Judicial Independence across Democratic Regimes: Understanding the Varying Impact of Political Competition

Abstract

One of the most prominent explanations of the creation and maintenance of independent judiciary is the “insurance theory” that proposes a positive relationship between political competition and judicial independence. But, does intense political competition inevitably lead to higher levels of judicial independence across all types of democracies? Conducting a large-N cross-country analysis over 97 democratic countries, this study shows that as democratic quality across countries changes, the impact of political competition on judicial independence changes as well. The empirical findings reveal that while in advanced democracies high levels of political competition enhances judicial independence, in developing democracies political competition significantly hampers the independence of the courts.

A judiciary that is insulated from legislative and executive influence as well as from other private interests is not only the fundamental principle of the rule of law but also the central precondition for good governance and consolidation of democracy. Independent courts serve as an effective mechanism that controls and constrains the operation and power of the legislature and executive. Independent judges, for instance, have the power to punish political authorities who abuse or misuse their position. On the other hand, through judicial review independent courts can declare legislative acts or government policies unconstitutional. Being insulated from electoral accountability and other political interferences, an independent judiciary may also produce counter-majoritarian decisions. But then why do the elected representatives of democratic countries construct an independent judiciary in the first place and try to maintain it even when the courts do not render decisions in conformity with their interests or policies?

According to the insurance logic of judicial independence, the politicians who face the possibility of losing power seek to limit their opponents by supporting judicial independence. This logic posits that the ruling elites, who expect to fall into minority status after elections, might want to strengthen the courts in order to protect their own rights and liberties once they become political minorities (Ginsburg 2003). In other words, the advocates of the insurance theory emphasize that in the long-run the incumbents may have long-term benefits under an independently performing judicial system. Independent courts are perceived by these incumbents as a mechanism that would protect them from the opposition's attack after future electoral change (Finkel 2008; Ginsburg 2003) or ensure that legally enacted policies continue to be implemented even after they leave office (Landes & Posner 1975). As a result, the proponents of the insurance theory argue that politicians offer independent courts when political competition is intense and incumbents' expectation of winning the future elections is low (Finkel 2008; Ginsburg 2003; Landes and Posner 1975; Ramseyer 1994; Stephenson 2003). Thus, attributing high levels of judicial independence to intense political competition, the advocates of this theory appear to envision a positive relationship between these two aspects.

Although the underlying logic of the insurance theory is quite appealing, it does not explain why we do not see high levels of judicial independence in all democratic countries with high levels of electoral competition. I argue that the cost-benefit analysis that the rational political elites have to undertake while choosing their judicial policies would reflect different trends across advanced and developing democracies. Hence intense political competition would not inevitably lead to high levels of judicial independence across all democratic countries. Thus we should not expect a similar impact of political competition both in advanced and developing democracies.

By advanced democracies I mean regimes where democratic values are fully consolidated and political processes are successfully institutionalized. In these types of regimes democracy and its rules are perceived to be “the only game in town” (Linz & Stepan 1996). Citizens and leaders conclude that no alternative form of regime has subjective validity. The party system is stable and the political parties have strong networks of grassroots organizations. Democratic values are highly internalized by the citizens. Individual rights and civil liberties are protected by the rule of law. Developing democracies, however, are regimes that meet the procedural minima for democracy but lack consolidation of democratic values and institutionalization of political processes. The weakness in protecting individual rights and civil liberties threatens the very existence of public opposition. Media is often controlled by the state and strongly supports the regime. The party system is underdeveloped and volatile. With high volatility, the entry barriers to new parties are lower, and the likelihood that personalistic politicians become the head of government is higher (Mainwaring & Zoco 2007).

In line with these differences between advanced and developing democracies, it seems logical to presume that the incumbents' cost-benefit analysis of offering independent courts might be different. Especially in developing democracies—which are characterized by high levels of corruption, weak party systems and high electoral volatility- the immediate short-term benefits that incumbents may obtain from interfering in judicial decisions may be higher than the long-term benefits that may be gained from high levels of judicial independence. Given the fact that in developing democracies citizens have lower levels of confidence in the judiciary; the media is highly controlled by the government; citizens have limited awareness and willingness to participate in politics; and that the political and civil rights of the citizens are not efficiently institutionalized, the power holders may be less fearful of public reaction than their associates in advanced democracies. For the politicians who aim to offer a dependent judiciary, this situation may lower the costs of intrusive behaviors. Hence, in the context of developing democracies when political competition is highly intense, the incumbents may be more inclined to interfere in judicial decision-making.

This article does not present political competition as the only factor and insurance logic as the only mechanism that accounts for high levels of judicial independence. It acknowledges that there are many contextual and institutional factors which account for high levels of judicial independence.¹ Yet the main objective of this article is to test the insurance logic of judicial independence and albeit indirectly show whether—and if so how—the impact of political competition on judicial independence changes across advanced and developing democracies. I suggest that while in advanced democracies political competition has positive impact on judicial independence, in developing democracies it has negative impact. In this regard, I develop an empirical model and test it across 97 democratic countries.

The article proceeds as follows: the first part of the study gives the theoretical framework about the varying relationship between political competition and judicial independence across advanced and developing democracies. The second part introduces the data, key variables and the empirical model. The third part is the empirical section where the main hypothesis is tested and results presented. The last part concludes the study by discussing both theoretical and practical implications of the findings.

