publisher colophon

The early 1990s saw a wave of competitive multiparty elections in Africa. These contests can be described as “founding” elections in the sense that they marked for various countries a transition from an extended period of authoritarian rule to fledgling democratic government. By the middle of the 1990s, this wave had crested. Although founding elections continued to be conducted in African countries that were latecomers to the political-reform bandwagon, they took place less frequently than earlier in the decade. Meanwhile, in countries that had experienced early regime change, expiring electoral cycles gave rise to a groundswell of “second” elections. Less glamorous than the landmark contests that gave birth to democracy, these events nevertheless held out the possibility that democratic routines might be deepened.

The consolidation of democracy involves the widespread acceptance of rules to guarantee political participation and political competition. Elections—which empower ordinary citizens to choose among contestants for top political office—clearly promote both sorts of rules. But analysts do not agree on the role that elections play in the consolidation of democracy. Some, like Samuel P. Huntington, use electoral criteria for measuring consolidation: the so-called two-turnover test. 1 Against such an approach, Terry Karl has raised the specter of a “fallacy of electoralism.” 2 As experience with “illiberal” [End Page 51] democracies shows, elections can coexist with systematic abuses of political rights and the disenfranchisement of much of the population. 3

I hold a middle view in this debate: while seeking to avoid the electoral fallacy, I try not to commit its antithesis—what Seligson and Booth call the “anti-electoralist fallacy” 4 —by assuming that elections never matter for democratization. I recognize that elections do not, in and of themselves, constitute a consolidated democracy. This end-state also requires civil rights and due process of law; checks on arbitrary executive power; civilian control of the military; and an independent press and civil society. In a consolidated democracy, citizens and politicians alike accept that this array of institutions is the only legitimate arrangement for governing public life.

But while elections and democracy are not synonymous, elections remain fundamental, not only for installing democratic governments, but as a necessary requisite for broader democratic consolidation. The regularity, openness, and acceptability of elections signal whether basic constitutional, behavioral, and attitudinal foundations are being laid for sustainable democratic rule. It is meaningful to study elections for the simple reason that, while you can have elections without democracy, you cannot have democracy without elections. If nothing else, the convening of scheduled multi-party elections serves the minimal function of marking democracy’s survival. The most immediate concern for many of Africa’s fragile new democracies is whether they will endure at all. By recording the occurrence of a second competitive election, we can at least confirm that democratic gains have not been completely reversed by executive fiat or military coup.

In assessing second elections, the empirical tasks are straightforward. The first concerns electoral quantity. Are second elections held? And if so, do they take place on time? The answers here help ascertain how strictly officeholders subject themselves to the rule of law. Incumbents who respect electoral schedules (rather than illegally altering election timing to increase their chances of holding on to power) acknowledge that good governance requires observance of at least some constitutional constraints.

Next come questions of electoral quality. Exactly how free and fair are second elections? Definitive judgments are difficult whenever the quality of elections varies across different stages of the process. For example, flawed voter-registration exercises or highly unfair campaigns may be followed by relatively open and free balloting. But to the extent that observers report gross deficiencies at any stage of an election, its integrity can be called into dispute. In addition, one must consider whether an election is boycotted by opposition parties. Whereas widespread involvement by various political parties probably indicates the absence of major electoral deficiencies, a boycott seems to signal a lack of agreement on the rules of the democratic game. Yet the quality of [End Page 52] boycotted elections can be ambiguous; we should remain alert to the possibility that a boycott, rather than reflecting a flawed electoral process, can be a ruse by opposition parties that have concluded that they stand no chance of winning.

This brings us to a third set of questions concerning electoral meaning. Because political power involves intangible elements like the legitimacy of government, elections are contested in symbolic as well as empirical terms. In new democracies, where elections are not yet fully institutionalized, contenders vie to win votes and seats but they also struggle to control the interpretation of outcomes. Elections that result in regime transition or leadership alternation are usually unequivocal, signifying a break with the past. Those in which incumbents retain power are harder to interpret because they involve judgments about whether the vote enhances or reduces a sitting government’s mandate. In part, an election’s legitimacy can be gauged by objective indicators such as the winner’s shares of votes or seats. But even these are subject to interpretation, for example in the light of turnout rates or considerations of campaign context and conduct.

In Africa, disputes about the meaning of multiparty elections tend to be directed externally at foreign aid donors, especially when funding decisions hinge on a verdict of “free and fair.” Because the stakes are high—the continuation of balance-of-payments support or project aid—disputes over interpretation come to the fore after the election. Winners deploy the informational and coercive instruments of the state to reinforce their claim to have received a mandate, while losers try to undermine this assertion by arguing that the process was rigged. The competition to assign meaning takes place before various audiences—elite and mass, domestic and international—that may apply different standards of judgment. Thus analysts are forced to interpret diverse claims of electoral quality, understanding all along that the contending parties seek to interpret the meaning of elections in order to promote their own interests.

