Leiden University
Political Science – International Relations and Organizations (IBO)
Economic Inequality and the Quality of Democracy
A statistical analysis on Sub-Saharan Africa in the period
between 1990-2012
Name:
Mirthe Voppen
Student number:
s1550160
E-mail:
m.j.h.voppen@umail.leidenuniv.nl
Date:
18-06-18
Word count:
8326
Bachelorproject 6: Inequality in Political Perspective
Supervisor:
Dr. B.K.S. van Coppenolle
1 Introduction
Inequality is something of all times and all places and is visible in many different aspects of society. This is very clear when looking at the actual division of incomes across the world. UNICEF found that the top 20 percent of the world population enjoys more than 70 percent of the total world income (e.g. see Graph I). Graph I visualises the growing economic inequality as a trend around the world. Although it shows a minor drop in the 21st century, the average Gini coefficient is closer to 100 than it is to 0. This means that the division of wealth in the world is closer to complete inequality than it is to complete equality (Bourgougion, 2015). And although some progress is made in the last years, UNICEF estimates that it would take another 800 years for the bottom billion to achieve (only) 10 percent of the global income (UNICEF, 2011).
With economic inequality rising in many parts of the world, the question arises what effect this has on politics. Although income inequality seems an economic problem, this cannot be separated from politics. Acemoglu and Robinson (2000) found evidence for the effect that (economic) inequality has on the extension of franchise. It was due to a growing level of inequality among different groups in society that the elites feared their positions in society and the growing urge for the lower class to extend their rights, which might lead to a revolution. The authors found evidence that the inequality among citizens became smaller after the franchise was extended and a country would shift towards more democratic principles (Acemoglu & Robinson, 2000, p. 1167). Even though growing economic inequality might have led to extending the franchise and the development of a democracy, it is not clear that a decline in economic inequality will lead to the occurrence of a qualitative democracy. Elections are nowadays held in almost all countries around the world. It is however not clear whether all elections also contribute to the quality of a democracy. To test whether or not elections will contribute to the quality of a democracy, this research will focus on the presence of free and fair elections as a requirement of a qualitative democracy. Elections are considered free when the rules of an election and the process leading up to the election are free. Fairness of an election refers to the events that happen on the election day itself (Bishop & Hoeffler, 2016). The relevance of choosing free and fair elections as a measurement of a qualitative democracy will be elaborated later on in this thesis. In this research, the specific region Sub-Saharan Africa is chosen to be analysed. The choice to look specifically at this region comes from the fact that from the 1990s and onward, there has been a huge democratization wave among African states. Although the processes did
2 not always lead to fully developed qualitative democracies, there were major changes among political structures in many countries (Osaghae, 1999). However, within the academic debate, there is less research on those less advanced democratic countries. There is a gap in the literature on the effect of inequality on democracies in less advanced countries. This research aims to fill this gap by focussing on the relationship of economic inequality and the quality of democracy in less advanced democratic countries in the Sub-Saharan Africa region.
This research is organized in the following way. First, an overview of existing literature on this research topic will be discussed. This section also provides a conceptualisation of different concepts and the legitimization of the case selection is elaborated. In the third part, the methodological analysis is presented and this is followed by the operationalization of different concepts. In the fifth part the results of the analysis are presented. A discussion/conclusion about the findings is presented in the end.
3 Literature Review
Since the third wave of democratization, many undemocratic countries made the transition towards democracies. Around the 1990s, many African countries made an attempt to become more democratic (Osaghae, 1999). Although democratization often seems like a predetermined path for the entire world, not all countries develop the same way. In order to understand the path countries take, it is interesting to ask the question why democracy ever occurred. There are multiple theories that explain why and how democracy occurred as a reaction to inequality. Melzer and Richard (1981) state that because of elites, who generally have much of the political and economic power, there is an unfair distribution of money and wealth. More inequality means that the poor are poorer in relation to their fellow citizens. For them, redistributive policies should be more attractive to have a more equal distribution among all citizens. The elite fear this median (poorer) voter when inequality is high because the median voter will demand for tax redistribution which will affect the rich elites. For those elites, redistributive policies will be more costly due to the fact that the richer ones will have to pay more in relation to the poorer ones (Melzer & Richard, 1981). A contradictory theory is proposed by Ansell and Samuels (2010), they found that a regime change towards a democracy is not due to the elite’s fear of the poor who will raise taxes, but that is the politically disfranchised but rising economic groups who want to have a say in what happens with their income and assets (p. 1544). These theories however provide us with information about the process of democratization. This process focusses on the transition of non-democracies turning into democracies. Democratization can be seen as a positive process for a country to achieve a good working democracy. However, the completion of becoming a democracy raises some questions. It must be stated that not all countries that transition towards a democratic constitution will have (completely) free and/or fair elections as an outcome of their transition. This research mainly focusses on the quality of existing democracies rather than on countries that are transitioning towards a democracy. Often, countries are considered democratic just because there are official elections. Unfortunately, even when there are elections, this does not necessarily mean that the country’s elections are free and/or fair. It is a shortcoming to just look at the question whether or not a country is a democracy. Therefore, it is key to assess
whether the country holds free and fair elections.
Many scholars define and operationalize the concept of democracy. One definition that is widely accepted in contemporary research is Lipset’s (1959) definition. He defines
4 democracy as “a political system which supplies regular constitutional opportunities for changing the governing officials, and a social mechanism which permits the largest possible part of the population to influence major decisions by choosing among contenders for political office” (Lipset, 1959 p.71). Another definition is given by Larry Diamond and Morlino (2005), he states that democracy has at least four requirements, 1) universal adult suffrage; 2) recurring, free, competitive and fair elections; 3) more than one serious party and 4) alternative sources of information. Some degree of civil and political freedom is necessary in
a system like this.
