Democratic representation and measurement thereof
Six factors that impact proportionality of representation
Proportionality of representation is usually measured by comparing party totals with vote counts. A proportional result is if a party with say 40 percent of the votes takes 40 percent of the seats, for example.
But there are several factors that may render such analysis flawed.
Sometimes the votes cast are not true polling of the population; other times some of the elected members are elected under special circumstances other than than getting a majority of the votes or getting the most votes, and so on.
1. Do all adults have the vote?
When women did not have the vote, the ability of the political system to represent the sentiment of all was severely limited.
And even after that when some races or groups were still denied the vote, the same held true to a lesser degree.
"Treaty Indians" did not get the vote in Alberta until 1971, for example.
And the minimum voting age also changed the complexion of the voters polled to elect representation
2. Do some voters have more votes than others?
When some voters cast multiple votes (as in Block Voting) while others cast only single votes as in FPTP, the overall party tallies may be unbalanced as regards measuring popularity of parties. This is especially true if a party runs none or few candidates where voters cast multiple votes.
Such was the case In the 1921 Alberta election. Each voter in Edmonton and Calgary could cast up to five votes and the UFA ran just one candidate in Edmonton and none in Calgary. Most of the votes cast in the cites thus did not go to the UFA. The UFA candidates mostly ran in rural districts where each voter could cast just one vote. So the UFA party received no more than one vote from any voter in the province, while the Liberal party for example received as much as five votes in Edmonton from each of some of the voters. We can't know how many gave all five of their votes to Liberal candidates, though but we do see that the total of votes received by Liberal candidates in Edmonton was more than the number of Edmonton voters who voted!
3. Are districts about the same size?
Are they balanced or do their boundaries not benefit one party over others?
Or by being un-balanced, such as being a concentration of voters normally scattered across the land, does a minority achieve its due share of representation that it might not achieve where there is no un-balance?
Or by being un-balanced, a very small voting block may elect one member, which even by just the one success alone is far more than its due.
If all districts are reflections in miniature of the overall society, the largest single group (even if not majority) will be over-represented. If only plurality is needed to take a seat, then one party, if is supported by the largest single group in every district, even if it is not a majority in any district, will take all the seats.
Such was the case in the 1987 New Brunswick election and in 1935 PEI election.
If a district through accident or purpose is dominated by a group not common overall, it will get more representation than is its due.
An Independent candidate elected in a district and not as part of a party's slate may be an example of this. The Independent candidate cannot be known to be supported by any more voters than voted for the person as an individual in the one district and therefore the seat would be more than is properly due.
For example, Clarence Copithorne, an Independent candidate, was elected in Banff-Cochrane in the 1967 Alberta election. He received 2400 votes, not even a majority in the district. The seat he won was an incredible return for his mere 2400 votes when the Liberal party won just three seats with 54,000 votes and the NDP with 80,000 votes won no seats at all.
If a district through accident or purpose is dominated by a group common overall but too thinly spread t get a seat normally, it will get the representation that it is due. This is what is sometimes called affirmative gerrymandering.
In 1909, Charlie O'Brien of the Socialist Party took the Rocky Mountain seat. But this actually made sense - workers made up a majority of the population of the province; Socialist/Labour voters cast 3 percent of the vote, which made the group due one seat. Interestingly, the charge was made that the district was drawn in such a way as to allocate one seat to the coal miners who dominated the area. Whether intended or not, that was the effect of the district being drawn how it was - and of the FPTP voting system used there.
The odd victories secured in district elections under FPTP were seen in the fortunes of a far-right party in Alberta in 1982. In February of that year, Gordon Kesler, a Western Canada Concept candidate, was elected in a by-election in Olds-Didsbury. The party received 4000 votes in this by-election. (In the Legislature he opposed official bilingualism and protested the introduction of the metric system.)
Later that same year in the 1982 general election, the WCC received 111,000 votes and took no seats. Kesler did not run for re-election and the WCC did not retain the Olds-Didsbury seat. Nor did take any others despite taking 12 percent of the vote, and thus being due ten or more seats.
4. Do all parties run candidates in all districts?
In the 1926 Alberta general election the UFA ran a candidate in 45 of the province's 49 rural districts but did not run in four rural districts. By dong so, it reduced the portion of the votes that it might have received, the amount that it probably would have received if it had run in every rural district. The UFA also only ran one candidate in Edmonton where five seats were being filled, ran none in Calgary where five seats also were being filled and ran none in Medicine Hat where two seats also were being filled.
