Printer Friendly

The strategy of dominant-party politics: Electoral institutions and election Outcomes in Africa.

1. Introduction

An ongoing problem in the study of African politics is why legislative elections there, whether democratic or semi-authoritarian, tend to produce party systems of the dominant or predominant variety. (1) Table 1 contains data on recent elections in a cross-section of African nations, (2) and the data show that the nations of Africa employ a distinct variety of electoral rules in their respective legislative elections. Despite this variety of electoral rules, we also see from the table that election outcomes tend to be similar with respect to the effective number of electoral and parliamentary parties they produce. (3)

The two narrow columns in the middle of Table 1 contain measures of the Effective Number of Electoral Parties (ENEP) and the Effective Number of Parliamentary Parties (ENPP), and while they appear to involve lower ENP values in the countries employing SMD plurality rules, these differences are overstated. (4) Differences across FTFP, Plurality and List PR systems on these two indicators are due to a single outlier, Benin, which registered very high values on its effective number of parties indicators. (5) When values for this country were eliminated and the averages recalculated, the results were nearly identical, where SMD Plurality systems had values of 2.49 and 1.96 (ENEP and ENPP respectively) and the PR systems had values of 2.51 and 2.06. (6)

These similarities are also revealed in data contained in the last column of Table 1, which contains the name of the largest vote getter in the elections listed in the first column as well as the percentages of votes and seats each obtained in the reported elections. As the data show, single party dominance is the prevailing pattern is these countries, and this is true regardless of whether single-party dominance is measured in terms of vote or seat shares. There are exceptions to be sure, but what makes this pattern interesting is that it exists in spite of the fact that the nations of Africa employed different electoral formulas. 7 Perhaps the most notable exception is the 2001 election in Zambia where the largest party, the Movement for Multi-Party Democracy, received only 28% of the vote but 46% of the seats. While this vote share is lower than what we typically associate with a dominant party system, the seat share is comparable to what we witness in the predominant party, minority government systems of Scandinavia.

The literature on electoral rules and party systems tells us that SMD plurality rules are associated with two large parties, while rules that involve some version of PR tend toward multiple electoral and parliamentary parties. (8) The electoral data in Table 1 , however, tell us that something other than Duvergerian imperatives are at work in the party systems of Africa. While some of the countries that use SMD Plurality rules, as well as those which have PR Systems, tend toward the expected number of aggregate parties, most countries reveal a stronger tendency toward dominant-party politics. (9) This is not to suggest that electoral rules are irrelevant in Africa or operate there in a completely unexpected manner, because this is not the case. Indeed, the changes in the effective numbers of electoral and parliamentary parties that accompanied the electoral system change made in Lesotho provide evidence for this. Rather, the point is that because single party dominance that is prevalent pattern there, regardless of type of electoral system in use, there must be other factors at work in these party systems.

Our purpose in this research is to identify these factors, illustrate how they work, and ultimately answer the question of how it is that different electoral rules produced relatively similar patterns of single-party dominance in the nations of Africa. While past scholarship has attempted to answer this question, most of it has done so only indirectly, that is, in the context of evaluating whether or not various African nations have crossed the threshold of democracy in their electoral politics. Given the authoritarian past of African nations, this is a reasonable approach to the problem of elections and party politics in these nations. (10) Nonetheless, it has contributed to us still being without a more focused explanation for why single-party dominance in the prevailing pattern in Africa.

Our approach to this problem is direct and begins with the fact that, to be the dominant actor in a nation's legislature, a party must obtain support from a majority of its voting population. How dominant parties accomplish this, however, will vary depending on two factors. First, it will vary depending on the characteristics of the nation's population, that is, whether it is relatively homogeneous or segmented and divided in terms of such features as ethnicity, language, culture, religion, and geography. In a homogeneous society, a dominant party simply has to obtain and keep the support of most voting members of a uniform society. In the case of a heterogeneous society, the dominant party can seek majority support in the same way, namely across all groups at the same rate, or it can achieve its dominance by seeking the support of some groups at the expense of others.

Second, the manner in which electoral dominance is achieved will vary depending on a nation's electoral rules, that is, whether candidates stand for office in single-member, plurality district systems or multi-member districts where winners are determined by some version of proportional representation. In each case, a dominant party has to be the largest holder of legislative seats, but it will approach this problem differently depending on how candidates stand for office and electors cast their ballots.

What is important about these two sets of factors is that they allow us to profile countries in terms of politically relevant groups and electoral institutions. Characterizing countries in terms of these two factors is useful for determining how dominant parties achieve their exalted status, because the bases on which single-party dominance rests vary depending on how countries are defined in terms of both politically relevant groups and electoral institutions. This is because different combinations of these two factors lead dominant parties to employ different strategies to achieve that status, and, to show how this occurs, we examine available data in the Afro Barometer surveys. (11) These data allow us to characterize the countries of Africa in terms of the nature and distribution of politically relevant groups which we can then combine with the electoral rules they use.

