Tuesday, August 6, 2013

How safe is Toronto for the OLP?

Two of the five Ontario by-elections that took place last week were held in Toronto. The city is considered the last fortress of the Ontario Liberals, as they won 17 of its 22 ridings in the 2011 provincial election and, whenever their numbers tank province-wide, they always seem capable of holding on to a majority of Toronto's seats.

But the by-elections put a bit of a chink in that armour. The Progressive Conservatives won Etobicoke-Lakeshore with 47% to 42% of the vote, after having placed second in the riding in 2011 by a margin of 51% to 29%. The Liberals held on to Scarborough-Guildwood with a 36% to 31% win over the Tories, while the New Democrats placed third with 28%. But the Liberals had taken 49% in 2011, compared to 29% and 19% for the PCs and NDP, respectively.

If similar swings in support take place throughout the city in the next provincial election, the Liberals could see themselves win only a handful of ridings in Toronto and booted out of government - without even losing a single seat outside the city.

Of course, by-elections are special beasts and the Tories and NDP won't be able to put up candidates with the name recognition of Doug Holyday and Adam Giambrone throughout the city. Turnout was lower and voters in these ridings were not at any risk of toppling a government, so the implications of the shifts in support that occurred should not be stretched too far. Nevertheless, let's take a look at what these by-elections represent in terms of the potential threat they pose to the Liberals' other Toronto ridings.

Somewhat surprisingly, the New Democrats may not pose the biggest threat to the Liberals in Toronto by this measure (through realistically, they probably do). The swing in support that occurred between the Liberals and NDP in Scarborough-Guildwood was worth 22.1 points - not enough to put the riding in the NDP's column but more than enough to swing four ridings over to the New Democrats from the Liberals: York South-Weston, Scarborough-Rouge River, Scarborough Southwest, and York West. That these are all clustered close together gives an indication of where the NDP need to put their efforts in the city in the next election.

Apart from these four, however, the New Democrats have little scope for growth. The next riding on the list is Etobicoke North, and there the Liberals placed ahead of the NDP by almost 27 points. It means most of the Liberals' Toronto ridings are safe from the NDP. However, few of those safe seats are outside of the range of the PCs.


The total swing between the Progressive Conservatives and Liberals in Etobicoke-Lakeshore was an impressive 26.4 points. That is enough to put 10 seats inside Toronto within the grasp of the Tories - a huge number. In two of them, Scarborough Southwest and Scarborough-Rouge River, the New Democrats are better placed than the Tories. But in the other eight the swing that occurred in Etobicoke-Lakeshore would give the PCs a win.

Now, some of these require almost every ounce of that swing: Scarborough-Guildwood, Eglinton-Lawrence, Don Valley East, and Etobicoke North were all won by the Liberals by more than 20 points. But a few ridings are ripe for the picking: York Centre was won by less than 10 points, while Scarborough-Agincourt, Willowdale, and Etobicoke Centre were won by less than 20.

Holyday's might have been too much of a boost for the Tories to replicate throughout the city. But, as the chart above shows, Etobicoke-Lakeshore was pretty far down the list on ridings the Tories could have won in Toronto.

If all went perfectly for the New Democrats and Progressive Conservatives in the city, as many as 12 ridings could swing over from the governing Liberals to the opposition parties. That alone would drop the Liberals to only 38 seats, against 45 for the PCs and 24 for the NDP. And, of course, if the Liberals are losing 12 seats in Toronto they are undoubtedly losing many more outside of the city. It certainly means a majority for the PCs and Official Opposition for the NDP.

If we take out the Holyday and Giambrone factors, though, the Liberals still look vulnerable. It would not take much for the NDP and PCs to each pick up two seats on their list, dropping the Liberals to 46 seats and only seven up on the Tories - seats that would be quite easy to pick-up elsewhere (six non-Toronto ridings were won by the OLP over the Tories by less than five points in 2011, for starters).

The by-elections are not necessarily harbingers of things to come, but the results in Etobicoke-Lakeshore and Scarborough-Guildwood certainly should put the Liberals on notice that they will have to fight to hold on to some of their Toronto seats in the next campaign. Having a Toronto-based premier in Kathleen Wynne will not be enough - and if the party is focusing efforts on holding their backyard it does not bode well for their chances outside of Toronto.

