Abstract
Deviations from presidential term limits are potentially damaging to democracies in low- and middle-income countries because they allow for the dangerous consolidation of authority. One way to address this problem is to determine the critical institutions able to protect term limits. By evaluating a set of relevant historical cases using qualitative comparative analysis, this study highlights the combinations of institutions associated with the successful protection of presidential term limits in the past. The study finds that independent judiciaries, legislatures, electoral commissions, and militaries, strong political parties, and vibrant civil societies, working together in various combinations, are the institutions that most effectively protect presidential term limits when they are challenged in democratic countries. Exploiting these findings, large-N qualitative analysis is then used to examine the hypothesized regularities of the optimal pathways identified by the QCA to explain the causal mechanism behind the protection of term limits in the study cases.
Keywords
Introduction
Given the demonstrable democratic backsliding of the 21st century,1,2 protecting presidential term limits and preventing deviations from them assume ever greater importance. Term limit amendment or abolition is a good indicator of democratic backsliding and often one of the last steps in the concentration of executive power and a slide into autocracy. 3 Historically, once a president has successfully subdued other competing centers of power, democracy gives way to increasingly authoritarian rule, as witnessed in countries such as Nicaragua, Uganda, and Venezuela, where term limits were originally enshrined in the constitution but eventually removed by long-serving presidents.
Recent years have witnessed increased attempts by presidents in low- and middle-income countries – in particular in Africa and Latin America – to extend their stays in power. 4 Numerous studies demonstrate that “leaders who have evaded term limits tend to oversee governments that are progressively more repressive, corrupt, and unstable”, 5 and term limit deviation demonstrates a strong correlation with democratic deterioration.6–8 Cognizant of this connection to democratic backsliding, it would be useful to know which democratic institutions are most instrumental in protecting presidential term limits when they are challenged.
One potential way to approach this problem is to unpack democracy into its core building blocks – the institutions that underpin democratic governance and protect a country’s citizens and its constitution from potential abuses of power. The term limits literature highlights a strong correlation between the level of democracy of a country and the protection of term limits.9,10 Examining specific institutions within democracies can pinpoint which particular combinations of institutions can best protect presidential term limits when they are challenged by aspiring autocrats.
This study finds that various combinations of key institutions – the judiciary, the legislature, the electoral commission, the military, political parties (both opposition and ruling parties), and civil society – are most effective in protecting term limits when they are challenged in low- and middle-income democracies. While the term limits literature (along with the broader democracy literature) has highlighted the role played by the judiciary,11,12 the legislature, 13 civil society,14,15 and opposition parties16,17 individually in democracy protection, the roles of the military and the electoral commission in democratic backsliding have not received similar attention. Furthermore, the conjunctural, complementary nature of these combinations of institutions has also not been sufficiently emphasized. Qualitative comparative analysis (QCA), a methodology uniquely suited to conjunctural causality, has increasingly been used in international relations studies in recent years. 18 However, QCA has been underused for democratization studies 19 and has not commonly been applied to the study of term limits. This study adds to the existing literature in using QCA to home in on the specific combinations of institutions that protect term limits and builds upon this approach in using large-N qualitative analysis (LNQA) to unpack the causal mechanism at work in the different cases in the study. A key finding of this joint QCA-LNQA approach is that the resistance of multiple different institutions, as well as different types of institutions, is critical to protecting term limits when they are challenged. Given the prevalence of democratic backsliding and the accompanying attacks on independent institutions in these and other countries, this research is both timely and potentially impactful.
Section 1 Presidential term limit deviations, provides a literature review for the protection of presidential term limits, identifies the key institutions included in the study, and offers initial hypotheses as to which combinations of institutions might be most impactful. Section 2 Methods, explains the methods used including the qualitative comparative analysis (QCA) and large-N qualitative analysis (LNQA) methodologies used in the paper and justifies the appropriateness, fit, and complementarity of these methods. Section 3, Data, presents the data, explaining the decisions behind the study data, and describing the approach to data collection. Section 4, Results from the QCA, presents the results from the QCA and offers a discussion of its findings, and Section 5, Results from the LNQA, presents the results from the LNQA. Section 6, Discussion of the international dynamics of term limit protection, provides a discussion of international applications of the paper, and Section 7, Conclusion.
Presidential term limit deviations
This article aligns with the vast majority of studies in the literature that support presidential term limits as an important constitutional provision to protect democracy and prevent the dangerous consolidation of power by the executive.20–23 Within that literature, numerous studies focus on the external factors or considerations that shape incumbent presidents’ decisions to leave peacefully at the end of their constitutionally-appointed terms or to try to remain in power, such as a president’s level of popular support 24 ; the amount of income generated by the office of the presidency25,26; and the possibility of post-office immunity versus prosecution.27,28
The key institutions involved in term limit deviations
While external factors like foreign aid and international pressure can certainly shape the context in which term limits are protected or altered, a country’s internal political dynamics factor more prominently in the ultimate fate of presidential term limits.29–31 Several studies provide insights as to which institutions might play important roles in protecting presidential term limits.
The two most common methods of securing additional presidential terms have been constitutional amendments, proposed through legislation, or even new constitutions altogether; therefore, the legislature and the judiciary represent key potential veto players and can provide powerful horizontal checks on presidential authority if they remain independent. When these branches of government work cooperatively, they present a significant check on executive power and a formidable obstacle to executive aggrandizement. Many presidential systems, which use the US Constitution as a template, feature these strong checks and balances, which are often eroded by executive aggrandizers seeking to modify or eliminate presidential term limits. Tellingly, results from empirical studies demonstrate that weaker legislative opposition and compliant judiciaries are associated with increased likelihood of attempts at term limit deviation.32–34
Anecdotal evidence from the cases in the study supports the conclusion that legislatures and judiciaries, working in conjunction, represent a strong protection for presidential term limits. In Nigeria 35 (2007), the independent legislature blocked President Olusegun Obasanjo’s attempt to push through a constitutional amendment to increase the number of terms; because the Nigerian Supreme Court was also considered independent, the president had no other pathway to term limit alteration and accepted defeat, peacefully stepping down later that year. In contrast, despite losing a popular referendum to extend term limits in Bolivia 36 (2016), President Evo Morales’ hand-picked Supreme Court declared that it was his constitutional right to run for re-election just 3 years later; the legislature, also under his control, offered no resistance.
