Pierre Poilievre Responds


Sometimes, when you howl at the moon, the moon howls back.

In this metaphor, Pierre Poilievre is the moon.  And I am some sort of wolfman.

Point being: the Minister of State for Democratic Reform has responded to my needling of his Fair Elections Act with an op-ed, sent to Maclean’s Magazine.

Poilievre makes some fair points in his letter, but, alas, I fear he’s cherrypicked some of the issues I’ve raised with my analysis of his position.

So, in the name of dialog and whatnot, I’ll go through his letter and see where we stand.

Vouching Problems Are Serious

ID laws should be, too

Canadians can use 39 forms of authorized identification to prove who they are and where they live.  If they fail to bring any of those to the polls, the law allows someone else to vouch for them.  Due to massive irregularities in the use of vouching, the Fair Elections Act would end the practice.

Kudos for expanding the list of ID.  Really good job.  Fantastic.  Pat on the back.

But let’s hold off on the “massive irregularities” for just a minute, here.

I have pointed to an Elections Canada-commissioned Compliance Review, as evidence of the problem.  It included audits of four ridings in which, the level of irregularities for vouching averaged 25 percent.”(P.15)

“Compliance Review” sounds so boring.  Let’s call it the Neufeld report.  And you can find the easy-to-read HTML version, here.

Now, freelance journalist Justin Ling, has questioned both this fact and its importance.

The fact comes from page 15 of the above-mentioned compliance review, authored in 2013 by Harry Neufeld, the former Chief Electoral Officer for the province of British Columbia.

Mr. Ling correctly points out that the national rate of irregularities is actually 42%.  The difference arises from the fact that the 25% number is an average of four ridings, and the 42% is a nationwide figure.  At worst, I have understated the problem.  Either way, both numbers are correctand both are shockingly high.

Yes, 25% is the observed irregularities rate for the 2011 election in Etobicoke Centre, as well as the byelections in Victoria, Durham and Calgary Centre.  42% is the Neufeld estimate for the entire country.  You’re welcome.

Are they important?  Mr. Ling argues that they are minor paperwork glitches having no impact on the integrity of the vote.

Yet that is not what the reports author found. Mr. Neufeld defined irregularities” as serious errors”—his wordsnot small details.  “An irregularity is a failure by an election officer to administer safeguards demonstrating that a voter is entitled to receive a ballot.” (p.63)

They are, by and large, paperwork errors.  Let’s look at the types of irregularities identified by the audit:

  • Instances where vouching was required and there is no record in the Poll Book clearly indicating both who the voter and who the voucher are: 90.4%
  • Number of vouchers who vouched for more than one voter: 0.07%
  • Instances where the voucher was registered to vote in a different polling division than the voter: 0.10%

To be clear, in the latter two categories, the person should not have been allowed to be vouched for.  That they were allowed was an error by the poll staff.  In the first, largest, category — there are plethora of reasons why this would happen, and we have no more details that what is given in the audit.

Now, you may be thinking “Hey, where’s the other 10%?”  (Or, if you are quite clever, “where is the missing 9.43%?”)  It’s not entirely clear, though I am pretty sure it’s because these numbers are estimates and are extrapolated from very, very small sample sizes from these three byelections.  The total number of irregularities associated with vouching that were identified in these ridings is a whopping 269 votes.  Thus, I suspect there is some mismatching numbers and imprecise figures.

Let’s pause for a second and consider something critical: of all those vouching-related irregularities, 192 came from the 50 polls chosen as a sample from the Victoria by-election.  That’s certainly curious.  When you remove Victoria from the mix and consider just the other two byelections, you get a more reasonable 15% irregularity rate.  Victoria, as the biggest sample size of the three, may have also tipped the scale when calculating the national estimate.  Considering that thousands of people vote in each poll, and there were less than 200 irregularities, inept staff or poorly-kept records in one or two of the polls may have contributed to inflating the numbers.  Remember, too, that this audit was done fully a year after the election — records may have been lost or destroyed.  That would be bad, but not as egregiously awful as the records never existing in the first place.

Zooming out, it’s worth noting that while Neufeld bent over backwards to emphasize the seriousness of these irregularities, also wrote that: “An ‘irregularity’ is a failure by an election officer to administer safeguards demonstrating that a voter is entitled to receive a ballot.  It is important to note that the failure of an election officer to administer these safeguards does not, in and of itself, mean that the voter was not in fact eligible to vote.” (Emphasis his.)

Got it? Moving on.

Mr. Ling correctly points out here that this failure does not automatically mean that the voter was not eligible to vote.  But by that logic, we would not require any identification rules whatsoever and instead put the onus on elections officials to prove a person is not eligible.  Someone could walk in, assert an identity and an address, and unless the officials could prove otherwise, the person would cast a ballot. 

We do not allow that, nor should we.

You’re right, Minister, we should not allow that.  In fact, that’s exactly what the Supreme Court of British Columbia said in Henry v. Canada, where they decided that this government’s voter identification laws should be allowed to stand.

What the court found was that the amendments to the Elections Act, which tightened requirements for same-day registration at the polls, are infringements on Canadians’ Charter-protected right to vote.  Luckily for the government, because the impairment is minimal — in that, theoretically, every Canadian can still vote — the constitutionality of the changes are saved under Section One of the Charter.

