A few more thoughts on STV
A few more comments on elections, and thanks to the comments on the post below.
First, Chris Lightfoot reminds me of Arrow's Theorem which makes the point I made about how there's no perfect electoral system. I had heard of it before, but it slipped my mind when I wrote the post, so thanks to Chris for reminding me. This page makes the case for the Condorcet system being satisfactory on the grounds of Arrow being unneccesarily stringent and indeed, in that small set of elections that can only return a single winner, it may be a better system than the Alternative Vote.
The other issue, raised by Matthew in the comments, is the idea of STV as a Proportional Representation (PR) system of election. Now, this may be more of a personal foible (after all, the ERS was originally called the Proportional Representation Society) but I'm not sure that one can classify STV as a PR system, at least within the common perception of 'PR'.
I think the issue stems from the general idea that the main 'problem' with FPTP is that it doesn't guarantee the delivery of a proportional result where proportionality is defined as the individual parties receiving a number of MPs broadly equivalent to their level of support in the country as a whole (that's not to say that it can't, of course). Thus, the alternatives to FPTP are often grouped together under the banner of 'proportional representation'. The various party list systems are proportional representation systems under this definition but I would argue that STV is not - I think an overall 'proportional' result is more likely under STV than FPTP, but it is not a necessary outcome of the system.
As I said below, the key difference between STV and party list systems is that where they deal with votes for parties as a whole across a region or country, STV deals with votes for individual candidates in a constituency in the same way as FPTP. It's important to remember that when one looks at the share of the vote in general elections currently, that assumes each vote has been cast for an individual because they are the representative of their party in that constituency. For the vast majority of voters, this is the case, but it's not universal (for instance, I'm sure that many of the people who elected the independents Martin Bell and Richard Taylor wanted a Conservative or Labour government, respectively, but did not want to vote for the individual party candidate in that constituency).
The same issue applies to STV elections in that it's hard to determine what would count as a 'proportional' result in an STV election. One could assume that whichever party a voter casts their first preference for is their preferred party - and I suspect this would likely be the case for the majority of voters - but any strict proportionality would have to assume that their immediate further preferences were for the other candidates of that party in the constituency.
As an example, assume that we're in a constituency that returns five members under STV. One candidate of Party X has a huge level of personal support in that constituency and receives 70% of the first preference vote. If the result was to be proportional then one would expect Party X to get 4 of the 5 MPs based on that level of support. However, if the support is for the individual rather than the party then when his surplus votes are transferred they will go to the candidates of the voters' parties of preference rather than Party X. Thus, while Party X receiving 4 MPs would seem the proportional result, it's not the result the system delivers because the voters don't wish that result. They elect the five candidates who they support the most, not the parties they support the most.
However, if one assumes that voters generally use their preferences within candidates of their preferred party - as I said before, this would likely be the case for the majority of voters and in the vast majorities of constituencies - then results will be broadly proportional, the degree of proportionality depending on the size of constituencies in the system - the larger they are, the more likely the overall result is likely to tend towards proportionality.
As a final minor point, I think that, if STV was to be introduced in Britain, I'd like to see constituencies returning five MPs in most cases, though there should be flexibility on that point - rural constituencies could return as few as three, lest the constituency becomes too large in terms of area to be credible, and urban ones could return as many as seven, especially when a larger constituency could then cover an entire city. However, I think five is the optimum size, as it would require a party or individual to get over 16% of the vote in that constituency to win a seat automatically. This would be enough, I think, to limit the risk of extremist candidates getting elected but it's still low enough to give a chance for smaller parties and independents who can attract later preferences from voters to get elected.
First, Chris Lightfoot reminds me of Arrow's Theorem which makes the point I made about how there's no perfect electoral system. I had heard of it before, but it slipped my mind when I wrote the post, so thanks to Chris for reminding me. This page makes the case for the Condorcet system being satisfactory on the grounds of Arrow being unneccesarily stringent and indeed, in that small set of elections that can only return a single winner, it may be a better system than the Alternative Vote.
The other issue, raised by Matthew in the comments, is the idea of STV as a Proportional Representation (PR) system of election. Now, this may be more of a personal foible (after all, the ERS was originally called the Proportional Representation Society) but I'm not sure that one can classify STV as a PR system, at least within the common perception of 'PR'.
I think the issue stems from the general idea that the main 'problem' with FPTP is that it doesn't guarantee the delivery of a proportional result where proportionality is defined as the individual parties receiving a number of MPs broadly equivalent to their level of support in the country as a whole (that's not to say that it can't, of course). Thus, the alternatives to FPTP are often grouped together under the banner of 'proportional representation'. The various party list systems are proportional representation systems under this definition but I would argue that STV is not - I think an overall 'proportional' result is more likely under STV than FPTP, but it is not a necessary outcome of the system.
As I said below, the key difference between STV and party list systems is that where they deal with votes for parties as a whole across a region or country, STV deals with votes for individual candidates in a constituency in the same way as FPTP. It's important to remember that when one looks at the share of the vote in general elections currently, that assumes each vote has been cast for an individual because they are the representative of their party in that constituency. For the vast majority of voters, this is the case, but it's not universal (for instance, I'm sure that many of the people who elected the independents Martin Bell and Richard Taylor wanted a Conservative or Labour government, respectively, but did not want to vote for the individual party candidate in that constituency).
The same issue applies to STV elections in that it's hard to determine what would count as a 'proportional' result in an STV election. One could assume that whichever party a voter casts their first preference for is their preferred party - and I suspect this would likely be the case for the majority of voters - but any strict proportionality would have to assume that their immediate further preferences were for the other candidates of that party in the constituency.
As an example, assume that we're in a constituency that returns five members under STV. One candidate of Party X has a huge level of personal support in that constituency and receives 70% of the first preference vote. If the result was to be proportional then one would expect Party X to get 4 of the 5 MPs based on that level of support. However, if the support is for the individual rather than the party then when his surplus votes are transferred they will go to the candidates of the voters' parties of preference rather than Party X. Thus, while Party X receiving 4 MPs would seem the proportional result, it's not the result the system delivers because the voters don't wish that result. They elect the five candidates who they support the most, not the parties they support the most.
However, if one assumes that voters generally use their preferences within candidates of their preferred party - as I said before, this would likely be the case for the majority of voters and in the vast majorities of constituencies - then results will be broadly proportional, the degree of proportionality depending on the size of constituencies in the system - the larger they are, the more likely the overall result is likely to tend towards proportionality.
As a final minor point, I think that, if STV was to be introduced in Britain, I'd like to see constituencies returning five MPs in most cases, though there should be flexibility on that point - rural constituencies could return as few as three, lest the constituency becomes too large in terms of area to be credible, and urban ones could return as many as seven, especially when a larger constituency could then cover an entire city. However, I think five is the optimum size, as it would require a party or individual to get over 16% of the vote in that constituency to win a seat automatically. This would be enough, I think, to limit the risk of extremist candidates getting elected but it's still low enough to give a chance for smaller parties and independents who can attract later preferences from voters to get elected.
Nick got away with this at 4:07 PM
0 Comments:
Post a Comment
<< Home