The majority of studies on corruption suggest that it is bad for economic and social development [1]. Corruption has also been shown to have a detrimental effect on tax revenue [2], economic growth [3], equality and poverty [4-7] and well-being [8]. In addition, corruption appears to deteriorate political trust as well as political legitimacy in various institutional settings [9-11]. These harmful effects serve to justify the implementation of incentive mechanisms that influence the behavior of politicians, anxious to stay in power. This is how, in a democracy, elections appear as an opportunity given to citizens to encourage incumbents to reduce corruption.

However, the increase in the level of corruption in the new democracies, on the one hand, and the re-election of a corrupt candidate in the consolidated democracies, on the other hand, gives renewed interest to the analysis of the links between corruption and elections. The relationship between these two variables has been the subject of a varied literature, both theoretical and empirical. On a theoretical level, the analyzes crystallize around two orientations.

The first looks at the influence of corruption on the electoral results of incumbents [12-14], on the one hand, and on electoral participation [15-17], on the other hand. It is commonly believed that high levels of corruption lowers voter turnout [18].

The second, for its part, studies the effects of elections on corruption. This approach broadens the notion of performance voting [19-21] to non-economic dimensions, such as “good governance” [22] and considers that the sanctions and rewards of citizens have incentive effects on the corruption of politicians. Analyzes of the effects of elections on corruption reveal two (02) main points of view. The first, developed by Ferejohn [21], Persson and Tabellini [23], Besley and Case [24], Besley and Burgess [25] predicts that elections encourage politicians to align with citizens' preferences and to abstain of any corrupt behavior. This approach suggests that corruption would decrease in consolidated democracies [26-27]. The second point of view starts from the observation that there are high levels of corruption in consolidated democracies and seeks to explain the absence of electoral sanctions for corrupt candidates. The justifications put forward include asymmetries of information between the voter and the candidate [12,28-30] the sharing of private benefits [31-34] functional corruption [35-36] and voter loyalty [22,37-38].

Empirically, work has been carried out in developed countries; then in emerging and developing countries. Regarding developed countries, Peters and Welch [15] assess the electoral impact of accusations of corruption on candidates in the United States House of Representatives contests in five elections from 1968 to 1978. They show that while most of candidates accused of corruption are re-elected, overall they seem to suffer a loss of 6 to 11% from their expected vote. Ecker, Glinitzer and Meyer [39] analyze whether and to what extent voters punish incumbents for high levels of corruption. Based on 20 elections in Western and Central Eastern Europe from Module 2 of the Comparative Study of Electoral Systems, they show that voters generally punish governments if they perceive high levels of corruption. More recently, Benito et al. [40] discuss the decisions made by politicians that can affect the level of municipal corruption. They use a sample made up of 358 Spanish municipalities with more than 20,000 inhabitants for the period 2004-2009. They find that when politicians want to be re-elected, corruption decreases.

Regarding emerging and developing countries, Callen and Long [41] study the relationship between political networks, weak institutions, and electoral fraud during parliamentary elections in Afghanistan. They find that candidates linked to officials in charge of aggregation receive an average of 3.5 fraudulent votes per polling station. 

Although previous studies have explored the effects of elections on corruption, most of them have been conducted in developed countries. In addition, studies dealing with the link between elections and corruption in countries in African countries do not control this relationship with the number of ministerial rotation. In developed countries, staff turnover in public administration has only recently been seen as an effective anti-corruption measure. Indeed, experimental studies [42-44] carried out in Germany and China support the hypothesis that the instruments of rotation reduce the level of corruption. The absence of such a control variable in studies conducted in African countries is explained by the difficulties encountered in constituting a sufficient time series. Our analysis will help fill this void by addressing the influence of elections on corruption in Gabon, while also integrating rotations in government.

Gabon, a country in Central Africa, is a favorable environment for at least two reasons: 

First, the 2019 ranking of the NGO Transparency International (TI) on the Corruption Perception Index (CPI) places Gabon 123rd out of 180 (with a score of 31 out of 100). In sub-Saharan Africa (continental and island), the country is in 20th place out of 49 countries assessed. Thus, Gabon comes first among the countries of continental Central Africa ahead of Angola (146th), Cameroon and CAR (both 153rd), Chad (162nd), Congo (165th), DRC (168th) and Equatorial Guinea (173rd). In the CEMAC zone, it is second only to Sao Tome and Principe (64th). According to the survey carried out in Gabon by the 10th Global Corruption Barometer Africa 2019 (co-publication carried out by TI and Afrobarometer), 75% of citizens questioned from September 2016 to September 2018 believe that members of government and police officers would be the most corrupt. While this perception is shared by 75% of citizens, 71% of them also perceive the President of the Republic, the Prime Minister and parliamentarians as key players in corruption. Moreover, 87% of citizens believe that the government is making no effort to fight corruption effectively. Survey results show that more than three-quarters of Gabonese citizens fear reprisals if they report acts of corruption; 

Second, according to the NGO Transparency International (TI), “A cross-analysis incorporating data on democracy around the world shows that a link exists between corruption and democratic health. In the CPI, well-established democracies score on average 75 points; imperfect democracies an average score of 49; hybrid regimes - exhibiting elements of autocratic tendency - score 35; and autocratic regimes record the worst results, their average score barely reaching 30 points”.

This reflection therefore analyzes corruption, with an emphasis on elections and the number of ministerial rotations. In particular, we are investigating whether the approach of the electoral period provides sufficient incentives to curb the corrupt behavior of politicians in Central Africa, and particularly in Gabon. In view of the small size of the sample, which only covers the period 2000-2019, we estimate a static model using the robust least squares method, after checking the stationarity of the variables. We show that the presidential and legislative elections and the press increase corruption in the country. Moreover, even if the number of turnover of members of the ministerial office is not significant, its sign shows that there is a risk that it increases the level of corruption in Gabon. The rest of the document is structured as follows. The second section presents the literature review on the links between election and corruption. Section 3 is dedicated to the model for determining the effects of the election on corruption. Section 4 presents the empirical results and the discussion. Finally, section 5 concludes.

Literature Review 

The literature on the links between election and corruption shows a double sense of causation. While some works look at the effects of corruption on elections, the more abundant ones deal with the effects of elections on corruption. 

