Party Positions in the EP -- p12
[Table 3 about here]
A few expected results for individual parties stand out from Table 3, such as the very
high importance attached to the environment for the Greens. Also, it is interesting that for the
UEN and EDD—the two right, Euro-sceptic party groups—the international issues were the
most important, with other policy dimensions of only middling importance.
The components of left and right in the EP
The results summarized in Figure 2 seem to suggest that two broad dimensions of policy
competition are present in the European Parliament. The first represents the classic national
policy issues associated with left and right, namely economic and social liberalism, and also a
bundle of relatively newer issues such as immigration and the environment. The second
dimension relates to the authority and institutions of the European Union itself. Substantively,
the question is whether the EU policy space is uni-dimensional or instead consists of two or
possibly more dimensions. The EP policy space has previously been described as uni-
dimensional with the traditional left-right or “regulation” dominating (Tsebelis and Garrett
2000; Kreppel and Tsebelis 1999) or one-dimensional with geo-political pressures defining
the principal axis of competition (Hoffman 1966, Moravcsik 1998). Other scholars, however,
have described the European policy space as consisting of two dimensions, a left-right
dimension composed of economic and socio-political issues from the domestic arena, and an
orthogonal dimension of EU integration versus national sovereignty (Hix and Lord 1997).
Variations on the two-dimensional characterization relate to whether positions on EU
integration are significantly correlated with left-right (e.g. Hooghe and Marks 2001, Gabel
and Hix 2004) or whether positioning on the two dimensions is independent.
In Table 4 we have used principal components factor analysis to group and separate the
constituent policy dimensional scorings into orthogonal factors. In order to explore the issue
of what policy dimensions were grouped with left and right, we also included the general left-
right dimension. Two factors clearly emerge (having eigenvalues well above 1.0), together
explaining more than 77% of the variance in specific policy placements. The last panel in the
table provides the varimax-rotated factor loadings for the eight constituent policy dimensions