Theoretical Reflection on Institutional Innovation and Performance

This section concentrates on the relationship between institutions and their environment by questioning two theoretical aspects explaining institutional efficacy or inefficacy and upon which I build further: on the one hand, the origin of institutional change and the factors that explain institutional autonomy, and on the other hand, the structuring capacity of institutions.

New-institutionalists (March and Olsen, 1989: 2–8) have identified many competing ways of conceiving of this relationship. Even though these ways vary according to the three new-institutionalisms (historical, sociological and rationalist), I concentrate on historical institutionalism, which, according to many authors, offers a more complex view of politics and the best potential for heuristic exchanges with the other components (Hall and Taylor, 1997: 494; Immergut, 1998: 28; DiMaggio, 1998).

Let us turn now to the first theoretical avenue this article is founded upon. Historical institutionalists explain institutional creation and change as products of critical historical moments that produce innovation (Krasner, 1984; Dobry, 2000). In this study of electoral commissions, two aspects may be distinguished analytically, though they are empirically related. The first relates to the process of institutional innovation and especially to the interaction between political actors, which is particularly important in the processes of democratization (O'Donnell, Schmitter and Whitehead, 1986; Guilhot and Schmitter, 2000). In historical institutionalist literature, though diverse and generally associated to external shocks (Krasner, 1984), critical moments leading to institutional change are seen as highly conflictual situations between actors struggling to define new rules and to design new institutions (Lecours, 2002: 12). In other words, in order to understand the genesis and status of electoral commissions, one must pay attention to power-based relationships between actors. As Immergut (1992: 85) demonstrated in her comparative study of policy making, “institutions are most certainly created by social actors engaged in a struggle for political power.” The issue of this battle may depend on many factors, but power relations are nonetheless crucial determinants. In addition, K. Thelen (2004: 31–33) points out that in most instances, the new rules and institutions reflect the asymmetry of power between key actors or coalitions engaged in the process of institutional building. The initial distribution of power during moments of innovation has a direct impact on the nature and prerogatives of the new institution, as I will demonstrate further on. Such a power politics view of institutional genesis and innovation has been fruitfully applied to understand the development of sub-state nationalism (Lecours, 2005); the political economy of skills in industrial countries (Thelen, 2004); the adoption of the European Community's Budgetary Treaty (Lindner and Rittberger, 2003); or the forging of bureaucratic autonomy (Carpenter, 2001). I propose to use this power politics view to understand the capabilities of electoral commissions and the diverging quality of electoral management in the case study countries. But because institutions are not congealed structures without a dynamic of their own (Immergut, 1992: 85) and because actors, their incentives and constraints vary across time, I propose to distinguish between the formal “given” prerogatives of institutions and the empirical capabilities they may acquire and that allow them to affect the political process.

This brings us to a second theoretical avenue (different from the innovative moment): that of institutional performance and its capacity to determine actors' behaviour and actions. Indeed, actors struggle over the definition of rules during moments of innovation, because they are aware that institutions are neither neutral (Steinmo, 1998) nor simple intervening variables, but rather canvasses of action and mechanisms for producing political results (March and Olsen, 1989; Pierson, 1994). This perspective does not conceal the importance of actors, their subjectivity and impact, nor does it pretend that institutions alone can explain everything (Pierson, 2001: 9–10). It does insist, however, on the strength of institutions and their capacity to impose specific political configurations and even structure people's beliefs and thinking (Hall and Taylor, 1997; Sultan, 1998: 47). Historical institutionalists have many ways of dealing with the source of this institutional strength. The simplest of these is that because institutions embody rules and norms, they function as structures which, in Giddens' words (1986), impose constraints and provide opportunities. Thus, the mere creation of institutions implies new stages for political conflicts and new distributions of power (Immergut, 1992: 85; Pierson, 1994: 31). I propose here to approach the way electoral commissions structure elections and democratization by focusing on the concept of path dependency (Levi, 1997; Dobry, 2000; Pierson, 2004) and, especially, on two of its complementary manifestations: self-reinforcing processes and lock-in effects. First, in this article, I use path dependency to refer to what Pierson calls “the dynamics of self-reinforcing or positive feedback processes in a political system.” This means that “once established patterns of political mobilization, the institutional ‘rules of the game’ and even citizens' basic ways of thinking about the political world will often generate self-reinforcing dynamics” (Pierson, 2004: 10). Then, once created, an institution “has a formidable capacity for its own reproduction across time” (Young, 1994: 2), and tends to buttress its power and persevere over time. Second, I use path dependency to explain lock-in effects, i.e., phenomena of resistance to change generated by pre-existing institutional structures and interests that are crystallized within them (Thelen, 2004: 33). Pierson (1994, also Pierson, 2000 and 2004) used this perspective to explain how Reagan and Thatcher failed to dismantle the welfare state in the US and Britain in the 1980s. In the US, the welfare state had produced lobbies, such as the American Association of Retired People, that mobilized to protect their own interests. Previous policies created incentives for actors who mobilized, thus preventing the Reagan administration from dismantling the welfare state. When self-reinforcing processes and lock-in effects combine—a phenomenon I will explore in some of the case studies—institutional autonomy is so important that, to borrow Young's words (1994: 2), they structure politics and are able to resist even “in the face of systematic efforts … to uproot prior forms and build new blueprints.”

