[1]                           By this Reference, the Government of Alberta invites the Court of Appeal of Alberta to say that the present electoral boundaries of Alberta are constitutionally valid. Despite some hesitation, the Court has again decided to refuse to condemn Alberta's electoral boundaries.

[2]                           The root of the problem before us is the long history of population shifts from other parts of Alberta to Metropolitan areas, particularly Edmonton and Calgary. As a result, the average electoral division in those cities contains 13 per cent more voters than the average of other divisions. The Canadian Charter of Rights and Freedoms guarantees those urban electors the right not to have the political force of their votes unduly diluted. We simply are unable to say, for lack of an explanation for the present boundaries, whether the dilution that exists today is "undue". There can be many valid reasons for disparity, but they do not include a fear of the future by electors whose electoral divisions might be subject to surgery to assure other electors their constitutional rights. Constitutional rights must be respected even if to do so is momentarily unpopular. We suggest a renewed effort for the next election.

[3]                           Pursuant to section 27(1) of the Judicature Act, R.S.A. 1980, Chapter J-1, as amended, the Alberta Lieutenant Governor in Council by Order in Council 215/93 referred the following questions to the Court of Appeal for hearing and consideration:

[4]                           The Order in Council asked the Court to assume that certain amendments would be made to the Electoral Divisions Statutes Amendment Act, 1993, S.A. 1993, c. 2. These amendments were required to correct technical and clerical errors in the legal descriptions of the electoral divisions and have now been enacted by the Electoral Divisions Amendment Act 1993, S.A. 1993, c.4.

[5]                           Prior to 1989, the Alberta electoral map demonstrated a major under-representation of urban voters. A Special Select Committee of the Legislature reported, in 1989, in a report embodied in the Electoral Boundaries Commission Act S.A. 1990, c.E-4.01, that:

[6]                           This imbalance had a long history. Over the years, some municipalities experienced huge population growth, and some did not. Some predominantly agrarian districts suffered significant drops in population. Revisions of boundaries did not keep pace with these demographic changes. Even after the establishment, more than three decades ago, of a tradition of independent electoral boundary commissions, the Legislature accepted, and forced upon the boundary commissions, one formula or another that required the maintenance of this differential.

[7]                           The following table indicates the extent of the under-representation of the two metropolitan areas for the years 1986, 1991 and 1992. For the purpose of this analysis, we put aside "special consideration" districts.

[8]                           A more equal distribution for 1992 would have reversed the distribution, and offered 41 seats to Edmonton and Calgary and 38 to the remainder of the province.

[9]                           Alberta sought in argument to soften the impact of this disparity a little by lumping some smaller urban centres with Edmonton and Calgary, areas that are adequately represented. The Liberal Party, in turn, sought to worsen the comparison by reliance on the enumerated population, and voting population, as opposed to the general census. The table just displayed is the fairest presentation of the problem.

[10]                       The 1989 statute, enacted after the 1986 census, created a Commission to recommend precise boundaries. It required further legislation to establish those boundaries. That statute, moreover, required the Commission to follow some rules: it provided for 83 seats in total, and assigned 43 seats to named municipalities. It also permitted five "hybrid" electoral divisions, which might contain parts of these named municipalities.

[11]                       This last rule makes little sense unless one recognizes that the 43 named municipalities contained all the major urban areas in Alberta, and 60 per cent of the population of the province lived in them. But 43 is only 51.8 per cent of 83. To achieve equal representation between urban and rural population, the named municipalities required roughly 50 electoral divisions, or seven more than the assigned number. Hence the importance of the five "hybrid" constituencies. By permitting the Commission to place urban population in them, the disproportions might have been significantly lessened.

[12]                       Alberta referred the 1989 statute to this Court for constitutional review in 1991. See Reference re Electoral Boundaries Commission Act 1991 ABCA 317 (CanLII), [1992] 1 W.W.R. 481 (Alta. C.A.). Comment by the Court was limited to "the general idea of the use of urban and non-urban categories as a criterion for the selection of boundaries, and the 40/43 split." The Court said:

[13]                       The Court, however, refused to make a declaration that the legislation offended the Charter. In so doing, it observed that the proposed changes, which included the hybrid constituencies, demonstrated "gradual but steady" change toward a valid approach.

