ABSTRACT∞
When transitional justice emerged as a field of practice amid the third wave of democracy, many hoped it would reinforce the rules and norms that prevent democratic decline. However, over half of all the democracies that transitioned from autocracy or experienced civil war in the last half-century have since undergone backsliding or breakdown. Is transitional justice partly to blame? This article theorizes six ways that transitional justice might contribute to democratic decline, or alternatively, to democratic progress. It then analyzes data from 118 democratic spells in 89 countries spanning 1970 to 2023 to evaluate which hypotheses are plausible. In line with theories focused on endogenous change, the findings suggest that transitional justice is actually associated with stronger electoral and judicial institutions, which reinforces the guardrails of democracy against autocratic reversion. Yet, while transitional justice might help prevent full democratic breakdown, it does not hinder democratic backsliding, particularly that which is driven by government restrictions on civic association.
INTRODUCTION
Transitional justice became an international policy concern amid the third wave of democratization.1 The fall of the Soviet Union, and the subsequent expansion of American democracy promotion, entailed renewed focus on political transitions, which necessarily brought a consideration of how states confront the worst misdeeds of their authoritarian past.2 Since 1970, the vast majority of transitional democracies – which we define as those that have emerged from a period of autocratic rule or that have faced armed intrastate conflict – initiated at least one type of transitional justice mechanism. These mechanisms, including prosecutions, reparations, truth commissions or bans from public office, varied substantially from one country to the next, but one could argue that the simultaneous diffusion of transitional justice that started in the 1990s constituted a global political movement. Though that movement’s agenda broadened to include postconflict peacebuilding and economic transformation, advancement of democracy and rule of law always remained the centerpiece. Indeed, many have speculated over the years that properly administered transitional justice invigorates liberal forces and helps regimes move from democratization to the consolidation of democratic rule of law.3 The present special issue on ‘afterlives’ is a good opportunity to revisit this often assumed relationship between transitional justice and democracy.
A glance at contemporary developments might invite skepticism.4 Despite the widespread adoption of transitional justice mechanisms over the last 30 years, the democratic gains of the late third wave have receded substantially. According to some experts, the world in 2024 is no more liberal than it was in 1989.5 Figure 1 depicts trends over the last 50 years. The dark gray bar graph shows the cumulative number of transitional democracies each year between 1970 and 2023. Of these 89 total countries, 70 pursued some form of transitional justice at some point over this period. Yet the number that later experienced democratic decline also began to rise considerably starting around 2004.6 As of 2023, 56 of these transitional democracies either backslid toward less liberal forms of governance or fully reverted to autocracy. Should this affect our judgments about the success of past transitional justice processes? Does the global decline of democracy represent a failure of transitional justice?
Transitional Democracies, Transitional Justice and Democratic Decline.
To these questions, many observers would likely answer ‘yes.’ For some, the transitional justice program, which emerged out of a flawed ‘transitions paradigm,’7 is elitist, excessively legalistic, neoliberal and neocolonial, serving to distract from local practices that might be more transformative.8 Because it is a byproduct of existing structures and institutions, its contributions to substantive democracy are either marginal or non-existent.9 For others, the uncompromising push for transitional justice is politically transgressive, and it may actually cause democratic decline by provoking backlash that can ultimately re-ignite violent conflict, unravel elite pacts or place unbearable strain on weak institutions.10 Recognizing ‘this dark side’ of transitional justice means ‘lowering expectations about the positive, causal connection that is often made between justice and democratization.’11
This article takes seriously the possibility that transitional justice is in part responsible for the recent ebb of liberal democracies around the world. Borrowing from and elaborating on prior work, we theorize an extensive list of causal pathways through which transitional justice might plausibly lead to democratic decline, and we derive six testable hypotheses. Briefly, these hypotheses predict that transitional justice leads to unfair elections, a weakened judiciary, limitations on civil association, greater political polarization, more robust antidemocratic social movements and less political engagement. Our aim here is theoretical synthesis. It is easy to find claims that transitional justice may have negative unintended consequences on governance, yet it is difficult to find an account that fits these claims into an overarching theory of democratic decline. Our first contribution is offering such a big-picture account.
Developing a theory, however, is not the same as believing in a theory. We are not attempting to normatively push the claim that transitional justice is bad for democracy. In fact, our previous work has shown that transitional justice can, under some conditions, lead to an easing of state-level repression because it sends signals to violent state agents that the old way of ruling, through surveillance and coercion, is no longer acceptable.12 In this respect, our scholarship might typically be grouped with those who find transitional justice to be a source of endogenous change toward greater liberal democracy and enhanced rule of law.13 Yet we also do not resolutely advocate this progress narrative. If anything, our theoretical priors are flat, meaning that we operate on the neutral assumption that transitional justice could either damage or reinforce democracy. For this reason, we pursue a strategy of two-way testing, where each hypothesis linking transitional justice to democratic decline is paired with an opposite, contradictory hypothesis that predicts transitional justice will lead to improvements in democratic governance. This allows us to assess a counter explanation for contemporary trends – that democratic decline is caused not by too much transitional justice, but by too little. If this were the case, the democracies undergoing autocratic reversal would be the very ones which never sufficiently heeded demands for justice, truth and reparation.
Following our theoretical presentation in the second section, the article turns to empirical assessment, which is the only way to evaluate competing interpretations of transitional justice’s long-term effects. The third section will present descriptive statistics on a sample of all transitional and conflict democracies between 1970 and 2023. Specifically, we compare practices in the part of the sample that remains democratic versus that part of the sample that undergoes democratic decline, defined as either autocratic breakdown or illiberal backsliding. These data show that transitional justice is far less common in democratic states that experience full autocratic breakdown, but just as common in states that merely backslide from liberal to illiberal forms of democratic governance. The fourth section then statistically tests contradictory hypotheses linking transitional justice to democracy. Ultimately, the article finds that transitional justice does not cause democratic breakdown, and in fact appears correlated with cleaner elections, greater judicial constraints and more democratic political engagement. However, transitional justice appears powerless to stop democratic backsliding, or de-liberalization, especially that which is defined by narrowing freedoms to political association. Finally, while transitional justice seems to exacerbate political polarization, this is not the avenue through which it leads to democratic backsliding.
LINKING TRANSITIONAL JUSTICE TO DEMOCRATIC DECLINE
Democratic decline includes both ‘backsliding’ and ‘breakdown.’ Though scholars and laypersons use these terms in myriad ways, we define backsliding as a process wherein a liberal democracy retrogresses to an electoral democracy. In this process of illiberalization, government leaders undermine certain rights protections or critical checks on executive power but maintain competitive elections that might still result in the alternation of rule. Breakdown, however, is when an electoral or liberal democracy fully reverts to an autocracy, meaning that the country ceases to hold competitive elections, and may even slip into non-elected forms of leadership selection. Both backsliding and breakdown are types of democratic decline, but they vary in terms of degree.
Democratic decline almost always occurs in one of these ways: through coups or through elite subversion.14 In the former, unelected armed actors capture the state apparatus by force and install an alternative autocratic government. In the latter, elected leaders chip away at democratic institutions until fewer meaningful checks on executive power remain.
We theorize that transitional justice could cause democratic decline either by inspiring elites to alter institutions in a less liberal direction or by leading to more antidemocratic public behaviors. In the former, transitional justice alters the strategic interactions of elites who control the way that democratic institutions function. If transitional justice inspires leaders to manipulate elections, remove judicial constraints on the executive or limit the freedom to engage in party formation or civic association, then it is a direct cause of backsliding or breakdown. In the latter path, transitional justice causes democratic decline by fraying the relationship between governing institutions and mass public behavior. If transitional justice leads to greater political polarization, more energized antidemocratic movements or diluted political engagement, then it creates the social conditions for backsliding or breakdown. Seeing beleaguered public support for the regime, rapacious leaders can take the opportunity to consolidate their own power at the expense of democratic institutions.
Figure 2 formalizes this theory, depicting the six possible causal pathways that we identify – three related to elite behavior and three related to public behavior. For transitional justice to cause democratic decline, it must either incentivize elites to attack or subvert democratic institutions or reshape collective behavior in ways that invite such antidemocratic attacks or subversions. Yet the question remains, how might transitional justice do this? And importantly, does transitional justice mainly push in the autocratic direction, or are there reasons to expect that transitional justice might actually prevent democratic decline? In the remainder of this section, we discuss the reasoning behind opposing hypotheses linking transitional justice to regime change.
Causal Pathways Linking Transitional Justice to Democratic Decline.
Elections
Many observers worry that transitional justice provokes backlash by – or at the very least fails to constrain – powerful actors with a rational interest in despoiling incipient democratic institutions. These actors include members of former autocratic networks, including the military, or emergent demagogues who see an opportunity to exploit regime weaknesses in a bid to become an unrivaled head of state. One major target for aspiring autocrats is the democratic regime’s signature mechanism of vertical accountability: free and fair elections. Electoral manipulation can involve legal actions like de-funding election monitoring bodies, engaging in discriminatory redistricting and enacting party proscriptions, or it can involve illegal forms like ballot stuffing, vote buying, intimidation and violent attacks on voters. In the face of efforts to achieve transitional justice, some political elites may employ these tactics to protect their own allies and marginalize pro-justice opposition.
Though undertheorized, there are reasons to expect that transitional justice may lead to the weakening of electoral institutions. First, transitional human rights trials have often been accompanied by threats from the military to end democracy in retaliation for being held criminally accountable. The case of Argentina is often cited here. Following the Trial of the Juntas in 1985, Argentine courts received a flood of criminal filings regarding disappearances and torture by soldiers and clandestine forces.15 Fearing accountability, the military threatened a coup against the new democracy, and President Raúl Alfonsín responded by supporting the Punto Final Law, which ended penal action against those who acted to ‘repress terrorism’ during the junta regime. There are other more obscure examples of this phenomenon as well. In 1987, Colonel Gregorio Honasan led a failed coup in the Philippines, in part as a response to new President Carazon Aquino’s attempt to prosecute soldiers for their actions in the fight against Communist insurgents.16 Prosecutions are not the only transitional justice mechanisms that can trigger right-wing actors to threaten democratic institutions. Salvadoran leaders reacted harshly to the publication of a UN-backed truth commission report in 1993, and demanded that an amnesty be passed protecting the military from future efforts at accountability.17 These instances point to the coup risk posed by transitional justice. Military elements and their allies might, seeking to preserve their own impunity in response to transitional justice efforts, simply ignore elections and overrun democratic institutions.