Founding Elections, 1989 to 1997

In Democratic Experiments in Africa, Nicolas van de Walle and I documented the nature of “early” founding elections in Africa. 5 These events were widespread, with 54 elections occurring in 29 countries during a brief five-year period of transition from 1990 to 1994 (see Table 1). More than half of these early founding elections—30 out of 54—reflected the will of the electorate, inasmuch as reputable election observers ruled them free and fair. They were marked by relatively high turnout (averaging 64.1 percent of registered voters) and convincing victories (with winners averaging 63.4 percent of valid votes in presidential elections). Most importantly, in a momentous [End Page 53] break with past patterns of leadership succession in Africa, these contests resulted in the peaceful ejection of sitting presidents in 11 countries (plus three more turnovers where incumbents declined to run). In sum, the postcolonial generation of African political leaders proved unable to survive truly democratic elections.

Table 1

Trends in Founding Elections in Sub-Saharan Africa, 1989–97

Chronology of Election No. of Countries No. of Elections Opposition Boycott1 Free and Fair2 Leadership Alternation3 Losers Accept4 Voter Turnout5 (% Reg. Voters) (mean) Winner’s Vote Share (Pres. Elecs.) (mean) Winner’s Seat Share6 (Pres. & Leg.) (mean)
Early (1989–94)7 29 54 6 (11.1%) 30 (55.5%) 11 (37.9%)* 32 (59.2%) 63.3% 61.4% 62.7%
Late (1995–97)8 11 15 11 (73.3%) 0 (0.0%) 1 (6.6%)* 0 (0.0%) 66.8% 69.1% 72.0%
All 40 69 17 30 12 32
All (Percent) (24.6%) (43.5%) (30.0%)* (46.3%) 64.1% 63.4% 65.3%


1. An opposition boycott occurs if any party withdraws in protest from the election. In most cases, boycotts are partial with some parties participating and others standing back.

2. The “free and fair” determination is based on the preponderance of judgments reported by international election observers and domestic election monitors.

3. Leadership alternation refers to electoral turnover of chief political executives.

4. Loser acceptance is established when minority parties do not mount a legal challenge to the election of a president, or when they accept parliamentary seats following a legislative election.

5. Wherever possible, turnout is measured as total valid votes as a percentage of registered voters.

6. Winner’s seat share refers to the proportion of seats won by the largest party in the legislature, which, for bicameral systems, is regarded as the lower house. Data on founding legislative elections are taken from the variable LEGSEATS in Michael Bratton and Nicolas van de Walle et al., Political Regimes and Regime Transitions in Africa: A Comparative Handbook (East Lansing: Michigan State University Press, 1996).

7. All other data on early founding elections are taken from Michael Bratton and Nicolas van de Walle, Democratic Experiments in Africa: Regime Transitions in Comparative Perspective (New York: Cambridge University Press, 1997), Tables 7 and 8.

8. Data on late founding elections were collected for this article from Africa Research Bulletin, Election Notes, Parliamentary and Presidential Elections Around the World, Elections Today, NDI Reports, Journal of Democracy, and Marches tropicaux.

* Numbers refer to countries rather than elections. Percentages are calculated using the number of countries as a denominator. All other figures and percentages refer to elections.

Since that time, multiparty elections have become commonplace in sub-Saharan Africa. By the end of 1997, only four countries in the region had failed to complete a competitive contest during the 1990s: Nigeria, Somalia, Swaziland, and Zaire. All others had held either first founding elections (40 cases) or regular competitive polls (5 cases). “Late” founding elections (i.e., those held after 1994) did not usually result in leadership alternation, however; more often than not, sitting presidents found ways to survive. As the 1990s progressed, leaders became adept at accommodating the international norm for competitive elections, while at the same time learning to manipulate them to their own ends. In general, the later founding elections were held in Africa, the poorer the quality of their conduct and the lower the likelihood that incumbents would lose.

Table 1 summarizes the features of the 15 late founding elections held in 11 African countries between January 1995 and December 1997. In some respects, these contests resembled the early founding round, for instance in terms of relatively high turnout (averaging 66.8 percent). In other respects, late founding elections revealed novel trends. There were opposition boycotts, for example, in 11 out of the 15 cases. Concomitantly, election observers were unable to endorse any of these elections as fully meeting international standards, calling into question the integrity of the polling, the campaign, or the electoral rules. This undistinguished record stands in marked contrast to the more than half of African elections that observers ruled free and fair during 1990–94.

