there must be a serious question to be determined;
irreparable harm will be suffered should the stay not be granted; and
the balance of convenience favours a stay.
In the constitutional law realm, it seems general practice to grant a stay pending a Crown appeal after a law is declared unconstitutional – seldom is there no merit to such an appeal, irreparable harm seems likely if the law is not enforced pending appeal, and the balance of convenience tends to favour the status quo. But in the June 23, 2014 decision Frank v. Canada, Sharpe J.A. of the Ontario Court of Appeal declined to grant a stay, and made it very explicit that there is no “presumption” of a stay in the aftermath of a declaration of unconstitutionality.
On May 2, 2014, Penny J. of the Ontario Superior Court declared that a provision of the Elections Act, which limited the rights of non-resident citizens to vote in elections to those who had been absent from Canada for less than five years, violated the right to vote guaranteed in s. 3 of the Canadian Charter of Rights and Freedoms, and was not a reasonable limit of that right under s. 1 of the Charter. Penny J. refused to stay his judgment and, shortly thereafter, four federal by-elections were called for June 30, 2014, in which 13 non-resident Canadians registered to vote as of June 16, 2014. The Attorney General appealed the judgment and sought a stay pending appeal.
Sharpe J.A. specifically rejected the Attorney General’s submission that there was a “presumption” that a stay should be granted in these circumstances:
 It is the case that very often, the public interest in the orderly administration of the law will tilt the balance of convenience in favour of maintaining impugned legislation pending the final determination of its validity on appeal: See, for example RJR-MacDonald Inc. v. Canada (A.G.),  1 S.C.R. 311 at p. 346[:]
In the case of a public authority, the onus of demonstrating irreparable harm to the public interest is less than that of a private applicant. This is partly a function of the nature of the public authority and partly a function of the action sought to be enjoined. The test will nearly always be satisfied simply upon proof that the authority is charged with the duty of promoting or protecting the public interest and upon some indication that the impugned legislation, regulation, or activity was undertaken pursuant to that responsibility. Once these minimal requirements have been met, the court should in most cases assume that irreparable harm to the public interest would result from the restraint of that action.
 However, I cannot agree with the Attorney General that there is a presumption approaching an automatic right to a stay in every case where a court of first instance has ruled legislation to be unconstitutional. As Lamer J. also held in RJR-MacDonald, at p. 343, that “the government does not have a monopoly on the public interest.” See also Bedford, at para. 73: “The Attorney General does not have a monopoly on the public interest, and it is open to both parties to rely upon the considerations of public interest, including the concerns of identifiable groups.”
 In my view, it is necessary to carefully review the particular facts and circumstances of this case in order to determine whether or not a stay is warranted.
Sharpe J.A. expressed serious doubts that the Crown’s appeal would succeed, but nonetheless held that it was “arguable” and had sufficient merit to pass the low threshold of serious issue to be argued. Sharpe J.A. further agreed that irreparable harm would occur if a by-election were decided by a few votes, and non-resident voters granted the right to vote by Penny J.’s judgment could have made the difference. Though he also found this possibility to be quite remote, and noted that the irreparable harm to those denied the right to vote would be equally present if a stay were refused: “Once the election has passed, the constitutional right to vote in that election will be lost forever” (para. 22).
Ultimately, the balance of convenience persuaded him against granting the stay:
 In my view, the balance of convenience in this case favours refusal of a stay. I reach that conclusion for the following reasons.
 First, this is not the typical case where a complex statutory scheme or administrative apparatus has to be dismantled or constructed in order to give effect to the trial judgment. In such cases, the balance of convenience will typically favour a stay to avoid the cost and disruption that would flow from implementing a new regime based upon a trial judgment that may need to be undone in the event of a successful appeal.
 In the present case, Elections Canada immediately took the minimal administrative steps required to permit non-resident citizens to vote in accordance with the decision of the application judge. If a stay is granted, Elections Canada will have to undo what it has already done. It is clear from the record that it may not be possible for Elections Canada to determine in time for the by-elections which non-resident voters who registered after the judgment would have been eligible before the judgment. The terms of the stay requested by the Attorney General recognize that difficulty and ask for a qualified stay that applies “unless Elections Canada is unable to determine” if those who registered meet the pre-judgment requirements. In addition, at least one non-resident has cast her ballot. To grant a stay in this case would require Elections Canada to rescind the registrations of up to 13 non-resident electors and claw back the vote of a citizen who may well in the end have the right to cast her ballot. Granting a stay in this case would not avoid the cost and inconvenience of prematurely erecting or dismantling a scheme – it would do the opposite.
 Second, this is not a case like Bedford [which concerned prostitution] where the trial judgment creates a legislative void in an area of activity that needs to be regulated in the public interest. Allowing the judgment to operate does not create a void or gap in Canada’s election law. Nor does the judgment radically alter the class of those eligible to vote. The Act already grants many non-resident citizens the right to vote. The judgment under appeal merely extends the right to a broader class of non-resident citizens.
 As counsel for the applicants pointed out, it is highly unlikely that the judgment will produce a floodgate of votes from disinterested and disengaged non-resident Canadians. We know that the number of newly qualified non-resident voters who had registered as of June 16 is 13 or fewer. The non-resident must be both determined and informed. He or she must first register and then obtain a ballot. The non-resident voter cannot vote by simply marking an X beside one of the listed candidates but must complete a special ballot that requires the voter to know and write in the name of an actual candidate.
 I conclude that the balance of convenience does not favour granting a stay in this case.