Second, lustration policies could also weaken electoral institutions. Lustration is a kind of ‘transparency regime’ that bans would-be officeholders from government posts because of their antidemocratic political allegiances, like involvement with secret police.18 This is often done for the sake of neutralizing political actors who might act in a way that is antagonistic to the new regime. However, in cases like the Czech Republic, lustration amounted to a de facto party proscription because it disqualified Communist party officials and members of the secret police from running in elections. The reason, according to reformer of the Czech security service Stefan Bačinsky, is that people from the old networks ‘certainly have the right to go with us to build the democratic society. But I am convinced that they should not lead us.’19 Pro-democratic disqualifications through lustration are similar in form to party bans that occur for antidemocratic reasons. Both are technically legal forms of electoral manipulation. Therefore, one may expect countries that have enacted officeholder bans to feature less robust elections because limitations on electoral competition become normalized (Hypothesis A).
Hypothesis A: Transitional justice is associated with more unfair electoral institutions.
Alternative A: Transitional justice is associated with less unfair electoral institutions.
Others might expect transitional justice to play an opposite role in new democracies, fortifying rather than endangering electoral institutions (Alternative A). The most notable way that transitional justice could improve elections is by sidelining cheaters. ‘Potential demagogues exist in all democracies,’ write Levitzky and Ziblatt, ‘But in some democracies, political leaders heed the warning signs and take steps to ensure that authoritarians remain on the fringes, far from the centers of power.’20 The biggest threat to electoral democracy emerges when ruling parties form political alliances with politicians who harbor a latent preference for autocracy.
Research suggests that transitional justice can help isolate enemies of democracy.21 First, prosecutions target autocrats for any of those actions they take to illegally manipulate elections, and in so doing, send a signal that future violations will also be punished. Greenstein and Harvey point to the example of South Korea, where former Presidents Chun Doo-hwan and Roh Tae-woo were prosecuted in 1996 for spending millions on graft and vote buying. A supplementary example is the case of Peru’s former President Alberto Fujimori, who was tried not only for human rights violations, but also for an extensive vote buying and blackmail operation run and videotaped by his spy chief.22 Second, bans from public office are perhaps better at eliminating legal forms of electoral manipulation, because they prevent pro-autocracy officials and bureaucrats from sabotaging democratic institutions from within.23 Through lustration, thousands of state agents can be reviewed for their prior involvement in human rights violations and other questionable activities, as they were in the Czech Republic. Nalepa discovers that this process improves the quality of democracy by limiting the extent to which secret authoritarian networks can actively participate in political competition.24 Therefore, one might expect that bans from office diminish the ability of pro-autocratic elites to cooperate to avoid compliance with electoral institutions.
The Judiciary
A second institution even more likely than elections to come under fire from aspiring autocrats is the judiciary. Courts are the main vehicle of horizontal accountability in a democracy; they are one of the primary checks on the power of the executive. Not only this, but courts are often tasked with implementing transitional justice policy. When called upon, independent judges must issue rulings that directly challenge the interests of executives and their allies, for instance when they preside over criminal prosecutions where the defendants are powerful members of the security community.
In this situation, pessimists may expect transitional justice to result in the hollowing out of judicial institutions, for three reasons. First, courts implementing transitional justice attract direct attacks on their independence in the form of public threats, bribery, impeachment and the turnover of personnel. These attacks can be preemptive or reactive. In Chile, General Augusto Pinochet preemptively warned the incoming civilian government in 1989 that ‘the day they touch one of my men, the rule of law ends.’25 In Sri Lanka, President Mahinda Rajapaksa reactively impeached Chief Justice Shirani Bandaranayake in 2013 for ruling against postconflict development projects that would be overseen by the president’s brother, but also in retaliation for the judge blocking other efforts by the Rajapaksa family to consolidate power.26
Second, when courts assert their positive judicial independence by issuing adverse rulings, they invite non-compliance from the executive.27 Though short of frontal assault, repeated non-compliance can decrease the legitimacy of the judiciary by slowly eroding its diffuse support.28 For instance, in 2013, the Nepalese Supreme Court ruled in the Basnet Judgment that the Government had run afoul of its international legal obligations by failing to create ‘appropriate, administrative, and judicial measures’ to pursue accountability for human rights violations.29 The judgment also stated that the provision of amnesties was unconstitutional. Nevertheless, the Government passed a Truth and Reconciliation Commission Act that included an amnesty, and generally proved ‘unwilling to comply with its [the Court’s] directives in this issue area.’30
Third, compounding these other problems, transitional justice caseloads can tax already overburdened judiciaries. Transitional and postconflict courts commonly operate in ‘institutionally imperfect’ environments, where they face a lack of resources and general support.31 Piling on hundreds or thousands of criminal or administrative cases for the courts to hear could ultimately lead to paralysis. Taken together, these lines of reasoning support Hypothesis B.
Hypothesis B: Transitional justice is associated with declining judicial constraints on the executive.
Alternative B: Transitional justice is associated with improving judicial constraints on the executive.
Optimists might expect transitional justice to be not a cause of declining judicial independence but a source of increasing returns to the judiciary (Alternative B). Case study researchers repeatedly note how transitional prosecutions have at times transformed and emboldened courts, leading to what Catalina Smulovitz calls the ‘discovery of law.’32 First, courts can serve as a site for renewed pro-justice activism, which results in pro-democracy institutional drift. In parts of Latin America and Europe, human rights advocates were able to mobilize the courts through rights to private prosecution, which ‘gives victims or their surviving relatives the opportunity to intervene in the pretrial and trial stages.’33 In these cases, private parties were able to achieve key successes that opened the path toward greater prosecutorial and judicial momentum over time.34
Second, transitional justice can serve as a focal point for efforts at institutional conversion. Cause-lawyers promoting transitional prosecutions in Argentina and other countries have acted in concert to change ‘cultures of legal interpretation’ and to turnover judicial personnel, ultimately reshaping the country’s courts.35 This kind of turnover could also be pursued directly through lustration policies, which are able to root out and dismiss pro-authoritarian judges, or truth commissions, which can advocate for reform of legal institutions through extensive recommendations.36
Third and finally, legalized transitional justice can serve a sociopedagogical function, strengthening norms against abuse by promoting ‘civil dissensus.’37 Through legal and administrative efforts at transitional justice, citizens in new democracies learn to deliberate and disagree productively over issues of national concern. While hardly any sector of society is completely satisfied with transitional justice given that outcomes are always suboptimal for at least one side, they help to transmit adversarial but nonviolent modes of dispute resolution. All this arguably strengthens judicial constraints on executive power.
Political Association
A final way that one might plausibly connect transitional justice to the decline of democratic institutions is through subversion of rights to political association. Civil society is a critical element of diagonal accountability in a democracy. And there are numerous ways that civil society organizations (CSOs), including nongovernmental organizations and free media, promote transitional justice. CSOs organize demonstrations, bring critical litigation and monitor the implementation of truth commissions and other transitional justice mechanisms.38 But these actions draw the ire of autocrats, who grow tired of persistent criticism and demands by CSO watchdogs. Increasingly in the last two decades, governments have chosen to pass laws restricting CSOs instead of meeting these demands. These laws limit foreign aid to local CSOs, make registration more difficult, censor publications and create a raft of other burdensome regulations.39 As Kenneth Roth writes:
By closing the political space in which civic groups operate – interfering with the funds that allow groups to organize, investigate, and agitate – autocrats are trying to suck the oxygen from organized efforts to challenge or even criticize their self serving reign.40
Research on the worldwide crackdown against CSOs is still developing, and this backlash is about more than transitional justice. However, in some cases, it may be possible to connect the dots between transitional justice and broader pressure on civil society. For example, human rights advocates in Guatemala have been under sustained attack for years. Dozens of human rights defenders have been murdered, and CSO offices have been burgled and ransacked.41 Much of anti-CSO sentiment is tied to human rights advocates’ support of postconflict and anticorruption prosecutions. For instance, former President Alejandro Giammattei supported the passage of Decree4-2020, which restricts and sanctions CSOs. Also, with the backing of the military and national police, he blocked the renewal of the Comisión Internacional contra la Impunidad en Guatemala, or CICIG, which was tasked with investigating corruption and former human rights violations in the country. This backlash, according to Amnesty International, was a:
reaction to the significant progress made since 2007 regarding access to justice in Guatemala, both in cases of crimes under international law committed during the armed conflict … and more recent cases driven by the CICIG.42
Evidently, the push for transitional justice in the country, which managed to produce a conviction of former General Efrain Rios Montt (later overturned), struck a chord. This example, combined with the overall trend toward narrowing space for CSOs, leads to Hypothesis C.
Hypothesis C: Transitional justice is associated with greater limitations on political association.
Alternative C: Transitional justice is associated with fewer limitations on political association.
More hopeful theorists might expect the opposite: that transitional justice could in fact lead to a flourishing of civil society and, in general, fewer legal restrictions on political association and assembly (Alternative C). First is through deterrence. Presumably, countries that are emerging from authoritarianism and violent conflict have already engaged in extensive repression of civil society. Punitive transitional justice goes some way toward denouncing this type of behavior and disqualifying those leaders who prefer repressive violence to other forms of accommodation. Research repeatedly shows that transitional human rights prosecutions are associated with greater respect for human rights over time, as well as less organized criminal violence.43 This deterrent effect could produce a more hospitable environment for civil society by narrowing the range of repressive options available to state agents. Second, some transitional justice mechanisms are a site for CSOs to articulate their demands against the government, some of which may involve rolling back prior restrictions. For instance, the Tunisian Truth and Dignity Commission (IVD) consulted extensively with CSOs, and thus reported on ‘security campaigns against all political groups and civil society organizations during the rules of Bourguiba and Ben Ali.’44 The commission went on to recommend that the government enact an Organic Law for Associations that respects nearly unlimited rights to form parties and other civic organizations.45 Should any of these kinds of recommendations exert influence on elites in government, it could conceivably result in more space for civil society.