Working in conjunction, a truly independent legislature and the judiciary present the most significant counterweight to presidential authority and protection for term limits. These horizontal institutions provide the most powerful and direct checks on presidential ambition given their constitutional mandates and authority. These contentions lead to the first hypothesis:
A key internal dynamic in any democracy is the party system. Countries with strong party systems are less likely to allow presidents to extend their terms in office given the incentives at play. From the perspective of the ruling party, internal political competitors will want their shot at the top office. From the perspective of opposition parties, they will want to maximize their electoral chances and ensure the most competitive environment possible (ideally without having to face a powerful incumbent). Therefore, actors in both groups have incentives to protect presidential term limits within the context of a strong party system. If the ruling party is under the sole control or influence of the president, however, that party is less likely to protect term limits.
The term limits literature confirms the importance of the role of both incumbent and opposition parties. The principal hurdle to securing an extension of term limits is the control of the president of their own coalition37,38; several other studies corroborate this finding.39–42 In addition to the president’s own party, opposition parties are also often the first line of defense against term limit extension or manipulation 43 and are key institutions in the fight to preserve term limits.44,45 Having a pluralist incumbent party along with strong opposition parties means that there will be legitimate challengers both from within and outside of the president’s party; because of the overlapping incentives of these challengers and their interests, the protection of term limits becomes more likely.
Again, the anecdotal evidence from relevant case studies highlights the ability of incumbent and opposition parties to protect term limits. In Zambia 46 and Malawi, 47 the fracturing of the president’s party alongside strong opposition parties thwarted presidential attempts to extend term limits. In Sri Lanka 48 and Senegal, 49 opposition parties took on and defeated incumbent presidents who fractured their own parties when they changed the constitution to be able to run for a previously-unconstitutional third term. This leads us to the second hypothesis:
Outside of the combinations of horizontal and party institutions mentioned in the previous hypotheses, other combinations involving these factors and others may effectively protect term limits. Diagonal institutions – like the media and civil society – are not directly a part of government but still impact their functioning and provide essential checks on government authority and policy. An independent media has been established in the term limits literature as an effective potential check on presidential power.50,51 Additionally, numerous scholars have found that an active civil society has served as an essential deterrent to presidents seeking to alter term limits.52–54 Comparative country case studies have also found that term limit challenges are much less likely to succeed where a free press operates55,56 and where vibrant civil societies forge coalitions with political parties and a range of different social actors to oppose potential deviations.57–59
These diagonal institutions are complementary to each other and to other democratic institutions as well. A free press can inform civil society about corruption or malfeasance within a presidential administration or even their plans to extend their constitutionally-appointed terms. In Benin, 60 Burkina Faso, 61 the Dominican Republic, 62 and Paraguay, 63 civil society played a central role in organizing and protesting against potential additional presidential terms after learning about it from the press; in each of these cases, the president backed away from or withdrew the challenge to term limits in the face of widespread popular opposition. Civil society protests may embolden the judiciary to take a strong stance to uphold the constitutionality of term limits. Media reports can trigger congressional inquiries or lead to corruption cases, as happened in Colombia when President Alvaro Uribe’s eventually unsuccessful bid for a third term took a hit when bribery and other scandals were uncovered by the press. 64
Other prominent institutions, such as the military and electoral commissions, that are loyal to the state and not a particular leader, may also serve to protect term limits in conjunction with other institutions. In consulting the Africa Executive Term Limits database, the military emerged as one of the most effective forces to stop term limit manipulation in the cases where it intervened. 65 At the opposite end of this spectrum, Harkness finds that “leaders who can count on the aid of a co-ethnic army are both more likely to challenge term limits and more successful in so doing”. 66 If a leader controls the legislature, the judiciary, and the military, as happened in Venezuela with Hugo Chavez, 67 they have the legal authority alongside the threat of physical force to extend their stay in power. Finally, independent electoral commissions that can conduct free and fair elections as well as monitor and uphold the integrity of electoral results may be critical to prevent term limit deviations. 68 An independent electoral commission in concert with a free press, for example, can publish the accurate results of an election to voters to prevent attempted tampering or election fraud on behalf of an incumbent president. This leads us to the third hypothesis:
While these hypotheses posit combinations of factors that should be effective in protecting term limits, they are not exhaustive or exclusive, as the final hypothesis suggests.
The parameters of the study
Within the universe of the study of term limits deviations, there are two distinct points of analysis. The first point happens when the president decides whether to challenge term limits in the first place. Stronger institutions naturally dissuade would-be term limit deviators, 69 therefore there are undoubtedly numerous cases of potential deviation that never happened because presidents did not think that they would succeed or felt that the barriers were too high. Several studies focus exclusively on the factors that influence a leader’s decision whether to attempt to deviate from term limits.70,71 The second point, however, occurs when a president has already decided to challenge term limits, and the subsequent analysis determines which factors lead to the success or failure of these challenges. The focus of this study is exclusively on the second point and the factors that can protect term limits when they have been challenged.
As an initial definitional note, this paper uses ‘term limit deviation’ to refer to any instance where an incumbent president attempts to alter the time that they can remain in office. While the term ‘violation’ is commonly used in the literature, sitting presidents often amend the constitution to remain in power, thus even though they have betrayed the spirit of the law, they have not broken the revised letter of the law as it now stands; therefore, this article uses the designation ‘deviation’.
Standards for determining term limit deviations and challenges
To establish a consistent standard, a term limit deviation is judged to have occurred “when a president is permitted for whatever reason, to serve a longer time than the maximum amount of time prescribed at the time of election to office” for at least 1 year or longer. 72 A president can accomplish this by several means: changing the number of legal terms allowed; restarting the clock on a first term by passing a new constitution; lengthening the term in office; or even eliminating term limits altogether. 73 For the purposes of this paper, all of the preceding methods of extending a president’s time in office are classified as term limit deviations.