Effects of Corruption on Elections

The effects of corruption on elections are discussed by discussing the influence of corruption on electoral results, on the one hand, and turnout rates, on the other.

Regarding the effects of corruption on the electoral results of incumbents, they are often based on data at the macroeconomic level and address variations in corruption between countries [12-14]. At the microeconomic level, analyzes are often linked to specific scandals and certain contexts [15,45-50]. Taking into account the agency relationships between politicians and voters [21,51] allows us to postulate that corruption decreases the probability of being reelected [52].

Regarding the effects on electoral results, recent research highlights the role of perceptions of corruption on political behavior, in particular its effect on electoral participation [15-17]. To this end, Mc Cann and Dominguez [18], then Stockemer et al. [17] show that high levels of corruption decreases turnout because high levels of perceived corruption lead to ratings of political authorities and the political system in general [11]. Yet others find a positive effect of corruption on participation [16]. Kostadinova [53], in turn, argues that both effects can occur: it is a direct and mobilizing effect of corruption on participation and an indirect and negative effect that affects participation through political effectiveness.

Effects of Elections on Corruption

Conceptually, this approach relates to the broad theoretical framework of performance voting models [19-21]. Performance voting is based on the simple idea that voters rate parties based on their achievements. This analysis assumes that government performance enters into a voter’s decision making in retrospect. Thus, performance voting is based on voters' ratings of parties' past performance, regardless of how they rate the parties' potential performance. Until now, existing studies on performance voting have mainly focused on economic outcomes [19-20,54]. The main reason is that voters generally agree on desired economic policy goals (economic growth and low unemployment and inflation) and the economic outcomes are easy to observe for the average voter. Economic voting allows voters to make rational decisions even if they are politically inattentive [54]. However, recent literature shows that voters also condition their choice to vote on non-economic dimensions, such as corruption.

The Objective Of (Re) Election, a Brake on Corruption 

The objective of (re) election constitutes a brake on corruption when there is a perfect alignment between the incumbent and the preferences of the voters, on the one hand, and an electoral system based on the majority system, on the other go. 

The Perfect Alignment of Incumbents to Voter Preferences: The Importance of Electoral Repression

The incumbent behavior model, developed by Ferejohn [21], then Besley and Case [24], Besley and Burgess [25] and Besley [55] predicts that the possibility of (re) election encourages politicians to align their preferences with those of citizens and refrain from any rent-seeking behavior. Indeed, incumbents who wish to be re-elected are therefore more likely to be less corrupt than those who are not eligible for re-election. It is often accepted that the possibility of competitive re-elections obliges democratic governments to take into account the preferences of citizens more than they would otherwise. Similarly, Persson and Tabellini [23], as well as the model of Ferraz and Finan [56], predict that when the races are close, the higher proportion of pivot voters prompts politicians to reduce the extraction of rents to guarantee re-election for a second term [57].

In a recent review, Klansja et al. [58] question whether voters sanction corrupt politicians. The authors incorporate rich experiences in choosing joint candidates in surveys in Argentina, Chile and Uruguay. Next, they test the importance of two contextual factors believed to lessen voter retribution against corrupt politicians: how widespread is corruption and whether it brings secondary benefits. Like other researchers, they find that corruption drastically decreases support for candidates. 

Electoral Systems and Corruption 

Rose-Ackerman [59], Kunicova [60] and Golden and Mahdavi [61] show that electoral rules can have variable effects on corruption.

Persson et al. [26], then Rudolph and Däubler [27] show that majority electoral systems are associated with lower levels of corruption than proportional ones, because they promote alternation of government and political competition. Additionally, Carey and Shugart [62], and later Kunicova and Rose-Ackerman [63], argue that in proportional representation, systems using closed lists are associated with more corruption than those using open lists. Finally, the size of the district could also play an important role, especially in proportional systems. Indeed, Chang and Golden show that in proportional open-list systems, corruption increases with the size of the district.

A fundamental assumption underpinning research on the effects of electoral rules on corruption is that voters punish corrupt politicians if given the opportunity. Yet the empirical evidence on electoral crackdown on corruption is mixed. Although there is ample evidence to suggest that corrupt activities do indeed cost incumbents a heavy electoral price [12-13,49,64-65], several results challenge this hypothesis. The re-election of corrupt politicians, likely to increase corruption, is not just a trait of developing countries characterized by weak political and economic institutions, but also concerns established democracies like France, Italy, Japan and the United States [22,30,66].

The Organization of Elections, a Factor of Corruption

Elections are a factor of corruption in terms of imperfect alignment between incumbents and voters' preferences, on the one hand, and tolerance for corruption, on the other.

Imperfect Alignment of Incumbents with Voters' Preferences 

According to Zimmerman [67], an agency problem exists between incumbents and voters. Indeed, the voters, the directors, elect a politician, an agent [21,51]. The interests of politicians and voters are not perfectly aligned, so incumbent politicians can use their power to pursue their own interests rather than those of citizens.

Peters and Welch [15] assess the electoral impact of corruption charges on candidates in the United States House of Representatives contests in five elections from 1968 to 1978. Despite the existence of various writings on corruption in American elections [22,35,68-70] and numerous laws enacted to reduce corruption during election campaigns, the authors argue that 'there are still few studies to assess the loss of votes due to the corruption accusation. Specifically, they suggest assessing the degree to which specific corruption allegations decrease the electoral success of accused candidates. They find that most of the candidates accused of corruption are re-elected, overall, but they appear to suffer a loss of 6 to 11% from their expected vote.

Alt and Lassen [52] argue that corruption could increase when the politician remains in power or when the politician does not stand in the next election. On the one hand, if the incumbent is "too sure" to stay in power, corruption would be high. On the other hand, if the incumbent is almost certain that they do not have a next period, for example due to time limits, they may be more corrupt. In other words, a politician who faces the possibility of re-election can exploit information asymmetry to increase the chance of re-election by behaving honestly, refraining from corruption. Thus, politicians who are encouraged to run in the next election are likely to be less corrupt than politicians who do not. However, politicians running for re-election may use several illicit means, such as illegal campaigning practices, fraud or intimidation, to increase the likelihood of being re-elected and continuing their illegal activities [71]. In other words, re-election can fuel corruption.

Chang shows that if politicians running for office are uncertain about their chances of winning elections, they will be more corrupt in order to fund campaigns to stay in power. 

