Theodore R. Hart was born, raised and educated in Manitoba. He currently resides in Winnipeg.

The first-past-the-post (FPP) system of election in single-Member constituencies is the only means used for election to legislative assemblies in Canada at this time. Occasionally, this system gives far too many seats to the leading party and produces an assembly with an opposition too few in number to fulfill adequately its proper function. It also tend to allow one party to dominate a wide area sometimes completely. This article proposes a new alternative, the Regional seats system which would forestall one problem and combat the other.

Aside from producing victories that sometimes are far too lop-sided, first-past-the-post tends to bolster regional speciality, awarding one party in an area nearly all of the seats in election after election.  This is unhealthy, for the legislature, the polity it governs, and the area itself.  The ruling party will inevitably reflect the interests mainly of the regions where it has done best and other regions may feel comparatively neglected.

In recent years several places have moved away from pure FPP to “mixed” systems that are in effectively modified Proportional Representation systems.¹ They share certain shortcomings.  In particular, they rarely feature single-party government: coalitions are the norm. Their model is Germany where the ratio of FPP to PR seats is 50:50.  New Zealand mimicked this, while in Scotland it is 57:43 and in Wales 67:33.

An instructive example is the 1999 election to Scotland’s parliament.  Labour won 73% of the FPP seats, a result that would be considered a huge victory in an Ontario provincial election.  Yet once the additional PR seats were distributed, Labour wound up with only 43% of the seats in Parliament and had to take on a coalition partner in order to govern.

Regional Seats System

A better alternative would be what I call the Regional Seats system of election.  In essence, it is a modification of FPP that retains its strengths while correcting some weaknesses. As Table A shows, it would have given Labour a majority in the Scottish election had it been in use there.  Indeed, like FPP it typically would yield single-party governments. Coalitions would be rare.

In one respect the Regional Seats (RS) system is similar to what the Jenkins Commission in 1998 recommended for reform of the UK parliament.  Jenkins foresees about 20% of parliamentary seats being filled by regional Members chosen separately from the other Members who are elected in single-Member constituencies.  He would have a large number of regions, each with one regional Member (though with a handful of exceptional regions having two).

The RS system would rely exclusively on one Regional Member per region. Thus a typical region would include -four single-Member seats and have one Regional Member.² The purpose behind having so many small regions is to break down regional exclusivity.  Under an RS system every small region is necessarily represented by Members from at least two political parties. This would be typical also of the Jenkins plan.  So one party could not completely dominate any small region let alone a larger area of several regions.

Table A: Scottish Parliament 1999
	Seats: MPR Results			Hypothetical With Regionals			% of all Seats
	FPP	PR	Total	FPP	RS	Sum	FPP only	with RS	MPR
Lab	53	3	56	53	1	54	73%	60%	43%
SNP	7	28	35	7	15	22	10%	24%	27%
LDP	12	5	17	12	0	12	16%	13%	13%
Con	0	18	18	0	2	2	0	2%	14%
Other	1	2	3	1	0	1	1%	1%	2%
Sums	73 + 56 = 129			73 + 18 = 91			100% 100% 100%
MPR- “mixed” proportional Parties are: PR - proportional distribution Labour, Scottish National FPP - first past the post Liberal-Democratic, and RS - Regional seats Conservative MPR is actual results; RS is based on actual results

Having 20% of a legislature comprised of Regional Members assures an opposition of functional size regardless of how overwhelming one party’s victory in the single-Member ridings might be.  This problem, peculiar to our provincial assemblies, is one that an RS system would forestall.  As Table B indicates many provincial legislatures have suffered from this phenomenon over the last four decades. The current case is in PEI which has but a solitary Opposition voice.

Table B: Huge Wins In Provincial Elections 1960-2000
Elections:	Gov’t % Seats	Size of Assembly	Members Non-Gov’t
New Brunswick 1987	100%	58	0
Prince Edward Island 2000	96%	27	1
Prince Edward Island 1993	97%	32	1
Prince Edward Island1989	94%	32	2
Alberta 1982	95%	79	4
Alberta 1979	94%	79	5
Alberta 1975	92%	75	6
Quebec 1973	93%	110	8
Newfoundland 1966	93%	42	3
Alberta 1963	95%	63	3

Let us see how RS would work. Suppose some province with an assembly of 60 seats has 12 Regional Members and 48 other Members.  Perhaps it is Saskatchewan or Manitoba or Nova Scotia.  The constituency map for the province would show 48 single-Member ridings grouped into twelve clusters of four seats each.

