Approval-Then-Condorcet Voting System

First of all...

If you are not an election method expert, you can safely ignore this page for now...

Since I have advertised this web site on the Election Method list, in february, I have received several private emails commenting on this proposal. Here is part of a public reply made to the same list:

Re: [EM] Educative Alternative Voting board...
From:  Anguo 
Date: 11/02/2004 9:14 pm

>       A few quick replies. First of all, that you for your
> work on educating people about Condorcet and alternative
> voting methods in general. It is my hope that eventually
> this issue won't be so obscure. 

Thank you for your encouragement. As far as I know, there 
isn't a single country in the world that uses Condorcet for 
its national elections. If we want this to change, we need 
a broad public support, and before that, a general public 
understanding of what Condorcet has to offer. 

There's nothing in my site that is new to the members of 
this list. 
I am no election method expert but just generally enthusiast 
about Condorcet. 
My site is more aimed at those who don't know yet Condorcet 
and those who have yet to fully understand what it has to 
offer. 

>       Last, I disagree with your assertion that the Condorcet
> method doesn't work well when there are a large number of
> candidates, and therefore I also do not think that your
> proposal for an approval then Condorcet procedure is
> necessary. What is your reasoning for wanting to avoid a
> situation in Condorcet where there are a lot of
> candidates? Is it just the hassle of printing out ballots
> that big? Maybe the answer is simply a computer
> interface. As far as voters truncating their ballots when
> the come to candidates whom they have never heard of,
> this doesn't create a problem in Condorcet. Candidates
> who a big chunk of the voters have never heard of are
> unlikely to win, and that is as it should be.

Not understanding that much, I just listened to the people I 
was in touch with when the coding of this patch just got 
started. One of those people was adamant that a Condorcet 
ballot, to be fully honest and compliant to the Condorcet 
method, should be filled out in full, with all the 
candidates ranked from the first to the last. I coded the 
BB accordingly.

Now, you are telling me that a truncated ballot is equally 
valid and that an election where truncating is allowed 
complies with the same criteria as an election where it is 
not.
In such a case, then you are right again: my 'solution' 
looses its "raison d'etre". A single round of Condorcet 
election would suffice.
If such is the concensus among all the experts, then I shall 
humbly withdraw my proposal. 

I tried to look information on the internet about the effect 
(or lack thereof) of truncating Condorcet ballots. I found 
none. 

Is it a case where there are two school of thoughts: 
truncate or not truncate, that is the question. ?
In such a case, I can modify the path so that both 
alternative are available in the Board (as it stands now, 
the board requires a fully ranked ballot). 
(I have a vague recollection of a city using Condorcet but 
requiring all the candidates to be ranked, but I am far 
from sure).

Or was I just listening to someone misinformed?
Then, I still have to change the code to reflect this.

Now, I am still waiting for more expert feedback. I have already changed the phpBB: there are now two Condorcet options. The original one where a full ranking is required (maybe the EM experts can help me find a better name?) and the traditional, simple condorcet where any ranking is acceptable (try both to see the differences in User Interfaces). The default and recommanded option is the newly coded simple condorcet.

As far as this proposal is concerned, I will consult knowledgeable experts and modify/delete it accordingly. Meanwhile, true to my desire to make of this site an educative site aimed at those who don't already know about Condorcet, I am spending the little available time I have on improving the phpBB modification and the related documentation. The latter should be made available online in the next few weeks.

Thank you for your help in promoting Condorcet

A simple solution

For major elections, where the number of candidates can be quite high, I advocate the use of a two rounds voting method. The first round, we use the Approval Method, to stream out the number of candidates. We retain the five most approved candidates for the second round. For the second round, we use Condorcet.

There should be a fair amount of time between the two rounds, about two months, so that the five approved candidates have time to clarify further their policies.

Why a two round system?

Condorcet is often recognized as the fairest election method but at the same time, it is also said that it can get too complicated should the number of candidates be too high. In order to retain the full benefit of Condorcet, the votes should not be truncated, meaning that all the candidates should be ranked from the favorite to the least favorite. It is ok to do so with four or five candidates, but doing the same thing with ten to twenty candidates can be a hassle and even unpractical. Moreover, Majority Voting favorises the candidates from the two biggest parties, discouraging smaller candidates to even take part in the election, but under a Condorcet Voting system, more candidates will come forward to participate in the election. Thus, the number of candidate can be fairly high: certainly above ten, and maybe up to a few dozens (it also depends on what the law says are the pre-conditions to be an official candidate).

The solution to this problem is obviously to take the best from each of the two voting methods: while keeping the Condorcet Method for the second round, with a limited number of candidates, we can use the Approval Method to select among a large pool of candidates. Approval is simple to understand and to use, and the voters are unaffected by the number of candidates. They can just ignoreall the candidates they don't know, and focus on approving those they know and like.

Electing well known candidates

Condorcet Voting allows for the citizens to voice their true opinion. This election method is much fairer but only if the voters use it to voice their complete, true preferences. That's why, in an ideal situation, voters should not be able to truncate their vote. It is however difficult to rank candidates one knows very little about. Imagine an election like the French 2002 Presidential Elections with 16 competing candidates, many of which were fairly new-comers on the political scene. It would have been nearly impossible to rank them all without understanding the sometimes subtle differences in their platform. It would have been especially futile to try to make a difference between the most extreme and least prefered candidates.

An election with a First Round Approval system would ensure that such a difficulty never arises. With Approval, many candidates can have scores of 50% Approval or more. It can be expected that the 5 candidates selected for the Second Round Condorcet election will all have approval ratings ranging from at the very least 30% to well over 50%. Such candidates would be well known to the public who should not have too much difficulties ranking them.

It is otherwise imporant that the candidate finally elected should be well known. In the second round, the candidates would all have received high approval ratings and they are sure to have enough media exposure.

This emphasis on having only well-known candidates participating in the second round is in no way detrimental to small, lesser known candidates. As seen earlier, in Majority voting the small parties never have a real chance to succeed. Voters shy away from them because they fear to help their least favorite candidate to win. Small parties are doomed to remain small, election after election, even those who have policies that could be very popular should they be given a fair chance. Even in the French Presidential election, which has a Two round majority voting system said to give smaller parties a chance to voice their opinions, it is expected that voters, especially the left-wing electorate, will abandon their favorite but smaller candidates in order to prevent candidates who are disapproved of by 82% of the electorate to be selected for the second round. That 2002 election, despite the high number of candidates, has reinforced the power of the two main parties and will limit the choice of the electorate in future elections. With Majority voting, it is a matter of all or nothing: either you are backed by one of the two big, powerful, wealthy party, or you've no real chance to succeed.

With Approval-then-Condorcet, small parties have a chance in the first round to voice their policies. If those policies are popular, people will approve them without any fear or calculating mind. A good showing, even if not enough to be selected for the second round, will ensure that they get better media exposure during the next election. With better mediatic support, the electorate will get to know them better and be able to decide if indeed their policies can be approved of. Thus, election after election, small but could-be popular parties have a real opportunity to prove themselves. It is only a matter of time (not that long) before they can participate in the second round. The process is much smoother, progressive and, eventually, fairer.

Back to election index.