Political Polarization
The three sets of hypotheses thus far all involve ways that transitional justice may cause elites to weaken rules governing vertical, horizontal and diagonal accountability. As depicted in Figure 1, however, there is another way that transitional justice could undermine democracy, by negatively affecting citizens’ attitudes, expectations and behaviors. One behavioral phenomenon that arguably creates conditions for democratic decline is political polarization. When societies ‘grow so deeply divided that parties become wedded to incompatible worldviews … stable partisan rivalries eventually give way to perceptions of mutual threat.’46 When political society becomes more derisive and antagonistic, it paves the way for the erosion of norms and invites subversion by autocratic demagogues who claim to have a solution to society’s ills.
Skeptics might argue that transitional justice, with its backward-looking emphasis on establishing blame for human rights violations, could easily devolve into a perpetually unsatisfied and polarized politics of ressentiment.47 While there is a ‘paucity of research on the correlation between transitional justice and polarization,’ a few case examples may illustrate the concern.48 In the former Yugoslavia, prosecutions for war crimes ultimately became a source of continued competition between ethnic groups, and publicized trials have created a platform for some high-profile defendants to develop a ‘war criminal cult.’49 This has only fueled previously existing inter-ethnic divisions.50 In many Eastern European states, former communists successfully framed lustration policies as mere expressions of ‘political vendetta,’ rather than earnest efforts to discover the truth and cleanse society of corruption. Writing on these policies, Csilla Kiss states that:
Politicians are often interested in the files when the information promises to potentially embarrass their opponents, and they are more eager to hide the information in cases where they might face embarrassment.51
And in Tunisia, opportunists like Abir Moussi, a populist and open proponent of former President Ben Ali, has significantly enlarged her following through ‘ruthless criticism of the IVD [truth commission], its work, its president and the very idea of reparation for victims.’52 Based on these and other examples, one might advance Hypothesis D.
Hypothesis D: Transitional justice is associated with greater political polarization.
Alternative D: Transitional justice is associated with less political polarization.
Alternatively, many have repeatedly predicted that transitional justice is a way of mending social fractures inherited from previous regimes (Alternative D). After all, citizens often demand transitional justice because polarization and antagonism already escalated to the point of social breakdown.53 In the postconflict order, altogether ignoring the past will create more problems than it will solve. As O’Donnell and Schmitter write:
it is difficult to imagine how a society can return to some degree of functioning which would provide social and ideological support for political democracy without somehow coming to terms with the most painful elements of its own past.54
However imperfect, and however dissatisfying, transitional justice cuts a middle path ‘between forgetting the past and engaging in violent retribution.’55 For instance, though it served as a lightning rod for criticism in the moment and in the three decades since,56 the South African Truth and Reconciliation Commission’s (TRC) amnesty-for-truth bargain seemed a necessary step to advance multiparty democracy after decades of apartheid brutality and violent resistance. The same could be said about Bosnia. According to UNDP, ‘People at all levels of society and from every part of the country see putting war crimes suspects on trials as a necessity with no serious alternative.’57 Likewise, Colombia’s extensive reparations program will likely help quell the understandable desire to seek maximal punishment of FARC and other rebel groups, a move that might only destabilize the 2016 peace settlement and reignite a decades-long civil war. Indeed, it is transitional justice’s potential to provide an outlet for channeling previously existing animosities that many find so promising.
Antidemocratic Movements
A second way that transitional justice could indirectly lead to democratic decline is by feeding into popular antidemocratic movements. This consists of peaceful or armed collective action that is opposed to democracy, meaning that followers are ‘not willing to respect constitutional provisions or electoral outcomes.’58 Transitional justice might propel antidemocratic movements in two ways. First, it can produce grievances among a group that feels underserved by democracy. In El Salvador, populist Nayib Bukele won a second term as president in 2024 despite norms against two-term officeholders. He did so by championing a brutal crackdown on gangs through emergency laws, which has led to the arrest of over 70,000 individuals. His supporters are legion, and they bristle at the human rights community for remaining fixated on prosecuting the decades-old El Mozote massacre instead of curbing gang violence. These ‘unwaveringly loyal’ supporters amount to an antidemocratic movement because they take little issue with Bukele’s extraconstitutional reign – he has referred to himself as the world’s ‘coolest dictator’ – nor his open hostility toward human rights and transitional justice.59
Second, transitional justice can itself be hijacked by autocratic movements. In Hungary, when Prime Minister Viktor Orbán’s Fidesz Party lost the 2002 elections to the Hungarian Socialist Party (MSZP), a Fidesz-backed newspaper released information that Orban’s successor, incoming Prime Minister Péter Medgyessy, had been a secret agent under the Communist regime. This Medgyessy scandal resulted from a cynical bid on the part of Orbán’s party to use secret files, which had been permitted under the policy of lustration, to discredit a political rival. This maneuver showed the autocratic tendencies of Orbán, who would rise to power again in 2010. Since, he has consolidated his rule as an ‘illiberal democrat,’ presiding over the demolition of executive checks, and pushing through the retroactive and unconstitutional trial of 92 year-old former Communist leader Béla Biszku for crimes committed in 1956.60 In Hungary, the hijacking of memory politics to serve a right-wing agenda, as well as the broader trend toward democratic backsliding, is embedded within a larger anti-democracy movement in the country. Ergo, Hypothesis E:
Hypothesis E: Transitional justice is associated with growing anti-democracy movements.
Alternative E: Transitional justice is associated with weakening anti-democracy movements.
Others would push back (Alternative E). Transitional justice in the Hungarian case was not the cause of anti-democracy movements. Instead, as Gábor Halmai writes, ‘the Hungarian population seemed not to be receptive toward legal constitutionalism in general, and the very formalistic approach of the rule of law in particular …’61 In fact, transitional justice in 1990s Hungary was rather weak. So too was transitional justice in El Salvador, where entrenched right-wing leaders and sympathizers met the findings of the truth commission and efforts at trials with denial and obstruction. That autocratic forces in some cases block transitional justice from making inroads does not mean that transitional justice causes anti-democracy movements to grow. In fact, new systematic evidence from survey experiments conducted in Guatemala and Chile demonstrates that exposure to transitional justice – in the form of prosecutions, reparations and memorialization – may in fact produce greater trust in democratic institutions and governance.62 If this is accurate, it means high-profile and well-implemented transitional justice mechanisms should remove grievances that fuel anti-regime agitation.
Political Engagement
A final way that transitional justice might indirectly lead to democratic decline is by depressing civic engagement. Critical commentators have identified two main ways this can happen. First, because transitional justice mechanisms are state sanctioned, and often highly legalistic, they redirect the energies of human rights CSOs away from civil disobedience, protest and other creative forms of demand making toward highly technical and statist forms of action, like legal reporting and litigation. The struggle for democracy gets tamed through the process of legalization.63 Second, in maintaining a preoccupation with fighting impunity for extreme civil and political rights violations, transitional justice mechanisms provide insufficient attention to the economic and social inequalities that animate the concerns of aggrieved citizens.64 In Egypt, for instance, trials of a handful of state agents under the post-Mubarak government only diverted attention away from the aims of the revolution.65 In contexts like Chile, Guatemala and South Africa, truth commissions presented ‘an interpretation of history as parable rather than as politics’ and ‘largely denied the [economic] conditions that brought them into being.’66 And in Timor Leste and Sierra Leone, survivors expected their basic needs to be addressed by transitional justice, only to be disappointed.67 This is a worrying possibility. Victims and survivors, expecting more and getting less from transitional justice, may lose the will to engage with new governing institutions altogether (Hypothesis F).
Hypothesis F: Transitional justice is associated with less political engagement.
Alternative F: Transitional justice is associated with more political engagement.
Optimists would expect the opposite, that transitional justice invigorates the democratic public (Alternative F).68 Sure, transitional justice cannot itself drastically reconfigure the political economy of any state. However, it can serve as a site for contestation, and raise the profile of human rights issues, which include economic redistribution in addition to the reparation of harm. Moreover, there are three further reasons to suspect that transitional justice catalyzes democratic political participation and engagement. First, the process of interacting with investigators, TRC commissioners and victim registries provides a platform to build CSO capacity. In Liberia, the Transitional Justice Working Group (TJWG) brought together nearly two dozen CSOs to lobby on behalf of the TRC Act, and this body engaged in extensive consultations with the Liberian TRC when it began operations.69 Given that Liberian civil society had been nearly decimated by a brutal civil war, these efforts were transformational. Second, micro evidence suggests that when ordinary citizens interact with transitional justice mechanisms, they are more likely to become reinvolved in political society. A study on the Chilean reparations program shows that former victims who received payments were more likely to register to vote, compared to those who had not received compensation.70 Third and finally, evidence suggests that bans from office, especially those that result in ‘bureaucratic turnover and renewal,’ increase the ‘trustworthiness of the government by changing its composition.’71 As trust increases, so too does the willingness of citizens to engage. In the end, it appears plausible that rather than sapping the will of citizens to mobilize, various types of transitional justice may in fact play a role in reconstituting social capital.
Toward Analysis
The aim of this section was theoretical exploration. Given the alarming trends away from liberal democracy in the last decade, it makes sense to ask whether transitional justice is part of the problem. However, the vast array of case-based inferences in the literature point in different directions, and the linkages between transitional justice and democracy are legion. Our contribution thus far has been to map six pathways – three direct and three indirect – through which transitional justice might feasibly contribute to democratic decline, and we presented the case on both sides of each causal hypothesis. We now turn to descriptive and correlative statistical analysis to assess which of these hypotheses are most plausible given the empirical record. The next section examines the relationship between transitional justice and democratic decline in general. Then, the fourth section will test the hypotheses linking transitional justice to the elite-led institutional and public behavioral changes identified in the discussion.
