Political Failure and Bureaucratic Potential in Africa

Introduction

Emerging discussions have posited that failures in politics drive much of the observed bureaucratic dysfunction. This is a departure from the popular notions that many of the problems with good governance stem directly from bureaucracies rather than the political spheres that enable them (Meier et al., 2019). Rarely has this discussion been timelier. Right-wing populism has been on the rise over the last half-decade in many developed countries while at the same time objective scores of governance are falling over time in their developing counterparts. Many of these challenges are further exacerbated by confusion and political muck-ups of the handling of the Covid-19 pandemic (Hatcher, 2020).

In this article, I focus on how political failure affects bureaucracies in developing countries—namely those on the African continent. The literature in this vein has suggested that political failure is a global phenomenon, and indeed it is; but the primary focus has been on the United States and other advanced democracies (Meier, 2020; Meier et al., 2019; Park, 2021). There are also relatively few studies that empirically examine the proposed relationships between different avenues of political failure and bureaucratic performance. This study represents an initial effort to this end with a focus on developing African countries in particular. I build on the previous work, but I also chart new territory as I raise a series of issues that are inherently unique to the developing context.

Proceeding in several sections, I first position this work in the broader literature on African bureaucracies and the emerging literature on political failure. From there, I discuss a series of testable expectations that I derive from the literature along with the data and methods I use to test these expectations for African countries. Finally, I discuss the findings and conclude by offering remarks on the limitations of the project and potential avenues for further analysis.

Political Failure and the Effects on Bureaucracy in Africa

Although most extant scholarship examining bureaucracies has focused on the wealthy, industrialized democracies (Gulrajani & Moloney, 2012), there is an ever-growing focus on developing countries—including those on the African continent. Bureaucracies in African countries are not monoliths, and it is a mistake to assume all of the countries group together on any given dimension of measurable governance. Indeed, African bureaucracies run a gamut of diversity across many facets of governance and statehood, and thus it is not always practical to assume issues that apply to one country, or even a region, should be applied to all. In some ways, however, this diversity is a boon for cross-national research as it allows for significant variation among independent variables of interest. As such, the discussions that follow are general when appropriate and qualified when necessary.

The emerging scholarship on political failure is wide in its breadth and application, still most of the discussion has stopped short of empirical analysis (Compton & Meier, 2017; Meier, 2020; Meier et al., 2019). In addition, much of the genesis for the work comes from what should be well-established bureaucratic systems within fully consolidated industrial democracies. This context is markedly different from much of what one might find when it comes to bureaucracies on the African continent. Examples in Meier et al. (2019), focusing on the United States, suggest that what should be a well-functioning, politically supported bureaucracy instead gets drug along with the political system at large through constant crises and dysfunction. Although there is certainly variation in these mechanisms among the developed democracies, the starting place for what constitutes a crisis or wane in function is much different for most African countries.

African countries have seen a wide range of typologies in governance since the waves of independence in the mid-twentieth century. From long-standing democracies like Mauritius and Botswana to strict autocracies like Eritrea, the range of governments on any scale of democracy is varied and changing over time. In some cases, countries have seen both ends of the spectrum multiple times in their relatively short history as independent and fully recognized states. The bureaucracy is no different. Looking retrospectively at the literature, it is clear that what was of interest to scholars in the 1960s and 1970s has changed dramatically to what we see today. In part, much of the early focus was on what was characterized as “first-generation” issues with bureaucracy (e.g., overexpansion, bloat, and responsiveness) whereas only later did issues more relevant to the present study appear into the literature (Goldsmith, 1999).

Indeed, early discussions in the literature on bureaucracies in Africa saw scholars debate whether challenges to effective governance were due to the bureaucracy being too “swollen” or, conversely, whether any woes to governance were due to undersized staffing (Diamond, 1987; Goldsmith, 1999). Earlier assumptions were that African bureaucracies, in general, were too bloated and overstaffed while later findings indicated that, on average, the opposite was the case for some countries. This sort of debate has an earlier foundation in the broader bureaucracy literature where discussions of the trade-offs between efficiency and size abound.

