Vote Yes to STV on May 17 (if you live in B.C., anyway)

You may have noticed that I added a (somewhat crude - anyone know a good drawing program?) button to the links on the right. This post explains why it's there (Note: there may be some repetition here for longtime readers, for which I apologize, but I wanted to try and get all the relevant information/arguments/resources together into one post)...

-----
Since the current B.C. Liberal government honoured a campaign promise and brought in fixed election dates we already know, 4 months ahead of time, that on May 17th, B.C. will be going to the polls for a provincial election. At this point, I am somewhat indifferent about which party gets elected (they all have serious problems) although that may change as the date approaches and we actually (hopefully) see some platforms.

But what I will be campaigning for is a 'Yes' vote on the referendum to change the electoral system. The Liberals honoured another election promise by forming a 'Citizen's Assembly' to study whether the electoral system in B.C. should be changed (In the 1996 election the Liberals got more votes than the N.D.P. but the N.D.P. still formed a majority government - this soured the Liberals somewhat on the old (First Past the Post) system).

After months of study, the Assembly recommended that B.C. switch to the 'Single Transferrable Vote' (STV) system which has been used in Ireland since it gained independence in 1922. So along with voting for a party on May 17th, B.C. residents will also vote on whether to switch to STV or to stick to the old system. In order to pass, 60% of people must vote in favour, and there must be a favourable (50% or more, simple majority) vote in favour in at least 60% of the provincial ridings.

I've been reading various people's opinion on whether to vote 'Yes' or 'No' and there seems to be some confusion on what the vote is about:

So, to clarify:

It is not a referendum on how you feel about the Liberal government.
It is not a referendum on the process which was used to generate the choice between STV and the old system.
It is not a choice between STV and your ideal electoral system.
Similarly, it is not a choice between STV and the Mixed Member Proportional System
It is not, or shouldn't be, a choice of which system you think will be best for the party you support.

So, if it's not all that, what is it?

It's simple, it's a choice between whether you think the old (First-Past-the-Post) electoral system or the proposed (STV) system is a better electoral system for B.C.

In order to answer this question we need to ask: what is an electoral system supposed to do? i.e. what makes an electoral system good (or bad)?

In a representative democracy, the electoral system is the mechanism by which all the votes cast (roughly 1.5 million in B.C.) are translated into a list of individuals (79 MLA's in B.C.) who will 'represent' the voters in making laws.¹

In order to do this well, the system needs to accomplish 3 things:

1) Proportionality: The distribution of representatives should be reflective of the distribution of votes. As an extreme example, if some party received 80% of the votes in an election and got 0 seats, we would say that the electoral system had failed. If they got 80% of the votes and 80% of the seats, we would say the system had succeeded.

2) Minimize Wasted Votes: As many people as possible should be represented by someone they voted for. For example, under the current system, if I vote for, say, Bill Derlago and he doesn't get a seat in parliament, there will be nobody in parliament representing my vote, and the system has failed to fully represent me.

3) Ensure Geographic Representation: In addition to each person having representation, each geographical region should have representation as well. For example, if after the B.C. election there were no MLA's from the Okanagan region, we would say the system had failed.²

In order to assess how well each of the two systems do on these 3 tests, we need to understand a bit more about how they work.

First Past the Post

The First-Past-the-Post system is quite simple (in fact simplicity is it's main selling point). The territory in question is divided into equally sized ridings and whoever gets the most votes in each riding is elected.

Proportionality: First Past the Post systems routinely fails to accurately reflect the popular vote in the selection of representatives. Generally speaking, the more parties there are, the worse it does. There are three factors which cause distortion under First Past the Post:

1) The rich get richer: In general, the more votes a party gets, the more it is overrepresented in the legislature. For example, in the last B.C. election, the Liberal party got 57% of the vote, but got 97% (all but 2) of the seats. In the previous election, the N.D.P. got 39% of the votes and got 52% of the seats (and formed a majority). The corollary to this of course, is that parties which get fewer votes are underrepresented. As an example, in the 2001 election in B.C. the Green Party got 12% of the votes (an almost identical percentage as the Bloc Quebecois got in the last federal election) and got 0 seats. Which leads nicely into factor 2)...