Diamond and Morlino (2005) also developed a definition of a good democracy. According to him, whether a democracy is indeed a good democracy depends on several requirements. Firstly, political and civil freedoms, secondly, popular sovereignty and thirdly, political equality. Vanhanen (2000) defined democracy as “a political system in which ideologically and socially different groups compete for political power, and in which institutional power-holders are elected by the people and responsible to the people” (p. 252). However, there is a lot of debate on how to measure the quality of a democracy. The quality of a democracy in linked the level of a democracy. The higher the quality of a democracy, the higher the level of a democracy is as well. But, these concepts should also be measured. Larry Diamond and Morlino (2005) characterises a quality democracy as one that “provides its citizens a high degree of freedom, political equality and popular control over public policies and policy makers through the legitimate and lawful functioning of stable institutions” (p. 11). Again, in this definition, control over public policies and policy makers through institutions links to the presence of some sort of election. A common character in all the different definitions is the role of the people to elect others through elections. The aim of this research is not to point out a perfect definition of what a democracy is or what it should consist of. Finding a perfect definition is not relevant for this research. It is however very interesting to see that all different definitions mentioned above include a common factor. The factors of representation and the ability to elect your representative are considered very important by all mentioned scholars. One indicator for measuring a democracy that will be used in this research is the presence of elections in countries. Elections are used as an indicator for the quality of a democracy and to measure whether or not a democracy can be described as qualitative or not. The question on how to measure democracies is debated by many scholars. Freedom House is an institution that created a measurement for democracies. This institute rates countries on the amount of political rights and civil liberties an individual has. Because the data of this institution was not
5 easily accessible, their data for measuring democracies is not used in this research (Freedom House, 2018). Another measurement to measure democracies is by looking at if countries have ‘free’ and ‘fair’ elections. ‘Freeness’ then refers to the rules of the elections and the process leading up to the elections, whereas ‘fairness’ refers to the events on the day of the election (Bishop & Hoeffler, 2016). The dataset that is provided by Bishop and Hoeffler (2016) will be used in this research to analyse if there is a relationship between the quality of a democracy and economic inequality. Their analysis uses ten different variables to measure whether or not an election, in a democratic country, can be evaluated as free and/or fair.
There are also different visions on how to define the concept of inequality. Blau (1977) defines the concept as follows: “the average difference in status between any two pairs relative to the average status”. Allison (1978) however introduces a formula in which complete equality, when all citizens have the same income, is zero and when there is no equality among citizens, the value can go to a maximum of 100. Often scholars choose measurements of inequality based on convenience and familiarity, or based on other “vague
methodological grounds” (Allison, 1978, p.865).
However, a commonly used concept to test whether or not there is (economic) inequality among people is by using the Gini coefficient. The World Bank (2018a) is one of the largest sources of information for this concept. Due to availability of the information of the Gini and the comprehensive explanation of this concept, Gini is chosen to be the best indicator for inequality in this research. Because the Gini provides a scale of 0-100, this helps the research with understanding different outcomes easier. The higher the number of Gini is, the higher the inequality of a country in a certain time is. Although the number itself might seem arbitrary, the indication it gives on a scale from 0-100 and in comparison with other
countries is very clear.
An alternative factor that might be considered to explain a possible relation between inequality and the quality of a democracy is a country’s GDP. The GDP, Gross Domestic Product, of a country explains a country’s wellbeing or wealth. GDP can be considered to be a factor that influences the attendance of free and fair elections in countries. Bishop and Hoeffler (2016) already find evidence that a higher GDP per capita income country is more
likely to have free and fair elections (p. 614).
A country’s economic wellbeing is hardly comparable to economic wellbeing in another country. For example, when looking at Tanzania and Kenya the GDP is comparable. Kenya is ranked 68th on the world ranking of GDPs with a GDP of 70.529 billion US dollar a
6 year, Tanzania is ranked 81st place respectively with a GDP of 47.340 billion US dollar (World Bank, 2018b). Although these values initially seem quite far apart, the yearly economic growth of both countries is relatively similar, approximately between 6%-7% in 2016 (African Economic Outlook, 2017a-b). When looking at the economic inequality rate, defined by the GINI coefficient, it seems that Kenya and Tanzania are actually ranked far apart from each other and Tanzania is stated to be more economically equal than Kenya (World Bank, 2018b). In this research, a country’s GDP is used as a control variable.
This research focusses on the effect of economic inequality on the quality of a democracy. A natural question that rises from this research question is which of the two factors (inequality or the quality of a democracy) came first. The origin of this debate lies within the question on why democracy initially occurred in countries. One argument on why democracy occurred in societies is given by Acemoglu and Robinson (2000). They find that extending the franchise to common people, as a form of democratization, had several reasons and effects. First of all, there was social unrest in many societies where elites were in power and were very wealthy. The common, mostly poorer citizens, felt left behind and wanted more distribution of wealth among all people, which meant that the elites would probably end up paying more taxes than they did before. This finding contributes explains that first there was inequality before a (qualitative) democracy occurred in countries. As a reaction to the inequality, democracies expended their franchise in order to represent more people and enjoy a higher quality of democracy.
Extending the franchise to a larger group of people had several reasons. The first reason was that in the enlightment view, new social values were developed by a larger group of people. The elites thought it was the right choice to give common people more influence. The second reason for extending the franchise might come from the political competition among different elites. If more people had the opportunity to vote, more competition among elites would be necessary to gain those votes. For the third reason it was argued that the middle class was a strong driving force of this extending franchise. The middle class was hoping to shift the future balance of power (Acemoglu & Robinson, 2000, p.1169). The interesting finding of this research however is that there was a drop of the inequality rate (Gini), right after the franchise was extended to a larger group of people in Great Britain in the 19th century. The reason for this was that more people voted for a redistribution of income
towards the poorer segment of society (Acemoglu & Robinson, 2000, p.1167). Inequality is both an economic and political issue. McCarty, Pontussen and Jonas
7 (2008) believe that proper democratic institutions have an equalizing effect on citizens’ wealth. According to them, in proper democratic institutions participation from citizens of different parts of society insures an inclusive distribution of wealth. Another debate on the relation between democratic quality and income inequality is raised by Muller (1988). He tests multiple hypotheses that link the quality of a democracy to income inequality. He finds that a high level of democracy is expected to produce a more egalitarian distribution of income. In this finding there should be no difference between countries which are ‘young’ democracies and ‘old’ democracies. (p.65). Another one of his hypotheses is that democratic institutions ensure a more egalitarian distribution of income among citizens. This finding is extremely relevant for this research because Muller finds that the relationship is not just from democratic institutions towards better distribution of income, the relationship appears also to be the other way around. Egalitarian political institutions (equal distribution of income) provide citizens with the opportunity to participate in the governing process and to engage in free and fair elections (Muller, 1988). This would eventually lead to a more equal distribution of wealth among citizens because the major part of the voters would vote for what is best for them (Muller, 1988, p.65). A reaction to Muller (1988) his statement on the effect of equal distribution of income on the participation democracy as mentioned above, is suggested by himself several years later (Muller 1995a). He then states that “the mortality rate of a democracy, given high income inequality, was 80%; whereas it was 4 percent, given low income inequality” (p. 990). This means that there is a relation between economic inequality and the quality of a democracy to endure and keep a certain quality while the democracy is lasting. This is, according to Muller, a causal relationship he finds with a logistic regression analysis. Bollen and Jackman (1985) however say that political democracy and economic inequality do not seem linked in any meaningful way. The relationship between factors, inequality and democracy are neither way linked according to Bollen and Jackman (1985). Many of the articles that suggest a positive effect of democratic institutions on equality use data from advanced democracies (McCarty & Pontussen, 2008, p.666). Less research is found on less advanced democratic countries. With focusing on a group of less advanced democracies in a certain region, this research tries to contribute in the academic debate on the relation between democracy and inequality in less advanced democracies. The focus of many scholars solely lies on how democratic qualities impact economic inequality. There are however few scholars who link the two variables the other way around (how economic inequality impacts democratic quality). An example of such research comes from
8 Houle (2009). He finds evidence for the statement that inequality does not harm the democratization process, but does harm the consolidation of a democracy. It might be that because of the two possible effects that inequality has on democracy. The two effects of inequality on democratization come from the fact that democracy is demanded by the population but, in the end, is conceded by the elite (p.597). First, inequality might decrease the willingness from the elite to democratize and second, inequality might increase the incentive of the population to contest the regime (p.597). The other factor that is suggested by Houle, that inequality harms consolidation of a democracy, is more interesting for this particular research. The article states that the relationship between inequality and democracy equals the relationship between wealth and democracy: inequality harms consolidation but has no net effect on democratization (p.590). The argument he uses on why inequality harms consolidation is that elites are more likely to stage coups in unequal democracies because these are believed to redistribute more and making the democracy more costly for the elites themselves (p.595). A counterargument is suggested by Acemoglu and Robinson (2000), who state that the elites possibly extended the franchise in order to avoid revolution or social unrest occurred by the people who wanted more distribution of wealth (p.1168). This argument then does find evidence that inequality affects the democratization process, as an
effect of extending the franchise.
