Monday, February 09, 2009

Cartoons Can Be Deceiving

Back on the anti-STV trail. Because I'm feeling generous I'll let my opponents define their system as they would see fit. Here's the vid:



Think the video has all the answers? Pop Quiz!

1. In the beginning of the video, it states that a person may choose to vote for only one candidate, in that case what happens when that vote needs to be redistributed?

2. How does BC-STV guarantee proportionality, province wide?

3. The video says every vote counts, how does the 4th place candidate (whose votes are never redistributed) have his votes counted any more than they would under FPTP?

4. What is the formula to determine the transfer value for votes transferred a second time?

5. How many times was a vote for the first place candidate counted? How many times was the votes for the 4th place candidate counted?

6. Should voting be this complicated?

Don't have the answers? Maybe switching voting systems isn't child's play after all.


We learned this fall that Canadians don't understand their parliamentary system as well as we'd all like. If you are going to vote in the upcoming referendum in BC, get informed, know what you are voting for.



Answers to Pop Quiz:

1. They don't get redistributed, because that's not possible. As I understand it, ballots that can no longer be counted are discarded and the threshold is recalculated excluding those ballots. So let's say 10,000 people vote in a riding with 3 MLA's the original threshold would be:

(10000/(3+1))+1=2501.

Let's say that after the first candidate is elected/dropped there are 200 ballots with no further preferences, the threshold would be recalculated as

(9800/(3+1))+1=2451.

In other words, the bar is constantly moving.

2. It doesn't. This is damn close to an out and out lie. In Ireland, where this system is used, results are nowhere near proportional. In the last election Fianna Fáil lost seats while increasing their share of the popular vote, something that shouldn't happen in a proportional system.

3. It's exactly the same as FPTP the votes are, in Fair Vote lingo, "wasted".

4. Well the original transfer value is calculated at

Transfer Value 1 = (Total Votes-Threshold)/Total votes

so the new value would be as follows:

Transfer Value 2 = TV1*((Total Votes for second candiate-threshold)/total votes for second candidate)

5. To avoid another formula with no values in it. Let's put it this way. The first candidate's voters got to have their opinion heard three times (albeit at different values), the fourth place candidate's voters got their opinion heard once.

6. No.

12 comments:

Wayne Smith said...

1. If you express only one preference, obviously your vote cannot be redistributed. However, your vote will still count exactly as much as it does under the current system.

2. Results in Ireland are in fact remarkably proportional, certainly far more than current results in Canada. Results are proportional because most voters will distribute their preferences by party. Voters can also distribute their preferences by any other criteria they choose, and this will also be reflected in the results. This is also a type of proportionality.

3. STV makes every vote count as far as this is mathematically possible. When the counting is done, there will still be some wasted votes, but on the order of 5-10%, not 40-60% as under the current system.

There is no perfect voting system, but some are better than others. STV is a good system. What we have now is a very bad system.

4. Not sure what the point of this question is, except to confuse people with a barrage of numbers.

5. More obscurantism. Each vote counts one full vote's worth, as far as is mathematically possible. How many times each ballot is handled is irrelevant.

6. STV voting is not complicated. You rank the candidates in order of preference. If that's too complicated for you, you probably shouldn't be voting anyway. Irish and Australian voters don't have problems with STV voting. Are you saying Canadians are too stupid to count, one two, three?

Aaron Ginsberg said...

1. As much as it does in the current system? No. Everyone's vote is counted evenly in the current system, you take a substantial hit relative to your fellow voters under STV.

2. I hate clauses that start with "most voters will..." because you have no idea how voters will act. Why on earth would a right wing Liberal voter rank any NDPers or vice versa in BC? In Ireland the two dominant parties have no major ideological differences, not true in BC. Also, how does a party see a 20% increase in support and a 65% increase in seats in a proportional system (as Fianna Fail did in Ireland in 2007)? Proportionality should mean in proportion not "everybody wins under this system so relatively even preferences will be represented relatively evenly."

3. I'm not the one claiming every vote counts, the cartoon TV guy did. I agree that if you care about the "wasted vote" crisis (about as real as the Easter Bunny), this system is better. However, don't say that this system fixes the imaginary problem. Only a PR based system with no thresholds does that.

4. You don't think it's important that people know how their vote is counted?

5. I get the mathematical argument for every vote being equal but statistics lie. Yes AFTER redistribution the transferred vote values will add up to 1. However, before redistribution the first preference is counted at full value and then is only reduced in value afterwords.