It is also important to define what constitutes an unsuccessful term limit challenge. Unsuccessful challenges either leave term limits unaltered or do not meet the specified standard (e.g., a president overturned term limits but did not remain for a full year in power as happened in Niger and Guinea). An unsuccessful challenge was determined to have occurred if there were deliberate, demonstrable attempts to change term limits that both did not yield any changes to the length or number of presidential terms and that reached the public eye and were reported by multiple media outlets.
The contribution to the literature
Given that democratic institutions work collaboratively in high-functioning democracies to reinforce one another, a study that investigates the conjunctural nature of possible paths to term limit protection across numerous cases is necessary. In seeking to understand how best to protect presidential term limits, there is a need for an approach that can effectively evaluate the universe of cases of term limit challenges in both breadth and depth. However, while there are numerous large-N regression analysis studies74–77 and small-N case studies78–82 on the protection of term limits, there is currently no such study that can account for the multiple paths and combinations of conditions that can lead to the desired outcome.
To address this gap, this study uses a qualitative comparative analysis (QCA) methodology that employs set-theoretic methods to determine the combinations of institutions 83 that protect term limits in the democratic countries in which they are challenged. In doing so, it hopes to address the causal complexity posed by this persistent threat to democratic governance by highlighting the pathways that are more effective at term limit protection. In further unpacking this causal complexity, large-N qualitative analysis (LNQA) is then used to unearth the causal mechanism at work in these cases to better understand the process through which presidential term limits are protected.
Methods
Qualitative comparative analysis
QCA and LNQA were chosen in combination with one another to provide complementary methods that together offer both breadth and depth. QCA was developed by the sociologist Charles Ragin to account for causal complexity using set-theoretical models to determine the combinations of variables that result in a particular outcome. 84 Of the two types of QCA – crisp set and fuzzy set analysis – fuzzy set was selected for this study because of its ability to account for greater nuance in the data. While crisp set QCA codes data in a binary fashion (as either 0 or 1), fuzzy set QCA allows for partial set membership (in addition to full or no membership) of cases for each variable. 85 All of the institutions selected as conditions for the study are measured as continuous variables, therefore the use of fuzzy set QCA best reflected the data and the use of partial set membership. 86 For a more detailed explanation of the QCA methodology and specific software version used, see Supplemental Appendix A.
Large-N qualitative analysis
LNQA is an emerging methodology in international relations and comparative politics developed to assess a larger set of relevant cases by examining hypothesized regularities and establishing causal mechanisms to account for them. 87 Regularities represent the likelihood that when we have outcome Y, we also see treatment X (known as Y conditions), or, as is the case in this study, the likelihood that when we have prescribed treatment X, we see outcome Y (known as X conditions). The hypothesized regularity (or generalization) represents the expectation of a “constant conjunction” between treatment X and dependent variable Y. 88 The regularity is calculated as the percentage of times that X results in Y and produces a value between 0 and 1. As with QCA, we are looking for regularities with values greater than 0.75 to establish strong causal generalizations. The methodology allows for multiple X treatments that can lead to the outcome Y and has therefore been used in conjunction with QCA before 89 due to its ability to identify conjunctural causation.
After the regularity of the treatment(s) and outcome has been determined, within-case studies of each case in the study (or a significant sample of cases) are used to develop a causal generalization that explains the relationship between the treatment and outcome. Because all or many of the cases in the sample are evaluated, we can be more confident that the generalizations developed accurately reflect causal processes for the entire population. 90
This study employs a specific approach called “causal process observation” developed by Haggard and Kaufman 91 in their comprehensive LNQA work Dictators and Democrats, which is itself based on the process-tracing case study methodology pioneered by George and Bennett. 92 Causal process observation (CPO) is novel and appropriate for this study as it examines “relatively rare events” (like presidential term limit challenges) and selects all of the relevant cases in a “cross-national panel model for closer scrutiny”. 93 CPO is applied to the cases within the scope to determine whether they fit with the causal mechanism laid out in the theoretical model. All 34 cases in the study were individually examined and coded using CPO, with a particular focus on the 20 cases in which presidential term limits were successfully protected to see the distribution of cases and key institutions within cases. Certain illustrative cases were examined in greater depth for their ability to illuminate the causal mechanism. It is important to note that not all cases need to conform to the causal mechanism for it to remain broadly valid, as there will be conforming cases as well as non-conforming cases.
Integrating QCA and LNQA
Though they are both case-based qualitative methodologies, QCA and LNQA take fundamentally different approaches to arrive at their findings. QCA uses cross-case causal analysis to determine necessary and sufficient conditions or (as in this study) combinations of conditions that lead to the desired outcome. 94 LNQA, on the other hand, uses within-case causal inference to explain causal mechanisms and then to determine the extent to which the hypothesized causal mechanism is present and working as predicted across the universe of relevant cases. 95 While they differ in their fundamental analytical approaches to the study of cases, QCA and LNQA have been used together in previous studies 96 to complement one another to better explain complex causation at work in specific research questions.
In this study, the research question asks which democratic institutions best protect presidential term limits. QCA identifies the combinations of conditions that protected term limits, and LNQA reveals the regularities of these combinations and the causal mechanisms at work through an in-depth examination of the relevant cases. 97 Therefore, working sequentially, QCA can determine effective combinations of institutions, and LNQA can then explain how these combinations work in the cases within the scope to produce the desired outcome. In identifying which combinations have worked and in understanding how they worked, we can better protect term limits.
Mahoney also used these methods complementarily when studying the comparative economic development of post-colonial Latin American countries. 98 He first used QCA to ascertain the combination of conditions that determined levels of economic development. After settling upon the most effective combination of conditions that produced development, Mahoney then investigated each case that conformed to the theorized regularity to trace and confirm the causal process. A similar sequence is used in this study to determine the conditions that lead to term limit protection through QCA and a better understanding of how they do so through LNQA.
Data
Both the QCA and LNQA methodologies require a set of relevant historical cases bound within a defined scope to support effective analysis. QCA in particular requires these cases and a list of conditions or variables through which to evaluate them. The initial data for the study came from a purpose-built dataset constructed from available data and secondary sources that examined all instances of challenges to term limits in low- and middle-income presidential regimes from 1989 to 2020. The research conducted for the original dataset found sixty-three cases of challenges to presidential term limits within this timeframe.
