[EM] Comparing ranked versus unranked methods

[Adam Tarr]

>Using d'Hondt's rule, this sort of offensive strategic manipulation
>by clever vote-splitting appears to be impossible... it seems obvious
>from playing with examples, although I'm having trouble coming up
>with a clean way to explain it.  So, it looks like d'Hondt might be
>the better choice for PAV for strategic reasons.

I think a slight addendum is deserved here.  All of the strategic 
manipulations I mention for PAV are also possible in any list PR that uses 
Webster's rule for allocation of seats.  These problems are not really 
unique to PAV.  Despite this, I think these problems are a much greater 
concern in PAV.

In list PR, the candidate lists for parties have to be formed well in 
advance of the election.  It would be very difficult for a party to foresee 
a specific voting breakdown, and split their party into two or more parties 
to take advantage of this.  Even if the voting is extremely predicable, 
this still invites some political fallout from splitting into two 
parties.  In general, the nature of list PR makes it very difficult to take 
advantage of the strategic manipulations made possible by Webster's rule.

In PAV, however, the party faction split need not be set up by an actual 
split of the party in advance of the election.  It can be done manually by 
the voters at the ballots.  Moreover, the party leadership need not be 
obvious about their strategy; they can merely tell their supporters "just 
vote for your favorite two or three candidates, no more" and they might 
expect roughly the desired breakdown.

There is already an incentive in PAV (and STV) to leave off any sure 
winners who you support, in order to make your vote count more strongly 
toward the marginal candidates you support.  The additional incentive to 
shorten your list if Webster's method is applied to PAV could corrupt the 
proportionality of the results.  For this reason, d'Hondt makes more sense 
in PAV.  Standard list PR does not give the voter the freedom to manipulate 
the results in this way, so I would support Webster's rule for a list PR 
system.

-Adam