Improving Representation in Progressive Organizations with the Single
Transferable Vote

By Dennis Pilon

Introduction

Increasingly, it seems, the demands for representation in progressive
organizations are becoming hopelessly complex.

On the one hand, progressives need to work to assure that their
organizations reflect and represent the diversity of Canadian identity.
On the other hand, we have divergent views on key hot button issues like
the environment, economic policy, and the role of government and the
market in social policy - these differences also need effective
representation.

For the most part, progressive forces today are not managing the process
of representing diversity very well. This is because we tend to rely on
old zero-sum methods of election. These 'all or nothing' approaches to
representation force different identity claims to compete against each
other, or with issue-oriented concerns, with lousy results for everybody.

What we need to do is replace our all or nothing methods of election with
a more proportional approach, one that can effectively facilitate the
diversity of representation we both want and need.

And here there is one proportional voting system that is particularly well
suited to our purposes: the single transferable vote (STV).

Proportional representation (PR) is most often associated with public
elections involving political parties. But any organization, even
political parties themselves, could benefit from a 'proportional
representation' of its members, their various identities and opinions. If
slates exist in an organization, STV will represent them proportionately.
If they don't, STV will proportionately represent whatever other interests
exist. For example, since the New York City School Boards adopted STV in
1971 there has been a fairly close proportional representation of the
city's race and ethnic diversity, even though these interests are not
organized as slates or parties.

It's time for progressives to take proportional voting seriously. In what
follows, we'll review how progressive organizations typically elect
representatives now, and why these methods perform so poorly. Then we'll
consider some of the ways in which organizations have tried to 'fix' the
problem, and the contradictory results they've achieved. Finally, we'll
sketch out how the STV form of PR works, and its potential as a
progressive method for electing representatives.

What progressive organizations use now.

Typically, organizations elect representatives to specific positions:
president, secretary, treasurer, etc. Sometimes the winner needs only a
plurality of the vote (more votes than any other candidate), sometimes
they require a majority. The latter is usually accomplished through a
process known technically as 'exhaustive balloting'. Essentially, if no
candidate gains a majority on the first ballot, the candidate with the
least votes is dropped from the contest and balloting is conducted again,
with the whole process repeated until one candidate does secure a majority
of the votes.

Some organizations do not elect representatives to specific positions.
Instead, they elect all representatives 'at large'. For instance, if a
board had ten positions, then voters would have ten votes, and the winning
candidates would be those ten with the most votes at the end of balloting.
Some at large voting rules require voters to cast all the votes they are
entitled to. With our example above, voters would have to cast ten votes
for different candidates or have their ballot declared spoiled and remain
uncounted. Those who support this rule believe that when a voter casts
less than the number of votes they are entitled to, they are 'plumping',
or gaining some kind of unfair advantage. However, it should be noted,
the application of this rule to at large elections is uneven. For
instance, elections in Ontario credit unions require it, but contests for
Vancouver City Council do not.

Finally, many organizations combine the two approaches above. Candidates
may then choose to run for specific positions like president and
treasurer, or seek one of a number of board seats 'at large'.

Why these approaches don't work.

Our conventional methods of electing representatives create 'zero-sum'
results, they are 'winner take all'. Zero-sum approaches to voting may
make sense when elected boards need to make decisions, but for the
elections themselves our first priority must be to effectively represent
our members. Here zero-sum voting rules fail miserably. Let's sketch out
why.

Let's say we have an organization with two broad streams of opinion that
need representation. Even if their respective support is fairly evenly
divided, our conventional voting approaches will produce very lopsided
results. In the example below, candidates for slate A and slate B have
roughly similar levels of support, but slate A wins all the
representation:

positions slate A slate B winner

President 55 45 A
Vice President 54 46 A
Treasurer 52 48 A
Secretary 54 46 A

The results appear even more stark where at large voting is used:

candidates votes cast status

slate A candidate 1 55 elected
slate A candidate 2 54 elected
slate A candidate 3 54 elected
slate A candidate 4 52 elected

slate B candidate 1 48
slate B candidate 2 46
slate B candidate 3 46
slate B candidate 4 46

On the face of it, given that each slate above represents around half the
membership, the most fair result would have seen each slate win two seats
each. As it stands, slate B supporters, nearly 50% of the group's
membership, have no representation at all. And this problem just gets
worse the more diverse an organization is. In the example below, three
slates contest an at large election with very unrepresentative results:

candidates votes cast status

slate A candidate 1 46 elected
slate A candidate 2 44 elected
slate A candidate 3 42 elected
slate A candidate 4 40 elected

slate B candidate 1 35
slate B candidate 2 34
slate B candidate 3 32
slate B candidate 4 30

slate C candidate 1 30
slate C candidate 2 26
slate C candidate 3 22
slate C candidate 4 19

Slates B and C, comprising between them more than 50% of the membership,
gain no representation at all in this at large election.