[14]                       The Government of Alberta then appointed an Electoral Boundaries Commission to make recommendations with respect to the specific electoral divisions. It held public hearings and conducted its work over a period of approximately a year and a half.

[15]                       The Commission produced an Interim Report in November, 1991. That report, no doubt sensitive to the points just made, proposed major changes to existing boundaries to reduce variations. This led it to heavy reliance on the hybrid rule.

[16]                       After public hearings about the interim report, the Commission was unable to arrive at a majority decision, and the Commission members instead released five separate "final" reports on April 24, 1992. It is clear from the Reports of the Commissioners that their difficulties arose because of the mandated imbalance, and because nobody in Alberta seemed to like the idea of a riding that mixed urban and rural voters. The Report said:

[17]                       The chairman endorsed the interim report. Two members proposed a new scheme, saying that the 10 per cent plus or minus variation must be abandoned for a far greater variation. Member Biggs said:

[18]                       In sum, the majority were of the view that the proposed scheme for hybrid ridings failed for lack of popular and governmental support. Moreover, a majority was of the view that, without recourse to that scheme, the continuing under-representation of urban voters would be unjustifiably high.

[19]                       On the day of that "final" report, the Legislative Assembly was within 20 months of the date within which a general election was required to be held pursuant to section 4 of the Charter. Yet no clear proposal for electoral boundaries was available. The Legislature on July 2, 1992 appointed yet another Select Special Committee on Electoral Boundaries. The Committee was required to report to the Legislative Assembly by November 15, 1992 for the stated purpose of enabling the Report to be debated during the 1992-93 winter sittings of the Legislature. The resolution appointing the Committee contemplated a membership of seven members of the Assembly. The opposition parties, however, did not appoint anyone. The Committee in the result comprised four members of the Assembly from the government side.

[20]                       The Motion establishing the Special Select Committee provided the following criteria to be considered by the committee:

[21]                       The Committee held no public hearings, but did hear from nine invited "consultants" [Report p.67-68] and 18 other persons or groups. [Report p.72]. It also reviewed the more than 1,000 comments offered to the Commission. And it recommended that an independent boundaries commission, with public hearings, be reestablished in the future.

[22]                       The Committee was appointed on July 2 and had its first meeting on July 29. The Committee submitted its report to the Speaker of the Legislative Assembly in November 1992.

[23]                       It proposed an entirely new scheme, with no increase in electoral divisions and reliance upon the latest Statistics Canada census data. That data indicated all the old trends toward centralization continued.

[24]                       The Committee scheme scrapped hybrid electoral divisions. It assigned one additional seat to Edmonton and two to Calgary, and assigned all the electoral divisions to precisely-defined boundaries. In the result, there were 44 divisions in the areas for which the 1989 Act offered 43. Those numbers, however, were not "diluted" by voters placed into hybrid divisions. In the result, the urban ridings were, on average, 13.4 per cent above average while all other ridings were, on average, 8.4 per cent under average. The metropolitan voters thus were less well represented than had been possible under the 1989 Act, and less well represented than suggested in the 1991 interim report of the Commission.2

[25]                       The Committee report offered no detailed explanation for the specific boundaries chosen save these comments:

[26]                       The Committee, in their very brief Report, stated [at p.72] that they also endeavoured to follow existing electoral and municipal boundaries and community lines within Edmonton and Calgary whenever possible.

[27]                       They suggested that re-distribution occur again only after the next national census:

[28]                       Finally, the Committee recommended:

[29]                       On January 25, 1993, Bill 55, which incorporated the recommendations of the 1992 Committee, was introduced in the House and debated January 25 to February 12, 1993. It was enacted as the Electoral Divisions Statutes Amendment Act, 1993, subject to proclamation. Certain technical and clerical errors in the Act were corrected by amendment in April, 1993, and the Act, as amended, was proclaimed in force on May 18, 1993, thus creating the electoral map which is the subject of this Reference.