Adopting such a historical institutionalist perspective allows me to go beyond a mere formalistic view of the electoral commissions and to provide a longitudinal, empirical and actor-centred explanation of their diverging levels of autonomy and efficacy.

Institutional Autonomy and Electoral Management Commissions: The Debate

In one of the few articles published on this subject, Shaheen Mozaffar (2002) offers a typology of African EMCs based on their different degrees of autonomy vis-à-vis the government. He proposes to measure autonomy according to four independent variables: the colonial institutional legacy; the political legacies of post-colonial regimes in terms of the frequency of previous elections; ethno-political fragmentation; and the presence or absence of political negotiations prior to the creation of the commission. He views the autonomy of EMCs mainly in legal terms, which he defines as their “constitutional and statutory provisions” (Mozaffar, 2002: 91–92).

His methodology is valuable, and he found high correlations between each independent variable and the dependent variable (autonomy). But his paper focuses on variables that go too deep. As Bratton and Van de Walle (1997) have stated, such variables are hardly able to explain empirical contemporary situations. The cases analyzed here do not follow the pattern Mozaffar describes. Therefore, this article is founded on different assumptions and hypotheses. First, with regard to the colonial legacy Mozaffar identifies as crucial, all of our cases, except Ghana, are francophone countries. Yet they present profound differences in electoral governance. Second, traditions of competition highly correlated with autonomy in Mozaffar's study run deeper in Senegal (one of the few African pluralist countries before 1990) and in Burkina Faso, yet these countries initially adopted an EMC that was less independent than that of Niger, a country that did not have this pluralist tradition. Third, and contrary to what Mozaffar's conclusions would suggest, Togo has the least independent commission but, at the same time, is the most ethnically fragmented country. Lastly, we need to go beyond the strictly legal view of the autonomy of EMCs that Mozaffar proposes if we are to find more encompassing and satisfying explanations. A broader analysis allows us to go beyond a perspective limited to the legal view and to establish a correlation between institutions and the trajectory of democratization.

Daniel Carpenter (2001: 17) provides perhaps the most precise definition of institutional autonomy in suggesting that “bureaucratic autonomy prevails when a politically differentiated agency takes a self-consistent action that neither politicians nor organized group interests prefer, but that they either cannot, or will not, overturn or constrain in the future.” I propose to search for the autonomy of EMCs by concentrating on two intertwined directions. First, autonomy certainly has something to do with legal provisions and, therefore, we must pay special attention to the organic relationships between institutions and governmental structures, especially the Ministry of the Interior, which is traditionally responsible for holding elections. But second and most importantly, we must adopt a more sociological approach to institutional autonomy. Institutional autonomy is empirical, a term that points out two different aspects here: on the one hand, it refers to the power relations in the critical moments of institutional forging, which in most instances determines its legal status. On the other hand, autonomy refers to the institution's capacity to enforce its rules.

A Legal Analysis of the Autonomy of EMCs

The creation of independent electoral commissions has now become a pattern in the emerging democracies of Africa, Eastern Europe and the former Soviet republics (Pottie, 2001: 133; Mozaffar, 2002: 90). Independent EMCs are viewed as efficient tools for fostering political participation and for lending legitimacy to electoral processes in a context of political renewal (Thiriot, 2004: 130–131). The study of the legal provisions of EMCs is necessary if one is to understand how these institutions function. But, at the same time, juridical regulations offer a static and idealistic view of the relationship between EMCs and government authorities that, especially in Africa's neo-patrimonial regimes, might be misleading. Legal provisions differ from effectiveness and, in the contemporary international environment in which democratic credentials have considerable weight, even repressive regimes tend to claim a democratic nature and concede formal powers they might never actually respect.