So of the 181,000 votes cast, the UFA were not able to take any of the 20,000 votes cast in Calgary nor any of the 2800 cast in Edson, 1700 in Grouard, 4500 in Lethbridge and 3700 cast in the Rocky Mountain district. If it had run candidates in these districts, it seems likely to would have taken a large portion of the total votes count and even it had not won any of those districts, its total vote count wold have given it the appearance of being that much more due the majority government it did in fact achieve in that election. (However such a total slate would have made the UFA party appear to be out for power, while not running candidates in Edson, Medicine Hat and Lethbridge left room for Labour to have a good chance at election by running candidates un-opposed by the UFA there. In these districts, Liberals and or Conservative ran against Labour but Labour at least did not have opposition from the UFA in these contests. And the UFA showed itself to be not out for every possible seat or advantage it could take.) In recent elections (those held since the 1970s), each major party runs as complete a slate as it can. And due to the near-random results of FPTP, that has produced an election where one party has taken all the seats (NB in 1987). Such a clean sweep is only a possibility if a party runs a complete slate. Mixed representation would be ensured if no party can run a full slate. But that kind of fairness does not feature in the politics we see in Canada today. (By near-random results I point to how FPTP produces many elections where a major party receives more than 10 percent of the vote but no seats at all.)
5. Is there an actual election held?
Do at least two candidates run candidates in a contest? If only one candidate runs, there is no need for polling to be conducted. No votes are cast. Thus election by acclamation reduces polling possible.
But once more than two are in a contest, the ability of X voting to yield majority rule is lessened. The more candidates that run, the more often a small minority may win the seat. It is common for five candidates to run in a district where a single seat is being filled. And it happens that 31 percent of the vote is enough to win a seat under FPTP sometimes. So a large number of candidates may create mis-representation, just as having a single candidate can do that as well.
Alberta in a very unusual move also allowed the automatic re-election of 11 MLAs in the 1917 general election. Any MLAs serving in the war were allowed to retain their seats without election being held. They were re-elected by acclamation but not because no one wanted to run against them but because no one was permitted to run against them. The un-democratic nature of this ensured that they sat until 1921 (none of them died in the fighting). But the reaction to such high-handedness may have contributed to the fact that none of the 11 were re-elected in 1921 when their seats were finally emptied.
The effect of automatic re-elections is seen if we analyse the 1917 election. Overall, the Liberal party received 34 seats out of 58 in the Legislature, 59 percent which compares not too badly to the party's vote share of 48 percent.
But if we look at only those seats where Liberal candidates competed, we see that they won 28 out of those 45 districts, 62 percent of those seats.
As well, Liberals won six seats by automatic re-election; and the Conservatives won five that way.
Overall, the Conservatives won 19 seats out of 58, 33 percent of the seats, with 42 percent of the vote. But if we look at only those seats where Conservative candidates competed, we see that they won 14 out of those 45 districts, 31 percent of those seats, a good 11 percent lower than its vote percentage.
Election by acclamation - where only one candidate runs with no competition - provide other instances where no actual polling is conducted, and where seat count percentages vary from vote count percentages in ways that might be surprising.
The effect of nine acclamations in the 1909 Alberta election was somewhat similar. All nine of these seats were won by the Liberals, with no votes being cast.
Overall, the Liberal Party won 36 seats out of 41, 88 percent of the seats, with 59 percent of the vote, almost 30 percent higher than its vote proportion.
But if we look at only those seats where Liberal candidates competed, we see that they won 27 out of those 32 districts, 84 percent of those seats, 25 percent higher than its vote proportion warranted.
Alberta had four acclamations in 1930, and then with the emergence of the CCF and Social Credit parties, there were only two more thereafter - one in 1952 and one in 1963.
The acclamation in 1963 - in Taber-Warner - was a sign of voter dis-interest and also contributed to a drop in the number of votes cast that year compared to the previous election, as discussed next.
6. Low voter turn-out
If voters do not use the voting privilege they have, representation may not be proportional to the sentiment held in the overall society. We see this especially in districts where the same party has won the seat numerous times and voters of other parties lose hope and abstain from voting.
In 1986, less than half the eligible voters in the Drayton Valley cast votes. This was one of the lowest turn-outs in the history of the province. This election was the fifth consecutive election where a Progressive-Conservative was elected in that district. The district had just been created five elections before. The next election a different person ran for the P-Cs. The new candidate received fewer votes than had been cast for the P-C candidate in 1986 but again won the seat for the party, although getting less than a majority of the votes cast in the district.
This sometimes affects turn-out across a province overall. This was so strong a phenomena in 1963 that there were actually fewer votes cast in 1963, about when the Social Credit had been in power almost 30 years, than had been cast in 1959. This despite the province having grown in population. Voter turn-out dropped from 64 percent in 1959 to 56 percent in 1963. There were 721,000 eligible voters in 1963 compared to 650,000 in 1959, an increase of 71,000, but 11,000 fewer votes were cast in the latter year.
This drop in turn-out had effect on the perception of proportionality.
The SC party received just lightly fewer voters in 1963 compared to 1959 but took about the same proportion of the total votes, about 55 percent. However the CCF in 1959 received 18,000 votes in 1959 and under its new label NDP, took more than twice that number of votes in 1963 but again won no seats. Judging by the seats count or by the percentage of votes won by the SC, you might think the NDP in 1963 won no more votes than the CCF had done in 1959, but the party's vote tally shows a rising fortune, which, as it happened, would eventually, much later, take the party to government status.
