The regionalization of parties’ support in the 2012 Alberta election both matches and defies the predictions of many commentators, myself included. What has come true is the north-south split between the PCs and Wildrose in rural Alberta. What has not materialized is the Edmonton-Calgary split, with both cities going overwhelmingly to the PCs.
Figure 1: Regional popular vote averages
Daveberta beat me to posting regional numbers, and this table gets at the same idea, but his compiles the aggregate vote totals in a region, whereas I compare popular vote averages across regions. I compiled similar tables for the 2008 election, which can be seen here. (See the bottom of this post for some notes on methodology.)
The second table reports the party’s percentage of the popular vote in each region relative to the provincial average. This table shows the areas of relative strength and weakness of each party, which gives a different perspective than just looking at their regional percentage of the popular vote without additional context.
Figure 2: Regional popular vote averages, relative to provincial average
The PCs, though having regional variations in their support, did well across the province. Their weakest areas were inner-city Edmonton and the Greater Calgary Region—areas that went NDP/Liberal and Wildrose, respectively. In southern Alberta, where the vast majority of the seats went Wildrose, the PCs were still competitive, and their loss is due to the Wildrose’s strength moreso than their own weaknesses. However, the most interesting feature (at least to me) is that, in both Calgary and Edmonton, there are significant variations in support between the sub-regions of the inner-city, the suburbs, and the surrounding metropolitan areas that are not technically part of the cities’ corporate boundaries. The PCs are stronger in metro Edmonton and weaker in metro Calgary by about the same magnitude.
The bedrock of the Wildrose’s support is in metro Calgary and rural southern Alberta. They also did better than their provincial average in both rural central and rural northern Alberta. The difference in between two regions is that the PCs were weaker in central Alberta than in northern Alberta, which resulted in the Wildrose sweeping central Alberta. The Wildrose’s weakest area, by far, is within the City of Edmonton. However, they were only a few points under their average in metro Edmonton, which is likely as a result of their alliance with local lobby groups, such as RETA, on property rights and power line issues. I believe the over-estimation of Wildrose support in both CMAs is at least partially attributable to polling firms using CMAs as their sampling area, as opposed to just the technical city boundaries. Such errors could lead one (myself included) to believe that the Wildrose could pick up seats in Edmonton (which they did not) and that they would sweep Calgary (which they also did not). Even in Calgary, where they won two seats and have several strong second-place finishes, the Wildrose were still only a few points above their province-wide average, and finished 10 points behind the PCs in both the inner-city and suburbs. The reasons for such are being debated, but Deborah Yedlin’s endorsement of the PCs suggests that, while Calgary has a stereotypically conservative and corporate outlook, Danielle Smith and the Wildrose could be seen as potentially bad for business, given their views and the degree of change that they plan to implement.
The Liberals’ strongest areas are the city of Edmonton and inner-city Calgary. Unsurprisingly, given the strength of the NDP in Edmonton, the Liberals picked up more seats in Calgary because they faced less competition from the NDP. That the Liberals were able to hang on to five seats demonstrates the importance of incumbency and local star candidates, and their three seats in Calgary are arguably held more on a personal rather than party basis.
Similarly, the NDP is strongest in inner-city Edmonton, and Brian Mason and Rachel Notley’s seats are two of the safest seats in the entire province. The NDP also did very well in Lethbridge, where their showing was likely bolstered by local star candidate Shannon Phillips. However, the NDP has a lot of work to do in even inner-city Calgary, where prominent centre-left Liberals David Swann and Kent Hehr likely attract many traditionally-NDP voters.
The Alberta Party’s strongest showing was in West Yellowhead, where leader Glenn Taylor ran. They ran credible local campaigns in Edmonton-Rutherford and Glenora, where they were able to come close to tying with the Liberals. Their eventual goal should be to displace the Liberals, though their showing in this election shows they still have much work to do before the next election.
The non-conservative parties have some soul-searching to do. While ChangeAlberta tried to coordinate the voters on the centre and the left, the tribal loyalties that bind partisans to their party are too strong. Moreover, the centre is getting a lot more crowded with the entrance of the Alberta Party, Brian Mason’s efforts to moderate the NDP, and, most importantly, Redford shifting the centre of gravity of the PCs to the centre. Both the NDP and the Liberals have been wiped off the electoral map before, and, if 1993 is an indicator, the only way for a non-conservative party to have a credible shot at forming government is if it obliterates its rivals on the centre and left.