2) Regional is better: In general, a party whose voters are more regionally concentrated will be overrepresented in the legislature. We see this most clearly in the national parliament where in the last election the Bloc Quebecois (whose support is concentrated in the French speaking parts of one province) gained 18% of the seats (54 seats) despite only getting 12% of the total vote. The corollary to this is that parties with broad-based support from all regions are underrepresented. The most extreme recent case of this was in the 1993 federal election when the nationally supported Progressive Conservative party got 16% of the vote but only got 0.6% of the seats (2 seats).

3) Luck: Finally, there is an element of pure chance in that a party which happens to win a lot of ridings by close margins will be overrepresented while a party which happens to lose a lot of ridings by close margins will be underrepresented - as happened to some extent in the 1996 B.C. election.

So, the First-Past-the-Post system gives unearned seats to the big, the regionally concentrated and the lucky. And by the same token it cheats the small, the broadly supported and the unlucky.

However, there are further, less obvious problems of proportionality in the First Past the Post system. Because of the way it operates, the First Past the Post system may distort the actual vote itself, by forcing people to vote for someone who is not their first preference.

Consider the case of Ross Perot or Ralph Nader. How many people would have preferred to vote for one of them but didn't because by voting for Perot they actually helped elect Clinton and by voting for Nader (instead of Kerry) they would help elect Bush.

Under First Past the Post, having two parties with a similar message is counter-productive, which discourages the formation of new parties, discourages voting for new parties if they do form, and discourages the continued existence of the new parties even if people do vote for them. Consider the case of the Reform/Alliance party and the Progressive Conservative party. The two parties had different messages which appealed differently to different people. But because they appealed to similar, overlapping groups of people, the two parties were effectively forced to merge in order to have any chance of electoral success.

The end result is that people are never really given the option of voting for the party that they would choose in an unbiased system, so the vote count itself is not proportional with the preferences of the population.

Wasted Votes: With only 2 parties, the proportion of those votes which are wasted (i.e. the percent of the population not represented by anyone they voted for) varies from a maximum of 50%, down to about 25% in really lopsided elections. As the number of parties increases, the number of wasted votes increases. For example, in a four party race, one candidate could win election with only 30% of the votes (with the other 3 getting 25, 25% and 20% respectively). In this case, 70% of the voters will be left unrepresented.

Geographical Representation: First Past the Post does a good job of this, and ensures that each riding is represented by one person.



Single Transferrable Vote:

Before explaining what STV is, it's worth considering what the opposite of First Past the Post would be. First Past the Post breaks the territory down into a number of regions (ridings) such that there are as many ridings as representatives and then chooses one representative from each riding. The opposite would be to only have one riding (equal in size to the whole territory) and then choose all your representatives from that one riding.

This opposite system is known as [Pure] Proportional Representation or Party List Proportional Representation (since it is normally implemented in such a way that who gets elected from each party is based on a list that each party puts together - although there's theoretically no necessity to implement it that way).

Pure Proportional Representation achieves (surprise, surprise) perfect proportionality, but does it by sacrificing geographic representation (and in the case of party lists, people are only really assured that they are represented by the party they voted for, not the individual). It is generally used in geographically compact places where geography isn't a big issue.

So, given that First Past the Post and Pure PR represent two extremes (One person elected per riding & perfect geographic representation vs. one riding from which everyone is elected & perfect proportionality) how do we compromise between them?

There are two ways. The first is to elect half (or some percentage) of the members from First Past the Post and half (or some percentage) from Pure PR. This approach leads to what is typically known as the Mixed Member Proportional system (MMP), the system which was recommended by the Law Commission of Canada and the one which is going to referendum in P.E.I. in November. MMP is a pretty good system and, if you ask me, it's about equally as good as STV in general and probably better for large territories with many regional parties (like the federal government in Canada).

In B.C. however, the Citizen's Assembly rejected MMP and chose the second way to compromise: fewer ridings with more than one person elected from each one - STV.