Strikingly, leadership turnover occurred in only one case after 1994. In Sierra Leone in February 1996, civilian leader Ahmad Tejan Kabbah displaced coup-maker Brigadier Julius Maada Bio (who did not run) in an election in which “a military connection was the kiss of death for vote-seekers.” 6 In all other cases of late founding elections, the incumbent was returned and election winners commonly increased their margins of victory (with vote and seat shares averaging 69.1 and 72 percent in presidential and legislative elections, respectively). Almost invariably, losers refused to accept the results.

How can this trend of declining quality of founding elections be explained? With the exception of Tanzania, late founding elections were all held in countries whose leaders had come to power by military coup or, in the cases of Ethiopia, Liberia, and Uganda, by dint of a guerrilla military victory. The person presiding over (and competing in) the election was a soldier who had doffed his uniform to run as a [End Page 55] civilian. His commitment was usually less to democracy than to shoring up his international and domestic standing by catering to the new expectation that African political leaders ascend to office via elections. Not surprisingly, the imperfect elections sponsored by such calculating leaders did little to advance the cause of democratization on the African continent. The lower quality of late founding elections was perhaps predictable: Because the most reluctant political reformers were the last to concede elections, they departed furthest from the democratic ideal.

Second Elections, 1995 to 1997

During the same period, 1995 to 1997, African countries that had embarked early on political reforms entered a second round of elections. Having completed early founding polls (with various degrees of success), these countries came due again for regularly scheduled polls. A list of 16 such countries and 23 second elections is presented in Table 2.

Table 2

Second Elections in Sub-Saharan Africa, 1995–1997

Country Date1 (Type)2 of Election Held on Time3 Opposition Boycott Free and Fair4 Quality Trend5 Leadership Alternation6 Losers Accept Voter Turnout (% Reg. Voters) Winner’s Vote Share (Pres. Elec.) Winner’s Seat Share (Leg. Elec.)
Namibia 12/7/94 (G) Yes? No Yes unchanged No Yes 76.1 74.5 73.6
Niger 1/12/95 (L) Yes No Yes unchanged Yes? Yes 35.0 34.9
Benin 3/28/95 (L) Yes No Yes unchanged Yes? Yes 75.9 24.1
3/3/96 (P) Yes No Yes? worsened Yes Yes 86.8 52.5
Côte d’Ivoire 10/22/95 (P) Yes Yes No worsened No No 56.2 95.2
11/26/95 (L) Yes No No worsened No Yes? 48.9 84.0
Cape Verde 12/17/95 (L) Yes No Yes unchanged No Yes 76.5 69.4
2/18/96 (P) Yes No Yes unchanged No Yes 43.4 80.0 (est.)
Comoros 3/6/96 (P) Yes No Yes? unchanged Yes? Yes 62.0 61.2
12/1/96 (L) Yes? Yes No worsened No No low 85.7
S~ao Tomé 6/30/96 (P) Yes No Yes improved No Yes? 67.5 52.2
Mauritania 10/11/96 (L) Yes No Yes? unchanged No Yes? 30.0 (est.) 88.6
12/12 /97(P) Yes Yes No worsened No No 74.0? 90.2
Madagascar 11/3/96 (P) Yes No Yes? worsened Yes Yes? 49.7 50.7
Zambia 11/18/96 (G) Yes? Yes No worsened No No 58.7 72.6 87.3
Ghana 12/7/96 (G) Yes? No Yes improved No Yes 77.9 57.4 66.0
Gabon 12/15/96 (L) No Yes No unchanged No No not available 76.6
Mali 4/30 & 7/20/97 (L) Yes Yes No worsened No No 21.0 88.4
5/11/97 (P) Yes Yes No worsened No No 28.4 95.9
Burkina Faso 5/11/97 (L) Yes? No No worsened No No 44.0 90.9
Cameroon 5/18/97 (L) Yes No No unchanged No No not available 60.6
10/12/97 (P) Yes Yes No worsened No No 81.0? 92.6
Kenya 12/29/97 (G) Yes No No unchanged No No 67.0 40.4 51.4
All (n=16) 23 22 8 7 2 8
All (Percent) (95.6%) (34.8%) (30.4%) (8.7%) (34.8%) 55.8% 7 70.4% 70.1%


1. For two-round elections and for elections that last more than one day, the date of the election refers to the first day of polling. Dates are given as month/day/year.

2. G stands for general election, P for presidential, and L for legislative. Elections are dubbed “general” if a presidential and legislative poll are held concurrently on the same day or days; otherwise, elections are listed separately.

