Trust in Colombia’s Justicia Especial Para La Paz: Experimental Evidence

Introduction

In November 2016, a peace agreement was signed between the Colombian government, headed by President Juan Manuel Santos, and the Fuerzas Armadas Revolucionarias de Colombia (Colombian Revolutionary Armed Forces, FARC). This peace agreement sought to put an end to a lengthy and violent armed conflict by brokering terms for the demobilisation of the oldest active guerrilla in the world. The peace agreement was all-encompassing and ambitious: one of its pillars was a specially designed transitional justice (TJ) system comprising a web of new institutions including a truth commission, an investigative unit to search for the disappeared, and a temporary criminal tribunal charged with judging crimes committed during the conflict. That tribunal, the Justicia Especial para la Paz (Special Peace Jurisdiction), known by its Spanish acronym “JEP,” began operations in early 2018. The twenty-eight judges have a maximum of fifteen years to adjudicate responsibility and punishment in the context of a decades-old conflict that cost more than 260,000 lives (Grupo de Memoria Histórica, 2013).

The creation of this high court, a modified version of what some call “war tribunals” in other countries, is part of a critical juncture in Colombian history. The JEP is a domestic, apex-level court that is an integral part of a contested TJ framework, whose deployment will be central to Colombian politics in the coming years. As I will explain in more detail below, the innovative, and controversial, nature of the restorative sentences this tribunal will hand down – which many perceive as too lenient – could represent a source of vulnerability in the eyes of citizens evaluating this new institution. Because the peace process and the creation of the JEP itself remain controversial topics that figure prominently in domestic political and electoral debates, understanding this court’s relationship to the Colombian public is crucial to assessing its political viability. In light of the JEP’s unique institutional characteristics, exploring this requires drawing from and working at the intersection of two bodies of literature: research on courts and public opinion, and work on TJ and public opinion. On the one hand, existing work on apex courts and public opinion emphasises the importance of individual’s support for democracy as well as their partisanship and prior direct experiences with the judicial system in shaping individuals’ attitudes towards high courts, including international war tribunals. On the other hand, the literature on public opinion and TJ suggests that the particulars of the conflict and individual-level factors (like ideology, ethnic identification, or victimisation) matter when understanding a person’s attitudes towards such types of measures. Understanding a transitional court such as the JEP asks us to move beyond.

This article relies on experimental evidence in order to investigate whether features of the case and the ruling play a role in citizens’ attitudes towards the JEP; in particular, whether the profile of the accused and whether or not they are sentenced to the most lenient of restorative justice measures have an effect. I expected respondents to trust the JEP less, and its decision, when it handed down a restorative, more lenient, and less punitive sentence. This expectation is borne out with regard to support for the decision. The group affiliation of the accused (whether he is a former FARC combatant or a military member) had no effect on citizen evaluations of the court.

Studying the public’s trust in transitional high courts is important to more fully understand the operation of these tribunals and, relatedly, the implementation of peace processes. Theory suggests that trust in courts and support can contribute to the active exercise of their independence and can be related to whether citizens accept rulings (Easton, 1975; Epstein et al., 2001; Gibson and Caldeira, 2003). Popular support may be even more crucial for transitional courts, that is, new courts that are meting out TJ (Arzt, 2006), like the JEP. Such tribunals often operate in deeply politicised environments, where their decisions are scrutinised by political actors, victims of the conflict, and civil society organisations. Distrust in the court could affect its credibility among the public and, in turn, foster political discontent and instability in the implementation of the peace process.

Furthermore, the creation of new high courts is a relatively infrequent event and we have few studies that shed light on the initial stages of these institutions, much less so in the Global South. For domestic tribunals, we have Gibson and Caldeira’s (2003) study of the establishment of the constitutional court in post-apartheid South Africa and their joint research with Vanessa Baird on European courts (Gibson et al., 1998). At the international level, Voeten (2013) has more recently studied the legitimacy of several supranational war tribunals and human rights international courts. Other research explores the determinants of mass attitudes towards TJ measures (e.g. Aguilar et al., 2011; Albarracin and Gamboa, 2018; Carlin et al., this issue; Gibson, 2002; Nalepa, 2012; Samii, 2013), but aside from Gibson and Caldeira (2003), I am not aware of another study that focuses on mass attitudes towards tribunals borne of transitional contexts.

