Adam Goldenberg is a Kirby-Simon Fellow at Yale Law School. He was a Liberal staffer on Parliament Hill and at Queen’s Park. Follow Adam on Twitter at @adamgoldenberg.
Last week, I argued elsewhere that the Supreme Court of Canada should overturn an old piece of judge-made election law, the magic-number test, which says that an election should only be voided if the number of tainted ballots exceeds the margin of victory:
“The magic number test implies that, in an election, who won matters more than who voted for whom. If it were otherwise, the margin of victory would be irrelevant; only the voting irregularities themselves, their number and severity, would decide whether or not a new election is justified. Instead, our courts have allowed efficiency to limit democracy—in the very cases where the latter is most in question. . . .
The magic number test is too simple, too crude. It fails to distinguish between ‘the result of the election’ and the outcome—the result being the final vote tally, the outcome being the winner.”
“The question for the Supreme Court, under a truly democratic interpretation of the Canada Elections Act, is not who should or should not have won, but rather whether the result — that is, the number of votes cast for each candidate — was affected by Election Day irregularities. Votes for Mr. Optiz and Mr. Wrzesnewskyj are not the only ones that matter. Why count every vote unless every vote counts?”
Some very smart people think I’m nuts. It’s wrong, they say, to distinguish between the result of an election and its outcome, particularly since, under section 524(1)(b) of the Canada Elections Act, “the result of the election” is what matters—under the Act, if “the result of the election” was “affected” by “irregularities, fraud or corrupt or illegal practices,” then a court can toss it out and order a new vote.
Some very smart people think that, unless we want to have a by-election every time a few ballots go awry, “the result of the election” must mean the winner, not the raw number of votes cast for each candidate. If the number of “irregularities” is smaller than the margin of victory, they say, then “the result of the election” cannot possibly be “affected.” In other words, some very smart people think the magic-number test works just fine.
Start with the Act itself. Section 524(1) indicates that “the result of the election,” means the number of votes cast for any candidate, not just one of the first- or second-place finishers. It reads as follows:
“Any elector who was eligible to vote in an electoral district, and any candidate in an electoral district, may, by application to a competent court, contest the election in that electoral district . . . “
If “the result of the election” were the same as the outcome—that is, if the “irregularities” in question could only “affect […] the result” if there were enough of them to change who won—then why would the Act allow any candidate, not just one of the two who could be “affected,” challenge the “result” in court?
My guess: it would not. The Act gives every candidate the legal standing to challenge the “result” because the “result” is something that affects every candidate; the “result of the election” is the final vote tally, which shows who voted for whom, not just who won.
But perhaps, as some very smart people claim, section 524(1) permits “any candidate in an electoral district” to “contest the election” in court because some candidates are not also “elector[s] who [were] eligible to vote” in the same constituency. True, there is no rule against running in a riding in which you do not live, and if the Act only allowed eligible electors to challenge irregularities, it could leave certain candidates powerless to contest the result of an election in which they themselves were on the ballot.
Fair enough, but the question yet remains: Why would the Act allow “any candidate” to contest an election “result,” if the “result” were not something that could affect any candidate?
You may not be convinced. But if you still believe that the Supreme Court should uphold the magic-number test, or if you still think that the Canada Elections Act says that a court can order a new election only if it finds enough irregularities to affect who won—if you agree, in other words, with some very smart people who say that “the result of the election” is the same as the outcome—then consider this imaginary example from the not-too-distant future:
In late 2012, Prime Minister Stephen Harper calls a by-election in Calgary Centre, to replace retiring Conservative MP Lee Richardson. On election night, the Tory candidate wins handily, while the Liberal hopeful finishes a distant third.
Two years pass. With six months to go until the 2015 federal election, the Liberals still lag in the West, despite the Pacific pedigree of their new leader, David Merner, the former president of the party’s British Columbia wing. Facing a Western wipeout, but desperate to keep Liberal hopes alive west of Winnipeg—and to help Merner keep his job—the rookie leader’s campaign staff come up a cunning plan: a surprise surge to second place in Calgary Centre, the site of the party’s humiliation two years before. Winning the riding would raise too many eyebrows, but a runner-up result in the Conservative heartland would still suggest, at the very least, that the party still has a pulse past the Prairies.