3. Elections that are called early or held on time according to the electoral timetable are scored as “Yes.” If elections are only slightly delayed, i.e., held within three months of the scheduled date, the score is “Yes?” This category is included in the calculation of the total number and percentage of elections that were timely.

4. “Yes?” refers to qualified “free and fair” judgments by observers or cases where observers did not agree on the quality of the election. This category is not included in totals.

5. Quality trend is measured by change, if any, between founding and second contests in reported judgments by observers on whether the elections were free and fair.

6. Leadership alternation refers only to presidential elections. “Yes?” indicates instances in which a new party (or party coalition) took over control of the national legislature as the result of a second election. This category is not included in totals.

7. Excludes questionable official turnout figures for presidential elections in Mauritania and Cameroon.

This list excludes countries whose democratic transitions were reversed by military coup before second elections could take place. Melchior Ndadaye’s elected government in Burundi survived all of four months before being overthrown by the Tutsi-led military in 1994. In Sierra Leone, the elected government lasted little more than a year before falling in 1997 to adventurer Johnny Paul Koroma. In Congo-Brazzaville, deposed dictator Denis Sassou Nguesso fought his way back to power with the help of the Cobra militia; second elections were canceled amid the fighting between rival ethnic factions. By the end of 1997, four of Africa’s 17 new democracies (Burundi, Congo-Brazzaville, Niger, and Sierra Leone) and one longstanding multiparty regime (Gambia) had succumbed to military takeovers. And the long-awaited [End Page 57] ouster of Mobutu Sese Seko in Congo-Kinshasa (formerly Zaire) came not after founding elections, but after the military victory of a leader who modeled his transitional regime on no-party principles. Thus any doubt that a backwash of democratic reversals had begun in Africa was surely dispelled by late 1997.

Yet soldier-politicians have not had everything their own way in the current era of global democratization. Military coups failed as often as they succeeded, in part due to a changing international environment in which both donor countries and neighboring African governments withheld recognition and support from unelected regimes. Attempted coups have been blocked by international interventions, for example, by the French in Comoros and the South Africans in Lesotho. Successful coups have been followed by political isolation. And since the ouster of coup-maker Koroma by Nigerian-led forces in Sierra Leone in February 1998, soldiers even run the risk of facing the forced restoration of elected leaders. Reflecting the widespread influence of the democratic idea, soldiers who seize power now find it essential to promise citizens and the international community that they will convene competitive elections as soon as possible.

Indeed, military rule is now exceptional in Africa. Thirteen of Africa’s 17 new democracies have retained civilian rule, and most have held second elections on time (see Table 2, column 3). Just one multiparty regime has so far failed to adhere to its electoral cycle, and then only temporarily. President Omar Bongo breached Gabon’s 1990 Constitution, abandoned a written accord between the ruling and opposition parties, and ignored a Constitutional Court ruling calling for National Assembly elections before 8 June 1996. Instead, he reappointed the prime minister whose term had expired and hinted that he would bring family members back into government. Commentators speculated that Bongo was worried that his party would lose the impending legislative elections, forcing him to appoint an opposition premier and relinquish control over the crucial ministries of finance, oil, and internal security. 7 In the end, however, Bongo allowed disputed elections to go forward six months late on 15 December 1996, and his party was returned with an increased majority.

The other 22 second elections in Africa’s civilian regimes have all been convened in timely fashion. In a few cases the elections were held several weeks after a due date (see “Yes?” in Table 2, column 3). In Zambia, the government and opposition could not agree on the starting date used to calculate the expiration of the electoral cycle; in Ghana, the delay was incurred against the wishes of the government so that presidential and parliamentary elections could be held on the same day. Yet even where incumbents exploited their discretion over scheduling to the utmost, elections were held with sufficient timeliness to satisfy at least the spirit of the law. In other cases, elections were actually called [End Page 58] ahead of time. For example, parliamentary elections were held prematurely in Niger in January 1995 when the ruling coalition broke up and the National Assembly was dissolved. In Madagascar, the impeachment of President Albert Zafy, a legislative decision that was later upheld by the High Constitutional Court, led to early presidential elections in November 1996. In all other eligible countries, second presidential and legislative elections were held precisely on schedule.

Unfortunately, the same cannot be said for local-government elections. In perhaps half of Africa’s new democracies, these were held either late or not at all. Again, Gabon was a main offender: Bongo postponed municipal polls on five separate occasions, saying each time that more time was needed for preparations. While incumbents had self-serving reasons for avoiding exposure at the polls, there was some truth to the claim that logistically demanding local elections were beyond the financial and administrative capacity of poorer African countries (though not of Gabon, which runs budget surpluses). In Malawi, local polling had to be put off because of a government fiscal crisis and international donors’ unwillingness to pay. Obviously, until elected local governments are established or restored, the task of building democracy in Africa will remain seriously incomplete.