Of course, not all organizational elections are conducted on the basis of
slates. And even where they are, strong, broadly popular individual
candidates may draw support from all sides and weaken how effective they
are. But these qualifications do not reduce or ameliorate the arbitrary
or unrepresentative nature of 'all or nothing' voting systems. Because
they are 'all or nothing', they tend to systematically over-represent the
largest voting bloc, regardless of how organized they are, and often to
the exclusion of everyone else.

How we've tried to fix them in the past.

In the last three decades, rising concern over representation in
progressive organizations has led to the adoption of many creative
'solutions' to the problem.

Sometimes organizations have relied on informal methods to improve
representational diversity. If an organization is small enough, moral
suasion and individual lobbying can be used, both to assure that a diverse
group of candidates run, and that others do not run against them. In
larger groups, the existence of slates is sometimes used as a vehicle to
achieve a balanced 'ticket', one that embodies all the main
representational concerns.

More formal methods have also been adopted by progressive organizations to
address representational deficiencies. Some set quotas that require
slates to offer a diverse set of candidates, or a specific minimum number
of candidates from a particular representational concern (eg a minimum
number of women candidates). Sometimes quotas apply to the results
themselves, either by instructing voters how to cast their ballots (eg
'you must cast two of your five votes for female candidates'), or by
altering the candidate order after balloting on the basis of a
representational priority (eg 'the top two female vote-getters will be
declared elected regardless of their position in the final standings').
Another formal method adopted to improve representational diversity is to
set aside a certain number of seats for designated groups, and only
members of those groups may run for the positions (eg people of colour,
gays and lesbians, etc).

Though some success has resulted from the adoption of these various
strategies, few would dispute that they have fallen far short of their
initial aims. And, over time, as new and more complicated
representational claims have emerged, these reforms have proven difficult
to adapt. Let's examine why.

Informal methods to fix representational problems often fail because they
rely too heavily on existing elites to manage the process, the very elites
whose unrepresentativeness sparked the concerns in the first place. Where
moral suasion is used to lobby candidates to run or ask others to stand
down, the elites' own view of the 'right' candidates, whether conscious or
not, will be foremost. Additionally, these kinds of informal methods can
lead to friction as some are asked to serve and others are passed over.
While often a necessary starting point in addressing problems with
representation, these kinds of informal approaches cannot be sustained
indefinitely.

On the other hand, where organizational elections regularly feature
competing slates of candidates, informal methods that focus on balancing a
'ticket', particularly in terms of the diversity of identity, can work
effectively. However, while slates can often balance different identity
claims, they tend to be more homogenous in terms of issues or 'ideology'.
And since we know that our 'all or nothing' voting methods typically
reward one slate with all the representation, we still have a problem.

Formal methods of addressing representational problems also face serious
limits.

First, they tend to mistake the nature of the problem. Quotas are often
introduced because their proponents believe it is the only way that voters
will nominate and vote for under-represented groups. But as we've
demonstrated above, our choice of voting system can also place
considerable constraints on voters, forcing them into zero-sum choices (eg
identity versus identity, identity versus issue) that more sophisticated
systems might allow them to avoid. Two decades of research on women's
representation bears this out. Public attitudes have not been the key
factor in whether women get elected or not; much more important have been
the strategic incentives different voting systems offer to parties and
party strategists.

Second, quota approaches, particularly designated seats, can act as a
ceiling that limits minority representation. New Zealand's Maori
population had four seats set aside for them in the national parliament in
1867; since that time Maori representation seldom exceeded four members.
But when that country switched to a proportional voting system a few years
ago Maori representation immediately shot up to their proportion of the
voting population. Designated seats under 'winner take all' voting rules
effectively froze the representation of the Maori minority for over a
century, despite changing circumstances and needs.

There is also a concern over 'who' votes for the candidates running for
designated seats. In many organizations all members do, but criticisms
have been raised that this ultimately allows dominant groups to determine
who minority representatives are.

Finally, formal methods of fixing representational problems through quotas
or setting aside seats also risk weakening the link between
representatives and those they represent. For instance, creating
designated positions can give organizations the illusion that they are
reaching out to different groups, but do little to assure that they are.
There is something to be said for representatives playing a mobilizing
role within organizations, because through the mobilization to get votes,
representatives learn what members want and may have to accountable to
them. And if candidates can't find the support within the organization,
they may have to work harder to bring them in.

Of course, there are times when all of the above approaches are both
appropriate and necessary. Sometimes quotas are required to get certain
representational concerns on the table; sometimes designated seats are the
only way to assure an organization starts working on representing
historically marginalized groups. But there are limits, both structural
and organizational, to how far these approaches can be applied.
Structurally, elected boards can only be so big before they become
unwieldy and unproductive - designated seats can't be added indefinitely.
Organizationally, memberships will eventually revolt if their electoral
choices too hemmed in by competing quotas. As new demands for
representation continue to mount, we've got to find a better way of
facilitating change and representing the diversity of our members.

