Introduction

Competitive elections are becoming institutionalized in Africa. But, so far, elections have proven an uncertain mechanism for guaranteeing the political accountability of political leaders. After all, African voters do not always eject corrupt incumbent rulers, instead re-electing them to multiple terms in office. And given the persistent dominance of strong national presidents, elected legislators seldom succeed in closely monitoring the performance of the executive branch. Thus, despite two decades of democratization across the sub-Saharan subcontinent, political executives in Africa continue to enjoy considerable room for decision-making maneuver with all the opportunities for corruption and maladministration that such discretion allows.

Why might this be? One reason is that democratic political accountability has numerous attitudinal, behavioral and institutional requirements of which popular participation in regular, open and honest elections is only one. While free elections may be a necessary condition for holding political leaders in check, such elections alone are rarely sufficient. Also required is a population of informed and active citizens who singly and collectively express demands for political accountability in the long intervals between elections. In addition, public officials can readily avoid responsibility for their actions unless habits of executive discretion are offset by systematic restraints from effective institutions of countervailing power. Using public opinion data, this article explores the strengths and weaknesses of elections and other institutional mechanisms in obtaining political accountability in Africa.

The article proceeds as follows. A first section distills essential features of the amorphous concept of accountability from debates on this subject in the literature. Subsequent sections use public opinion survey data to measure key dimensions of democratic political accountability and to describe the institutionalization of elections and popular preferences for accountability in Africa. The article then proceeds by testing whether elections are related to democratic political accountability and, if so, how. A final section concludes that, in Africa, to the extent that elections matter to accountability at all, they seem to operate indirectly by strengthening legislative autonomy and oversight rather than directly by voters disciplining leaders at election time.

The concept of accountability

Mark Bovens notes that the term ‘accountability’ too often serves as an easy synonym for good governance, ‘a conceptual umbrella that covers various other distinct concepts such as transparency, equity, democracy, efficiency, responsiveness, responsibility, and integrity’ (Bovens, 2007: 449). Preferring a narrower approach, he defines accountability as a relationship in which an agent has an ‘obligation to explain and justify conduct’ to a principal (see also Miller, 2005). In a political democracy, the key relationship is between voters as principals and public officials as their agents. At the core of democratic political procedures is the expectation that citizens can compel officials to be responsive regarding issues of governmental power, public expenditure, and popular needs (Moncrieffe, 1998, Goetz and Gaventa, 2001).

Adsera and Boix state the issue in the plainest terms: ‘How well any government functions hinges on how good citizens are at making their politicians accountable for their actions’ (Adsera and Boix, 2003: 478). In a cross-national empirical analysis, they demonstrate that, ‘the presence of a well-informed electorate in a democratic setting’ explains between one-half and two-thirds of the variance in the levels of government performance and corruption (ibid: 445). Moreover, high quality elections and high levels of voter awareness matter more to accountable governance than other often-cited factors — including economic development, social structure or constitutional arrangements.

Even when the concept of accountability is considered in distilled form — as democratic political accountability (Przeworski et al, 1999) — it nevertheless remains multifaceted. Two theoretical distinctions from the comparative politics literature enable further specification.

First, in a landmark contribution, Guillermo O’Donnell distinguishes vertical from horizontal accountability (O’Donnell, 1994). Vertical accountability comes from below. It is exercised mainly through elections, which provide citizens with intermittent opportunities to reward or punish incumbent leaders. In some jurisdictions, dissatisfied voters may even petition for the recall of their elected representatives and vote for replacements. Moreover, in all democratic regimes, citizens can seek vertical accountability on a sustained basis between elections by making regular claims against elected leaders through political parties, civic associations and the mass media. When citizens organize themselves into voluntary groups to press collectively for accountability, some analysts have been tempted to distinguish a distinct form of ‘social’ accountability (Goetz and Jenkins 2001, Malena et al, 2004). We prefer the view, however, that popular calls for executive restraint — a demand-driven process — render ‘social accountability mechanisms as an example of vertical accountability’ (Stapenhurst and O’Brien, 2007: 3).