Legal autonomy depends on three criteria: organic separation from government (Mozaffar, 2002; Diarra, 2003), functional autonomy (Diarra, 2003), and the extent of the juridical control over the electoral process (Mozaffar, 2002; Diarra, 2004). These criteria vary from one country to the next and, in some of the countries studied here, variations can be found from one election to the next. Benin is the only country where these criteria have not varied significantly since the creation of the electoral commission in 1995 (Banegas, 2003: 2002). In Benin, the commission has always been independent and has always had total control over the entire process, from the compilation and revision of the voters' lists to the declaration of the results. In Niger, the commission has always satisfied the criteria of formal autonomy (from the creation of an independent national electoral commission in 1996 to the 2004 elections), except during the authoritarian rule of General Baré, who tried to weaken the institution, but succeeded only for a brief period in 1996. This authoritarian experience has paradoxically illustrated institutional dynamism. Despite attempts to dismantle it, the electoral commission (seen by opposition parties as the only acceptable means of electoral management) proved its ability to resist change, as evidenced by the massive mobilization of political parties and civil society in a call for its re-establishment (Mamoudou, 2000; Gazibo, 2004). Such a process of institutional lock-in and self-reinforcement has been described in the study of welfare state institutions (Pierson, 1994: 39–40), and proves that it is crucial to go beyond a merely juridical view of autonomy.

In contrast, Togo has had different versions of the electoral commission. The 1993 national electoral commission was all but autonomous. Its legal attributions included three aspects: controlling the availability of the elections' material; controlling the fairness of the process; and coordinating the observation of elections by foreign observers (Pilon, 1994: 141). Legally, compilation and revision of voters' lists, the production of electoral documents, the material organization of the elections, and the declaration of the results were under the control of the government (Africa South of the Sahara, 2002: 1059). In 1997, a new national electoral commission was created. Yet despite the fact that it was claimed to be legally independent (article 72), in reality it was not so, since the crucial attributions we have mentioned in the previous pages remained under governmental control. Even under such conditions, the government felt obliged to intimidate the members of the commission during the June 1998 presidential elections. Commission members resigned and the government exploited this vacuum to take full control of the process, and declared that Gnassingbe Eyadema had won the elections.

After these flawed presidential elections and the subsequent boycott of the legislative elections by the opposition, negotiations were scheduled in 1999 between the government and opposition parties. They agreed to create an independent electoral commission. However, as presidential elections approached, the Parliament, dominated by Eyadema's Rassemblement du peuple togolais (RPT), unilaterally decided to modify the consensus-based electoral law. Two points were particularly important. The limitation stipulating a maximum of two five-year terms was abandoned so that Eyadema could run indefinitely, and the commission's legal attributions were negated, for example regarding the presence of members of the opposition parties. As the US State Department put it: “Eyadema ran against four opposition party leaders and one independent candidate on June 1st, and his RPT party declared victory, claiming 57.22% of the vote. The election was marred by voters' inability to access the registration cards, and the government's failure to investigate allegations of irregularities, including intimidation of opposition party monitors and the stuffing of ballot boxes” (US State Department, 2004).

The cases of Cameroon and Senegal are less tumultuous and quite similar. Cameroon has indeed adopted a Senegalese model, but the more authoritarian nature of the Cameroonian regime accentuated governmental control, as in Togo. In Senegal, as Ould Ahmed Salem (2004: 150) has shown, none of the criteria were met. The Observatoire nationale des elections (ONEL) was created in 1996 and “simply charged … with supervising and verifying without actually being in charge either of the elections' material organization, or of the broadcasting the results” (Ould, 2004:150). In Cameroon, it was in December 2000 that the party in power and its allies approved a law permitting the creation of the ONEL by 126 out of the National Assembly's 180 representatives. Its legal attributions are numerable (17 attributions) and meet the criteria of juridical control over the electoral process. This includes the compilation and revision of voters' lists, the production of electoral documents, the control of the availability of the elections' material requirements, the control of the fairness of the electoral process, and so on. But organic and functional autonomy were completely absent. This is why five opposition parties boycotted the creation of the ONEL, denouncing the president's nomination of members and the total lack of financial autonomy and calling for the creation of a truly independent electoral commission.

Burkina Faso adopted a national commission of electoral organization (CNOÉ) in 1996. From a legal point of view, CNOÉ was not autonomous: its activities were carried out under the supervision of the Ministry of the Interior, its members were appointed by the government, and the control over the electoral process was limited. In 1997, important amendments were made to enhance the autonomy of the commission. Members are no longer simply appointed. Once designated (6 from the opposition parties, 6 from the government, and 14 from civil society), they elect their president and other members of the bureau. But as Englebert (1999: 218) states: “The compilation and revision of voters' lists were to remain under the control of the Ministry of Territorial Administration and Security.” The legislative elections held in May 1997 were reported to have been widely manipulated, since the presidential camp won 101 of the 111 parliamentary seats. The government was then free to propose, in 1998, the creation of an independent electoral commission, as presidential elections were scheduled for November. The main opposition parties decided to boycott the commission as well as the presidential elections in protest that the commission was not given control of the compilation and revision of the voters' lists or the production of electoral documents (Englebert, 2000). The third step in the evolution towards the creation of an independent commission was taken in April 2000. This time, the legal attributions of the commission were considerably expanded to encompass control over the whole process, including the electoral lists. The commission also has a complete organic autonomy (article 2 of electoral law). Even the composition of the commission confirms this tendency. With six members each, the government and the opposition parties are equally represented, while the trade unions, traditional and religious communities, and other civil society organizations shared the 15 remaining member positions.