The Wildrose also has some figuring out to do. For whatever reason, there was a seismic overnight shift the night before the election. (That would make a great poem, if someone wants to write it: “Twas the night before E-Day…”) It is likely that the pressures of being Official Opposition will lead to a moderation of their stances on the contentious social issues that contributed to the implosion of their momentum. It is ironic that the Wildrose themselves did a far better job of proving the PCs’ ‘not-worth-the-risk’ messaging than the PCs themselves ever did. Even if they were able to sweep suburban Calgary, the Wildrose will still need to win a few seats in Edmonton and a handful of seats in rural northern Alberta if they want to form government, which means reigning in some of the more polarizing opinions of the social conservative element in their base. The other irony is that the Wildrose did not learn from Decore that contentious social issues tend to destroy campaigns They would do well to learn another lesson from the Liberals: pushing the leader out because she came up short is not a good idea. If the internal tensions between the libertarians and social conservatives explode to a full-scale civil war in the Wildrose, they will suffer the same fate as the Alberta Liberals after 1993 and walk the road to self-decimation.
The PCs were weakened, but reports of their impending demise were greatly exaggerated. Redford has lost many of Ralph Klein’s loyalists and has brought in many of her own people. If anything, she probably could have lived with losing a few more seats, but the PCs have still managed to reinvent themselves. Moreover, they are in a strong strategic position of dominating the centre of the spectrum. While Alison Redford and Stephen Carter dismiss the importance of left and right, they do so because they are simultaneously both and neither. Given that the NDP will always say spending is too low, and the Wildrose will always say that spending is too high, the PCs can point to the dissatisfaction of both sides and say that a proper compromise was reached. This presents a dilemma for the Liberals, who are being crowded out by the PCs. The Liberals either have to move left and take out the NDP, or dig in on the centre and find a way to cauterize the bleeding to the PCs. Moreover, the fledgling Alberta Party, which hopes to become the new centrist option, faces an even greater challenge in a reinvented and firmly centrist PCs party than they ever did in their original opponent of the Liberals.
Canada has tended to violate Duverger’s Law, which states that single-member-plurality (a.k.a. first past the post) electoral systems tend to be characterized by two-party competition. We have four years to figure out if this four-party configuration will hold, but the academic literature is onside with Alberta’s history in that a multiparty configuration is exceptional rather than the norm. Fortunately, there is much interest in the outcome of this election, so political scientists, pollsters, and pundits have four years to figure out what happened to make sure their 2016 predictions are a little bit closer.
A quick note on methodology:
Elections Alberta has no official definition of which ridings make up each region in Alberta. Lacking access to academic journals, I have not seen a current academic treatment of the political regions of Alberta. In an effort to follow some precedent, I follow the Wikipedia article’s classification with some exceptions:
- Each city is comprised of three sub-regions: inner-city, suburbs, and the ridings of the surrounding municipalities that comprise each major city’s census metropolitan area. This includes Airdrie, Chestermere-Rocky View, and Highwood in Calgary. ‘Metro Edmonton’ includes Sherwood Park, Strathcona-Sherwood Park, St. Albert, Spruce Grove-St. Albert, Stony Plain, Fort Saskatchewan-Vegreville, and Leduc-Beaumont.
- I find the sub-regional classifications within the two Census Metropolitan Areas to be questionable, but I stick with them, given the lack of established alternative classifications.
- Banff-Cochrane was previously classified as ‘Western and Central AB’, and I classified it as such in my previous post. It has since been moved to ‘Southern AB’, which makes much more sense.
- Each region outside of the two CMAs has been subdivided into ‘urban’ and ‘rural’ categories. In these cases, the ‘urban’ section refers to the principal regional centres, i.e. Grande Prairie, Fort McMurray, Red Deer, Lethbridge, and Medicine Hat.
The reported percentages are the mean of the percentage of the popular vote each party receives across all ridings in a region. This is calculated by adding the party’s percentage of the popular vote in each riding and dividing by the number of ridings. (This is slightly different from using the raw vote totals to calculate a regional average, but the two methods produce similar enough results, and my method is a lot quicker to execute.)