Finally, this research is important for Colombia’s current political juncture. The JEP’s task is a difficult one, and it began performing amidst growing political scepticism with regard to the agreements themselves (Botero, 2017; Matanock and García-Sánchez, 2017), heightened political controversy around their implementation, and growing citizen distrust in democratic institutions (Observatorio de la democracia, 2017). The JEP itself has been at the centre of harsh criticisms: opponents of the peace agreements sought to stop its justices from taking office, tried to stall the approval of the bill that gave the tribunal its by-laws, supported a referendum to revoke the JEP’s mandate, and sponsored a bill to modify its structure and functions (Bermúdez Liévano, 2019). None of these measures have succeeded thus far, but they are an indication of the saliency of the JEP in the current juncture. Studying the early stages of this tribunal and how it relates to the public contributes to building a much-needed empirical baseline that will allow us to place into context and better understand an institution that will be central to domestic politics in the coming decades. ¹

The main finding has interesting implications for the study of courts and public opinion, and specially for research on transitional courts and their relationship with their public. Following prior work, this article distinguishes between two types of support for the court: specific support and diffuse support. This distinction was first sketched by Easton (1975); it differentiates between the degree of support a respondent affords to a given decision (specific support), and the degree to which that respondent trusts the institution (diffuse support) – in this case, the JEP. To Easton, specific and diffuse support are orthogonal in older, established courts; if these tribunals make an unpopular decision, they can rely on a reservoir of goodwill towards the institution to offset citizen discontent with specific rulings. I focus on the effect of ruling characteristics on support because, for younger courts, in the absence of direct experience with the tribunal or information about its track record, these could matter more for legitimacy-building processes. The results indicate that support for the decision is influenced by some features of the ruling in the case of the JEP, whilst none of the treatments had an effect on trust in the institution. The question that this raises is whether the independence between both kinds of support will hold over time for the JEP. The JEP operates in a very different environment than that of the established tribunals that Easton as well as Gibson and Caldeira studied. Like most young courts, the JEP lacks a reservoir of goodwill; like other transitional courts, it deals in particularly unpopular and controversial topics. In this context, we might expect that, over the medium term, specific rulings could negatively influence institutional support for this court.

The article proceeds as follows. The next section discusses the literature on trust in high courts and attitudes towards TJ, putting the two in dialogue with reference to the JEP in Colombia, to develop some expectations. The “Experimental Strategy and Data” section discusses the research design and the sample. The last two sections present the results and draw our attention to the implications of this first experimental study of trust in Colombia’s TJ tribunal.

Trust in High Courts, and Attitudes towards Transitional Justice and the JEP

Though we know a lot about how citizens perceive judicial institutions, particularly in the context of Europe and the United States, we have little information on whether the same rules, or which ones, apply to newly emerging judicial institutions in the aftermath of peace negotiations. I will draw on the literature on trust in courts as well as that on public opinion and TJ mechanisms in order to develop some expectations about how Colombians evaluate the JEP.

Previous research has identified a series of elements that play an important role in understanding to what extent citizens trust judicial institutions and, in particular, high courts. Features of the judicial system itself or of the individual’s experience with it – such as sentencing expediency, performance (Malone, 2010), and perceived fairness – have been shown to influence trust in courts (Mondak, 1991; Tyler and Sevier, 2014). With regard to high courts more specifically, scholars have shown that the public’s trust in the US Supreme Court (Gibson and Caldeira, 2009) and in high courts in Europe (Gibson et al., 1998) is rooted in an individual’s experience with, commitment to, and support of democratic institutions and processes. In Colombia, support for democracy as well as perceptions of levels of corruption also play a role in accounting for citizens’ trust in the Constitutional Court (Botero, 2020). In his research on supranational courts, Voeten (2013) also found that levels of trust depend on the pre-existence of positive attitudes towards existing domestic institutions.