Within this initial set of cases, the countries can be broken down by democratic status into democracies and non-democracies. To delineate democracies from non-democracies, I rely on a recent study by Kasuya and Mori 99 that empirically determined that the optimal cut-off point for democracies using different measures of democracy; according to their results, the ideal cut-off point for V-Dem’s data was 0.42 for the Electoral Democracy Index in the year in which the term limit challenge occurred. When I applied this cut-off point to the 63 cases, 34 qualified as democracies whereas 29 did not. Among the 29 cases that did not qualify, there were no cases in which term limits were successfully protected. This validates the decision to focus exclusively on democracies for the factors that might be critical to protecting term limits, though a further study that examined institutional dynamics within authoritarian regimes could prove insightful. The full list of all 63 cases where term limits were challenged from 1989 to 2020 can be found in Supplemental Appendix B.
Examining all of the 63 cases of term limits challenges, the vast majority – approximately 80% – occurred in Africa and Latin America. Figure 1 shows all of the countries in the world in which term limits were challenged and which countries experienced term limit protection and deviation (or both at different times).
Map of all countries where term limits were challenged, 1989–2020.
To further bound the cases, the year 1989 was used as a point of embarkation because it marked the end of the Cold War and the concomitant proliferation of democracies, democratic constitutions, and presidential term limits as a constitutional feature, in particular in Sub-Saharan Africa. 100 Only low and middle-income countries were included in the study because presidential term limits have rarely been challenged in high-income democracies and never during the time period in question according to the data. In the coming years, studies that include high-income countries may be warranted given the democratic backsliding that is occurring in long-standing democracies like the United States and others; however, for the purposes of this study, we focus on the existing universe of cases in low- and middle-income countries. Furthermore, this narrower group of countries is likely to be more comparable in terms of the relative maturity of their political institutions and as such may face similar democratic challenges and hurdles and may warrant similar protective measures and strategies.
Selection of institutions for inclusion in the QCA
The aforementioned studies from the literature review provided the foundation for the theories that were developed to explain the combinations of institutions that prove most effective in protecting presidential term limits. Using this literature as a guide, eight institutional factors were selected to serve as the analytical lens through which to evaluate the thirty-four historical cases: independent legislatures, judiciaries, electoral commissions, and militaries; strong opposition parties; divisions within the ruling party; vibrant civil societies; and a free press. The full justification for the inclusion of each institutional factor and its link to presidential term limits is presented in Supplemental Appendix C.
Data for the institutional factors in the QCA
The data used in the QCA was drawn from publicly-available sources including the Varieties of Democracy (V-Dem) Project, Freedom House, and Reporters Without Borders. V-Dem’s Version 13 Dataset indices were used to measure independent legislatures, independent judiciaries, independent electoral commissions, independent militaries, strong opposition parties, division within the ruling party, and vibrant civil societies. 101 The exact indices or combinations of indices used to measure each institutional factor can be found in the table in Supplemental Appendix D. Freedom House’s Freedom Press Index was used to measure the independence of media from 1989 to 2018, 102 and Reporters Without Borders data was used from 2019 to 2020 103 after Freedom House’s Index was discontinued as neither dataset covered the entire range of years in the study. The raw data for each of the 8 institutions for all 34 cases can be found in Supplemental Appendix E.
Calibration of the QCA data
With a full set of cases, components (the institutional factors), and data, we can begin to conduct the fuzzy set QCA. The first step is to calibrate the data and determine whether conditions are fully in, partially in, or fully out of the membership set which correspond to fuzzy set scores of 1, 0.5, and 0, respectively. The direct calibration scores as well as the data source for each condition are provided in Table 1.
Fuzzy set calibration for outcome and conditions.
Inputting the raw data along with the direct calibration scores into the fs/QCA software (version 4.1) yields the calibrated data for all of the components across the different cases on a scale of 0–1 that can then be used to construct a truth table. The table of calibrated data scores for every case and component can be found in Supplemental Appendix F.
Analysis of necessity and sufficiency (truth table)
The first step in conducting a QCA is determining if there are any necessary or sufficient conditions 104 that lead to the outcome. A necessary condition exists when a factor is present in every outcome or leads directly to the desired outcome and is determined through an analysis of necessary conditions calculated through the fs/QCA software. To qualify as a necessary condition, a component should have a 0.90 score of consistency for its presence or absence. 105 None of the eight institutions included in the study reached this threshold. The empirical results for the analysis of necessary conditions are listed in Supplemental Appendix G.
To determine sufficient conditions (or more aptly, combinations of conditions) that can lead to the desired outcome, the fs/QCA software constructs a truth table using the calibrated data. The truth table displays all of the possible combinations of conditions and shows the level of consistency, raw coverage, and unique coverage (concepts defined and explained in Supplemental Appendix A) for each of these combinations. Given that this study uses eight conditions, there are 2 8 possible combinations for a total of 256. Table 2 shows the abbreviated truth table containing only the 23 rows that have empirical cases associated with them.
Truth table for the outcome term limit protected.
LG: independent legislature; JD: independent judiciary; EC: independent electoral commission; ML: military; OP: strong opposition party; DV: division within ruling party; CS: vibrant civil society; FP: free press; TLP: term limit protected; the cases in bold have >50% membership in the outcome. The logical remainder cases are not shown in this abbreviated truth table but are available in Supplemental Appendix G.
To prevent against unwarranted assumptions, all logical remainders (solutions that do not have cases associated with them but have a consistency above 0.5) that contained only one condition toward the outcome were excluded. It follows that all remaining logical remainder rows that contain at least two conditions have been treated as plausible counterfactuals. 106 The full truth table with all 256 rows, including all logical remainders and plausible counterfactuals can be found in Supplemental Appendix G.