Likewise, Ferraz and Finan [72] study how re-election incentives affect political corruption. Using a dataset of Brazilian municipalities, they find that politicians with re-election incentives are much less corrupt than politicians without re-election incentives.

In the same vein, Beylis et al. [73] show that mayors facing re-election inducements are much less corrupt than mayors who cannot stand for re-election. Therefore, by increasing the horizon of the candidate, the incentive to engage in corruption decreases.

Callen and Long [41] study the relationship between political networks, weak institutions, and electoral fraud during parliamentary elections in Afghanistan. Starting from the existence of the different ways of manipulating elections, including bullying voters, stuffing boxes and changing the vote totals after the vote, they are interested in whether candidates exploit the links (their connections) with election officials to add fraudulent votes during the aggregation process (aggregation fraud). To conduct this analysis, the authors combine data on political relationships between candidates and election officials, a nationally controlled assessment of new surveillance technology, and direct measures of aggregation fraud. They find considerable evidence of aggregation fraud in favor of logged in candidates. In fact, candidates linked to officials in charge of aggregation receive an average of 3.5 fraudulent votes per polling station (around 13.7% of the average for their polling station).

Ecker, Glinitzer and Meyer [39] analyze whether and to what extent voters punish incumbents for high levels of corruption. Drawing on the literature on economic voting, they argue that corruption performance voting varies with factors at the individual level and institutional context. The authors' empirical analysis is based on individual, macroeconomic data, clarity of responsibilities, and government economic performance. The data collected to estimate the model concerns 20 elections in Western and Central Eastern Europe from module 2 of the comparative study of electoral systems. The results show that voters generally punish governments if they perceive high levels of corruption.

For De Vries and Solaz [38], democratic elections are supposed to play a crucial role in the fight against corruption among civil servants. Yet the empirical evidence to date is mixed and suggests that the electoral punishment for corruption is rather light. Hence, the purpose of their article is therefore to provide an overview of the conditions under which voters punish corruption. For De Vries and Solaz [38], three key stages make it possible to understand the absence of electoral sanctions for corrupt candidates: (1) the acquisition of information, (2) the attribution of blame and (3) the behavioral response which underpin a retrospective vote based on corruption. A breakdown of one or more of these certain stages may lead to an absence of electoral sanction for corruption.

Benito, Guillamon, Rios and Bastida [40] discuss the decisions made by politicians that can affect the level of municipal corruption. Specifically, they study whether local politicians' incentives for corruption are influenced by the salary they receive and / or their intention to run in the next election. The sample used includes 358 Spanish municipalities with more than 20,000 inhabitants for the period 2004-2009. The authors use two different estimation methods, ordinary least squares and logit models, to analyze the influence of politicians' salaries and / or intent to be re-elected on corruption. The corruption indicator used is the total number of urban corruption cases (reported in the press) in each municipality during this period. They find that, first, relatively higher salaries do not reduce politicians' incentives for corruption and, second, when politicians want to be re-elected, corruption decreases.

Voter Tolerance for Corruption 

Two (2) work groups are identified: while the first group suggests an information asymmetry between honest citizens and corrupt candidates; the second group, on the other hand, is based on the assumption that citizens are as corrupt as candidates. 

Regarding the first group of works, Rundquist et al. [22] and Rundquist and Hansen [74] examine the reasons which would lead voters to support corrupt politicians. A commonly accepted explanation is that voters are poorly informed or poorly informed about candidates. Thus, Bryce [28] and Levin [29], Ferraz and Finan [12], Chang et al. [30] emphasize information asymmetries in political agency relations. Indeed, a voter may be honestly unaware of the candidate's corrupt activities. The behavioral assumption in this first argument is that once citizens are made aware of corrupt behavior, they prefer uncorrupted politicians to corrupt ones. This hypothesis is found in the work of Perssone and Tabellini [23], Ferraz and Finan [56], Kurer [75], Chang, Lederman et al. [76]. For them, the votes for corrupt politicians therefore come from the information asymmetries that affect voters and politicians.

The second group of work starts from the observation that there are cases where voters support a candidate, knowing that he has been convicted of a crime. Lack of knowledge therefore cannot explain all the cases in which voters vote for a corrupt candidate.

For this reason, studies by Wilson [31], Gosnell [32], McCook [33], then Banfield and Wilson [34] suggest that voters can knowingly vote for a corrupt politician by exchanging their vote for a material advantage, by particularly economic rewards [77] [78]. Therefore, the electoral coercion of corrupt politicians cannot be accomplished because of the “venality” of the voters, themselves capable of being corrupted. According to Manzetti and Wilson [77], voters are affected by the provision of private goods by corrupt politicians.

There are two variations of this explanation. In the first case, voters' behavior stems from their subjective assessment of the benefits associated with these private goods. If these benefits outweigh the utility generated by the vote for these uncorrupted politicians, voters would support them. In the second variant, voters are trapped in a bad balance. In such a scenario, corrupt politicians are re-elected due to the uncertainty of future election results. This group would promote non-corruption to corruption; however, as they receive private property from an incumbent, they have no incentive to vote for their own candidate. Indeed, they fear that the worst-case scenario will happen (i.e. re-election of the corrupt politician without their support, in which case they would lose access to benefits; [75].

Berg et al. [35], then Rogow and Lasswell [36] argue that citizens vote for corrupt politicians because voters see corruption as "functional" to society. While some theorists see corruption as a "disease" or "pathology", others see it as a necessary evil. Citizens, in some political cultures, including some subcultures in the United States [79], see corruption as a reality, a normal practice in the exercise of public service (and private enterprise). Corruption is thus perceived by voters as a "necessary evil" in the development of public policies [31] [80]. 

Rundquist et al. [22] suggest that the success of corrupt or allegedly corrupt candidates is that citizens weigh corruption charges in their voting decision along with other matters.

For Anduiza et al. [37], then Solaz et al. [38], it is not correct to assume that every voter uses corruption as the sole criterion for choosing a candidate, regardless of partisan ideas or other loyalties within the group. If the voter values the honesty of the government, then the belief that the candidate is corrupt may be enough to crystallize that voter’s decision to vote against that candidate. If, on the other hand, a voter prefers the candidate's political party or sticks only to other relevant issues, the voter can discount any accusations of corruption against the candidate and vote for him in any case. This surrender is particularly easy to do when the corruption charges have not been proven. Furthermore, when voters are indifferent to accusations of corruption, they would ignore them when voting. 