African bureaucracies, and indeed African governments in general, have long been marked by a poor reputation—often the result of corruption (Gould & Mukendi, 1989; Mbaku, 1996; Peters, 2021; Werner, 1983). Applications to bureaucracies can include corruption both in the bureaucracy itself and in the political sphere that influences it. Sources of corruption can include ethnic and kinfolk preference (Apter, 2015), lack of national interest (Gould & Mukendi, 1989; Mbaku, 1996), and self-enrichment (Agbese, 1992) among others. Although corruption has expanded and contracted over time, it remains a concern of high priority for organizations who aim to measure good governance (Mo Ibrahim Foundation, 2020) and scholars alike. In some cases, the issues with effective bureaucracy are so severe that there is an outsized dependence on political patronage and intractability when it comes to professional development (Szeftel, 2000). These challenges are also prone to mar by corruption.

There is also an emerging discussion in the literature on the effects of modern populism on the bureaucracy (Peters & Pierre, 2019; Wood et al., n.d.). By definition, this is a political-related challenge to bureaucratic effectiveness, although there is a question to what extent the rise of populism is a political failure. That discussion is outside the scope of the present study, but the recent work certainly takes a dim view of the value of populism for anything in democracies—especially the impact on bureaucracy and governance (Peters & Pierre, 2019). Despite any ambiguity about the normative aspects of populism, it would not be controversial to suggest that the personalistic and clientelistic nature of politics in many African countries lends itself to at least some degree of populism.

Operationalizing Political Failure in Africa

The previous work on political failure has primarily presented narratives that suggest multiple routes for developing additional empirical work. I take those discussions and create a series of testable propositions that allow for an initial test of how failure can be tracked and treated in developing (and in this case, African) countries. The broad expectation for the empirical section that follows is that better functioning political climates yield higher quality bureaucracies in the form of increased capacity for effectiveness. In narrower terms, I identify several dimensions on which to conduct initial tests for these relationships in Africa. Taken together, I operationalize these propositions using a unique combination of data sources which are outlined in the following section.

I first consider political corruption. A corrupt political climate should be expected to lead to severe consequences for bureaucratic effectiveness and equitable distribution of services (Ionescu et al., 2012). Corruption has long been associated with African countries, and indeed in some cases, it is endemic. For example, scholars have demonstrated that impoverished people are much more likely to experience corruption in their interactions with street level bureaucrats (Justesen & Bjørnskov, 2014). Others have shown that corruption hinders economic and political development; and once entrenched, corruption undermines any efforts at reform and policy innovation (Mbaku, 1994; Mulinge & Lesetedi, 1998). The expectation here is that as corruption increases there will be a reduction in bureaucratic capacity.

The next dimension identified in the literature is the independence and impartiality of the judiciary. This was raised as an important consideration in Meier et al. (2019), and it is one of the clearer theoretical applications of this broader discussion on failure. Absent an ability to arbitrate disputes and place some check on the executive, countries face an uphill battle for keeping any political failures from significantly affecting the bureaucracy. Indeed, as noted by Ayodeji and Odukoya (2014), the judiciary in many African countries is plagued by bribery and corruption.

Finally, I propose that political pluralism and inclusion are important parts of this puzzle for African countries. Institutional arrangements designed to be inclusive are less common for African countries, and thus those governments that have managed to involve the diverse interests of highly fractured societies should be expected to pass this political effectiveness along to their bureaucracies. A breakdown in inclusion often leads to dire outcomes for African countries, and there should be little doubt that drastic social upheaval and, in the worst cases, severe violence can lead to political failures that could do nothing but inhibit the effective function of the bureaucracy.