6. Australian voters don't use STV they use IRV. While similar, not the same thing. I don't think British Columbians are stupid. I just like simplicity and direct accountability in my electoral system.

Willem said...

Hi Aaron,

That video is actually totally accurate, although it doesn't always make clear the idea behind STV, which is why I prefer the following less technically accurate but more qualitatively rich description of STV: http://www.slideshare.net/STVYesCampaign/smartier-way-to-vote

Bruce

Willem said...

Something to consider as well - under STV (or another more representative system federally) you wouldn't have Liberals shut out of Alberta or Conservatives out of all the big cities. You wouldn't have a huge separatist party dominating Quebec either. See:

http://dangrice.com/?q=node/226

Anonymous said...

Aaron always the perpetual critic...

If you don't like STV do you think the status quo is GOOD FOR THE COUNTY?

Here'a a pop quiz for you:
1) Is it good to see the government of the day switch from being city-dominated in one election to have NO MEMBERS from Toronto, Montreal or Vancouver the next?

2) Is it good to see the government of the day switch from having essentially NO representation from Alberta, Saskatchewan and Manitoba to being western dominated and then back again? (if the Libs win the next election I GUARANTEE they'll have 4MPs max from those 3 provinces).

3) Is it good that there is NO HOPE of patterns 1) and 2) changing over at least the next 3 elections regardless of who wins?

4) Is it good to see the Bloc Quebecois get 50 seats with 900,000votes while the Green Party who got over a million gets none?

So what's your solution to these things Aaron? Or do you just want to keep complaining?

Anonymous said...

Aaron,

You take no 'hit' to your vote with STV. In fact, your vote is much more likely to help you elect a candidate you support. The essence of BC-STV is that you get elected if you win a certain number of votes - 17,000 on average across BC. This is the basis of its proportionality.

Over 81% of voters will likely directly contribute to determining the outcome of the election, vs only 36% in 2005 with First Past the Post. Similarly, while barely 50% of voters in 2005 ended up with an MLA they voted for, over 90% will likely do so with STV. I'm fine if you want to say that 'nearly every vote counts' or 'with STV, over twice as many votes will count as with FPTP'.

I think you've misinterpreted Wayne - he was arguing that voters would 'distribute votes by party' - ie, that they'd give their votes primarily to their preferred party. The fact that these votes will coalesce on the preferred candidate(s) from each party are what give STV its proportionality.

Speaking of which, you've misrepresented what happened with Fine Gael in 2007 (not Fianna Fail) - in 2002, they had 22.5% of the vote and 18.7% of the seats, an under-representation of 4%. In 2007, they increased to 27.3% of the vote, and won 30.9% of the seats - an over-representation of 3.6%. These are among the most extreme discrepancies that occur under STV - deviations of up to a maximum of about 7% between a party's share of the popular vote and their share of seats. The average value is about 3.5%, compared to BC's average deviation of about 20% (remember the Libs getting 97% of the seats on 57% of the vote - that's a 40% deviation!). 40% in BC will win you anywhere from 20% to 70% of the seats - how logical is that?

One other minor point: Australians use STV for their Senate and in at least one house in all states but one.

Aaron Ginsberg said...

Mike, when you're ready to talk about the issue at hand instead of reciting Fair Vote talking points about federal politics, post something. I will say, as I've said before, that the burden lies with those advocating a particular form of change to defend it and prove it not just to be an alternative but a better alternative. Change for the sake of change is folly.

Antony, glibly, it depends what the meaning of is, is. First on the Ireland thing, you are right on the parties, however:

51/31=ca 1.65 or 65% increase
27.3/22.5= ca 1.2 or 20% increase

Those are the percentages. Now in terms of percentage points, that's a different matter. Under PR with those popular vote totals the seats would have gone from 37 to 45 which is an almost identical increase to the popular vote. That's proportionality.

Proportionality, traditionally, is a measure applied to parties. You get x% of the popular vote, you get x% of the seats. When people advocate Proportional Representation (PR) that's what they're talking about. Any other definition is really a bastardization of the term. I understand the definition your giving, but its difficult to accept because the proportion is only 17,000 votes within one mega-riding:1 MLA. It has no application to defeated candidates or parties in general. In other words, a defeated candidate receiving 8,500 votes does not get 1/2 a vote in the legislature and a party receiving 170,000 votes doesn't necessarily receive 10 seats in the legislature.