The JEP is an apex-level domestic court (on par with Colombia’s other three high courts: the Supreme Court of Justice, the Council of State, and the Constitutional Court) that is part of a separate temporary structure aimed at providing TJ. This prominent position in the judicial institutional hierarchy and its function as heart of the TJ framework make it unique. The differences between the JEP and other apex courts suggest important limitations to what we can apply from the study of support for regular judicial institutions to the study of a young transitional court like the JEP. First, the institution is so young that the public has no prior experience to rely on when making their evaluations. Second, citizens are unlikely to personally encounter the court, so their evaluations are less likely to be based on direct experience and will probably rely on other cues. Third, conflict and violence are still present in Colombia; they remain a central part of the public discussion overall, and of discussions around the JEP in particular. Thus, neither direct experience with the institution or with the judicial system nor long-standing attitudes towards democracy will be the most salient consideration for citizens when evaluating a tribunal in a post-conflict setting such as this.

In studying the JEP, we stand at a crossroads between apex courts and TJ institutions. Since we cannot simply look to the literature on trust in high courts for clues, we must broaden our scope. Research on mass attitudes towards TJ mechanisms is a burgeoning field where one insight stands out: the particulars of the context matter, as do the features of TJ mechanisms. Because the court is so young and because of the political controversy around the leniency of its sentences, I suggest that we look at the kind of sentences themselves and the profile of the defendants before the JEP (particularly whether they are former guerrilla combatants or military members) as important elements in understanding how citizens evaluate this young court. To explore these issues in the Colombian context, I begin by recounting the approach to TJ in the peace accord with the FARC, and the basics of the JEP’s task and some of its features, to discuss how these may influence the public’s perception of the court.

Experimental Strategy and Data

To test my expectations, I embedded a survey experiment in the Barómetro de Las Américas Colombia 2018 nationally representative survey (Observatorio de la democracia, 2017). The face-to-face survey was fielded between September and December 2018 by IPSOS and coordinated by the Observatorio de la Democracia at Universidad de los Andes in Bogotá, Colombia. The sample included adults (N = 1663) over 18 years of age.

Mine is a factorial design with four different treatments. Thus, the sample was randomly divided into four groups of about 411 individuals each (Supplemental Appendix Table A1), which were balanced across covariates (Supplemental Appendix Table A2). All respondents were presented with a vignette that introduced the JEP and then described the profile of a hypothetical man who confessed to homicide and was found guilty by this tribunal. All four groups were exposed to the same profile except as it pertains to two details, which were randomised: (1) the affiliation of the accused, which varied between a soldier and a former FARC combatant, and (2) the sentence he received from the court (jail time or house arrest while doing landmine clearing). The vignette read as follows:

Given the recent date of creation of the JEP and for transparency, the vignette started off with a brief statement introducing the institution. The profile of the accused and the decision on his case is brief and the name used is typical of an average Colombian male, the sex of the majority of members of both organisations mentioned in the experiment. The affiliations were chosen to exemplify the two groups of accused most likely to come before the JEP: military members and former guerrilla members.

The complex reality of the Colombian conflict and the nuances of the JEP’s actual sentencing framework posed some particular challenges for the design of this experiment. The accused confessed to homicide in all iterations of the vignette, a crime that is (unfortunately) common in the Colombian context. I chose a crime that could have likely been committed by both a former guerrilla as well as a member of the military. I shied away from crimes under international humanitarian law because some of them are more associated with the FARC (like kidnappings) and all, as I explained earlier, fall under a special category that is not subject to the JEP’s special sentences.

In choosing the sentences for this hypothetical accused, my aim was to be as faithful as possible to the actual potential sentences as defined by the legal framework governing TJ under the 2016 peace agreements and as we knew them at the time in which the vignette was designed (2018). I wanted the experiment first and foremost to help us understand the Colombian case better and chose to prioritise that the treatments be as accurate as possible. This had several implications. To begin with, the agreements provided some sentencing guidelines, which narrowed the options. Since the Havana accords and the TJ system they designed are guided by principles of restorative justice, as mentioned earlier, the JEP’s own sentences are reduced. Indeed, the maximum jail time if found guilty by the JEP is eight years – unless the accused takes up arms again or fails to make reparation to the victims and tell the truth, in which case the sentence can be as long as twenty years.