By contrast, horizontal accountability refers to constitutional checks and balances on the political executive from other institutions of government — like the legislature and courts. While the role of the courts is to hold the executive legally accountable, parliaments in a democracy are empowered to command political accountability. Together these formal institutions provide ongoing oversight during a government’s term of office. They may be supplemented by additional agencies of restraint, including, but not limited to, anti-corruption commissions, supreme audit institutions, ombudsman offices, human rights commissions, and independent central banks. In short, a bevy of separate and relatively autonomous institutions can invoke obligations on the executive branch to explain and justify conduct by requiring its agencies to ‘report sideways’ (Stapenhurst and O’Brien, 2007: 1). In these cases, the lines of accountability run laterally among institutions at the apex of the state.

A second important conceptual distinction refers to political actors rather than political institutions. Richard Rose and colleagues adopt the demand and supply language of microeconomics in order to differentiate popular demand for political goods from the supply of such goods delivered by political elites (Rose et al, 1998). These authors have in mind the political good of democratic governance. But it is hardly a conceptual stretch to apply the same terminology (of supply and demand) to political accountability, which we have noted is commonly seen as a defining characteristic of democratic governance.

On the one hand, the electorate demands accountability by trying to make political leaders answer for their official deeds. The effectiveness of the demand side of the accountability quest depends in good part on the transparency of government operations and the extent to which information on government performance is disseminated through the mass media, including social media. Moreover, the effectiveness of demand for accountability hinges on the extent to which citizens mobilize to become active agents in promoting their own interests both through elections and through extra-electoral forms of political participation such as contacting leaders or joining with others to raise issues. As noted earlier, demand for democracy requires a well-informed citizenry that regards itself as rightfully empowered to play a role in its own self-governance.

On the other hand, politicians seek to maximize their own freedom of action. Motivated by the drive to maintain political office, and to enjoy its attendant perquisites, they prefer to insulate themselves from excessive popular claims. In pursuit of their own political and economic interests, incumbent leaders find advantage in managing mass demands and limiting the amount of accountability they are willing to supply. Everything equal, the power differential between principals (voters) and agents (elected leaders) can be expected to incur a deficit of political accountability in which demand usually exceeds supply.

A cross-tabulation of institutions and actors generates a two-by-two matrix for framing an investigation of political accountability (see Figure 1). The logic of this framework is that political actors (both mass and elite) secure (or fail to secure) accountability depending in part on the structure of opportunities and constraints inherent in the institutional context in which they operate. From an institutional perspective (column headings), accountability may be secured in one of two ways: either vertically by competitive elections and other mechanisms of popular representation; or horizontally by a constitution that specifies a separation of governmental powers and procedures for oversight of executive decisions. From the perspective of actors (row headings), collectivities of citizens make demands for political accountability from below, but the elites who occupy state office usually have the largest say in determining the actual level of supply.

Figure 1 

A Matrix of Political Accountability.

Using this framework, we derive four critical elements that together comprise a holistic and mutually reinforcing system of political accountability. As illustrated by the cell labels in Figure 1, these elements are:

  1. Demand for vertical accountability
  2. Supply of vertical accountability
  3. Demand for horizontal accountability
  4. Supply of horizontal accountability

Each element is necessary for a system of political accountability; but like elections, none is alone sufficient. Rather, accountability is most likely to be realized only when all four essential elements are present.

The data

We now operationalize these elements using data from the Afrobarometer, an African-led, independent, comparative survey research instrument. The survey documents the public mood on issues of democracy and governance, currently in 34 African countries.1 Trends in popular attitudes about political accountability have been tracked in 16 of these countries between 2002 and 2012.2 The purpose is to assess the current and evolving views of African citizens towards the various proposed components of democratic political accountability and to consider why, to date, elections have not proven to be a fully effective device for holding African political leaders to account.