DESCRIPTIVE DATA ANALYSIS
Setup
Our analysis starts with basic descriptive statistics. First, we assembled a sample of all democracy years following (1) regime transition from autocracy or (2) the outbreak of a violent intrastate conflict, between 1970 and 2023.72 This produces 118 postautocratic or conflicted democracy spells in 89 countries. To arrive at postautocratic spells, we draw on the Transitional Justice Evaluation Tools (TJET) database. TJET codes democratic transitions based on three prominent scholarly democracy datasets: Polity5; the Boix, Miller and Rosato (BMR) dataset; and V-DEM’s Episodes of Regime Transformation (ERT).73 The database registers a transition start year when at least two out of the three democracy datasets indicate a significant change from autocracy to democracy.74 To determine whether a democracy ever experienced intrastate (including internationalized) conflict, we use the UCDP Armed Conflict Dataset.75 Democracy spells are included in our dataset even if they never pursue transitional justice, and they remain in our dataset every year after a transition or after the onset of violent intrastate conflict until the year 2023. If a regime undergoes full democratic breakdown, in effect changing from a democracy to an autocracy, it falls out of the analysis dataset.
Second, we divided these spells into three groups: (1) those that never underwent a democratic decline; (2) those that eventually backslid from liberal to electoral democracy; and (3) those that eventually suffered democratic breakdown, resulting in a reversion from liberal or electoral democracy to any form of autocratic government. Backsliding and breakdown are derived from the V-DEM ERT’s ‘regimes of the world’ measure.76 Of the 118 spells in our database, 48 (40.7 percent) experienced no decline; 15 (12.7 percent) experienced backsliding and 57 (48.3 percent) experienced democratic breakdown. In just two countries – Hungary and Serbia – did a democratic regime first backslide and then later break down. In the end, of all transitional democratic spells since 1970, the majority (71 of 118) underwent some form of democratic decline.77
Differences in Context
For each of the three outcome groups – no decline, backsliding and breakdown – we tabulate the mean values of a series of important contextual variables (Table 1, Rows 1–5). Regarding the contextual comparison, one can note that the democratic spells resulting in breakdown are clearly more unstable. First, these spells last, on average, 14 years less than those with no decline. However, spells that end in backsliding actually last just as long as those with no decline.78 Second, breakdown spells had a higher average number of coups and coup attempts than those that experienced no democratic decline.79 Still, coups do not appear to be the most important mechanisms of breakdown, as only 13 spells ending in breakdown also had coups in the same year. On this basis, one could infer that subversion of institutions is by far the most common form of democratic regression. Third, the average number of years with intrastate conflict activity was higher in spells with democratic breakdown (3.04) than in spells with no democratic decline (1.83). Relative to the average duration of spells, countries with very brief periods of democracy usually featured prolonged intrastate conflict. Spells with backsliding experienced the highest average number of years with active conflict, but this is due to a single outlier, Israel; without the Israeli spell, the number of conflict years in backsliding democracies was almost nil (0.07, in parentheses in Table 1).
Descriptive Statistics For Democratic Spells through 2023.
| Variable | No Decline | Backsliding | Breakdown |
|---|---|---|---|
| total number* | 48 | 15 | 57 |
| average duration of democratic spell* | 26.29 | 33.53 | 12.30 |
| average number of coups (incl. attempts, data through 2022)* | 0.21 | 0.00 | 0.61 |
| average number of years with active internal conflict (data through 2022) | 1.83 | 3.27 (0.07) | 3.04 |
| average legacy of violence score | −0.59 | −0.65 | 0.17 |
| average number of years with amnesties | 0.67 | 1.20 | 0.86 |
| average number of transitional human rights trials of state agents* | 15.69 (8.00) | 35.13 (6.50) | 2.30 (2.21) |
| average number of transitional human rights trials of high-ranking state agents | 5.19 (2.11) | 13.67 (1.79) | 0.40 (0.38) |
| average number of individuals convicted in transitional trials* | 44.38 (13.45) | 165.73 (10.86) | 2.84 (2.89) |
| average number of high-ranking individuals convicted in transitional trials* | 3.98 (1.64) | 20.20 (1.71) | 0.37 (0.38) |
| average truth commission publicness index | 0.71 | 0.40 | 0.44 |
| average proportion of spell-years with reparations policy | 0.38 | 0.40 | 0.22 |
| average proportion of spells with bans from office | 0.41 | 0.59 | 0.28 |
| Variable | No Decline | Backsliding | Breakdown |
|---|---|---|---|
| total number* | 48 | 15 | 57 |
| average duration of democratic spell* | 26.29 | 33.53 | 12.30 |
| average number of coups (incl. attempts, data through 2022)* | 0.21 | 0.00 | 0.61 |
| average number of years with active internal conflict (data through 2022) | 1.83 | 3.27 (0.07) | 3.04 |
| average legacy of violence score | −0.59 | −0.65 | 0.17 |
| average number of years with amnesties | 0.67 | 1.20 | 0.86 |
| average number of transitional human rights trials of state agents* | 15.69 (8.00) | 35.13 (6.50) | 2.30 (2.21) |
| average number of transitional human rights trials of high-ranking state agents | 5.19 (2.11) | 13.67 (1.79) | 0.40 (0.38) |
| average number of individuals convicted in transitional trials* | 44.38 (13.45) | 165.73 (10.86) | 2.84 (2.89) |
| average number of high-ranking individuals convicted in transitional trials* | 3.98 (1.64) | 20.20 (1.71) | 0.37 (0.38) |
| average truth commission publicness index | 0.71 | 0.40 | 0.44 |
| average proportion of spell-years with reparations policy | 0.38 | 0.40 | 0.22 |
| average proportion of spells with bans from office | 0.41 | 0.59 | 0.28 |
Variables not used in models presented in Figure 3. Transitional justice variables extend through 2020.
Descriptive Statistics For Democratic Spells through 2023.
| Variable | No Decline | Backsliding | Breakdown |
|---|---|---|---|
| total number* | 48 | 15 | 57 |
| average duration of democratic spell* | 26.29 | 33.53 | 12.30 |
| average number of coups (incl. attempts, data through 2022)* | 0.21 | 0.00 | 0.61 |
| average number of years with active internal conflict (data through 2022) | 1.83 | 3.27 (0.07) | 3.04 |
| average legacy of violence score | −0.59 | −0.65 | 0.17 |
| average number of years with amnesties | 0.67 | 1.20 | 0.86 |
| average number of transitional human rights trials of state agents* | 15.69 (8.00) | 35.13 (6.50) | 2.30 (2.21) |
| average number of transitional human rights trials of high-ranking state agents | 5.19 (2.11) | 13.67 (1.79) | 0.40 (0.38) |
| average number of individuals convicted in transitional trials* | 44.38 (13.45) | 165.73 (10.86) | 2.84 (2.89) |
| average number of high-ranking individuals convicted in transitional trials* | 3.98 (1.64) | 20.20 (1.71) | 0.37 (0.38) |
| average truth commission publicness index | 0.71 | 0.40 | 0.44 |
| average proportion of spell-years with reparations policy | 0.38 | 0.40 | 0.22 |
| average proportion of spells with bans from office | 0.41 | 0.59 | 0.28 |
| Variable | No Decline | Backsliding | Breakdown |
|---|---|---|---|
| total number* | 48 | 15 | 57 |
| average duration of democratic spell* | 26.29 | 33.53 | 12.30 |
| average number of coups (incl. attempts, data through 2022)* | 0.21 | 0.00 | 0.61 |
| average number of years with active internal conflict (data through 2022) | 1.83 | 3.27 (0.07) | 3.04 |
| average legacy of violence score | −0.59 | −0.65 | 0.17 |
| average number of years with amnesties | 0.67 | 1.20 | 0.86 |
| average number of transitional human rights trials of state agents* | 15.69 (8.00) | 35.13 (6.50) | 2.30 (2.21) |
| average number of transitional human rights trials of high-ranking state agents | 5.19 (2.11) | 13.67 (1.79) | 0.40 (0.38) |
| average number of individuals convicted in transitional trials* | 44.38 (13.45) | 165.73 (10.86) | 2.84 (2.89) |
| average number of high-ranking individuals convicted in transitional trials* | 3.98 (1.64) | 20.20 (1.71) | 0.37 (0.38) |
| average truth commission publicness index | 0.71 | 0.40 | 0.44 |
| average proportion of spell-years with reparations policy | 0.38 | 0.40 | 0.22 |
| average proportion of spells with bans from office | 0.41 | 0.59 | 0.28 |
Variables not used in models presented in Figure 3. Transitional justice variables extend through 2020.
Finally, spells with full democratic breakdown had much higher levels of human rights violations in the past than those without decline and those with backsliding. This is evident when examining the average scores on TJET’s Legacy of Violence Index, created to comparatively assess potential demand for transitional justice within 174 countries over time. Values on the Legacy of Violence Index capture how widespread human rights violations have been in each country since 1949, compared to the level of human rights violations of all other countries; higher values indicate more widespread violations. The index builds on the well-established latent physical integrity measure first introduced by Schnakenberg and Fariss,80 which was designed to assess the protection of core, non-derogable human rights – the rights to life and bodily integrity – over time.81 Because higher scores indicate higher levels of state violence over the course of history, one can clearly see that legacies of abuse are associated with the propensity for a transitional democracy to revert to autocracy.
Differences in Transitional Justice
We also present data on eight transitional justice measures, averaged by the different spell outcomes (Table 1, Rows 6–13). These measures are derived from the new TJET database of transitional justice mechanisms spanning 1970 to 2020.82 The bivariate comparisons suggest that transitional justice is generally associated with less democratic breakdown.
First, no-decline spells had, on average, fewer amnesties offered to state agents or rebels than breakdown spells. However, spells with backsliding had the most amnesties. While amnesties are often seen as necessary to bring former autocrats or rebels to the negotiating table, they are also exceptional laws that stand in the way of regular functioning of the criminal justice system. Thus, many view them as obstacles to the development of rule-of-law institutions.
Second, data on transitional human rights trials draw a stark comparison. We include human rights prosecutions of state agents with nexus to democratic transitions and intrastate conflict; they include proceedings with charges for conduct during the prior autocratic regime or transition and conduct in past or ongoing conflict. No decline spells had much higher average numbers of such transitional trials in general, and trials of high-ranking state agents specifically, than breakdown spells. However, spells with backsliding had by far the highest average numbers of trials. This is due in large part to the inclusion of Chile in this pool of cases. Chile has the largest numbers of trials and convictions in the TJET database, and Argentina is second. According to V-DEM data, Chile backslid from liberal to electoral democracy in 2019, due to widespread protests, which were met with repressive tactics and restrictions on the right to association. When we remove Chile and Argentina from the data, the averages for spells without democratic decline and with backsliding are very similar, as shown in parentheses. The comparison is similar for convictions.83 Overall, these data suggest that lack of prosecutions is associated with democratic breakdown, but states with a high number of trials may be just as prone to democratic backsliding as they are to maintaining liberal democratic institutions.