Results of the qualitative comparative analysis
Fuzzy set QCA results
The intermediate analysis (the optimal of the three models according to many QCA experts107–109) yields three unique pathways to term limit protection. Equifinality appears to be at work given that multiple pathways to the same outcome (term limit protected) exist, and QCA is particularly suited to addressing this phenomenon. 110 Not surprisingly, there is some overlap between the three pathways in regard to the specific institutions that lead to term limit protection with an independent judiciary, legislature, electoral commission, and military, division within the ruling party, and a vibrant civil society appearing in multiple solutions. None of the solution sets overlap exactly with the first two hypotheses espoused in the introduction; however, there are several institutions from these theories that feature in these sets – notably legislatures, judiciaries, and ruling parties – with two that were not part of a predicted combination: independent electoral commissions and militaries. Paths 2 and Path 3, with their all-hands-on-deck solutions, support the theorized equifinality espoused in Hypothesis 3. The overall solution has a consistency of 0.826 and a coverage of 0.443; therefore, the three paths offer a highly consistent explanation even if they only account for around half of the set-membership values for the outcome. The fact of lower coverage is a limiting factor in the QCA, but this is hopefully mitigated by the in-depth study of cases in the accompanying LNQA. These three pathways are illustrated in Table 3.
Solution for the outcome term limit protected.
The black circles represent the presence of a condition while the circle with an X represents its absence. Cases in bold have >0.50 membership in the outcome.
The solution terms for the complex and parsimonious QCA solutions can be found in Supplemental Appendix H.
Combinations of institutions most effective in protecting term limits
The first pathway to protect term limits includes the combination of an independent judiciary, electoral commission, and military alongside a strong opposition party. This combination of factors does not square neatly with any of the initial hypotheses. However, it is noteworthy that each of these factors possesses strong institutional measures to force compliance with term limits individually. Courts can rule additional terms unconstitutional; electoral commissions can refuse to list an incumbent candidate who has run afoul of term limits or refuse to rig elections; militaries can remove term limit violators from office or disobey unlawful orders to keep them in power; and opposition parties can vote down constitutional amendments to extend or abolish term limits.
When we assess their combined, joint effectiveness, we get an idea of how these institutions reinforce one another. An independent judiciary can provide legal protection and cover for opposition parties (and civil society organizations, for that matter) and can uphold the integrity, funding, and mission of the electoral commission. An independent military can ensure the rule of law in abiding by court decisions and not giving in to executive whims or unlawful executive orders. An independent electoral commission can ensure that opposition party candidates have a fair chance to contest elections and take their turn in power. Finally, a strong opposition in place offers a viable alternative to the president and party in power that can seek lawful paths to power under the guarantee of the rule of law enshrined by an independent judiciary and enforced by an independent military. Examples of this combination of institutions successfully protecting term limits can be found in Supplemental Appendix I along with examples of the repercussions of their absence.
The second pathway again features an independent judiciary and military but also includes an independent legislature, division within the ruling party, a vibrant civil society, and a free press. This solution set validates both the first and third hypotheses. Hypothesis 1 predicted the importance of horizontal checks on presidential power. Both of these institutions are present here, and, consistent with the theorized effect, afford significant protection to term limits when they maintain their independence from the president. Working hand-in-hand, an independent judiciary can protect the constitutionally-proscribed powers of the legislature and limit the extent of executive aggrandizement. And an independent legislature can safeguard the judicial appointment process to ensure that only unbiased, qualified judges are confirmed.
When these institutions are further reinforced by an independent military, a vibrant civil society and a free press, this combination of factors becomes even more potent. Civil society organizations (CSOs) can reinforce the power of the legislature and judiciary to stand up to the president while the military can protect them by refusing to disperse protests or commit violence against peaceful demonstrators. A free press can check presidential power in bringing to light abuses of power and episodes of state corruption that in turn lead to CSO demonstrations. And division within the ruling party is a potentially strong check on the president from within their own party that is reinforced by the presence of two co-equal branches of government. Examples of the presence and absence of the institutions in this solution can again be found in Supplemental Appendix I.
The third pathway includes most of the same elements from the second pathway for their presence (legislature, judiciary, division within the ruling party, and civil society) while substituting an independent electoral commission for an independent military. Equifinality is again at play with a slightly modified collection of factors working collaboratively to protect term limits. This pathway also features the absence of a free press. And while a free press is undoubtedly a force for transparency and accountability, the presence of other strong, independent institutions can compensate for its absence. Again, see Supplemental Appendix I for an example.
Robustness checks
The two most substantial robustness checks involved running a crisp set QCA model in addition to the fuzzy set model and alterations to the cut-off points for the direct calibration of the key components. Both checks were important to determine if and how they altered the results discussed in this section.
As previously mentioned, there are two different QCA approaches: fuzzy set and crisp set analyses. Unlike the fuzzy set analysis that uses continuous, calibrated measurements from 0 to 1, a crisp set QCA codes each variable as 0 or 1 in a binary fashion for all of the components for each case. A value of 0 indicates that the institutional factor is not present while a value of 1 indicates its presence. The cross-over points from the fuzzy set analysis were used as the cut-off values to determine the absence (scored as 0) or presence (scored as 1) of a condition in the crisp set analysis. The truth table for the crisp set QCA can be found in Supplemental Appendix J.
The results from the crisp set QCA are highly similar to those of the fuzzy set QCA. Most of the optimal pathways again include a combination of independent judiciaries, electoral commissions, and militaries along with strong opposition parties. Other pathways that protected term limits included independent legislatures, division with the ruling party, and vibrant civil societies. The intermediate, conservative, and parsimonious solutions for the crisp set QCA can be found in Supplemental Appendix K.
As an additional robustness check, the cut-off points for all factors were slightly lowered in one model and raised in a subsequent model in the direct calibration to see if it changed the results of the fuzzy set QCA. When the cut-off points for the calibrations were changed to make them lower and higher, almost all of the same institutions (judiciary, legislature, electoral commission, civil society, military, and opposition party) remained in the optimal pathways to term limit protection. However, when the cut-off points were lowered, a free press entered into some of the optimal solutions. When the cut-off points were increased above the original values, an independent military disappeared from the optimal solution. Based on the data and the judgments made as to the optimal cut-off points, the original analysis remains the most plausible, but a free press should not be dismissed as a potentially salient institution for term limit protection, as the third hypothesis suggests.