Finally, Persson and Tabellini [26] focus on supply factors. For these authors, voters may find it costly to vote for their own candidate if they have strong preferences that differ from voters' ideal points (ideological, religious, kinship preferences, etc.) in the other dimension of society relevant question.

Model for Determining the Effects of Elections on Corruption 

We draw on the analysis of Benito, Guillamon, Rios and Bastida [40] which aims to address the influence of re-election on corruption. The dependent variable of their model is the level of corruption, measured by the number of cases of corruption in each municipality during the period 2004-2009. The independent variables are broken down into variables of interest and control variables. The variables of interest retained by the authors are salary and re-election, while the control variables are: population, income inequalities, level of education, urban income, debt, unemployment and tourism. Presented as an ordered Logit model, the relationship is estimated by the ordinary least squares method.

Considering the absence of individual data, we set ourselves apart from Benito et al. [40] by estimating the following macroeconomic model:

With:

The dependent variable retained is corruption. There are several definitions of corruption, but the most widely used is "the misuse of public office for personal gain, whether material or political." It is a symptom of the dysfunctional relationship between state and citizens which results in bribes, extortion and nepotism [81]. It is commonly conceptualized as "a bribe to a public official for a benefit which may or may not be earned" or "the sale by officials of government property for personal gain" [82-85]. According to the World Bank, “corruption is the act of using one's position as a manager of a public service for one's personal benefit” and according to the Asian Development Bank: “Corruption is any behavior by which agents in the sector public and private obtain improper and illicit enrichment, whether it is personal or for the benefit of relations, or whether they provoke it by taking advantage of their position”. According to Transparency International, "Corruption is the abuse of delegated power for private gain." The Parliamentary Assembly of the Council of Europe defines corruption as "the use and abuse of power for private gain". The definition given by the Multidisciplinary Group on Corruption (GMC) of the Council of Europe is slightly different: “corruption is unlawful retribution or any other behavior towards persons with responsibility in the public sector or the public sector. private, which contravenes the duties which they have by virtue of their status as State agent, employee of the private sector, independent agent or other report of this nature and which aims to procure undue advantages of whatever nature, for themselves or for a third party ”. These definitions show that the legal limits of the phenomenon are difficult to pinpoint with precision.

Due to the illegal and clandestine nature of corruption, it is difficult to obtain “hard data” on these activities or to measure them empirically [86-87]. Despite these difficulties, quite some research institutions, international organizations and even private venture firms have attempted to create proxies to assess countries' levels of corruption since the early 1980s [88]. At the macroeconomic level, the most commonly used corruption indices are: 

• The “Corruption Perception Index” (CPI) which is a widely cited measure created by Transparency International (TI) and covering data for different sample sizes since 1995. As a composite based measure From "poll polls," the CPI index ranks countries according to the average score each country is assigned by business executives, economists and financial journalists in various independent polls. This subjective estimator ranks more than 180 countries (since 1994) on a scale of 0 to 10, with 0 indicating the most corrupt nations, while 10 indicating the entirely clean ones;

• The "Guide International Country Risk" (ICRG) index is assembled by the Political Risk Services (PRS) Group monthly on the basis of expert opinions. It is a predictive tool for international investments, as it analyzes financial, economic and political risks. Specifically, it captures the likelihood that senior government officials will demand special payments and the extent to which illegal payments are expected across levels of government in each country [85,89]. The ICRG measure ranks countries from 0 to 6, with lower scores indicating higher corruption. As the only index providing annual data for more than 100 countries from 1984 to 1997, the ICRG measure is one of the most widely used indices of corruption [85,90-95];

• The World Bank's (WB) "corruption control" (CC) index reflects a perception. It indicates the extent to which public power is at the service of private interests and provides information on the phenomena of “capture” of the State by elites and private interests. Its performance score from 0 to 100. The highest score reflects the best situation; 

• The CGI Index is compiled by the World Economic Forum (WEF). This index covers 12 “pillars” that determine the quality of a country's institutions and economic productivity and provides a comprehensive picture of the competitiveness of 140 countries. It ranks the level of corruption with a scale of 1-7;

• The Institute for Management Development index (IMD index) is published in the World Competitiveness Yearbook and also aggregates data from surveys of global business leaders. This index ranks 61 countries according to their levels of corruption; In our analysis, we use the Corruption Perception Index (CPI) as a measure of the level of corruption, with regard to the availability of data. The key independent variable in our model is election. It is a binary variable taking the value 1 during an election period (presidential and legislative elections) and 0 outside. In order to isolate the effect of our variable of interest (election), it is necessary to control for the effects of other factors on corruption. Among the control variables retained, we have the population, level of education, debt and unemployment. Variables such as income inequalities, urban income and tourism were removed from the initial model due to the lack of statistical data in this area. In addition, we set ourselves apart from the existing literature on the relationship between election and corruption by integrating a new « ministrerial staff rotation » control variable still used very little in recent work, namely, the number of « ministrerial staff rotation ».

According to Billger and Goel [91], population can have two opposing effects on corruption. Indeed, a greater concentration of the population increases the possibilities of interaction between potential corruptors and those who receive the bribes, which results in more corrupt deals. On the other hand, a concentrated population can also increase the number of supervisors and have a dissuasive effect. Therefore, we would expect a positive impact of the population on corruption if the first effect is predominant [52,92]. However, the opposite result would be expected if the second wins [91,93].

We also suggest that the state budget can influence the level of perception of corruption in the country. Indeed, public expenditure is structured in such a way as to facilitate the capture of rent. Thus, unlike current spending, presenting large-scale projects or increasing public capital spending are opportunities created by politicians and bureaucrats to increase their rents and benefits [84,94-95]. In addition, various studies [93,96] have studied the influence of public revenues on political corruption. Using policy agency theories, they argue that an exogenous increase in tax revenue can increase the incidence of a general measure of corruption.