Our efforts to address this problem will proceed as follows. We first review explanations for dominant party politics in Africa, revealing how they inadequately address the possibility that the sources of single party dominance may be different across the nations of the continent based on how those countries are defined in terms of electoral institutions and socio-political factors. We also use the existing literature to generate a set of hypotheses about how different electoral rules and political conditions lead us to expect the support bases for Africa's dominant parties to be different. After this, we then begin to map how the sources of single-party dominance are indeed different across the different nations. This mapping will begin with the extent to which voters in the countries we examine are supporters of a political party, after which we then categorize nations by electoral institutions and politically relevant groups. We then map the different bases of dominant-party support to determine if party support patterns across each nation's regions, ethno-linguistic groups, and other socio-political divisions correspond to our expectations.

2. Single Party Dominance in Africa: Explanations

In earlier analyses of single-party dominance in Africa, (12) the emphasis was on cultural and historical factors, especially as they relate to the level of a nation's political development. (13) In more recent studies conducted in the African multi-party elections period, the emphasis has been on political and institutional factors, specifically, how electoral institutions and ethno-linguistic and other socio-economic divisions within in a country are exploited by the ruling government elites to maintain their dominant electoral statuses. Moreover, in more recent analyses, scholars have typically explored the operation of political and institutional factors in two different ways. They have either looked at such factors through case work, that is, by examining the results of single elections in specific countries, (14) or they have defined measures of these factors, constructing large-N aggregate data sets and then estimated the impact of the factors they deemed important. (15)

We have learned much from these two types of studies, but there are certain limitations with existing studies that behoove us to proceed in a different direction. In the single-election, single-country studies, we have learned about why election outcomes turned out the way they did in selected cases. For example, we learned that it was the manner in which Lesotho's SMD plurality system advantaged the Lesotho Congress of Democrats (LCD) that led it to obtaining 79 of 80 available legislative seats in the 1998 election. Indeed, compared to the parties of the opposition, the LCD possessed a generally large support base across the all of the nation's regions, allowing it to win against a fragmented opposition in virtually of the nation's single-member districts. (16) In the case of the 1995 legislative election in Tanzania, we see a different set of factors operating. Specifically, the overwhelming victory for the ruling CCM Party (Chama Cha Mapinduzi) was explained by the advantages that party enjoyed by virtue of its having been in the governing position going into the election. (17)

While these explanations are well done and offer insight into why the results of these two contests turned out the way they did, they are not generalizable across the entire set of elections that we need to understand if we are to explain why elections in Africa have so often produced single-party dominance. This is not a criticism of the work provided by the authors of these and similar studies, because they were not intended to show how the results of specific elections explain the phenomenon of single-party dominance across the African cases in which it exists. This may suggest that the large-N, aggregate data analysis of Mozaffar et. al. (2003) offers a better, more general explanation for single-party dominance in Africa. To be sure, the Mozaffar et. al. (2003) analysis is general and offers many insights--insights that will guide the analyses we offer below, but it alone does not adequately answer the question of whether or not dominant party politics is the same across the nations of the African continent.

There are essentially two reasons for this, and the first, to be completely fair, concerns the fact that this research was not designed to address the question as we have posed it here, that is, whether or not the sources of single party dominance are the same across the nations of Africa. Mozaffar et. al. (2003) use aggregate data which tell us about the political, institutional, and social characteristics of the nations they examine, but to determine if variation in these factors is related to different electoral strategies on the part of dominant parties requires additional data. Specifically, since different strategies should manifest themselves in different patterns of support across politically relevant groups, it is necessary that the aggregate data of the kind used in Mozaffar et. al. (2003) be combined with individual-level data like those found in the Afro Barometer surveys. Another limitation of the Mozaffar et. al. (2003) analysis for our purposes concerns the dependent variable, the Effective Number of Parties--both electoral and parliamentary. While the effective number of parties has become the standard way to calibrate the number of relevant political parties in a political system, it is not well suited to capturing single-party dominance in Africa. Again, consider the data presented in Table 1 and the discussion we provided above about the phenomenon of single-party dominance in Africa. While the Effective Number of Parties measure indicates the presence of a dominant party in some countries, it does not in others. (18) For example, the dominant party of Botswana (BDP) increased its seat share between the 1999 and 2002 legislative elections, but the ENPP measure suggested that the party system in Botswana moved closer to a two-party format.

Overall, these issues suggest that, to answer the question of why elections in Africa tend to produce single-party dominance, even across different institutional formats, we need to combine individual-level and aggregate data. This is because only with both levels of data will we be able to determine whether or not the bases of single-party dominance are the same or different across the party systems of Africa. This process begins with examining countries' electoral institutions to assess how they are related to the politically relevant groups in that country that are then used by dominant parties to achieve and maintain their electoral dominance. Mozaffar et. al. (2003) have argued that, for most African nations, exclusive support from a single group alone is not sufficient to guarantee the electoral success of a dominant party. This means that group fragmentation alone "has no significant effect on the structure of party systems in Africa's emerging democracies," (19) and that, to understand why dominant parties exist, we must also examine how countries are divided regionally and how electoral institutions aggregate politically relevant divisions.