[30]                       These electoral divisions could remain in place for many years, as the Act mandates their review following the next census, which does not take place until 2001.

[31]                       In March 1993, when the Electoral Divisions Statutes Amendment Act had received third reading and royal assent, the Town of Lac La Biche took proceedings to prevent the Lieutenant Governor in Council from proclaiming the legislation in force, primarily on the grounds that the boundaries of the Lac La Biche-St. Paul electoral division offended the Charter right to vote of the citizens of the Town of Lac La Biche. The Court of Appeal refused to grant an interim stay pending determination of the applicants' Originating Notice requesting a declaration that the Electoral Divisions Statute Amendment Act was of no force or effect. See Lac La Biche (Town) v. Alberta (1993), 1993 ABCA 105 (CanLII), 102 D.L.R. (4th) 499. Subsequently the Town and other applicants elected to discontinue the Queen's Bench action without having the merits heard, but intervened in this Reference.

[32]                       Also in 1993, there was a general election based upon the new electoral map. The Premier, however, both before and after election promised to proceed with this Reference.

[33]                       In this Reference, the Town of Lac La Biche objected to the creation of a constituency that disconnected the Town of Lac La Biche from the forest region to the north. The Town and immediate environs are now joined to the electoral division south of it, called St. Paul. The Town complained that its separation from the forest region fails to recognize its economic, cultural and geographic ties with the forest industry and the oil and gas industry of the north, and that it has little in common with the agricultural economy and agrarian community centred on St. Paul. In sum, the complaint is that the new boundaries disrupt an existing community.

[34]                       The Alberta Association of Municipal Districts and Counties supported the new boundaries. The Association submitted that the over-representation of rural areas is justified by the fact that it is more difficult to represent rural as opposed to urban ridings.

[35]                       In contrast, the Liberal Party of Alberta, the New Democratic Party of Alberta and the Alberta Civil Liberties Association objected to the under-representation of the urban voter on the basis that this violates section 3 of the Charter. The Alberta Civil Liberties Association also argued that under-representation of voters in the inner areas of Calgary and Edmonton constitutes systemic discrimination against members of certain disadvantaged minority groups, namely the disabled, women, single parents, the elderly, immigrants, the poor, and the unemployed who inhabit those areas in disproportionate numbers, and is contrary to section 3 and section 15(1) of the Charter.

[36]                       The New Democratic Party of Alberta submitted that the new electoral boundaries violate section 3 of the Charter on a further four grounds, namely:

[37]                       In the 1991 Reference we offered this summary of the constitutional rights held by all Albertans (Reference Re Electoral Boundaries Commission Act, Alberta, 1991 ABCA 317 (CanLII), [1992] 1 W.W.R. 481 (Alta. CA).):

[38]                       Before this panel, we heard no suggestions for a re-statement of these rules. The argument, rather, turned over what amounts to "undue" dilution.

[39]                       Alberta took the position that every boundary-writer must balance many factors, and thus this Court should not intervene unless this artful balancing was clearly wrong. Its argument rested essentially on this statement in the affidavit of Mr. Bogle:

[40]                       Alberta invited us to affirm the legislation as, on the facts before us, reasonable compliance with the Charter:

[41]                       In Carter,3 Madam Justice McLachlin stated at page 189:

[42]                       In the 1991 Alberta Reference, this Court concluded at page 488:

[43]                       As has been observed, the New Democratic Party and the Town of Lac La Biche contended that, in some of its choices, the Committee was clearly wrong. Some of this argument turned on the criteria for any "special" ridings with disproportionately small populations. Critical reference was made to the two smallest electoral divisions, which happened to be the seats of the chairman and the vice-chairman of the Committee.

[44]                       The Liberal Party also raised the impossibility of approval of an electoral distribution where no specific reasons are supplied. Its point was that, with these reasons, it is not possible to know that irrelevant considerations were not given play, let alone whether the balancing of relevant factors was clearly wrong. Thus, the absence of the detailed reasons left Albertans, and the Court, with no assurance that the constitutional rights of all Albertans had been respected.

