The Determinants of the Success of Transitions to Democracy

Abstract

The article provides an analysis of the determinants of the success of transitions to democracy based on a combination of qualitative comparative analysis of Poland, the Czech Republic, Slovakia, Hungary, Slovenia, Croatia, Bosnia & Herzegovina, Serbia, Montenegro and Macedonia; and hierarchical clustering. The key finding is that one can reveal configurations of several factors which jointly determine either continuous or disrupted transitions to democracy.

Notes

We are grateful to the anonymous referees for their comments and the Slovenian Research Agency for financial support.

During the 1990s the classical dichotomy of parliamentary and presidential constitutional systems was overcome by a continuum of types combining various elements of the two (Shugart 1993; Kasapović1997). For the purpose of our empirical research we distinguish between parliamentary, semi-presidential and presidential systems.

UN Statistical Division, available at: http://esa.un.org/unsd/mbsdemo/mbssearch.asp, accessed 15 January 2001; World Health Organisation, available at: http://www.who.int/whr/1999/en, accessed 1 March 2009; World Bank, available at: http://www.worldbank.org/data/countrydata.html, accessed 15 January 2001; International Monetary Fund, available at: http://www.imf.org/external/pubs/ft/weo/, accessed 15 January 2001; Freedom House, available at: http://www.freedomhouse.org/template.cfm?page=439, accessed 12 March 2009; Election around the world, available at: http://www.Electionworld.org, accessed 12 October 2004; Agora—national election results, available at: http://www.agora.stm.it/elections, accessed 13 January 2001.

See Table 1 as well as the tables in Appendix 1.

Source available at: http://www.freedomhouse.org/template.cfm?page=439, accessed 12 March 2009.

For further discussion, see Miller (1997); Goati (1999, 2000).

From 1992 the Federal Republic of Yugoslavia only included Serbia (together with Vojvodina and Kosovo) and Montenegro. On 4 February 2003 it was renamed Serbia & Montenegro. On 3 June 2006 Montenegro declared its independence on the basis of the 21 May 2006 national referendum. Serbia declared its independence on 5 June 2006.

In particular, social indicator data are missing for Bosnia & Herzegovina, which was heavily involved in war. In the cases of Montenegro and Serbia only rarely were specific data available separately for each country.

Ward's method is distinct from other methods of hierarchical clustering in using an analysis of variance approach to evaluate distances between clusters: at each step it seeks to join clusters (objects) so as to minimise the increase in the Error Sum of Squares (see Ward 1963).

Since the QCA was developed (Ragin 1987), these principles and technical forms of presentation have been in use by various authors including Amenta and Poulsen (1994), Goertz and Mahoney (2005) and Rihoux and Ragin (2009).

A ‘truth table’ is a table of configurations (lines of the table) which present different combinations of conditions (independent variables) associated with a given outcome (the dependent variable) (Ragin 1987, pp. 87–89). In our case, the truth table would be a simplified version of Table 2—it would consist of six lines presenting six configurations which could be found from observing the 10 countries. The six lines (expressions) in Table 3a are in fact another form of presentation of a truth table.