In light of this, the persistence of dominant party systems in Africa can be understood only as the result of the individual and joint action of two variables, namely politically relevant groups and the country's electoral institutions. With respect to the former, what is important is not simply mapping the size and fragmentation of politically relevant groups but also how geographically concentrated or dispersed they are throughout a country. These characteristics will help dictate what a ruling party has to do to capture a level of support sufficient to allow it to maintain its status as dominant party. Concerning the latter, electoral rules are the primary institutions through which votes from a country's politically relevant groups are aggregated and translated into legislative seats.

In his discussion of the origin of the electoral rules used by the nations of Africa, Mozaffar (2004) has isolated five institutional patterns. (20) The first two institutional patterns include countries that employ SMD plurality rules and those that employ some form of majority runoff. The second two institutional patterns include different versions of proportional representation, while the final institutional pattern includes countries with mixed systems. The important question for our purposes is why these different institutional patterns exist, that is, what is the political significance of the patterns we witness, and the answer rests with the manner in which a country's politically relevant groups are defined and distributed.

The answer to this question constitutes the variable we are to explain, that is, how do Africa's ruling parties achieve their electoral dominance. To answer this question, we need to test the hypothesis of whether or not, to achieve a majority of votes and seats sufficient to dominate a country's legislature, a ruling party devises an electoral strategy that depends on the size, fragmentation, and geographic concentration of its country's politically relevant groups. In other words, do such parties choose electoral institutions that best assist it in its strategy to achieve and maintain their electoral dominance. This means dividing its support among groups and aggregating the votes of its supporters in a way that leads to its candidates obtaining the dominant share of electoral seats.

To test this hypothesis, we assume that the variety of electoral institutions we see in Africa were put into place because they constituted the rules best suited to assist ruling parties in their strategies to maintain electoral dominance in light of the characteristics of their respective country's politically relevant groups. This means that, while we see different patterns of dominant party politics in Africa, countries employing certain electoral institutions should be characterized by similar patterns of politically relevant groups. This is because, to maintain electoral dominance, they need to divide group support and aggregate supporters' votes in ways that correspond to group characteristics. To show this, we begin by mapping the countries of Africa in terms of basic party support patterns and then seeing if these basic patterns differ by the type of electoral rule used for legislative elections.

3. Party Support, Electoral Institutions, Opposition Strength

The data in Table 2. are taken from recent Afro Barometer surveys and include surveys for seventeen elections in thirteen countries. The data reveal that an average of 57% of respondents stated that they supported a political party. The lowest party support rate was found in the 2000 Uganda election where only 24% of those polled indicated that they had a party they supported, while the highest was found in Malawi, where those supporting a party comprised nearly 81% of the sample. If the understanding of single-party dominance in Africa we have discussed above is correct, these are dramatic differences in party support should be related to the type of electoral system being employed in a country.

While party support is a complicated phenomenon that is explained by a number of individual and institutional factors, our expectation is that party support rates should be generally higher in countries that employ some form of proportional representation rule. (21) Dividing the sample into nations that employ majority/plurality rules and those that use some form of PR and recalculating mean party support rates confirmed this. Average party support rates for the PR systems was nearly ten percentage points higher than those nations that used plurality/majority rules. (22) Moreover, as expected, for those countries where support rates were below 45%, none employed a proportional representation rule. On the other hand, a number of countries where support rates were high employed plurality/majority rules. This tells us that while party support is certainly connected to electoral institutions, it is also related to something else, namely the characteristics of politically relevant groups in the society and how they are connected to ruling parties.

In addition to differences in rates of party support, the nations in the table can also be distinguished in terms of the electoral strength of the political parties that oppose those that are in the governing position. To capture this difference, we calculated how much larger the governing party's support level was compared to its strongest challenger. (23) These calculations produced three categories of countries with the first group defining countries where opposition parties were fairly strong relative to their governing counterparts. This group includes Ghana where the governing party was 1.5 times larger than its largest challenger, Malawi where the ruling party was 1.9 times larger, and Botswana where the governing party is 2.0 times larger than the strongest challenger. On the other extreme, there were several countries where the governing party was substantially stronger than its strongest challenger. These include Mozambique where the governing party had 17.8 times more support than its strongest challenger, and Tanzania, Uganda, and Zambia where governing parties enjoyed nearly ten times more support than their strongest challengers.

With these data we can construct a classification of African countries that distinguishes them based on their overall levels of party support and the electoral strength of the largest opposition party. This classification scheme is presented in Figure 1, and it includes only those countries where differences on the two dimensions were clearly different. In the upper left quadrant are Ghana, Malawi, and Botswana, countries where party support levels are high and the largest opposition party is strong relative to each nation's governing party. In the lower right quadrant are Uganda and Zambia, countries where party support levels are rather low and opposition strength relative to each nation's governing party is weak. On the off diagonal are three other countries. In the lower-left quadrant of the figure is Nigeria where party support is low but where the opposition is relatively strong. In the upper-right quadrant are Tanzania and Mozambique where party support levels are relatively high but opposition strength is rather low.

