third kind of uncertainty—over the speciWcs of electoral system eVects—may well
deter change but it may not be the only factor at work in shaping the decision to
change system or not in established democracies.
One of the biggest roadblocks to change is surely that it requires winners under
the current system to consider altering a system under which they have won. Benoit
( 2003 ) argues that we can expect to see electoral system change when the people
who have the power to do so see a way to improve their seat share under alternative
electoral arrangements. The obvious point is that change in an electoral system—
especially if its eVects are uncertain—essentially asks current winners to run the
risk of losing. Parties and candidates are typically reluctant to do that to themselves
and so uncertainty will impact whether or not players think they can win more
seats. In the meantime, the current incumbents are doing justWne out of the
current system.
Self-interest, then, provides a major obstacle to change not least because reform-
ers may be tempted to renege. Reform promises, for example, were long a part of
the platform of the Parti Quebecois (PQ) in Quebec. After winning only a handful
seats in the provincial legislature, despite winning 24 and 30 percent of the vote in
the 1970 and 1973 elections respectively, the PQ promised to incorporate PR into
Quebec’s electoral system. When the PQ won power in 1976 , premier Levesque set
up a Ministry for Parliamentary and Electoral Reform whose mandate included
consideration of alternate voting systems for Quebec, but proposals were shelved.
Milner’s explanation for this failure of reform refers in part to the uncertainty of
members of the PQ over the eVects of any change and also the belief of many PQ
parliamentarians that they, themselves, were safe from electoral loss because of the
strength of their standing within their own districts even though the party itself was
low in the polls (Milner 1994 ). Similarly, while the UK Labour Party’s commitment
to electoral reform was put into practice for European elections, city elections, and
elections to the assemblies in Wales and Scotland, they were not put in place for
general elections to the national parliament—the elections that matter.
In addition to demonstrating the infrequency of electoral change, Colomer’s
Wgures also show a trend towards ‘‘increasingly inclusive, less risk formulas...
[from majority systems to] mixed systems and to proportional representation’’
(Colomer 2005 , 4 ). Proportionality should make change even less likely by giving
more parties a stake in the current system. Under majoritarian systems losing is
both deWnitive (‘‘winner takes all’’ after all) and likely to aVect relatively large
numbers of candidates and parties—the main actors—in the system. Under
proportional systems, however, not only are more players likely to be included in
government but also a wider variety of opinions and parties are much more likely
to be elected to some role in the system to begin with (Powell 2000 ). That is,
proportionality may well create a broad enough group of winners or stakeholders
to make subsequent change harder. Once systems drift towards proportionality it
may be hard to move back away from it.
586 shaun bowler