Political Competition and Judicial Independence: Theoretical Framework

Although the literature on judicial independence is characterized by various conceptual debates about the meaning of judicial independence,² a judge is independent when, “… she does not face undue external or internal pressure (as say from hierarchical superiors) to resolve cases in particular ways” (Rios-Figueroa & Staton 2009:12). On this account, a judge is independent when she can take decisions based on her own preferences and interpretation of law. Thus, judicial independence refers to independence of the judicial system from external political, economic and social influence, and to the ability of individual judges to make independent decisions based on their own interpretations of law. In line with this meaning of judicial independence, two of its characteristics are evident. The first is “impartiality” and refers to the idea that judges will base their decisions on law and facts (Shapiro 1981). A second trait of independence is “political insularity” (Fiss 1993) and refers to the condition that judges should be protected from political interference that might affect their impartiality. While identifying judicial independence, one should recall that the courts do not operate in vacuum. A number of exogenous factors will influence the judges' opinions and will have varying impacts on their impartiality and insularity (Larkins 1996). Although constitutional protections are presented as critical determinants of judicial independence, the independent performance of the courts cannot be achieved unless politicians and political factors construct the appropriate context.

One of the central debates in judicial politics literature is about whether—and if so how—the political institutions constrain the judicial decision-making. In this regard, a vast body of literature focuses on the relationship between political competition and judicial independence. While the proponents of the insurance theory focus on the relationship between the politicians and the courts from the political decision-making perspective—which aims to understand the politicians' calculations and decisions to maintain judicial independence, another strand of research focuses on the same relationship from the judicial decision-making perspective (see Epstein & Knight 1998; Ferejohn & Weingast 1992; Gely & Spiller 1992; Marks 1989). Formulating their analyses through separation-of-powers models, the basic idea of the latter group of studies suggests that the concentration of political power across the branches of government forces judges to behave strategically in order to avoid having their decisions overturned or to prevent some political sanctions. In line with this logic, some scholars designate political fragmentation as a proxy for political competition and assert that when political power is highly fragmented the judiciary would be more independent³ (Chavez 2004; Harvey & Friedman 2006; Iaryczower, Spiller, & Tommasi 2002; Ríos-Figueroa 2007).

Nevertheless, political fragmentation is just one of the mechanisms through which political competition affects the performance of the courts. The basic feature of this mechanism entails that political fragmentation reduces the capability of incumbents to interfere in judicial decision-making because the dispersion of power makes it more difficult to obtain the political support to curtail the autonomy of judges. Although political fragmentation or high number of veto players/sanctioning players may help to control and constrain the incumbents' intervention in the judiciary, the real independence cannot be achieved without a real intention among the political elites. In this regard, electoral competition appears as another key mechanism through which political competition affects the incumbent politicians' preferences for maintaining judicial independence.

Under a democratic regime the ruling government can only maintain its power through re-election, but intense electoral competition increases the probability of losing its office. Hence, the extent of competition between politicians affects the policy choices of the incumbents. According to one strand of research, respecting the independence of the courts may increase the politicians' expected payoff. This logic has led some scholars to think of judicial independence as a form of political insurance that incumbents buy to reduce the cost of being out of office. Thus political insurance is perceived by the incumbents as protection from the opposition's attack or preservation of policy stability after future electoral change.

Ginsburg (2003), for instance, argues that when political incumbents expect to win future elections, they have little incentive to empower the judiciary. However, when political competition is intense and the incumbents have low expectation of retaining their positions, they are more likely to support an independent judiciary through which they will be able to challenge the policies of the incoming government. Extending Ginsburg's thesis to Mexico, Finkel (2008) contends that the 1994 judicial reform in Mexico, with its introduction of new judicial review powers and independence guarantees, was motivated by the ruling party's fear of losing power. Thus, political incumbents delegated power to courts to preserve their rights in case they were to later become the opposition. This argument foresees that political incumbents will give up current opportunities to attack opponents through the courts in exchange for insurance that they will not be attacked once they find themselves in the opposition. This logic necessitates high levels of trust between the political actors. However, in developing democracies which are characterized by deep-mistrust among political actors, the credibility of this commitment would attenuate (Popova 2010).

According to Landes and Posner (1975), on the other hand, independent courts are likely to ensure that legally enacted policies continue to be implemented even after the politicians who put them in place leave the office. The scholars argue that incumbent politicians who pressure the courts will not be able to attract interest groups to support their policy proposals because interest groups would know that the policy will not endure after those politicians leave the office. Yet assuming that interest groups would value long-term policy stability over short-term benefits of short lived policies is proved to be empirically wrong in the context of developing democracies (Hellman 1998).

Adopting a similar logic to Landes and Posner (1975), Ramseyer (1994) argues that by appointing judges who are ideologically close to them, the incumbent politicians increase the probability that the judiciary would prevent future incumbents from completely scrapping their policies. In Ramseyer's account, the main objective of the incumbent politicians is ensuring policy stability and enlarging their influence during periods when they are out of power. Yet, they do this at the cost of decreasing their influence over policy while they are in power. In this regard Ramseyer (1994: 742) says “… because politicians will have to run the country with independent judges that their predecessors appointed, they will necessarily have less impact over policy while in office.” This argument assumes that incumbent politicians would value future policy stability and they would expect to be in politics for a long time. Yet, in developing democracies, where the rule of the game is not fully established, the incumbent politicians may not care much about the future. Thus they may not prefer to sacrifice their current policy control in order to increase their future policy control.

Finally, including forward-looking politicians as a necessary condition for inter-temporal calculus, Stephenson (2003) offers a formal model of the mechanism presented by Ramseyer (1994) and finds empirical support for his prediction. Focusing on 159 countries he argues that judicial independence is sustainable only when the political system is competitive and contends that political parties find judicial independence less attractive as they do not fear losing in the next election.