The Quality of Second Elections

Did the quality of second elections, like that of the late founding contests with which they overlapped, also decline? The best standard for comparison is a country’s own recent history. Here the findings are less encouraging, with performance worsening in 11 out of 23 cases (see Table 2, column 6). Indeed, as a group, Africa’s liberalized regimes experienced fewer acceptable elections during 1995–97 (30.4 percent) than in their own founding round in 1990–94 (55.5 percent).

Take some examples. In Côte d’Ivoire, President Henri Konan-Bédié introduced a new electoral code stipulating that a presidential candidate had to be both born of Ivorien parents and resident in the country for at least five years, thereby effectively sidelining Alassane Outtara, his only serious rival. Bédié marshaled his main assets—the ruling party’s patronage machine, the paramilitary gendarmerie, and backing from France—to engineer landslide victories in both the presidential and legislative polls. As in Zambia, where similar doctoring of the rules took place, the incumbent appeared determined to “rig an election he would probably have won anyway.” 8 In Comoros, where the presidential election of March 1996 went relatively smoothly, the legislative election of December 1996 was a disaster. Voter turnout was reportedly low following the arrest of two former prime ministers and a series of arson attacks in the run-up to the election. 9 In a searing symbol of electoral decline, the central administrative building in [End Page 59] Moroni was burned to the ground, destroying ballot boxes and other electoral materials.

Although incumbent parties were returned to power in these cases, the poor quality of the elections cast doubt upon the legitimacy of the resulting governments. In Côte d’Ivoire and Comoros, these outcomes represented a measure of institutional continuity, since neither of these countries had strong democratic records. Zambia, on the other hand, had held a model founding election in 1991 and therefore enjoyed a much more promising start. It stands as perhaps the clearest example of the trend of declining quality of second elections in the sub-Sahara region. The Zambian case encapsulates many of the trends evidenced in other new African multiparty regimes, including the disqualification of leading candidates, the spotty coverage of voter registration, the lack of internal democracy in ruling parties, the abuse of government resources during the campaign, and the growing hostility of governments toward watchdog groups. 10

In another unfortunate precedent from one of Africa’s most promising new democracies, the second legislative elections in Mali in April 1997 were badly bungled. Polling stations failed to open, voting materials were in short supply, and inaccurate voter lists prevented voters from casting their ballots. The organization of the elections was so poor that the courts ultimately annulled the results, and the election had to be rerun four months later. Suspicions that these obstacles were not entirely accidental led the two main opposition alliances to announce boycotts; incumbent Alpha Oumar Konaré therefore won the second presidential election of May 1997. Although he got 96 percent, the 28 percent turnout made it a less-than-ringing endorsement.

Africa-wide trends, however, obscure a couple of exceptions. In Ghana and also in the island republic of S~ao Tomé and Príncipe, the quality of elections improved as dubious earlier proceedings were succeeded by more open and transparent contests. In S~ao Tomé, the second presidential election of July 1996 was the first to feature more than one candidate. In Ghana, observers uniformly praised the conduct of the December 1996 second elections, singling out the independence and professionalism of the national Electoral Commission and the determination of voters peacefully to exercise their political rights as “positive step[s] forward in the strengthening of Ghana’s democracy and its electoral process.” 11 Interestingly, although incumbents were returned to office in these improved second elections, the losers accepted the results. While hardly conclusive, this suggests that Africans are beginning to focus less on the removal of individual strongmen and more on the creation of lasting electoral institutions.

Generally, second elections were not marked by leadership alternation; clear presidential turnover occurred in only two cases (see Table 2, column 7). 12 In Benin in March 1996, Nicéphore Soglo was [End Page 60] ejected by the same voters who had swept him to power in one of Africa’s landmark democratic transitions just five years earlier. Soglo (though reportedly privately bitter at the result) honorably hailed the principle of multiparty elections: “My greatest consolation comes from the conduct of my fellow countrymen . . . all the people of Benin were the winners.” 13 In an equally close but more sharply disputed contest, leaders alternated in Madagascar in a drawn-out electoral process. Having tried to delay the second round of voting and charging malfeasance in vote counting, former president Zafy eventually conceded defeat in early 1997. In both these cases, second elections restored former strongmen who had been ousted in founding elections. Yet while Kérékou of Benin proclaimed that he had been “born again,” not only as a Christian but also as a democrat, Ratsiraka of Madagascar showed less indication of having changed his autocratic spots.