[45]                       It is one thing to say that the effective representation of a specific community requires an electoral division of a below-average population. That approach invites specific reasons, and specific facts. The constitution of Canada is sufficiently flexible to permit disparity to serve geographical and demographic reality.

[46]                       It is quite another to say that any electoral division, for no specific reason, may be smaller than average. In the 1991 Reference, we affirmed the first, not the second. We affirm again that there is no permissible variation if there is no justification. And the onus to establish justification lies with those who suggest the variation.

[47]                       We heard considerable argument over whether there was, in Alberta, an adequate factual basis for a differential. Reference was made to "distance from the Dome", and the extra travel demands upon some members. Others countered by pointing to the excellent air and road network in Alberta, and the fact that the Calgary divisions are as distant from Edmonton as many non-urban divisions. Reference was made also to the strong rural tradition that most voters know their members, and as a result make great demands upon them. One urban member responded by saying that urban voters who know their members are equally demanding; the problem is that it is difficult to get to know one's members in a large urban riding. Also, urban members will often meet with, and speak for, people who reside outside the division but who have a community of interest with electors. These are all difficult factual issues.

[48]                       Similarly, it is one thing to say that a variation of, say, 10 per cent up or down may be common. It is quite another thing to assume that, for any electoral division, this variation is, without more, "permissible".

[49]                       During argument on this Reference, we heard the view that this Court had, in the 1991 Reference, blessed a general discrepancy between rural and urban divisions. We did not. Nor did we accept the suggestion that areas in the province with a below-average density of population might expect, without more, to have electoral divisions with below-average populations.

[50]                       We did accept that the principles of fair representation might well lead to deviation. But we insisted that these must be established on a division-by-division basis. And we complained that the pre-set divisions in the 1989 Act had not been justified for that very reason.

[51]                       While a delicate balancing of many sometimes conflicting factors is an inevitable part of the task of boundary-writing, and a reviewing Court should not intervene unless the effort is clearly wrong, it is impossible for us to say that the effort here meets a Charter challenge when we do not know with any precision the reasons for the boundaries under review. Therefore, it is not possible for us to answer the first, and key, question.

[52]                       In its argument Alberta said:

[53]                       We accept that approach. But where is the explanation that what was done was needed? The Legislature offered no reasons, but essentially adopted the recommendation of the Select Committee. Before us, Alberta equated the Committee's reasons with those of the Legislature. We do not know with any certainty or detail what those reasons are.

[54]                       The Committee Report said [p.67] that transcripts of its minutes would be available. But none are before us. Instead, Alberta supplied, a long affidavit from the Chairman, Mr. Bob Bogle, attempting to explain, in his words, some of the reasoning of the Committee. We do not know if the other members concur. He was cross-examined on the affidavit. This method of attempting to establish facts in a Reference is quite unsatisfactory, for obvious reasons we elaborated upon during oral argument.

[55]                       In cross-examination on his affidavit, the Chairman of the Committee said:

[56]                       Unfortunately, as we have observed, those reasons were not supplied. Mr. Bogle's affidavit tells us:

[57]                       In fairness, we do observe that he went on to offer his detailed reasons for the decisions about some impugned electoral divisions, being those along the southern boundary of Alberta and those in the Athabasca region.

[58]                       Nevertheless, we simply do not know, with precision, why the many variations were permitted. In this case, what was needed was a riding-by-riding justification. It has not been offered. Unlike the Supreme Court of Canada in Attorney General of Saskatchewan v. Carter [supra], we are not persuaded otherwise by the material before us. We therefore cannot answer the question. Because of this, it equally would not be fair to deal with the specific attacks on specific electoral divisions made by the Town of Lac La Biche and the New Democratic Party.

[59]                       While we hesitate to make a firm statement in the absence of detailed reasons, we fear there may well have been regard for an irrelevant consideration. This serious concern proceeds from this basic idea: we do not think it a correct approach to the Charter to exclude unpopular alternatives simply because they are unpopular.

