Ideological Voting on Chile's Constitutional Tribunal: Dissent Coalitions in the Adjudication of Rights

Abstract

In this article, we examine the relationship between judicial behavior on the Chilean Constitutional Tribunal and the political background of its judges since the constitutional reforms of 2005. We first examine judges’ positions on rulings and find that some distinction has emerged among judges with different political backgrounds and between partisan members and nonpartisans. Notably, these distinctions vary across subject matter and case type. Second, we examine the judges’ behavior in nonunanimous cases using a multidimensional scaling analysis and find that the pattern of dissent coalitions is consistent with a general separation between the judges with center-left and right backgrounds. Finally, we examine several cases to illustrate the patterns on the Tribunal in this period. We conclude that some ideological differences on the Tribunal have emerged while a broadly “political” pattern of judicial dissents has so far not occurred.

Notes

1. We use the term individual rights to broadly mean human rights (usually defined by international law) and civil rights (usually defined by domestic law).

2. For an important recent exception, see Sánchez et al. (2011).

3. Peretti asserts that the dangers are offset by safeguards against overly political courts; safeguards that may be present when courts are involved in statutory interpretation where Congress can change the law to counter the Court but may be absent when courts are involved in constitutional review where Congress would need to overcome the onerous process of amending the constitution to have the final word.

4. Dargent makes a similar assessment of the voting rule for appointing Peruvian judges to Peru's Constitutional Tribunal. He argues that judges selected by two-thirds majority of a unicameral Congress require “political negotiation” to reach an agreement (Dargent 2009: 266). He argues that reaching such an agreement makes appointees more independent. We similarly argue that Chilean judges selected by the Senate and Chamber requiring confirmation by two-thirds of the Senate would make these judges more moderate and more independent from any one political interest in the legislature.

5. While Cea, Peña, and Navarro are regarded as conservatives by observers, it should be noted that all of these judges were appointed by either the CSN or the Supreme Court and therefore have been appointed by the mechanisms with the least direct connection to the political parties.

6. As the data are structured into individual-vote-level observations, we employ case-level random intercepts to account for factors affecting all votes on a case.

7. Based on his analysis for the Court, he finds “In only a few instances did a justice in one wing find himself dissenting in company with a justice from the other wing. This fact would seem to indicate that there were indeed ‘underlying differences of gospel’ in terms of which decisions in practically all of these controversial matters were given. … Locating the justices along a single attitude scale in terms of relative liberalism or conservatism would adequately account for the judicial disagreements manifested during that period” (Prichett 1948: 33). It should be noted that Pritchett's analysis is based on the 15 percent of non-unanimous decisions given by the Court during this period.

8. These types of coordinates are usually called “ideal points” (e.g., Martin and Quinn 2002).

9. This distinction is necessary only because a number of cases (generally, abstract review) have more than one set of dissents and these sets are not necessarily representing a similar perspective. Thus, such cases actually reflect more than one “vote” in the data structure for this section, allowing us to present these coalitions as substantive agreements.

10. Note that here we are including cases with only a single dissent—“lopsided cases”—that inflate the appearance of unidimensionality relative to numbers that might be more comparable to scaling larger voting chambers. The standard fit statistic Average Proportional Reduction in Error (APRE), which accounts for the bias of lopsided votes, is only 0.64 for a one-dimensional model, compared to 0.84 for a two-dimensional model. This indicates a substantial improvement in the predictive value. For comparison, the recent US Supreme Court produces an APRE of 0.76 and 0.89 (Poole 2005).

11. Note that all members are jointly scaled in this period and depicted on the graph as such, despite that one (Carmona) was appointed to replace another (Correa). Individuals appointed since 2010 are not included in this analysis due to insufficient data for scaling purposes.

12. For instance, the following combinations of Concertación judges appear most frequently together: (1) Correa Sutil, Fernández Baeza, and Vodanovic, (2) Correa Sutil and Fernández Baeza, (3) Correa Sutil and Fernández Fredes, and (4) Carmona and Fernández Fredes.

13. The juez de garantía was a judgeship created under the criminal law reform. A single judge or juez de garantía ensures that defendants’ rights are well respected and decides issues relating to pretrial detention prior to transferring the case to another court for oral trial. The oral trial proceeds before a three-judge panel.

14. “La contradicción de ambas normas es palmaria y la Carta Fundamental no autoriza que un imputado puesto a disposición de un juez permanezca privado de libertad por la sola voluntad de un fiscal del Ministerio Público” (Case #1001: 39).

15. Another example of a well-predicted case where the Concertación appointees find against the protection of individual rights is Case #605. This case involved Chile's Tax Code section 116 that allowed regional tax directors to make decisions regarding individuals’ tax disputes. Similar cases pervaded the Tribunal's docket after the constitutional reforms transferred concrete review cases to the Tribunal. The majority in Case #605 and similar cases (generally consisting of Colombo, Bertelsen, Vodanovic, Fernández Baeza, Peña, and Navarro) found that the code section was unconstitutional because it resulted in an improper delegation of judicial power to regional tax directors and because it violated individuals’ rights to equal protection under the law. In this and related well-predicted cases, the dissent consisted of Correa-Sutil and Fernández Fredes who often dissent together. They argued that Section 116 was constitutional because it allowed the tax directors to undertake administrative not judicial functions.

This group of tax decisions culminated in Case #681 declaring Section 116 of the Tax Code unconstitutional creating its first precedent. As such, the Tribunal found inadmissible similar tax cases after the date of Case #681 on the grounds that this issued had previously been decided in the precedent-setting case.

16. The provision that was amended several times prior to the Tribunal receiving the law was drafted in an attempt to prevent private institutions from profiteering on education. Although not mentioned by the Tribunal, many teachers protested another provision of Article 46, which allowed university graduates who did not possess teaching qualifications to teach in areas related to their degree in secondary schools for up to five years (Estrada 2009).

17. Another poorly predicted part of this case involved the constitutionality of Article 10 of the proposed law dealing with the rights and duties of individuals involved in the education system. For this, conservative Bertelsen and more liberal Vodanovic joined to find this provision to be outside the Tribunal's abstract review powers. These judges believed that the Tribunal should not have made any findings as to constitutionality of this section.

18. The case received a great deal of national and international attention.

19. In a separate dissent, Navarro, who rarely dissents, also objected to the majority's finding of unconstitutionality.