Everywhere else in sub-Saharan Africa, sitting leaders weathered second elections. Most often, voters reaffirmed their support for political reform by reelecting leaders who had first obtained office during the regime transitions of the early 1990s. Providing that voters had not been entirely disillusioned by an incumbent’s economic management, they generally sought continuity in political leadership in the aftermath of turbulent interludes of regime transition. Additionally, voters probably chose to give incumbents another term in office before passing judgment on whether the government had delivered on its promises.

The Meaning of Second Elections

What, then, was the meaning of second elections? To be sure, sitting governments and their opponents differed in their views of these events, as indicated by the increased frequency of opposition boycotts, usually to protest an incumbent’s efforts to bend electoral rules or monopolize electoral resources. As Table 2 (column 4) shows, protesting parties stayed away from nearly a third of second elections, a higher rate than in founding elections (24.6 percent), especially early ones (11.1 percent). Opposition boycotts were a surefire way to call the integrity of an election into question. Table 2 confirms that every boycott of a second election was accompanied by unfavorable reports on that election from observers and monitors. Not unexpectedly, there was also a strong correlation between boycotts and the refusal of losers to accept the results of elections.

Especially where boycotts occurred, the announcement of results was immediately followed by a new competition to assign meaning to elections. The default position for election losers—sometimes supported by watchdog nongovernmental organizations (NGOs)—was to level allegations that the elections had been rigged, whether or not [End Page 61] electoral fraud could be definitively proven. Reelected governments responded to these allegations by pointing to peaceful polls and to participation by multiple candidates and parties as evidence that democratization remained fundamentally on course. Both sides directed their appeals to the international community, which is the principal source of funds for both government and NGO development budgets.

For example, an intense struggle over the meaning of second elections consumed Zambia’s political discourse during the weeks following the November 1996 vote. Both President Chiluba’s victorious Movement for Multiparty Democracy (MMD) and the opposition parties had much at stake in the verdict. The MMD stood to lose the balance-of-payments support that it needed to prevent the economy from slipping back into the stagnation of the Kaunda era. The opposition parties, trounced at the polls, stood to lose the leverage over the government that they derived from the sympathy of the donor community. In press conferences and newspaper editorials, each side aggressively presented its own version of how the elections should be read.

Other participants expressed their own interpretations. International donors chose to make Kaunda’s exclusion the centerpiece of their objections, correctly asserting that no election could be fair when a major opposition candidate was barred from competing. Somewhat overlooked in the discourse, however, were the views of those who voted. By largely ignoring the opposition boycott and granting the MMD a second term, most Zambians indicated that they hoped that the “old man” (Kaunda) would quietly step aside from politics. Many also supported the idea that candidates whose ancestries were not “authentically Zambian” should be barred from running for president and saw nothing wrong with altering the Constitution to uphold this principle. 14 Thus different audiences judged the election differently. What was undoubtedly a flawed election from the standpoint of most outside critics was far more acceptable to most Zambians.

At a minimum, opposition boycotts signaled a lack of full agreement on the rules of the political game. Flawed second elections thus had contradictory implications for democracy: While stimulating civil society and mobilizing it against manipulation by incumbents, they undercut consensus on procedures for constituting governments. In most African countries, opposition boycotts of flawed elections raised the alarm that even elected leaders would stoop to political abuse in order to remain in power. In the final analysis, however, an opposition’s rejection of a flawed election did not deny the legitimacy of elections per se. If anything, the boycotters’ emphasis on electoral violations (rather than on the fact that elections were being held in the first place) served to underscore the basic agreement emerging among [End Page 62] all participants that elections are the only acceptable institutional device for choosing top leaders.

Participation and Competition

For the most part, leaders have not chosen to meddle with rules of political participation as a means of maintaining power. Civilian leaders reluctantly accept that citizens have political rights to be enfranchised, to vote, and to be consulted about policy in some fashion. Cases may be found—in Kenya and Cameroon, for example—of deliberate tampering with voter rolls in order to exclude voters from known opposition strongholds. More widespread are tardy election planning, inadequate systems of voter identification, and outdated voter registers. Yet despite opposition claims, these problems are just as likely to be attributable to the state’s fiscal and administrative weakness—and to growing voter indifference—as to premeditated interference on the part of politicians.

Instead of trying to control participation, incumbents have doctored the rules of political competition by restricting entry to candidacy for office. In Africa’s personalistic political regimes, such rule changes usually involve the disqualification of principal rivals for the presidency, though dissident ruling-party candidates for parliamentary seats may also be sidelined. This finding confirms that the biggest challenge in getting to democracy, especially from the mass-mobilizing one-party systems that were so common in Africa, is not so much the expansion of political participation as the introduction of genuine political competition. It also suggests that, although a normative consensus may be emerging in Africa around the principle of broad popular participation, there is still fundamental disagreement on the rules for open political contests.