The institutionalization of elections

To all appearances, Africans are robustly attached to competitive elections as a preferred method for choosing leaders. Across 34 African countries in 2011–2013, an average of 84 percent of survey respondents associate themselves with the view that ‘we should choose our leaders in this country through regular, open and honest elections.’ Only a small minority (14 percent) demurs that ‘since elections sometimes produce bad results, we should adopt other methods for choosing this country’s leaders.’

In some countries, like Tunisia (94 percent) and Ghana, Senegal and Algeria (all 93 percent), almost everyone prefers elections, and in no country apart from Lesotho (63 percent) and Malawi (73 percent) does popular support for elections fall below three quarters of the adult population. Moreover, high average levels of popular support for elections have held steady over the past decade (see Figure 2). Even a 2012 military coup in Mali did not shake citizens’ faith in elections; in December of that year, more than eight out of ten Malians continued to regard a competitive poll as the best method for selecting leaders (Coulibaly and Bratton, 2013). These data can be taken as evidence that, at least from a popular perspective, competitive elections are now institutionalized in Africa as the only legitimate mode of national leadership succession.

Figure 2 

Popular Support for Elections.

But the quality of elections in Africa remains strained. There is considerable variation across countries when survey respondents are asked to rate ‘the freeness and fairness of the last national election.’ In 2012, the proportions of the adult population thinking that the relevant contest was ‘completely free and fair’ or ‘free and fair with minor problems’ averaged 65 percent, ranging from 91 percent in Senegal to just 31 percent in Kenya (in reference to the 2007 elections) (see Figure 3). In the nine countries where fewer than half of all respondents granted the latest elections a clean bill of health, political elites are apparently much less attached to free elections than are ordinary citizens. Under these circumstances, therefore, the institutionalization of low-quality elections is just as likely to lend resilience to authoritarian rule as to help to consolidate democracy (Bratton 2013a, Bogaards 2013).

Figure 3 

Election Quality.

Trends in political accountability

Apart from preferring elections, are ordinary Africans also attached to political accountability? For the first time in 2012 the Afrobarometer asked a direct question on this subject: ‘Which of the following statements is closest to your view?’ Either (a) ‘It is more important to have a government that gets things done, even if we have no influence over what it does.’ Or (b) ‘It is more important for citizens to be able to hold government accountable, even if that means it makes decisions more slowly.’

On average, a majority (52 percent) opted for the latter statement, which indicates a basic popular predisposition toward responsible government, even when posed as an alternative to effective government (see Figure 4). A significant minority (43 percent) however, prefers governmental effectiveness. In addition, there is very substantial cross-country variation. Citizens in Botswana express the strongest general claim for accountability (74 percent, versus 24 percent for effectiveness — a 50 point gap). But in Niger and Madagascar, there is a substantial preference for effective over accountable government (by margins of 27 and 25 points, respectively). In this regard, and in contrast to near-universal popular espousal of free elections, a cross-national consensus has not yet developed around the virtues of political accountability.

Figure 4 

Preference for Accountability vs. Effectiveness.

But why, after two decades of competitive elections, are Africans not more wholeheartedly committed to a political culture of accountability? Is it because citizens, as voters, doubt their own limited capacity to hold leaders responsible for the actions they take in office? Or does the problem lie with institutions other than elections? For example, do citizens consider that written constitutional rules — such as the oversight powers of legislative and judicial bodies — are adequate for disciplining errant leaders? Or do they lack the requisite information to know whether leaders are performing accountably?

To address these questions, we use public opinion data to explore the matrix of political accountability developed in the first section of this article.

