Kultrun and Nonviolence Festival: Election Simulation Results

Ballot for Riding 3

Ballot for Riding 3

(Our apologies for getting this out so late.)

On July 11 and 12 the Waterloo Region chapter of Fair Vote Canada re-ran its Mixed-Member Proportional Representation (MMPR) election simulation. This time, we staffed booths at two polling stations: one at the Kultrun festival on July 11 and 12 in uptown Waterloo, and one at the Nonviolence Festival on July 12 at Victoria Park in Kitchener. Although the simulation was identical to the one we ran during the Multicultural Festival, the results were quite different, and quite educational in understanding how First-Past-the-Post (FPTP) voting systems compare against MMPR.

There were 40 valid votes cast at the Nonviolence festival (including one declined ballot) and 85 cast during the Kultrun festival, for a grand total of 125 ballots.

The biggest news of this election is that there was a tie in the riding results: out of 18 valid votes cast in Riding 3, five riding votes were cast for Chocolate (of the Ice Cream Party) and five were cast for Avocado (an independent candidate!). These results held up even after a tense recount of the ballots. Therefore, the winner of the riding had to be determined by flipping a coin, which hardly seems democratic, but does seem like an appropriate metaphor for how FPTP systems work. Avocado won the coin toss, and thus won Riding 3.

This tie creates a few different “what if” scenarios to explore. First, let’s present and explore the official election results under both FPTP and MMPR. Next, we can look at an alternative universe where Chocolate Ice Cream won the coin toss instead of Avocado. Finally, we can look at different ways to allocate list seats for our MMPR election, since we — hopefully unintentionally! — once again neglected to show the closed lists for the MMPR election.

Election Results: Avocado Wins!

Riding seats for both FPTP and MMPR are determined by who has the most votes in the riding. Since Avocado won the coin toss in Riding 3, the riding winners are:

  • Riding 1: Pear (Fruit Party), 37.9% of the riding vote
  • Riding 2: Watermelon (Fruit Party), 32.0% of the riding vote
  • Riding 3: Avocado (Independent), 27.8% of the riding vote
  • Riding 4: Steak (Beef Party), 28.6% of the riding vote
  • Riding 5: Salmon (Sandwich Party), 42.9% of the riding vote
  • Riding 6: Zucchini (Independent), 33.3% of the riding vote

Out of the 125 votes cast, 42 of them helped elect a winning candidate, which means 66.4% of the votes were wasted in the riding vote.

Here is a chart illustrating the FPTP winners:

Chart of FPTP results

The Fruit Party would have been the big winner in this election, although independent candidates did remarkably well (which is very infrequent in real FPTP elections).

Another interesting result was that even though the Ice Cream and Pie parties did well in the Multicultural Festival election simulation, neither of these parties won riding seats in this election. As we shall see, the party vote results tell a very different story.

Now let us consider the Mixed-Member Proportional election, which takes party votes into account. Once again, we used the “list-free” method of determining list seat winners, which you can read about in the Multicultural Festival writeup. Here were the list seat winners (which are in no particular order):

  • List Seat 1: Chocolate (Ice Cream Party)
  • List Seat 2: Mint Chocolate Chip (Ice Cream Party)
  • List Seat 3: Cookie Dough (Ice Cream Party)
  • List Seat 4: Cherry (Pie Party)

Here is a chart illustrating the overall winners:

Chart of MMPR results

It may seem strange that the Ice Cream Party wins so many list seats, but it shouldn’t. Many people supported the Ice Cream Party with their party vote, but the party won no individual riding.

Under the “list-free” selection system, if the Ice Cream Party had won only a single list seat, that seat would have happened to go to Chocolate — the very candidate that had lost the coin toss in Riding 3. This is because Chocolate had a greater percentage of its riding vote than any other member of the Ice Cream party. In other words, the tie result would not have been that catastrophic for the Ice Cream party under MMPR as it was under FPTP (where the Ice Cream Party would have had no representatives elected at all).

The two independent seat wins skew the proportionality of this result quite a bit, and the Pie party is the big loser as a result. This is similar to the skew in the Multicultural Festival result, when the Ice Cream Party was the big riding winner. The underlying problem is that there are not enough list seats in our small election to make the overall results proportional. One way to address this is to add list seats when there are overhangs that cause disproportionality, but the Ontario Citizens Assembly from 2007 chose not to include overhang seats in its proposal.

Hypothetical Election Results: Chocolate Ice Cream Wins!

Now let’s look at a situation in which Chocolate Ice Cream had won the coin toss in Riding 3. What would be different?

In this case, the riding winners would be the same, except for riding 3:

  • Riding 1: Pear (Fruit Party), 37.9% of the riding vote
  • Riding 2: Watermelon (Fruit Party), 32.0% of the riding vote
  • Riding 3: Chocolate (Ice Cream), 27.8% of the riding vote
  • Riding 4: Steak (Beef Party), 28.6% of the riding vote
  • Riding 5: Salmon (Sandwich Party), 42.9% of the riding vote
  • Riding 6: Zucchini (Independent), 33.3% of the riding vote

The number and percentages of wasted votes would be identical in this case, although the specific voters who elected winners would change.

Here is a chart illustrating these results graphically:

Chart of FPTP results

This makes a big difference to the Ice Cream Party, because it gets pretty good representation. The Pie party is the big loser here, as are voters who supported Avocado in Riding 3.

How about when we take party votes into account, running an MMPR election instead?