Further considerations and caveats
Because of the considerable overlap and interconnectivity between many of the aforementioned institutions, questions of endogeneity are inevitable and to some extent justifiable. The extent to which one institution bleeds into another and the degree to which they can be considered truly independent, siloed entities is certainly debatable. When institutions are attacked by illiberal democrats or aspiring autocrats, they rarely attack just one institution. Democratic deterioration happens not just to one institution but to many at the same time. The legislature and judiciary may have their independence undermined while the president attempts to co-opt the military to ensure its loyalty to the regime (as in Venezuela under Chavez). Furthermore, the weakening of one institution can have an effect on the ability of others to resist. For instance, if the judiciary is weakened, it undermines protections for civil society and the free press, as they are likely to see their freedoms and civil liberties dwindle or vanish (as in Belarus under Lukashenko). These interrelated and compounding attacks necessarily make the identification of the specific institutions that are essential to protecting term limits more difficult.
While possible endogeneity is a valid critique of this study and other QCA studies more broadly, the attempt to home in on the combinations of institutions that may better protect term limits and merit closer protection themselves nonetheless retains its value as a guide for pro-democracy actors. And if other, complementary institutions are inadvertently strengthened through the protection of those that are deemed more critical, this would not represent a bad outcome or collateral effect.
Results of the large-N qualitative analysis
Regularity of the treatments and outcome
The first step in conducting an LNQA is to determine the regularity of the outcome given the treatments. The outcome – the protection of presidential term limits – remains the same as in the QCA. The treatments are the combinations of institutions that lead to the outcome. In the previous section, the QCA pinpointed the most effective pathways in protecting term limits: 1. independent judiciaries, electoral commissions, and militaries and strong opposition parties and 2. independent legislatures and judiciaries, division within the ruling party, and vibrant civil societies (the second and third pathways overlap for these key institutions, therefore they do not require separate analyses). While these two pathways merit specific attention given the results of the QCA, it is important to note that they are not the only ones that can lead to the outcome, as equifinality is at play in the context of term limit protection.
Calculating the regularities of these two pathways yields results that are higher than our established threshold of 0.75 as a first step toward establishing meaningful causal generalizations. The first pathway yields a regularity of 0.88 while the second pathway has a regularity of 0.83. When these combinations of institutions are not present, the regularity drops sharply to 0.33 for other combinations. Table 4 shows these results along with the associated conforming and non-conforming cases.
Regularities for optimal pathways (X treatments) for term limit protection (outcome Y).
X1 represents the pathway that includes independent judiciaries, electoral commissions, and militaries and strong opposition parties. X2 represents the pathway that includes independent legislatures and judiciaries, divisions within the ruling party, and vibrant civil societies. X3 represents all other combinations of components that do not include X1 or X2 pathways.
Numbers do not add up horizontally because of case overlaps in solution sets.
The causal mechanism
The next step in the LNQA process is to determine the causal mechanism at work. Examining the process of challenges to term limits and whether they are ultimately protected or altered reveals several key junctures in the causal chain. While institutions certainly can and should be shored up and strengthened in low- and middle-income democracies before they are challenged, 111 the process of term limit protection is often a reactive one: the incumbent president attempts to stay in power by altering or ignoring existing constitutional mandates to step down at the end of their appointed term(s). Once this challenge reaches a public and actionable stage, institutions generally respond by either acquiescing or resisting. If no substantial institutional resistance emerges, either because the president has total control over them or they are too weak to mount an effective opposition, the term limit challenge sails through. However, in all the democracies included in this study save one (Argentina in 1994), at least one institution contested the challenge.
The next phase of term limit protection involves individual institutions or, more likely, combinations of institutions responding to this challenge. This response can be led by intra-governmental or extra-governmental institutions. Intra-governmental responses involve actions taken by the horizontal institutions (the legislature or the judiciary) or challenges from the party system, be they from within the president’s own party or from opposition parties. These intra-governmental challenges can be highly effective since the legislature, judiciary, electoral commission, and political parties have specific, powerful tools to halt the challenge directly: a legislature can vote down an amendment to change the constitution (Malawi in 2002), a supreme or constitutional court can rule additional terms unconstitutional (Colombia in 2010), an electoral commission can refuse to buckle to incumbent pressure to disqualify potential challengers or rig the election (Benin in 2006), or political parties can block these efforts politically or legislatively (Paraguay in 2017). The data from the cases discussed in the following section confirms this hypothesis.
Extra-governmental responses can come from civic organizations (civil society or the media) or the military. While the military is technically a part of and run by the government, it is a unique entity that operates in its own silo and is ideally separated from the realm of politics, though the degree to which this is true varies considerably from democracy to democracy. Militaries have rarely intervened to protect term limits (though they did in Niger in 2009), but their political neutrality and non-action is a key factor in protecting term limits. Civil society organizations and protests have proven effective at preventing presidents from altering term limits (Burkina Faso in 2014). And while a free press has not protected term limits by itself, it represents a potentially important element in a mix of institutions that can successfully resist executive aggrandizement by bringing important information about corruption or malfeasance to light (Bolivia in 2016). Finally, many responses employ both intra- and extra-governmental-led responses. These combinations, as the QCA demonstrates, often prove to be the most successful given the efficacy of the all-hands-on-deck approach highlighted earlier.
Once democratic institutions have responded to the term limit challenge, the president can either capitulate to the demands to respect the constitution, or they can double down on the challenge. If the president capitulates, term limits remain protected. If the president doubles down, three causal pathways emerge, and the cooperation and combinatory power of institutions can be critical to the eventual results. The first pathway unfolds when a president is strong enough to circumvent the institutional resistance (Ecuador in 2008) or simply ignore it (Belarus in 1996); term limits are then altered or even eliminated. The second pathway unfolds when institutions with the constitutional or martial authority to force the president to relent (the legislature, judiciary, the electoral commission, or the military) are able to do so (Honduras in 2009). Finally, the third pathway unfolds when institutions without this explicit authority (the ruling or opposition parties, civil society, or the media) can nonetheless put sufficient pressure on the president to give up the attempt to secure additional terms (Zambia in 2001). Figure 2 offers an illustrative view of the causal mechanism of term limit protection.
Causal mechanism for term limit protection.
