Sunday, August 30, 2009

The Magic Numbers

While some political parties are figuring out whether it would be feasible to force a fall election, I figured that it would be interesting to determine some “magic numbers” using the data from the most recent election. Where are the parties going to get their votes to win seats? Perhaps you have noticed that the Green Party may run May on Vancouver Island conceding that she is unlikely to defeat MacKay in Central Nova and it is important for the Greens to establish a “beach-head” in parliament. If she gets in, does that legitimize the party and attract even more new voters? From a policy standpoint the Greens are closer to the NDP than any other party in terms of ideology. With platforms like “minimum guaranteed incomes”, it was a noticeable shift towards socialist ideals. But if May thinks bleeding the NDP is the Green route to multiple seats, even if they siphoned off 1/3 NDP voters nationally, they still have no likely seats (other than perhaps May herself depending on where she runs). If they managed to attract 40% of NDP votes, now they are looking at 3 or 4 seats. In the event of that improbable demographic shift, they would be Greens sitting in a Tory Majority parliament.

Stephane Dion, like May, tried to move to the left in an attempt to outflank Layton, but in doing so a portion of his support in the center peeled off to the right. I get the science behind the strategy, that if 100% of the NDP vote were absorbed into the Liberal party, the LPC would win about 166 seats to the 97 Tory. I’m sure Liberal strategists have lost many a night’s sleep pondering schemes where the NDP would cease to exist as a formal political party. They could still have fan groups on Facebook, but they would not run candidates. That’s the dream of Liberal foot soldiers and operatives. The tragic flaw in the strategy when observed in practice is that shifting policy to the left to court that niche demographic turns out to be very unpopular in the center, and that’s where I believe the majority of people are. If a full half of NDP voters split Liberal, it would still be a Tory minority parliament with the LPC 14 seats back.

Michael Ignatieff is clearly aiming at the center and as such is weakening his left flank. If the NDP could manage to court 20% of Liberal voters, that is the threshold where the NDP would attain an equal number of seats to the Liberals. For an NDP minority government, you’d need catastrophic and virtually impossible losses in the LPC, and an NDP majority is virtually mathematically impossible. Assuming the NDP can hold reasonably steady in the next election, the Tories only need 7% of the Liberal vote to ascend to majority status. The Liberals would need 15% of the Tory vote for a tie at 108 seats. The roads to victory for the Liberal Party come with a very low probability of success rate. The number of different ways that the Conservatives can win a majority is significantly higher and with a much more likely probability of occurring than the Liberals winning a minority. The Tories have a first and goal on the Liberal 10 yard line and the Liberal D is desperate to force a turnover.

Predicting seat counts is actually really easy if you have a spreadsheet that contains the voting results riding by riding, region by region. When I predict changes in seat counts, I can quickly identify which ridings they will be. I would almost advise Ms May to move to Toronto Danforth and take on Jack Layton. If the Greens are to hold seats in parliament by eating into the left wing of the NDP, Layton’s seat is one of the most probable outcomes, aside from the headache of going head to head with Simple Jack. It would be an entertaining fight, unlike the futility of fighting a MacKay in Central Nova.

**Note that when I say 7% of the Liberal vote, I am not speaking in terms of percentage points in popular vote, instead it is a percentage of people who voted for the party. Out of every 100 Liberals, only 7 need to be converted to the Tories to form a majority.


  1. iceman: I have a feeling that any future election is going to be decided, not by people changing votes, but by people staying home. Could you try running the numbers based on potential vote losses from each party?

  2. That depends on which voters are more likely to stay at home. Tory voters have displayed greater attendance, and I don't know if you could accurately predict apathy rates. It is a parameter that is difficult to quantify. People become more likely to vote the older they get, where young people are less likely to care about who wins the election. Voter turnout declined in 2008, and it appears that the left wing suffered the worst in the apathy epidemic.

  3. Have you run numbers where the Liberals who stayed home last time actually come out next time?

  4. See how grear an increase on national vote percentage your spreafsheet says it would take to turn Kitchener-Waterloo blue using the 2006 results. You will conclude it would never happen. It did. You can ave fun playing around ith your spreadsheet, but reality can be much different.

  5. Since you asked, predicting Kitchener would have been really easy. The Tories won 35% of the popular vote in Ontario in 2006. Polling had them in the 39% range prior to the 2008 election. 4% of the popular vote in a riding with roughly 66,000 votes is 2600 new votes. The Tories received 3000 new votes to win the riding. The Liberals lost 10,000 votes, and as best as I can tell they lost 3,000 to the Tories, 3,000 to the Greens, and 4,000 stayed home. If apathetic Liberals were to return to the polls next time, that riding would very likely swing back to them. It is nice to know that frustrated Liberal staffers read my blog.

  6. It is very difficult to forecast apathy, but based in Ontario the Liberals lost votes to the Greens and the Tories. Because we have secret ballots, we don't know who changed their minds in what proportion and who would have voted what had they infact shown up. By my best estimate 15% of Ontario Liberals did not vote. If you magically added 15% to each Liberal vote total in Ontario, then 5 seats would swing back Liberal. A whole 5 seats. The idea that the Liberals can win an election by exciting the unexcited alone is a poor strategy on which to base a decision to force an election.