This is a post by contributor Yonatan L. Morse, based on his article ‘Presidential Power and Democratization by Elections in Africa’ that will be published in the journal Democratization
In traditional studies of democratization, elections are generally the end phase of a prolonged process of liberalization and political opening. However, in recent years political scientists have also entertained the idea that elections might actually be the starting point of a process of democratization. In foundational work on Africa by Staffan Lindberg, he contended that repeated consecutive elections could create self-reinforcing mechanisms that deepened democracy over time. This approach is intuitively appealing for an era in which elections are commonplace, yet many countries still fail to live up to democratic standards. And expectedly, this thesis has been subject to quite widespread replication, scrutiny, and criticism.
In new research, now published online by the journal Democratization, I engage with the democratization by elections thesis in Africa, and argue that repeated elections can induce some forms of democratic behavior but face real limitations when formal presidential powers are strong. This is because under certain conditions strong presidentialism reinforces incentives for elections to become opportunities for clientelistic exchange, rather than moments of self-expression. Powerful presidents that control legislative agendas, access to political appointments, and the purse strings, might lead certain actors to behave more democratically during elections, but not necessarily to develop more robust notions of citizenship. This holds true in Africa because levels of economic development and inequality reinforce the role of clientelism as a central way elites and citizens access their government.
A caveat is in order here first. If the democratization by elections thesis has been so heavily scrutinized (in Africa and elsewhere), what is there to add to the debate? Other studies have generated, at best, mixed results. For instance, in Latin America democracy was restored in the 1980s after periodic interludes of authoritarianism. Therefore, many of the indicators of democracy simply jumped back to their prior levels, and have in fact declined since in many countries. Most importantly, in many countries repeated elections seemed to reinforce rather than undermine authoritarianism. Referred to as electoral or competitive authoritarian regimes, in these cases repeated elections appear to offer incumbents the ability to reshuffle their coalitions, gather information about their levels of support, and generate international legitimacy. In one study of Africa, the authors found that democratization by elections could only truly be found in a handful of cases.
The problem with previous studies is that they often mischaracterize what the democratization by elections thesis is actually about. Lindberg makes a crucial distinction between the “process of democratization” and a “transition to democracy.” Regimes can show improvements in specific indicators of democracy, while not necessarily transitioning to a new regime. Indeed, autocratic regimes can exhibit more or less democratic behavior. For instance, when actors participate more, compete more effectively, or appear to accept the legitimacy of the election process, this is a sign of democratic progress. Specifically, for Lindberg this is evidence of how elections create self-fulfilling expectations. Elections might also lead to improvements in other realms of democratic life like the protection of civil liberties. This indicates some form of socialization by elections, whereby citizens learn from election experience to demand voice in other realms of life. Using this more limited definition of democratization yields quite different results from previous studies.
My contribution is therefore to stress which factors condition the impact of repeated elections on much more specific democratic outcomes. I gathered information on 679 African elections since 1990, and combined this information with data from the Varieties of Democracy (V-DEM) and Presidential Power (PRESPOW) datasets. These data offer new ways to explore both numerous democratic outcomes, and to compare and contrast the extent of formal presidential power across Africa. The V-DEM data includes measures of electoral participation, competition, and legitimacy. But, it also includes indexes of many non-electoral elements of democracy like the protection of civil and private liberties, civil society participation, and equal protection under the law. I controlled for numerous other factors like executive years in office, levels of economic growth and development, foreign aid, ethnic heterogeneity, religion, and historic experiences with democracy.
A key utility of this study is its use of formal measures of presidential power in Africa. In many studies of African politics the focus has often been on the various ways in which presidents violate constitutions and operate through parallel informal institutions. This approach is mistaken for a number of reasons. First, it is equally clear that African presidents routinely amend constitutions, which means that the formal powers of presidents are not trivial. Second, using formal measures of presidential powers limits the risk of endogeneity in a study. For example, if a president unconstitutionally repeals legislation, this is often coded as both a violation of the democratic process and stronger informal presidential power. It is difficult to know what factor is influencing what factor. By focusing on the formal attributes of presidents, this risk of conflation is mitigated.
The analysis shows that improvements in the election process do not depend on levels of presidential power. Using Lindberg’s criteria, with more experience African elections become more participatory, competitive, and legitimate. This validates the notion that elections reinforce actors’ expectations and conditions them to accede by the rules of the game if they want to get ahead. On the other hand, presidential power significantly conditions the impact of repeated election on civil and private liberties, civil society participation, and equal protection under the law. When presidents are formally strong, repeated and consecutive elections limit the ability of elections to socialize more participatory and democratic behavior. These results hold up to various statistical models, and even the inclusion of a measure of the unfairness of the election.
This corresponds with expectations regarding the intersection of presidential power and clientelism in Africa. When levels of access to a system of spoils define the political game, and when presidents control that access, elections become devoid of deeper civic meaning. Rather, actors decide to accept electoral processes because fighting the system keeps them excluded. These results do not reject the democratization by elections thesis, but rather shed light on its limitations. Moreover, it also corroborates that the problem of democratic progress is not only due to the fact that elections themselves are unfair. In many cases the playing field remains heavily tilted toward incumbents, but clientelism and powerful presidents exist in diverse settings and exert an independent impact on democratic outcomes. It is not enough to just get the elections right, the disproportionate formal powers of presidents need to be tempered too.