As a result, the insurance theory suggests that the long-term benefit that the incumbents expect to reap from an independent judiciary is either an insurance against any future political harassment or protection of the existing policies. Although the underlying logic of the insurance theory is quite appealing, it does not explain what happens when a hostile government confronts a newly empowered court. Why would incumbents believe that the next government will continue to support the independence of the judiciary? The insurance theory assumes that the incumbents make an inter-temporal calculus and expect to be in politics for a long time to credibly commit to such an arrangement. But especially in newly democratized countries the deep seated mistrust among the politicians would undermine the credibility of this commitment. Thus, while the proponents of the insurance theory emphasize the long-term benefits of an independent judiciary, they seem to neglect the short-term benefits of providing a subservient judiciary. In other words, under certain conditions the incumbents might not give up current opportunities to harass opponents through the courts in exchange for the long-term benefits of an independent judiciary. For that reason, a closer scrutiny of the politicians' cost-benefit calculations in offering subservient courts would provide a theoretical explanation about whether—and if so, why—the impact of political competition on judicial independence may change across advanced and developing democracies.

The Cost-Benefit Analysis of Pressuring the Judiciary

The politicians' immediate consideration usually hinges on the upcoming elections so that they tend to make decisions that will bear fruits in election results. In this regard, when political competition is intense, the outgoing incumbent—current government that faces a high probability of being replaced in the upcoming elections—can reap immediate benefits from interfering in the judiciary. The major benefit of a subservient judiciary would be its aptitude to increase the incumbent government's probability of remaining in office. In other words, a subservient judicial system can help the incumbent government to maximize its chance of re-election by weakening the social credibility, financial and even legal standing of the opposition parties.

Especially in developing democracies where political parties lack well-developed grassroots organizations and stable financing, a few court decisions can cause considerable damage. Through one single trial a court can undercut a party's campaign budget and close down a party's newspaper that will have negative impact on that party's legitimacy. Nevertheless, in advanced democracies the same court decision might have a smaller effect on established parties that have a strong network of grassroots organizations to energize their base (Popova 2010). Moreover, many parties in developing democracies are used only as vehicles for their leaders to participate in elections. Thus, a court decision to remove the party leader from the ballot could destroy the whole party (Popova 2010). In some developing democracies such as Turkey, the Constitutional Court may also frequently close down political parties, ban their members from political participation or cut their budget to an extent that would directly affect their very existence.⁴ Yet, in advanced democracies where personalistic parties are not the common trend and where the political processes are fully institutionalized these kinds of judicial decisions would not generate the same impact. Thus, in the context of developing democracies the immediate short-term benefits that incumbents can obtain from offering a politically dependent judiciary may be higher than the long-term benefits that power holders may reap from supporting the independent performance of a judiciary.

Under highly competitive political environment, while an incumbent government tries to decide between the long-term or short-term benefits of dependent courts, it should also take into consideration the corresponding costs. Rebolledo and Rosenbluth (2010), for instance, emphasize that in the countries with a relatively short experience with democracy, voters have only weak incentives to monitor politicians and to punish them at the polls for irregular political actions. Thus it will not be wrong to argue that the costs of pressuring the judiciary are closely related to the overall democratic structure of the society.

The major potential cost that an incumbent government would have to deal with due to its interference in the judiciary is public backlash. Thus if incumbent politicians expect that a strong public reaction would follow any attempt to pressure the judiciary, they will refrain such actions. Especially as political competition increases the incumbent politicians would be more sensitive to public backlash. But in order for the citizens to efficiently hold the officials, who intervene in the judiciary accountable, the electorate (1) has to be informed about the wrongdoings of the incumbents and able to discern the executive's attempts at interfering in the judiciary (2) has to have high levels of confidence in the judiciary and (3) has to be capable and willing to punish the incumbent.

In order for a public enforcement mechanism to work, the court cases and the relationship between the courts and political actors must be sufficiently transparent (Vanberg 2005). As an individual becomes better informed about the political processes, and her/his interest in such processes increases s/he forms better informed preferences and attitudes. The more informed the individual, the more s/he understands how the political process works. And this increased knowledge about the system is expected to generate stronger attitudes about the independence of the judicial system. In this sense, an independent and free media is the most crucial mechanism that would help to provide transparency and increase the public awareness. Yet, one should recall that in developing democracies media is largely controlled by the government and is itself an object of attack (Simon 2004). In some developing democracies media is also almost inexperienced in investigative journalism (Waisbord 2002). Because of all these reasons, in developing democracies the level of transparency is low and this aspect decreases the public awareness of the wrongdoings of the incumbent politicians.

A strong public belief in the courts' legitimacy is another important factor that will make incumbents perceive public backlash as a credible threat. Many scholars have pointed out judicial legitimacy as the key for an independent and powerful judiciary (e.g., Caldeira 1986; Gibson 1989; Gibson, Caldeira, & Baird 1998; Murphy and Tanenhaus 1990). Accordingly, Staton (2010:13) argues that: “If we continue to assume that public preferences constitute the primary incentive for political action in the elected branches, then we can conclude that the public will influence the choice to respect judicial decisions.” In this regard, if the electorate has low confidence in the judiciary, it may tolerate the political interference and the judges will lack the leverage to exercise authority. In contrast, if the electorate is unwilling to accept any interference in the judiciary, the judges will have the leverage to influence policy outcomes effectively (Carrubba 2003; Stephenson 2004). Yet, does public confidence in the judiciary changes across advanced and developing democracies? Looking at the public confidence in the justice system across 49 countries, a recent study finds that in advanced democracies confidence in the judiciary is higher than the confidence levels in developing democracies⁵ (Aydın & Sekercioglu 2010). Hence, in developing democracies where the society does not hold strong confidence in the judiciary, an incumbent's attempt to create subservient courts may not lead to considerable public backlash.