Vertical accountability

Demand for vertical accountability

In an earlier study using 2005 data, we argued that popular demand for vertical accountability is weak in sub-Saharan Africa (Bratton and Logan, 2009). We showed that ordinary Africans displayed low levels of participation between elections, were more oriented toward local than national government, and were inclined to delegate discretion over important policy decisions to elected officials in the capital city. In short, Africans were ‘voters but not yet citizens’ who had yet to claim their democratic right to monitor elected political leaders.

What has happened to demand for vertical accountability since then? The key indicator asks respondents, ‘Who should be responsible for making sure that, once elected, legislators do their jobs?’3 Is it the executive branch, the national assembly, a political party, or (most importantly) the voters themselves?

Across 16 countries in 2005, the proportion who thought that ‘the voters’ should demand accountability stood at 32 percent (see Figure 5). By 2012, the proportion in these same countries had risen to a slightly more respectable 37 percent, which suggests that people are beginning to acknowledge their own role in holding leaders to account. Respondents grant responsibility to voters most often in Malawi (67 percent), Kenya (66 percent) and Zambia (61 percent).

Figure 5 

Vertical Accountability.

But it must be noted that, across all 34 countries around 2012, only 34 percent saw the voters as the prime movers in ensuring accountability. A majority (56 percent) still delegates this responsibility to others, namely institutions like the presidency (31 percent), the parliament (17 percent), or a political party (8 percent). At the extreme end of the spectrum, an extraordinary 81 percent of the electorate in South Africa sees no role for voters in holding leaders responsible. They are joined by more than three quarters of all adults in Mauritius, Namibia, Nigeria and Sierra Leone. In sum, growing demand for vertical accountability is more than offset by the continued willingness of many in the electorate to grant the political class the authority to monitor itself. In this regard, elections do not automatically lead to accountability because they are not buttressed by a strong demand for vertical accountability.

Supply of vertical accountability

Turning from the demand side to the supply side, results are even less encouraging. A key indicator of the supply of vertical accountability is whether citizens think that elected legislators ‘try their best to listen to what people like you have to say.’4 Popular evaluations of leadership responsiveness have always been low: in 2005 just 24 percent across 16 countries thought that their members of parliament ‘often’ or’ always’ listen; by 2012, the average had fallen to just 15 percent across the same 16 countries, and 13 percent across all 34 countries surveyed. In only two countries – Botswana and Swaziland – did even a minimal threshold of one in four respondents (25%) say that their representatives listen.

In the most extreme examples, popular assessments of leadership responsiveness have dropped 36 percentage points in Tanzania (down from 53 percent who said MPs were responsive in 2005 to 17 percent in 2012) and 32 points in Namibia (from 47 percent to 15 percent). Even in Benin, a country that usually scores well on aggregate democracy indices, just 3 percent of adults now think that their national assembly deputies often or always listen, down sharply from 17 percent in 2005. Local government councilors do only slightly better, with an average of 18 percent of survey respondents saying that councilors listen. But ratings have also dropped sharply: in 2005, across 16 countries, 32 percent gave councilors positive marks, compared to just 19 percent in these same countries circa 2012.

Sharp declines in the perceived supply of vertical accountability again suggest that democratic reforms, such as competitive elections, have not succeeded in guaranteeing a reliable supply of vertical accountability. On the contrary, as years pass under democracy, citizen satisfaction with the responsiveness of elected leaders has actually declined. In other words, even though weak popular demand for vertical accountability is gradually growing, the supply by political elites remains at low levels and may even have shrunk over the past decade. The most positive interpretation of these trends is that Africans are becoming ‘critical citizens’ (Norris 1999, Norris 2011): as they develop higher expectations of leadership performance, they grant lower scores for responsiveness to elected representatives. A more negative construal would infer that elected leaders are gradually learning that democratization does not threaten their job security; instead, they face limited risks at the polls even if they neglect constituency service.