Data from the case studies
Because we are mainly concerned with the protection of term limits, the primary focus remains on the 20 cases in which term limits were successfully protected. Within these 20 cases, 20% featured the use of exclusively intra-governmental responses, 15% exclusively extra-governmental responses, and 65% both types of responses. Moreover, 80% of the cases witnessed the involvement of more than one institution in the protection of term limits, with 40% of cases seeing the involvement of three or more key institutions. The analysis of each case, categorizing the type of institutional response along with the key institutions involved in the protection of term limits, can be found in Supplemental Appendix L.
In the 14 cases in which term limits were not protected, 36% featured exclusively intra-governmental-led responses, 7% exclusively extra-governmental-led responses, 50% both types of responses, and 7% no response. Tellingly, in only one of these cases did more than two institutions respond (Namibia in 1999). A broader institutional response is highly correlated with and is more effective at term limit protection in the study cases: this is a key finding of this study. The analysis of each case of unsuccessful protection can be found in Supplemental Appendix M.
Key case studies
As an additional part of the within-case analysis conducted for the LNQA, I return to the two optimal pathways to term limit protection identified by the QCA. Causal process observation further illustrates how two cases – Nigeria (2007) and Honduras (2009) – help to explain and flesh out this process. The Nigerian case, which illustrates the second optimal pathway, is included below while the Honduras case, which illustrates the first optimal pathway, can be found in Supplemental Appendix N.
In addition to these two conforming cases that highlight the causal mechanism at work in the two optimal pathways, I have included five additional cases in Supplemental Appendices O and P to showcase different examples of both conforming cases (Zambia, Malawi, and Benin (2006)) and non-conforming cases (Brazil, Nicaragua), respectively. Examining a variety of cases is important to demonstrate alternative causal pathways according to best practice suggested by Haggard and Kaufman. 112 Finally, I include cases from different regions (Central America, Sub-Saharan Africa, and South America) as well as from low-income countries, low-middle-income countries, and middle-income countries to ensure regional, temporal, and development-level variance.
Nigeria (2007)
Olusegun Obasanjo first rose to prominence as a general during Nigeria’s brutal civil war but managed to emerge from the conflict as a harmonizing figure for his messages of national and ethnic unity. 113 Obasanjo burnished his democratic credentials and endeared himself to Nigerian democrats by ending military rule in 1979 and ensuring a peaceful handover of power to a civilian-led government. 114 In the subsequent decade, he was one of the few leaders of national stature to challenge the brutal military rule of General Sani Abacha, earning him four years in prison. 115
Obasanjo won what is considered to be Nigeria’s first truly democratic election for president in 1999, and a wave of optimism surrounded his win. Despite his national appeal, he was still an ethnic Yoruba and Christian, so his power base remained stronger in the southern region of Nigeria, and his candidacy and regime encountered more opposition from the Muslim region in the north. Overlapping ethnic and regional tensions in Nigeria have persisted since the country’s inception. In 2007, after serving two terms as president, Obasanjo was term-limited, but there was skepticism as to whether there would be a democratic transfer of power. Upon taking office, he ensured that he had personalized control over his People’s Democratic Party (PDP), appointing loyalists to key positions. 116 Furthermore, because he maintained direct access to the oil revenues coming from the Ministry of Petroleum, he was able to ensure control over key financial resources 117 for political purposes.
For additional context, it is important to remember that despite his democratic credentials, Obasanjo was still the former head of a military-led government in the late-1970s. Given his military background, he was often resentful of any legislative oversight or scrutiny and therefore had a contentious relationship with the legislature, which he considered an ‘irritant’. 118
Inciting incident
In 2005 and early 2006, Obasanjo’s people and party began to mount a pressure campaign to revise the 1999 constitution that would include an amendment to increase the number of presidential terms from two to three. 119 Immediately, numerous opposition legislators voiced their objection to this proposal. 120 Moreover, civil society organizations, which had been clamoring for a review of the constitution for liberal democratic ends, were also largely against the amendment to term limits. 121 In a strategy common to other presidents in a similar position, Obasanjo was careful to stick to the party line in public about serving out his constitutional terms, while behind the scenes, he consistently used his proxies and “suspense, subterfuge, intimidation, and bribery” 122 to push for a third term.
Causal mechanism at work
Causal process observation reveals how the Nigerian case conforms to the stipulated mechanism. The challenge to term limits was made official when Obasanjo’s People’s Democratic Party (PDP) formally introduced the previously-mooted constitutional amendment in the Senate in February 2006. 123 As a part of the initial institutional response, opposition members of the legislature pushed back against a possible third term 124 in an intra-governmental-led response (a horizontal challenge). At the same time, the announcement provoked demonstrations in the Muslim northern region of the country and editorials against the action in prominent newspapers in an extra-governmental-led response (civic resistance).
Despite this initial multi-front resistance, Obasanjo doubled down, using his position as president and leader of the majority party as well as his access to the country’s considerable oil wealth to pursue his objective. He pressured members of his PDP to support the constitutional amendment and bribed these and other legislators. 125 However, this pursuit of a third term also disturbed the perceived ethnic balance of power within Nigeria as it was assumed (or perhaps tacitly agreed) that the presidency should pass to a Muslim from the north after eight years in the hands of a Christian from the south. 126 Even Obasanjo’s vice president, himself a northern Muslim, came out publicly against a possible third term, though as a potential presidential candidate, he also had personal as well as ethnic and regional interests at stake. The division with the ruling party, attributable to the regional dynamics and federalist structure of Nigerian government, can therefore be considered as an additional intra-governmental institutional constraint on presidential power as well, given the resistance of certain governors and other important players in regional politics.
In response to the president doubling down, the institutional cooperation represented by the collective efforts of the legislature, civil society, the media, and the division within the ruling party successfully resisted this attempt at term limit alteration. While ultimately it was the legislature’s defeat of the constitutional amendment that forced Obasanjo’s compliance with term limits, the combined forces of these institutions helped to win the day and successfully protect term limits.
Key take-aways
In an illustration of the second optimal solution of the QCA, the combined impact of the legislature, civil society, and division within the ruling party (and to a lesser extent, the press and the regional governments) proved successful in protecting term limits. While the judiciary did not play a key role, it was considered independent and could not therefore be manipulated by Obasanjo as a means to sign off on an extension of term limits. In the final analysis, the Nigerian example does conform to the causal mechanism.
