Patterns of Global Democracy Promotion: Centrality in DINGO Networks,1981–2015

Abstract

Despite the long-recognized role of international organizational networks in spreading global norms, including democracy, democracy-promoting international nongovernmental organizations (DINGOs) remain understudied. This visualization addresses this gap by plotting nations’ degree centrality within various DINGO networks over time from 1981 to 2015, thereby quantifying, for the first time, the configuration of nonstate democracy promotion networks. The results indicate that all networks are extremely dense, and nations’ mean centrality increases over time. Although dispersion tends to decrease over time, particularly after 2000, relatively high dispersion persists for one network: civil liberties. Thus, although more nations are increasingly integrated within DINGO networks overall, this trend is not uniform. The authors suspect that this difference reflects nations’ growing disillusionment with an enterprise that condemns civil liberties only when geopolitics allow and the subsequently declining traction of civil liberties norms. The results suggest a pivotal yet potentially controversial role of DINGO networks and motivate further research exploring their effects.

How does democracy proliferate? Many point to the importance of international organizations in disseminating global norms, such as democracy, to national contexts (Meyer et al. 1997). As nations become more plugged into organizational networks, the number of pathways through which norms travel expands, accelerating diffusion (Paxton, Hughes, and Reith 2015). Given this, it is surprising that we do not know more about how democracy-promoting international nongovernmental organizations (DINGOs) ¹ contribute to the spread of democracy. Our visualization fills this gap by exploring the architecture underlying DINGO networks using social network analysis, thus providing novel insights about (1) DINGOs and their relations with states and (2) their structural capacity to foster democratization, while (3) providing a more complete picture of the democracy promotion enterprise.

Drawing upon an original DINGO dataset, Figure 1 uses violin and dot plots to visualize the distribution in nations’ degree centrality between 1981 and 2015 across four DINGO network categories. We present (1) the overall DINGO network (“All”) and three DINGO subnetworks promoting different components of democracy ² on the basis of conventional definitions (2) civil liberties, (3) rule of law, and (4) democracy, general (Møller and Skaaning 2012). A trend line (red) plots the mean normalized degree for each category across each year of observation.

Figure 1.

Degree distribution within DINGO networks, 1981 to 2015.

Each network contains a set of “nodes” (i.e., countries) connected by “edges” (i.e., shared DINGO memberships), which enable democratic norms to travel across borders. Countries with more total edges, quantified as degree centrality, therefore experience an expedited flow of democratic norms, to which they become increasingly socialized by DINGOs (Finnemore and Sikkink 1998). To enable comparisons across network categories, we normalize country degree scores from 0 (a country has no edges) to 1 (a country has all possible edges) and plot their distributions by year. Although we considered alternative centrality measures, we prioritize degree because it best enables us to compare the overall distribution of centrality scores across time and network category (rather than, for instance, drawing specific between-country comparisons on the basis of relative positionality, for which eigencentrality is a better fit) (Paxton et al. 2015). Finally, we plot trend lines for graph density. As a network-level measure of integration, graph density measures the proportion of existing edges in a network relative to the total number of possible edges. Because we normalize degree, and because our networks are undirected (i.e., norms flow across nations in both directions), graph density in this case is equivalent to a network’s mean degree. Ranging from 0 (no connections exist) to 1 (all possible connections exist), graph density therefore indicates a network’s average interconnectedness, and thus, its overall propensity for norm diffusion. Further discussion of our methods, alternative metrics, and descriptive tables can be found in the Supplemental Information.

Results

Considering Figure 1, all networks are highly interconnected throughout the period, with most nations clustering toward the upper range of values–as the top-heavy, long-tailed violins and density trend lines indicate. Substantively, this suggests that DINGO networks are extremely dense and cohesive, provide abundant opportunities for interaction and efficient information sharing, and enable democratic norms to flow with ease.

On one hand, this makes sense given the might of the democracy promotion enterprise and the high premium the international community places on democratic norms (Carothers 2008). Because democracy is “the only game in town,” it is simply too costly for legitimate nations to refuse relations with DINGOs, leading to high integration. On the other hand, these results seems surprising; how is global democracy in decline (Diamond 2015) when DINGO networks are so densely interwoven? Within well-connected networks, members closely monitor one another, creating immense pressure to conform to convention, often despite its (in)efficacy (DiMaggio and Powell 1983). In DINGO networks, this may perpetuate adherence to highly standardized models of democracy, which are often ill-fitting and promote superficial democratization (Carothers 1999; Hyde 2011).

Differences in distribution by DINGO category further complicate this narrative. Variability within the rule of law, general democracy, and overall DINGO networks generally decreases over time (barring extreme outliers) as nations converge toward high centrality, rendering high consolidation by the early to mid-2000s (i.e., most countries are well connected and fewer are “left out”). Yet for civil liberties, comparatively more nations remain on the outskirts. Although convergence does occur until 2000, this trend is reversed thereafter as the distribution slightly rewidens. For instance, minimum degree across the period of observation increases from approximately 0.30 to just under 0.80 for general democracy but remains relatively static at about 0.55 for civil liberties. Thus, the flow of civil liberties norms is possibly less fluid, which may stunt diffusion. Current trends in democracy corroborate this narrative. In its report “Freedom in the World,” Freedom House (2018) reported that “civil liberties around the world deteriorated to their lowest point in more than a decade.”

Although we cannot be certain why these differences emerge, we offer one possible explanation. When America launched its war on terror in the early 2000s, it used democracy promotion as justification, all while violating the rights and liberties of countless citizens abroad. This historical turning point illustrated to the world that democracy promotion had once again become a geopolitical game rather than a normative ideal, illustrating firsthand that some governments need only “talk the democracy talk,” without “walking the democracy walk.” This hypocrisy struck a major blow to Western democracy’s reputation (Carothers 2008) and may have prompted nations to disengage from civil liberties DINGOs as act of defiance. However, unable to rebuke democracy entirely without losing legitimacy, nations made a strategic choice: remain embedded within networks that require broad, surface-level, or difficult-to-monitor democratic commitments (i.e., overall, general democracy, and rule of law, respectively) and disengage from those perceived as illegitimate (Zakaria 1997). Considering the current global climate, these trends may reflect what is recognized today to be an apparent and rising global hostility toward Western democracy and the broader liberal international order (Ikenberry 2018).

Conclusions

This visualization documents longitudinal trends in nations’ degree centrality across four DINGO networks, which are highly dense and interconnected. Although variation in centrality decreases over time for most networks, dispersion remains relatively high for civil liberties, particularly after 2000. Future studies can expand on this work by empirically investigating how different DINGO networks help or hinder democratization.

Supplemental Material

sj-docx-1-srd-10.1177_23780231231221462 – Supplemental material for Patterns of Global Democracy Promotion: Centrality in DINGO Networks, 1981–2015

Supplemental material, sj-docx-1-srd-10.1177_23780231231221462 for Patterns of Global Democracy Promotion: Centrality in DINGO Networks, 1981–2015 by Jessica Kim and Andrew Collins in Socius