I chose to compare two reduced sentences standing at different ends of the punitive versus restorative continuum within what the JEP’s own sentences would actually allow for, but did not entail also having to consider whether there was recidivism or a lack of commitment to clarifying the truth for the victims. (These last two are the conditions under which, aside from committing certain types of crimes, the longer jail sentences would be triggered.) Invoking such actions, which are associated with not fulfilling the promises acquired by demobilising, might raise a different set of concerns in the minds of respondents. In this case, my interest in using examples that were as close as possible to the JEP’s sentencing scheme meant that the two sentences would be jail versus house arrest along with restorative/reparative work. It is worth noting that though it would have been cleaner from the design point of view to compare jail versus house arrest, that option is not accurate: an alternative form of confinement and restorative work go together in the JEP’s sentencing scheme.

That being said, at the time of writing, the specific details of what exactly would be the different possible forms of restorative work had not been announced. I decided to use in the vignette the example of restorative work most commonly used in the media at the time, which was mine clearing. The JEP published a list of accepted restorative work activities in May 2020, ⁵ which included mine clearing.

After hearing the profile, respondents were asked two questions that inquired about different aspects of their trust in the court: the extent to which the respondent supported the decision (specific support) and the extent to which they trusted the JEP (diffuse support). These two items are my dependent variables of interest. The conceptual distinction between trust in the institution and trust in the specific ruling is standard in the literature on courts and public opinion, dating back to Easton (1975). Each taps into a distinct dimension of support for and trust in a given tribunal. The two questions were measured with a 1–7 scale where 1 = not much and 7 = a lot (recoded to 0–6). Randomisation of the four treatments guaranteed that all four groups were identical on observable and unobservable on average. Thus, any systematic difference in the average responses to the questions about support for the decision and trust in the JEP provides an accurate estimate of the effect of the accused’s affiliation (military or former FARC) and of the type of sentence (jail or alternative) on respondents’ evaluations of the JEP.

Other variables used as covariates or to test for heterogeneous effects include “education” in two formats – a continuous version and a dummy that distinguishes between those with high school-level education and those with college and above; “ideology,” measured on a 1–10 scale (left–right) and a categorical version that distinguished between left, centre, and right; “victimisation,” measured in two ways – first, according to type, with two dummies based on questions inquiring whether a family member had been displaced or disappeared and according to the victimiser (whether the perpetrator of violence had been a member of the FARC or a paramilitary or other); “uribista” affinity, measured in two ways – (1) with two dummies based on different combinations of questions inquiring about the likelihood of voting for uribista candidates to different public positions, and (2) with an item that inquired about the respondent’s closeness to the Centro Democrático political party. Finally, I also tested for approval of the peace process.

Results and Discussion

Before I present the results, it is helpful to describe and contextualise the levels of trust in the JEP that we observe in this sample. At the time of writing, the JEP had been in operation for under two years. Two pieces of information are useful to make an initial approach: describing how respondents in this sample perceive the JEP and briefly discussing support towards the peace agreement.

Mean trust in the JEP in this sample was 2.83 (over 6) and the mean support for the decision was 2.86. Figure 1 places mean support for the JEP in context by comparing it with that of other key institutions in Colombia. Two things stand out. First, while respondents’ average trust in the JEP is not in the upper bounds, it is not the lowest of all the institutions presented in the survey. Overall, trust in political institutions is not at the high end in Colombia – not surprising, if one takes into account recent public opinion trends, which register a general decline in trust in political institutions and democracy (see Observatorio de la democracia, 2017). Second, it is worth noting that trust in other judicial institutions is at similar levels to those of the JEP, and even a bit lower: trust in the Constitutional Court and in the legal system as a whole have means of 2.59 and 2.51, respectively.

OPEN IN VIEWER

Figure 1

Mean Trust in Institutions in Colombia. FARC: Fuerzas Armadas de Colombia; JEP: Justicia Especial para la Paz.