Third, the data clearly hint that truth commission publicness may provide protection from democratic decline. The truth commissions publicness index ranges from zero, in cases without a truth commission or where a truth commission does not hold public hearings and does not make a final report publicly available, to two, in cases of truth commissions doing both. We developed a measure of truth commission publicness for this analysis because most theories about the impact of truth commissions on democracy presume that they are public affairs that sometimes conduct well-publicized hearings and provide publicly consumable recommendations for reform. The average publicness index is higher in no-decline spells than in spells with democratic backsliding or breakdown.
Fourth, reparations are also associated with less democratic decline. We examine the average proportions of spell years that contained reparations policies offering compensatory payments to individuals who had experienced harm. While it would be ideal to possess a measure for how extensively each and every reparations policy was implemented, in practice these are very difficult data to collect. In diagnostics, we discovered that one of the main sources of variation is simply whether the policy actually paid out, which is what our measure indicates. In the end, no decline and backsliding spells contained a far higher proportion of compensatory reparation years than did those spells that eventually broke down.
Finally, ban from office is the one transitional justice mechanism that follows a different pattern. Though it is far less prevalent in those cases that experienced democratic breakdown, it is even more common in cases that later undergo backsliding. We analyze the average proportions of spell years with operative bans from future government employment or political office. The average proportion of spell years with at least one ban policy was larger in no-decline spells than in those with democratic breakdown, but it was by far the highest in backsliding spells. This is likely because Eastern European and post-Soviet countries, which implemented lustration policies and bans from office at various levels of severity, have featured several cases of backsliding.
Overall, while the descriptive data presented here cannot sustain causal inference,84 these bivariate trends strongly suggest a pattern. Transitional democracies with more extensive transitional justice seldom break down, but they do frequently backslide. In fact, transitional justice is just as prevalent in transitional states that remain liberal democracies as those that eventually illiberalize. One implication is that, moving forward, the field should focus less on whether transitional justice causes reversion to autocracy, and more on whether it leads to illiberal democracy. The multivariable models in the next section take an initial step in this direction, statistically comparing the effects of transitional justice on the institutional and behavioral links that we identified in the second section. With these models, we can answer the question: if transitional justice leads to backsliding, how does it do so?
TESTING THE HYPOTHESES
As theorized in the second section, transitional justice may play a role in democratic decline through three direct or three indirect pathways. We presented six hypotheses (A–F) and their alternatives. The following analysis utilizes Varieties of Democracy (V-DEM) indicators to test those hypotheses. Table 2 summarizes those measures. To arrive at the figures, we take the mean of the indicators in each individual spell, and then average those means over all the spells in each of the three sets. For all variables, higher values indicate a greater prevalence of the phenomenon being measured.
Average Outcome Statistics for Democratic Spells through 2023.
| Variable | No Decline | Backsliding | Breakdown |
|---|---|---|---|
| unfair elections index | 0.27 | 0.13 | 0.42 |
| weakened judicial constraints index | 0.25 | 0.12 | 0.37 |
| limitations on association index | 0.16 | 0.12 | 0.22 |
| political polarization | −0.15 | −0.83 | 0.01 |
| anti-democratic movements | 0.15 | 0.18 | 0.15 |
| diluted political engagement | −0.74 | −0.82 | −0.55 |
| Variable | No Decline | Backsliding | Breakdown |
|---|---|---|---|
| unfair elections index | 0.27 | 0.13 | 0.42 |
| weakened judicial constraints index | 0.25 | 0.12 | 0.37 |
| limitations on association index | 0.16 | 0.12 | 0.22 |
| political polarization | −0.15 | −0.83 | 0.01 |
| anti-democratic movements | 0.15 | 0.18 | 0.15 |
| diluted political engagement | −0.74 | −0.82 | −0.55 |
Average Outcome Statistics for Democratic Spells through 2023.
| Variable | No Decline | Backsliding | Breakdown |
|---|---|---|---|
| unfair elections index | 0.27 | 0.13 | 0.42 |
| weakened judicial constraints index | 0.25 | 0.12 | 0.37 |
| limitations on association index | 0.16 | 0.12 | 0.22 |
| political polarization | −0.15 | −0.83 | 0.01 |
| anti-democratic movements | 0.15 | 0.18 | 0.15 |
| diluted political engagement | −0.74 | −0.82 | −0.55 |
| Variable | No Decline | Backsliding | Breakdown |
|---|---|---|---|
| unfair elections index | 0.27 | 0.13 | 0.42 |
| weakened judicial constraints index | 0.25 | 0.12 | 0.37 |
| limitations on association index | 0.16 | 0.12 | 0.22 |
| political polarization | −0.15 | −0.83 | 0.01 |
| anti-democratic movements | 0.15 | 0.18 | 0.15 |
| diluted political engagement | −0.74 | −0.82 | −0.55 |
The first three indicators are focused on institutional protections. Higher values on all these indicators means that liberal democratic institutions are more troubled. UNFAIR ELECTIONS is the inverse of V-DEM’s Clean Elections Index, which assesses to what extent elections are free from ‘registration fraud, systematic irregularities, government intimidation of the opposition, vote buying, and election violence.’85 Second, WEAKENED JUDICIAL CONSTRAINTS is the inverse of V-DEM’s Judicial Constraints on the Executive Index, which assesses ‘[t]o what extent does the executive respect the constitution and comply with court rulings, and to what extent is the judiciary able to act in an independent fashion?’86 Finally, the LIMITATIONS ON ASSOCIATION variable is the inverse of V-DEM’s Freedom of Association Index, which measures to what extent CSOs are allowed to form and operate and political parties are allowed to participate in elections. When we examine averages across three sets of cases – those that experienced no decline, those that experienced backsliding and those that underwent full breakdown – an intriguing pattern emerges. While breakdown cases have significantly higher scores than those cases with no democratic decline – indicating that democratic institutions are under greater strain – the cases that experienced backsliding score lowest. This means that, before these transitional democracies de-liberalized, they actually had cleaner elections, more judicial constraints and greater freedom of association than those democracies that did not decline.
The remaining measures address political behavior within society. V-DEM’s POLITICAL POLARIZATION variable refers ‘to the extent to which political differences affect social relationships beyond political discussions.’87 A polarized society consists of antagonistic political camps that are less likely to compromise with each other. According to the political polarization averages, spells ending in breakdown were the most polarized, while those with backsliding were the least polarized. V-DEM’s CSO’s Anti-Systems’ Movement indicator assesses to what extent there are civil society movements that oppose and aim to change the current political system. ANTIDEMOCRATIC MOVEMENTS measures a specific variant of this anti-system movement activity, which is defined as unwillingness to ‘play by the rules of the democratic game,’ ‘to respect constitutional provisions or electoral outcomes’ or ‘to relinquish power (under democratic auspices).’88 Intriguingly, antidemocratic movements are relatively constant across all types of spells and are only slightly stronger in those cases that backslide. Finally, the DILUTED POLITICAL ENGAGEMENT indicator is the inverse of V-DEM’s Engagement in Independent Political Associations measure, which categorizes the share of the population that is regularly active in independent political interest associations. The higher the values on Diluted Political Engagement, the smaller the share of the population that is politically engaged. As with polarization, democratic spells that end with backsliding fare best. They have the lowest scores on this variable, meaning they feature more political engagement than spells with no decline, and far outpace spells that experience breakdown.
Models
To test the hypotheses, we implement fixed-effects regressions of the three institutional outcome variables (Unfair Elections, Weakened Judicial Constraints and Limitations on Association) and three behavioral outcome variables (Political Polarization, Antidemocratic Movements and Diluted Political Engagement).89 For this analysis, the sample includes country-year observations for postautocratic and conflicted democracy spells until 2020 because the TJET data end that year.90 Each model incorporates transitional justice variables described in Table 1. To account for the weight of past practice, we include cumulative sums of amnesty laws and human rights trials of high-ranking state officials. We chose high-ranking prosecutions because these are more likely to attract the attention of other elites and the democratic public. The truth commission scale and the binary reparations and bans from public office measures are carried forward from their first instances.
The models also include five control variables. First is time, which counts the years since a spell entered the sample. Second are two variables to account for economic wealth: log of latent GDP PER CAPITA estimates and an indicator of percentage GDP GROWTH (divided by 10).91 Third, to control for conditions giving rise to demand for transitional justice, we include the cumulative number of years with ONGOING CONFLICT, and the LEGACY OF VIOLENCE index. In additional models (see Supplementary Appendix), we include total foreign aid received to account for international linkages potentially impacting on democracy promotion and transitional justice; these do not alter the findings. A table showing scales for the outcome measures and the covariates is available in the Supplementary Appendix.
Results
We present the results of the models with coefficients plots, which show the estimated effects for a one-unit change in the variables with 95 percent confidence intervals.92 The effects are statistically significant if the confidence intervals do not touch the zero line. If the estimates are to the left of the zero line, it means they are negatively correlated with the dependent variable; if they are to the right, it means they are positively correlated.
Figure 3 presents the results from tests related to the effects of transitional justice on democratic institutions and on public behaviors. The top left panel shows that Hypothesis A, that transitional justice will be associated with more unfair elections, is mostly disconfirmed. Only one of the variables, high-ranking trials, is positively associated with unfair elections. While this is some evidence that prosecutions are threatening to elites, the effect is very small and substantively negligible. Though this is some validation for theories linking trials to pushback from elites, it is not of a sufficient magnitude to suggest that elites initiate wholesale attacks on electoral institutions in response to prosecutions. On balance, Alternative A seems the more plausible expectation: transitional justice in the form of bans from office and truth commissions is associated with less unfair electoral practices, and in significant fashion. Encouragingly, this model finds these effects while also validating other known relationships; for instance, states with ongoing conflict are more prone to electoral manipulation, while wealthier states are more likely to hold cleaner elections.
Models of Transitional Justice and Institutional and Public Behavior Outcomes.