Because of the limited data that is available on the impact of inequality on the quality of a democracy, this research aims to answer the question “What effect does economic inequality have on the quality of democracies in Sub-Saharan Africa -in the period between 1990-2012”.
The choice to focus on a specific region of the world is based on different aspects. First of all, this region, consisting of 48 countries, is chosen because of its interesting political developments since the 1980s. The wave of democratization spread great hopes for democracies after a long undemocratic past (Osaghae, 1999). Secondly, it is interesting to compare a certain region (Sub-Sahara Africa) with the rest of the world. When using a global data set, as Bishop and Hoeffler (2016) did by looking at 165 countries (see Table I), it is possible that this gives a distorted picture on the actual distribution of free and fair elections in the world. They find that from the 890 observed elections, most of them (469) are both free and fair, whereas only 212 are neither free nor fair. Almost half (50%) of the elections were free and fair and around 25% of them were not free and not fair. In this same dataset, when looking at the numbers of free and fair elections in the
Sub-9 Saharan region in the period of 1990-2012, the number of both free and fair elections is much lower. From the 152 observations in that time and region, only 37 elections are found both free and fair (see Table II). As a result, almost a quarter of the observations are coded as free and fair (25%). When looking at the amount of both unfree and unfair election, a number of 71 elections are found in the observations. This results in about 50% of the 152 included elections in Sub-Sahara Africa in 1990-2011.
Table I. Overview free and fair elections worldwide (1975-2012)
Unfair Fair Total
Unfree 212 169 381
Free 40 469 509
Total 252 638 890
Table II. Overview free and fair elections Sub-Saharan Africa (1990-2012)
Unfair Fair Total
Unfree 71 40 111
Free 4 37 41
Total 74 77 152
Contradictory to the overall global numbers that Bishop and Hoeffler (2016) provided for all their observations in the dataset, the percentages on free and fair elections differ a lot in the African region in this specific timeframe. This research will not search for evidence why the percentages of free and fair elections in Sub-Sahara Africa differ so much from the overall world percentages. It is however very interesting to know how it is possible that the amount of free and fair elections in this region is much lower than in general over the world. Further research on the reasons of this difference would be very interesting and helpful. Due to the different percentages and findings in numbers of free and fair elections, this research only focusses on the Sub-Saharan countries in the period of 1990-2012 to see if there is a link between the economic inequality of a country and its quality of the democracy. The timeframe of 1990-2012 is chosen for this research because the African continent, and specifically Sub-Saharan Africa, made an attempt to become democratic with elections, starting in the 1990s. Muller (1995b) provides a theory on the effect that economic inequality has on the
10 level of democracy. Muller explains that there is a relation between income inequality and economic development, stating that less inequality increases economic developments. He finds that the relationship between income inequality and the quality of the democracy is a negative one, meaning that the higher the income inequality, the less qualitative the democracy will turn out. This finding is very helpful in this research. His theory, saying that high economic inequality has a negative effect on the quality of a democracy will be used to analyse the research question of this research. The aim is to use this theory to find evidence for the significant relationship that the degree of economic inequality has on the quality of a democracy. Whereas Muller (1995b) analyses the effect of income inequality on the quality of a democracy in liberal democracies, this research focuses on democracies that are in the Sub-Saharan region. Another difference between this research and Muller’s research is that he links the independent variables to different scores of democracies. This research focusses on the presence of free and fair elections as an indicator of a qualitative democracy. His approach to test the relationship between the level of economic development and the level of democracy lies within a cross-national analysis with middle income countries. He states that intermediate levels of economic development are associated with high levels of income inequality. After a statistical analysis in 58 middle income countries, he finds support for his hypothesis that income inequality has a negative effect on the level of democracy. Muller’s theory (1995b) however uses a dichotomous approach that a country either is a democracy or is not a democracy. This research does not aim to not answer the question whether a country is a democracy or not, but looks at the quality of a democracy on the basis of elections. This demarcate is chosen because testing the quality of a democracy is easy to operationalize when there are certain concepts to check. Concepts such as free and fair elections can be tested by certain requirements set by Bishop and Hoeffler (2016). Because of this, the degree of economic inequality is chosen to be linked to the presence of free and fair elections as a measurement of a qualitative democracy. Taking into considerations the arguments of all of the authors mentioned before there is one hypothesis that will be tested in this research: economic inequality has a strong negative effect on the level (quality) of a democracy (Muller, 1995a, p.991). There are different possible outcomes of this hypothesis that will lead to a different conclusion. The hypothesis will be tested with statistical regression analyses between the degree of economic inequality and the quality of a democracy. In a logistic regression analysis, hypotheses can be tested.