Based on the work of Freille et al. [98], Jorgensen and Hunt [99], Osaghae and Osaghae [100], Signé and Korha [101], then Hamada et al. [102], we also integrate press freedom as an explanatory variable of corruption. Indeed, it is widely believed that a partially or completely free press is an important part of the process of detecting criminals and therefore has a deterrent effect on corruption [103-104]. In addition, restrictions on the free press lead to higher corruption [98,105-106]. However, while most agree that the press plays a role in detecting corruption, there is some reason to suggest that its effectiveness may be overestimated [107-108]. The press can orchestrate and publish bogus campaigns and accusations against politicians if these stories are likely to gain the favor of the ruling party, public attention and increase sales. Or, just as important, journalists and the press can also be corrupt and choose not to report the evidence.

Finally, we introduce the variable « ministerial staff rotation » in our model in order to take into account "top corruption" (particularly, that of the executive). According to Transparency International (2019), citizens' perception of corruption among senior public officials has increased in recent years. In addition, in Gabon, 2019 was particularly marked by scandals relayed by the media, leading to the arrest of the former director of the cabinet of the Presidency of the Republic and of the ministers close to him for embezzlement of public money and money laundering. In 2017, a former prime minister of economy, trade, industry and tourism, and former minister of investment promotion, transport, public works, housing, tourism and planning was jailed for alleged corruption and embezzlement of public funds. While not caused by corruption scandals, this succession of cabinet shuffles gives some incumbents the possibility of embezzling public funds.

Unlike Benito et al. [40], we estimate the following model: 

With:

The data used to estimate the relationship between election and corruption come from the database of Transparency International (TI), the World Bank and the Official Gazette of the Gabonese Republic. The series cover the period 2000-2019. The first is the Corruption Perceptions Index (IPC) calculated by Transparency International (TI). Low values of the CPI correspond to a high level of corruption, while high values of the CPI correspond to low levels of corruption.

The World Bank (WB) database provides variables such as the rate of population growth (pop), oil revenues (roil), public capital expenditure (dcap) and the population index freedom of the press. The latter variable takes on a low value when the government exercises direct control over all news media, and a high value on which media are free. Regarding data relating to « ministerial staff rotation » (rem), they are taken directly from the Official Gazette of the Gabonese Republic. 

The estimation of the regression [2] is done by Robust Ordinary Least Squares (OLS), after making the variables stationary. This formulation presents to correct the problems of heteroskedasticity of the residues.

Table 1 shows that the elections, the number of « ministerial staff rotation », the population growth rate and the freedom of the press are stationary at level. In addition, the corruption perception index, public capital spending and oil tax revenues are stationary in first difference. These last variables must therefore be differentiated to make them stationary.

Table 1: Stationarity Tests

Note: This table presents the results of the tests applied to the variables in level I (0) and in first difference I (1), using a model with constant (c), then with constant and trend (c and t). The numbers correspond to the different test statistics. The present first generation tests are based on a null hypothesis of non-stationarity and an alternative hypothesis of stationarity. When the number is preceded by (*) then the null hypothesis is rejected at the threshold of 1%; for (**) the null hypothesis is rejected at the 5% threshold; for (***) it is rejected at the 10% threshold. Source: author's calculations.

Table 2: Determinants of the Corruption Perception Index

Note: The long-term determinants of well-being are represented here. Student's tests on each coefficient are based on a null hypothesis of significance at zero and an alternative hypothesis of non-significance at zero. When the coefficient is preceded by (*) then the null hypothesis is rejected at the threshold of 1%; for (**) the null hypothesis is rejected at the 5% threshold; for (***) it is rejected at the 10% threshold. Endogenous variable (ipc_t) R-squared: 0.4593

Figure 1: Estimation par les MCO

Figure 2: Test d’hétéroscédasticité de Breusch-Pagan-Godfrey

Figure 3: Estimation par les Moindres carrés robustes

Figure 4: Test de Normalité des résidus

Figure 5: Stationnarité des résidus

Figure 6: Autocorrélation des résidus

The estimation of the regression [2] by ordinary least squares, with the stationary variables, gives the results presented in Figure 1. However, the Breusch-Pagan-Godfrey test shows the presence of heteroskedasticity of the residuals (Figure 2). In order to correct this problem, we carry out a robust estimation (Figure 3). The tests for normality (Figure 4), stationarity (Figure 5) and autocorrelation (Figure 6) show that the residuals generated are Gaussian white noise. The results of the robust estimation are presented in Table 2 below:

The results of Table 2 reveal that the elections have a negative and significant influence (at the 1% threshold) on the corruption perception index, which amounts to saying that the electoral period favors the increase in corruption in Gabon. . Such a result supports the idea defended by Rundquist et al. [22] and Rundquist and Hansen [74], who found that voters do not always sanction corrupt candidates. Indeed, the results obtained show that during an election period, the corruption perception index decreases, which corresponds to an increase in corruption of 0.3798%, all other things being equal. It therefore appears that in Gabon, the presidential and legislative elections allow the maintenance of corrupt or allegedly corrupt candidates who participate in maintaining a system of corruption in the country. There are several reasons for this result. The first is the information asymmetry that exists between voters and candidates. The results show that press freedom has a significant negative influence at the 5% threshold on the CPI, meaning that an increase in the press freedom index of 1% results in an increase in corruption of 0.02181 %, all other things being equal. Thus, the Gabonese press does not play a role in detecting corruption, as suggested by Graber [107], Pharr and Putman [108] and Vaidya. She tends to orchestrate and publish accusations against politicians if the stories are likely to gain the favor of the ruling party and public attention to increase sales. Or, just as important, Gabonese journalists and press can also be corrupt and choose not to report the evidence. Aside from information asymmetries, Gabonese citizens are guided by partisan votes. They are led to vote for a corrupt or presumed corrupt candidate as long as he is in the political party.

In order to reduce corruption in the country, Gabon should promote a democratic system supported by a free press independent of electoral results and elections carried out in full transparency. In such a system, Gabonese citizens, fully informed about the actions of the candidates and the harmful consequences of corruption on confidence in institutions, well-being, equality and poverty, etc., will have the possibility of sanctioning or to vote for a corrupt or presumed corrupt candidate, according to their preferences. 

In addition, the results presented in Table 2 show a negative but not significant relationship between the number of ministerial staff rotation and the corruption perception index. The same is true for the rate of population growth and revenues from the petroleum sector. These four variables deserve special attention because there is a risk that citizens mistakenly view them as contributing to the increase of corruption in the country. Regarding public capital expenditure, the positive and insignificant relationship with the corruption perception index. This means that when making public investments, there is a risk that citizens will mistakenly view these achievements as evidence of less corruption in the country.

