Cumulative Voting in Chandler

Putting off "how elections work" for one more post, this article about city spending on the new City Hall and Arizona Ave. improvements becoming the dominant issue in Chandler's election got me thinking about how these sorts of things would play out under cumulative voting.

Let's suppose that spending really is the decisive issue in this election, and that other factors like campaign spending, endorsements and incumbency play no role. Let's also suppose that most voters will vote for candidates pledging to reduce spending. Finally, let's take the candidates positions on the two projects above as indicative of their view of city spending.

Three candidates, opposed to both projects, are in the "spending is inherently evil" category: Roe, Taylor, and Weninger. One candidate, Trinity, is in the "spending can do good things" category. The other two, Hartke and Wallace, are both rhetorically and historically somewhere in the middle. Suppose voters are also divided into three camps, too. 50% are flatly anti-spending, 30% are pro-spending, and 20% are in the middle (not crazy, but worried higher spending will lead to higher property taxes while they're already hurting). For simplicity's sake, there are 100 voters.

Under the current election system (requiring 50%+1) we would expect the three anti-spending candidates to win. The anti-spending 50% all vote for Roe-Taylor-Weninger. The pro-spending 30% vote for Trinity along with Hartke-Wallace (the next closest to their preferences). And the middle 20% vote for Hartke-Wallace and then spread their third votes evenly across the other four candidates.* The results would look like this:

Candidate	Anti 50%	Middle 20%	Pro 30%	Total
Roe*	50	5		55
Taylor*	50	5		55
Weninger*	50	5		55
Hartke		20	30	50
Wallace		20	30	50
Trinity		5	30	35

* = winner

Now, let's re-do the election using cumulative voting (each voter can distribute their 3 votes however they'd like). We assume that voters have strong preferences and will cast a second/third vote for their preferred candidate rather than support a less-than-ideal candidate. Voters have no means of choosing between preferred candidates, and will choose to divide their votes evenly among them.

Anti-spending voters will divide their votes among their 3 preferred candidates. The pro-spending voters will each give all 3 of their votes to Trinity. The middle 20% will split their votes evenly between Hartke and Wallace (half will give 2 votes to Hartke, half to Wallace):

Candidate	Anti 50%	Middle 20%	Pro 30%	Total
Roe	50			50
Taylor	50			50
Weninger	50			50
Hartke		30		30
Wallace		30		30
Trinity*			90	90

Trinity would win, along with two of the anti-spending candidates. Note that the pro-spending side breaks through (and not the middle) both because it is the largest minority and because it consolidates its support behind one candidate. Suppose the middle 20% organized to only support Hartke. He would claim all 60 of their votes and win a seat as well. The proper response of the anti-spending 50% would be to only support 2 candidates, returning to equilibrium at 1-pro-spending candidate and 2 anti-spending candidates emerging from the election.

The anti-spending side still holds the policy reins, so the cumulative scenario is functionally different only in that it guarantees at least one opposition faction a public forum for its opposition. The strong incentives to organize with like-minded others created by a cumulative system also means more people organizing around city elections, which can't be a bad thing.