Borda Count

[Blake Cretney]

Jurij Toplak wrote:

> Just a comment on the oppinions about Borda count:
> 
> In Slovenia we use Borda system to elect two of our parliament
> representatives. They are elected in singlemember districts. (Other 88
> members of parliament are elected using combination of Droop and D'Hondt
> method).
> 
> As a teacher of constitutional law I use a following example to show
> students possible anomalies of Borda system:
> 
> Borda system is one of rare systems where a person can be elected even if
> practically noone votes for him.
> For example, we have two strong candidates, a right party (RR) and a left
> party (LL) candidate. They both enjoy 50% support. In addition there is one
> more, totally unknown candidate, let's call him Billy. Nobody knows him and
> nobody wants to vote for him.
> Voters of RR candidate, of course, rank candidate RR as n. 1, and because
> they do not want LL to get elected, they rank him as n. 3. They rank Billy
> as n. 2. Voters of LL do the opposite thing.
> 
> Imagine, that there is 100.000 voters and half of them votes for LL and
half
> of them for RR. LL receives 200.000 points (50.000*3+50.000*1) and RR
> receives 200.000 points. Although nobody voted for Billy, he also received
> 200.000 points (100.000*2).
> 
> It is enough for candidate Billy just to vote for himself and he wins.

Interesting.  I wonder, though, what is meant by "nobody voted for
Billy".  Perhaps only one person voted for Billy as their favourite,
but many voted him as their second choice.  To say that nobody voted
for him suggests that only the first choices are real votes.

Of course, the argument against Billy is that he is totally unknown,
and therefore that Borda (or Condorcet) might cause a total unknown to
win.  We are meant to recoil in horror at the thought of unknowns
winning elections.  And I suspect we would all agree that that would
be a terrible result. 

However, the example seems to assume that while we would naturally
choose an election method based on keeping out unknowns, that the
voters themselves are eager to rank unknowns between the leading
candidates.  In so doing, they claim that they would prefer an unknown
to their second favourite known candidate.

Is there any reason to believe that other voters are less concerned
about electing unknowns than we are?  Is this the experience of
Slovenia?

As well, in the above example, Billy would be a very likely winner
during the campaign.  This would tend to focus media attention on him.
 He would then no longer be an unknown.

---
Blake Cretney