Data & Method

Data

To examine the impacts of political failure on bureaucratic capacity, I begin with a unique dataset assembled by the Ibrahim Index of African Governance (henceforth IIAG). These data are a compilation of indices, measures, and indicators from reputable sources such as The Varieties of Democracies Project, The World Justice Project, The World Bank, The UN, The World Economic Forum, Global Integrity, and other international/regional organizations (Mo Ibrahim Foundation, 2020). Although the data include a range of variables across several dimensions of governance, my focus is only on those variables relevant to the present study. Although earlier versions of the data were compiled going back to 2000, the most recent version begins the cross-sectional time series in 2010 and ends in 2019. It includes all 54 African countries. It is important to note that the data include countries that are not labeled as democracies by the various studies that do such classifications.

There are myriad discussions in the literature on measuring bureaucratic capacity and performance. The challenge in conducting cross-national work is that not all bureaucracies are easily measured along the same metrics given inherent differences in country size, centralization, government type, etc. As such, I opt for the multi-dimensional measure of public administration quality constructed by the IIAG that is described as follows: "This sub-indicator assesses the extent to which civilian central government staff is structured to design and implement government policy and deliver services effectively." The measure is scaled into a composite score from similar measures developed by both the World Bank and the African Development Bank as part of their assessments for international aid and lending. The measure is scaled to range from 0–100. The base components of this measure are created as part of the “Country Policy and Institutional Assessment” process undertaken annually by the World Bank (Gelb et al., 2004). This involves a series of comparative country assessments and elite evaluations by country experts. The average values for the index (2010–2019) are shown in Figure 1. For the remainder of the study, I interchange the terms capacity, performance, and quality with the general assumption that the meaning is the same: that is that any increase in these dimensions leads to a higher functioning bureaucracy. This is a similar approach as that taken by Wimpy et al. (2017) when using this same measure as a predictor instead of an outcome.

The IIAG includes several potential variables that allow for both the examination of the potential for political failures to impact the bureaucratic performance and the various confounding variables that may also affect the bureaucracy. All of these measures are normalized by IIAG to range from 0–100. To begin, I include a measure of corruption in state institutions. This measure is calculated by the IIAG and sourced from variables produced by the World Justice Project (Botero et al., 2020) and the Varieties of Democracy Project (Coppedge et al., 2021). The World Justice Project is an organization aimed and bolstering the rule of law around the globe. It generates an annual index that develops data and measures to this end. The Varieties of Democracy Project is a social science project aimed at modernizing and contextualizing the measure of democracy worldwide. The component indicators of this measure assess the extent to which government officials accept unauthorized financial payments or abuse their offices for personal gain.

I also include a measure that captures executive compliance with the rule of law. This variable makes sense in the African context given the executive-centered nature of many African governments. The measure is calculated by IIAG and is also sourced from The World Justice Project and the Varieties of Democracy Project, and it captures a series of indicators that measure the extent to which executives accept court rulings, abide by constitutions, and commit to peaceful transfers of power. The next measure captures impartiality in the judicial system. It is a measure sourced from variables calculated using the Africa Indicators from Global Integrity (Global Integrity Team, 2021) and the variables calculated by the Varieties of Democracy Project. Global Integrity collects and disseminates data on governance worldwide. This measure examines the impartiality of judicial appointments and independence in judicial processes.

As mentioned above, given that the IIAG includes all African countries irrespective of regime type, it is important to consider the degree to which more democracy affects bureaucratic capacity. Although this relationship may not always be a perfect one (Meier, 1997), the fact remains that much of the relevant literature emphasizes bureaucracies in democratic environments over those elsewhere. Despite this emphasis more broadly, the literature on transitions in Africa suggests that the extent to which the transition to democracy improved bureaucratic performance was minimal at best (Olowu, 1999; Szeftel, 1998).