With the exception of Mozambique, all countries in Figure 1 use SMD plurality rules in their respective legislative elections. This is important with respect to the strategies political parties follow to achieve electoral dominance. In SMD Plurality systems, for a political party to achieve electoral dominance it is necessary that it receive votes sufficient to have its respective candidates win a majority of districts. In a PR system, on the other hand, a political party must be sure that it is not only the largest vote getter but also that it support rate is sufficient to give it sufficient seats to be the dominant parliamentary party. While these differences in election institutions encourage somewhat different strategies to achieve electoral dominance, exactly what these differences are will depend on the electoral strength of the opposition and how politically relevant groups are characterized and distributed throughout a country.

Concerning the former, when parties of the opposition are strong and represent a realistic governing alternative, the dominant party must maintain all the support it can get which means that it can ill afford to alienate large numbers of politically relevant groups, especially if they are well represented in the nation's population. This is a particularly acute problem when elections are held under SMD Plurality rules where electoral institutions encourage opposition groups to amalgamate into large unified opposition parties to avoid the Duvergerian problem and challenge large governing parties.

With such cooperation, strong opposition candidates can become direct challengers to candidates from the dominant party, which makes it imperative that dominant parties go to great lengths not to alienate many politically relevant groups. Naturally, this problem is directly proportional to the electoral strength and unity of the opposition. Electorally weaker opposition parties allow dominant parties to be somewhat more exclusive in their respective strategies in the sense that the loss of some small number politically relevant groups is not as problematic as when the opposition is electorally strong. While an electorally strong and unified opposition is a concern, even when PR rules are used, it is somewhat less of a concern in SMD Plurality systems. This is because, being kinder to smaller parties, PR systems can often keep oppositions more fragmented by encouraging smaller electorally relevant groups to put up their own party lists and eschew electoral cooperation.

Given that the nations in Figure 1 are different on these party support and opposition strength characteristics, we would also expect that they will be different in terms of the nature of group support for their respective dominant parties. While the nations under consideration are not uniform in terms of how politically relevant groups are defined and distributed, (24) we still have certain expectations about how the above-defined differences will influence patterns of dominant party support. On the one hand, in nations with high levels of overall party support and strong oppositions, we would expect dominant party support to be more uniform across politically relevant groups. On the other hand, as opposition strength wanes and overall levels of party support decline, we expect that this uniformity of dominant party support across groups to diminish and a more distinctive group basis of party support to emerge.

Using data from the available Afro Barometer surveys mentioned above, we turn to this mapping in the next section. The Afro Barometer surveys allow us to produce such a dominant party support mapping because these surveys all contain questions that allow us to divide respondents into groups distinguished by region or district of the country in which they live, ethno-linguistic characteristics, and party support. (25)

4. Dominant Party Support: Ethno-linguistic and Regional Characteristics

Our mapping of dominant party support in Africa begins with those countries where overall levels of party support were high and support for the opposition was also high, Botswana, Ghana, and Malawi. (26) We see from the data in Table 3 that these three countries are different in terms of the level support the dominant party in each received in the Afro Barometer surveys we examined. Botswana's and Malawi's parties each were about ten percentage points stronger than was Ghana's which tells us that the electoral strategy followed by the former two was superior to that followed by Ghana. To distinguish the strategy followed by each party, we analyze more closely how the support characteristics of each can be distinguished in terms of the groups that support each country's dominant party and the regional/district distribution of governing party support.

We first examined the extent to which ethno-linguistic and regional differences explained support levels for each country's dominant party. This was accomplished by crosstabulating the regional and ethno-linguistic variables with the party support variable and then producing correlations and tests of statistical significance. Given that the ethno-linguistic and regional differences are essentially categorical variables, the lambda coefficient was used as the correlation coefficient while the chi square was used to determine statistical significance. In the third column of Table 3, we see that region and ethno-linguistic group are statistically significant explanations for differences in support for dominant parties but that the level of association with respect to these categorical variables in generally weak. This is true for all three countries, but it was least true for Malawi where the lambda coefficients were highest.

While these measures tell us something about how these three countries are different with respect to the ethno-linguistic and regional basis of support for each's dominant party, they do not tell us if the dominant party in each country followed a different electoral strategy. To make this determination, different measures are necessary, and these are contained in the last two columns of the table. The first measure is what we refer to as the average deviation measure, which was created by subtracting the national support rate for each country's dominant party from the rate at which various ethno-linguistic and regional groups supported the dominant party. We see that Botswana and Ghana are similar in that there is more variation across regional groups than there is across each country's ethno-linguistic groups. Malawi, on the other hand, is different in that there was much substantial variation across both regional and ethno-linguistic groups.

These differences are interesting but they still do not tell us enough about the different electoral strategies dominant parties pursued. To determine just how different electoral strategies were, we need another measure. This measure is found in the last column of the table and involves determining whether the dominant parties under consideration here focused on a few regional or ethno-linguistic groups or sought electoral support more broadly. To develop such a measure, we first had to determine whether a group's dominant party support rate was above or below the national support rate. Groups that were above the dominant party's national support rate were classified supporters and those below that rate were classified as non-supporters. We then took all support groups and non-support groups and calculated their average size in the sample. In other words, we averaged the sample proportions of all support and non-support groups, and produced these averages for both ethno-linguistic and regional groups. The results show that dominant parties in each of our three countries followed different electoral strategies.

