Party and Party Dyad Effects: The Network Structure of Government Survival.

• In Event: Poster Session: Formal Political Theory; Political Methodology & Canadian Politics
  In Poster Session: Poster Session: Political Methodology

Abstract

Government durability is vital for a functioning state. One indicator of government durability is the survival time of governments. Previous work on the determinants of government survival has focused on the impact of country and government features and, for the most part, has disregarded the governing actors – political parties. This is unfortunate because, after all, parties play a pivotal role in political decision-making, and it is likely that their features impact the governments of which they are part. In most Western parliamentary democracies, no single party controls the absolute majority of seats, which is why coalition governments form and create inter-organizational networks of coalescing parties. Dyadic features of the ensuing relationships may likewise impact coalition governance, which has also been overlooked by previous research on government survival.
The present paper addresses these shortcomings in two ways. I explicitly model the nesting of parties in governments and countries as well as the network of coalescing parties. This is accomplished by incorporating an aggregation function into survival analysis, making it possible to aggregate both party effects and effects of dyads of parties. The modeling approach allows me to subsequently examine the impact of party and dyadic features on government survival in a theoretically and statistically sound manner.
The examined party feature is the extra-parliamentary organization of parties. While most coalition theories operate under the simplifying assumption that parties can be treated as unitary actors, the organizational structure of political parties likely has consequences for the governance of coalition governments. Leaders of hierarchically organized parties with centralized decision-making processes have great latitude in decision-making. In contrast, democratically organized parties with decentralized decision-making processes offer their party members ample opportunity to participate, making it more difficult for party leaders to reach and maintain inter-party agreement. I therefore expect that governments containing democratically organized coalition parties exhibit higher termination hazards.
Next to party properties, dyadic features may also impact government survival. It has been argued that the history of cooperation among government parties is related to government survival as increasing amount of mutual trust, information about each other, and some form of asset specificity reduce bargaining complexity. While others have included in their models whether the succeeding government is a direct continuation of the previous one, I introduce a social capital perspective by examining the years of previous cooperation and discount them for the time the cooperation dates back.
I draw on the Political Party Database Project, which describes the extra-parliamentary organization of 122 parties from 19 countries. Using a multilevel approach to social network analysis, I explicitly model that parties are nested within (often multiple) governments, both parties and governments are nested within countries, and that coalition parties maintain ties among each other. This is accomplished by representing the effect of parties and dyads of parties as weighted sums of fixed and random effects. The approach resolves the theoretical problem of how parties collectively impact government survival and statistically accounts for unobserved party, dyadic, and country features that otherwise bias the estimates. The Cox model, predominant in the government survival literature, disregards all hierarchical and network structure, a major defect that is overcome with this approach.
The results from this study indicate that intra-party democracy has a negative effect on government survival. While intra-party democracy serves many purposes, there is evidence to suggest that it impairs coalition governance. Katz and Mair 1995 have argued that plebiscitary methods of intra-party democracy are used by the leadership as a device to manipulate members for their own ends and to overcome the resistance of middle-level activists and organized factions. We do not find evidence for this notion as empowering ordinary party members seems to confine party leadership. Examining dyadic features, we find that the accumulated social capital indeed decreases the government termination hazard. The analyses also reveal more unexplained variance at the party than at the country level, vindicating more recent efforts to integrate intra-party politics into explanations of government survival.
To conclude, this paper addresses important lacunae in government survival research by explicitly modeling the interdependence structure of coalition governance. The proposed approach is not limited to government survival and has the potential to advance our understanding of other areas characterized by network structures, such as multi-party wars, treaties, alliances, and international organizations.

Author

• Ben Rosche, Cornell University