Horizontal accountability

Demand for horizontal accountability

If vertical accountability falls short, then perhaps horizontal pressures can help to fill remaining accountability gaps. The Afrobarometer currently measures popular demand for horizontal accountability by asking people to choose between two statements: either (a) ‘The members of parliament represent the people, therefore they should make the laws for this country even if the president does not agree;’ or (b) ‘Since the president represents all of us, he should pass laws without worrying what parliament thinks.’ The former option is held to indicate demand for horizontal accountability.

Across 34 countries in 2012, we find that two thirds (65 percent) expect parliament to make the laws rather than the president. A majority calls for parliamentary initiative in 31 of 34 countries. But there are notable exceptions, including especially Algeria, where barely more than one third of citizens (37 percent) prefer a strong legislature, compared to 80 percent or more in Benin, Côte d’Ivoire, Liberia and Uganda. Across the 16 countries for which time series data are available, there has been essentially no change: 65 percent expect parliament to play the leading role in both 2005 and 2012 (see Figure 6).

Figure 6 

Horizontal Accountability.

These high and steady levels of demand for horizontal accountability are consistent with the earlier observation that voters tend to see accountability as someone else’s responsibility rather than their own. Moreover, it is obviously easy for citizens to endorse an idealized model of parliamentary sovereignty and oversight, with faith in formal proceedings perhaps exceeding the real level of effective demand (Nijzink et al, 2006). Also, citizens may lack the requisite level of information about procedure or policy to reliably determine whether parliament is doing an adequate job at exercising its oversight responsibilities. But it is nonetheless encouraging that majorities in most places seem to favor institutional checks and balances, even if people put more confidence in these mechanisms than their own knowledge or the actual performance of African parliaments warrants in practice. This brings us to the supply side.

Supply of horizontal accountability

Is demand for horizontal accountability matched by equally high and sustained levels of perceived supply?

The Afrobarometer measures the perceived supply of horizontal accountability by asking, ‘in your opinion, how often does the president ignore the courts and laws of this country?’ On average in 2012, a solid majority (59 percent) of all respondents across 34 African countries said that the president ‘never’ or ‘rarely’ ignores the laws and courts. People were especially likely to report an accountable president in Benin, Ghana, Mauritius and Tanzania (all above 80 percent), and majorities report an accountable president in 24 of 34 countries. In contrast, citizens see the leaders of Egypt (28 percent accountable), Morocco (30 percent) and Tunisia (35 percent) as strongmen who regularly make unilateral decisions.

On average, therefore, the perceived supply of horizontal accountability across Africa falls short of the level of expressed demand (59 percent versus 65 percent). As with vertical accountability, citizens want more horizontal accountability than political elites are willing or able to supply, although in this case the gap is much narrower. But an accountability gap persists regardless of whether the relevant institutional arena is periodic elections or the foundational laws of a constitution.

There is a silver lining to this cloud: across the 16 countries tracked since 2002 the trend in the perceived supply of horizontal accountability is somewhat positive, and the demand-supply gap has almost closed (see Figure 6). Across the 16 countries for which we have data over time, the proportion of survey respondents who think the president ‘rarely’ or ‘never’ ignores the law5 rose 9 percentage points between 2005 and 2012, from 55 percent to 64 percent, almost matching the 65 percent who demand horizontal accountability in these same countries. There is apparently some merit to the argument that, incrementally and in some respects, formal rules are beginning to matter in sub-Saharan Africa (Posner and Young, 2007).

From elections to accountability?

Does the institutionalization of elections bear any relationship to these patterns of political accountability? Recall that the motivation for this article was the concern that, so far, elections have proven an uncertain mechanism for guaranteeing accountability. If a relationship exists, is it positive or negative? Is it coincidental or causal? And through which mechanisms — vertical or horizontal, demand or supply — are elections and accountability linked, if at all?

As a preliminary step in analysis, Table 1 reports bi-variate statistical correlations between popular support for elections on the one hand and the various dimensions of accountability on the other. The results suggest various possible connections.

Table 1

Is Support for Elections Related to Accountability?