The most notable contrast rests between Botswana and Ghana. In the case of the former, we see that dominant party supporters in that country tended to come from the more populous politically relevant groups. This is true for the country's ethno-linguistic groups, but it was even more notable for the country's regional groups. Specifically, while the Botswana Democratic Party tended to rely on somewhat larger ethno-linguistic groups, it was much stronger in the regions of the country that are more populous. (27) Data for Ghana, on the other hand, revealed a dominant party following a very different strategy. Specifically, the dominant party there tended to have the support of ethno-linguistic and regional groups that were smaller than those that support the opposition. (28) Non-supporter regional groups tended to be about twice a large as supporter groups, but non-supporter ethno-linguistic groups were more than seven times larger. These observations leave us with the conclusion that the strategy followed by the Botswana Democratic Party, one that relies on groups that are more populous in the country overall, is more electorally effective.

What is interesting about Malawi, on the other hand, is that it is different from both Botswana and Ghana. It is different from Botswana in that the regional groups are nearly identical in terms of party supporters and non-supporters while, in terms of ethno-linguistic groups, supporters were far more represented in the country's larger groups. This is what we would expect because, in Malawi, the Chewa group is more than seven times larger than any other ethno-linguistic group, and it would be difficult to be a dominant party without support from a significant proportion of members of this group. In this sense, the United Democratic Front of Malawi followed a limited size strategy.

The remaining country in the lower portion of Table 3 is Nigeria which is different from the previous three countries because party support levels there are lower overall. This is witnessed in the fact that the nation's dominant party, the People's Democratic Party, received support from less than 25% of the survey's respondents, but it is also different in that, while Nigeria's opposition is categorized as being strong, it is the weakest of the four dominant parties in the two left quadrants of Figure 1. (29) This suggests to us that the dominant party of Nigeria is under less pressure to build a strategy of electoral dominance than its counterparts in Botswana, Ghana, and Malawi. The data in the Table 3 indicate that ethno-linguistic and regional groups are statistically significant factors explaining party support, but they also tell us that these groups are only weakly correlated with differences in dominant party support levels. On the other measures that we created to map electoral strategies, we see in the data that there is less deviation in terms of dominant party support levels across Nigeria's ethno-linguistic and regional groups. This is also reflected in the fact that differences in the size of groups that dominant party supporters and non-supporters rely on are fairly close.

These patterns lead us to conclude that perhaps lower overall party support levels and a somewhat weaker opposition mean that dominant parties can rest more solidly on such differences. In other words, lower levels of party support overall and a weaker opposition means that the dominant party has some advantages that do not require it to be as concerned as dominant parties in the other countries that we examined above to build a winning strategy. However, to examine this idea more thoroughly, we turn next to dominant parties in countries in countries where the opposition is clearly weak, and the data for this analysis are presented in Table 4. This analysis begins with the two countries in the lower right hand quadrant of Figure 1 , namely countries with low levels of party support overall and weak oppositions.

Data on these two countries, Tanzania and Mozambique, are contained in the upper portion of Table 4 and reveal immediate differences and similarities with the countries already examined. Concerning the former, we notice that support for Tanzania's and Mozambique's dominant parties are the highest of any in our analysis. Support for the dominant party of Tanzania is nearly 2/3 of the sample and, for the Frelimo of Mozambique, nearly 60%. In spite of this dramatic difference, we notice that the relationship between ethno-linguistic and regional groups and support for these two countries' dominant parties is similar to that found in the other countries that have already been examined. Specifically, while ethno-linguistic and regional groups are related to differences in support for dominant parties in these two countries, the relationship is weak when measured in terms of correlation coefficients. As in the previous cases, the principal reason for this is the fact that both countries are highly fragmented in terms of politically relevant ethno-linguistic and regional groups.

These two countries are quite similar in terms of the extent to which regional groups vary from their respective national dominant party support rates, but they are quite different in terms of this variation across ethno-linguistic groups. Such ethno-linguistic group variation is much higher in Mozambique than it is in Tanzania. Moreover, the dominant parties of these two countries seem to have followed slightly different strategies in order to maintain their respective levels of electoral dominance. In Mozambique, the dominant party relies overall on ethno-linguistic and regional groups that are on average larger than those on which their opponents rely. Having said this, it is important to note that, while larger, support groups for Mozambique's dominant party are not substantially larger than groups which support the opposition, but this is not threatening to the governing party of Mozambique because it is significantly larger than the largest opposition challenger. The same is true for the dominant party of Tanzania, which is nearly ten times larger than its most significant challenger. Consequently, it is for these reasons that the party is not threatened by relying on generally smaller ethno-linguistic groups than its challengers.