Now the appellants made a pretty convincing case: they figured that, yes, tightening ID restrictions could very well disenfranchise voters — if it weren’t for vouching.

They argued, writes the court, “that this procedure provides yet greater accessibility and a ‘failsafe’ measure because it enables those without acceptable identification to cast a ballot if vouched for by another elector.”

And who was that appellant?

The Government of Canada.

I would be curious to know what changed between that 2010 case, when Canada argued that vouching was one of the main reasons why the Elections Act was constitutionally sound, and now, when Canada alleges that vouching weakens the system.

There have been no legislative changes to the Act since 2010, and the only real contestation of the vouching process came from the Liberal Party, in trying to oust Conservative MP Ted Opitz.

So that’s curious.

ID requirements exist for a reason.  Failure to properly identify voters as so often happens with vouching allows people to vote when they are not eligible or to vote more than once.

There is no evidence from Elections Canada of this happening in the last decade.  I know the Minister likes to cite an episode of satirical Quebecois TV show Infoman as evidence that voters voted more than once, yet seeing as he hasn’t brought fraud charges against them, or launched an investigation, I am skeptical.

So.  I dunno what to tell you about that.

It should not be a guessing game.  Yet that is what vouching has become in nearly half of cases where it is used, if you believe facts in the Elections Canada compliance report.  Serious errors, of a type the courts consider irregularities that can contribute to an election being overturned, were found to occur in 12 percent of all Election Day cases involving voter registration, and 42 percent of cases involving identity vouching.” (P. 6).

It shouldn’t be a guessing game, which is perhaps why there is some confusion as to why the Minister didn’t adopt the recommendations of both Neufeld and Elections Canada as they pertain to fixing those irregularities by improving the electoral lists, implementing quality control at the voting locations, moving towards well-trained and professional (rather than political) staff, and eliminating the archaic and unnecessary notion of ‘polls.’

With due respect to the Minister, it’s a little disingenuous to cherrypick the parts of a report that makes it sound like your position is a good one, and ignore the recommendations that emphatically contradict your bill.

The latter point is particularly devastating.

Woah woah woah, let’s leave the snarky hyperbole up to the professional, here.

Judges dont overturn election results for nothing.  Yet the report says that vouchingirregularities which occurred 50,735 times (p.69) are serious enough for a judge to consider doing so.

Actually, we have no idea if irregularities occurred 50,735 times — that’s an estimate.  We only know of 269 irregularities in four ridings, because there’s never been a proper pan-Canadian study done it in — there’s only been one audit done, a year after the fact, because of a political challenge to an election result.

We also know, from the Opitz case, that not every irregularity identified constitutes an invalid vote.  So it’s flatly wrong to suggest that we have any idea of the number of votes that would be overturned, if challenged.

Further, judges do not go around flipping over elections like the Incredibly Litigious Hulk.  To contest an election, you need a very considerable amount of proof, and you will almost certainly need to survive each level of the judicial elimination process, all the way to the Supreme Court.  There, you will need to prove that each vote you are contesting is, in fact, so improperly cast that it must be nullified.  From there, the results will only be overturned if, and only if, the number of irregular votes matches or dwarfs the margin of victory.

In the two elections where vouching has become necessary, only one riding has been contested — and the challenge failed.

So let’s not overstate the danger, here.

But that is not all.

But wait! There’s more!

While vouching is theoretically intended to extend the right to vote, it may rob people of that very same right.  When courts invalidate votes because of serious irregularities, there is significant chance that legitimate ballots can be invalidated along with them.  In other words, a contaminated process can deprive even honest votes of being counted.

Right, which is why the courts have created such a rigorous and impossibly hard process to do so, that it would only happen in the most extreme cases, in the closest of ridings.

So why not simply administer vouching more competently, instead of ending it altogether?

Like what was recommended both by Elections Canada and its independent auditor.

The answer is that quality assurances seem to do little to improve the situation, according to Neufeld.

Who supports fixing the process, not eliminating it.

During two of these elections, quality assurance programs involving Onsite Conformity Advisors (OCAs) were applied.  However, vouching irregularities still averaged 21 percent during the OCA monitored elections.  This indicates that overly complex procedures cannot be remedied simply by improved quality assurance.” (P.15)

Right, which is why he recommended a more extensive and substantive reform to quash all irregularities, while still maintaining and improving accessibility.

It is important to let every eligible voter cast a ballot.  That is why the Fair Elections Act allows another day of advanced voting and requires Elections Canada to advertise voter ID requirements.


Every citizen is entitled to the voting franchise.  Let us never forget that a fraudulent vote has the same mathematical effect as denying someone their constitutional right to cast a ballot.  It cancels out the ballot of an honest person.

The use of “honest person” is interesting here because, from all the data we have, there is no reason to believe that any of these irregular votes were cast by anyone other than ‘honest people.’

We have no evidence of vouching being used for voter fraud.


None at all.

Not any.

Now, that is serious.  Our voter identity laws must be too.

Yes.  I agree.  Serious business.

Serious enough that I think this requires real consideration.



The cases for and against vouching

The Fair Election Act Goes to Committee

Democratic Reform Take Two

Building Poilievre’s Electoral Fraud in the Sky


Follow Justin Ling on twitter: @Justin_Ling


Share this article