Support for Elections

Demand for Vertical Accountability
   (Voters should be responsible for making sure that MPs do their jobs)
.032**

Supply of Vertical Accountability
   (MPs often or always try their best to listen to ordinary people)
-.008

Demand for Horizontal Accountability
   (Parliament should make laws rather than the president)
.162**

Supply of Horizontal Accountability
   (President rarely/never ignores laws and courts)
.061**

First, individuals who prefer to ‘choose leaders through regular, open and honest elections’ are likely also to demand vertical accountability.6 They are inclined to think that ‘the voters’ — rather than presidential, parliamentary or party leaders — should actively try to ensure that legislators do their jobs. Although the correlation between attitudes to elections and accountability between elections is positive and statistically significant, it is not very strong. And we cannot be certain that the causal direction of the observed relationship runs from elections to accountability, even though that route seems more plausible than the reverse. Those caveats having been noted, however, there is at least trace evidence that African citizens who prefer competitive elections are more likely to demand vertical accountability.

Second, however, popular support for elections has no apparent connection to the perceived supply of vertical accountability. Individuals who prefer elections report no gain in the extent to which legislators ‘try their best to listen’ to ordinary voters. Instead, the correlation between these two variables is negative and statistically insignificant. Apparently, the perceived supply of vertical accountability is entirely independent of a voter’s attachment to elections. This result allows the important inference — subject to further causal testing — that, in the African countries studied to date, elections are occurring without subsequent improvements in perceived levels of vertical accountability.

Third, consider the link between popular support for elections and demand for horizontal accountability. A strong, positive and significant relationship is evident here. In other words, individuals who favor elections as the best method of choosing leaders also intend that the parliament should hold the executive in check during the process of making and implementing laws. This observation suggests that voters regard the functions of elected legislators as being more expansive than merely the provision of patronage-based constituency service (Barkan, 2009). They seem to see roles for elected legislators in proposing and passing laws and exercising oversight on the executive branch, ‘even if the president does not agree’ (Mattes, 2014).

Fourth and finally, there are also hints of a supply response. An individual’s support for regular, competitive elections as the preferred method of leadership selection is positively and significantly related to his or her perceived supply of horizontal accountability. In other words, voters evidently think that elected legislators are able to reduce the extent that the president ignores parliament and ‘does what he wants.’ Although, in Africa, the institutional development of independent legislatures certainly lags behind the institutionalization of elections, there is trace evidence here that, at least in the minds of the electorate, parliaments are becoming stronger.

These results suggest, however, that, if there is a positive link between elections and accountability, it is mainly indirect. Instead of empowering voters themselves to obtain vertical accountability directly from their elected legislators, elections are seen to strengthen the legislature’s capacity to exercise horizontal checks on the executive branch. The question remains, however, whether such indirect forms of accountability are as effective as direct ones. Does it reflect an eagerness among voters to delegate responsibility for checking executive power mainly to elected political elites, thus abdicating their own rights and roles in the policy process beyond voting in elections? At minimum, the somewhat unexpected and relatively strong results on horizontal accountability require further interrogation and interpretation, which we now undertake.

Interpreting accountability

A pattern of accountability in which ordinary Africans rely more heavily on elected leaders to hold each other accountable than on their own participatory roles as civic watchdogs begs several lingering concerns. It is unclear whether voters are knowingly surrendering the opportunity to perform as political principals or whether they simply feel powerless to do more than simply cast a ballot. It certainly seems doubtful that all citizens have a sound grasp of the formal institutional relationship between president and parliament under a democratic constitution. Nor does it seem plausible to assume that many ordinary people possess enough reliable information to judge whether any arrangements for parliamentary supremacy or separation of institutional powers is actually working as intended.