In this article we have studied the relationship between elections and corruption. We were particularly interested in Gabon, a small economy in Central Africa, by integrating a new control variable, often absent in the literature on African countries, namely: the number of ministerial staff rotation. After performing two stationarity tests on the variables, we estimated the direct influence of elections on corruption by the robust least squares method. More specifically, the results highlight a negative and significant relationship between the elections and the corruption perception index. Thus, the presidential and legislative elections promote corruption in the country. It also appears that the Gabonese press is an instrument manipulated by corrupt candidates to win the elections. Moreover, the number of ministerial staff rotation does not have a significant effect on corruption, but its negative sign shows that there is a risk that it will increase the level of perceived corruption in the country.

1. Rothstein, B. The Quality of Government: Corruption, Social Trust, and Inequality in International Perspective. Chicago: University of Chicago Press, 2011.

2. Pani, M. “Hold your nose and vote: Corruption and public decisions in a representative democracy.” Public Choice, vol. 148, no. 1-2, 2011, pp. 163–196.

3. Mauro, P. “Corruption and growth.” Quarterly Journal of Economics, vol. 110, no. 3, 1995, pp. 681–712.

4. Chong, A. and C. Calderon. “Institutional quality and income distribution.” Economic Development and Cultural Change, vol. 48, no. 4, 2000, pp. 761–786.

5. Gupta, S. et al. “Does corruption affect income inequality and poverty?” Economics of Governance, vol. 3, no. 1, 2002, pp. 23–45.

6. You, J.S. and S. Khagram. “A comparative study of inequality and corruption.” American Sociological Review, vol. 70, 2005, pp. 136–157.

7. Uslaner, E. Corruption, Inequality and the Rule of Law. New York: Cambridge University Press, 2008.

8. Tay, L. et al. “Detrimental effects of corruption and subjective well-being: Whether, how, and when.” Social Psychological and Personality Science, vol. 5, no. 7, 2014, pp. 751–759.

9. Della Porta, D. “Social capital, beliefs in government, and political corruption.” In Disaffected Democracies: What’s Troubling the Trilateral Countries, edited by S.J. Pharr and R.D. Putnam, pp. 202–228. Princeton, NJ: Princeton University Press, 2000.

10. Seligson, M.A. “The impact of corruption on regime legitimacy: A comparative study of four Latin American countries.” Journal of Politics, vol. 64, no. 2, 2002, pp. 408–433.

11. Anderson, C.J. and Y.V. Tverdova. “Corruption, political allegiances, and attitudes toward government in contemporary democracies.” American Journal of Political Science, vol. 47, no. 1, 2003, pp. 91–109.

12. Ferraz, C. and F. Finan. “Exposing corrupt politicians: The effects of Brazil’s publicly released audits on electoral outcomes.” Quarterly Journal of Economics, vol. 123, no. 2, 2008, pp. 703–745.

13. Krause, S. and F. Mendez. “Corruption and elections: An empirical study for a cross-section of countries.” Economics and Politics, vol. 21, no. 2, 2009, pp. 179–200.

14. Choi, E. and J. Woo. “Political corruption, economic performance, and electoral outcomes: A cross-national analysis.” Contemporary Politics, vol. 16, no. 3, 2010, pp. 249–262.

15. Peters, J.G. and S. Welch. “The effects of charges of corruption on voting behaviour in congressional elections.” American Political Science Review, vol. 74, no. 3, 1980, pp. 697–708.

16. Stockemer, D. and P. Calca. “Corruption and turnout in Portugal – a municipal level study.” Crime, Law and Social Change, vol. 60, no. 5, 2013, pp. 535–548.

17. Stockemer, D. et al. “Bribes and ballots: The impact of corruption on voter turnout in democracies.” International Political Science Review, vol. 34, no. 1, 2013, pp. 74–90.

18. McCann, J.A. and J.I. Dominguez. “Mexicans react to electoral fraud and political corruption: An assessment of public opinion and voting behavior.” Electoral Studies, vol. 17, no. 4, 1998, pp. 483–503.

19. Key, V.O. The Responsible Electorate. Cambridge, MA: Belknap Press, 1966.

20. Kramer, G.H. “Short-term fluctuations in U.S. voting behavior, 1896–1964.” American Political Science Review, vol. 65, no. 1, 1971, pp. 131–143.

21. Ferejohn, J.A. “Incumbent performance and electoral control.” Public Choice, vol. 50, no. 1–3, 1986, pp. 5–25.

22. Rundquist, B. et al. “Corrupt politicians and their electoral support: Some experimental observations.” American Political Science Review, vol. 71, no. 3, 1977, pp. 954–963.

23. Persson, T. and G. Tabellini. Political Economics: Explaining Economic Policy. Cambridge, MA: MIT Press, 2000.

24. Besley, T. and A. Case. “Does electoral accountability affect economic policy choices? Evidence from gubernatorial term limits.” Quarterly Journal of Economics, vol. 110, no. 3, 1995, pp. 769–798.

25. Besley, T. and R. Burgess. “The political economy of government responsiveness: Theory and evidence from India.” Quarterly Journal of Economics, vol. 117, no. 4, 2002, pp. 1415–1451.

26. Persson, T. et al. “Electoral rules and corruption.” Journal of the European Economic Association, vol. 1, no. 4, 2003, pp. 958–989.

27. Rudolph, L. and T. Däubler. “Holding individual representatives accountable: The role of electoral systems.” Journal of Politics, vol. 78, no. 3, 2016, pp. 746–762.

28. Bryce, J. The American Commonwealth, vol. 1, edited by L.M. Hacken. New York: G.P. Putnam's Sons, 1959.

29. Levin, M.B. The Alienated Voter: Politics in Boston. New York: Holt, Rinehart and Winston, 1960.

30. Chang, E.C.C. et al. “Legislative malfeasance and political accountability.” World Politics, vol. 62, no. 2, 2010, pp. 177–220.