Finally, what is immediately noticeable about the remaining two parties in the table is that, while they are in countries with dominant party systems, they are supported by much lower proportions of respondents according to the samples we have used here. Despite this difference, we see that ethno-linguistic and regional groups are statistically significant factors explaining differences dominant party support rates, but, like the other countries that have been examined, the correlations between groups and dominant party support are weak. We also notice from the data in Table 4 that group support rates for dominant parties vary much less from national support rates than the other countries examined thus far. We see that, for Uganda which has the lowest dominant party support rate of any country examined here, differences in average group size between dominant party supporters and non-supporters are negligible. This is not unexpected given that the ruling party has nearly ten times more support than its strongest challenger. In the case of Zambia, the situation is somewhat different in that the ethno-linguistic groups that support the dominant Movement for Multi-party Democracy are larger than those that support the opposition, but, at the same time, the regional groups that support the dominant party are smaller than those that support the opposition. Again, given the dominance that the ruling party has over its strongest challenger, this is not unexpected.

This paper has shown that ruling parties in Africa rely on different strategies to achieve their electoral dominance and that electoral institutions and the characteristics of politically relevant groups in the country help explain these differences. While this kind of mapping has not been provided in the existing literature, there are still a number of questions that remain unanswered. These include the connection of other groups, like occupational and religious groups, to dominant party support which were not assessed in this paper. Also, a deeper explanation for why different countries rely on different groups for their electoral dominance is necessary as is an evaluation of the effectiveness of the different group strategies. This paper's use of Afro Barometer data has provided an initial foray into these questions about why dominant party politics is so prevalent in Africa. These Afro Barometer data can be used to answer other important questions regarding the patterns of party politics that define the nations of Africa, but these efforts will have to be the subject of future research.


Almond, Gabriel and James S. Coleman (1960) The Politics of Developing Areas. Princeton: Princeton University Press.

Bogaards, Matthijs (2004) "Counting Parties and Identifying Dominant Party Systems in Africa," European Journal of Political Research. 43: 173-197.

Bratton, Michael and Nicholas vande Walle (1997) Democratic Experiments in Africa:Regime Transitions in Comparative Perspective. New York: Cambridge University Press.

Cowen Michael and Liisa Laakso (eds.) (2002). Multi-Party Elections in Africa. Oxford: James Currey.

Friedman, Edward and Joseph Wong (2008) Political Transitions in Dominant Party Systems: Learning To Lose. London: Routledge.

Golder, Matthew and Leonard Wantchekon (2004) "Africa: Dictatorial and Democratic Electoral Systems since 1946," in Josep Colomer (ed.) Handbook of Electoral System Design. London: Palgrave.

Laakso, Markku and Rein Taagepera (1979). "Effective Number of Parties: A Measure with Application to West Europe," Comparative Political Studies. 12; 1: 3-27.

Lapalombara, Joseph and Myron Weiner (eds.) (1966) Political Parties and PoliticalDevelopment. Princeton: Princeton University Press.

Lijphart, Arend (1997) Electoral Systems and Party Systems: A Study of Twenty-Seven Democracies, 1945-1990. NY: Oxford University Press.

Mozaffar, Shaheen, James R. Scarritt, and Glen Galaich (2003) "Electoral Institutions, Ethnopolitical Cleavages, and Party Systems in Africa's Emerging Democracies", American Political Science Review. 97; 3: 379-390.

Mozaffer, Shaheen and James R. Scarritt (2005) "The Puzzle of African Party Systems", Party Politics. 11; 4: 399-422.

Sartori, Giovanni (1976) Parties and Party Systems: A Framework for Analysis. New York: Cambridge University Press.

Southall, Roger and Roddy Fox (1999) "Lesotho's General Election of 1998: Rigged or de Rigeur," Journal of Modern African Studies. 37; 4: 669-696.

Taagepera, Rein and Matthew Soberg Shugart (1989) Seats and Votes: The Effects and Determinants of Electoral Systems. New Haven, CT: Yale University Press.

Van Crananburgh, Oda (1996) "Tanzania's 1995 Multi-Party Elections: The Emerging Party System," Party Politics. 2; 4: 535-548.

(1) On the difference, see Sartori (1976).

(2) The countries in the table were selected because they represent a cross-section of the different types of electoral systems that are employed in Africa. In other words, they have been selected for the purposes of illustration only and are not meant to represent a randomly drawn sample of cases.

(3) The former refers to the numbers of parties (electoral and parliamentary) while the latter refers to the fact that most party systems are of the dominant/predominant variety. On the effective number of parties, see Laakso and Taagepera (1979). On why electoral institutions do not work in Africa as they do in the advanced electoral democracies, see Golder and Wantchekon (2004) and Mozaffar (2003).

(4) Values were calculated using data provided in Mozaffar, Scarritt, and Galaich (2003) from which we calculated the average ENP values for SMD plurality systems and for those countries using some form of PR. The values for the PR nations were 4.05 and 2.93 and, for the SMD Plurality nations, 2.49 and 1.96 respectively for the Effective Number of Electoral Parties and the Effective Number of Parliamentary Parties.

(5) Comoros was also an outlier for countries employing SMD Plurality rules with an ENEP of 14.9 and an ENPP of 5.28. This outlier was not included in our calculations.

(6) These results concur with the data presented in and conclusions offered by Golder and Wantchekon (2004).

(8) These are referred to as Duverger's Law (SMD Plurality Systems) and Duverger's Proposition (PR Systems). For elaborations on these relationships, see e.g., Lijphart (1997) and Taagepera and Shugart (1989).

