,416)
Suzanne Deveau (2,192)
Richard Brand (1,763)
Carli Irene Travers (1,619)
Pauline Farrell (1,482)

It is more probable that John Nuraney would not have been elected under STV, while somebody like Pietro Calendino, would be. Anti-STV advocates are very correct in pointing out that the average voter, seeing 15 names on the ballot, would cringe. However, if they know one or two people, they can mark them down, without having to rank all 15 candidates.

The strongest candidates in an STV system, however, would be the ones that could gain cross-partisan support – maybe there is a voter that is leaning toward the NDP, but like some of the left-wing principles of candidate X. So the voter marks “1″ next to Chuck Puchmayr and “2″ for Pietro Calendino, while he marks “3″ for Joyce Murray. Candidates that are able to garner enough cross-partisan support will do much better in STV systems than they would presently.

The smartest question I have seen so far in the comment forms was the following, from Mike:

Has anyone explored the game theory involved in parties choosing the number of candidates to run in a district?

This has been academically explored, and also practically explored in Ireland’s elections. It is very similar to situations that municipal slates face when they decide how many candidates to run for city council – they want to run as many candidates as the seats they think they can capture plus a tiny amount more. But if you go over that, then the candidates will be competing against each other for the same vote base, which will compromise their chances. The method of counting votes in STV mitigates this (which is the biggest difference between municipal council voting vs. STV), but the campaigning that would take place in the meantime would be more damaging.

To explore this question in its entirety, however, would take a lot of words. It is not an easy question to answer.

2 Comments

First BC-STV Referendum poll

Posted April 22, 2009 by Sacha Peter
Category: Polls, Referendum

No STV has been the “first past the post” to publish the full results of an Ipsos-Reid poll on BC-STV. The pro-STV side alluded to conducting their own poll (via Angus Reid Strategies), but they never replied to my inquiries with respect to source data.

The summary results (link, local copy), and the detailed data tables (link, local copy) are made available. There were 800 people surveyed, from March 24 to 30, 2009.

60% of people sampled (478 people) did not know there was a referendum coming up in this provincial election; slightly more women than men.

Despite this, 324 (41%) said they would lean toward or vote for FPTP and 340 (43%) said they would lean toward or vote for STV. The remainder (135, 17%) indicated they would not vote or do not know how to vote.

Not surprisingly, most people know about the existing electoral system, while most people do not know about the STV system (even if the referendum were to pass, I wouldn’t think the numbers would change that much in 2013!).

Something that slipped into the summary results (last page) was the union affiliation of the sample; 29% were affiliated with a union, 70% were not. Finally, 21% of the survey had a household income of $100,000 and above, something I thought was a little strange.

The FPTP/STV 41/43 result normalizes to 48.9% / 51.1%, if you expand this out to 100%.

Also ambiguous is the question “How likely is that you will change your mind and vote for the other option on Election Day?”, as it does not separate between FPTP and STV supporters – this was a critical question that either was not asked, or was omitted from the survey details.

I would also like to take this time to thank No STV for releasing the poll results.

3 Comments

Re-Modelling BC-STV election results

Posted April 19, 2009 by Sacha Peter
Category: Referendum

Based on a comment from Tom indicating my original analysis wasn’t correct (of which Tom was correct), I have re-done my modeling for the BC-STV results, in terms of how 2005’s voting would translate into results given the 2009 proposed STV electoral boundaries.

The new results are the following:

1. Assuming no cross-partisan support:
Result: BC Liberals 43, NDP 39, Green 3

2. Assuming Green voters support all candidates, and then 75% transfer their ballot to the NDP, and 25% transfer their ballot to the BC Liberals:
Result: BC Liberals 40, NDP 42, Green 3

Details are found on the Referendum page.

Please keep in mind the standard caveats that these results in no way would model reality had the 2005 election used STV to begin with; people will change their actions based on the voting system they are dealt with, and as a result the vote likely would materially change. Also, the simple equations I used to model these election assume equal preferences for individual candidates within parties.