Under STV you could have, for example, a riding from which 5 people would be elected (this riding would be 5 times the size of the old ridings if you wanted to keep the number of representatives the same). Voters would rank candidates starting with #1 and continuing up to #5. After that, anyone who gets enough first place votes (more than 1/6th of the total, since only 5 people can get more than 1/6 of the votes) is elected (if someone has more votes than they need, then their 'extra' votes are redistributed among the remaining candidates) and if nobody has enough votes to get elected than the person with the fewest first place votes is knocked out and the second place votes on their ballots are redistributed. This process repeats until 5 people are left standing.

It's a bit complicated although it becomes pretty clear if you watch the cute animation here (note: what the Aussie's call Proportional Representation in the animation is what we call STV).

So how does STV do on our three criteria for an effective electoral system?

Proportionality: While not as good as an MMP system or a Pure PR system, STV does much better than First Past the Post. The more people are elected from each riding, the more it resembles pure PR and the better the proportionality.

Minimize Wasted Votes: Because it allows you to express more than one preference, STV is probably the best system for making sure people have someone in parliament who they voted for, since even if their first choice was a fringe candidate with no hope of being elected, once that candidate is eliminated, their second preference will be counted.

Ensure Geographic Representation: Under STV every representative, represents a specific geographic area and every area has (more than one) representative. Some might say this is a disadvantage vs. First Past the Post since the one-to-one connection between citizen and geographical representative has been broken. Others may see it as an improvement since now voters have a choice of more than one representative from their area they can go to in case of trouble, giving them more choice. Personally, I'd say it's a draw, but that's a judgment call.

So on our 3 main criteria, STV is clearly superior on two and about even on the other. This is why the button I added to my site says, 'STV for True Democracy' - because true democracy means an assembly which accurately represents the votes cast and which doesn't waste the votes of half the population.

--
Other Issues:

Besides the 3 primary tests I identified, there are a variety of other issues which can be considered in comparing electoral systems:

Ease of entry for new parties:

First Past the Post: Very difficult. A new party typically needs to get at least 30% of the vote in a single riding in order to elect someone. What percent of the total vote they need in order to get someone elected depends on how regionally concentrated they are. That's why most new parties which form and are able to win seats under First Past the Post are regional protest parties.

Of course even these regional protest parties generally don't last all that long since they don't have the support to form a government and under First Past the Post, they just take votes from any other party with similar policies. So not only can they not gain any power, their presence tends to lead to the party they are protesting against taking power (Perot=Clinton victory, Nader=Bush victory, Reform=Liberal government) This is why there's even a 'law' which says that First Past the Post systems naturally lead to only having two political parties.

STV: Easier. Strategic voting is no longer a problem and a new party will start getting close to gaining seats as its support reaches double digits (roughly - it depends on how many members are elected per riding - the more, the better the chances for small parties).

Support for Extremist Parties: Mildly easier under STV but since you still need at least 10% of the vote (more in most cases) to get a seat, it's a question of how extreme a view something is if 1 in 10 people support it as their first preference.

Turnout: First Past the Post - probably the worst system, due to the high percentage of wasted votes, large numbers of safe seats where the outcome isn't in doubt, the problem of strategic voting and a lack of choice among candidates.

STV: Its complexity may discourage some people, and the percentage of spoiled ballots may be slightly higher, but the far greater choice, combined with no problem with strategic voting, no/fewer safe seats and fewer wasted votes would lead one to believe that turnout would increase under STV vs. First Past the Post.

Choice Among Candidates from the same Party: First Past the Post: none - if you want to vote conservative in Calgary West in the Federal Election you have to vote Rob Anders, the party tells you who to vote for.

STV: Yes - how many choices you have depends on how many people are elected from your riding. If 5, then you would have up to 5 choices of Conservative candidates to vote for.

Complexity: First Past the Post: Probably the simplest system you could come up with other than just flipping a coin.

STV: More complex. It is easy to vote (just put a 1 beside your first choice, a 2 beside your second choice, etc.) but the process for determining who gets elected in each riding is somewhat complicated (in order to make it as fair as possible).