Finally, one should recall that incumbents would perceive public backlash as a credible threat only when the citizens are capable and willing to punish the politicians who attempt to pressure the judiciary. Yet, having enduring memories inherited from their countries' previous authoritarian regime, the people living in developing democracies are usually accustomed to political interference in judicial affairs. The incumbents still have access to the mechanisms for pressuring the judiciary because it is quite difficult to root out these types of informal channels (Solomon & Foglesong 2000). The existence of these informal channels and memories of political intervention in the judiciary lower the cost of implementing an attack on judicial independence. A good example for this discussion might be Argentina under the Presidency of Menem. After a peaceful transfer of power at the end of a competitive election, the Menem administration publicly stated that a judiciary that was able to resist Menem's economic reforms would ruin any chance for economic recovery (Larkins 1996). So, Menem turned his attention to pacifying the Court. When questioned why he was not trying to improve the judiciary's capacity to protect the rule of law, Menem responded: “Why should I be the only president in fifty years who hasn't had his own court?” (Walker 2006:784).

A society's strong commitment to individual freedom and protection of liberal rights would also affect the individuals' willingness to punish the political elites who would intervene in the judicial decision-making. High levels of political participation and interest in politics indicate a higher engagement with the political system, a better understanding of the political processes and a higher motivation and/or ability to evaluate the system. Contrarily, the lack of these democratic values would not only lower the willingness of the citizens to punish the incumbents who intervene in the judiciary but would also lower the legitimacy of the judicial system.⁶ Hence, in developing democracies where the society does not hold strong democratic values, an incumbent's attempt to create subservient courts may not lead to considerable public backlash.

For all these reasons, in developing democracies the incumbents, who decide to pressure the judiciary, may have less fear of public backlash compared to their counterparts in advanced democracies. Contrary to the insurance theory that envisions the benefits of judicial independence would outweigh the related costs; in developing democracies the benefits of subservient courts may outweigh the related costs. In a simplified manner, the logic of the argument can be illustrated by the following specific examples.

Consider for example, Pakistan, wherein a number of opposition parties called for the resignation of President Musharraf to ensure free and fair elections. In October 2007, however, Pakistan's Electoral College re-elected Musharraf to a new five-year term in a controversial vote that many called unconstitutional. The interesting turning point in this sequence of events is the fact that five months before this flaw re-election, Musharraf had dismissed the country's Chief Justice (Musharraf 2008). Since in its recent history the Supreme Court's rulings damaged Musharraf's standing and credibility, it can be said that he tried to restore the public support by attempting to create a subservient court. Moreover, the powerful criticisms from the opposition parties signaled an imminent threat for Musharraf who decided to use whatever tools were available, including judicial manipulation, to stay in office. As a result, the case of Pakistan shows that in a developing democracy—which is characterized by political crises, institutional weakness and where two-thirds of the public is unable to provide a meaning for the term “democracy”⁷ —intensified political competition would lead the incumbent leader to curtail the independence of the judiciary in order to use it as a tool to remain in power.

Ecuador can be given as another example in this regard. In 2004, a group of opposition deputies signed a petition to create a committee in order to investigate certain charges against President Gutierrez who was accused of corruption. However, the impeachment trial request against the President was unsuccessful. After this incidence, claiming that the Supreme Court was loyal to his political opponents, President Gutiérrez and his congressional allies dismissed 27 of the Supreme Court's 31 judges and replaced them with their own political allies (BBC News, 9 December 2004). The interesting turning point in this sequence of events is the public rioting that started after the new Supreme Court justices dropped corruption charges against two former presidents. After those riots, the Ecuadorian Congress ousted President Gutierrez (BBC News, 17 February 2005). The President was arrested and detained on charges of endangering national security but he was released in 2006 after a judge dismissed the charges. Yet, the judicial independence in Ecuador is still under siege.⁸ As a result, the Ecuadorian case reveals three important aspects that characterize the relationship between the courts and political actors in developing democracies. First, in these types of democracies the incumbents perceive the creation of a subservient court as an important mechanism to remain in power and fight with corruption charges. Second, given that the Ecuadorian society has not fully consolidated the key values of democracy,⁹ in his cost-benefit calculation in pressuring the judiciary President Gutierrez does not appear to perceive the public backlash as a credible threat. Third, the public protests which were successful at ousting the President showed that the public backlash can be an efficient and credible control and constrain mechanism.

As a result, I suggest that the impact of political competition on judicial independence changes across advanced and developing democracies. And the main hypothesis to be tested is:

Hypothesis: While in advanced democracies political competition has a positive impact on judicial independence, in developing democracies it has a negative impact.

In the case of the main hypothesis being verified, it would be safe to conclude that in developing democracies as political competition increases, the incumbent governments tend to manipulate the judiciary and use it as a mechanism for re-election. At this point assuming that political competition has similar impacts on judicial independence across all types of democracies would be quite misleading.

Data, Measurement, and Model

There are many aspects emphasized by the scholars as possible determinants of judicial independence. These aspects range from institutional characteristics of the judiciary to the external environment (political, economic, and social) within which it operates. Yet by focusing on the impact of political competition, this study extends on only one of the explanations of judicial independence. Controlling for other possible determinants, in this section I develop and test an empirical model that, albeit indirectly, aims to show whether the impact of political competition changes across advanced and developing democracies.

Suggesting that the effect of political competition on judicial independence is mediated by the quality of democratic performance; the current study uses the 2000–2008 data for 97 democratic countries to test its hypothesis. Following Robert Dahl's (1971) Polyarchy, I consider as democratic the countries in which regimes hold elections and the opposition has some chance of winning and taking office. Thus, adopting a minimalist definition of democracy¹⁰ the sample is composed of regimes in which the executive and the legislature are both chosen in contested elections. Using the Cheibub, Gandhi, and Vreeland (2010) database, I apply this definition to the countries which the data indicates that have elected executive and legislative bodies and a legislature which is composed of multiple political parties.