Discussion of the LNQA results
In assessing the broader LNQA results with the aid of the specific lessons of the seven in-depth case studies, certain key takeaways emerge. The first and perhaps most essential conclusion of the LNQA is that the protection of term limits is most effectively accomplished through institutional cooperation – in effect, safety in numbers. The regularity data from all of the cases highlights the relative success of institutional cooperation, and the in-depth case studies hammer home this point. When the judiciary can count on military enforcement (or at least non-intervention), its decisions are more impactful and binding. When civil society can demonstrate or advocate under an umbrella of judicial protection and without the fear of military crackdowns, protests are bigger and more vibrant and CSOs more vocal and insistent. When opposition parties feel that they are competing on a level playing field, buoyed by the support of fellow institutions, they are more willing to stand up to the president and party in power to fight for access to and alternation in power.
A second takeaway emphasizes how the QCA and LNQA provide a complementary view of both conjunctural causation and equifinality in protecting term limits. The QCA highlights conjunctural causation in the various optimal pathways that lead to term limit protection, while the LNQA details how the cooperation of multiple institutions is much more effective in fostering this same protection. Furthermore, both of these methods showcase the equifinality inherent in term limit protection as there is no one surefire prescription for protection, rather different combinations of institutions and institutional resistance can lead to the desired outcome. Both of these ideas are essential takeaways for democratic actors wishing to protect term limits: there is no one magic formula or silver bullet, and all avenues and institutions should be availed upon in the defense of term limits specifically and democracy more broadly. While a broad coalition of institutions and opposition actors can be more unwieldy and difficult to manage, the benefits that these combined efforts and coordination offer outweigh the possible downsides.
As a final caveat and limitation of the study, it is important to note that even when the specific combination of institutions conforms to one of the optimal QCA pathways, that this is still not a guarantee that the causal pathway identified by the LNQA will hold. There are cases that demonstrate that even when the institutions theoretically align, there are ways around term limits, and aspiring autocrats are often clever and persistent. However, what these complementary methodologies do provide are the best (if not foolproof) strategies to pursue to term limit protection based on the available data: multiple institutions from inside and outside the government working together to uphold the constitution and ensure the rule of law.
Discussion of the international dynamics of term limit protection
Though the focus of this paper is largely comparative, there remain systemic lessons to draw from the totality of these cases and the emerging trends. International relations scholars have sounded the alarm about the democratic decline of recent decades and have responded with their own remedies and prescriptions to halt or even reverse it. This study can help to glean the overarching commonalities in the efforts of aspiring autocrats to evade term limits as well as the successes of democratic actors in thwarting them, along with the specific institutions that merit sustained protection. One possible limit to the external validity of this study is that it focuses exclusively on low-and middle-income democracies; the direct applicability of its conclusions may be limited to this group, as extrapolations to high-income countries or authoritarian regimes would be necessarily less tenable and more speculative in nature. Nonetheless, established democracies should continue to meet the threat of executive aggrandizement by bolstering their democracy promotion efforts abroad (and certainly not cutting back on or halting them), given their proven (albeit limited) success. Studies have demonstrated that democracy aid to vulnerable countries lowers the risk that presidential term limits are successfully circumvented. 127
While the factors studied in this article examine exclusively internal variables, the protection of term limits in the face of the broader global trend of democratic backsliding remains an important consideration. The recent rash of military coups in the Sahel demonstrates the profound inter-regional (and potentially international) effect of democratic backsliding from one country to its neighbors. In El Salvador, the example of Nayib Bukele conveniently side-stepping the one-term presidential limit 128 in 2024 and eventually eliminating term limits entirely 129 in 2025 might embolden his Latin American colleagues to do the same. Other leaders of countries faced with persistent security challenges, like Brazil, Honduras, or Mexico, might be tempted to follow his example of intimidating the legislature 130 and civil society, 131 removing and replacing high court judges, 132 and using the military to enforce crackdowns on civil liberties and political rights. 133 And this precedent, as Venezuela and Nicaragua demonstrate, is dangerous to both Salvadoran and international democracy.
The key lesson of this study – that different combinations of intra- and extra-governmental institutions best protect term limits – tells us that safeguarding these institutions is vital for the survival of democracy and that an all-hands-on-deck approach is best. Therefore, a wide variety of international actors including democratic countries, pro-democracy NGOs, and multilateral institutions should actively support independent legislatures, judiciaries, militaries, and electoral commissions, strong political parties, and vibrant civil societies, if they are to be able to resist democratic backsliding and promote presidential turnover and alternation in power.
Conclusion
The findings of this study suggest that term limits are best protected (when challenged) through certain combinations of institutions: 1. the judiciary, the electoral commission, the military, and opposition parties; 2. the legislature, the judiciary, division within the ruling party, and civil society; and 3. different combinations of intra- and extra-governmental institutions that combine to resist executive aggrandizement. A recent paper on democratic backsliding highlighted three key, recurring attacks by would-be autocrats: “loosening horizontal constraints, limiting political and civil liberties, and undermining the electoral system”. 134 These attacks target many of the same combinations of institutions indicated as essential to protect term limits in this study. Reinforcing and safeguarding these key institutions should help to protect term limits, erect roadblocks to deter aspiring autocrats, and thereby aid in halting potential democratic backsliding.
Given the interaction effect and causal complexity involved, further studies that investigate the specific ways in which these institutional factors play off or reinforce one another may prove valuable. While country context undoubtedly matters, further medium- or small-N studies that examine this causal complexity in the protection of presidential term limits would be a welcome addition to the term limits literature.
Supplemental Material
sj-docx-1-ire-10.1177_00471178251410728 – Supplemental material for Protecting presidential term limits from aspiring autocrats: a qualitative comparative analysis and large-N qualitative analysis
Supplemental material, sj-docx-1-ire-10.1177_00471178251410728 for Protecting presidential term limits from aspiring autocrats: a qualitative comparative analysis and large-N qualitative analysis by Bill Gelfeld in International Relations
Footnotes
Funding
The author received no financial support for the research, authorship, and/or publication of this article.
Supplemental material
Supplemental material for this article is available online.
Notes
Author biography
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