If transitional justice leads to democratic decline, it is not by encouraging elites to cheat elections. How about subversion of the judicial branch? The middle left panel of Figure 3 presents data that also disconfirm Hypothesis B, that transitional justice leads to fewer judicial constraints on the executive. No transitional justice variables correlate to weakened judicial constraints. And as before, bans from office and truth commissions are in fact associated with greater levels of horizontal accountability through the judiciary, perhaps because the former prevents autocratic sympathists from holding high-level positions in the courts, while the latter push for judicial reforms. High-ranking trials and reparations are statistically insignificant, as are amnesties. One should also note that Time is associated with fewer judicial checks on the executive, which suggests that across all transitional democracies over time, courts are losing power relative to other branches of government.
The bottom left panel on political association displays the most significant negative relationships we discover between transitional justice and eroding democratic institutions. Hypothesis C is partially confirmed: even controlling for the increased risk of restrictions over time, prosecutions and truth commissions are both correlated with limitations to political association. However, reparations, bans from office and amnesties are correlated with fewer restrictions to the freedom of association. It is hard to know why the effect of transitional justice policies differs. Given the balance of evidence, this suggests that the one plausible linkage between transitional justice and the de-liberalization of democratic institutions is the rollback of freedoms of association. One primary example of this relationship is Chile, which has had four truth commissions and around 400 human rights prosecutions, but which in 2019 began restricting CSO activity in response to massive and paralyzing street protests. Whether Chilean truth commissions and human rights trials caused those restrictions in any meaningful way is a separate question for additional research. But the point remains: if one is looking for linkages between transitional justice and direct institutional subversion, then one should further investigate how transitional justice might lead elites to narrow civic space.
How is transitional justice associated with features of political society that might precondition a country to illberalize? The top right panel of the figure presents results related to political polarization. This model mostly confirms Hypothesis D, that transitional justice increases political polarization. High-ranking trials and bans from office are associated with more polarized democratic societies, and truth commissions and reparations are also positively correlated with polarization. The results in the middle right panel, however, disconfirm Hypothesis E, that transitional justice fuels antidemocratic movements. Alternative E is far more plausible. Trials and reparations are actually correlated to less antidemocratic movement activity.
The bottom right panel displays the results from the final model testing the relationship between transitional justice and political engagement. Here, Hypothesis F does not perform well in comparison to Alternative F. Trials, truth commissions and reparations are all associated with less-diluted political engagement, and the effect of reparations is very robust. This validates other micro-level research that identifies a relationship between compensation through reparations and voter registration.93 Transitional justice, it seems, might very well inspire the public to get involved.
What inferences can be drawn from this second set of models concerning the impact of transitional justice on citizen behavior? First, transitional justice mechanisms might very well exacerbate political polarization, which is plausible given the bitter divisions that animate responses to prosecutions and other measures in countries like Guatemala and Bosnia. At the same time, transitional justice is related to slightly less vigorous antidemocratic social movements and greater political engagement.
Second, it is hard to discern whether this combination of factors is a recipe for democratic backsliding. Table 2 shows that transitional democratic spells that undergo backsliding actually feature less polarization, slightly more antidemocratic movement and greater political engagement. Yet it seems untenable to aim for more polarized and less engaged publics to protect liberal democracy. Third and finally, the findings on bans from public office warrant more extensive analysis. If we examine the 15 countries that experienced democratic backsliding (Table 3), five of them had lustration policies at some point (Czechia, Hungary, Latvia, Lithuania and Poland). These cases are less politically polarized, but they have more antidemocratic political activity. This points to the possibility that some post-communist publics are simply in agreement that liberal constitutional government is not the most desired form of government. If so, backsliding might be explained by the general popularity of the illiberal democratic model in former Communist states.94
Cases of Backsliding.
| Country | Backsliding Year |
|---|---|
| Chile | 2019 |
| Cyprus | 2023 |
| Czechia/Czechoslovakia | 2020 |
| Ghana | 2015 |
| Greece | 2022 |
| Hungary | 2010 |
| Israel | 2023 |
| Latvia | 2013, 2016 |
| Lithuania | 2016 |
| Poland | 2016 |
| Portugal | 2023 |
| Serbia | 2013 |
| Slovakia | 2019 |
| South Africa | 2013 |
| Trinidad and Tobago | 2021 |
| Country | Backsliding Year |
|---|---|
| Chile | 2019 |
| Cyprus | 2023 |
| Czechia/Czechoslovakia | 2020 |
| Ghana | 2015 |
| Greece | 2022 |
| Hungary | 2010 |
| Israel | 2023 |
| Latvia | 2013, 2016 |
| Lithuania | 2016 |
| Poland | 2016 |
| Portugal | 2023 |
| Serbia | 2013 |
| Slovakia | 2019 |
| South Africa | 2013 |
| Trinidad and Tobago | 2021 |
Cases of Backsliding.
| Country | Backsliding Year |
|---|---|
| Chile | 2019 |
| Cyprus | 2023 |
| Czechia/Czechoslovakia | 2020 |
| Ghana | 2015 |
| Greece | 2022 |
| Hungary | 2010 |
| Israel | 2023 |
| Latvia | 2013, 2016 |
| Lithuania | 2016 |
| Poland | 2016 |
| Portugal | 2023 |
| Serbia | 2013 |
| Slovakia | 2019 |
| South Africa | 2013 |
| Trinidad and Tobago | 2021 |
| Country | Backsliding Year |
|---|---|
| Chile | 2019 |
| Cyprus | 2023 |
| Czechia/Czechoslovakia | 2020 |
| Ghana | 2015 |
| Greece | 2022 |
| Hungary | 2010 |
| Israel | 2023 |
| Latvia | 2013, 2016 |
| Lithuania | 2016 |
| Poland | 2016 |
| Portugal | 2023 |
| Serbia | 2013 |
| Slovakia | 2019 |
| South Africa | 2013 |
| Trinidad and Tobago | 2021 |
CONCLUSION
Has transitional justice led to democratic decline? Our answer is a qualified ‘no.’ Transitional justice has more likely served as a firewall against full reversion to autocracy, though it has not hindered democratic backsliding.
How do we reach this conclusion? The article presents a theory of six pathways – three direct and three indirect – through which transitional justice mechanisms might lead to democratic decline. This alone is a contribution. Thus far, research has offered several expectations in piecemeal fashion, without bringing competing hypotheses together under the same umbrella. Our theory is rooted in the idea that democratic decline happens primarily through elite-led institutional subversion, which is enabled by certain patterns of public behavior. On balance, the evidence we analyze suggests that transitional justice is actually associated with institutional improvements, but with a couple of exceptions that expose vulnerabilities. The first is that trials are very slightly associated with more unfair elections, and the other is that trials and truth commissions are both associated with rollbacks to the formal rights to political association.
One way of synthesizing the findings in our analysis is to think of transitional democracies over the last 50 years as two different worlds. World A is represented by 57 (of 118 total, or 48.3 percent) democratic spells that end in breakdown, or autocratic reversion. This set is fragile, with a legacy of extreme violence and a greater propensity for coups and intrastate war. This half of transitional democracies prone to breakdown often experience double transitions; they undergo regime change away from autocracy and experience intrastate conflict. These countries are particularly vulnerable. Governments in World A are less willing to offer much by way of transitional justice. This is a problem because transitional justice seems capable of reinforcing the guardrails of democracy that prevent autocratization. Our analyses show that bans from public office are associated with improved elections, judicial constraints and freedom to association. Truth commissions too are correlated to greater judicial constraints, and reparations are also correlated to fewer limitations on association.
Could it be that transitional justice is merely a byproduct of preexisting conditions? Is a lack of transitional justice inevitable in this more troubled world of inevitable democratic breakdown? Some deeper structures are apparent. However, these structures are not wholly determinative; it is entirely possible that countries predisposed to breakdown can have transitional justice, and this can be transformative. For example, Liberia transitioned to democracy in 2006, it has regularly held successful elections and it has not yet experienced a breakdown. This could owe in part to the capacity building of the TRC and other sincere efforts at justice following a brutal civil war.95 Indeed, those countries that implement transitional justice, on average, experience endogenous change toward more vigorous vertical and horizontal accountability – even if they are preconditioned to experience democratic breakdown.
World B consists of 61 democratic spells (of 118 total, or 51.7 percent) that either experience no decline or backslide from liberal to electoral forms of democratic governance. This set of regimes exists in a context of less extreme instability and has a great deal more post-authoritarian or postconflict transitional justice. Intriguingly, within this half, transitional justice is equally common in the 15 cases that experience backsliding as in the 48 cases that experience no democratic decline at all. This suggests that transitional democracies with transitional justice are still quite capable of illiberalizing. And when backsliding occurs following transitional justice, it is most likely to come in the form of government restrictions on freedom of association. This is a chink in the transitional justice armor: it cannot deter elites from closing civic space.
Finally, the relationship between transitional justice and public behavior calls out for more research. Transitional justice is associated with more politically engaged publics that feature fewer antidemocratic movements, but transitional justice is also followed by much greater political polarization. This might be a paradox of democracy: when people support democracy, and they are more active in civic life, they are bound to grow more dissatisfied and find ideological enemies with whom to spar.96 Whether extreme polarization is democracy as usual or a death rattle for future democratic governance is hard to know. But it is worth noting that the 15 cases which backslid in our sample actually had less polarization and more engagement than those that experienced no decline. Here, the arrows point in too many directions to draw strong inferences.
What we can say is this: while the analysis in this article shows transitional justice is not much to blame for democratic decline in general, it also finds that transitional justice may be powerless to stop backsliding in particular. The data from our sample suggest that, with time, democratic institutions generally erode. Yet what causes gradual retrogression remains a mystery. Prior to illiberalization, the countries that backslide score on average the lowest on unfair elections, weakened judicial constraints and limits to political association measures. What are the preconditions, then, for backsliding? And does transitional justice play a role in the processes of de-liberalization? These are the questions that must be answered in future research.
SUPPLEMENTARY MATERIAL
Supplementary material is available at The International Journal of Transitional Justice online.
Footnotes
Samuel Huntington, The Third Wave: Democratization in the Late Twentieth Century (Norman: University of Oklahoma Press, 1991); Ruti Teitel, ‘Human Rights in Transition: Transitional Justice Genealogy,’ The Harvard Human Rights Journal 16(69) (2003): 69–94.