In this search for EMC's legal status, Ghana is an atypical case, as a National Commission for Democracy was created in 1982 and charged with signing up voters, holding elections and defining the electoral districts. In 1991, an electoral commission was created. Much criticized by the opposition, which called the victory of outgoing president Jerry Rawlings in the general elections into question, it was replaced in 1992 by the national electoral commission (EC). The EC organizes, conducts and supervises all the elections and referenda. The Ghanaian commission has complete independence from the government. Its chair members have permanent tenure of office and are subject to the same conditions of service as Superior Court judges. Changes brought about by this new commission included the introduction of a computerized electoral roll and voter identity cards (Jeffries, 1998).

This brief overview of the legal autonomy of EMCs is useful in emphasizing the diverging juridical rules that govern electoral processes across time and space, especially the organic and functional positioning of the institution vis-à-vis the government and the extent of its control over the electoral process. Formal analysis is insufficient, however, because it offers only a static and official approach. Moreover, it captures neither the day-to-day process leading to this legal status, nor the real powers invested in the commissions. Finally, it does not inform us of the effectiveness of the rules governing the elections. Therefore, we need to push the analysis further.

An Empirical Search for the Autonomy of EMCs

An analysis of the political battles leading to the creation of EMCs—as stated by new institutionalists—is a better indicator of the extent of EMC autonomy. It not only reveals why the formal status of some EMCs is better in terms of autonomy, but also the extent of the real control EMCs gain over the electoral process. This non-codified control greatly affects the organizational performance in general (Nee and Strang, 1998: 709).

In fact, important differences may exist between formal and empirical prerogatives. Some EMCs may be formally independent but empirically powerless, while others may be supposedly subordinate but over time be able to strengthen their capacity far beyond legal prescriptions, a process of unintended consequences well known to new-institutionalists (Pierson, 1993; Cortell and Peterson, 2001). The empirical autonomy of EMCs is the result of power politics between political actors engaged in the political struggles that arise during the process of democratization.

Different configurations of power relations (depending on the resources each camp possesses and the degree of polarization in the political scene) create particular interactions among actors and determine whose political agenda will prevail. To grasp the different possibilities shaped by the symmetry or asymmetry of forces prior to institutional forging, I asked several questions (see also Karl and Schmitter, 1991 on modes of transition): Is the process unilateral (one actor prevailed) or multilateral (compromise between actors)? Who prevailed (incumbents or outsiders)? Or, in the case of a negotiated process, did one camp dominate?

Responding to these questions, using our case studies as examples, led me to propose five possible modes of institutional design exposed in Figure 1: the consensual; the incumbent-led interactive; the outsider-led interactive; the imposed; and the managed.

EMC Empirical Forging Modes

These modes are not always stable, nor do they always produce the same type of institutions. On the one hand, a mode may be co-constructed (interactive) at one point and then cease to be later on, as political circumstances and power relations fluctuate in the same country. On the other hand, we may expect these modes to lead to different degrees of institutional autonomy. As we will see in the empirical insights explored below, the imposed mode provides a greater level of autonomy from government, but is at risk of being controlled by outsiders. The outsider-led interactive mode is similar to the imposed mode in terms of autonomy, but reinvigorated incumbents might be tempted to reverse the situation. The opposite managed mode produces EMCs that are presumably weak and subordinate to the government. The incumbent-led interactive mode is very similar but the EMC may gain some autonomy when an incumbent's strength is altered. The consensual mode may produce autonomous EMCs, but may also be politicized if the political parties manage to expel civil society members from the institution, as illustrated by the mixed electoral commission of the Central African Republic. However, in the cases studied, this consensual mode provides a high level of autonomy and seems to be the most favourable to (or rather an indication of) democratic consolidation in the long run.

The outsider-led interactive mode of EMC forging has dominated in Benin and Niger, where it was initially imposed by outsiders. The interactive nature of the process is rare during the initial moments of democratic innovation, and happens only later. The first reason has to do with the absence of institutional learning, and notably concerns countries—like Benin and Niger—that are first to democratize, and go on to create and transform their institutions. Another reason common to all countries is the existence of an institutional inertia caused by a path-dependency phenomenon that opposes change in political routines. Ministries of territorial administration are usually responsible for holding elections and any attempt to innovate (especially to create an independent electoral commission) is seen as a politics of disempowerment that provokes concern and resistance.