11 Methodology
The aim of this research is to provide evidence for the relationship between the degree of economic inequality and the quality of a democracy. By combining data from the Gini index on income inequality and data about free and fair elections from the dataset by Bishop and Hoeffler (2016) we will look for a significant relationship between both factors. With a large-n alarge-nalyses there will be differelarge-nt variables. The ilarge-ndepelarge-ndelarge-nt variable is the degree of economic inequality and the dependent variable is the level of democracy. When one takes all the elections from Sub-Saharan African countries from 1990-2012, there are 186 cases to select (Bishop & Hoeffler, 2016). Unfortunately, not all data that was provided by Bishop and Hoeffler was useful because some data were not available or coded as unsure. For several elections in unstable countries, for example Somalia, not all observations could provide all information necessary due to the unsafe developments during that election or due to the lack of information provided by governmental organs (Bishop & Hoeffler, 2016).
For the data on the Gini coefficient, two sources were used. The World Bank (2018a) and the African Development Bank (2018) with socio-economic information on Africa provided almost all the information needed. For some specific election years there was not always enough data available. Therefore, for some cases, the average of two years that lied in front and beyond the certain year, provided information for the specific year. Because the African Development Bank (2018) provided a graph on the development of a country’s Gini coefficient, it was possible to check whether or not the average was a possible outcome. For example, for the Central African Republic data on the Gini was only available for 1993 and 2005 although there was an election in 1999. Because 1999 lies perfectly in the middle between 1993 and 2005, the average of the given data on those countries is used to guess the value of the 1999 Gini coefficient. The cases where it was not possible to take the average of other years were taken out of the analysis. Due to some lack of information on the Gini coefficient, there were a few cases that had to be dismissed in the analyses. Eventually, there are 152 cases where all necessary data that is available to do an analyses. The 152 cases represent 44 countries of the Sub-Sahara Africa region. This still results in a large-n analysis. It is the aim to find enough variation among the data from different countries in the timeframe. An overview of a country’s Gini and GDP (World Bank, 2018a-b) in the year of the election is given in Table III and IV. These tables show the amount of variation there is between countries during timeframes of approximately 10 years. As one can see in the tables,
12 the categorization of a country’s Gini and GDP shows the variation there is among the different cases. Besides, the tables perfectly show that a higher GDP does not always means less inequality, e.g. South Africa. This country has a high GDP (over 25 billion USD a year) but also a very high degree of economic inequality (around 60).
Table III. Descriptive statistics: countries categorized by Gini Gini Country (1990-2000) Country (2001-2012) Gini between
0-15
Gini between 16-30
Ethiopia, Sao Tome en Principe (2)
Gini between 31-45
Burundi, Cameroon, Cote d’Ivoire, Djibouti, Ethiopia, Ghana,
Madagascar, Mauritania, Niger, Senegal,
Tanzania, Uganda (12)
Benin, Burkina Faso, Burundi,
Cameroon, Chad, Democratic Republic of the Congo, Cote d’Ivoire, Djibouti, Equatorial Guinea, Gabon, Gambia, Ghana, Guinea, Liberia, Madagascar, Malawi, Mali, Mauritania, Mauritius, Niger, Nigeria, Senegal, Sierra Leone, Sudan, Tanzania, Togo, Uganda, Zimbabwe (28)
Gini between 46-60
Burkina Faso, Central African Republic, Gambia, Guinea, Guinea Bissau, Kenya, Mali, Mozambique, Nigeria, South Africa, Zambia (11)
Cape Verde, Central African Republic, Comoros, Republic Congo, Guinea Bissau, Kenya, Lesotho, Mozambique, Rwanda, Seychelles, Zambia (11)
Gini between 61-75
Botswana, Equatorial Guinea, Lesotho, Malawi (4)
Botswana, Namibia, South Africa (3)
Gini between 76-90
Gini between 91-100
Total 27 (not all countries have elections 1990-2000)
13 Table IV: Descriptive statistics: countries categorized by GDP
GDP Country (1990-2012) Country (2001-2012)
GDP < 1 billion Burundi, Djibouti, Equatorial Guinea, Gambia, Guinea Bissau, Lesotho (6)
Cape Verde, Comoros, Djibouti, Gambia, Guinea Bissau, Liberia, Sao Tome en Principe (8)
GDP between 1 – 5 billion Botswana, Burkina Faso, Central African Republic, Guinea, Madagascar, Malawi, Mali, Mauritania, Mozambique, Niger, Tanzania, Uganda, Zambia (14)
Burundi, Central African Republic, Lesotho, Madagascar, Malawi, Mauritania, Niger, Rwanda, Seychelles, Sierra Leone, Togo, Zimbabwe (12)
GDP between 5.1 – 10 billion
Republic Congo, Ethiopia, Ghana, Kenya, Senegal (5)
Benin, Botswana, Burkina Faso, Chad, Gabon, Guinea, Mali, Mauritius,
Mozambique, Namibia, Zambia (11)
GDP between 10.1 – 15 billion
Cameroon, Cote d’Ivoire (2)
Equatorial Guinea, Senegal, Uganda (3)
GDP between 15.1 – 20 billion
Democratic Republic of the Congo, Cote d’Ivoire, Ethiopia, Ghana (4) GDP between 20.1 – 25
billion
Cameroon, Kenya, Tanzania (3)
GDP > 25 billion Nigeria, South Africa (2) Nigeria, South Africa, Sudan (3)
Total 29 (not all countries have
elections in 1990-2001)
44
To uncover a relationship between the two variables Gini and quality of a democracy, an empirical analysis must be done. With a logit regression analysis, the research aims to find a significant relation between the degree of economic inequality and the quality of a democracy. To increase the complexity by searching for causal relations, the adding of control variables is necessary. Control variables are necessary to include in this research to check if the possible relationship between economic inequality and the presence of free and/or fair elections is significant or whether other independent variables have an effect on the presence of free and/or fair elections as well. The control variables in this research are; 1) Natural
14 resources rents, 2) Aid, 3) Trade, 4) the presence of an observer, 5) presence of executive constraints, 6) presence of a constitutional term limit and 7) a country’s GDP (See appendix I for further clarification). These variables are chosen due to their relevance in the research by Bishop and Hoeffler (2016). They test a hypothesis focused on the effect resources have on democratization. They find evidence for the claim that countries with a high resource dependency are less likely to experience free and fair elections (Bishop & Hoeffler, 2016, p.614). The seventh variable GDP was not included in the research of Bishop and Hoeffler, but is added to give a broad overview of a country’s wealth and the possible relation it has to the quality of a democracy. Bishop and Hoeffler believe that countries with higher per capita income are more likely to have free and fair elections. In this research however, the seven control variables are used to give a broad outcome of the analysis. To include natural resources, economic factors and political factors, the analysis will be comprehensive in answering to the question what effect economic inequality has on the quality of a democracy and maybe finding other relationships among factors. Besides the substantive argument of using these specific control variables, there is also a practical argument. Due to availability of data and providing certain guidelines for the statistical research, this research uses the data that are provided by Bishop and Hoeffler (2016). The use of control variables will be further explained by evidence that is discussed in the results section.
