Constructing the 2013 German Political Spectrum

Abstract

The 2013 German political spectrum is constructed like the 2017 political spectrum is constructed in Chapter 9. For this purpose, we use the data about the positions of 28 German parties on 38 policy issues shortly before the 2013 Bundestag (federal) election. They are arranged into 28 38-dimensional vectors understood as the parties’ policy profiles, and the correlation between them is regarded as a proximity measure. The political spectrum— a contiguous party ordering where neighboring parties have close profiles — is constructed by applying the principal component analysis (PCA) to the matrix of correlations between the parties’ profiles. As in Chapter 9, the resulting ordering is exactly the left–right axis rolled into a horseshoe-like arc. The far-left and far-right ends approach each other, although they remain somewhat distant. For comparisons, contiguous party orderings are constructed using four other models. Among other things, the highly realistic political spectrum obtained this way demonstrates the flexible applicability of the framework studied.