In Benin, the National Conference held in February 1990 has come to be known as a laboratory for democratization in Africa, for it has led to profound institutional reforms, during which the creation of an EMC was not even discussed. In 1990, opposition parties and civil society coalesced and pressured President Kérékou to accept a multi-party system and elections. The conference elected a new prime minister, Nicephore Soglo, who became the dominant figure in the transitional government, granting Kérékou with honorific status only. Because the ancien régime was dismantled, there was no need for an independent commission. However, the situation changed four years later. The first national autonomous electoral commission was created by the Parliament in 1995, against the will of Soglo, president at the time. During his first term, his relationship with the other parties deteriorated and he began to show authoritarian tendencies (Banegas, 2003: 209). His opponents then began to be sceptical of his will to hold fair elections (Gbégnonvi, 1995: 61). The Parliament, dominated by opposition parties, decided to create an independent commission, a decision fiercely opposed by the government, which argued most notably that such electoral reforms are the government's, rather than the Parliament's, prerogative. Finally, the Constitutional Court, whose decisions are legally binding, decided that the Parliament does indeed have the legal prerogative to create an autonomous commission and that such an institution helps to secure free and fair elections and to protect civil rights (Ouinsou, 2000: 8). The creation of the EMC is the product of political struggle. Since then, interactive debates have existed during each moment of reform, but the attributions have never been questioned. The commission has gained strong legitimacy and become consensual. It is now difficult for any actor to erode its empirical autonomy, as the 2006 presidential elections have shown.

In Niger, the process has been less linear and suggests a greater importance of the balance of forces. The first democratic elections were organized by an electoral commission subordinate to the Ministry of Territorial Administration. The national conference, which imitated that of Benin, did not discuss the creation of an EMC. The point of departure was identical to that of Benin, and was characterized by the coexistence of a marginalized ancien régime and a powerful opposition that controlled the government. But as the first elections approached, the cohesion of the opposition parties began to erode. The Haut Conseil de la République, an institution that replaced the Parliament during the transitional period following the national conference, and which was fiercely opposed to the government, created a commission for the supervision of the elections (Maignant, 2000: 87). The March 1, 1993 elections were considered free and fair (Bratton and Van de Walle, 1997: 197) and were won by a coalition of parties holding the majority in the Parliament against the former single party. This coalition survived only 16 months, because Premier Mahamadou Issoufou resigned in September 1994 in a climate of political distrust and dispute with President Mahamane Ousmane. Issoufou signed a new alliance with the former single party. President Ousmane lost his parliamentary majority and called for new elections. In this polarized political situation, the outsiders were no longer ready to accept governmental control over the process. After many political battles, the opposing parties agreed to create a national electoral commission. From a legal perspective, the commission was not totally independent from the government. It was quite politicized, as its members came essentially from competing political parties. But as Mamoudou (2000: 24) puts it, the commission was so powerful that people used to talk about the “Independent Electoral Commission.” This is because at that point, an institutional self-reinforcing process was already in progress and became obvious later on, in July 1996. After the January 1996 military coup led by General Baré, the opposing parties decided to create an independent national electoral commission. This initiative was a simple formalization of the empirical autonomy gained by the previous commission, whose personnel were kept to serve in the new commission. This decision did not emerge from sharp-tongued debates, however, because on the one hand, the formerly divided political parties were now united against the candidacy of General Baré and, on the other hand, Baré was not opposed to the creation of the commission and created a managed National electoral commission because he was confident that he would win the July 1996 presidential elections. But when the results appeared to be dramatically unfavourable to Baré, the military junta dismantled the commission (Gazibo and Decoudras, 1997). It was only after two years of political battles between opposition parties and the military junta that a second independent commission was instituted in 1998 to manage the municipal elections of 1999. Those elections were well organized, but General Baré's partisans were unwilling to accept defeat. Violent actions were taken against the local commissions, yet this did not discourage committee members (CENI, 1999: 51). In April 1999, the Supreme Court decided the annulment of the elections in the majority of the circumscriptions won by the opposition parties. The situation resulted in political turmoil, a military coup led by Major Daouda Mallam Wanké, and the re-establishment of an autonomous electoral commission. This commission organized the general elections in the summer of 1999. After a brief attempt by the government to modify its legal provisions in 2003, a consensual agreement was finally reached and the commission was able to again organize free and fair elections in 2004.

The managed mode of building an EMC dominates in the cases of Burkina Faso and Togo. Such a situation is not surprising. In these countries, the power relations have been largely favourable to the government, with the exception of a few short periods. However, the empirical autonomy of the electoral commission appears to be better in Burkina Faso than in Togo.

