Judicialization of Politics by Elected Politicians

This study argues that the proximity to a general election would affect the frequency of the opposition parties’ referrals to the constitutional court. This effect is hypothesized to be conditioned on the opposition parties’ prediction of the upcoming election results. To test this theory, I constructed an original data set including all acts promulgated by Turkish Parliament and all cases that were brought to the constitutional court by the opposition parties during 1984–2011. The results show that once the opposition party believes that it will lose the election, it increases its referrals to the court as election approaches.


Introduction
The formation of higher courts with extensive jurisdiction led to the expansion of judicial power in many countries. By reviewing laws and decrees under the rubric of constitutionality, constitutional courts began to serve as a control and constraint mechanism against the other branches of government. Accordingly, it is stated that courts "have developed into powerful institutional actors or policymakers" (Shapiro and Stone-Sweet 1994, 401) who "direct the making of public policies" (Tate 1995, 28). This increasing involvement of the courts in addressing and resolving public policy questions and political controversies is often referred to as the judicialization of politics (Hirschl 2006).
Arguing that the judicialization of politics does not derive from a single cause, many scholars have tried to account for the institutional and political features that are necessary to create and sustain this phenomenon (e.g., Gauri and Brinks 2008;Ginsburg 2003;Hirschl 2006;Stone-Sweet 2000;Tate and Vallinder 1995). Despite their differences, all of these scholars suggest that, at the bare minimum, the judicialization of politics requires a relatively independent judiciary. This proposition implies that if judges can make decisions independent of any external factors, individuals or political oppositions will be more likely to bring their policy claims against the government to the court. This does, however, raise the question of why, in developing democracies where the judiciary is not fully independent, political opposition parties frequently bring public policy questions before the constitutional court and thus judicialize politics.
Are the political opposition parties willing to refer a bill to the court to increase their public vote share or secure more seats in Parliament? Or perhaps they act mainly to further public awareness, foster public discussion, and create pressure for social and legislative change? Or, perhaps by referring bills for abstract review, the political elites of the opposition simply want to place obstacles in the way of an incumbent government's ability to function. Although an opposition party might be motivated by a number of different political objectives, it will not always have the necessary incentives to systematically refer every major piece of legislation for abstract judicial review. For this reason, it is worth exploring under which conditions the opposition parties more frequently turn to the constitutional court.
This research suggests that opposition parties are rational actors who tend to use litigation as a tool to differentiate their policy preferences from the policies enacted by the incumbent government. Accordingly, it is hypothesized that the temporal proximity of a general election and the opposition parties' expectation about the election results will affect their referrals to the constitutional court. Specifically, once the opposition party believes that it will lose the upcoming election, it will begin to increase its referrals to the constitutional court to highlight the policy differences on a constitutional level, which in turn may increase public awareness and its vote share. However, once the opposition party foresees itself winning the upcoming election, it will use litigation less frequently as the costs of pursuing litigation will outweigh its marginal benefits. For the remainder of this paper, this phenomenon will be referred to as "strategic litigation theory." Turkey presents an ideal case for developing and testing this strategic litigation theory. Similar to the other European countries, 1 the Turkish judicial system enables the political opposition parties to challenge the constitutionality of legislation and public policies before the constitutional court. Despite some flaws in its judicial system, 2 elected Turkish representatives have frequently resorted to litigation as a part of their political activity over the last three decades. Nevertheless, within legislative periods, there are some interesting variations in the litigation practices of the opposition parties that need to be explained. To test the strategic litigation theory in the Turkish context, I constructed an original data set, including all acts promulgated by the Turkish Parliament between the years 1984 and 2011, and analyzed all cases that were brought to the constitutional court by the main opposition parties during the same time period.
Building an original data set to explore the conditions that account for the variation in the opposition parties' referrals to the constitutional court in Turkey is of crucial importance for understanding the conditions that trigger the judicialization of politics in other developing democracies where the constitutional courts exercise abstract review. This study not only speaks to broader debates surrounding the judicialization of politics in developing democracies but also seeks to integrate the Turkish Constitutional Court-which remains understudied at this point-into the large body of existing research on comparative political institutions and judicial politics.
The essay proceeds as follows: The first part introduces the literature on the judicialization of politics and constitutional review in developing democracies. The second part introduces the theoretical framework about the use of strategic litigation by the political opposition parties. The third part gives information about the Turkish judicial and political system, as well as the institutional design of judicial review in Turkey. The fourth part introduces the data, empirical model, and key variables. The fifth part is the empirical section where the main hypotheses are tested and the results are presented. The last part will conclude the study by discussing both the theoretical and practical implications of the findings.

