In all our analyses, the dependent variable is the presence or absence of a given electoral law and the independent variables are region, economic development, degree of democracy, and cultural heritage.
Our universe consists of the 237 countries and territories listed in the Comparative Data set that provides a systematic collection of how countries manage their elections (being accessed on the ACE Electoral Knowledge Network). It covers the field of electoral administration around the world including older election administrations (Europe and Latin America), newer (Eastern Europe, Pacific, some places in Africa, Middle East) and those who start from scratch (as Afghanistan or Iraq). Related territories are included in that collection; it should be noted however that there is a lot of missing data about them. We provide information about the rules according to which elections take place in both tiny and huge territories, poor and rich, free and unfree environments.
We have selected nine aspects of electoral legislation which seemed to be interesting and important and about which sufficient information was available.
On each dimension, we first document the variety of electoral laws while identifying the most popular practices. We then determine whether the propensity to adopt or not to adopt a given rule depends on the region, the country’s wealth, its degree of democracy, and its colonial heritage.
Our starting point is geography. We have retained the six regions according to the classification used by the ACE Electoral Knowledge Network: Africa, Americas, Asia, Middle East, Pacific, and Europe. We determine whether some rules appear to be more popular in some regions than in others. When such differences occur, and they do occur rather frequently, we are not in a position to specify why it is so. This could be because of cultural, economic or political differences between these regions. Deeper analysis would be required to flesh out the meaning of these regional variations.
Regional differences are inherently interesting, and this is why we systematically start with a simple description of patterns across the continents. As noted above, the exact meaning of these differences is ambiguous, and this is why we analyze them separately from the other factors. To the extent, however, that these regional variations do not seem to be related to the economic or political factors that are also considered, we are inclined to suppose that they reflect the impact of the culture and/or the social networks that link the political elites within a given region. We offer these conjectures, but they should be treated cautiously. Finally, we should point out that the number of observations for the Middle East (14 at the maximum, often less than 10 because of missing data) is quite small, and that generalizations concerning that region should be taken with extra caution.
After an examination of geographical variations, we look at the potential role of three factors that could affect the nature of electoral laws: the level of economic development, the degree of democracy, and colonial heritage.
The first factor is the economy. The indicator is GDP per capita in 2002 US dollar power purchasing party. The source is the cross-national comparative data The Quality of Government Dataset (version 2008) by Jan Teorell, Sören Holmberg and Bo Rothstein. We have data on 174 countries; the range is from 520 to 61,190, with a mean of 9087 and a standard deviation of 9925. Since Seymour Martin Lipset, a leading scholar of democracy, who wrote in 1959 a pioneer work on Economic Development and Political Legitimacy, a vast literature has established that economic development is a strong antecedent of democratization, and it is thus “natural” to verify whether electoral laws are systematically different in rich and poor countries.
The second factor is the degree of democracy. We use the political rights scores given each year by Freedom House. These scores range from 1 (most free) to 7 (least free). We use the mean score obtained between 2001 and 2006. These scores have been standardized so that the minimum equals 0 for those 13 countries that got the worst score of 7 every year and 1 for those 55 countries which were given the best score of 1 every year. Strong democracies constitute 29% of the cases for which we have information. Our objective is to determine whether democracies conduct elections differently from non democratic or authoritarian countries.
One may wonder whether it is the degree of democracy that leads to the adoption of a given electoral law or the reverse. It must be acknowledged that causation may run both ways. On the one hand, a country that has become “democratic” would be inclined to modify its electoral laws in some direction. On the other hand, a country that has adopted a given set of rules would be more likely to be construed as “democratic” by the international experts who rate political rights across the world. While both causal directions are possible, we believe that casuality runs mainly from degree of democracy to the adoption of specific laws. Our belief is based on the assumption that experts evaluate the extent of political rights first and foremost on the basis of concrete indicators of freedom of speech and the observed degree of competition than on the text of the laws.
The third factor to be considered is colonial heritage. We distinguish former French (27 countries), Spanish (19 countries), and British (66 countries) colonies. In their analysis of election laws in 63 contemporary democracies (“Establishing the Rules of the Game: Election Laws in Democracies” 2004), Massicotte, Blais and Yoshinaka found that former colonies tend to adopt the same electoral rules as their earlier mother country, and we wish to establish whether this pattern holds more generally.
Given the exploratory nature of this research the statistical analyses have been kept as simple as possible. We have dichotomized the dependent variables, and so we have simply distinguished those countries that did or did not adopt a given rule. For each rule, we provide two pieces of information. We first show the bivariate correlation between regions and the existence or absence of a rule. We then indicate in a summary table whether the presence or absence of a rule is significantly related or not to level of economic development, degree of democracy, and colonial heritage. We have performed multivariate analyses including these three factors and the summary tables tell whether there is a statistically significant correlation.
We should mention at the outset that the correlations between our independent variables are generally modest. The strongest correlation (.66) reflects the fact that an overwhelming majority of former Spanish colonies are located on the American continent, and as a consequence it is sometime difficult to distinguish the effects of these two factors. All other correlations are under .5, except that between Europe and GDP per capita, which is exactly .50.
The findings we report and the interpretations that we suggest are necessarily tentative. We are conscious of the fact that some of the information is incomplete and in some cases (we hope only seldom) perhaps even inaccurate, given the time span from the last validation of the data (for some countries the data has been confirmed only in 2001, for others it has been validated more recently). We examine a limited number of factors that may “explain” why certain rules are adopted in some countries and not in others and we test basic models. Our work is merely suggestive, and our hope is that it will trigger further and more elaborate analyses.
Despite all these limitations, we believe that it is fruitful to get some sense of different electoral laws that prevail in the world, to take stock of the amazing variation that exists, and to point out the factors that seem to be related to these variations.
Click here for the next chapter "Legal Framework"