To capture at least part of the impact of increased democracy on bureaucratic performance, I include a measure of how democratic the elections are in a given country. This measure is calculated by IIAG from indicators created by both the Varieties of Democracy Project and the African Electoral Index from the Ghana Center for Democratic Development (Ghana Center for Democratic Development, 2020). The measure is constructed using expert evaluations of the free and fairness of elections, transfers of power, and electoral openness and participation. Finally, I include a measure that considers political pluralism in a polity. This is of particular importance in countries with numerous parties and ethnic groups that are systematically excluded. This measure is constructed by IIAG from sources compiled by Global Integrity and the Varieties of Democracy Project. This measure captures the extent to which out of power and opposition parties are free to participate in political processes and the media.

Figure 1.Public Administration Quality in Africa (2010–2019)

I also include a series of control variables for which my expectations are relatively agnostic given the high degree of variation on country-level indicators across the continent. The first is a measure of corruption in the public sector, which captures the potential for corruption in the bureaucracy itself. This measure is constructed from indices created by the World Economic Forum, The World Justice Project, and the Varieties of Democracy Project. It captures theft, bribery, and embezzlement in public sector jobs. This variable is distinct from the state institutions variable above as it focuses on positions most likely to be tied to various levels of the bureaucracy whereas the above measure of corruption primarily relates to elected or autocratic officials. I include country-level measures of GDP per capita, population, and rurality as additional controls. The GDP, population, and rurality data come from the World Bank (World Bank, 2021). Summary statistics for all variables used in the analyses are presented in the appendix.

There is always a tradeoff when using data that are constructed as indices, indicators, and based on elite and expert evaluations. I deal with that in two ways. First, most of the variables used in the analyses that follow are comprised, at least in part, of relatively objective information as part of the indicator creation. For example, whether or not a transfer of power was peaceful is typically not in dispute. I also employ an additional indicator as a secondary dependent variable that draws on mass public opinion as an assessment of public administration quality. I discuss this in further detail in the results section below.

Method

Given that the IIAG variables are normalized to range from 0–100 I employ linear regression as my primary empirical test. This is not without potential pitfalls, however, given the cross-sectional and temporal dimensions of the data. Although it is often typical to use unit fixed effects in this sort of scenario, there is also emerging scholarship that suggests that both this approach and two-way fixed effects are problematic in these types of data (Kropko & Kubinec, 2020; Plümper & Troeger, 2019). As such, my analyses begin with linear models in which I employ clustered standard errors (around country) and time fixed effects to control for any common shocks or trends in the data. The inclusion of a time-lagged (one year) dependent variable instead of time fixed effects does not substantively change the results. The full model specification can be represented as:

$$\text{PAQ}_{jt}=\alpha+x^\prime_{jt}\beta+\delta_t+\epsilon$$

where PAQ represents the level of public administration quality for a given country, j, in a given year, t. $x^\prime_{jt}$ is a vector of the predictor and control variables discussed above for each country and year. The time (yearly) fixed effects are denoted by ${\delta}_t$, which simply represents a vector of dummy variables for each year after 2010.

Results

The model results are presented in Figure 2. Tabular results are presented in the appendix. The estimates are plotted on their original IIAG or observed scale with 95% confidence intervals included as estimates of uncertainty. Before proceeding with a discussion of the findings, it is important to reiterate that the scaling of the variables means that higher values typically represent more desirable outcomes. For example, a score of 100 on public sector corruption indicates that a country would have the least amount of observed corruption in that year.

Figure 2.Primary Model Estimates Examining the Impacts of Political Failure on Public Administration Quality

To begin, countries with more democratic elections are associated with higher levels of public administration quality. This means that political failures in which elections are undermined, thwarted, delayed, or otherwise conducted in a less than democratic fashion should be expected to negatively impact the performance of bureaucracies in developing countries.

The executive compliance with rule of law indicator fails to achieve statistical significance at the accepted threshold. This is somewhat contrary to what was expected as this is one of the distinct political failures we observe in recent examples from industrial democracies like the United States. On the other hand, African executives can be exceptionally powerful relative to other branches of government, and the high number of non-democracies may be driving this effect. Put differently, this could be a feature and not a bug as it relates to the political context across much of the continent. There are positive effects for the impartiality of the judiciary, political pluralism and inclusion, and political corruption. That is to say, in all three cases, more impartiality, more pluralism, and less corruption are associated with higher quality public administration. Taken together, these results suggest that bureaucracies in developing countries can be severely hampered and undermined by political failures when they are present.