It is worth delving deeper into this pattern beyond mean trust and looking at the distribution of responses in terms of categories. Figure 2 graphs trust in a subset of Colombian political institutions as grouped into three categories of the original seven-point scale: those who do not trust the institution, those who are indifferent (respondents who placed themselves in the middle categories of the scale), and those who trust the institution. As we can see, 42 per cent of the sample do not trust the JEP – a figure that is similar, but still lower, than the 45 per cent and 51 per cent that state they do not trust the Constitutional Court or the judicial system, respectively. The JEP has higher positives (39 per cent) than any of the institutions in Figure 2, even higher than those of the Catholic Church. Interestingly, though the Catholic Church has higher overall mean trust (3.5), upon close inspection it becomes clear it also has a much higher proportion of N/A (not applicable) responses. Taken together, Figures 1 and 2 paint a picture of low trust in judicial and political institutions among Colombians. In this scenario, while levels of trust in the JEP are not high, they are on the one hand not as low as those of other institutions, and on the other, not as low as those of such a young and controversial institution could be.

OPEN IN VIEWER

Figure 2

Trust in Institutions (Categories).

Public support for negotiated peace has been relatively high in recent years in Colombia, always between 60 per cent and 70 per cent in the years 2004–2016 (see Tellez and Montoya, in this volume; and Observatorio de la democracia, 2017, though citizen confidence in the accords differs across groups and may be shifting, overall. Indeed, as García-Sánchez and Caviedes, as well as Carlin et al. (this issue) and others suggest, support for the peace agreement varies widely across regions, partisan lines, exposure to violence, and depending on the specific components of the accords (Branton et al., 2019; Garbiras-Díaz et al., 2019; Tellez, 2019). More recently, public opinion polls suggest that perceptions of progress in the implementation of the accords are negative, with 70 per cent of those interviewed stating that the implementation is not going well (Gallup I and Gallup, 2020: 2020).

My first hypothesis was that respondents who were assigned to the profiles with restorative sentences (house arrest plus mine-clearing duties) would be less likely to trust the JEP and have lower levels of support for the decision. The results support H1 only with regard to support for the decision. Figure 3 shows the main treatment effects as measured through an Ordered Least Squares regression (OLS). ⁶ Being assigned to the alternative sentences condition reduced support for the decision by 0.26 and is significant at p < .05. Figure 4 graphs the mean value for support for the decision across the different experimental groups.

OPEN IN VIEWER

Figure 3

Change in Support for Decision and Trust in the JEP (OLS). FARC: Fuerzas Armadas de Colombia; JEP: Justicia Especial para la Paz.

OPEN IN VIEWER

Figure 4

Average Support for the JEP’s Decision by Experimental Group. FARC: Fuerzas Armadas de Colombia; JEP: Justicia Especial para la Paz.

The fact that trust in the institution (diffuse support) and support for the decision (specific support) do not go in lockstep is in line with what Easton (1975) posited early on about trust in older, more established high courts. He hypothesised that greater or lower trust in the institution need not go hand in hand with correspondingly greater or lower trust in the decisions it makes. Citizens’ evaluations of both constructs could and would most likely be orthogonal to each other. In further support of this point, note that the two variables – trust and support for the decision – correlate at a relatively low level in this sample: 0.52. This finding – namely, that support for the decision behaves differently from trust in the institution and is lower when respondents are presented with a defendant who is sentenced to house arrest doing de-mining – is consistent across different model specifications and tests. It is worth noting because the JEP as a young, TJ tribunal is in a dramatically different position than would be an established court. In the medium term, unpopular decisions may begin to hurt institutional trust (diffuse support).

In order to test my second hypothesis, I assessed the effects of being assigned to the former FARC profile, regardless of sentence type. The data do not lend support to this hypothesis (Supplemental Appendix Tables B1 and B2). Additionally, I analysed whether being assigned to specific combinations of both conditions (profile of the accused and type of sentence) had any effect: whether, for example, being assigned the profile of a former FARC combatant who was sentenced to house arrest doing de-mining work (a restorative sentence) had any effects on the dependent variable of interest. Taking into account how unpopular the FARC is, one could expect that more lenient sentences for those defendants would result in less support for the JEP. Neither this nor any of the other three specific combinations had an effect (Supplemental Appendix Tables B3−B6).