===================================
Proportionality of voting system used
And apart from the factors listed above there is the "normal" mis-representation created by the system itself. Those types of skewed results are briefly described next.
They originate from three main causes: - Wasted votes reduce proportionality. The waste arises in FPTP from no party except the leading one receiving any representation in any district. - Artificial placement of votes also in one election gave an advantage to a party with that "special power."
- Where transferable votes are used, vote transfers between candidate of different parties may cause election results that vary from original party standings.
The normal waste of votes produced by FPTP is discussed in any work on the need for PR so I won't get into that here now. But note that in some elections as much as 69 percent of the votes cast in a district receive no representation and are wasted.
Artificial placement of votes
The Edmonton Bulletin reported how at a local political meeting a farmer/activist pointed out how the "perfidious Wartimes Election Act allowed the soldier vote to be placed where they government chose." This gave the government an advantage and we see that the government took more seats than its percentage of the votes. (Edmonton Bulletin, Dec. 2, 1921)
The Borden Unionist government's ability to place solders' votes meant that in Edmonton East, Frank Oliver took the most votes of votes cast by Edmonton voters but lost his seat when the federal government placed the soldiers' vote.
Borden's Unionist party took 57 percent of the votes but took 153 of the 235 seats in the House of Commons, 65 percent of the seats. In part the party's large percentage was artificially created by extending the vote to female relatives of soldiers serving the army (but not to women who had no male relatives serving) and denying the vote to some naturalized citizens who had already had the vote if they originated from countries that were at war with Canada in WWI and had only been in Canada since 1902. Thus many /some German-Canadians, Turkish-Canadians, Ukrainian-Canadians, etc. were denied the vote in 1917.
Overall, the changes (and the wave of immigrants that came ot Canada between 1911 and the war's start in 1914, which was 400,000 in one years alone) meant that 1,888,702 votes were cast in 1917; in the previous election (1911) 1.3M were cast.
Alberta was just slightly more pro-Unionist than the national average (61 percent versus 57 percent as per votes cast) but all but one of the Alberta 11 MPs elected in this election were Unionist. The party was thus due only six or seven of them but took 10.
Luckily the kind of vote and voter manipulating seen in 1917 has not been seen in any other federal election.
FPTP waste of votes means false-majority governments are normal
But under FPTP it is normal for the leading party to get 10 percent more of the seats than its proportion of the vote. This sometimes produces a false majority government - where a party with only a minority of the votes attains power through a majority government.
Transferable votes can lead to dis-proportionality as measured by party votes
Meanwhile even in systems where proportional representation is the goal, vote transfers may yield seats counts that are not proportional, at least not according to original vote placement anyway.
When a candidate is eliminated in STV or AV systems, that person's votes are transferred to the other candidates. Often at least some of the votes go to each other candidate still in the race, thus many go to candidates of other parties. This is not a mistake - it is evidence of he voters' complete liberty to allocate back-up preferences just as he or she desires and not to be bound by party lines unless desired. Cross-party transfers are particularly seen when the last remaining candidate of a party is eliminated. No intra-party transfer is possible. Votes belonging to voters who vote along party lines and want their vote to go to no other party under any circumstances will not have any usable preferences marked on them at this stage. They will be moved to the "exhausted vote" pile. Among the eliminated candidate's votes will likely be found some votes cast by voters with more open minds than that. Those votes may bear usable back-up preferences. Those votes will be transferred to candidates of other parties. And perfectly in line with the voters desire those votes may aid in the election of a candidate of a party to which the vote was not originally placed. And sometimes even votes cast by voter with open minds will go into the "exhausted vote" pile. Because it could be that the voters although crossing party lines marked all his or her back-up preferences for candidates who had been elected or eliminated. Only a voter who ranks every candidate or every candidate except one can be assured that the vote will remain outside the "exhausted vote" pile. There is no guarantee that any vote, even one with preferences marked for all the candidates will be used to elect someone. For example a vote cast for a candidate who was popular enough to avoid being eliminated but not popular enough, even after vote transfers, to be elected in the end will not used to elect anyone. It will not even be transferred but will stay with candidate it was first marked for until the end when it will be ignored. Votes like that, and votes that somewhere along the way were transferred to that last unsuccessful candidate(s), plus any votes in the "exhausted vote" pile, are the only one ignored in an STV or AV election. The wasted votes, usually no more than about 20 percent of the votes, plus cross-party transfers cause the degree of dis-proportionality that is produced in STV elections. But that dis-proportionality when it does occur is ameliorated by the fact that in most STV elections (elections where more five or more members are being elected) he representation elected is mixed, with no one party taking all the district's seats and that a majority of the votes were used to elect the members, both assurances not seen in the FPTP elections held across a city in historical elections.
For instance in 1959 the Social Credit party took all the provincial seats in Edmonton, far more than its due based on vote counts.
===========================================
留言