28 comments:

  1. I will be interested to see if the by-elections are a better predictor of the next election this time than they have been in the recent Ontario past. During the 38th and 39th Legislative Assemblies (2003-2011), there were 15 by-elections. Of those, four resulted in party changes: Hamilton-East (2004) went from OLP to NDP, Parkdale-High Park (2006) went from OLP to NDP, York South-Weston (2007) went from OLP to NDP and Haliburton-Kawartha Lakes-Brock (2009) went from PC to OLP. Of those, both York South-Weston and Haliburton-Kawartha Lakes-Brock reverted back to their pre-by-election party at the next general election.

    ReplyDelete
    Replies
    1. The two ridings that flipped in byelections in BC between 2009 and 2013 both reverted back to BC Liberal on election day too.

      I would hold off on the doom and gloom based solely on the byelection results.

      Delete
    2. There are plenty of anecdotal counter-examples as well!

      Delete
    3. Indeed. It's almost as if the byelection results tell us nothing at all one way or the other :p

      Delete
    4. As with polls generally, by-election results tell you nothing in isolation. You can only read the implications of a poll or a by-election by viewing them in context. So, plenty of pundits and poll-watchers insisted that Mulcair's 2007 by-election victory in Outremont meant nothing for the future of the NDP or even of Mulcair himself in Quebec. However, a reading of the mood in the street... a recognition of the specific shifting political orientation of the riding and of the province generally gave a different sense of what might happen. No one predicted (or would have predicted) the magnitude of what happened in 2011, but a sizable increase in support for the NDP in QC was predictable based on the combination of poll/by-election results and a careful reading of the political mood in the province.

      Delete
    5. "careful reading of the political mood in the province. "

      And therein lies the real answer. Byelection polls really tell us noting about the majority view. Just the views within that contested riding.

      Thus as a preview of bigger results not likely !

      Delete
    6. Reading the mood in the street would still have gotten it wrong in BC though. Sure these patterns are easy to pick out in hindsight, but I don't believe it's possible to make any sort of meaningful prediction of future performance based on byelection results.

      Mulcair's breakthrough in Outremont was certainly contributed to the NDP's victories in Quebec, but it wasn't the only factor. Piss-poor campaigns by the Bloc and the Liberals, along with a very strong performance by Jack Layton at the debates and in the media all contributed to the NDP breakthrough in Quebec. Outremont didn't guarantee any of those things.

      Similarly, whether or the next election turns out to be a triumph or a catastrophe for the Liberals in Ontario remains to be determined. We don't know how well Wynne will govern or campaign going forward. We don't know how competent or incompetent her opposition will be (though history suggests incompetence for certain circles). We can't know these things today, nor can we infer them based on a handful of byelections.

      Delete
  2. I think it is not right to speak of swing because of the low turnout. I suspect the factor much more relevant in the by-elections is that a lot of OLP supporters stayed home than any change in how they voted in 2011.

    How will all these people that did not vote in the by-elections act in the next general election?

    ReplyDelete
    Replies
    1. In Scarborough-Guildwood at least, there was some swing. The NDP got 800 more votes in the by-election than they did in the general despite turnout being down eleven percentage points.

      Delete
  3. I've yet to see any correlation which is actually mathematically accurate between byelection results and following general election results.

    IMO there is NO real correlation !!

    ReplyDelete
    Replies
    1. Over the 39 Legislative Assemblies that have been held and then dissolved, 11 times the government has lost more seats than it gained through by-elections, and of those 11, six of those times the government fell. 13 times, the government gained more seats than it lost through by-elections, and of those 13, the government survived 10 times in the next election.

      What it seems is that, on a cursory overview, losing more seats than a government gains is not a good predictor of the government falling, but there may be some value in whether the government gains more seats than it loses as a predictor of government survival.

      The data sets are small, though, so its hard to draw any conclusion with confidence.

      Delete
    2. The data sets are small, though, so its hard to draw any conclusion with confidence.

      Which just backs up what I said. Thanks

      Delete
    3. I wasn't trying to disagree with you. I was trying to support you!

      Delete
    4. That's actually really interesting TS. It'd be interesting to see if there are some more meaningful trends to be teased out from a larger sample of other legislatures in the country.