With few exceptions, second elections also confirm that polling procedures and vote counting are not principal sources of electoral malfeasance in Africa. If rigging occurs, it does so long before polling day in the form of vote-buying and political intimidation. Indeed, (notwithstanding election-day debacles in Mali and Tanzania,) observers increasingly distinguish between the growing efficiency and effectiveness of polling administration on the one hand and persistent problems with election rules or campaign conduct on the other. Incumbents who are intent on retaining office have found “wholesale” rule changes (concerning who competes) to be a far more effective means of controlling outcomes than seeking to influence votes individually at the “retail” level (i.e., who participates).

Comparisons between Table 1 and Table 2 show clear trends in political participation and political competition. Voter turnout fell between founding and second elections (from an average of 64.1 to a [End Page 63] mean of 55.8 percent). To some extent, this is to be expected as elections become routine and as their stakes get smaller. Yet in some countries (like Niger and Mali) turnout rates were low enough to indicate an electorate seriously disengaged from the political process. And, in other countries (like Mauritania and Cameroon), incumbent strongmen revived the discredited expedient of inflating turnout figures, a practice that only added fuel to their opponents’ rejection of election results. Yet in most places, election observers and monitors commented that voters displayed great seriousness of purpose; at least a core of African citizens seemed to have become attached to political rights and wished to keep exercising them regularly.

Political competition also declined as winners increased their margins of victory from first to second elections. Whereas the winning candidates or parties averaged 63.4 percent of the presidential votes and 65.3 percent of the legislative seats the first time around (see Table 1), these figures increased to 70.4 and 70.1 percent respectively in the second elections (see Table 2). To be sure, leadership alternation was associated with close races, suggesting that vigorous contestation was becoming routine in places such as Benin and Madagascar. But where incumbents were returned, it was usually with increased majorities, and often thanks to opposition boycotts (as in Côte d’Ivoire and Mali).

Indeed, the trend is toward entrenchment of incumbent presidents and domination by ruling political parties. In the 13 countries that have held second presidential elections, the winners beat their rivals by a margin of two to one in no less than seven cases (see Table 2, column 10). In the 14 countries that have held second parliamentary elections, ruling parties now enjoy two-thirds majorities in nine (see Table 2, last column). Under these circumstances, legislatures are more likely to act as agents of powerful presidents than as checks on their authority.

Survival, Not Consolidation

Half a decade after founding elections reached a peak in Africa, many more countries have retained civilian rule than have succumbed to military intervention. These countries have begun to establish a pattern of relatively competitive and participatory contests for top political offices. As such, second elections have probably helped some democracies to survive on the African continent. Yet, wherever elections have revealed persistent disagreements about basic political rules—witness the increasing frequency of opposition boycotts since 1994—they have not helped to consolidate democracy.

On a purely formal level, presidential and legislative (though not municipal) elections are taking place in Africa’s surviving civilian multiparty systems with acceptable punctuality. There is a growing sense among political elites that they cannot avoid going through at least [End Page 64] the motions of competitive elections if they want to retain a semblance of legitimacy. Recent boycotts, moreover, have often led their proponents into the political wilderness, thus reinforcing the impression that elections are the principal game in town. And the intensity of postelection disputes in second elections serves to underscore the importance universally attached to elections.

By the same token, the quality of multiparty elections in Africa is far from perfect—and getting worse. Military dictators like Nigeria’s Sani Abacha are devising ever more cynical formulas for avoiding truly competitive founding elections. And in second elections, civilian presidents have not hesitated to use executive power to rig the rules against their opponents. The more dominant the elected ruling party and the more secure its legislative majority, the greater its latitude in this regard. In contrast to electoral volatility and incumbent turnovers in second elections in Eastern Europe and Latin America, dominant parties in Africa have usually been able to reinforce their supremacy the second time around. After a period of turbulent transitions, African politics is returning to an institutional legacy of “big man” rule, and the electoral alternation of leaders is again becoming abnormal.

To assess the prospects of Africa’s surviving electoral democracies, analysts must distinguish those that are slowly dying (like Zambia by 1997) from those that are gradually consolidating (like Ghana by 1997). And one should never underestimate the difficulty of democratic consolidation on the African continent. So far, among African countries, only Mauritius has satisfied even the most minimal (and excessively electoral) conditions for consolidation set by the “two-turnover test.” It is not certain that countries like Benin and Madagascar, each of which has seen one post-transition turnover, will also tread this path. After all, the ruling parties of Africa’s longest-surviving multiparty systems—Botswana, Senegal, and Zimbabwe—have never lost at the polls. In this regard, Zambia seems to represent a far more common trajectory. There, a dominant party, aided by an opposition boycott but checked by international criticism, obtained an imperfect renewal of its domestic mandate in a second election.