A closer scrutiny of the literature shows that judicial independence is generally studied under two main categories: “de jure” and “de facto.” De jure judicial independence refers to the institutional guarantees outlined in constitutions and contains issues such as the tenure of a judge, the nomination process, and salary protections. De facto judicial independence—that is the dependent variable of our empirical model—focuses on judicial behavior and tries to discern whether and how the formal rules are implemented in practice (see Rios-Figueroa & Staton 2009). Yet, there is no direct way to measure objectively the level of de facto judicial independence. For 134 countries the World Economic Forum's Global Competitiveness Report (Porter & Schwab 2003) provides standardized and relatively comprehensive subjective assessment of judicial independence. This indicator measures the experts' opinion regarding the independence of the judiciary in their own countries through the following question: Is the judiciary in your country independent from political influences of members of government, citizens, or firms? Countries are coded in a scale ranging from 0 (no-heavily influenced) to 10 (yes-entirely independent). Calculating the average judicial independence index of each country for the period 2000–2008, judicial independence in our sample varies from 0.64 in Venezuela to 9.09 in Denmark. The mean value of judicial independence in the sample is 5.04 with a standard deviation of 2.04. Of the 97 countries 36 had judicial independence under 4, while 33 had values between 4 and 7; and only 28 had values above 7.

In order to test the hypothesis suggesting that the impact of political competition on judicial independence changes across advanced and developing democracies, I model political competition and the level of democracy as the key independent variables of the study.

Political Competition

Although political competition could take different forms, the most common framework involves electoral competition, in which politicians or parties must compete for public support via elections. In order to measure political competition a variable that proxies for parties' subjective assessment of their probability of controlling the government is needed. The theoretical framework of this article—the insurance theory—suggests that political competition generates uncertainty which in turn provokes the incumbents to support the creation of an independent judiciary (Finkel 2008; Ginsburg 2003). Yet, challenging the insurance theory I argue that in developing democracies while political competition is intense and the incumbent party has lower chance of winning the upcoming election, a subservient judicial system can help the incumbent government to maximize its chance of re-election. For that reason our measure of political competition should be able to capture the political uncertainty and demand for insurance in both multi-party and two-party systems. For instance, one might argue that the “effective number of parties”¹¹ measure—which is frequently used to measure political competition—correlates with political uncertainty. A smaller number of parties in the parliament would indicate a higher chance of each party to capture seats in government and this would signify a lower political uncertainty. But what the proponents of the insurance theory imply by the notion of political uncertainty is the condition where an incumbent party has low chance of winning the upcoming election. Thus the “effective number of parties” measure would not capture the political uncertainty in a legislature dominated by two equally large parties.

For all these reasons I measure political competition by taking the difference between the percentage of seats of the winning party (or bloc of parties) and the percentage of seats of the runner up in the legislature. Regardless of the number of parties in the legislature, a smaller difference between the seat shares of these two parties (or blocs of parties) would indicate a higher competition and a higher uncertainty about the upcoming election results. On the other hand, the higher differential between the seat shares, the more certain will be the leading party/bloc that will end up in power. This measure would also capture the extent to which there is a dominant party. But still the presence of electoral competition means that even the most dominant and popular party faces a relatively higher chance of losing power than it would under a one-party system.

The data for political competition is drawn from the IFES Election Guide (2010) for the period 2000–2008. For the reason that there is a direct inverse relationship between “the differential between the seat shares” and “political competition” (political competition increases as the difference between the seat-shares decreases) I create a variable of political competition that takes the inverse value of the difference between the seat shares. The measure of political competition is normalized between 0 and 1, “0” indicating minimum political competition and “1” referring to maximum political competition.¹² For each country the average political competition for the given time period (2000–2008) is calculated. In our sample the values of the political competition variable ranged from 0.036 to 0.989 with a mean of 0.777 and a standard deviation of 0.199. The data was heavily skewed towards the higher end of the scale. Only 13 countries had political competition lower than 0.5.

Level of Democracy

I take the level of democracy as an indicator that differentiates between advanced and developing democracies. “Level of democracy” refers to a rough categorization of democratic regimes ranging from strong democratic regimes where democratic values are fully consolidated and political processes successfully institutionalized (advanced democracies) to democratic regimes that meet the procedural minima for democracy but lack consolidation of democratic values and institutionalization of political processes (developing democracies). The Freedom House (FH) measurement scale is used as a tool to operationalize this abstract classification of democratic regimes. The FH scale becomes an agreeable tool for categorizing democratic countries according to the institutionalization and protection of the political rights and civil liberties of citizens.

The use of the FH measure in the current analysis, however, necessitates careful thinking. First, one of the sub-scores of the FH index is the rule of law that in addition to other legal protections also measures judicial independence. In this regard, including judicial independence in both sides of the equation will create fundamental methodological problems and lead to biased and inconsistent estimates. Second, the FH index also includes the measure of the electoral process as well as political pluralism and participation. Using these measures in the analysis will also lead to multicollinearity problem since one of the key independent variables of the model is political competition. Thus, I recalculated the FH index to obtain a democratization level index that is exogenous to judicial independence and political competition. I excluded the rule of law, electoral process, political pluralism and participation and functioning of government sub-scores that are used in calculating the FH index. I recalculated the democracy index by summing up each country's points only on the subcategories of Freedom of Expression and Belief, Associational and Organizational Rights, and Personal Autonomy and Individual Rights. In total each country's democracy point ranges between 14 and 44. Thus having a continuous character the recalculated FH index is normalized between 0 and 1. In our sample the values for level of democracy variable ranged from 0 (Cameroon) to 1 (7 countries) with a mean of 0.716 and standard deviation of 0.247. The data was heavily skewed towards the higher end of the scale.

Control Variables

In the empirical model I also explore the robustness of our estimates to the inclusion of other potential judicial independence determinants which are related to the executive-legislative arrangement, type of the legal system, and the constitutional status of courts.