Chart of MMPR results

The List Seat winners would have been as follows:

  • List Seat 1: Cherry (Pie Party)
  • List Seat 2: Pumpkin (Pie Party)
  • List Seat 2: Mint Chocolate Chip (Ice Cream Party)
  • List Seat 3: Cookie Dough (Ice Cream Party)

As it turns out, there would have been further controversy with respect to the list seats filled by the Pie Party. Both Pumpkin Pie and Apple Pie won the same percentage of votes in their ridings, and they tied for second place in the party lists. Again, a coin toss was used to resolve the winner.

We can see that proportionality makes a big difference now. The Pie party gets good representation, as does the Ice Cream Party. The Fruit Party and Beef Party remain underrepresented (in part because the number of ridings are so small), and sadly the Avocado independent is the big loser. In real elections, independents tend to fair poorly under FPTP (because they do not have party support) and MMPR alike (because MMPR depends on party support to achieve proportionality).

On the other hand, if two party candidates end up in a tie under the MMPR system with list-free seats, then it is likely that both candidates will win seats. This is not necessarily the case with closed party list MMPR systems, however. We will explore closed party list systems next.

Closed Lists

The MMPR electoral system proposed by the Ontario Citizen’s Assembly in 2007 used closed party lists, where the order of list seats was determined before the election. Understandably, many people disliked this aspect of the system because it put control of the party lists more with the parties and less with the voters. However, it is important to emphasize that even an MMPR system with closed party lists will be more proportional (and almost certainly more representative) than the first-past-the-post system we use now. Furthermore, switching from an MMPR system with closed party lists to one with open lists or a list-free selection system is much easier than jumping from FPTP to any kind of proportional election system.

Once again, we neglected to publish the party lists before the election, but for fun we can explore the results of the election if we had used them. Let’s consider the “real” election result, in which Avocado won the coin toss. Then the allocations of list seats would have been as follows:

  • List Seat 1: Ice Cream Party
  • List Seat 2: Ice Cream Party
  • List Seat 3: Ice Cream Party
  • List Seat 4: Pie Party

This is the same party allocation as above, but the winning list seat candidates might be different. As it turns out, here are the party lists we intended to publish for these two parties.

Ice Cream Party:

  1. Vanilla
  2. Chocolate
  3. Pistachio
  4. Butterscotch
  5. Neapolitan
  6. Wild Cherry
  7. Fudge Ripple
  8. Maple Nut
  9. Rocky Road
  10. Coffee

Pie Party:

  1. Key Lime
  2. Lemon Meringue
  3. Apple
  4. Cherry
  5. Blueberry
  6. Mincemeat
  7. Pumpkin
  8. Pecan
  9. Custard
  10. Raisin

Not all of the candidates on these party lists ran for riding seats, and in some cases (such as Mint Chocolate Chip or Cookie Dough Ice Cream) candidates that ran for local ridings did not appear on the party list. It is not even necessary that candidates that run in ridings appear ahead of those who do not — for example, Key Lime Pie was not a local candidate in our election, but appeared first on the Pie Party list. Parties are permitted to choose party lists in whatever manner they wish, so long as they publish these lists before the election and submit the details about how these lists were selected to Elections Ontario. (We did not simulate this reasoning in our election simulation.)

In this case, neither the Ice Cream nor Pie parties won any riding seats, so the list seats would have been allocated using the top names of their lists:

  • List Seat 1: Vanilla (Ice Cream Party)
  • List Seat 2: Chocolate (Ice Cream Party)
  • List Seat 3: Pistachio (Ice Cream Party)
  • List Seat 4: Key Lime (Pie Party)

Say there was some election in which Chocolate had won its riding seat but the Ice Cream Party still deserved three list seats (which was not the case in either scenario above). In that situation, the three list seat winners would have been: Vanilla, Pistachio, and Butterscotch.

Note that there is no guarantee that Chocolate would have won a list seat under a closed list system. It just happened to appear high in its party list.

Further Exploration

Full details of this election are available in this zipfile (458.9 kBytes). This file contains spreadsheets which calculate winners under the two scenarios. You can also see how the votes broke down over the two polling locations.

Soon we will publish a toolkit containing the PDF files and source files we used to create this election simulation, so that you can run a similar election yourself. Alternatively, if you would like the chapter to help you run this simulation for a group of people (for example, a classroom) then contact us and we can help you out.

Election Simulation at Kultrun

On Saturday, 13 July 2014 the Fair Vote Canada Waterloo Region Chapter ran another Election Simulation at the Kultrun music festival in Uptown Waterloo.

Sharon Sommerville shows Elisabeth Daub how an MMPR ballot works, then Elisabeth votes for her favorite food and food group.

This weekend Fair Vote Waterloo had a second polling booth for the Simulated Election at the Nonviolence Festival. The results will be combined, the ballots tallied, and results should be available by Wednesday, 16 July 2014.

Vote this weekend using MMP!

logo

Nonviolence Festival

Hello FairVoting Friends,

This weekend FVC -WR will be at Kultrún, the world music festival at Waterloo Square and the Nonviolence Festival on the island at Victoria Park in Kitchener. Both are wonderful festivals.

FVC – WR is doing a demonstration election using MMP with two polling stations, one at each festival!

Kultrun logo

Kultrún

The mock election illustrates that PR isn’t complicated or difficult to understand. We had a demonstration election as the Multicultural Festival. People had fun & found MMP to be helpful in solving the problem of liking the candidate but not the party and vis versa. The results from this weekends election will be posted on our website a couple of days after the festivals.

You see the Multicultural Festival Election Simulation Results on our website.

Join us at Kultrun & the Non-Violence Festival to cast your vote using Mixed Member Proportional.

See you soon,
Sharon
for FVC -WR