The Judicial Trilemma visits Latin American Judicial Politics

The Judicial Trilemma, by Jeff Dunoff and Mark Pollack, studies the dynamic relations between accountability, transparency, and independence, and suggests that designers can only maximize two of these three values at once. They can create a court that has high levels of (1) independence and accountability, (2) transparency and independence, or (3) accountability and transparency, but only at the cost of having a low level of the third value. The article explores these ideas using four different international tribunals, but its insights are not limited to international courts. Domestic designers also have to decide what levels of accountability, independence, and transparency their courts should have, and in making a decision they will face the Judicial Trilemma and confront the hard choice of selecting primarily two out of three values.

correct, shorter terms might actually enhance judicial independence, and longer terms might undermine it. This reasoning is consistent with the replacement of one of Ecuador's Supreme Courts. The attacks against this Court were triggered by the actions of one judge. These actions did not warrant dissolution of the whole Court. However, the absence of a mechanism to hold individual judges accountable increased the President's inclination to dismiss the whole Court. These collective sanctions are not exceptional. Court purges also occurred in Peru and Argentina, two countries that also guarantee life tenure to its judges. In Argentina, mass court purges occurred in 1955, 1966, and 1976. 4 In April 1992, President Alberto Fujimori dissolved both the Supreme Court and the Constitutional Tribunal. 5 In Chile, as one Supreme Court Justice admitted, if the Court had challenged the Government, Pinochet would have closed it down. 6 The domestic settings described here are very different from the international ones in the Judicial Trilemma. 7 However, similar arguments could be made about some international courts. The lack of short and renewable terms to discipline individual judges might cause a state to respond by disciplining the whole court. Unable to terminate an individual judge's tenure or refuse his reappointment, a state could ignore the court's decision, refuse to pay its financial contributions, or denounce the treaty that granted the court its jurisdiction. If this is a real risk, long and nonrenewable terms might undermine judicial independence, and short and renewable terms might enhance it.
In sum, the Judicial Trilemma predicts that courts with high levels of independence and transparency will necessarily have low levels of accountability. However, experiences in Latin America show that judges with longer terms face higher risks than judges with shorter ones. These risks are of a different kind than a veto or a refusal to reappoint. In many cases, they have made courts more dependent than they would have been if there were mechanisms to hold individual judges accountable.

High Levels of Accountability and Independence Can Increase Transparency
From 1992 to 1997, terrorism cases in Peru were tried by faceless tribunals that were designed to protect judges from revenge by terrorist groups. The system was created to increase the number of condemnatory sentences against terrorist suspects who were frequently acquitted by judges too afraid to condemn them. The initiative was a failure. What happened? Instead of attacking individual judges, terrorist organizations like Shining Path escalated violence against all judges and prosecutors in the judicial system. Soon enough, everybody leaked information about the true identity of the judges to protect their lives and jobs. 8 Faceless judges also operated in Colombia for more than nine years. The system's goal was also to protect judges from attacks by "narcos" and "guerrilleros." Like in Peru, faceless judges also increased the number of condemnatory sentences against criminals. However, when the case dealt with a criminal who had the connections and power to force or lure people to leak, the "anonymous" judges were exceptionally lenient. 9 These examples demonstrate that anonymity is a mere parchment barrier if there are not enough incentives to make it self-enforcing-in other words, if all the players in the game do not have an incentive to keep the information confidential. There is no such incentive in a system that tries to maximize accountability: a state needs to know who the judges are and what they did to hold them accountable. If it does not have that information, it will tempt or force the judges to disclose their colleagues' identities.
There are many differences between the domestic setting described here and the international setting studied in the Judicial Trilemma, but a similar logic could apply in some international domains. Suppose that a powerful country loses an important case two votes to one. The three judges will soon face reelection, and the country wants to punish the judges who ruled against it and reward the judges who did not. However, the country does not know who was in the majority and who dissented. How can it get that information? The state knows that it has to reward or punish someone if it wants to send the right message to the judges in the future, but to reward or punish the wrong person will not accomplish that goal. In this situation, it has three options. It can make an informed guess about who the judges are and reward or punish based on that guess. It can threaten to randomly reward only one judge and veto the reappointment of two. It can also threaten to veto the reappointment of all of the judges on the panel. The three strategies have the same goal: to elicit information from the judge who is afraid of being wrongly punished or from others who want to prevent wrongful punishments of their colleagues, surely with the expectation of reciprocity if they ever face similar circumstances. If these judges reason as some faceless judges combating terrorism in Peru and Colombia did, they will secretly leak the information to avoid it.
In sum, the Judicial Trilemma predicts that courts with high levels of accountability and independence will necessarily have low levels of transparency. However, there is no way to hold someone accountable without knowing who he is and what he has done. Accountability requires identifiability. In this position, a state has a variety of mechanisms to get that information, and the judges' dominant strategy-knowing that the state has to punish someone if it wants to send the right message to future adjudicators-is to leak that information to prevent a wrongful sanction.