Turning to the control variables, the most notable finding is that countries with less corruption in the public sector experience higher levels of public administration quality. This is in line with what we would expect if the variables are truly capturing their intended concepts. There is a very small positive effect for population, but that is not too interesting beyond the status as a control variable. The GDP control variable is not statistically distinguishable from zero. Finally, rurality is a positive and significant predictor of capacity. I can only speculate as to the reason for this, but one possibility is that rural countries have a more decentralized or localized bureaucracy. In any case, this finding deserves more consideration in a study of African bureaucracies more broadly.

Comparing Results to Citizen Survey Measures

Recent iterations of The IIAG Index have included a section called “Citizens Voices” that takes information from the Afrobarometer surveys of numerous African countries (Afrobarometer Data, 2021). Included in that is a composite variable that measures public perception of public administration. Unlike the measure used above that collects information from both elite surveys and other indicators, this measure aggregates several response options from the Afrobaramoter into a measure that is much closer to people in a given polity. The Afrobarometer is a mass public opinion on political and social attitudes survey that, for the measures used here, covered 36 African countries in 2016 (Afrobarometer Data, 2021).

The IIAG describes the citizen voices measure as follows: “This indicator assesses citizens’ perceptions of how easy it is to obtain identity documents and access essential household government services.” This type of measure represents an initial effort at “measuring citizen feedback” as discussed in Meier (2020). The survey responses were measured using a four-point scale and the questions used to construct the measure are as follows:

"In the past 12 months have you tried to get an identity document like a birth certificate, driver’s license, passport or voter’s card, or a permit, from government? [If yes] How easy or difficult was it to obtain the document you needed?

In the past 12 months have you tried to get water, sanitation or electric services from government? [If yes] How easy or difficult was it to obtain the document you needed? [If yes] How easy or difficult was it to obtain the service you needed?

There is a clear distinction here between mass citizen perception and the elite opinions that make up the public administration measure discussed above given that the correlation for these two indicators is only moderate albeit significant (r = 0.471, N = 336, p = .000). Although this survey aggregation is somewhat different from the capacity measure I use above, it is also capturing actual attitudes from the citizens whom bureaucracies are designed to serve. I use this variable as an outcome in an otherwise identical model to the above as an additional test and robustness check. These results are displayed in Figure 3. Tabular results are presented in the appendix.

Figure 3.Secondary (Afrobarometer) Model Estimates Examining the Impacts of Political Failure on Public Administration Quality

The results here are generally less striking given the few variables that are distinguishable from zero. On the other hand, the two significant variables to note are public sector corruption and political corruption. Both of these are probably more immediately observable to citizens on the ground than the other measures that may be more easily understood by elites or international institutions. The other aggregate measures are either too far removed from citizens on the ground or need more time to make an impact on the citizenry to be discernable. Nonetheless, these findings more definitely underscore the importance of corruption as a political failure that predicts the deliverable side of bureaucratic function. As with the primary model presented above, rurality is a positive and significant predictor of bureaucratic capacity.

It is worth noting the limitations and challenges to inference for this citizens’ voices measure. To begin, the number of countries included in the Afrobarometer is far less (36) than the totality of countries in the analyses above, and inclusion in the survey project cannot be assumed as entirely random. This is also an aggregated measure that represents individual-level opinion. A more ideal approach would be to examine additional survey measures as predictors and move any country-level measures to a random effects, multilevel framework. Finally, the Afrobarometer is not conducted every year in every country. This means that some form of imputation is needed to incorporate the measure in the index. Although this may not be an inhibitor to inference, more investigation and testing is needed to determine the extent to which this approach affects and conclusions one might draw from these types of data.