31. Wilson, J.Q. “Corruption: The shame of the states.” Public Interest, vol. 2, 1966, pp. 28–38.

32. Gosnell, H.F. Machine Politics: Chicago Model, 2nd ed. Chicago: University of Chicago Press, 1967.

33. McCook, J.J. “The alarming proportion of venal voters.” Forum, vol. 14, 1892, pp. 1–13.

34. Banfield, E. and J. Wilson. City Politics. Cambridge, MA: Harvard University Press, 1963.

35. Berg, L.  et al. Corruption in the American Political System. Mooristown, NJ: General Learning Press, 1976.

36. Rogow, A.A. and H. Lasswell. Power, Corruption and Rectitude. Englewood Cliffs, NJ: Prentice-Hall, 1963.

37. Anduiza, E. et al. “Turning a blind eye: Experimental evidence of partisan bias in attitudes toward corruption.” Comparative Political Studies, vol. 46, no. 12, 2013, pp. 1664–1692.

38. De Vries, C.  et al. “Voter coordination and the punishment of corruption.” Working paper, Department of Government, University of Essex, 2017.

39. Ecker, A.  et al. “Corruption, performance voting, and the electoral context.” European Political Science Review, vol. 8, no. 3, 2016, pp. 333–354.

40. Benito, B. et al. “Can salaries and re-election prevent political corruption? An empirical evidence.” Revista de Contabilidad – Spanish Accounting Review, vol. 21, no. 1, 2018, pp. 19–27.

41. Callen, M. and J.D. Long. “Institutional corruption and election fraud: Evidence from a field experiment in Afghanistan: Dataset.” American Economic Review, 2015. http://dx.doi.org/10.1257/aer.20120427.

42. Abbink, K. “Staff rotation: A powerful weapon against corruption.” Mimeo, 1999.

43. Abbink, K. “Staff rotation as an anti-corruption policy: An experimental study.” European Journal of Political Economy, vol. 20, no. 4, 2004, pp. 887–906.

44. Bühren, C. “Staff rotation as an anti-corruption policy in China and in Germany: An experimental comparison.” Jahrbücher für Nationalökonomie und Statistik, vol. 240, no. 1, 2020, pp. 1–18.

45. Welch, S. and J.R. Hibbing. “The effects of charges of corruption on voting behavior in congressional elections, 1982–1990.” Journal of Politics, vol. 59, no. 1, 1997, pp. 226–239.

46. Vivyan, N.  et al. “Representative misconduct, voter perceptions and accountability: Evidence from the 2009 House of Commons expenses scandal.” Electoral Studies, vol. 31, no. 4, 2012, pp. 750–763.

47. Basinger, S.J. “Scandals and congressional elections in the post-Watergate era.” Political Research Quarterly, vol. 66, no. 2, 2013, pp. 385–398.

48. Praino, R.D.  et al. “The lingering effect of scandals in congressional elections: Incumbents, challengers, and voters.” Social Science Quarterly, vol. 94, no. 4, 2013, pp. 1045–1061.

49. Winters, M.S. and R. Weitz-Shapiro. “Lacking information or condoning corruption: When do voters support corrupt politicians?” Comparative Politics, vol. 45, no. 4, 2013, pp. 418–436.

50. Wagner, M.  et al. “Partisan bias in opinion formation on episodes of political controversy: Evidence from Great Britain.” Political Studies, vol. 62, no. 1, 2014, pp. 136–158.

51. Barro, R. “The control of politicians: An economic model.” Public Choice, vol. 14, 1973, pp. 19–42.

52. Alt, J.E. and D.D. Lassen. “The political economy of institutions and corruption in American states.” Journal of Theoretical Politics, vol. 15, no. 3, 2003, pp. 341–365.

53. Kostadinova, T. “Abstain or rebel: Corruption perceptions and voting in East European elections.” Politics and Policy, vol. 37, no. 4, 2009, pp. 691–714.

54. Fiorina, M.P. Retrospective Voting in American National Elections. New Haven, CT: Yale University Press, 1981.

55. Besley, T. “Political selection.” Journal of Economic Perspectives, vol. 19, no. 3, 2005, pp. 43–60.

56. Ferraz, C. and F. Finan. “Electoral accountability and corruption in local governments: Evidence from audit reports.” Institute for the Study of Labor Discussion Paper No. 2843, 2007.

57. Mesquita, B.  et al. The logic of political survival. Cambridge, MA: MIT Press, 2003.

58. Klasnja, M. et al. “When do voters sanction corrupt politicians?” Journal of Experimental Political Science, Cambridge University Press, 2020, pp. 1–11.

59. Rose-Ackerman, S. “Political corruption and democracy.” Connecticut Journal of International Law, vol. 14, no. 2, 1999, pp. 363–378.

60. Kunicova, J. “Democratic institutions and corruption: Incentives and constraints in politics.” In International Handbook on the Economics of Corruption, edited by S. Rose-Ackerman, pp. 140–160. London: Edward Elgar, 2006.

61. Golden, M.A. and P. Mahdavi. “The institutional components of political corruption.” In Routledge Handbook of Comparative Political Institutions, edited by R. Ruiz Rufino and J. Gandhi, pp. 404–420. New York: Routledge, 2015.

62. Carey, J.M. and M.S. Shugart. “Incentives to cultivate a personal vote: A rank ordering of electoral formulas.” Electoral Studies, vol. 14, no. 4, 1995, pp. 417–439.

63. Kunicova, J. and S. Rose-Ackerman. “Electoral rules and constitutional structures as constraints on corruption.” British Journal of Political Science, vol. 35, 2005, pp. 573–606.

64. Fackler, T. and T. Lin. “Political corruption and presidential elections, 1929–1992.” Journal of Politics, vol. 57, 1995, pp. 971–993.

65. Klašnja, M. “Increasing rents and incumbency disadvantage.” Journal of Theoretical Politics, vol. 28, no. 2, 2016, pp. 225–265.

66. Reed, S. “Political corruption in Japan.” International Social Science Journal, vol. 149, 1996, pp. 395–405.

67. Zimmerman, J. “The municipal accounting maze: An analysis of political incentives.” Journal of Accounting Research, vol. 15, Supplement, 1977, pp. 107–144.

68. Benson, G.C.S. Political corruption in America. Lexington: Heath Lexington Books, 1978.

69. Gardiner, J. The Politics of Corruption. New York: Sage, 1970.

70. Peters, J. and S. Welch. The effects of charges of corruption on voting behavior in congressional elections. University of Nebraska, Lincoln, 1978.

