Chapter 3. Quantitative Analyses
3.1 Operationalization and Measurement
國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Chapter 3. Quantitative Analyses
3.1 Operationalization and Measurement
The unit of analysis for this study is country-year. I collected data for 19 European countries from 1990 to 2018. For most observations, the beginning year is the year when the given country showed interest to become member or officially applied for membership, and the end year depends on when the country successfully obtained the membership or candidacy status.
For countries that never obtained the membership or candidacy status within the research scope of this thesis, the end year is 2018 (Bosnia and Herzegovina; Kosovo).
Croatia and North Macedonia are the exceptions for my coding rule in my dataset. Because their candidacy status was granted one year right after the year when they applied for
membership, which indicates that the criteria for getting candidacy was not demanding at that time, I consider the year when they successfully obtained the candidacy as their beginning year. Moreover, because North Macedonia has not obtained the EU membership within the research scope of this thesis, the end year for North Macedonia is 2018. Table 1 shows the observations included in my dataset.
‧
Table 1. Country-Year Observations in the Dataset
Country Time-span EU Membership Status at the End Year Region
Albania 2009-2014 Candidacy WB
Bulgaria 1995-2007 Member CEE
Bosnia and Herzegovina 2008-2018 Non-Member WB
Cyprus 1990-2004 Member MED
Czech Republic 1996-2004 Member CEE
Estonia 1995-2004 Member BL
Croatia 2004-2013 Member WB
Hungary 1994-2004 Member CEE
Lithuania 1995-2004 Member BL
Latvia 1995-2004 Member BL
North Macedonia 2004-2018 See notes below WB
Malta 1996-2004 Member MED
Montenegro 2008-2010 Candidacy WB
Poland 1994-2004 Member CEE
Romania 1995-2007 Member CEE
Serbia 2009-2012 Candidacy WB
Slovakia 1995-2004 Member CEE
Slovenia 1996-2004 Member CEE
Kosovo 2008-2018 Non-Member WB
Notes:
1. BL (Baltic), CEE (Central Eastern European), Mediterranean (MED) and Western Balkans (WB).
2. North Macedonia applied for EU membership in 2004 and obtained the candidacy status in 2005. Because the time for getting the candidacy is short, I consider the year 2004 as the beginning year for the empirical analyses.
In addition, because North Macedonia has not obtained the EU membership within the research scope of this thesis, I consider the year 2018 as the end year.
The dependent variable for the empirical analysis is EU accession. It is a dichotomous variable for indicating whether a countries obtains the candidacy for joining the EU or obtains the EU membership. It is coded 1 if a country obtained an EU membership or obtained a candidacy status for EU membership, and 0 otherwise.
The main independent variable for this study is Political Corruption (lagged by one year).
Considering various measurements of corruption, I chose political corruption based on various reports that emphasize that the Western Balkan countries are captured by corrupt politicians who are linked by organized crime, and this is documented by different
international reports that consider political corruption as a major concern among EU officials on the enlargement process (Ben-Meir 2019). For political corruption, I collected data
developed by Varieties of Democracy (V-Dem) Institute. The variable is measured by the
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
political corruption index, which entail questions such as: How pervasive is political corruption? This index includes six types of measurements of corruption, covering areas of levels of polity, focusing on different aspects such as executive, legislative and judicial corruption. On the executive aspect, the index focuses on bribery and corruption due to embezzlement. Furthermore, the index measures corruption on the highest level
(leaders/cabinet), and in the public sector. The index measures petty and grand corruption;
briery and theft (Coppedge et al. 2019).
Previous literature suggests that the level of economic development of countries is an important determinant for EU accessions (Katchanovski 2011; Mattli and Plümper 2002;
Schimmelfennig 2002). Therefore, I include a logged transformation of GDP per capita in USD to control for the probability that countries with higher levels of economic development are more likely to become members of EU. The data were gathered from the World
Development Indicators (World Bank 2020).
In addition, studies of EU enlargement suggest that the more democratic a candidate country, the more likely that it obtains an EU membership (Schimmelfennig 2002; Mattli and Plümper 2001; Selck and Deckarm 2013; Katchanovski 2011). Therefore, I include the level of
democracy to control the probability that more democratic countries tend to become members of the EU. This variable is measured by the Freedom House Index, a composite index of civil liberties and political rights constructed by Finkel, Pérez-Liñán, and Seligson (2007). The value of this variable ranges from 1 to 13, with a higher score indicating that country is more democratic. Thee data are from the Freedom of the World data from Freedom House (2020).
Moreover, I include a variable of the percentage of population that is Muslim (Pew Research Center 2017) as a control variable in the empirical models. As Katchanovski’s (2011)
contend, the European countries where the majority of the populations are Protestant and Catholic tend to have a higher chance to obtain EU membership, compared to countries whose population is predominantly Muslim or Orthodox Christian. In this sense, it is possible that countries with more Muslim population might be less likely to join the EU. Last, I
control for the number of population (in millions) for considering that country size might matter for the possibility of joining the EU.
‧
國立 政 治 大 學