Our previous 'solutions' have overlooked the most essential part of the
problem - the 'all or nothing' nature of our voting systems. If it is
better representation we want, then we have to change the logic of our
methods of electing those representatives. All or nothing is a
understandable principle when one faces a zero-sum choice (ie 'do we allow
smoking in this room or not?') but it fails miserably in 're-presenting'
the diversity of an organization in elections. Instead, what we really
want is a proportional representation of our diverse identities and
interests. And that means we need to consider voting systems that operate
proportionately.

Proportional voting for progressive organizations: a new 'solution'

Proportional representation (PR) is a pretty straightforward idea; if you
get the votes, you should get the seats. If a minority view in an
organization represents ten percent of the members, and there is a
ten-member board, under PR they should be able to elect one
representative. In fact, with PR, minorities can't be denied their
representation if they have sufficient voting strength. This is because,
unlike all or nothing voting systems, successful candidates do not require
a plurality or majority of all the votes cast to gain election, they only
need a proportion of the overall vote. The proportion itself is usually a
function of the size of the board - with a four member board the
proportion of the vote needed for election would be approximately
twenty-five percent, with ten members it would be ten percent, and so on.

All kinds of organizations in Europe use proportional voting for their
internal elections. Where competition between organized slates is common,
most use a 'list' form of PR. Here getting to a proportional result is
easy. If a slate gets twenty percent of the vote, it will get twenty
percent of the seats, with the actual candidates gaining election being
the first twenty percent of the names on their ordered list. Progressive
organizations here could adopt a PR-list system and get excellent
proportional results. However, one potentially limiting feature of
PR-list is that it requires the existence of slates to work (under this
system seats can only be assigned proportionately to groups, not
individuals). Of course, not all organizations feature slates in their
elections. As such, PR-list won't suit every groups' need for
proportional voting.

A more flexible form of PR, one that creates proportional results whether
there are slates or not, is the single transferable vote (STV).

With STV, voters rank individual candidates numerically, their first
choice as '1', their second as '2', and so on. Voters keep on ranking
candidates for as long or as little as they like. Voters could restrict
their rankings to members of one slate or, alternatively, priorize a
particular issue or identity and rank candidates accordingly, regardless
of which slate they belong to.

As should be apparent, voting with STV is pretty simple. Getting to our
results, however, is a bit more complicated. When balloting is completed,
votes are tabulated on the basis of all the first choices marked. Some
candidates will probably have obtained a high enough proportion of the
vote, or 'quota' as it is called in STV, to be declared elected. The
portion of their vote above the quota is then redistributed to other
candidates - this is how STV creates proportional results. For example,
let's say a candidate has obtained twice the proportion of votes needed to
get elected. What this means is that the candidate really only needs half
of each vote they have received. The other half of each of their votes
can now be transferred to the second choices that have been marked on
them. Thus one vote may ultimately help a number of candidates gain
election. When the transferred surplus from elected candidates is no
longer enough to elect anyone else, our attention shifts to transferring
the votes of the least popular candidates, and so on. This process
continues until all the positions are filled. All this used to be done by
hand but now computer programs can do all this calculating and counting.

For the members of progressive organizations, STV allows for some pretty
sophisticated voting. When the results are in, we will discover that most
voters contributed to the election of a number of candidates, either by
taking direction from slate organizers or marking preferences with their
own ideas of what is best for the organization. Either way STV-elected
boards will represent a diversity of identities and ideas. STV also gets
around the thorny issue of majorities defining (by their voting strength)
who minority representatives will be - if minorities have the votes they
can't be denied their choice of representative.

Conclusion

As our society becomes more diverse, we need to adapt our organizations to
respond to and represent these changes. To do so effectively, we've got
to examine the ways we try to represent diversity. So far, we've been
lurching from one half measure to another, always patching up some
'solution' but never really getting to the root of the problem. It's now
time for a root and branch approach to change. Diverse representation
simply cannot be reconciled with 'winner take all', 'all or nothing'
approaches to voting.

We've got to give up these old, familiar ways of electing representatives,
and adopt methods more in line with our commitments to diversity and
inclusion. Proportional voting is the very embodiment of these new
values. Where all or nothing voting often leads to lopsided and skewed
results, a proportional allocation of representation can't help but
'include', it can't help but be 'diverse'. And the STV form of PR
particularly gives members, both individually and in groups, a great deal
of power to shape the representative process, either by forming or not
forming slates, or by initiating new claims for representation.

Of course, voting systems alone do not create any results. People still
have to act to better representation regardless of how we count the votes.
Nor can a change in voting system solve all problems of representation.
Progressive organizations will still have to address how inequalities of
all kinds affect and diminish the democratic opportunities for some, and
enhance those of others. But the choice of voting system might contribute
a great deal to how we take up these challenges, how divisive they become,
and how much reward we reap from our ongoing efforts. For these reasons
alone, the question of voting system reform in progressive organizations
deserves our serious consideration.

Back to top