Model based clustering of political finance regimes: developing the regulation of political finance indicator

Abstract

Political finance literature lacks a common framework for classifying regulatory systems. As these tools are influential in the identification of generalizable relationships, studies assessing political finance in areas such as corruption, competition, and electoral outcomes, often present case specific findings. Using updated International IDEA data, the application of a Multiple Correspondence Analysis and Model Based Clustering framework presents a variable to measure levels of regulation; the ‘Unregulated’, ‘Partially Regulated’ and ‘Strongly Regulated’ system types; and statistics for assessing the certainty of each country’s classification. Applying this methodology to a 180-country sample represents an improvement on previous studies which, due to data limitations, have often used reductive methods and limited sampling. In closing, the ‘Regulation of Political Finance Indicator’ is introduced via Multinomial Logistic Regression, where analyses from prior literature are revisited. Avenues for further study are provided, which may seek to identify generalizable relationships in the areas described above, while also looking to produce ongoing panel data.