The principal components of electoral regimes
Cite this dataset
Wiesner, Karoline; Bien, Samuel; Wilson, Matthew (2024). The principal components of electoral regimes [Dataset]. Dryad. https://doi.org/10.5061/dryad.np5hqc030
Abstract
A critical issue for society today is the emergence and decline of democracy worldwide. It is unclear, however, how democratic features, such as elections and civil liberties, influence this change. Democracy indices, which are the standard tool to study this question, are based on the a priori assumption that improvement in any individual feature strengthens democracy overall. We show that this assumption does not always hold. We use the V-Dem dataset for a quantitative study of electoral regimes worldwide during the 20th century. We find a so-far overlooked trade-off between election capability and civil liberties. In particular, we identify a threshold in the democratisation process at which the correlation between election capability and civil liberties flips from negative to positive. Below this threshold we can thus clearly separate two kinds of non-democratic regimes: autocracies that govern through tightly controlled elections and regimes in which citizens are free but under less certainty -- a distinction that existing democracy indices cannot make.
README: The principal components of electoral regimes
Description of the data and file structure
We provide two .csv files containing the results of the principal component analysis that we applied to the V-Dem dataset.
The two files contain the following data:
pca_loadings.csv
This file contains the so-called loadings of the 24 principal components (PC1-PC24, in columns) in terms of the 24 V-Dem variables (in rows).
The table of loadings can be understood as a rotation matrix for transforming the original coordinate system of the V-Dem variables into the PCA coordinate system, i.e., each entry is a linear coefficient representing the significance of a V-Dem variable for a principal component. The loadings already reflect the rescaling of the principal components.
A detailed description of each of the 24 V-Dem variables can be found in the V-Dem Codebook https://v-dem.net/documents/38/V-Dem_Codebook_v14.pdf.
selected_components.csv
This file contains a time series of selected variables on a country level from 1900 to 2021.
Explanation of each column:
- country_name - official name of the country
- country_text_id - international unique three-letter abbreviation of the country name
- year - year of assessment
- EDI - Electoral Democracy Index: an established index to compare the "democraticness" of countries, curated by experts.
- PC1 - the first principal component of our PCA, i.e., the linear combination of the 24 V-Dem variables that exhibits the largest variance.
- PC2 - the second principal component of our PCA, i.e., the linear combination of the 24 V-Dem variables that exhibits the second-largest variance.
- PC2_pos - the positive sub-component of PC2, i.e., the loading-weighted sum of V-Dem variables that have a positive loading in PC2.
- PC2_neg - the negative sub-component of PC2, i.e., the loading-weighted sum of V-Dem variables that have a negative loading in PC2.
Sharing/Access information
Data was derived from the "Varieties of Democracy" (V-Dem) project, version 12 (2022): https://www.V-Dem.net.
Methods
We use version 12 (2022) of the V-Dem data (https://www.V-Dem.net) and apply standard principal component analysis (PCA). Following standard procedure, we normalized each V-Dem variable (i.e. centered it to a mean of zero and rescaled it to a variance of one) prior to performing PCA. For better readability of the plots, we rescaled all principal components uniformly such that the first component has a maximum absolute value of one (i.e. its values are bounded by [-1,1]) while preserving the mean of zero for all components. We further re-oriented each component such that its strongest loading is positive.