How might presidentialism and parliamentarism affect judicial independence? For instance, Ackerman (1997) suggests that presidentialism is good for courts by providing them with a role as an arbitrator among law-making powers. Because of the potential for institutional divergences between the executive and legislative branches, it is stated that the presidential systems support judicial activism. Ginsburg (2003) argues that these divergent policy views can be ameliorated by the presence of a powerful and independent constitutional court. In addition, division of power between branches can allow the courts to exercise greater independence in their rulings because attacking the court may be more difficult in systems where passage of legislation requires the cooperation between two separate political bodies. On the other hand, some scholars argue that a strong president can control the actions of judges through acts of coercion. Hayo and Voigt (2007), for instance, observe that presidential systems enjoy lower levels of judicial independence compared to parliamentary systems. According to this line of thinking strong presidents are more easily able to retaliate against the courts for unfavorable decisions. Likewise some scholars show that concentrated executive power is a significant factor in explaining judicial subservience in the post-communist countries (Herron & Randazzo 2003). According to this line of thinking, strong presidents impose substantial constraints on judicial independence.

Measuring the relationship between the executive and legislature as a form of government, I include executive-legislative arrangement in the empirical model as a control variable. Indicating whether a country has a parliamentary, presidential or semi-presidential regime, the data for executive-legislative arrangement variable is taken from the World Bank Political Institutions database (Beck et al. 2001). Parliamentary systems are coded as 2, semi-presidential systems are coded as 1 and presidential systems are coded as 0.

Another system based explanation that is elaborated in the literature as a possible determinant of judicial independence is the type of the legal system. In their study, La Porta et al. (1998) find that the quality of law enforcement differs across legal families. They take “efficiency of the judicial system” provided by the Business International Corporation as a proxy for the quality of law enforcement and find that countries with common-law tradition have more efficient judicial systems. On the other hand, Djankov et al. (2003) empirically demonstrate that the legal origin explains about 40 percent of the variation in legal formalism. In the data for legal systems, civil-law legal tradition is coded as 0 and common-law legal tradition is coded as 1. In our sample 68 countries had common-law and 28 had civil-law legal system.

For an accurate inference about the independence of a judicial system, one should also take into consideration the institutional guarantees of the judiciary stated in legal texts. In the literature, these guarantees are designated as de jure judicial independence and generally comprise issues such as tenure of a judge, nomination process, and salary protections (Feld & Voigt 2003). The degree of institutional protections in a given country is assessed by looking at its constitution. I formulate de jure judicial independence as an additive index by summing up three variables from the Comparative Constitutions Project (Elkins, Ginsburg, & Melton 2007) that provides information on the characteristics of written constitutions for 192 countries. These variables are judicial independence,¹³ judicial review¹⁴ and protection of judicial salary.¹⁵ The additive index is normalized between 0 and 1. Thus, the countries with maximum de jure judicial independence refer to higher levels of constitutional guarantees of the judiciary. Finally, to control for the arguments stating that judicial independence and political competition are merely the result of having more financial resources, I use the natural log of the average per capita gross domestic product for the period 2000–2008. I use the data from the World Bank's World Development Indicators (2008-2008). In our sample the ln GDP per capita (corrected for purchasing power parity) has a minimum value of 6.48 (Niger) and maximum value of 11.24 (Luxemburg) and a standard deviation of 1.22.

To correct for potential omitted variable bias, all these control variables are included in the empirical model. If the relationship between de facto judicial independence and political competition is merely the result of economic development; having a presidential system, common-law legal tradition, and strong constitutional protections for the judiciary then controlling for these variables should eliminate the statistical significance of political competition on judicial independence.

Empirical Analysis and Results

Although the available data for judicial independence has longitudinal character, our key independent variables are time invariant (e.g., level of democracy) and rarely changing (e.g., political competition changes with national elections). The lack of variation in the level of democracy variable may not constitute a big problem since that variable is expected to condition the effect of political competitiveness. Yet the bigger problem is the lack of variation in the political competition variable. For this reason conducting panel data analysis is not appropriate. Using ordinary least squares (OLS) regression, as is appropriate for continuous dependent variable (Wooldridge 2002), I aim to test whether the impact of political competition on judicial independence changes across countries with different democratization levels. Yet one should bear in mind that, since the empirical analysis does not directly take into consideration the changes to judicial independence that follow from changes in political competition, the analysis provides an indirect test of the mechanism that is proposed in the theoretical section.

I estimate two interaction models that would help to delineate whether and to what extent the impact of political competition on judicial independence changes across advanced and developing democracies. In these models the dependent variable is de facto judicial independence score. In the first model, I include only the key variables (political competition; democracy level and political competition interacted with democracy level). In the second model, I control for the legal system, executive-legislative arrangement, de jure judicial independence and the GDP per capita.

Both models presented in Table 1 show that the impact of political competition on de facto judicial independence reveals a significant variation across advanced and developing democracies. The coefficient on the interaction term is positive and statistically significant in both models, suggesting that the beneficial effect of political competition on judicial independence is greater for countries with very high levels of democratization compared to the countries with very low levels of democratization. The negative estimate of political competition shows that when the democratization level is “0” the net impact of political competition on judicial independence is significantly negative, whereas when the democratization level is “1” the impact of political competition is significantly positive.

Table 1. OLS regression on judicial independence index

Note: *** <0.001, **<0.05, *<0.10.