      Delete
    5. It took me probably 45 minutes to pull together the Ontario data, so doing it country-wide would be a bit of a project, but I'd certainly be interested in the results if someone wanted to take the time.

      Delete
  4. The other thing is the PC win was only really attributable to a couple of Ford factors.

    Mayors second in command and the strenuous campaigning of both Rob Ford and the "Ford Nation" in that riding.

    Face facts that that can't be done in a General Election as there is simply too much territory to cover then !!

    ReplyDelete
  5. BTW - I agree with some of you commenters that too much should not be read into the by-election results (ahem, I actually said so in the post above).

    On the flip side, however, if the Liberals followed some of the above comments to their logical conclusion - that they should learn NOTHING from the results - they are certain to be on the road to defeat. What the results mean is not certain, but they certainly mean more than nothing.

    ReplyDelete
    Replies
    1. I completely agree. There is def a lot to learn here. I think the best thing that the OLP can learn from this is who is unsatisfied with them? Why did people in Etobicoke choose to vote against government? Why is Windsor so happy with the NDP?
      They may not find specific riding issues, but overall there is a lot to learn for their platform.

      Delete
  6. I am really looking forward to the possibility of so many three-way races in the city, particularly in Scarborough and York.

    ReplyDelete
  7. While governments often recover after adverse by-elections swings we must remember that last week's swings were HUGE. The Liberal overall vote fell from 47% to 29%.This shows the Liberals have a long road ahead of them, one they are unlikely to navigate successfully with an election within a year.

    It is also interesting that the PC vote held and in fact increased in four ridings. TC Norris has applied the by-election swings provincewide and projects 65 PC, 30 NDP and only 12 Liberals.

    http://tcnorris.blogspot.ca/

    It has been pointed out on other blogs that both Ottawa South and Scarborough Guildwood are among the 11 seats in Ontario that the federal Liberals retained in 2011. The provincial Liberals therefore should not be too complacent about holding onto two of their most rock-bottom seats.

    ReplyDelete
    Replies
    1. The "overall vote" is only useful if these five ridings were a representative sample of the vote in the last election, and turnout was roughly similar. Neither of those things holds true. You absolutely cannot apply the "by-election swing" province-wide as if it is a valuable predictor of the province-wide swing. Eric has also amply demonstrated that a province-wide swing model is simply not sufficiently accurate to make predictions. If it was, he wouldn't need the kind of model that he has developed.

      I say all of this as a New Democrat, most certainly not as a Liberal. I just do my best not to let my partisan inclinations get in the way of my analysis.

      Delete
  8. And according to todays Toronto Star there is a strong movement underway within the PC Party to dump Hudak ASAP !!

    Not a moment to soon !!

    ReplyDelete
    Replies
    1. Hudak is one of God's gifts to the Ontario Liberals. Turfing Hudak would benefit the PCs, not the Liberals IMHO.

      Delete
    2. I agree, Ryan. Hudak has an inspired gift for snatching defeat from the jaws of victory.

      Delete
    3. Ryan I totally agree re Liberals.

      Now minus Hudak how would the PC's do?? That's the big question. I think unless they do some major policy changes not much better at all.

      Delete
    4. Éric,

      I think Peter has it exactly right. Cosmetic changes -- of the leader or upper echelon just won't cut the mustard. Policy is where it's at. They seriously need to become more centrist.

      Look at Harper, was it a mistake to move right when he got the majority? For my money, it sure was and it could very well cost him government in 2015.

      Delete
  9. Perhaps a dumb suggestion, but could you not take the results of these bi-elections as small regional "polls" and update the general numbers for Ontario? I know there would be issues with that, but it may give us an idea of where the neighboring ridings could be swinging too, kind of like a reverse regional trend. Just a thought.

    ReplyDelete
    Replies
    1. I don't think that will work. By elections are much more concerned with the situation in the riding. i:e: When's the subway coming etc.

      A general election throws up a whole different selection of issues.

      Thus we really can't compare them.

      Delete

COMMENT MODERATION POLICY - Please be respectful when commenting. If choosing to remain anonymous, please sign your comment with some sort of pseudonym to avoid confusion. Please do not use any derogatory terms for fellow commenters, parties, or politicians. Inflammatory and overly partisan comments will not be posted. Please keep discussion on topic.