8
1998.
8
The sample includes 162 formation opportunities, and over 200,000
potential governments. The vast number of potential governments in the data set is
due to the fact that there is a large number of parties represented in some of the
included legislatures (the number of parties vary between three and fourteen).
9
The first five hypotheses are all based on the assumption of pure office-seeking
actors. Besides information on which parties are included in government, we need
information on the distribution of seats to evaluate these theories. We here use data
on government composition and distribution of seats provided by Müller and Strøm
(2000).
10
Using these data we can determine if a potential government controls a
majority of seats in the legislature, if it is minimal winning or even minimum seats
winning, if the largest party is included, and the number of parties included in it.
In order to evaluate the three policy-based hypotheses, we also need information
on parties’ policy positions along some key dimension. This key dimension is
typically assumed to be a left-right dimension. To measure parties’ ideal policy
positions, we will here rely on data derived from two of the expert surveys that
have been administered during the last decades; the Laver and Hunt (1992) expert
survey administered in 1989 (used only for Luxembourg), and the Huber and
Inglehart (1995) survey administered in 1994.
11
Using this information we can
identify minimal connected winning coalitions, and we can determine the
ideological range of a coalition. The median party hypothesis is also measured
using these data. The median party controls the median legislator, which is found
by adding the parties’ seats from left to right and finding the median. The
incumbent administration variable is a dummy variable, which describes if a
potential government was the previous cabinet or not. Information to create this
variable was drawn from the data set provided by Müller and Strøm (2000).
––––––––––––––––
8
We count a change of cabinet whenever: 1) general elections are held, 2) there is a change in the
party composition of the cabinet, 3) there is a change of Prime Minister, and we only consider
minority situations (formation opportunities where no single party holds a majority of seats).
9
We have a much larger number of potential governments than Martin and Stevenson (2001), as their
full sample consisted of 33,000 potential governments whilst their number of formation opportunities
(220) was higher than ours. This is due to the fact that we work on an updated data set – the last
opportunities they analyzed dated back to the mid-1980s – and the fact that political systems since the
mid-1980s have been much more fragmented than in the pre-1970s period, causing an exponential
increase of potential governments.
10
Updated for post-1998 government formations by the authors.
11
The Laver and Hunt survey (1992) asks country experts to place the parties on a number of scales
ranging between 1–20, and the Huber and Inglehart survey (1995) asks country experts to place the
parties on what they consider to be a left-right scale ranging between 1–10. We have standardized the
scales to vary between 0 and 1. To come to terms with the fact that we had missing values for some
parties, we have performed a small survey with country experts. We however still have some missing
values for some parties (e.g. “others”). A drawback with using these data instead of manifesto data
(which is the main available alternative), is that the expert survey data do not vary over time, that is,
we have to assume that the parties’ policy positions have stayed stable between 1970 and 2000. There
are however several reasons for using expert surveys instead of manifesto data. One reason is that
party manifestos are likely to reflect the image that parties want to give to the voters rather than their
“true” policy positions, a view that has recently been analyzed by Pelizzo (2003: 87). Parties may also
anticipate future coalitions when writing their program, which makes manifesto data unsuitable to use
when studying coalitions (also see Dumont & Bäck forthcoming; Bäck 2003: chapter three; Dumont
2002, 2004). Moreover, if the main advantage of studying the contents of party manifestos is to
provide measures for each election year, the empirical exercise provided by Budge and Klingemann
(2002) does not convince us to use these more dynamic data, as they show that parties’ locations
remain fairly stable and they rarely leapfrog each other on a left-right scale.
Convention