Presidential power and democratization by elections in Africa

ABSTRACT

Do elections in and of themselves provide mechanisms for democratization? The “democratization by elections” thesis has been challenged, yet scholars still differ over its substantive effect. Some of the disagreement is over the specific outcome of interest, with proponents advocating for a narrower definition of “democratization”. Others want to know more about the factors that condition how elections impact on democracy. This article addresses both points by demonstrating that in Africa the extent of formal presidential power significantly shapes the ability of repeated elections to socialize more broadly democratic behaviour in the form of greater civil and private liberties, more civil society participation, and wider egalitarianism. Using recently available data on African presidents and the democratic qualities of regimes, the article demonstrates the ongoing influence of presidential power in Africa and provides some previously unstudied constraints on the democratization by elections thesis.

KEYWORDS:

• Elections
• Africa
• presidential power
• socialization
• electoral autocracy

Disclosure statement

No potential conflict of interest was reported by the author.

Notes

1 Schedler, The Politics of Uncertainty.

2 Platner, “Is Democracy in Decline?”

3 Edgell et al., “When and Where do Elections Matter?”; Lindberg, “The Power of Elections in Africa Revisited”.

4 Arriola, “Patroange and Political Stability in Africa”; van de Walle, “Presidentialism and Clientelism in Africa’s Emerging Party Systems”.

5 Rustow, “Transitions to Democracy”.

6 Lindberg, Democracy and Elections in Africa; Lindberg, “The Power of Elections in Africa Revisited”.

7 Bratton, “Second Elections in Africa”.

8 Lindberg tested the impact of repeated elections on civil liberties by measuring the change in Freedom House’s civil liberties score in the years immediately preceding an election. Furthermore, he regressed the civil liberties score in 2006 against the number of cumulative elections, while controlling for economic development and performance, religious and ethnic diversity, popular mobilization, international pressure, and share of oil exports in the economy.

9 Di Palma, To Craft Democracies; Huntington, The Third Wave.

10 Jensen and Wantchekon, “Resource Wealth and Political Regimes in Africa”.

11 Bratton and van de Walle, Democratic Experiments in Africa.

12 Keefer and Knack, “Does Social Capital Have an Economic Payoff?”

13 McCoy and Hartlyn, “The Relative Powerlessness of Elections in Latin America”.

14 Kaya and Bernhard, “Are Elections Mechanisms of Authoritarian Stability or Democratization?”

15 Lust-Okar, “Legislative Elections in Hegemonic Authoritarian Regimes”.

16 Teorell and Hadenius, “Elections as Levers of Democratization”.

17 Roessler and Howard, “Post-Cold War Political Regimes”.

18 Levitsky and Way, Competitive Authoritarianism Hybrid Regimes after the Cold War; Schedler, The Politics of Uncertainty.

19 Blaydes, Elections and Distributive Politics; Magaloni, Voting for Autocracy Hegemonic Party Survival and Its Demise in Mexico.

20 Gandhi and Lust-Okar, “Elections under Authoritarianism”.

21 Geddes, Why Parties and Elections in Authoritarian Regimes.

22 Bunce and Wolchik, Defeating Authoritarian Leaders in Postcommunist Countries.

23 Howard and Roessler, “Liberalizing Electoral Outcomes in Competitive Authoritairan Regimes”; Donno, “Elections and Democratization in Authoritarian Regimes”.

24 Brownlee, “Harbinger of Democracy”; Miller, “Democratic Pieces”.

25 Lynch and Crawford, “Democratization in Africa 1990–2010”.

26 Cheeseman, “African Elections as Vehicles for Change”.

27 Bogaards and Elischer, “Competitive Authoritarianism in Africa Revisited”; Bogaards, “Reexamining African Elections”.

28 Lindberg, “Confusing Categories, Shifting Targets”.

29 Lindberg, Democracy and Elections in Africa.

30 Edgell et al., “When and Where do Elections Matter?”

31 Bratton, “Where Do Elections Lead in Africa?”

32 Bratton and van de Walle, Democratic Experiments in Africa; Chabal and Daloz, Africa Works; Jackson and Rosberg, Personal Rule in Black Africa.

33 Arriola, “Patronage and Political Stability in Africa”; van de Walle, “Presidentialism and Clientelism in Africa’s Emerging Party Systems”.

34 Rakner and van de Walle, “Opposition Parties and Incumbent Presidents”.

35 Kitschelt, “Linkages between Citizens and Politicians in Democratic Politics”.

36 Wantchekon, “Clientelism and Voting Behavior”.

37 Takougang and Krieger, Africa State and Society in the 1990s.

38 Studies show that presidential power varies significantly across political context (Metcalf, “Measuring Presidential Power”; Siaroff, “Comparative Presidencies”) and that some measures of formal presidential power are associated with weaker party systems and a lower number of effective presidential candidates (Hicken and Stoll, “Are All Presidents Created Equal?”; Elgie et al., “Proximity, Candidates, and Presidential Power”).