There are two control variables that Bishop and Hoeffler (2016) use that will not be used in this research. First the presence of a presidential election and second the 4th wave of democratization. These are left out in the analysis because (almost) all cases within this analysis have presidential elections and are placed within the fourth wave of democratization. There would be no variation among these variables, which is why those two are not used in
the analysis.
To test whether or not there is a significant relationship between the degree of economic inequality (independent variable X) and the quality of a democracy (dependent variable Y), a logistic regression is used to test the hypothesis economic inequality has a strong negative effect on change in the level of democracy (Muller, 1995a, p.991).
A regression analysis investigates the relationship between two or more independent variables and a single dependent variable and will fit the aim of this research; to test what effect economic inequality (IV) has on the quality of a democracy (DV) (Argyrous, 2015, p.258). In doing a multivariate regression, the amount of variation that is left over to be explained by the error term will be much smaller than in a bivariate model, because there are just more
15 variables that might explain a relationship. In this research, the multivariate logistic regression analysis tries to explain the attendance of free and fair elections in countries based on different independent variables. For this research, the independent variable, economic inequality, is valued by the Gini-coefficient. This number between 0-100 shows the degree of economic inequality within a country. This is a numeric variable with a clear scale measure. By doing a logistic regression, it is important to understand that the dependent variable (Y) is a dichotomous variable. The dependent variable, the quality of a democracy, is transformed into a dichotomous variable. This means the value can be either 0 or 1. Because there are more than two options for an election to be free of fair, (both free and fair, only free not fair, only fair not fair), we split the analysis into three parts. The first model in the analysis will be between 1: both free and fair elections and 0: neither free nor fair elections. The second analysis will check the relationship between 1: free elections and 0: all that are non-free elections (this analysis is not measuring whether elections are fair). The third analyses will check the relationship between 1: fair elections and 0: all that are non-fair elections (this analysis is not measuring whether elections are free).
The logistic curve that predicts the change of X on Y is shaped in an S-curve. This shows that the probability that Y will happen, will not go either under 0 or above 1 because the change that X has effect on Y can never be smaller than 0 and higher than 1. There is no easy interpretation of the effect of X on the probabilities. To interpret the logistic regression, it is easier to interpret it by the odds. The odds Ratio is the change between two odds with a change of one. The Odds Ratio is always the same when X is changing. B is the change in the log of odds. B shows how the logit changes when X increases by 1. By transforming the probability (0 to 1) to odds, the range is transformed from 0 to plus infinity. To transform the range from minus infinity to plus infinity, the odds have to be transformed to natural logarithms. The transformation is called logit.
Operationalizing
To operationalize the concept of democracy, this research uses the measurement of ‘free’ and ‘fair’ elections. Competition and participation are concepts that are also included in the definitions of democracy (Vanhanen, 2000 & Diamond and Morlino, 2005). ‘Freeness’ means that all adult citizens must have the right to register and vote as well as have the right to establish and join parties and campaign freely within the country (participation) (Bishop &
16 Hoeffler, 2016). Fairness’ means that there must be equal treatment for all equals (Bishop & Hoeffler, 2016, p.609). The authors have created 10 variables to measure whether elections are free and fair. The criteria that determine whether or not an election is considered free/fair or free and fair are 1) legal framework, 2) electoral management bodies, 3) electoral rights, 4) voter register, 5) ballot access, 6) campaign process, 7) media access, 8) voting process, 9) role of officials and 10) counting of votes. If an election would fulfil the criteria, it will be coded 1 and if the election does not fulfil the criteria, it will be coded 0. The precise criteria that are linked to the 10 variables can be found in the appendix II, this information is additional information that Bishop and Hoeffler provided with their data analysis from 2016. The database that is created by Bishop and Hoeffler (2016) is based on 1114 elections in a total of 169 countries in the years 1975-2012. For each case concerning this research, this thesis coded 152 cases as free and fair, free or fair by checking whether or not the 10 criteria
would match the specific election.
To measure the degree of economic inequality within countries, this research will use the Gini index as indicator for measuring economic inequality within countries in certain years. In order to answer the research question ‘What effect does economic inequality have on the quality of democracies in Sub-Saharan Africa -in the period between 1990-2012? The Gini coefficient is a necessary source to test this relation... The Gini will hereby serve as an independent variable in the analysis. Data from the Gini index is based on primary household survey data, obtained from government statistical agencies and World Bank country departments (World Bank, 2018). The data on economic inequality will be linked to the data on ‘free’ and ‘fair’ elections to hopefully see if there is a significant relation between the degree of economic inequality within a country and the level of democracy. The terms economic and income inequality are used interchangeably in this research. “Gini index measures the extent to which the distribution of income among individuals or households within an economy deviates from a perfectly equal distribution … The Gini index measures the area between the Lorenz curve and a hypothetical line of absolute equality, expressed as a percentage of the maximum area under the line. Thus a Gini index of 0 represents perfect equality, while an index of 100 implies perfect inequality.” (Index Mundi, 2018). With using this definition of the Gini, the research will look for a possible relationship between this data on economic inequality and the quality of a democracy. The Gini coefficient is a very clear measurement of the inequality rates. A number between 0-100 easily shows the scale where a country is placed on the basis of its inequality.
17 Results
The data that is provided below will give the necessary information to test the hypothesis: ‘economic inequality has a strong negative effect on the quality of a democracy’. Because the scale of the Gini goes from 0-100 and 0 means more equal than 100, the term negative might be a bit inconvenient, in essence the hypothesis means that when the value of the Gini gets higher, closer to 100 (more unequal), it will effect the quality of a democracy negatively. When primarily looking at the relationship between the degree of inequality, measured by the Gini coefficient as an independent variable and the quality of a democracy, measured by the attendance of free and fair elections as dependent variable, it shows that there is indeed a significant relationship amongst those two (Table V). The degree of inequality in a country does affect the quality of a democracy when only looking at the Gini as explanatory variable. The odds ratio of the free and fair elections model is 2,720.050 = 1,0513. This means that
when the Gini of a country grows with 1 unit (this means that the Gini gets a lower value, getting closer towards 0 on the 0-100 scale, i.e. becoming more equal), the odds that an election will be free and fair is approximately 5% higher than it was before.