Under an RS system no candidates and no party run directly for the regional seat.  What voters see, and what the news media report, are the 48 FPP contests.  Each regional seat is filled later with one of the losing candidates from the FPP races within that region.  Consequently, the Regional Member to be will also have been knocking on doors to solicit votes and will have the same direct connection to voters as do the FPP winners.

The ballot the voter sees is the customary one.  The voter marks it once for one candidate for that constituency.  And the candidate with the most votes wins the seat.  From a voter’s perspective an RS system appears identical to a pure FPP system.

On election night, however, in addition to seeing who wins in each of the 48 FPP races, the voter also sees which party wins each regional seat.  The procedure used is simple.  Tally the votes of the losing candidates for each party within the region, and the party with the largest tally wins the regional seat.  Votes for winning candidates are ignored.  (After all, those voters got who they voted for.)  Votes for independent candidates are also ignored because only an official party can win a regional seat.  Table C is an example.

Table C: Example of a Six-Member Region
Hypothetical Region	---- % of votes for: ----
Nova Scotia 1999	Lib	ND	PC	winner
Chester-Ste Margarets	21.2	33.4	w	45.3
Lunenburg	26.3	22.2	w	51.5
Lunenburg West	w	19.8	34.8	45.5
Queens	20.8	24.0	w	55.2
Shelburne	41.5	16.0	w	41.5
Tallies:	109.8	116.3	34.8
Regional Seat = New Democratic Party

So on election night, besides knowing who won in each of the 48 single-Member constituencies, it is also known which party won each of the 12 regional seats.  Therefore, it is known which party has the largest number of seats overall and will form the government.

What is not known on election night are the names of the Regional Members.  The leader of the party winning a regional seat has the right but is not obliged to nominate someone to be the Regional Member.  Eligible nominees are limited to unelected candidates who ran for that party in that region at the latest general election or subsequent by-election.  None else qualifies: not an independent candidate, nor someone who ran for another party, nor someone from outside the region, nor any successful candidate. But with one exception: that party leader, if otherwise not a member, could nominate himself or herself.³

Typically, the party leader would have between two and four defeated candidates to choose amongst for the nomination.  This fact has several consequences.  Firstly, it means that each Regional Member will owe a personal debt to his or her leader and is likely to be more loyal than other Members might be — especially in a party that has done more poorly than expected, for those elected might blame the leader for the poor showing.

Secondly, it allows for greater balance in a party’s legislative caucus.  An extreme example is the outcome of the 1999 Saskatchewan election where New Democrats mostly won urban seats and the Premier had to appoint an urban representative as Minister of Agriculture. Under an RS system he would have had several rural Regional Members for that important post including the Minister who was just defeated. Meanwhile, the Opposition includes a lone urban member but would have had several more as Regional Members.

Ontario’s 1999 election outcome would also have benefited from using an RS system.  As Table D indicates a better balance would have been obtained there for both Government and Opposition.  Each would have had at least one Member from every region in the province, which is not now the case.

Thirdly, it means that some prominent politicians who were defeated and would otherwise be gone from the legislature might return as Regional Members.

Similarly it sometimes happens that a leader will suffer personal defeat even as the party wins power.  Had an RS system been in place in Alberta in 1989, then Don Getty could have become the Regional Member for Edmonton South.  Instead, he had to induce a loyal backbencher to resign so that he could run in and win the subsequent by-election in that safe seat. Two examples are David Peterson in London in 1990 and Roy Romanow in Saskatoon in 1982.

The same fate befell Clyde Wells in Newfoundland that year, and it nearly struck Gary Filmon in Manitoba the following year.  Federally, Mackenzie King twice had to seek a safe seat because of suffering personal defeats. So did Tommy Douglas.

The same procedure would well serve a new party leader who has no seat in the assembly but seeks one before the next general election.  Stockwell Day, Joe Clark and Jean Chrétien all entered the House of Commons in this manner.  How much simpler it would be to merely induce a Regional Member of the party to resign so that the leader could take his place!  No needless by-election and no waiting for the Prime Minister or Premier to call it.