On Election Day, everything goes according to plan. Liberal voters from neighbouring ridings register to vote in the targeted riding with fraudulent identification, as Liberal partisans disguised as Elections Canada officials “forget” to fill out the necessary paperwork. The ballots are counted, and the pundits are perplexed; the Liberals have come second in Calgary Centre—far behind the Conservatives, but well ahead of the NDP. David Merner becomes the first Liberal leader since John Turner to keep his job after an election defeat.
In this hypothetical election, the outcome is obvious: the Tory candidate is the winner. But is that the same as the result?
Say the final vote tally looks something like this: 30,000 votes for the Conservatives, 26,000 votes for the Liberals, and 24,000 for the NDP. If, under the Canada Elections Act, “the result of the election” is the same as its outcome, then the NDP’s votes are all but irrelevant, as is section 524(1)(b); the number of ballots tainted by “irregularities, fraud or corrupt or illegal practices” would have to exceed 6,000—the difference between the Conservatives’ vote count and the NDP’s—before a dispute involving the third-place finisher could possibly “affect the result.” If there were more than 4,000 voting irregularities, then a court could still use the magic-number test, but to benefit the Liberals, not the NDP.
Now, say the number of voting irregularities—ballots cast for the Liberal candidate as the result of “irregularities, fraud or corrupt or illegal practices”—is 2,500. The NDP candidate could still challenge the election in court, under section 524(1) of the Canada Elections Act, but to no avail; the number of tainted ballots still does not exceed the Tory candidate’s margin of victory, as the magic-number test requires. The challenge would fail. The final vote tally would stand.
Tell me: In this example, is “the result of the election” the same as the outcome? You may still believe that the latter matters more than the former, but the two are certainly not the same.
If the Supreme Court upholds the magic-number test when it decides the Etobicoke Centre election dispute, then a future court, faced with the Liberals’ Calgary Centre caper, would be unable to order a new election.
But if the Supreme Court rejects the magic-number test—and states clearly that “the result of the election” means the final vote tally, not the winner—then a future court would have some wiggle-room in such an exceptional case, as well it should.
We tell ourselves that, in a democracy, every vote matters. Our laws should say the same.
In the comments below, a reader asks—as have several others on Twitter —for an alternative test to the magic-number rule. After all, if “the result,” under the Act, means the number of votes cast for each candidate, then any number of tainted ballots could be enough for a court to overturn an election. Rather than hold a by-election every time a voter registration certificate goes missing, some other standard ought to apply.
Let me suggest one: reasonableness.
Ignore, if you will, the sound of everyone who has ever gone to law school groaning in unison; for centuries, and to law students’ collective chagrin, lawyers and judges in the English-speaking world have put “reasonable limits” on just about everything. Reasonableness shows up first-year Contracts and Torts, and even in the Canadian Charter of Rights and Freedoms, section one of which puts reasonable limits on the rights and freedoms that follow.
If the Supreme Court overturns the magic-number test, it can substitute a standard of reasonableness in its place, and set “reasonable limits” on a court’s determination that some number of irregularities “affected the result of the election” under section 524(1)(b). The meaning of “affected the result” is already limited by the magic-number test—moving that internal threshold to a standard of reasonableness seems, well, reasonable.
But what is reasonable? When can “irregularities, fraud or corrupt or illegal practices” reasonably be said to have “affected the result of the election?” A court might put the question this way:
Are the proven instances of “irregularities, fraud or corrupt or illegal practices” sufficiently severe and numerous so as to bring the democratic process into disrepute, such that a new election is a reasonable remedy?
If the answer is “yes,” then the court should follow section 531(2) of the Canada Elections Act, and invalidate the result. If the answer is “no,” then it should not.
Hang on, you say. Surely this formula gives judges far too much leeway. Surely we need a bright-line rule, based on simple arithmetic.
But broad judicial discretion is already written into the Act. Section 531(1) provides that “[t]he court may at any time dismiss an application if it considers it to be vexatious, frivolous or not made it good faith.” Such a decision cannot even be appealed.
When the facts may differ dramatically from one dispute to the next—and where the integrity of our democracy is potentially at stake—why would we handcuff our courts to a single, simplistic rule? We pay judges to make tough decisions. We should let them.
At the end of the day, most cases will produce similar judgments, whether or not the Supreme Court keeps the old test. A disputed election in which there were 50 proven irregularities and a margin of victory of 15,000 votes would almost certainly flunk both the magic-number test and the proposed reasonableness standard.
But in exceptional cases, like the hypothetical Calgary Centre controversy described above, the reasonableness standard would at least give judges the power to decide on a case-by-case basis. The magic-number test would not.