[60]                       It is one thing to say that, on the facts and for a given community, a deviation becomes a practical necessity; it is quite another to say that existing deviations must remain because significant numbers of voters otherwise will be unhappy. The only fit response to that, in general, should be to remind voters, with Burke, that "the people never give up their liberties but under some delusion".

[61]                       While some deviations in Alberta are no doubt inevitable, we see evidence that the practical necessities raised by the principle of effective representation did not, alone, guide the hand of the legislators. On the contrary, what seems to have motivated this scheme at least in part was the acknowledgment that, whether or not some disparities were warranted, change would be made slowly so as not to offend unduly the political sensibilities of some electors.

[62]                       The boundaries before us, at least in part, seem to be a response to widespread protest from those Albertans who live in farming communities. The record before the Court in this case is no different in that regard from the record before us in 1991. We then offered this summary of the materials:

[63]                       Mr. Bogle, in cross-examination on his affidavit, said the elected officials had detected public support for the status quo because,

[64]                       The Chairman added that "… the first priority would be to respect existing constituency boundaries, if possible…". This is, of course, a simple way to assuage the concern of some voters.

[65]                       The new electoral map clearly shows the result of that approach. For example, it was common ground before us that the population figures indicated the need, in the absence of any special considerations, to reduce the number of divisions in southern Alberta by two. Mr. Bogle acknowledged this in his affidavit (paragraph 37a) but explained that the committee chose instead to reduce the number of divisions by one, despite the fact that a further reduction would eliminate one of the smallest divisions in the province, which, by happenstance, was that for which he was then the sitting member. One reason he gives in his affidavit for this decision was that a further reduction "would have meant a sudden and substantial reduction in the level of representation." That is, we observe, exactly the concern of some electors. The concern, we feel constrained to add, of other electors, those in Metropolitan Alberta, was that their existing inadequate level of representation would remain reduced.

[66]                       With respect, this very natural concern of an elected official for the "comfort zone" of a vocal portion of the electorate is not a valid Charter consideration. The essence of a constitutionally-entrenched right is that it permits an individual to stand against even a majority of the people. Put another way, Canadians entrenched certain traditional rights for minorities in the Constitution because they do not trust themselves, in all times and circumstances, to respect those rights. The fact, then, that a significant number of Albertans do not like the results of an equal distribution of electoral divisions is no reason to flinch from insisting that they take the burden as well as the benefit of democracy as we know it.

[67]                       It may be thought that we took a different view in the 1991 Reference about the disaffection of non-urban electors. The Court there refused to make a declaration of invalidity because:

[68]                       We accepted that reasoning, but only as a reason for judicial restraint in the face of error, not as a valid consideration for a boundary-writer. We saw "gradual but steady" movement towards a Charter-supported approach. (Whether we thus invoked section 1 or the power to withhold a remedy is an issue we stepped past for that case, and will again for this.) We saw, in the 1989 scheme, a considerable closing of the gap. In coming to that conclusion, we relied on the power of an electoral commission to use the hybrid division. On that basis, we saw evidence of enough progress to warrant a refusal to act. But this was consideration of appropriate judicial restraint, not approval of regard for irrelevant considerations.

[69]                       In that regard, the legislation now before us does not offer even the same assurance as did the 1989 law. Much was made of the fact that the disparity is now 44-39, but this is without any further amelioration by the use of the hybrid riding. This is not progress.

[70]                       As we have said, the origin of the problem before the Legislature is the historic imbalance in the level of representation between agrarian and non-agrarian populations in Alberta. Each year this problem worsens, because each year urban populations increase and non-urban populations decrease. We call this a problem because it impacts significantly on the right to vote of urban Albertans. This cannot be permitted to continue if Alberta wishes to call itself a democracy. The courts, and the people, have rejected the notion of mechanical one-person, one-vote equality. That does not mean we can or should accept significant disparities without reasoned justification just because some members of the population resist change.

[71]                       There are only three possible solutions to the historic disparity: first, a mixing of urban and non-urban populations in electoral divisions of equal size, second, more seats over-all, or third, fewer non-urban seats.