Because of moments of political unrest (such as the assassination of journalist Norbert Zongo) that have allowed civil society to pressure (and weaken) the Compaoré regime (Ouédraogo, 1999: 163), the EMC in Burkina Faso evolved gradually towards more autonomy even if, paradoxically, the process remained largely unilateral. The first demands for a fairer electoral process were made in 1991. As the first general elections approached, the then national union government disintegrated. The main opposition leaders decided to boycott the elections to protest the government's unilateral adoption of the electoral law. A few months earlier, these leaders had failed to force the Compaoré regime to hold a national conference, such as those that had taken place in Benin, Niger and elsewhere. The presidential and legislative elections were easily won by the incumbent regime. As Englebert (1995: 154) points out, “Compaoré's opponents alleged widespread malpractice” and called for a new and fairer electoral law.

The revision of the electoral law occurred after the promulgation of the Law of November 14, 1996, which allowed for the creation of the CNOÉ. The opposition parties wanted the commission to be independent, whereas members of the majority-led government, who prevailed, simply wanted the commission to have some autonomy while still under the control of the Ministry of Administration and Territorial Security. All this was taking place at a time when the social climate was deteriorating. In December, students started a long strike in the capital city of Ouagadougou that would profoundly affect the Burkinabé government. According to Christopher Wise (1998), the violent repression that the protesting students encountered attracted sympathy from the population, the media and workers' unions. The workers' confederation and the human rights organization supported the strikers. Burkina was at the brink of a social explosion.

Only to calm the tensions (which were exploited by the opposition), the Parliament, dominated by Compaoré supporters, decided to change electoral laws. Without granting independence to the CNOÉ, the law allowed the election of the commission's president and deputy commissioners by their peers. They were no longer simply appointed by the government. Article 37 of the Constitution, limiting the number of mandates to two, was abolished, but the opposition rejected the new law. The 1997 legislative elections were won by the incumbent party, which obtained 101 out of the Parliament's 111 seats and established an independent electoral commission without the opposition's participation.

The commission was composed of 26 members. Six came from the presidential majority, six from the opposition, and the remaining 14 from civil society, including unions, religious groups and traditional chiefs (Englebert, 2000: 225). But the members were not in charge of the establishment of the electoral lists or the distribution of cards. The opposition parties radicalized their position by boycotting this commission and the elections.

President Compaoré began a seven-year term that went unchallenged by the weakened opposition. But the assassination of journalist Norbert Zongo in 1998 provoked an unprecedented mobilization throughout the country, which threatened to overthrow the regime. In order to calm tensions, Compaoré appointed a “group of wise people” who were to recommend institutional reforms. The opposition refused to participate. In April 2000, the National Assembly adopted all the proposed modifications of the electoral laws. From then on, the mission of the electoral commission was redefined to cover “the organization and supervision of all electoral and referenda operations.” The commission was said to enjoy “autonomy in all matters of organization and functioning.” Its tasks range from supervision and establishment of voter lists and electoral cards, to declaring the provisional results and forwarding them to the Supreme Court.

The opposition was weak and plagued by its own divisions. However, the 2000 commission emerged from a power politics situation, with civil society playing an important role. This explains its astonishing degree of legal independence, in spite of the fact that the commission had organized only one presidential election, which took place in November 2005. In contrast to the previous elections, opposition parties took part in this election, a crucial test to see whether there were strong challengers. Compaoré let the commission free, but managed to weaken the opposition leaders by using two main strategies of division and corruption. Notably, he gave financial “support” to some of the opposition leaders. From the moment the story was “opportunistically” revealed during the campaign, the opposition parties began to criticize each other and the public's confidence in them fell dramatically. It was against 11 weak and divided candidates from the opposition that Compaoré won the elections, with 80.35 per cent of the votes. We can analyze this situation in two ways: either we look at the result and interpret it as a failure of democracy, or we consider the changing strategies of President Compaoré as imposed by the institutional constraints. As it is costly to dismantle or openly violate the electoral commission, even powerful actors, like Compaoré, have to devise other ways of winning.

In contrast, in Togo, the authoritarian regime of the late President Éyadéma (he passed away in February 2005) never faced serious challenges, except during the early years of political transition, when the national conference proclaimed its sovereignty, elected a new prime minister and scheduled elections (Nwajiaku, 1994: 440). But Éyadéma rapidly regained confidence as his military supporters began a widespread campaign of violence and killings against opposition leaders (Pilon, 1994: 137). In December 1993, the military attacked Premier Kokou Koffigoh's office, kidnapped him and brought him to Éyadéma's palace. A few hours later, he appeared on national television and announced a new government, dominated by the old guard of the president's supporters. Even before the first elections, the game was already over. Since then, the electoral process has been a long story of abuses and intimidations.