Table V. Logistic regression analysis economic inequality and quality of a democracy Model 1 Model 2 Model 3
Free and Fair Free Fair (Constant) -3.441 -1.866 -1.140 (1.067)*** (3.013) (1.024) Gini coefficient 0.050 -0.039 0,002 (0.022)** (0.069) (0.022)
-2L 163,698 35,586 175,194 Cox and Shell’s R^2 0.032 0.002 0.00 Nagelkerke R^2 0.048 0.011 0.00 N 152 152 152
Note: binary regression coefficient with standard errors in brackets. *** p < 0.01, ** p<0.05, * p<0.1
To test whether or not this relationship is a spurious relationship, it is helpful to add control variables in the analyses. Control variables control the given relationship with other
18
independent (explanatory) variables.
When adding different control variables, it is clearly that the relationship between different variables differed (see Table VI). There are also dummy variables added in the analysis to see whether or not a specific group of years might influence the presence of a free and/or fair election. The dummy groups are: 1. Years 1990-1995, 2. Years 1996-2001, 3. Years 2002-2007 and the fourth group will be the reference group. The reference dummy group (years 2008-2012) is left out the analysis, this means that the variables Y19901995, Y19962001 and Y20022007 will be compared to the elections in the years 2008-2012.
The first analysis (Model 1) between multiple independent variables, including the Gini coefficient, and the dependent variable both free and fair elections resulted in table VI. The regression analysis, analyses the relation between income inequality and both free and fair elections, the results show that there is a positive relation between the Gini and the presence of a free and fair elections. The odds ratio that if the Gini would gain 1 unit on the 0-100 scale (a country becomes more equal, the value gets closer to 0), the presence of a free and fair election would grow is 2,720,033 = 1,033 = 3,3%. Although this seems like a positive
relationship, there is no significance outcome found for the Gini coefficient in Model 1. Although there is also no significance for the Gini in comparison to free elections (Model 2), it is interesting to see that the outcome shows a negative value. The odds ratio (2,72−0,101 = 0,903. This shows that if the Gini would drop with 1 unit on the scale from
0-100 (a country becomes more equal), the odds of a free election would decrease by
approximately ten percent.
The presence of executive constraints in Model 1 do show a significant relationship with the presence of free and fair elections (p<0,01). This means that the presence of an executive constraint has a significant effect on the presence of free and fair elections in a country. It is interesting to see that in both Model 2 and 3, none of the results are found significant to the presence of free and fair elections, not even the presence of executive constraints which had a very strong (p<0,01) relationship on the free and fair elections (Model 1) but it has no significant relation on the presence of a free election. When looking at the different timeframes of the elections, incorporated by dummy variables, it is interesting to see that in all the models, the timeframes before 2008 are compared to the cases after 2008. Samples in the years 1990-1995 (Y19901995) score lower in comparison to the reference group which means that it is less likely free and fair elections would happen in the 1996-2001 timeframe. In Model 1, the presence of free and fair elections
19 in 1996-2001 is scored lower than in the reference group. The logistic regression analysis gave the opportunity to test the hypothesis on a strong negative effect between economic inequality and the quality of a democracy. Unfortunately, there is no significant relationship found between the Gini and all three possible elections. Therefore the hypothesis that was tested in this research to support the research question is rejected.
20 Table VI. Logistic regression analysis economic inequality and quality of a democracy (control variables included)
Model 1 Model 2 Model 3 Free and fair Free Fair
(constant) -4,863 -13,676 -0,306 (1,674)** (4562,483) (1,264) Gini 0.033 -0,101 0,008 (0.030) (0,084) (0,026) GDP -0,003 -0,307 0.001 (0,005) (0,287) (0.005) Natural resources rents -0.035 -0,130 -0.021 (% of GDP) (0.027) (0,175) (0.017) Aid 0.011 0,100 0.007 (% of GDP) (0.016) (0,095) (0.015) Trade 0.004 0,038 0.000 (% of GDP) (0.004) (0,042) (0.005) Observer present -0.225 -1,359 0.149 (0/1) (0.464) (1,392) (0.402) Executive constraints 1.627 -0,114 -0.323 (0/1) (0.502)*** (1,426) (0.423) Constitutional term limit 1.269 -0,353 -0.694 (0/1) (0.725)* (1,4581) (0.473) Y19901995 -0,199 19,451 -0,614 (0,760) (4562,479) (0,634) Y19962001 0,358 0,717 -0,410 (0,623) (4562,479) (0,543) Y20022007 0,438 18,721 -0,238 (0,604) (4562,479) (0,505) -2L 137,281 20,746 170,192 Cox and Shell’s 𝑅2 0,187 0,102 0,032
Nagelkerke 𝑅2 0,279 0,469 0,047
N 152 152 152
Note: binary regression coefficient with standard errors in brackets. *** p < 0.01, ** p<0.05, * p<0.1
21 Conclusion
This research focusses on the relation between economic inequality and the quality of a democracy. In order to answer the research question: “what effect does economic inequality have on the quality of democracies in Sub-Saharan Africa -in the period between 1990-2012”, a literature review with research on different aspects on this topic, is presented. After considering different literature on the relationship between (economic) inequality and the quality of a democracy, the hypothesis ‘economic inequality has a strong negative effect on the level (quality) of a democracy’ (Muller, 1995a, p.991) was chosen to be tested in regard to help find an answer to the main question of this research. Many countries considered themselves a democracy, but one clear definition of what a democracy precisely is and should consist of, often remains absent. Interesting is that in many different definitions that are used to define ‘democracy’, a returning factor is the presence of participation and competition. In most of the definitions, participation from people and competition between different candidates / parties is mentioned as a requirement for having a qualitative democracy. Although many countries describe themselves as a democratic country, not all countries would be considered as a qualitative democracy. Measuring this quality of a democracy is done by using the presence of free and/or fair elections. Free elections means that all adult citizens must have the right to register and vote as well as have to opportunity to vote for something they want, fair elections refer to the equal treatment of equals on the election day. This definition includes the factors of participation and
competition as well.
