Other implications of the consequences of Duverger focus on the issues associ-ated with proportionality and representation. To the extent that proportionality
provides more parties then this provides both more choice to voters and, in
principle, a party voice for a wider range of interests than may be expressed by
just two parties. The wider range of views is especially important to those who are
concerned with minority rights and descriptive representation (see Amy 2000 and
Birch 1971 for general discussion). The concern is not simply a concern for opinion
minorities but for demographic—and most especially racial and ethnic—
minorities. The empirical example here is that of South Africa’s choice of PR in a
racially divided society. A less successful example might be the use of a single
transferable vote system of proportional representation (STV-PR) in Northern
Ireland in the face of sectarianism there. Others, especially those who favor systems
that allow voters to choose individual candidates, stress ‘‘substantive’’ representa-
tion and the activities of the representative on behalf of his/her voters regardless of
whether voter and representative share demographic traits. These kinds of
concerns tend to focus on giving voters the ability to reward and punish represen-
tatives but also to allow voters not just to actex postbut also to select candidates
ex anteon some desirable behavioral attribute. Just how realistic it is to expect
voters to perform these functions is a matter of some concern (see Fearon 1999 for
theoretical discussion; Cain, Ferejohn, and Fiorina 1987 for empirical evidence).
Nevertheless, diVerent electoral systems imply diVerent mixes of the normative
components of representation. One implication buried within Duverger’s Law is
that representation of interests between diVerent interests or groups in society will
be conducted and brokered within a small number of large parties rather than
between parties who must form coalitions in order to govern.
The implications in terms both of governability and also of representation are
related to the bias of the electoral system in favor of or against particular parties. For
the cases of SMSP, especially when compared to list proportional representation,
the bias towards bigger parties and against smaller ones seems especially
pronounced.
Within the class of proportional systems we can see similar, though much more
muted aVects, depending on the particular counting rule involved. Exact propor-
tionality is hard to achieve. It is arithmetically easier to achieve with larger district
magnitudes and large elected assemblies but even with a district magnitude (the
number of seats to be elected) of 100 it is hard to achieve exact proportionality.
There are always fractions of vote shares that cumulate to remainders. How these
fractions and leftovers are distributed can shade the outcome towards larger or
smaller parties, too. In the end the eVects are much less pronounced than those
implied by Duverger’s Law but at the margin the choice of ‘‘largest remainder’’ or
‘‘highest average’’ may well change the results by one or two seats (see Farrell 2001 ,
71 – 9 for an especially clear discussion). This may sound small potatoes but it does
shift who wins and who loses.
electoral systems 581