The first electoral commission was instituted in 1993, after a long round of negotiations between the government and opposition leaders in Ouagadougou. Facing a weak and divided opposition, the government prevailed, maintaining control over voter registration and the publication of electoral lists (Africa South of the Sahara, 2002: 1059). As Pilon (1994: 142) states, the CÉNI seems to have limited its mission to observation and has left the rest—the preparation of the electoral lists, the organization of elections, and the declaration of the results—to the government. After pushing the opposition to a boycott, manipulating the elections, and influencing the Constitutional Court, Éyadéma “won” with 96.49 per cent of the votes and also took the majority in the Parliament.

Subsequently, new battles took place around the electoral commission. But, as Jennifer C. Seely (1998) points out, “the opposition had little influence on important parliamentary decisions, including the composition of the Constitutional Court and the National Electoral Commission that would govern the next elections.” In fact, a new commission was created in 1997 to organize the upcoming elections, which were held in 1998. The commission was considered independent, but registration on the electoral lists, the preparation, and then the distribution of electoral cards remained entrusted to the Ministry of Territorial Administration. Moreover, two days after the presidential elections of June 1998, the commission's president and four members who represented the government resigned, as if at random. With this provoked paralysis of the commission, the Minister of Territorial Administration announced that his office would undertake recounting the votes and checking the ballots. Éyadéma was once again declared winner of the presidential elections, with 52.13 per cent of the votes. The legislative elections of March 21, 1999, annihilated opposition parties. As a result of having boycotted the elections, they were no longer represented in the National Assembly. Éyadéma's RPT won 77 out of the 79 seats.

President Éyadéma was unchallenged when he initiated an inter-Togolese dialogue with the opposition parties in 1999. Thanks to the mediation of the four “facilitators” from the European Union, the International Organization of the Francophonie, France and Germany, an agreement was reached, but it was only adopted by the Parliament after considerable resistance. The parties planned the creation of an independent national electoral commission, which would organize, supervise and control elections. The president also promised to respect the two-term mandate limit (article 59 of the Constitution). But once again, just before the presidential elections of 2003, General Éyadéma, confronted with a weakened and divided opposition, violated this agreement. On February 8, 2002, the Parliament dominated by his party unilaterally modified 34 articles of the electoral law. With regards to the independent commission, new provisions stated that decisions would, from now on, be taken by a simple majority, and no longer with a four-fifths majority. Its numbers were reduced from 20 to 10 members, five from the majority and five from the opposition. The new code also stipulated that if the independent commission was unable to fulfill its tasks, the Supreme Court (nominated by Éyadéma) would appoint a committee of five magistrates to replace it. Obviously, the agreements became obsolete and Éyadéma again “won” the June 2003 presidential elections with an official score of 57.78 per cent. It is now clear, following Éyadéma's death, that a truly autonomous and effective commission did not emerge. Faced with reprobation from the international community for the unconstitutional military nomination of Faure Éyadéma to replace his father, incumbents went as far as to change the Constitution in order to appoint a sympathetic temporary president, government and institutions. It was therefore an all but autonomous commission that declared Faure Éyadéma elected, with more than 60 per cent of the votes in the April 2005 elections. The incumbents are so strong and the institutional forging process so unilateral that the electoral commission will remain managed and its autonomy null as long as the power relations remain unaltered.

The incumbent-led interactive mode of EMC forging is rare. This is logical, given that, as Bratton and Van de Walle (1997: 84–86), and Lindberg (2001: 185–186) have shown, in Africa, neo-patrimonial leaders try to remain in control as long as they can. They hardly initiate political reforms, which usually occur only in response to pressure from below. In these regimes, characterized by “winner takes all” politics, pacts are unlikely and both insiders and outsiders of the system struggle over rules and institutions in an instrumental way (Bratton and Van de Walle, 1997: 88). In an incumbent-led interactive process, one may expect the resulting commission to be weak and subordinate to the government, as the Cameroonian 2004 Office national des élections (ONEL) shows. The incumbent president Biya reacted to the opposition parties' claims for an independent electoral commission by calling for the creation of the ONEL. Nonetheless, the extent of its legal status and the appointment of its members remained totally under his control. In Senegal, as Vengroff and Magala (2001:132) stated, “the setting up of an independent national election commission (ONEL) to observe and assess the fairness of elections could not have taken place without Diouf's personal support and the mediating role he played between the various political parties, including his own.” However, the Senegalese case is different from the others, since the country has been pluralist since 1976 even if, as Ould Ahmed Salem (2004: 150) has shown, elections have always been flawed. By the mid-nineties, opposition parties asked for an independent commission because the ONEL was under the direct control of the head of state.