In all political regimes (including new democracies), the meaning of incumbent victories is more difficult to interpret than the meaning of historic voter realignments. Especially in a “big man” political culture, it is unclear whether the reelection of an incumbent constitutes the extension of a leader’s legitimacy or the resignation of the electorate to his inevitable dominance. For these reasons, the meaning of Africa’s second elections will necessarily be murkier than that of the watershed contests of the early 1990s. This round of elections poses special challenges to analysts, who must cut through the rhetorical appeals of both winners and losers. When all is said and done, however, the [End Page 65] fact that intense political struggles are being waged over the convening, the conduct, and the meaning of second elections is proof positive that, in Africa, the institution of elections is beginning to matter.

Michael Bratton

Michael Bratton, professor of political science at Michigan State University, is spending 1998 in South Africa. He is the coauthor, with Nicolas van de Walle, of Democratic Experiments in Africa: Regime Transitions in Comparative Perspective (1997). An early version of this article appeared in a chapter, coauthored with Daniel Posner, in State, Conflict, and Democracy in Africa, edited by Richard Joseph, copyright © 1998 by Lynne Rienner Publishers, Inc. Parts of the original chapter are used with permission of the publisher.


1. Samuel P. Huntington, The Third Wave: Democratization in the Late Twentieth Century (Norman: University of Oklahoma Press, 1991), 266–67. According to this standard, the consolidation of democracy occurs whenever the winners of a founding election are defeated in a subsequent contest, and the new winners themselves later abide an electoral turnover. Actually, the “two-turnover test” may be a misnomer since its fulfillment actually calls for three elections, potentially involving three alternations.

2. Terry Lynn Karl, “Imposing Consent: Electoralism and Democratization in El Salvador,” in Paul W. Drake and Eduardo Silva, eds., Elections and Democratization in Latin America, 1980–1985 (La Jolla, Calif.: University of California-San Diego, Center for International Studies, 1986), 9–36.

3. Fareed Zakaria, “The Rise of Illiberal Democracy,” Foreign Affairs 76 (November–December 1997): 22–43.

4. Mitchell A. Seligson and John A. Booth, Elections and Democracy in Central America, Revisited (Chapel Hill: University of North Carolina Press, 1995), 18.

5. Michael Bratton and Nicolas van de Walle, Democratic Experiments in Africa: Regime Transitions in Comparative Perspective (New York: Cambridge University Press, 1997), 6.

6. “Sierra Leone: Falling Out Parade,” Africa Confidential, 29 March 1996, 4.

7. “Gabon: Have Petrol, Will Travel,” Africa Confidential, 4 October 1996, 6.

8. “Côte d’Ivoire: No Contest,” Africa Confidential, 20 October 1995, 7.

9. “Low Turnout of Voters in Comoros Islands Poll,” Reuters, 2 December 1996.

10. See Michael Bratton and Daniel Posner, “A First Look at Second Elections in Africa, with Illustrations from Zambia,” in Richard Joseph, ed., State, Conflict, and Democracy in Africa (Boulder, Colo.: Lynne Rienner, 1998).

12. In Comoros, a new head of government was elected, but under circumstances where the incumbent (Djohar) did not run again (see “Yes?” in Table 2, column 7).

13. “Benin’s Ex-President Says His Loss Is a Win,” Christian Science Monitor, 22 August 1996.

14. Of the 19.1 percent of Zambians who disputed the quality of the 1996 elections, just 16.4 percent (i.e., 3.1 percent of the national population) said they did so because of the exclusion of Kaunda from the presidential race. Most cited other reasons. See Michael Bratton, Philip Alderfer, and Neo Simutanyi, “Political Participation in Zambia, 1991–1996: Trends and Determinants” (Michigan State University Working Papers on Political Reform in Africa, No. 16, East Lansing, 1997), 7.

11. National Democratic Institute, “Preliminary Statement by the NDI International Observer Delegation to the December 7 Elections in Ghana” (Accra: NDI, 10 December 1996). See also “Interim Report of Domestic Election Observers” (Accra: Network of Domestic Election Observers [NEDEO], undated); and International Foundation for Election Systems, “Supporting the Electoral Process in Ghana: Results of the 1996 Presidential and Parliamentary Elections” (Washington, D.C.: IFES, January 1997).

Additional Information

Print ISSN
Launched on MUSE
Open Access
Back To Top

This website uses cookies to ensure you get the best experience on our website. Without cookies your experience may not be seamless.