In
response to yesterday's column, which called for reference
to the Supreme Court to determine the constitutionality
of the proposal from the Citizens' Assembly, a former
member of the Assembly wrote with an observation that
should be included in any brief to the court. You won't
find it in the report of the Assembly, but the mathematics
of how the single transferable vote works makes STV
inherently unfair.
For
the sake of discussion, assume that a boundaries commission
is able to determine new constituencies so that each
riding has exactly 30,000 voters per MLA. Also assume
that each riding has 80% voter turnout, which means
24,000 votes cast per MLA in each riding. Under these
assumptions, 48,000 people would vote in a riding that
had 2 MLAs and 168,000 people would vote in a riding
that had 7 MLAs, but in both cases the number of voters
per MLA would be the same. Here is how BC-STV would
work. A quotient is determined, which is the number
of votes necessary for a candidate to be declared elected.
The quotient is the number of valid votes divided by
one plus the number of members to be elected. So in
this example, the quotient in the 2 member riding is
16,000; the quotient in the 7 member riding is 21,000.
What's fair about that?
Under
STV, even with perfect equality between the ridings
in terms of the number of voters per MLA, candidates
in the smallest ridings would be declared elected with
5,000 fewer votes than candidates in the largest ridings.
If we work backwards to adjust the number of voters
so that it would take the same number of votes to get
elected regardless of the size of the riding, then we
would violate the basic premise of equality, roughly
the same number of voters per MLA.
It
will be interesting to see if the courts believe that
the inherent inequality between ridings of different
sizes satisfies the constitutional requirement for effective
representation, such that one citizen's vote is not
diluted compared to another's. Anyone who works through
the mechanics of how BC-STV would work, can see that
there is no way to make boundaries for multiple-member
ridings of vastly different sizes (2 - 7) without having
them be unfair when the effectiveness of votes is compared
between ridings of different sizes. That runs contrary
to the case law that applied section 3 of the Canadian
Charter of Rights and Freedoms to the determination
of constituency boundaries.
Defenders
of STV will rush to point out that the number of voters
per MLA varies in the current system, but the courts
have ruled on that and any redistribution that is found
to violate the Charter can be overturned by the courts.
If you combine the problems of unequal population-to-MLA
ratios between urban and rural ridings with the inherent
unfairness of STV, the difference in the number of votes
it takes to get elected becomes much greater. Instead
of 5,000, as shown in the above example, the difference
can be over 20,000. There is nothing fair about that.
With the current single-member ridings, attempts are
made after every two elections to redraw boundaries
to make them fairer. In the case of STV, there is no
way to fix the problem because you either have unfairness
with respect to the number of voters per MLA, or you
have unfairness with respect to the number of voters
it takes to elect an MLA. It is not possible to remedy
both problems.