(9) SMD Plurality system countries like Botswana, Mali, and Tanzania are much more dominant-party than two-party systems, while PR Systems like Namibia and Burkina Faso in 1997 are less multi-party than dominant-party systems.

(10) There are scholarly investigations that focus specifically on Africa's pattern of dominant party politics, and we discuss these in more detail below. But for discussions of African electoral politics in the context of democratization, see e.g., Bratton and vande Walle (1994) and Golder and Wantchekon (2004).

(11) The following surveys have been made available for secondary analysis and, thus, are the ones we use in this research. Specifically, they are as follows:

Botswana (1999), Cape Verde (2002), Ghana (1999 and 2002), Lesotho (2000), Malawi (2000), Mali (2001), Mozambique (2002), Namibia (2000 and 2001), Nigeria (2000 and 2001), Tanzania (2001), Uganda (2000 and 2002), Zambia (1999), and Zimbabwe (1999). We are grateful to those who made these data available and note that all interpretations of the data are our own.

(12) The reference here is to the early post- independence years.

(13) See e.g., Almond and Coleman (1960) and LaPalombara and Weiner (1966).

(14) See e.g., Southall and Fox (1999) on the 1998 legislative election in Lesotho, and the essays in Cowen and Laakso (2002). See Mozaffar (1997) for a comparative discussion of election system choice in Senegal. For a more recent study of the problem of transition in dominant-party systems, see Freidman and Wong (2008).

(15) The best examples of this kind of study is found in Mozaffar (2002) and Mozaffar, Scaritt, and Galaich (2003).

(16) Again, see Southall and Fox (1999).

(17) See e.g., van Cranenburgh (1996).

(18) See especially Bogaards (2004) for an in-depth discussion of this issue and alternative measures of single-party dominance.

(19) Mozaffar, et. al., (2003) p. 384.

(20) The discussion that follows is taken from Mozaffar (2004), pp. 420-426. See especially Table 24.1.

(21) This is because under PR rules, small electoral parties have a higher chance of achieving parliamentary representation than in SMD plurality systems, keeping supporters from abandoning their first choice as readily as they would.

(22) The average support rate for the PR countries was 64% while the average party support rate in the Plurality/Majority countries was 55%., but a difference of means test showed that the difference was not statistically significant. This was most likely due to the small number of cases (N = 4) for the PR systems in the sample. When the difference of means test is recalculated using the same parameters but with a larger number of cases in simulated data, the difference is statistically significant.

(23) These calculations were made using data from the Afro Barometers.

(24) See e.g., Scarritt and Mozafer (1999) and Mozaffer, Scarritt, and Galaich (2003).

(25) There are certainly other ways to categorize respondents into politically relevant groups. Such categorizations include most notably religion, occupation, and gender. Because of their strong relationship to each other, this study focuses on ethno-linguistic divisions and regional or district differences.

(26) The average percentage of party supporters across the three countries was 71%. We defined an opposition as being electorally strong as long as the dominant party possessed no more than two times as many supporters in the Afro Barometer surveys.

(27) An electoral strategy that relies on fewer but larger groups is referred to as a size strategy.

(28) An electoral strategy that relies on more but smaller groups is referred to as a quantity strategy.

(29) The People's Democratic Party is nearly three times as large as the next largest opposition party.

Dennis P. Patterson

Texas Tech University

Leslie Fadiga-Stewart

Delta State University
Table 1. Party Systems and Election Systems in Selected
African Countries (Recent Elections)

Country/ Election ENEP ENPP Dominant Vote/Seat
Elec. Year System % Party

Botswana Dem.Pty. (1)
 1999 54%/68%
 SMD, Plu. 2.7 1.4 Dem.Pty. (1) 52%/77%
 2002 SMD, Plu. 2.8 1.6

Burkina Faso CDP2
 1997 List, PR 3.6 1.8 69%/87%
 CDP2 50%/52%
 2002 List, PR 3.8 3.3

Cameroon CPDM3
 1997 MMD/Plu. 2.9 2.6 NA/61%
 MMD/Plu. 3.3 CPDM3
 2002 2.2 NA/83%
Lesotho L. C. Dem. (4)
 1998 SMD, Plu. 2.2 1.0 61%/98%
 MMP, PR 2.6 L. C. Dem. (4) 55%/65%
 2002 2.2
 1994 SMD, Plu. 2.7 2.7 UDF
 SMD, Plu. 2.8 NA/48%
 1999 2.7 UDF 47%/48%
Mali Alliance (3)
 1997 Majority 2.4 NA/87%
 H 2002
 2002 Majority 1.3 NA/49%
Namibia SWAPO (5)
 1994 List, PR 1.7 1.7 740/0/740%
 1999 List, PR 1.7 1.7 SWAPO (5) 76%/76%
Nigeria PDP (6)
 1999 56%/59%
 SMD, Plu. 2.3 2.3 PDP (6) 55%/62%
 2003 SMD, Plu 2.6
Senegal 2.2