Another important aspect of this research is the measuring of economic inequality. The Gini coefficient provides a proper indication of a country’s distribution of wealth among its citizens. The inequality rate is used as an explanatory factor to the presence of free and/or fair elections and is controlled by different other variables. The logistic regression analysed different independent variables to its effect on free and/or fair elections and found no significant evidence that there is indeed a strong negative effect of economic inequality on the level of a democracy. However, the outcome of the analysis did show that for both free and fair elections and for fair elections, the effect of a more equal distribution of wealth (a lower Gini), did have a positive effect on the presence of the elections. However, for the presence of free elections, the Gini coefficient has a negative effect on the presence of free elections, meaning that a lower Gini coefficient (closer to 0, more equal) actually has a negative effect on the presence of fair elections. It must be stated
22 that for all of the three models, no significant evidence is found. What effect does economic inequality have on the quality of a democracy in Sub-Saharan Africa in the period 1990-2012? It seems that the effect of a county’s Gini rate does not have much effect on the presence of free and/or fair elections. Due to the absence of significant evidence in the data analysis, the effect of a country’s Gini on the quality of its democracy cannot be properly concluded in this research. Looking at the numbers does give an insight in what direction the relation of inequality and free and/or fair elections goes. Although the hypothesis is rejected due to the lack of significance, the relationship mentioned by Muller (negative effect of inequality on quality of a democracy) does seem valid when interpretation the details. When the Gini value dropped (gets closer to 0) the presence of free
and fair and fair elections will be more likely.
This outcome does however raises a broader question on what effect inequality might have on other aspects of politics. Is it possible that inequality might affect the outcome of an election in politics, rather than affecting the election itself? Further, the outcomes of this research could be used in different ways. This research can be the starting point of further research. Further research could focus more specifically focus on how the Sub-Saharan African region could have such different percentages of free and/of fair elections, compared to the world wide statistics. It is interesting to look at more specific reasoning for this, and to compare the presence of free and fair elections in different regions of the world with each other.
23 Literature
Acemoglu, D & Robinson, J. (2000). Why did the west extend the franchise? Democracy, inequality, and growth in historical perspective. The Quarterly Journal of Economics
(November) 1167-1199.
African Economic Outlook. (2017a). Tanzania. Retrieved at
http://www.africaneconomicoutlook.org/en/country-notes/tanzania on 25-03-18
African Economic Outlook. (2017b). Kenya. Retrieved at
http://www.africaneconomicoutlook.org/en/country-notes/kenya on 25-03-18
African Development Bank (2018). Socio-economic database 1960-2019. Retrieved at http://dataportal.opendataforafrica.org/bbkawjf/afdb-socio-economic-database-1960 2019 on 19-05-18
Allison, P. (1978). Measures of Inequality. American Sociological Review (43), 865-880.
Ansell, B & Samuels, D. (2010). Inequality and Democratization: A Contractarian Approach. Comparative Political Studies 43(12), 1543-1574.
Argyrous, G. (2014). Statistics for Research. London, Sage Publications
Bishop, S & Hoeffler, A. (2016). Free and fair elections: A new database. Journal of Peace
Research 53(4), 608-616.
Blau, P. (1977). Inequality and Heterogeneity – A Primitive Theory of Social Structure. Free Press - USA.
Bollen, K & Jackman, R. (1985). Political Democracy and the Size Distribution of Income. Amercan Sociological Review (50)4, 438-457.
Bourguignon, F. (2015). The Globalization of Inequality. Princeton University Press
Diamond, L & Morlino, L. (2005). Assessing the Quality of Democracy. Baltimore: Johns Hopkins University Press.
Freedom House. (2018). About Freedom in the World. Retrieved from https://freedomhouse.org/report-types/freedom-world on 18-05-18
24 Houle, C. (2009). Why Inequality Harms Consolidation but Does Not Affect
Democratization. World Politics 61(4), 589-622.
Index Mundi (2015). Gini Index – Worldbank Estimate. Retrieved from
https://www.indexmundi.com/facts/indicators/SI.POV.GINI/map/africa on 14-05-18
Lipset, S. (1959). Some Social Requisities of Democracy: Economic Development and Political Legitimacy. American Political Science Review 53(1), 69-105.
McCarty, Nolan and Pontusson, Jonas. 2011. ‘The Political Economy of Inequality and Redistribution’, in Nolan, Brian, Salverda, Wiemer and Smeeding, Timothy M. eds. ‘The Oxford Handbook of Economic Inequality’, Oxford: Oxford University Press Melzer, A & Richard, S. (1981). A Rational Theory of the Size of Government. Journal of
Political Economy 89(5), 914-927.
Muller, N. (1988). Democracy, Economic Development, and Income Inequality. American
Sociological Review (53)1, 50-68.
Muller, N. (1995a). Income Inequality and Democratization: Reply to Bollen and Jackman. American Sociological Review 60, 990-996.
Muller, N. (1995b). Economic Determinants of Democracy. American Sociological Review
(60)6, 966-982.
Osaghae, E. (1999). Democratization in Sub-Saharan Africa: Faltering Prospects, New Hopes. Journal of Contemporary African Studies 17(1), 5-28.
Unicef (2011). Global Inequality report – retrieved from
https://www.unicef.org/socialpolicy/files/Global_Inequality.pdf on 21-05-18
Vanhanen, T. (2000). A New Dataset for Measuring Democracy, 1810-1998. Journal of
Peace Research 37(2), 251-265.
World Bank. (2018a). Gini coefficient and economic data. Retrieved from
http://iresearch.worldbank.org/PovcalNet/povOnDemand.aspx on 25-04018 on
18
World Bank. (2018b) GDP, current US Dollar (data for all countries necessary. Retrieved from https://data.worldbank.org/indicator/NY.GDP.MKTP.CD on 14-06-18
25 Appendix I. Definitions control variables (Source Bishop and Hoeffler, 2016)
Variable Definition
Free&fair election Dichotomous variable, takes a value of one if the election was free and fair, source: Bishop&Hoeffler (2016)
Free election Dichotomous variable, takes a value of one if the election was free; out of election conduct variables 1-7 at least four have a value of one, source: Bishop&Hoeffler (2016)
Fair election Dichotomous variable, takes a value of one if the election was fair; out of election conduct variables 8-10 at least two have a value of one, source: Bishop&Hoeffler (2016)
GDP A country’s Gross Domestic Product. This is a numeric value. This is not copied from Bishop & Hoeffler (2016)
Aid Overseas Development Assistance as a share of GDP, averaged over the past three years, percentage of GDP, source: WDI 2014
Trade Trade (imports plus exports) as a share of GDP, averaged over the past three years, percentage of GDP, source: WDI 2014 Observers present Dichotomous variable, takes a value of one if the variable
observerrating takes a non-missing value, source: Bishop&Hoeffler (2016)
Executive constraints Dichotomous variable of executive constraints, takes a value of one if xconst≥5, source: Marshall et al (2013)
Constitutional term limit Dichotomous variable, takes a value of one if there is a constitutional term limit, source: DPI
Y19901995 Dummy variable. All elections within this timeframe are coded 1, all others 0. (This is no data from Bishop and Hoeffler, 2016) Y19962001 Dummy variable. All elections within this timeframe are coded 1, all others 0. (This is no data from Bishop and Hoeffler, 2016) Y20022007 Dummy variable. All elections within this timeframe are coded 1, all others 0. (This is no data from Bishop and Hoeffler, 2016)
26 Appendix II. 10 variables on which elections are considered to be free and/or fair (Source: Bishop and Hoeffler, 2016)
Ten Variables of Election Quality: Each of our election quality variables is based on a number
of criteria, if these were fulfilled we code the variable one and zero otherwise. We were unable to code each criterion for every election and our dummy variable summarizes our assessment. If we had insufficient evidence to reach an assessment we coded as -22.
