Winnipeg’s mayoral race, how the middle is being won and lost
From the beginning, there have been two front-runners in the race for mayor of Winnipeg: Judy Wasylycia-Leis, a left-wing former NDP cabinet minister and MP, and Gord Steeves, a former Liberal and city councillor distinguished from his time on council as a moderate.
Past behaviour, Dr. Phil teaches, is the best predictor of future behaviour. Not so in this race. To the contrary, both Wasylycia-Leis and Steeves have conducted themselves in this race in ways that one might find surprising given their histories as elected officials. Despite her history on the left, Wasylycia-Leis has run for the most part a quiet, centrist campaign. Steeves, in sharp contrast to his history as a compromising moderate, has run a campaign marked by stridently right-wing policy proposals. What can explain these seemingly surprising campaign strategies?
Political scientists may tackle this puzzle by marshaling the median voter theorem. The theorem states that on a one-dimensional political spectrum from extreme left to extreme right with a normal (bell-curved) distribution of views, the preference of the median (centrist) voter will prevail under majority rule voting systems. The theorem has obvious implications for how candidates conduct themselves in election campaigns. Most notable of these implications is the “rush to the centre”: candidates face incentives to converge on the median voter in order to corral the significant number of votes in the centre of the spectrum.
How does the theorem do at predicting the candidates’ behaviours in this campaign? For Wasylycia-Leis, it performs well. She started the race as a perceived front-runner and, after tossing some policy scraps to her base on the left, promptly moved on to the mushy centre, claiming it and the vast expanse of the ideological spectrum to the left as her own.
In two-candidate races, candidates can converge on the centre and neglect their supporters on the left or right, safe in the knowledge that more extreme voters have no one better to cast their ballots for (in simple spatial models, people vote for the closest candidate to them even if the distance is substantial). But in multi-candidate races, candidates can upset centrists’ plans by invading from both the left and right. The entrance of neophyte candidate Robert Falcon Oullette raised the prospect of an invasion from the left, in which case Wasylycia-Leis would be flanked and could not count on the support of left-wing supporters. But Wasylycia-Leis’ campaign seems to have predicted that Oullette’s attempted invasion would not succeed, and recently polling demonstrates that they were largely correct.
For Steeves, the story is more complex, but the theorem still performs well. The key distinction between Wasylycia-Leis and Steeves is the nature of the competition they faced on the left and right respectively. Once the campaign began in earnest, two high-profile candidates—Brian Bowman and Paula Havixbeck—appeared on the right side of the political spectrum. This was not so much an invasion as an occupation, and Steeves could not hope to win as a moderate against Wasylycia-Leis while flanked to his right.
Steeves was forced into a two-stage strategy.
What to do? Steeves was forced into a two-stage strategy. In the first stage, he sought to consolidate support to his right and thus remove Bowman and Havixbeck as credible challengers. Having done so, Steeves could then in the second stage converge on the centre to challenge Wasylycia-Leis over the median voter.
Thus, Winnipeggers were surprised to see a formerly centrist, moderate city councillor making right-wing, headline-grabbing policy announcements once the campaign began in earnest. This was Steeves in Stage 1, tossing red meat to conservatives in order to consolidate his support on the right side of the spectrum. But it now appears that Steeves will never have the opportunity to move on to Stage 2, as recent polling demonstrates his attempts to remove Bowman as a credible challenger on the right have not been successful.
The past stated beliefs of both Wasylycia-Leis and Steeves perform poorly when it comes to predicting their behaviours in this race. It is clear that both candidates have had to makes changes to their own personal politics in response to the incentives of this particular race. In contrast, the central voter theorem performs very well at predicting their behaviour once the actual structure of the race is taken into account. Winnipeg’s 2014 civic election appears to substantiate Anthony Downs’ famous and provocative claim that “Parties [or, in this case, candidates] formulate policies in order to win elections rather than win elections in order to formulate policies.”