[This project utilized Tableau and R programming code. The R script, Tableau file, and data are available upon request.]
Introduction
The unceasing debate surrounding the state of global democracy has spurred a scholarly quest to redefine what constitutes a truly democratic society. Unsettled by the limitations of existing indices and driven by a conviction that the health of democracy cannot be captured through electoral processes alone, I embarked on a mission to devise a new measure. This report is the culmination of that mission, reflecting an ambition to construct a more inclusive and representative democracy index that embraces broader dimensions such as civil liberties, equitable governance, and the strength of rule of law. It is an attempt to shine light on the nuanced complexities that define democratic systems, offering a fresh perspective on the ongoing discourse of democracy's trajectory worldwide.
“Democracy arises out of the notion that those who are equal in any respect are equal in all respects; because men are equally free, they claim to be absolutely equal.” — Aristotle
Report
My objective was to expand beyond the liberal democracy index, integrating additional facets of democratic participation and rights not captured by the index. I incorporated metrics for civil society involvement, equal protection, equitable access, resource distribution, and political corruption. I further deconstructed the Polyarchy and Liberal indices into their component indices, such as judicial constraints and freedom of association, to create my custom measure. This approach equally emphasizes these mid-level indices, thereby dampening the focus on electoral democracy and placing slightly more emphasis on rights-based measures compared to the original liberal democracy index.
Overall Trends
Both the unweighted and population-weighted versions of my custom democracy measure exhibit parallel trends; however, population weighting moderates the surge in democracy during the 1990s, reflecting the influence of smaller countries that improved during that era. Conversely, the recent decline in democracy is accentuated in the population-weighted measure, suggesting that this downturn is predominantly influenced by larger countries, as the decline is barely noticeable in the unweighted version but significant in the weighted one.
Change in Regime Classifications
Throughout most of the examined period, my measure captured alternating cycles of democratization and autocratization. However, since 2019, the trend has consistently veered toward autocratization. Given that my measure is generally skewed toward more optimistic assessments of democracy, this persistent shift toward autocratic tendencies is particularly concerning for the prospects of democratic endurance.
Worldview of Democracy
My measure generally rates countries as more democratic compared to V-dem's classification, driven largely by the reclassification of numerous electoral autocracies as electoral democracies, and the transformation of some electoral democracies into liberal ones. In Africa, my measure eliminates the category of closed autocracies entirely, introducing a few nations as liberal democracies where there were previously none. Additionally, countries like Argentina and Brazil now attain the status of liberal democracies within my assessment.
Association with Liberal Democracy
The association scatterplot reveals that under my custom democracy measure, all regime types are rated more democratically than under V-Dem's liberal democracy index. The increase is more moderate for liberal democracies, largely because they are already at the high end of the democratic scale. In contrast, closed autocracies receive a slight uplift despite scoring low on V-Dem's index due to their significant deficiencies in democratic attributes. Overall, my measure portrays a more positive view of global democratic trends compared to the V-Dem index.
The Pearson correlation test results indicate that there is a very strong and statistically significant positive correlation between liberal democracy index and my custom democracy measure. The relationship is linear, and the correlation coefficient of approximately 0.93 suggests that as one variable increases, the other variable also increases correspondingly. The test provides very strong evidence that the correlation observed in my sample is not due to chance.