39 Ekeh, “Colonialism and the Two Publics in Africa”.

40 Posner and Young, “The Institutionalization of Political Power in Africa”.

41 Barkan, “African Legislatures and the ‘Third Wave’ of Democratization”.

42 Van Cranenburgh, “‘Big Men’ Rule”.

43 Doyle and Elgie, “Maximizing the Reliability of Cross-National Measures of Presidential Power”.

44 There is some disagreement in the literature over how to measure presidential power. One approach examines constitutions and looks at the actual powers that presidents have (Frye, “A Politics of Institutional Choice”; Mainwaring, “Presidentialism, Multipartism, and Democracy”). However, others examine the problem from the perspective of executive constraint (Beck et al., “New Tools in Comparative Political Economy”), or the number of checks and balances (Henisz, “The Institutional Environment for Infrastructure Investment”). A third approach looks at legislative power as the primary constraint on executives (Fish, “Stronger Legislatures, Stronger Democracies”).

45 Metcalf, “Measuring Presidential Power”.

46 There is also a second index (PRESPOW2) that employs principal component analysis that underweights attributes of presidential power that vary significantly from others. The PRESPOW2 data are tested in the supplemental materials (Appendix D).

47 Van Cranenburgh, “‘Big Men’ Rule”.

48 The PRESPOW data do not cover the following multiparty African countries: Chad, Equatorial Guinea, Ethiopia, Lesotho, Liberia, and Rwanda.

49 Another advantage of the V-DEM data is that they include a larger timespan than Lindberg’s original study, which generally ended data collection in 2006. Since 2006, military coups have ended regular electoral practice in a number of countries: Burkina Faso, Central African Republic, Côte d’Ivoire, Guinea, Madagascar, Mali, Mauritania, and Niger. Further, many countries that were unstable in the original study have held second and in some cases third consecutive elections, as in Angola, Burundi, the Democratic Republic of the Congo, Sierra Leone, and the Republic of the Congo. Finally, there are now far more countries that have reached the critical fourth election threshold. In 2006, only 14 countries had held at least four consecutive elections; by 2015 that number had increased to 28.

50 Beck et al., “Taking Time Seriously”; Steenbergern and Jones, “Modeling Multilevel Data Structures”.

51 To test the impact of model selection the supplemental materials (Appendix F) include results from regression analysis which look only at civil liberties protection using a standard ordinary least squares (OLS) estimation, a fixed effects estimation, a dynamic panel estimation using the generalized method of moments (GMM), and a Hausman–Taylor model with a lagged measure of civil liberties as an endogenous variable.

52 As is common procedure, the growth model includes the measure of time and time squared in the fixed part of the model, but only time in the random part of the model. Likewise, the models assume an unstructured covariance structure.

53 Rabe-Hasketh and Skrondal, Multilevel and Longitudinal Modeling Using Stata.

54 The Freedom House data have frequently been criticized over opaque aggregation methods and unclear conceptual categories (Munck and Verkuilen, “Conceptualizing and Measuring Democracy”). Recently Freedom House began publishing disaggregated scores, which allows researchers to make finer-tuned distinctions in the effects they aim to study. Unfortunately, there are no disaggregated data from Freedom House prior to 2006.

55 In the original democratization by elections study Lindberg made a distinction between consecutive executive and legislative elections. This meant that he counted cumulative executive and legislative elections separately. There were theoretical reasons for this, mostly to do with the elite learning effects. However, in the most recent iteration of this research this distinction is no longer made.

56 The economic data are from the World Bank’s Development Indicators dataset (World Bank, World Bank Development Indicators).

57 Data on executive years in office are from the World Bank’s Database of Political Institutions (Beck et al., “New Tools in Comparative Political Economy”).

58 Posner, “Measuring Ethnic Fractionalization in Africa”. The data on religion are from the Association of Religion Data Archives (thearda.com). In the supplementary materials (Appendix E) I use Alesina et al.’s index of ethnic fragmentation (Alesina et al., “Fractionalization”) as an alternative measure of ethnic diversity. Generally, Posner’s index of politically significant groups reports much lower levels of fractionalization.

59 Alesina et al., “Public Goods and Ethnic Divisions”.

60 Fish, “Islam and Authoritarianism”.

61 Gerring et al., “Democracy and Human Development”. The Freedom House civil liberties index and political rights index, which captures more straightforward indicators of democracy, correlate across the sample by 88%. Therefore, Freedom House is not a useful predictor for the model. The Polity IV score offers data from outside Freedom House and V-DEM and addresses the historical record of democracy prior to the inauguration of elections in the early 1990s.

62 These simulated results were produced using Stata 15 and the margins command.

63 This index has a −0.16 correlation with the PRESPOW index, indicating that the two are not strongly related. Weaker presidents often manipulate elections, while stronger presidents often allow more competitive elections.

64 Diamond, “The Rule of Law versus the Big Man”.

Additional information

Notes on contributors

Yonatan L. Morse

Yonatan L. Morse is Assistant Professor of Political Science at the University of Connecticut. He researches the intersection of authoritarian politics, elections, and democratization in sub-Saharan Africa. He is currently working on a book titled How Autocrats Compete that compares the role of ruling parties during electoral authoritarian contestation in Africa.