Paige Arthur, ‘How “Transitions” Reshaped Human Rights: A Conceptual History of Transitional Justice,’ Human Rights Quarterly 31(2) (2009): 321–367.
David Gairdner, Truth in Transition: The Role of Truth Commissions in Political Transition in Chile and El Salvador (Bergen: Chr. Michelsen Institute, 1999); Amy Gutmann and Dennis Thompson, ‘The Moral Foundations of Truth Commissions,’ in Truth v. Justice, ed. Robert Rotberg and Dennis Thompson (Princeton: Princeton University Press, 2000), 22–44; Cynthia M. Horne, ‘The Impact of Lustration on Democratization in Postcommunist Countries,’ International Journal of Transitional Justice 8(3) (2014): 496–521; A. James McAdams, ed., Transitional Justice and the Rule of Law in New Democracies (Notre Dame: University of Notre Dame Press, 1997); Juan Mendez, ‘Accountability for Past Abuses,’ Human Rights Quarterly 19(2) (1997): 255–282; Elin Skaar, Camila Gianella Malca and Trine Eide, After Violence: Transitional Justice, Peace, and Democracy (Abingdon: Routledge, 2015).
For a discussion of long-standing skepticism concerning transitional justice and democracy, see redundant Monika Nalepa, After Authoritarianism: Transitional Justice and Democratic Stability (New York: Cambridge University Press, 2022), 5.
Vanessa A. Boese et al., ‘Autocratization Changing Nature?’ Democracy Report 2022 (Gothenburg: Varieties of Democracy Institute (V-DEM), University of Gothenburg, 2022).
We use the phrase democratic decline because it is a general name for the phenomenon of interest that avoids some terminological confusion by the inconsistent application of other terms. See Thomas M. Keck, ‘Erosion, Backsliding, or Abuse: Three Metaphors for Democratic Decline,’ Law & Social Inquiry 48(1) (2023): 314–339.
Thomas Carothers, ‘The End of the Transition Paradigm,’ Journal of Democracy 13(1) (2002): 5–21.
Rosemary Nagy, ‘Transitional Justice as Global Project: Critical Reflections,’ Third World Quarterly 29(2) (2008): 275–829; Simon Robins, ‘Transitional Justice as an Elite Discourse,’ Critical Asian Studies 44(1) (2012): 3–30; Mohamed Sesay, ‘Decolonization of Postcolonial Africa: A Structural Justice Project More Radical than Transitional Justice,’ International Journal of Transitional Justice 16(2) (2022): 254–271.
Leslie Vinjamuri and Jack Snyder, ‘Law and Politics in Transitional Justice,’ Annual Review of Political Science 18(1) (2015): 303–327.
See Huntington, supra n 1. Jack Snyder and Leslie Vinjamuri, ‘Trials and Errors: Principle and Pragmatism in Strategies of International Justice,’ International Security 28(3) (2003): 5–44.
Omar G. Encarnación, ‘Transitional Justice at 40: A Critical Appraisal,’ Current History 123(849) (2024): 7. See also Susan Thomson, ‘The Darker Side of Transitional Justice: The Power Dynamics Behind Rwanda’s Gacaca Courts,’ Africa 81(3) (2011): 373–390.
Geoff Dancy and Oskar Timo Thoms, ‘Do Truth Commissions Really Improve Democracy?’ Comparative Political Studies 55(4) (2022): 555–587.
Cath Collins, Post-Transitional Justice: Human Rights Trials in Chile and El Salvador (University Park: Pennsylvania State University Press, 2010). Geoff Dancy and Verónica Michel, ‘Human Rights Enforcement from Below: Private Actors and Prosecutorial Momentum in Latin America and Europe,’ International Studies Quarterly 60(1) (2016): 173–188. On ‘endogenous change,’ see James Mahoney and Kathleen Thelen, ‘A Theory of Gradual Institutional Change,’ in Explaining Institutional Change: Ambiguity, Agency, and Power, ed. James Mahoney and Kathleen Thelen (New York: Cambridge University Press, 2010), 1–37.
Steven Levitsky and Daniel Ziblatt, How Democracies Die (New York: Crown, 2018).
Snyder and Vinjamuri, supra n 10.
Clayton Jones, ‘Philippine Coup Attempt Highlights Military Problems,’ Christian Science Monitor, 31 August 1987; Seth Mydans, ‘Coup Crushed in Philippines: In a Two-Crisis Week, Aquino Shows Her Nerve,’ The New York Times, 30 August 1987.
Priscilla Hayner, Unspeakable Truths: Confronting State Terror and Atrocities (New York: Routledge, 2001).
Nalepa, supra n 4 at 14. We exclude purges from our analysis. Purges are similar to lustration policies, but they dismiss current leaders and/or rank-and-file members of government agencies, rather than vetting them from future public posts. The reason that we do not consider them here is that we are particularly interested in legal or institutional forms of transitional justice. Purges, or dismissals, are often carried out through extra-legal means.
Roman David, ‘Lustration Laws in Action: The Motives and Evaluation of Lustration Policy in the Czech Republic and Poland (1989–2001),’ Law & Social Inquiry 28(2) (2003): 401.
Levitsky and Ziblatt, supra n 14 at 20.
Claire Greenstein and Cole J. Harvey, ‘Trials, Lustration, and Clean Elections: The Uneven Effects of Transitional Justice Mechanisms on Electoral Manipulation,’ Democratization 24(6) (2017): 1195–1214.
Clifford Krauss, ‘Tapes Spy Chief Left Behind Scandalize Peru,’ The New York Times, 3 February 2001.
Greenstein and Harvey, supra n 21.
Nalepa, supra n 4 at 246–252. See also Milena Ang and Monika Nalepa, ‘Can Transitional Justice Improve the Quality of Representation in New Democracies?’ World Politics 71(4) (2019): 631–666.
Peter Kornbluh, ‘Chileans Confront Their Own 9/11,’ The Nation, 4 September 2013.
‘Sri Lanka President Sacks Chief Justice Bandaranayake,’ BBC News, 13 January 2013; Rehan Abeyratne, ‘Judicial Appointments in Sri Lanka: A Politicized Trajectory,’ in Appointment of Judges to the Supreme Court of India, ed. Rehan Abeyratne (Oxford: Oxford University Press, 2018), 255–266.
Lisa Hilbink, ‘The Origins of Positive Judicial Independence,’ World Politics 64(4) (2012): 587–621.
Yonatan Lupu, ‘International Judicial Legitimacy: Lessons from National Courts,’ Theoretical Inquiries in Law 14(2) (2013): 437–454.
Renée Jeffery and Bikram Timilsina, ‘Re-Democratising Nepal: Transitional Justice and the Erosion of Judicial Independence,’ Contemporary Politics 27(5) (2021): 550–571.
Ibid., 565.
Pablo de Greiff, ‘Theorizing Transitional Justice,’ in Transitional Justice: NOMOS LI, ed. Rosemary Nagy, Melissa S. Williams and Jon Elster (New York: NYU Press, 2012), 31–77.
Catalina Smulovitz, ‘The Discovery of Law: Political Consequences in the Argentine Case,’ in Global Prescriptions: The Production, Exportation, and Importation of a New Legal Orthodoxy, ed. Yves Dezalay and Bryant G. Garth (Ann Arbor: University of Michigan Press, 2002), 249–275; Naomi Roht-Arriaza, ‘After Amnesties Are Gone: Latin American National Courts and the New Contours of the Fight against Impunity,’ Human Rights Quarterly 37(2) (2015): 341–382.
Dancy and Michel, supra n 13 at 176.
Collins, supra n 13.
Ezequiel González Ocantos, ‘Persuade Them or Oust Them: Crafting Judicial Change and Transitional Justice in Argentina,’ Comparative Politics 46(4) (2014): 479–498.
Onur Bakiner, ‘Truth Commission Impact: An Assessment of How Commissions Influence Politics and Society,’ International Journal of Transitional Justice 8(1) (2014): 6–30. See also Nalepa, supra n 4.
Mark Osiel, Mass Atrocity, Collective Memory, and the Law (New Brunswick: Transaction Publishers, 1997).
David A. Crocker, ‘Truth Commissions, Transitional Justice, and Civil Society,’ in Truth v. Justice: The Morality of Truth Commissions, ed. Robert Rotberg and Dennis Thompson (Princeton: Princeton University Press, 2000), 99–121; Kelebogile Zvobgo, ‘Demanding Truth: The Global Transitional Justice Network and the Creation of Truth Commissions,’ International Studies Quarterly 64(3) (2020): 609–625.
Kendra Dupuy, James Ron and Aseem Prakash, ‘Hands Off My Regime! Governments’ Restrictions on Foreign Aid to Non-Governmental Organizations in Poor and Middle-Income Countries,’ World Development 84 (2016): 299–311; Kristin M. Bakke, Neil J. Mitchell and Hannah M. Smidt, ‘When States Crack Down on Human Rights Defenders,’ International Studies Quarterly 64(1) (2020): 85–96.
Kenneth Roth, ‘The Great Civil Society Choke-Out,’ Foreign Policy, 27 January 2016.
‘Statement of Concern on Online Threats against Guatemalan NGO amidst Wider Crackdown on Activists,’ Global Witness, https:///en/press-releases/statement-concern-online-threats-against-guatemalan-ngo-amidst-wider-crackdown-activists/ (accessed 28 May 2024).
‘Guatemala: Last Chance for Justice: Dangerous Setbacks for Human Rights and the Fight against Impunity in Guatemala,’ Amnesty International, https://www.amnesty.org/en/documents/amr34/0611/2019/en/ (accessed 28 May 2024).
Hun Joon Kim and Kathryn Sikkink, ‘Explaining the Deterrence Effect of Human Rights Prosecutions,’ International Studies Quarterly 54(4) (2010): 939–963; Geoff Dancy et al., ‘Behind Bars and Bargains: New Findings on Transitional Justice in Emerging Democracies,’ International Studies Quarterly 63(1) (2019): 99–110; Guillermo Trejo, Juan Albarracín and Lucía Tiscornia, ‘Breaking State Impunity in Post-Authoritarian Regimes: Why Transitional Justice Processes Deter Criminal Violence in New Democracies,’ Journal of Peace Research 55(6) (2018): 787–809.