[72]                       The immediate cause of the problem faced by the Legislature in 1993 was the conviction, by the Commission and by the Committee, and by most of the parties before us, that the people of Alberta simply would not accept the idea that agrarian and non-agrarian populations would both feel adequately represented in the same constituency. We accept this sentiment as a dramatic example of the importance of the idea of effective community representation. That eliminates the first of the three possible solutions, that being the one that Alberta advanced to this Court in 1991 as the saviour of the traditional electoral scheme.

[73]                       The second possible solution, more seats, would fly in the face of the refusal of the people of Alberta to support an increase. We suggested this in 1991. Before us on this appeal, the two parties with the greatest electoral support in the recent election opposed it. They say that the people oppose it, including the people in Edmonton and Calgary. We accept that as fact, but observe that yet another possible solution is gone.

[74]                       The third, and last, is a reduction in the number of non-urban electoral divisions. But that raises the natural and understandable reluctance of voters in the less populous ridings to accept the "massive surgery" that would be needed to create equity in the absence of an increase in seats. But, if one spurns this solution, none remains.

[75]                       The people of Alberta must understand that this last is the only solution unless they soften their attitude towards the other two. We re-affirm that popular opposition to "massive surgery" is not a reason to ignore the breach of the right to effective representation by widespread and significant imbalance in voting power.

[76]                       We again invoke the need for judicial restraint about interference in the electoral process. We do not think the existing inadequacy is large or glaring enough to invalidate the existing legislation. To do so would be a major disruption in the electoral process. In 1993, Alberta had a general election based on these boundaries. We do not see the democratic value in creating a political crisis.

[77]                       Moreover, while urgency was not offered as an excuse for error here, the fact is that the Legislature had a serious problem as a result of the failure of the Commission to achieve a majority proposal. It seems to have been common ground that the 1986 boundaries failed to meet minimal Charter standards, and that some new boundaries were essential. And it seems equally clear to us that these new boundaries needed to be in place well before the constitutional mandate of the Government expired in 1993. While we hesitate to use the term crisis, there was a serious problem for the Legislature to address.

[78]                       There was certainly no time left for the wholesale review made necessary by the abandonment of the hybrid division. That review must identify communities, in every sense of the word. It must look in depth at social history as well as demography and geography. Moreover, that review is unlikely to be effective unless the reviewer gives ordinary Albertans ample opportunity to come forward and describe the communities of interest they see in their lives. It is time-consuming and not inexpensive, but essential to a healthy democratic life.

[79]                       We accept also that part of the problem may have been a lack of clarity in the 1991 decision, which seems to have led some to think that we accepted, as a valid Charter factor, a sensitivity to the complaints of certain voters about the practical effect of enforcement of Charter rights.

[80]                       In the result, we again have decided to withhold any Charter condemnation. We do, however, wish to say more precisely what we meant by "gradual and steady" change. We think that a new and proper review is essential before the constitutional mandate of the present government expires, and, we hope, before the next general election. We reject any suggestion that the present divisions may rest until after the 2001 census.

[81]                       Because of the above result, it was not necessary for us to consider the standard of review in a case where electoral boundaries are drawn by a partisan group. We accept that there is no authority to say that the Charter mandates an independent commission. And we accept the "clear error" test for review of the decision of such a tribunal. It does not follow, however, that the same level of deference is appropriate when the author of the boundary is some person, or group of persons, who is not insulated from partisan influence, and who may be tempted to engage in some traditional political games, like gerrymandering or log-rolling. Compromise, and partisan advantage, may be very much a part of the political process, and we in no way disparage it. But we might not do justice to the Charter if we failed to offer a fresh assessment where before there was no independent review.

[82]                       We therefore answer the questions in the Reference in this way:

[83]                       We are unable to say because we lack enough information. We do think, however, that there may have been regard for irrelevant considerations.

[84]                       Unpopularity of Charter rights is not a valid factor when assessing Charter rights.

[85]                       While unpopularity of Charter rights is not a valid s.1 consideration, the idea of achievement of full electoral rights over time, but with all deliberate speed, warrants judicial restraint.