Dissenting Opinions and Judicial Transparency
Between 2000 and 2004, President Hugo Chavez appointed new judges across Venezuela. The courts were being packed mostly with "jueces Chavistas," but the government also needed to know which judges appointed by previous regimes deserved to be reappointed or promoted. Judges across the country started to "systematically" dissent from every important decision against the government in an attempt to signal their loyalty to President Chavez's regime. These dissents were part of the evidence that three illegally dismissed Venezuelan judges presented against Venezuela in a 2004 case at the Inter-American Court of Human Rights. Their argument was that the judges who had signaled their loyalty through these dissents had been confirmed or promoted, whereas judges who had ruled against the government or had not issued such dissents had been illegally removed. 10 It is not clear if dissents make a court more transparent. The suggestion might be correct in an international setting, where judges sincerely act as the states' trustees and exercise their authority in a nonpolitical manner. However, in many domestic settings dissents have been used as a political tool that has only made the judicial process more opaque. If judges believe that dissents can harm a court's legitimacy, the threat of a dissent can force judges to compromise their best-considered judgments to avoid that outcome. If judges are personally averse to criticisms, they might avoid dissenting to prevent criticisms against (that is, dissents from) their own opinions. As was seen in the case of "jueces Chavistas," dissents can also be used to avoid a punishment or gain a reward. The Judicial Trilemma identifies transparency with the possibility of identifying each judge's vote and opinion. If the goal of transparency is to promote an exchange of sincere professional opinions, transparency and accountability may work at cross-purposes. If the real reasons behind dissents are strategic, accountability has the perverse effect of making the process less transparent.

High Levels of Accountability and Transparency Can Increase Independence
In 1981, Argentina had one of the most accountable judiciaries in the world: an endless cycle of new government-new courts had repeatedly violated the judges' life tenure and independence. 11 General Jorge Rafael Videla, the dictator, had appointed most of the current judges, and few thought judges appointed by a dictatorship could turn against it. In 1982, following the public loss of faith in the government after the defeat by the British over the Falkland Islands, decisions against the government almost tripled. 12 The country was starting its transition to democracy and judges were strategically defecting. They were signaling their willingness to change sides. 13 Like Argentinian judges during the 1980s, even judges in the most accountable judiciaries can find the elbow room to act independently. No selection system is perfect and many judges have exploited its various defects to gain increased independence. For example, suppose a legislature or a judicial council appoints judges by majority vote, and no member or party has enough votes to form a majority. If power is symmetrically distributed within a five-member council, there are ten different coalitions that could appoint a judge. In this situation, a judge can gain more independence by threatening to defect to a new coalition. If the current appointers restrict her too much, there are nine other options to negotiate better terms. In turn, if the current coalition believes this is a credible threat, it will submit to the judge's demands.
The Judicial Trilemma predicts that courts with high accountability and transparency will necessarily be courts with low independence. However, unless appointments are made by unanimity, judges can exploit the imperfections of the selection mechanisms-especially majority rule-to gain more independence. Furthermore, to the extent high levels of transparency make judicial votes and opinions public, judges are allowed to credibly commit to agreements they could not reach if no way existed to observe their behavior. Thus, to the extent transparency allows the monitoring of compliance, it might actually increase the judges' bargaining and power.
One court that has exploited the imperfections of its selection mechanisms to gain independence is the "hyperactive" sala constitucional of the Supreme Court of Costa Rica. 14 By Dunoff and Pollack's definition, the sala constitucional would be an accountable and transparent institution. Its judges are elected by the legislature for a period of eight renewable years. However, the sala constitucional is one of the most independent courts in the region and one of the few that systematically enforces constitutional rights and arbitrates interbranch conflict. 15 How, then, did the sala constitucional achieve its independence? The answer lies in the way political power is distributed in the Legislative Assembly. By default, unless a majority vetoes them, judges are automatically reappointed. There is no single party with enough votes to veto a judge. 16 In other words, a judge who is up for reelection can be supported by different coalitions. The fact that multiple coalitions can support a judge's reelection gives her more independence than if only one could.
In sum, to know the real level of judicial accountability, we need to know how power in these bodies is distributed. Judges accountable to bodies in which power is fragmented will have more independence than judges accountable to bodies in which power is concentrated.

Expanding the Model: Three Players and One Court
The theme binding these examples together is the presence of a third strategic player: the rational and self-interested government that, well after the court is created, appears in front of it as an actual or potential litigant. The incorporation of this third player is essential in any theory of judicial design in Latin America. Most of the confrontations between courts and other branches of government have happened precisely because the officials who appointed the court are not the same government officials who at a later time appear in front of it as litigants. Even if they are the same, various factors require them to be treated as two different entities. For example, at the time of creation, a state might decide that an independent court will serve its interest best given the information it has about the cases that will likely come before the court. However, at a later time, once the state observes the cases that have actually come before the court, it might regret its decision and conclude that a more accountable institution is the better choice. The problem is worse when there is an actual change of government. For instance, at the time of creation, Party 1 appointed all the judges. At a later time, Party 2 wins the elections and finds itself subject to the judgments of a court appointed by its political adversary. In these circumstances, the government may try to influence the court even though doing so would violate its initial agreement to respect the court's independence or the confidentiality of its procedures. 17

Conclusions
Are domestic courts different from international courts? If so, how different are they? In this response I have tried to show how the Judicial Trilemma would operate in some Latin American contexts. I have given examples of faceless judges in Peru, court purges in Ecuador, insincere dissents in Venezuela, and judicial strategy in Argentina and Costa Rica to show how Trilemma's logic could, in some circumstances, be perversely self-defeating. Despite their differences, some similarities could exist to make the logic of these examples apply to international tribunals. If changes in power or the passage of time can alter a government's preferences, a state could well regret its decision to delegate power to an international court and then decide to use other mechanisms available to influence the court's behavior. If this is correct, some states might be willing to use options of exit or voice to undermine their initial compromise to respect the court's institutional features.