Existing research on public opinion and TJ suggests that a number of contextual and individual factors may matter. To better understand these results, I performed a series of additional tests, including for heterogeneous effects, to see whether there were differences across various characteristics. I tested for the moderating effects of education (as a way to measure the potential influence of political sophistication), victimisation (given contradictory findings on the effects of this variable on the literature on Colombia, as mentioned earlier), ideology, attitudes towards the peace agreement, and two variables measuring party identification especially as it relates to affinity with uribismo. The latter is important because the political movement known as uribismo (after Alvaro Uribe, former president and leader of the right-wing Centro Democrático party) has been very critical of the peace process in general, and the JEP in particular. One might expect (and other research has found [see, e.g. Garbiras-Díaz et al., 2019]) that those who feel closer to the Centro Democrático or those who self-identify as uribistas might have a more negative view of the peace accords and its components.

With regard to education, the evidence is inconclusive. Splitting the sample, the data suggest that among respondents with lower education levels, there is less support for the decision (Supplemental Appendix Tables C1 and C2 for the analysis). On average, those who are less educated express 0.30 points less support of a decision with restorative sentencing when compared to one that entailed jail time, a result that is significant (p < .05). Among those with higher education levels, we observe no change in their level of support for the court’s decision, regardless of the sentence type. However, interacting education and both treatments of interest show no effect, which suggests that we do not have enough evidence to conclude that respondents’ education plays any consistent role.

As discussed earlier, existing work on attitudes towards peace and TJ has shown that victimisation can be a crucial element when seeking to understand attitudes towards TJ. The evidence with regard to Colombia is, however, mixed. I ran a number of tests, with different operationalisations for victimisation – including exposure to different types of violence and different victimisers – and none proved significant (see Supplemental Appendix Tables C7–C18). This surprising result is in line with Nussio et al. (2015), who concluded that victims and non-victims did not exhibit different attitudes towards TJ in Colombia.

In the Colombian context, another important explanatory variable could be affiliation or ideological affinity with uribismo. Uribistas tend to be more right wing, hawkish and culturally conservative. As the Centro Democrático (CD) party was in consistent and open opposition to the peace agreements, one could easily imagine that they could have strong disapproving views on TJ mechanisms that appear lenient with the FARC. There is no conclusive evidence that the treatments had different effects on uribistas as measured through willingness to vote for an uribista candidate. Tests for ideology did not yield significant results either (see Supplemental Appendix F).

However, when uribismo is measured through self-reported closeness to the Centro Democrático (Supplemental Appendix Tables C19–C26), those who report feeling closer to this party trust the JEP and support the decision more when compared to those respondents in the sample who are not close to this party. This result is surprising, but since it is not consistent with the ideology and uribista tests, it suggests that the “closeness to the Centro Democrático” item is measuring something more flexible and slightly different. Recall that party identification is low and fluid in Colombia; according to the biannual surveys carried out by the Observatorio de la democracia (2017), between 2006 and 2016, an average of only 28 per cent of respondents stated that they identified with a party (lowest value 22 per cent in 2016 and highest 37 per cent in 2010). What we observe is an electorate that is particularly susceptible to personalistic leadership, and–when it comes to the Centro Democrático–to Alvaro Uribe’s charismatic draw. One possibility is that people report feeling close to his party because of Uribe's popularity, but that not all those who report closeness to the CD are fully consistent, right-wing, hard-core uribistas in terms of their attitudes and voting preferences. Hard-core uribistas are highly unlikely to view leniency with the FARC in a positive light. But there might be a portion of those who report some closeness to CD who are more susceptible to the treatments, because they are not full uribista ideologues and have a higher propensity to being swayed. ⁷

Finally, though support for the peace process with the FARC is positively associated with greater trust in and support for the JEP’s decisions, as one would expect, interaction terms are not significant and the main results hold under these specifications (Supplemental Appendix Tables C27–C29).