The conclusion, from a strictly partisan perspective, is fairly obvious:

1. If you are a BC Liberal supporter and want to see higher chances of BC Liberal majority governments, you would be inclined to vote against STV.
2. If you are the NDP, and you want to run in minority or near-minority government situations, at the cost of sacrificing your once-in-four terms in a majority government, you would want to vote for STV, but if you want the majority government chance (historically when the right-wing vote fragments) then you would be inclined to vote against STV.
3. The Greens (and all other parties) would always want STV no matter what.

Partisan interests are not necessarily congruent with non-partisans that support a particular party. A lot of people that are “disgusted with all parties and candidates” would be attracted to STV, as it would represent a change in candidate selection to a system that is more similar to those seen in at-large municipal government elections.

The primary feature of STV is that in an abstract sense, it allows for a dual selection of candidate and nomination for those voters that care about individual candidates that run. The primary disadvantage is that most of the population will not understand how the vote is counted at the end of the election. Later on, I will discuss which facts the proponents and opponents to STV are most “spun” to give a false message.

3 Comments

Sun 2009: BC-STV out in force

Posted April 19, 2009 by Sacha Peter
Category: Referendum

Just a small anecdote – after finishing the Sun Run, I noticed that there were a few people dressed in the blue “BC-STV” shirts and handing out literature near BC Place after the race. This was one of two political messages I saw during the race; the other was around the 3-4km mark, where there were some people protesting about the treatment of men in child custody/divorce cases.

8 Comments

BC-STV Referendum, what-if scenarios

Posted April 18, 2009 by Sacha Peter
Category: Referendum

(Note: Most of this article was re-worked to correspond to a subsequent analysis since my initial analysis was incorrect.  The original article is still here, except with a line through the text.)

The political math behind the BC-STV electoral system is a lot different than the standard first-past-the-post system. It is also inappropriate to assume the electoral results would be the same had STV been implemented for 2009; however, I will do so here anyway with that caveat.

The math behind STV is the primary argument against implementing the system – it makes the voting system very difficult to explain in a few words (despite what proponents may claim). The concept of “quota” and “transfer surplus” is also difficult to explain to a non-numerically oriented person, but for somebody that can handle the numbers, the system of STV is very elegant.

I have attempted a paper-napkin version of trying to establish what the Legislature would look like if STV had been established in 2005, using the 2009 STV boundaries. The results are the following (note the present system would have gave approximately 48 BC Liberals and 37 NDP members using present electoral boundaries and the First-Past-The-Post system):

1. Assume all voters are partisan animals and do not transfer their ballots to other parties:
Results: BC Liberals 46, NDP 39
Results: BC Liberals 43, NDP 39, Greens 3

2. Assume all voters are partisan animals and only the Green party supporters transfer their ballots wholly to the NDP:
Results: BC Liberals 40, NDP 45
Results: BC Liberals 40, NDP 42, Greens 3

I have posted the electoral district breakdowns on the Referendum page. Realize these calculations come with a lot of assumptions, and there is some guesswork involved, but I think it would be an accurate projection.

I also think that the Green party, based off of 2005 results, would still be hard-pressed to elect a single member of in the legislature if STV is implemented – but they would get awfully close in getting a seat in the Capital region. If STV is implemented for 2013, it is likely that enough voters will shift their first preferences to a major Green candidate (e.g. Jane Sterk) and push her over the quota and get a seat in the Legislature. The next seat would likely be one in Vancouver West.

I also find it odd, given these results, that the primary opponents of STV are David Schreck and Bill Tieleman, both NDP supporters.

19 Comments

Free-for-all, weekend of April 4-5, 2009

Posted April 4, 2009 by Sacha Peter
Category: Irrelevant and Irreverent, Referendum

This is the “weekend post”, which is intended to be more for general conversation.

This weekend’s conversation starter will be: The referendum on STV vs. First Past the Post has received very little coverage so far, other than the anecdotal news that both sides are consulting with polling agencies to see which message(s) will sell the best. How will you be voting in the referendum, and why are you voting that way? What marketing campaign do you think will be most effective for either side?

Feel free to post your comments on any topic dealing with provincial politics – it can be off topic as well and you will not incur the wrath of moderation unless if you post outright spam.

19 Comments