Party Control of Electoral Process: First Past the Post: Limited to choosing who runs in each riding.

STV: Even more limited since voters can choose among the various candidates nominated by a party in a certain riding. If they don't like one of the people who wins nomination, the other candidates will be elected first. This makes it difficult for party leaders to parachute in' outsiders.

Likelihood of Majority Government: First Past the Post: High - because parties with more votes get disproportionately more seats under First Past the Post, majority governments are more likely. This is seen as a plus by those who believe that strong government is better government (up to a point - they generally don't support dictatorship) and who don't mind skewing the results to achieve it.

STV: Lower, because STV is generally proportional, in order to gain a majority the party would need to get close to 50% of the total vote. This doesn't happen very often (although it could, if that's what voters wanted).

Some people express concern that in a minority government situation, a small party could hold the balance of power (like the Bloc does in our current Federal government). Of course, this doesn't mean that a small party can get whatever it wants. Even if it does enter into a partnership with the ruling party, it is generally a junior partner and is lucky to get progress on a couple of it's key issues (those least objectionable to the ruling party, generally). Really, the only change is that the small party holding the balance of power (and the people who voted for them) won't be completely ignored like they would be under a majority situation.

Representation of Women (+ Ethnic Groups, etc.): Although women make up half the population, they have historically made up a much smaller percentage of assembly members under First Past the Post systems. Party list systems (including MMP) can increase the number of women in parliament, basically because the party leaders decide who gets elected, so if they want, they can decide that a certain percentage of those people should be women.

However, like First Past the Post, STV require candidates to both win a nomination for a riding and then win a seat in the riding itself so it isn't likely to increase the number of women elected. In fact, since STV is both more competitive and offers voters more choice than First Past the Post, it may lead to fewer women: if, that is, women are not being elected due to an unwillingness to compete or because voters don't want to vote for women.

Level of Change from Past: First Past the Post: None, First Past the Post is the past.
STV: STV would be a fairly significant change although, if you believe this book (a fairly famous work, which probably deserves a post or two itself), any impact on government effectiveness is likely to be limited - change would be limited to the composition of the assembly. Of course, if you don't like change for any reason, it makes sense to vote for the status quo.

----
Conclusion:

Simple enough?

If not, let's just say that just about anyone who has studied electoral systems seriously (like the Citizen's Assembly) has concluded that STV is better than First Past the Post and the last few decades have seen a lot of countries switch away from First Past the Post and hasn't seen any (that I know of) switch to it. There are reasons for this, and those reasons are what I tried to capture in this post and they're why I'll be voting yes on May 17th.

-------
Notes:

¹For those of you more comfortable in the symbolic world, the electoral system could be considered: F(v) = a, where v = the votes cast, a = the composition of the elected assembly and F is the function (electoral system) which takes v as input and produces a as output. And yes, I am a bit of a math geek.

² Some may wonder why geography is so important but other possible criteria for representation (class, gender, occupational class etc.) are ignored. If so, you should go read Systems of Survival by Jane Jacobs. As she explains, the work of government (and what makes it ethically different from commerce and trade) is that it involves the control and management of physical territory. Hence the importance of territorial representation in the parliament.

--------

Some Resources:

Tons of relevant information at the site for the Citizen's Assembly

Additional info on First Past the Post, and Single Transferrable Vote, at Wikipedia

Dave Pollard had an excellent post on STV over at How to Save the World

The Electoral Reform Society has been studying what electoral systems work the best longer than anyone.

The best illustration of how STV works that I've seen can be found in a flash animation
put together by South Australia's State Electoral Office

One thing to note is that what the Aussie's refer to as Proportional Representation in the animation, we call Single Transferrable Vote, and what they call 'Exclusion (Bottom's Up)' is often referred to as Instant Runoff voting here.

A report on electoral reform from the Law Commission of Canada.

Fair Vote Canada is a group that was been working on improving the electoral systems across Canada.

Here is the same Vote Yes to STV campaign website that the button links to.

Finally, if you think I'm long-winded on this topic, don't go here. Within that site, you can also find this concise summary of the experience with STV in Ireland
----