For these reasons, it is necessary to explore possible reasons why, in the minds of African publics, elections are seen as a better guarantor of horizontal than vertical accountability. Several factors may be at work, including the quality of elections, the partisan identity of voters, and the availability of practical information that voters have about the actual nature of executive-legislative relations. In order to test and weigh these possible explanatory factors, Table 2 presents a multivariate regression model of the perceived supply of horizontal accountability. Predictor variables include popular perceptions of the quality of elections, whether voters identify with the party that won the last election (winners), and a proxy indicator (level of education) for access to relevant information.7 Also included in the model are controls for standard demographic variables (age, gender, residential location, and poverty)8 and for the other dimensions of accountability not presently targeted for explanation.

Table 2

Multivariate Model of the Supply of Horizontal Accountability.

B S.E. Beta Sig

Constant 1.877

Female -.003 .011 -.001 .791

Rural .045 .012 .023 .000

Education -.017 .003 -.039 .000

Age -.001 .000 -.001 .850

Poverty -.039 .005 -.049 .000

Support for elections -.058 .006 -.054 .000

Good quality elections .166 .006 .182 .000

Demand for vertical accountability -.099 .012 -.050 .000

Supply of vertical accountability .000 .007 .000 .955

Demand for horizontal accountability -.027 .006 -.026 .000

Winner (partisan of ruling party) .119 .004 .174 .000


Adjusted R square = .228

Note. Method is OLS regression. Country fixed effects included but not shown. Madagascar, Morocco and Swaziland not included due to missing data.

One would expect mass attitudes about the supply of horizontal accountability to vary with prevailing political conditions. As indicated earlier, the quality of elections is measured by a question that asks survey respondents, ‘on the whole, how would you rate the freeness and fairness of the last national election?’ Response categories range from ‘not free and fair’ through ‘major problems’ and ‘minor problems’ to ‘completely free and fair.’ Voters have good reason to expect greater responsiveness from representatives who gain office in an open contest than from those who are elected by irregular methods. They are also more likely to anticipate an assertion of parliamentary prerogatives — to develop a legislative agenda and impose constraints on the presidency — in the wake of free and fair elections. By contrast, if elections occur on an uneven playing field, the composition of the legislature is more likely to represent the president’s supporters than an independent opposition. As a result, voters would probably doubt the effectiveness of any subsequent effort to exercise institutional checks and balances. One would therefore expect that citizens would see a lower supply of horizontal accountability in the wake of poor quality elections.

The results in Table 2 are consistent with these expectations. In the public mind, there is a particularly strong positive association between a free and fair election on the one hand and, on the other hand, the perceived likelihood that the president will act within the bounds of the rule of law. In other words, citizens consider that a good quality election sets the stage for a supply of horizontal accountability. Among the various factors considered, the perceived quality of the last national elections is the best single predictor. One possible interpretation of this striking result is that good elections are seen to empower legislators to do a better job of monitoring and checking the executive branch.

Similarly, one might expect citizens to perceive a greater supply of horizontal accountability if they identify politically with the party that won the previous election. By the same logic, they might doubt the independent power of the legislature to check the executive if their own preferred party does not hold a majority in the parliament. As a shorthand method of measuring partisan identification, we calculate a partisan trust gap that distinguishes ‘winners’ from the ‘losers’ in the previous election. This indicator measures the difference in levels of popular trust expressed for the ruling party and opposition parties. A positive score indicates higher trust in the ruling party; a negative score means higher trust in opposition parties; as such, the indicator distinguishes electoral winners from electoral losers.9

While a number of factors play roles in predicting an individual’s perception of the extent of horizontal accountability, partisan identification has almost as much explanatory power as the quality of elections. In other words, and as expected, partisans of the ruling party are much more inclined than opposition supporters to think that the president is acting in accordance with the laws of the land as promulgated by the legislature and enforced by the courts. In this regard, perceptions of the supply of horizontal accountability take on a distinctive hue that depends on the partisan lens through which citizens view their leaders.