Nevertheless, it is impossible to evaluate the statistical significance of the political competition effect simply from reviewing the coefficients and their standard errors presented in Table 1 (Brambor, Clark, & Golder 2006). The marginal effect of political competition may not have a statistically significant impact on judicial independence over the entire range of democracy levels. In other words, the hypothesis tests summarized in Table 1 are useful and interpretable only when the level of democracy is equal to “ ‘0' ” or “1.” In order to better evaluate the statistical and substantive effects of political competition, we require additional analysis. Figure 1 shows the predicted level of judicial independence across the range of political competition for two hypothetical democratic countries. Estimates for the hypothetical developing democracy whose level of democratization is set to 0.2 (based on the recalculated FH index that ranges between 0 and 1) are represented on a solid line. On the other hand, the estimates for the hypothetical advanced democracy whose level of democratization is set to 0.9 are represented on a dashed line. Except their democratization levels, both countries share the same legal system and executive-legislative arrangement as well as the same level of de jure judicial independence and GDP per capita.¹⁶

Figure 1. Predicted Level of Judicial Independence. Note: Shows predicted level of judicial independence across the range of political competition.

Note that the solid line slopes downward and reflects the negative effect of political competition, whereas the dashed line slopes upward and reflects the positive of political competition. Accordingly, for our hypothetical developing democracy, the predicted level of judicial independence drops from approximately 6 points to 2 points across the range of political competition. On the other hand, for our hypothetical advanced democracy the predicted level of judicial independence increases from approximately 4 points to 7 points. The distance between the solid line and the dashed line is precisely the effect the political competition on judicial independence, conditional on a democratization level. These estimates are consistent with our argument suggesting that while in advanced democracies the political competition has positive impact on judicial independence, in developing democracies it hampers independence of the courts. Yet we do not know whether the differences are statistically distinguishable from zero.

To better convey how the effect of political competition on judicial independence changes across countries with different democratization levels, Figure 2 shows the marginal effect of political competition, with confidence intervals around the estimated effect. The graph is created by using the estimates from the interaction model and the full sample of democratic states. The black line in the figure represents the change in judicial independence associated with flipping the political competition variable from 0 to 1. The dashed lines surrounding the solid black line indicate the 95 percent confidence interval around that change. Figure 2 shows that the marginal effect of political competition on judicial independence does considerably vary across advanced and developing democracies. Indeed the effect of political competition is positive and significant for countries with relatively high democratization levels. It becomes statistically insignificant as democratization level decreases but for countries with very low levels of democratization it becomes statistically significant and negative.

Figure 2. Marginal Effect of Political Competition on Judicial Independence. Note: Displays marginal effect of political competition on judicial independence for different values of level of democracy measure.

The marginal effect of political competition on judicial independence appears to be significantly positive in countries with high democratization levels (between 0.8 and 1), whereas in countries whose recalculated FH index ranges between 0 and 0.3, the effect of political competition on judicial independence is significantly negative. These empirical results show that in countries with very low levels of democratization, political competition has significantly negative impact on judicial independence.

Contrary to the insurance theory that envisions long-term benefits to incumbents in creating independent courts, the intuition behind this empirical finding is that, in developing democracies under intensified competitive political environment, power holders might have high short-term benefits and low costs in creating politically dependent and subservient courts. In line with these empirical results, comparing the Russian and Ukrainian courts during the 2002–2003 parliamentary campaigns, Popova (2010) observes that the Russian politicians had a weaker incentive to interfere in judicial decisions compared to the Ukranian politicians. Indicating that the Ukranian elections were more competitive than the Russian elections,¹⁷ the author presents the varying levels of political competition across these two countries as the key determinant of the varying levels of judicial independence. As a result, Popova (2010) states that in developing democracies, political competition (a) increases the benefits to incumbents of dependent courts, (b) fails to increase the costs of exerting pressure on the courts, and (c) increases the number of court cases whose outcomes matter to incumbents. The incumbents who face stronger competition and a higher probability of losing the next election are claimed to be more likely to try to control and constrain the judiciary.

Regarding the explanatory power of control variables, Table 1 demonstrates that all else equal, the judicial independence in democratic countries with common-law legal system is substantially higher compared to the judicial independence levels of democracies with civil-law legal origin. This finding is in line with La Porta et al. (1998)'s study where they observe that judicial independence is empirically strongly associated with common-law legal origin.

In parliamentary democracies, the executive is an agent of parliament, accountable to the majority of the parliament and subject to being ousted at any time by vote of no confidence. Yet, in presidential democracies legislature and executive are agents of voters and so not accountable to one another (Shugart & Carey 1992). In line with these characteristics of the parliamentary systems, the empirical results suggest that parliamentary systems are inclined to have higher independent judiciaries compared to presidential systems. As regards the impact of political regime on judicial independence, Table 1 shows that the independence of the judiciary tends to be significantly higher in parliamentary democracies compared to the independence levels of the courts in presidential regimes.

Moreover, Table 1 specifies that the constitutional protections of the judiciary do not significantly explain a country's actual judicial independence level. Thus, based on our empirical findings one can argue that if a country would include in its constitution some legal reservations regarding the independence of the judiciary, such institutional protections would not guarantee higher levels of judicial independence. This finding is in line with the arguments stated by the scholars who acknowledge that actual practices do not always follow the rules (e.g., Chavez 2004). Thus, on paper judges may enjoy formal guarantees of tenure or salary but in practice they may face removal by the incumbent politicians before the end of their terms (Feld & Voigt 2003). Finally, Table 1 shows that, all other variables held constant, the countries with higher levels of GDP per capita have significantly higher levels of judicial independence.