The House of Commons would have looked considerably different had a Regional Seats system been in place for either of the last two general elections. Table E shows that the Liberals would have gained seats where they are habitually weak; the PCs would have been stronger in 1997; and Ontario would have had MPs from all major parties that ran there, both times.

Table D: Ontario is 1999 with Regional Seats added                                                                    Number of Seats                                                               % of Seats
	PC	Lib	ND	Total	PC	Lib	ND
Group A- FPP seats Regionals Total for Group A	47 0 47	5 13 18	0 0 0	52 13 65	90% 72%	10% 28%	0% 0%
Group B - FPP Seats Regionals Total for Group B	12 11 23	30 2 32	9 0 9	51 13 64	24% 36%	59% 50%	18% 14%
Both A and B Regionals Total for Ontario	59 11 70	35 15 50	9 0 9	103 26 129	57% 54%	34% 39%	9% 7%
Note: The 13 regions in Group B consist of the most northern three, most eastern three, plus Winsdsor-Kent, Niagara, Hamilton and four in Metro Toronto. The 13 regions in Group A are contiguous and include part of Metro Toronto, some east thereof and nearly all of southwestern Ontario.

Another benefit of having Regional Members is that it heightens political awareness.  After all, in one-quarter of the constituencies two of the candidates who ran there in the last election will be sitting in the assembly, one as the riding’s Member and one as its Regional Member.  Each from time to time will be sending mailings to their constituents, describing from their differing political perspectives what is going on in the legislature.  Come the next general election, they will both likely be running.  Which would make for some interesting contests.

In fact, the existence of regional seats would heighten political campaigning overall.  Consider a region dominated by one party who will take every FPP seat.  What about the other parties?  For them, the difference between garnering 15 - 20% of the vote across the region or averaging 20 - 25% may mean getting the regional seat or seeing a rival party snare it.  So they have an incentive to try hard in every seat in the region even though none of their candidates seems likely to be elected directly.

In more competitive regions, as Table C shows, if two parties have wins in a region it is easier for a third to bag the regional seat.  Yet it is just as likely that a party who wins a seat in the region will also take the regional seat.  So all major parties have an incentive to try harder in all seats within such a region.

Table E: Distribution of 60 Regional Seats for Federal Elections
1997                                                                      2000
	Lb	Rf	BQ	PC	ND	Lb	AL	BQ	PC	ND
21 Ontario	-	9	-	8	4	-	18	-	1	2
18 West	11	2	-	3	2	12	3	-	1	2
15 Quebec	7	-	4	4	-	7	-	8	-	-
6 Atlantic	3	1	-	1	-	2	1	-	3	-
60 Canada	21	12	4	16	7	21	22	8	5	4

One aspect of campaigning hard is the effort and resources put into individual campaigns.  Another is the presentability of all candidates within the region, whether or not the party’s central campaign managers expect them to win.  With an eye on taking the regional seat, the parties will take greater care in candidate selection.  And with the carrot of likely being nominated the Regional Member should they fail to win directly, quality candidates in some cases will be easier to recruit.

Finally, sometimes a third or fourth party would get more seats under RS than it would under pure FPP, though this occurs only occasionally. But unlike a “mixed” PR system, Regional Seats does not encourage the formation of minor parties or narrow focus by too readily awarding them seats. RS is much kinder to the large umbrella parties which characterize our political scene.

Conclusion: A Cure for Rampant Regionalism

For the third federal election in a row, party standings fail to reflect the diversity of voters’ intentions in much of the nation. This skewed outcome is a direct result of using the first-past-the-post (FPP) system of election in single-Member constituencies. The Regional Seats system would moderate such skewed results and cure rampant regionalism. It is a better alternative to FPP than any “mixed” PR system and is equally valuable in provinces with a similar problem, such as Manitoba, Saskatchewan, Quebec and Ontario. Moreover, it would always yield an Opposition of functional size, something FPP alone does not.

Notes

1. See: Andrew Reynolds: “Electoral Systems Reform in the UK” in Henry Milner ed. Making Every Vote Count: Reassessing Canada’s Electoral System; Peterborough ON, Broadview Press, 1999.

2. Alternatively, five plus one with 17% of the assembly being Regional Members instead of 20%.

3. The Crown’s representative has the obligation to accept the nomination, provided the nominee is an eligible one and the nominator is unambiguously the current leader of the party with the right to make the nomination. Otherwise, it should be refused.