 1998 Mixed 2.4 1.8 Socialist 50%/66%
 SOPI 50%/74%
 2001 Mixed 3.2 2.1
Tanzania CCM (7)
 1995 SMD, Plu. 2.2 1.5 60%/81%
 SMD, Plu. CCM (7) NA/91%
Zambia MMPD (8)
 1996 SMD, Plu. 2.4 1.3 49%/52%
 SMD, Plu. 5.6 MMPD (8) 28%/46%
 2001 3.0
Zimbabwe ZANU-PF (9)
 2000 SMD, Plu. 2.2 2.0 49%/52%
 SMD, Plu. 2.0 ZANU-PF (9) 60%/65%
 2005 1.9

(1) Botswana Democratic Party

(2) Congress for Democracy and Progress

(3) Cameroon People's Democratic Movement

(3) Alliance for Democracy in Mali

(4) Lesotho Congress for Democracy

(5) South West Africa People's Organization

(6) People's Democratic Party

(7) Chama Cha Mapinduzi

(8) Movement for Multi-Party Democracy

(9) Zimbabwe African National Union-Patriotic Front

Source: Calculated by the Authors from IFES

Table 2. Party Supporters and Non-Supporters
in Selected African Countries

Country Survey Percent Percent
Botswana Year Affiliated Not

 1999 73.3% 16.8%
Cape Verde 2002 47.6% 47.8%
Ghana 1999 67.3% 32.7%
 2002 62.3% 32.7%
Lesotho 2000 51.3% 42.1%
Malawi 2000 80.9% 17.8%
Mali 2001 56.1% 42.3%
Mozambique 2002 59.5% 33.5%
Namibia 2000 71.0% 29.0%
 2001 77.3% 22.7%
Nigeria 2000 36.8% 63.2%
 2001 42.4% 53.9%
Tanzania 2001 79.2% 20.8%
Uganda 2000 24.3% 69.5%
 2002 48.0% 51.3%
Zambia 1999 36.4% 62.6%
Zimbabwe 1999 39.6% 55.7%

Source: Complied by the authors from the Afro
Barometer Surveys.

Table 3. Dominant Party Support in Countries with Strong

 Dominant Basis of Average Average
 Party Party Deviation Sample
 Support Support From Size
 Level National Political
 Lambda Support Groups**
 Chi (2) Rate * Supt./
High Party Supt. ~Supt.

Botswana 49.3%
 Group .031 6.1% 15.6%/
 p<.001 10.1%
 Region .012 13.2% 22.3%/
 p<.000 6.1%
Ghana 38.3%
 Group .039 8.4%
 p<.000 4.2%/
 Region .079 13.1% 30.9%
 p<.000 7.9%/
Malawi 46.1% 14.5%
 Group .116 24.6% 26.2%/
 p<.001 5.2%
 Region .143 27.8% 6.6%/
 p<.000 6.7%
Low Party
Support Levels
Nigeria 23.6%
 Group .061 7.6% 9.6%/
 p<.000 11.9%
 Region .029 11.6% 2.2%/
 p<.000 3.1%

* This is Calculated by subtracting the national support rate for the
dominant party from the group or region support rate.

** This is calculated by taking the average size of groups (percentage
of respondents in the sample) for those groups who are above and
those who are below the national support rates.

Source: Compiled by the Authors from various Afro Barometers

Table 4. Dominant Party Support in Countries with Weak

Dominant Basis of Average Average
Party Party Deviation Sample Size
Support Support From Political
Level National Groups **
 Lambda Support Supt./
 Chi (2) Rate * ~Supt.

Party Support

Tanzania 63.1%
 Group .001 6.1% 5.4%/
 p<.000 12.8%
 Region .028 8.2% 5.1%/ 5.2 %
Mozambique 58.8% p<.000

 Group .017 12.1% 7.9%/ 5.3 %
 p<.000 10.1%/ 8.5%
 Region .042 8.8%
Low p<.000
Party Support
Uganda 19.6% .020 5.7% 7.7%/
 Group p<.000 6.6%
 .026 10.1% 3.2%/ 3.8%
 Region p<.000
Zambia 30.7 .017 3.5% 15.2%/
 Group p<.000 0.5%
 .017 6.3% 7.9%/ 11.3%
 Region p<.000

* This is Calculated by subtracting the national support rate for the
dominant party from the group or region support rate.

** This is calculated by taking the average size of groups (percentage
of respondents in the sample) for those groups who are above and
those who are below the national support rates.

Source: Compiled by the Authors from various Afro Barometers

Figure 1. Classification of African Party
Systems by Party Support and Opposition

 Opposition Strength

 Strong Weak

High Party Ghana Tanzania
Support Malawi Mozambique

Low Nigeria Uganda
COPYRIGHT 2009 Central European University
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2009 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Patterson, Dennis P.; Fadiga-Stewart, Leslie
Publication:CEU Political Science Journal
Article Type:Report
Geographic Code:60AFR
Date:Apr 1, 2009
Previous Article:Authoritarianism in the making? The role of political culture and institutions in the Albanian context.
Next Article:Electoral reforms and democratic consolidation in Nigeria: the electoral act 2006.

Terms of use | Privacy policy | Copyright © 2018 Farlex, Inc. | Feedback | For webmasters