Sensitivity Tests and Empirical Challenges

Conceptually and methodologically this article distinguishes the level of political competition in a country from that country's level of democracy. In other words, I do not undertake high levels of political competition as a characteristic that could be ascribed only to advanced democracies. Nonetheless, one may still intuitively argue that political competition is closely related to the level of democracy. Thus it is important to show to what extent the measures of these two concepts are correlated with each other. When I check for the correlation coefficient¹⁸ between these two measures, I find that the sample correlation between political competition and the level of democracy is Corr (X, Y) = +0.46. This value of the correlation does not indicate a strong relationship between political competition and the level of democracy measures used in the analysis. Recalling the theoretical framework of the empirical analysis, one should also remember that by the political competition measure my purpose is to infer the political uncertainty that refers to the chance of the incumbent party to capture seats in the incoming government. In this regard, I believe that it is far from obvious why we should expect a strong connection between political uncertainty and the level of democracy.

The fact that the key variables are measured by reputational indices means that sensitivity testing is particularly important to see whether the results are robust to different operationalizations of the main variables. As a sensitivity test, I conduct the same empirical analysis by using different measures of the key independent variables. For instance, as an alternative measure of political competition I use the political competition indicator used in the Polity dataset (Marshall, Jaggers, & Gurr 2010). The original values for this measure range between “0” and “10”. This measure is recalculated so as to vary between “0” and “1” so that “1” refers to the highest and “0” refers to the lowest level of political competition. On the other hand, as an alternative measure of democracy level, I use the World Bank Governance Indicator of Accountability and Voice (Kaufmann, Kraay, & Mastruzzi 2010). This may also be a good proxy for democracy level because it aggregates the results of numerous expert reputational surveys and captures the perceptions of the extent to which a country's citizens are able to participate in selecting their government, as well as freedom of expression, freedom of association, and a free media. In the original dataset the Accountability and Voice indicator ranges between −2.5 and +2.5. The measure is recalculated so as to change between “0” and “1” so that “1” refers to the highest level of democracy. Including the usual controls the empirical models are re-tested by using these alternative measures of political competition and democracy level. The empirical findings (see Appendix A) show that the coefficients of the key variables remain statistically significant and maintain the same sign even after these alternative measurement choices. All these results strongly support the hypothesis that political competition has varying impact on judicial independence across advanced and developing democracies. But it is important to underline two important caveats. First, the dependent variable of this study—de facto judicial independence—is derived from subjective evaluations and if these evaluations are correlated with other aspects of a country, the results might be biased. Yet, to my knowledge there is no better data that provides a direct and objective measure of de facto judicial independence in a cross-country setting.

Second, the key independent variables of the empirical model—political competition and the level of democracy—might be endogenous to judicial independence. Because of the absence of reliable instrumental variables (which have to be correlated with political competition and the level of democracy respectively but not judicial independence) I cannot rule out the possibility of reverse causation. Nevertheless I believe that it is important to consider the theoretical basis for the concern of endogeneity.

One might argue that in the countries where the judiciary is insulated from any external interference, the level of democracy would be higher (e.g., higher protection of freedom of expression and individual rights).¹⁹ But one should recall that I am not looking to estimate the impact of the level of democracy on judicial independence. Instead, my purpose is to measure the impact of political competition on judicial independence which is conditioned by the level of democracy. Thus the issue that could be problematic is the reverse causality effect of judicial independence on political competition. However it is far from obvious why the impact of judicial independence on political competition would be stronger in countries with high democratization levels. If judicial independence increases the level of political competition for whatever reason, one would expect that it would do so similarly in advanced and developing democracies. Yet I am aware of no theoretical model that explains how this could be so. Thus despite all these caveats of the study, the consistency of the empirical results with the predictions of the model increases our confidence that the model is accurate.

Concluding Remarks

Assuming that putting pressure on the judiciary is very costly, the proponents of the insurance theory present judicial independence as an efficient insurance mechanism. They argue that in democratic countries electoral competition functions as a sanctioning device through which the electorate is able to replace politicians who intervene in the judiciary. Thus, attributing high levels of judicial independence to intense political competition, the advocates of this theory appear to envision a positive relationship between these two aspects. This article, however, empirically shows that as democracy level across countries changes the impact of political competition on judicial independence changes as well. While in advanced democracies high levels of political competition enhance judicial independence, in democracies with very low democratization levels, political competition significantly hampers independence of the courts. In line with these empirical findings, the positive relationship between judicial independence and political competition that is suggested by the insurance theorists appears not to grasp the situation in developing democracies. At this point assuming that political competition has similar impacts on judicial independence across all types of democracies would be quite misleading.

This article proposes that the cost-benefit analysis that the rational political elites have to undertake while choosing their judicial policies would reflect different trends across advanced and developing democracies. For that reason high levels of political competition would not inevitably lead to high levels of judicial independence across all democratic countries. In the context of advanced democracies, for instance, the cost of meddling with judicial decision-making practices would be quite costly for the incumbent politicians. In advanced democracies free and independent media provides considerable transparency so that the public becomes aware of the wrongdoings of the government. In these types of regimes people also have high levels of confidence in the judicial system. Moreover they are highly interested in politics, have higher education levels and tend to actively participate in political or civic activities. Because of all these aspects, in an advanced democracy if an incumbent government decides to put pressure on the judiciary and create subservient courts, it will probably face fierce public backlash. In other words, the cost of putting pressure on the judiciary or meddling with its decision-making practices would be quite costly for power holders in these types of regimes. Thus, the incumbent politicians who face the challenge of being ousted from power would prefer to secure their long-term interests under the insurance of an independent judiciary.

In developing democracies, however, this article suggests that the costs of pressuring the courts are lower but the benefits of pressure are higher. In these types of regimes as political competition increases, the incumbent governments tend to manipulate the judiciary and use it as a mechanism for re-election or protection of power. In other words political competition appears to hamper the independent performance of the courts. This situation reveals an important dilemma for the democratic consolidation efforts in the developing democracies. As long as the public does not start to act as a credible control and constrain mechanism in these countries, this dilemma would stay irreconcilable and we should not wonder why the majority of developing democracies struggle with the rule of law despite electoral competition.