Conclusions

This article contributes to our understanding of how citizens in countries dealing with the aftermath of violence perceive the institutions devised to adjudicate on the atrocities of conflict. It takes a first step towards explaining trust in Colombia’s JEP, the tribunal borne out of peace negotiations between the government and the FARC. In line with research that suggests that we need to look beyond courts and the judicial system itself to understand public opinion about courts (Clark and Kastellec, 2015), I posited that to understand Colombians’ evaluations of the JEP, we should turn to the kind of sentences it hands down and to the affiliation of those who stand accused before it. Results show that restorative justice sentences, specifically the most lenient kind that the JEP will hand down, have a negative effect on support for the decision when compared to a more punitive decision, including jail time for the accused. Whether the defendant belonged to the FARC, as compared to being a soldier, had no effect on evaluations of this tribunal.

The fact that specific support and diffuse support behave independently for the JEP raises two issues with regard to the relationship between both components of trust: what can we expect for the JEP in the near future and what this particular case can teach us about trust in courts more broadly. As the results show, both dimensions of trust do not go together for the JEP, a young, transitional court. One could speculate that these two constructs are independent at this stage of the JEP’s functioning precisely because it is early, and citizens lack information and experience with the JEP. Other features, like the kind of sentence, are taking precedence in their evaluations. Going forward, we might well see a lack of support from certain groups hurting institutional trust over time. How so? In her work on trust in Central American judicial systems, Malone (2010) suggested that the deleterious effects of crime on citizens’ evaluations of specific decisions would, over time, hurt diffuse support (i.e. institutional trust), locking these institutions in a kind of vicious cycle. She highlights how external factors can influence trust in the long term. Gibson and Caldeira’s (2009) positivity bias thesis suggests another possible explanation for such negative feedback; institutional trust in high courts is, partly, a function of long-term exposure to their work and to the symbols associated with high courts themselves. Right now, the JEP is still a new institution building its track record. Citizens are at the early stages of being exposed to its activities and decisions. They may not like the decisions, but the lack of priors and information on the JEP’s work allows for a divorce between the two components at this early stage. That might well change in the coming years.

Trust is about long-term exposure to the presence, process, and symbols associated with the institution, but it is also a running tally of the support that rulings generate among different constituencies (Gibson et al., 1998). My research as well as Carlin et al.’s work (this issue) suggest that restorative justice mechanisms and sentencing tend to generate problems for specific trust as well as for the legitimacy of the JEP and the peace agreement overall in Colombia. In that sense, the JEP faces the same difficulties that similar TJ bodies have faced in other countries – a mandate to hand down decisions that the public finds particularly unpalatable. Over time, the accumulated effect of negative evaluations of specific decisions might hurt institutional (diffuse) trust in the JEP. Future research should explore and track how this develops.

Putting these results in dialogue with other work about trust in courts indicates we should incorporate time and ruling characteristics more into our way of thinking about this relationship. It might be that the relationship between diffuse support and specific support could be described using a curvilinear function with respect to a court’s age/salience, that is, it is orthogonal when the court is both very new/relatively unnoticed and also when it is very old/established, but less so in between. ⁸ In the meantime, this research suggests that specific decisions are likely to play a role in shaping institutional support. This is in line with Forero Alba and Rodríguez Raga (2019) study of the Colombian Constitutional Court in which they find that different frames for legal arguments can shape support for a decision and may ultimately impact diffuse support for the tribunal.

The results on the negative effect of restorative measures on support for the decision go in line with much of what we know about attitudes towards TJ. Such work has shown that these can change over time, depending on whether the systems are perceived to deliver on what they promised. According to Backer’s (2010) panel study on South Africa, “two factors that clearly contribute to evolving attitudes toward amnesty are changes in impressions of fairness concerning the amnesty policy and assessments of the extent of individualised truth recovery.” Where truth recovery was absent, respondents expressed worse evaluations of the system. Similarly, Gibson’s findings indicate that the combination of different TJ policies, such as like reparations and truth-telling, could improve support for the outcome. Colombia’s peace agreement contemplates an integral truth, reconciliation, and reparation process. Future avenues for research include exploring how perceptions of different reparation mechanisms, different sentences, as well their levels of implementation influence the attitudes and the lives of Colombians. Making the truth, reconciliation, and reparation process a reality is important for victims, and it might also be crucial for the institutional well-being of the structures meant to guarantee it.

Supplemental Material