years—in Italy, New Zealand, and Japan—attracted a great deal of attention as
special cases. Colomer ( 2004 ) presents an encyclopedic study of electoral system
change. In general there are not many examples of change. Colomer gives a total of
eighty-two changes from the nineteenth century. Of these, just fourteen are
considered to have taken place in the present democratic period (Colomer 2004 ,
57 ). Given the (increasing) number of democracies and the number of years
involved, examples of change are few and far between. It is hard to come up with
satisfactory priors of how often we might reasonably expect to see electoral systems
change or how often the opportunity for change would come up. With, say, twenty
to thirty democracies over a thirty- toWfty-year time period this would seem to
give somewhere between 60 and 150 opportunities to change electoral systems if we
are willing to assume that each country has a chance to change its electoral system
every decade (i.e. every two electoral cycles). If a more reasonable timeframe is
once every generation (thirty years) then theWgures drop to between twenty and
Wfty opportunities to change. But theWgures presented above are lower by far than
these numbers or, at the least, suggest that most attempts at change fail.
One of the main patterns to explain, then, is the striking absence of change in
electoral systems. Andrews and Jackman ( 2005 ) note the importance of uncertainty
as a deterrent to change (Colomer 2004 , 6 ; Shvetsova 2003 ). They identify three
kinds of uncertainty at work in electoral reforms.
- 1 Uncertainty over the Number of Political Parties
At moments of constitutional choice—in the wave of democratizations in Eastern
Europe in the 1990 s—it was far from clear who the players were going to be. Parties
other than the Communist party typically did not exist in any recognizable or
organized form and so the identity, party programme, and size of successor parties
was not known to electoral engineers. Instability in party systems and blocs in the
early post-Wall years did little to help this uncertainty settle down and so allow
accurate gaming. Explanations of electoral choice in early stages of democratization
can easily assume that parties and players are more uniWed and cohesive than might
actually be the case. In fact, the early electoral arrangements pretty much determine
which parties exist to play the game of institutional choice in the next round. But
even in established democracies it may not be entirely clear what the impact of a shift
in electoral system will bring about by way of new entrants into the system. New
Zealand’s change from single-member simple plurality to a mixed-member propor-
tional system (MMP) provides a good and remarkably well-documented ‘‘real
world’’ example of this point and the following one (Vowles 1995 ; Vowles et al.
2002 ; Boston, Levine, Mcheay, and Roberts 1997 ; Remington and Smith 1996 for a
Russian example; Bawn 1993 for a German example).
electoral systems 583