71. Nyblade, B. and S. Reed. “Who cheats? Who loots? Political competition and corruption in Japan, 1947–1993.” American Journal of Political Science, vol. 52, no. 4, 2008, pp. 926–941.

72. Ferraz, C. and F. Finan. “Electoral accountability and corruption: Evidence from the audits of local governments.” American Economic Review, vol. 101, 2011, pp. 1274–1311.

73. Beylis, G.  et al. Understanding corruption: Theory and evidence from the audits of local governments. United States: UCLA and UC-Berkeley, 2012.

74. Rundquist, B.S. and S. Hansen. “On controlling official corruption: Elections vs. laws.” Unpublished manuscript, 1976.

75. Kurer, O. “Why do voters support corrupt politicians?” In The Political Economy of Corruption, edited by Arvind K. Jain, pp. 63–86. London: Routledge, 2001.

76. Lederman, D. et al. “Accountability and corruption: Political institutions matter.” Economics and Politics, vol. 17, no. 1, 2005, pp. 1–35.

77. Manzetti, L. and C. Wilson. “Why do corrupt governments maintain public support?” Comparative Political Studies, vol. 40, no. 8, 2007, pp. 949–970.

78. Fernandez-Vazquez, P. et al. “Rooting out corruption or rooting for corruption? The heterogeneous electoral consequences of scandals.” Cambridge University Press, vol. 4, May 2016, pp. 379–397.

79. Elazar, D.J. American Federalism: A View from the States, 2nd ed. New York: Thomas Y. Crowell Company, 1972.

80. Ford, H.J. “Municipal corruption.” Political Science Quarterly, vol. XLIX, 1904, pp. 673–686.

81. Alatas, S.H. The sociology of corruption: The nature, causes and prevention of corruption. Singapore: Donald Moore Press Ltd, 1968.

82. Nye, J.S. “Corruption and political development: A cost-benefit analysis.” American Political Science Review, vol. 61, no. 2, 1967, pp. 417–427.

83. Rose-Ackerman, S. Corruption: A study in political economy. New York: Academic Press, 1978.

84. Shleifer, A. and R.W. Vishny. “Corruption.” Quarterly Journal of Economics, vol. 108, 1993, pp. 599–617.

85. Svensson, J. “Eight questions about corruption.” Journal of Economic Perspectives, vol. 19, no. 3, 2005, pp. 19–42.

86. Lancaster, T., and G. Montinola. “Comparative political corruption: Issues of operationalization and measurement.” Studies in Comparative International Development, vol. 36, no. 3, 2001, pp. 3–28.

87. Levy, D. “Price adjustment under the table: Evidence on efficiency-enhancing corruption.” European Journal of Political Economy, vol. 23, 2007, pp. 423–447.

88. Blackburn, K. et al. “The incidence and persistence of corruption in economic development.” Journal of Economic Dynamics and Control, vol. 39, no. 12, 2010, pp. 2447–2467.

89. Knack, S. and P. Keefer. “Institutions and economic performance: Cross-country tests using alternative institutional measures.” Economics and Politics, vol. 7, no. 3, 1995, pp. 207–227.

90. Alesina, A. and B. Weder. “Do corrupt governments receive less foreign aid?” American Economic Review, vol. 92, no. 4, 2002, pp. 1126–1137.

91. Billger, S. and R. Goel. “Do existing corruption levels matter in controlling corruption? Cross-country quantile regression estimates.” Journal of Development Economics, vol. 90, no. 2, 2009, pp. 299–305.

92. Meier, K., and T. Holbrook. “I seen my opportunities and I took 'em: Political corruption in the American states.” Journal of Politics, vol. 54, no. 1, 1992, pp. 135–155.

93. Hill, K. “Democratization and corruption: Systematic evidence from the American states.” American Politics Research, vol. 31, no. 6, 2003, pp. 613–631.

94. McChesney, F.S. “Rent extraction and rent creation in the economic theory of regulation.” Journal of Legal Studies, vol. 16, 1987, pp. 101–118.

95. Goel, R.K. and M.A. Nelson. “Corruption and government size: A disaggregated analysis.” Public Choice, vol. 97, no. 1, 1998, pp. 107–120.

96. Brollo, F. et al. “The political resource curse.” American Economic Review, vol. 103, no. 5, 2010, pp. 1759–1796.

97. Brollo, F. et al. “The political resource curse.” American Economic Review, vol. 103, no. 5, 2013, pp. 1759–1796.

98. Freille, S. et al. “A contribution to the empirics of press freedom and corruption.” European Journal of Political Economy, vol. 23, 2007, pp. 838–862.

99. Jorgensen, W. and J. Hunt. “Journalism, accountability and the possibilities for structural critique: A case study of coverage of whistleblowing.” Journalism, vol. 13, 2012, pp. 399–416.

100. Osaghae, E. and V. Osaghae. “Rethinking democratic governance and accountability in Africa.” Forum for Development Studies, vol. 40, no. 3, 2013, pp. 393–412.

101. Signé, L. and K. Korha. “Horizontal accountability and the challenges for democratic consolidation in Africa: Evidence from Liberia.” Democratization, vol. 23, 2015, pp. 1254–1271.

102. Hamada, B.I. et al. “Press freedom and corruption: An examination of the relationship.” Global Media and Communication, 2019, pp. 1–19.

103. Ahrend, R. “Press freedom, human capital and corruption.” Working Paper, vol. 2002-11, Delta, 2002.

104. Stapenhurst, R. The Media’s Role in Curbing Corruption. Washington, DC: World Bank Institute, World Bank, 2000.

105. Brunetti, A., and B. Weder. “A free press is bad news for corruption.” Journal of Public Economics, vol. 87, 2003, pp. 1801–1824.

106. Chowdhury, S.K. “The effect of democracy and press freedom on corruption: An empirical test.” Economics Letters, vol. 85, 2004, pp. 93–101.

107. Graber, D. “Press freedom and the general welfare.” Political Science Quarterly, vol. 101, 1986, pp. 257–275.

108. Pharr, S. and R. Putnam. Disaffected Democracies: What's Troubling the Trilateral Countries? Princeton University Press, California, 2000.