Australians use IRV to elect their lower house, which is what I was talking about. A different form of STV is used in the senate.

Anonymous said...

Aaron based on that logic you ought to stop criticizing the Conservative government entirely because "the burden lies with those advocating a particular form of change to defend it and prove it not just to be an alternative but a better alternative"

I think a lot of Liberals spend more time criticizing the Conservatives that promoting the Liberals as a positive alternative is that wrong? Isn't that what you've done?

The defense of a system of PR is that it resolves the 4 issues I lined up. Why don't answer my question as to whether you think the status quo is good for the country or whether it might be good to CONSIDER electoral reform at the national level (note I'm not getting into the debate on BC STV because I don't live there and nor do you).

Aaron Ginsberg said...

Mike, I think that M. Dion made a fairly clear argument for what the Liberals would do differently in the last election. A scandal, like the Cadman Affair, is not the same as changing centuries of tradition and rightfully different standards are applied.

In terms of national electoral systems? I've yet to see a system that as accurately represents the regional diversity of this country as the current one does. I think it is important that a vote for an NDP candidate in St. John's East is cast for different reasons than a vote for an NDP candidate in Vancouver East. A vote for a Conservative in Calgary is also very different than a vote for a Conservative in Quebec. Even a region based system would have difficulty accommodating small but distinct regions like Labrador or Nunavut.

Willem said...

Hi Aaron,

I find a lot of what you write doesn't make sense to me or at least requires justification. For example, how does our current system represent the regional diversity of this country, and why wouldn't the proposals on the table be able to do better?

Electoral reform has been looked into in great depth by the CA, the LCC, etc, and they recommended change.

Bruce

Willem said...

Aaron, what specific criteria do you think are important that voting systems should be compared using?

Anonymous said...

Aaron - are you saying you value proportionality? My impression is that you're actually opposed to it. In the latter case, there's little point in arguing about the extent to which STV is proportional. I will certainly grant that it's not 'perfectly' proportional, but the CA realized that British Columbians didn't want 'perfect' proportionality through the normal mechanisms for providing it (ie, Israel-style list party systems). STV balances regional focus with proportionality - almost everyone helps elect someone local from their preferred party, and, to the extent that there's any local disproportionality, you can contribute to the election of someone from a different party who most shares your views. This whole notion of 'party proportionality' overlooks the substantial benefit STV offers of allowing voters to choose specific candidates who best match what the voters actually want in their representative, whether this is party affiliation, issue expertise or familiarity with geographic, gender or ethnic issues. 'Party Proportionality' is but one element in a broad range of considerations.

On the off chance that you do actually value proportionality, are you seriously contending that a system that decreases the average disproportionality from close to 20% to 3.5% is not an improvement? (especially when most other MMP-type systems, as in Germany or New Zealand, offer similar disproportionalities). Again, I don't dispute your calculations. I'm sure you could envision a scenario in which a small party has their share of the vote unchanged at about 3% nationwide in some mythical country over two elections under STV and in one election they win 0 seats and in another they win 1. In this scenario, a 0% change of the vote would have resulted in an infinite improvement in their representation. How shockingly unproportional!!! What a travesty of a voting system!!! Keep FPTP at all costs to avoid this, Aaron would say. :) In contrast, most reasonable people would see this 3% discrepancy in the party's share of seats vs the popular vote as a normal result of the way STV works. It would just be smoke and mirrors to refer to this as an infinite% change in representation - technically correct, but irrelevant.

And I would just emphasize that you have picked the one of the most extreme examples of relative change in seats to vote share that I'm aware of. In BC, the NDP once steadily dropped from 45% to 43% to 41% to 39% of the vote over four elections. They lost the first two and won the second two. From 1986 to 1991, they went from 42.6% of the vote to 40.7%, increasing their seats from 32% to 68%, so their ratio of seat change to vote change was 36%/-1.9% = -1900%!!!!!!!! If you're arguing that this is a proportional change, I don't know what I could possibly say to you. Such swings are actually quite normal under FPTP. If you really want, I could dig out dozens of similar examples. I strongly doubt this is the most extreme example I could find.

All views expressed in this blog are those of the author and the author alone. They do not represent the views of any organization, regardless of the author's involvement in any organizations.

All comments are the views of the individual writer. The administrator reserves the right to remove commentary which is offensive.

The author is not responsible for nor does he support any of the advertisements displayed on the page