Truth & Dignity Commission, ‘Final Comprehensive Report: Executive Summary,’ May 2019, 162, https://truthcommissions.humanities.mcmaster.ca/wp-content/uploads/2020/11/Tunisia-Truth-and-Dignity-Commission-Report_executive_summary_report.pdf (accessed 27 May 2024).
Ibid., 437.
Levitsky and Ziblatt, supra n 14 at 116.
David Rieff, In Praise of Forgetting: Historical Memory and Its Ironies, reprint ed. (New Haven: Yale University Press, 2017).
Ines Stasa and Arian Dedej, ‘Transitional Justice as a Tool for Polarization in Albania,’ European Journal of Economics, Law and Social Sciences 7(3) (2023): 13.
Izabela Steflja, ‘The Production of the War Criminal Cult: Radovan Karadžić and Vojislav Šešelj at The Hague,’ Nationalities Papers 46(1) (2018): 52–68.
Janine Natalya Clark, ‘The Limits of Retributive Justice: Findings of an Empirical Study in Bosnia and Hercegovina,’ Journal of International Criminal Justice 7(3) (2009): 463–487.
Csilla Kiss, ‘The Misuses of Manipulation: The Failure of Transitional Justice in Post-Communist Hungary,’ Europe-Asia Studies 58(6) (2006): 927.
Olfa Belhassine, ‘Tunisian Transitional Justice in Danger, Warns Civil Society,’ JusticeInfo.Net (blog), 5 October 2020, https://www.justiceinfo.net/en/45584-tunisian-transitional-justice-in-danger-warns-civil-society.html (accessed 27 May 2024).
Laurel E. Fletcher and Harvey M. Weinstein, ‘Violence and Social Repair: Rethinking the Contribution of Justice to Reconciliation,’ Human Rights Quarterly 24(3) (2002): 573–639.
Guillermo O’Donnell and Philippe C. Schmitter, ‘Transitions from Authoritarian Rule: Tentative Conclusions about Uncertain Democracies,’ in Transitions from Authoritarian Rule, ed. Guillermo O’Donnell, Philippe C. Schmitter and Laurence Whitehead (Baltimore: Johns Hopkins University Press, 1986), 30.
Lavinia Stan, ‘The Vanishing Truth? Politics and Memory in Post-Communist Europe,’ East European Quarterly 40(4) (2006): 391–392.
Mahmood Mamdani, ‘Amnesty or Impunity? A Preliminary Critique of the Report of the Truth and Reconciliation Commission of South Africa (TRC),’ Diacritics 32(3) (2002): 33–59; David Backer, ‘Watching a Bargain Unravel? A Panel Study of Victims’ Attitudes about Transitional Justice in Cape Town, South Africa,’ International Journal of Transitional Justice 4(3) (2010): 443–456.
Clark, supra n 50 at 471.
Michael Coppedge et al., ‘V-Dem Codebook v14,’ Varieties of Democracy (V-Dem) Project, 2024, 204–205.
Camilo Freedman, ‘“Unwaveringly Loyal”: Why El Salvador’s Bukele Is Poised for Re-Election,’ Al Jazeera, https://www.aljazeera.com/news/2024/1/31/unwaveringly-loyal-why-el-salvadors-bukele-is-poised-for-reelection. (accessed 29 May 2024).
Gábor Halmai, ‘Rule of Law Backsliding and Memory Politics in Hungary,’ European Constitutional Law Review 19(4) (2023): 602–622.
Ibid., 619–620.
Laia Balcells, Valeria Palanza and Elsa Voytas, ‘Do Transitional Justice Museums Persuade Visitors? Evidence from a Field Experiment,’ The Journal of Politics 84(1) (2022): 496–510; Risa Kitagawa, ‘From Political Violence to Political Trust? How Transitional Justice Affects Citizen Views of Government,’ International Studies Quarterly 67(1) (2023): sqad013.
Elke Zuern, The Politics of Necessity: Community Organizing and Democracy in South Africa (Madison: University of Wisconsin Press, 2011).
Zinaida Miller, ‘Effects of Invisibility: In Search of the “Economic” in Transitional Justice,’ International Journal of Transitional Justice 2(3) (2008): 266–291; Karen Engle, Zinaida Miller and D. M. Davis, eds., Anti-Impunity and the Human Rights Agenda (Cambridge: Cambridge University Press, 2016).
Reem Abou-El-Fadl, ‘Beyond Conventional Transitional Justice: Egypt’s 2011 Revolution and the Absence of Political Will,’ International Journal of Transitional Justice 6(2) (2012): 318–330.
Greg Grandin, ‘The Instruction of Great Catastrophe: Truth Commissions, National History, and State Formation in Argentina, Chile, and Guatemala,’ The American Historical Review 110(1) (2005): 48. See also Kadar Asmal, ‘Truth, Reconciliation and Justice: The South African Experience in Perspective,’ The Modern Law Review 63(1) (2000): 1–24.
Lia Kent, ‘Rethinking Transitional Justice: Lessons from East Timor,’ in State, Society, and Governance in Melanesia (Canberra: Coral Bell School of Asia Pacific Affairs, College of Asia and the Pacific, Australian National University, 2013); Gearoid Millar, ‘Local Evaluations of Justice through Truth Telling in Sierra Leone: Postwar Needs and Transitional Justice,’ Human Rights Review 12(4) (2011): 515–535.
Dancy and Thoms, supra n 12.
E. Pajibo, ‘Civil Society and Transitional Justice in Liberia: A Practitioner’s Reflection from the Field,’ International Journal of Transitional Justice 1(2) (2007): 287–296.
Elsa Voytas, ‘More than Money: The Political Consequences of Reparations,’ OSF, 2024, https://osf.io/preprints/osf/akz26 (accessed 1 June 2024).
Horne, supra n 3 at 503.
We do not include non-transitional democracies or autocracies in our analysis dataset, though transitional justice often occurs in both contexts. See Mara Redlich Revkin, Ala Alrababah and Rachel Myrick, ‘Evidence-Based Transitional Justice: Incorporating Public Opinion into the Field, with New Data from Iraq and Ukraine,’ Yale Law Younal 133(5) (2024): 1401–1801.
Carles Boix, Michael Miller and Sebastian Rosato, ‘A Complete Data Set of Political Regimes, 1800–2007,’ Comparative Political Studies 46(12) (2013): 1523–1554; Polity Project, ‘Polity 5: Political Regime Characteristics and Transitions, 1800–2018,’ Technical report (Vienna: Center for Systemic Peace, 2020), www.systemicpeace.org (accessed 2 February 2024); Michael Coppedge et al., ‘V-Dem [Country-Year/Country-Date] Dataset’ (Varieties of Democracy Institute, University of Gothenburg; Varieties of Democracy Project, 2023).
For data and details, see https://transitionaljusticedata.org (accessed 5 December 2024).
Nils Petter Gleditsch et al., ‘Armed Conflict 1946–2001: A New Dataset,’ Journal of Peace Research 39(5) (2002): 615–637; Shawn Davies, Therése Pettersson and Magnus Öberg, ‘Organized Violence 1989–2022, and the Return of Conflict Between States,’ Journal of Peace Research 60(4) (2023): 691–708.
Amanda B. Edgell et al., Episodes of Regime Transformation [ERT] Dataset (v2.0) Codebook (Gothenburg: University of Gothenburg, 2020).
A full list of spells and outcomes is available in Supplementary Table A2 of the Appendix.
Using the V-DEM ERT regime age measure, we consider democratic regime age going back to 1900.
Coup data are taken from J. M. Powell and Clayton L. Thyne, ‘Global Instances of Coups from 1950 to 2010: A New Dataset,’ Journal of Peace Research 48(2) (2011): 249–259.
Keith E. Schnakenberg and Christopher J. Fariss, ‘Dynamic Patterns of Human Rights Practices,’ Political Science Research and Methods 2(1) (2014): 1–31.
The Legacy of Violence index is a cumulative and discounted average of this latent physical integrity measure since 1949. It is cumulative because for each country in the dataset it includes every available prior year in the annual calculation, and it is discounted because it weighs the recent past more heavily than the distant past. For more details, see https://transitionaljusticedata.org/en/faq.html (accessed 5 December 2024).
Geoff Dancy et al., ‘TJET Database of Transitional Justice Mechanisms, 1970–2020,’ (Toronto: Transitional Justice Evaluation Tools, 30 April 2024), https://transitionaljusticedata.org (accessed 5 December 2024).
Here we consider the numbers of individuals ever convicted in human rights trials, regardless of whether they were later acquitted in appeals.
First, these associations do not take into account other factors which may better explain democratic decline. Second, as in other research fields, selection effects present a thorny problem for assessing the effects of transitional justice, because both the choices to implement transitional justice policies and the absence of democratic decline may be explained by other variables.
Coppedge et al., supra n 58 at 2.
Ibid., 54.
Ibid., 232.
Ibid., 205.
The models of institutional outcomes have 2,125 observations, and the models of polarization, antidemocratic movements and engagement have 2,086, 1,840 and 2,050 observations respectively.
This takes the sample from 118 democracy spells to 116 democracy spells.
Christopher J. Fariss et al., ‘New Estimates of Over 500 Years of Historic GDP and Population Data,’ Journal of Conflict Resolution 66(3) (2022): 553–591.
The full model results are available in Supplementary Table A3 of the Appendix.
Voytas, supra n 70.
Halmai, supra n 60.
Charles Wratto, ‘Liberia Transferred Power Peacefully Again: 3 Reasons the Calm Is Holding, and One Red Flag,’ The Conversation, 23 January 2024, http://theconversation.com/liberia-transferred-power-peacefully-again-3-reasons-the-calm-is-holding-and-one-red-flag-221656 (accessed 1 June 2024).
Robert A. Dahl, ‘A Democratic Paradox?’ Political Science Quarterly 115(1) (2000): 35–40.
Author notes
Associate Professor, Department of Political Science, University of Toronto, Toronto, Canada. Email: geoff.dancy@utoronto.ca
Research Associate, Department of Political Science, Toronto, Canada. Email: oskar.thoms@utoronto.ca
∞ We are funded by Global Affairs Canada Grant No. 211298
Comments