In principal-agent relations, such as those between voters and elected representatives, the principal’s ability to monitor the agent also depends in good part on the availability of reliable information about the agent’s actions. We wonder whether voters know enough about the performance of MPs in parliament to dependably judge whether they are effectively exercising their oversight functions vis-à-vis the executive. As a proxy for such information we consider the voter’s level of formal education on the assumption that more educated people have the ability to access and comprehend a wider range of knowledge. Education is measured in a ten-point scale from no formal education to a post-secondary academic or professional qualification.

Importantly, Table 2 reveals that an individual’s level of education is negatively related to the perceived supply of horizontal accountability. The relationship may not be strong, but it is statistically significant. This result implies that the more voters are well informed about what MPs actually do in the parliament, the less likely they are to think that horizontal accountability is being supplied. Inversely, those voters who have the least capacity to monitor the behavior of elected leaders tend to give them the highest marks for obtaining horizontal accountability. This worrisome result is borne out by the fact that rural dwellers, who are generally less educated than urbanites, are significantly more likely to report high levels of horizontal accountability.

Conclusion

In a valuable recent contribution to the literature on African politics, Arriola writes:

Although the formal elements of democracy – elections, parties and legislatures – have been widely adopted in the (African) region, these institutions have not brought about their desired effects in most countries. Two decades of experimentation with multiparty competition have not necessarily rendered governments more accountable, obliged leaders to become more responsive to citizens, or induced their legislatures to more vigorously restrain the actions of executives (Arriola, 2013: 3).

This article offers a slightly more nuanced and somewhat more sanguine interpretation from the perspective of public opinion in 34 African countries.

True, the results from the Afrobarometer survey confirm that citizens see elected leaders as essentially unresponsive to mass demands for vertical accountability. The weakest link in the chain of accountability remains the perpendicular one between largely passive electors and evasive legislative agents. The direct line of vertical accountability appears to be activated only under the condition of free and fair elections, and then mainly for electoral winners. Moreover, little evidence can be found that the gap between demand and supply of vertical accountability is closing over time — in fact, it appears to be getting wider. In this regard — and with reference to the puzzle with which this paper began — the institutionalization of elections has yet to result in the direct accountability of elected representatives to their constituents.

On the other hand, the African citizens surveyed by the Afrobarometer seem to think that elections strengthen the institutional autonomy of parliament, thus enabling a greater measure of horizontal accountability. On the supply side, almost twice as many citizens think that horizontal rather than vertical accountability is being supplied. In interpreting these results, we note that persons who think the last elections were free and fair and identify with the ruling party are most likely to perceive a reliable supply of horizontal accountability. As such, elections are connected to accountability. But the main connections are indirect: only via high quality elections and obliquely through institutions like parliament. The connection is also politically mediated, that is, mainly in the eyes of those partisan voters who emerge as winners in the last electoral contest.

Finally, we return to Adser and Boix’s observation that political accountability requires ‘the presence of a well-informed electorate in a democratic setting’ (Adser and Boix, 2003: 479). As we have argued, the democratic setting of competitive multiparty elections is now largely institutionalized across African countries. What has still to be achieved is a well-informed electorate who can accurately judge the performance of elected leaders in office. In their direct relationships with representatives, Africans are clear that, because leaders ‘don’t listen,’ vertical accountability is in short supply. In this regard, elections have not yet produced accountability.

Between elections, voters nevertheless delegate a good deal of authority to these self-same leaders and grant them the benefit of the doubt when it comes to horizontal accountability. If voters were well informed about executive-legislative relations, then a case could be made that elections facilitate indirect forms of accountability. But we have noted that the perception that legislatures are a growing institutional constraint on Africa’s powerful presidents is a sentiment held disproportionally by rural voters with low levels of schooling. Citizen with more accomplished educational backgrounds, and therefore greater access to information about the limited capacity of legislators to hold executives to account, are more skeptical that the supply of horizontal accountability is on the rise.