Variable 1: Legal framework
Criteria:
a. Citizens are constitutionally guaranteed the right to vote
b. Citizens are constitutionally guaranteed the right to run for office
c. Laws governing the electoral process are not changed just before the election d. Elections are held at regular intervals
Additional Information: We assumed that these criteria have been met if a comprehensive observer report is available and does not state otherwise. Coded as 0 if the following cannot vote: expats, prisoners, military. Coded as 0 if the clergy cannot vote (but not if they can vote but cannot stand). Coded as 0 if the eligibility requirements include past political actions of candidates, e.g. engagement in a previous regime, where these actions did not entail human rights violations or subvert an existing democratic regime.
Variable 2: embs - Electoral Management Bodies
Criteria:
a. Election boundaries are set so that no candidate/party is favoured (no gerrymandering) (de facto)
b. EMBs are held accountable to election law and abide by it c. EMBs are independent and impartial
d. EMBs have sufficient time to organize elections (i.e. no snap election)
e. Decisions made by and complaints made to the EMBs are subject to review and possible reversal
Additional information: Coded as -33 if no relevant body exists.
Variable 3: Electoral rights
Criteria:
a. Equal suffrage is in place for citizens of voting age (e.g. no voter group is systematically disadvantaged) (de facto)
27 b. Equal and effective access to polling stations is in place
c. Any limitations on voting are based on internationally recognizable and acceptable norms
d. Voters have been informed effectively about how and where to vote
Additional information:We assume these criteria were met if a comprehensive first hand observer report is available and does not state otherwise, and turnout is over 50%
Variable 4: Voter register
Criteria:
a. Voter registers are up-to-date for the election taking place
b. Voter registers are accurate: without false names, lack of correct names of individuals, inclusion of name of non-eligible voters (e.g. the dead or children) and multiple entries c. Voters are able to easily and effectively register to vote and can meet the necessary
requirements on time Additional Information:
Coded 0 for registration of under 80%. For over 80%, coded as 0 if there is evidence of deliberate malpractice.
Variable 5: Ballot access
Criteria:
a. Citizens eligible to stand are able to compete in the election (de facto) b. Parties/Candidates get equitable treatment when applying for office
c. Any rejections of candidature are based on internationally recognizable and acceptable norms
d. No one candidate gets over 75% of the votes Additional Information:
We assume that these criteria were met if a comprehensive first hand observer report is available and does not state otherwise, and there is no other reason for doubt.
Variable 6: Campaign process
Criteria:
a. No violence, bribery, intimidation or any other unequitable treatment of voters occurs during the process
b. No violence, bribery, intimidation or any other unequitable treatment of candidates occurs during the process
28 c. Campaigns are free from government interference and the candidates are able to freely
express themselves by holding rallies, etc. d. Campaign Finance:
i. Prohibition on use of government resources other than that provided to all candidates ii. Without massive financial advantages for the incumbents
Additional Information: Where there is ongoing violence and intimidation against political dissidents initiated by the government on account of their political views, we coded 0. However, we did not code as 0 for ongoing violence by armed forces that do not pertain to
the citizen’s political views.
We did not code as 0 on the basis of generalised terrorist violence, if citizens are not in more danger of this violence if they choose to vote or engage in political activity than if they
abstain (note to self, Bhutan, Honduras).
We did not code as 0 on the basis of limited and isolated violent incidents that are ambiguously election-related, e.g. between clashing gangs where motivations are usually complex. We assumed that these criteria were met if a comprehensive first hand observer report that covers the campaign period is available and does not state otherwise.
Variable 7: Media access
Criteria:
a. All parties/candidates are provided with access to the media
b. All parties/candidates have equitable treatment and time on government owned media and the ruling party does not get disproportionately large media coverage in the name of news/editorial coverage
c. Freedom of speech is preserved Additional Information:
We did not code as 0 on the basis of complaints that all larger parties, including non-ruling, received more media coverage than significantly smaller parties.
Variable 8: Voting process
Criteria:
a. Votes are cast by secret ballot
b. Voters are practically limited to one vote per person (de facto) c. Adequate security is in place for both the voters and the ballots
d. Balloting is done without ballot box stuffing, multiple voting, destruction of valid ballots, officer voting, or manipulation of votes cast outside polling place
29 Additional Information:
We coded as 1 if independent sources state as free and fair without giving details, as this is a minimal condition. We did not use this guideline if opposition claim fraud.
We did not code as 0 for the following cases:
Reports of procedural irregularities in vote process if they are deemed by the reporter have neither a) been intentionally fraudulent, nor b) distorted the outcome.
Anecdotal reports of violations of ballot secrecy if their presentation suggests that they are anomalous or estimated to occur in less than 10% of cases, and they are not maliciously intended.
Variable 9: Role of officials
Criteria:
a. The officials adhere to the election procedures (e.g. they have been trained adequately and know which procedures to follow; they do not interfere in the voting process and file complaints made to them, etc.)
b. Unauthorized persons are barred from entering the polling station (e.g. Army members)
c. No campaigning is done within the polling station
d. Transparency is in place: all parties are able to have observers in the station e. International Election Observers can view all parts of the voting process Additional Information:
If there we reports of ignore inadequate training which were deemed not to have affected the outcome we coded as 1.
We assumed these criteria were met if both voting and counting are held to have been entirely trouble free.
Variable 10: Counting of votes
Criteria:
a. Tabulation of votes can be tracked from polling stations up through intermediate centres and to the final processing station
b. Entire counting process is observed by more than one group
c. No rules on what constitutes a valid ballot that favour one candidate/party
d. No evidence for fraud in any way (e.g. no inflation of election results by polling officials, no tampering with the ballot boxes during the counting or movement, etc.) Additional Information: We coded as 1 if independent sources state as free and fair without giving details, as this is a minimal condition. We did not use this guideline if opposition claim
30 fraud.
We did not code as 0 for the following cases: Procedural irregularities in the counting process if they are deemed by the observer to have neither a) been intentionally fraudulent, nor b) distorted the outcome.