Conclusion

In this article, I have taken several propositions from the nascent literature on political failure and tested them in a developing context. The results suggest that variables capable of predicting true political failures indeed predict bureaucratic implications. Put simply, countries with competitive, less corrupted, and inclusive political environments are more likely to demonstrate higher levels of bureaucratic capacity. Taken together, the above results indicate that many challenges to effective bureaucratic function in Africa have their roots in the political sphere. While some countries can mount effective bureaucracies that rank comparably with more developed counterparts elsewhere, many remain woefully hobbled by corruption in multiple political domains and a lack of inclusion and pluralism. Normatively, we typically assume that democracies are preferable to the alternative, but to what extent this is a relevant indicator in and of itself for the outcomes in the present study needs further consideration. It is quite clear that any number of political reforms, whether more democratic or not, would lead to improved bureaucratic performance and a public administration workforce that is better positioned to deliver for the citizens they serve.

These findings add to the expanding literature on the implications of political failure on the quality and performance of bureaucracy. The results also demonstrate the somewhat unique nature of the African context. Even in the presence of more autocratic executives (something fairly drastic for industrial democracies) the openness and quality of other dimensions of the political landscape still predict a portion of how well positioned the bureaucracies are to perform basic functions and provide services.

This study is not without limitations. The focus on African countries means that some important parts of the political failure are simply left out of this discussion due to fundamental lagging in political development across the continent. The primary dependent variable is, by definition, a measure of performance that may capture as much bureaucratic pathology as it does political failure. Some bureaucracies may be primarily focused on chasing performance as it is measured here and that is not easily controlled for given the limitations of available data. Another issue relates to measurement and the construction of comparative indices. In some cases, it becomes challenging to tease out those things that are included in one index from potential inclusion, or at least correlation with other measures that could end up on the opposite side of an equation. I try to avoid this in the macro-level analyses by focusing exclusively on the quality of public administration operationalized as bureaucratic capacity. Future studies should be wary of this when including, for example, the Worldwide Governance Indicators as dependent variables as a proxy for capacity. Such indicators capture much more, and would likely be highly related to any potential predictors.

There are numerous avenues for additional exploration in this project. There are several propositions from the literature on political failure that remain untested. Not the least of these is the degree to which the currents of populism within a country either cause or interact with political failures that inhibit effect bureaucratic governance. There is also an individual-level component of these questions that needs further investigation (Meier, 2020). The above secondary analysis presents only an initial consideration of individual-level examination and much more can be done on this front on either side of the equation. Finally, there are numerous ways in which these data could be subset, interacted, and expanded to find the boundaries of findings presented above. For example, there could be value in simplifying this discussion by focusing entirely on countries categorized as “democracies” by the various indices designed to that end. There are also interesting opportunities for data collection as it applies to developing countries. Two important arenas would be expanding extant datasets on populism and measures of the size of bureaucracies. These data were not immediately available for this initial iteration of these analyses, but I encourage further data collection and dissemination to these ends.

There is also potential for research that focuses on the role of path dependence in the forms of both colonial and traditional heritage impacts on modern bureaucracies in Africa. This broader issue was raised in Peters (2021), and Wimpy et al. (2017) provide at least some evidence that colonial legacy can moderate the impact of public administration quality on health outcomes. It is possible these legacy issues play a role in moderating the impact, and types, of political failures as they affect African bureaucracies. In sum, the analyses in this paper are only the beginning of exploring the how these relationships vary across the continent and extend to other developing settings more generally.

The study of bureaucracy in Africa is not new, nor is the concept of examining the various political challenges to good bureaucratic governance among the states on the continent. The interplay, however, between political failures and bureaucratic capacity, has heretofore gone empirically unexamined in this context. I have bridged some of that gap here, but certainly encourage more effort on the empirical side of the broader discussion of political failure and the challenges to good governance in Africa and beyond.

Acknowledgements

I acknowledge no funding for this manuscript. I thank the KJPS editor, anonymous reviewers, and William P. McLean for helpful comments on this project. Any remaining errors are my own.