However, because institutions have an existence of their own, autonomy may be achieved even in the incumbent-led interactive mode. Senegal's ONEL has demonstrated this through the defeat of the incumbent president and the 2000 political turnover (Diop et al., 2000). Even in the rather managed case of Cameroon, incumbents were forced to change the electoral law in order to secure victory. Facing a fragmented opposition, they enforced a “first-past-the-post” rule in order to prevent President Biya from going on to a second round, which allows opposition parties to unify against the incumbent.

The case of the electoral commission of Ghana also confirms that once established, even incumbent-led interactive commissions may have a dynamic of their own. Since 1996, the successive presidential and legislative elections have been free and fair despite initial fears that the Rawlings Regime refuses to concede defeat (Gyimah-Boadi, 2001; Nugent, 2005).

The consensual mode has been observed only in the later stages of institutional existence. It occurs when the democratic process begins to become institutionalized and political struggles over rules are routine. This is why even Benin's EMC is characterized as a consensual one only from 2001 onward, Ghana's EMC since the 2000 alternation and Niger's EMC since the last elections of 2004.

Conclusion: EMCs, Institutional Autonomy and Democratization

Several conclusions emerge from this comparative study. First, the degree of autonomy of electoral management commissions is mainly in the balance of power, as the contrasting cases of Benin and Togo illustrated. As Carpenter (2001: 15) points out, to be autonomous an institution needs to meet several conditions, including political differentiation. When a commission rises out of circumstances in which no political wing is powerful enough to control the whole process, the chances are greater that this commission will enjoy a more substantial degree of autonomy, as the case of Benin illustrated. When incumbents remain strong despite ongoing political liberalization (as in Burkina Faso and Cameroon) and, especially, when they use systematic violence to crush their opponents (as in Togo), commissions may exist and even have important formal prerogatives. Yet, they are certainly not truly autonomous and definitely not effective.

Second, a commission's status changes as the balance of forces evolves. Burkina Faso best illustrates the vital importance of the balance of forces in determining the degree of autonomy of EMCs, and this balance's fluctuating nature. There, the previously unchallenged Compaoré regime agreed to create an autonomous commission, but only after it had been faced with threats from widespread mass mobilizations. It may have only been a tactic devised to avoid the collapse of the regime. Nonetheless, institutions have an existence of their own (Skocpol, 1985; Pierson, 2004), and they tend to crystallize socio-political interests and create lock-in effects (Steinmo, Thelen and Longstreth, 1992: 9; Pierson, 1994). In other words, though there is no guarantee, there are chances that the commission's autonomy will be confirmed over time, in spite of the fact that this regime is unwilling to accept defeat.

Third, and in the same vein, the comparative processes of building EMCs tend to confirm the validity of the previous hypothesis about the existence of an institutional self-reinforcing process. This is particularly true in cases of institution building using the interactive or imposed modes. The case of Niger provides the best illustration, to borrow Pierson's (1993) words, because there “the effects [i.e., the electoral commission] became the cause.” In other words, once an institution has been established, it will eventually increase its ability to affect its environment, therefore fostering its own autonomy. To paraphrase Crawford Young (1994: 2), we can say that once institutionalized, an EMC gains a huge ability to reproduce across time, even in the face of systematic assaults by authoritarian rulers trying to weaken it. At some point, a path-dependent process takes place and it becomes very difficult for would-be authoritarian leaders to revert the process, as most of the cases (especially Benin and Niger, but also Senegal and Ghana) prove.

Finally, the emphasis on the bases of autonomy and on the modes of building EMCs highlights why commissions (in Niger and Benin) can shape their environments and reinforce the democratic process, or remain ineffective in the electoral process and even go so far as to legitimize questionable and disputed situations (as in Togo or Cameroon). At this point in the discussion, it seems plausible to assert that a correlation exists between the process of forging EMCs, their level of autonomy and the course of democratization. It is not surprising that in the Freedom House Ratings (see appendix below), Togo and Cameroon have been consistently classified as “not free,” while Niger and Burkina Faso have been mostly designated “partly free.” It is not surprising that Benin, which has always had an autonomous EMC, has been consistently classified as “free” since 1993 and that Ghana has been classified as “free” since the 2000 presidential turnover, which made clear that the electoral commission has buttressed its autonomy and was capable of holding free and fair elections. This topic invites further research, as does the question of the larger possible impacts of the electoral commission's autonomy.

Of course, electoral commissions are not the only variable affecting democratization. Many scholars warn that institutions are neither isolated factors nor variables that work in isolation (Pierson, 2001: 9; Pierson and Skocpol, 2002: 707). Yet it is obvious that holding free and fair elections in Africa—and thus, the very quality of electoral democracy—depends largely on the existence of an autonomous electoral commission, independent of political parties and able to organize legitimate elections which, although not sufficient, constitute the foundation of democracy.