The Politics of Judicialization: Theories and Puzzles
The initial studies on the judicialization of politics have defined the judicialization phenomenon as a transfer of decision-making rights from the legislative and executive bodies to the courts (Tate and Vallinder 1995). In this regard, the judicialization of politics refers to the processes by which "courts and judges come to make or increasingly dominate the making of public policies that had previously been made by other governmental agencies" (Tate 1995, 28). In all of these conceptions of judicialization, the courts are presented as decision-making bodies that make policies or shape the political landscape. However, the judicialization of politics is not a mere displacement of policy decision-making rights. It is a dynamic process that is based on the strategic interaction between the judiciary and other political actors (Gauri and Brinks 2008;Hirschl 2006). Moreover, the judicialization of politics does not solely depend on courts and their final decisions. It is a continuous process that consists of different stages, 3 and each stage involves a choice made by one or more strategic actors (Gauri and Brinks 2008).
To understand what triggers the judicialization process in the first place, one should note that cases do not come to the courts at random; rather, that litigants bring them to the courts. In this regard, litigation that can be conceptualized as placing the cases on the court's docket is the initial stage of the judicialization process, and as potential litigants, opposition parties play an important role in this process. Focusing on the litigation aspect of the judicialization process, this study aims to account for why and under which conditions the political elites who are part of the policy-making process would choose to go to the court instead of resolving those policy issues through traditional political forums. When a political opposition party decides to refer a bill to the court for its annulment, is it because the opposition party truly disagrees with the legal standing of the bill? Or is it for some other reason?
The conventional response states that the "losers" of a policy-making process decide to turn to the courts because they are unable to achieve their policy goals through political means (Stone-Sweet 2000). The recent theories about the empowerment of courts and the institutional design of judicial review also assume that political losers turn to the courts to protect their policy preferences. Ginsburg (2003), for instance, argues that political elites who are confronted with electoral uncertainty and foresee themselves being ousted from power will favor an independent judiciary as an insurance mechanism. According to this insurance logic, the ruling majority, who expect to lose the elections, believe that independent courts would protect their rights and policies once they become political minorities (Ginsburg 2003;Landes and Posner 1975). Hirschl (2004) proposed a similar theory of judicialization that he calls "hegemonic preservation." He argues that the establishment of a judiciary with strong judicial review power is a hegemony-preserving strategy adopted by political elites who foresee losing their grip on political power. Using the judicial review mechanism, they will be able to transfer the contentious issues to the court when their worldviews and policy preferences are increasingly challenged in majoritarian decision-making arenas (Hirschl 2004).
At first glance, the loser argument and insurance theory seem quite appropriate because they fit very well within the traditional view of the courts as defenders of minorities against the potential abuses of the incumbent government (Ryan 2010). However, they pose certain challenges in explaining what triggers the judicialization of politics in the first place. First, the insurance theory assumes that it is only through an independent judiciary that the Parliamentary minorities or political opposition will be able to protect their preferences, rights, and interests. Based on this assumption, one might argue that the opposition parties would decide to bring a contentious policy issue before the higher court when the judiciary is relatively independent. However, if that is the case, then why would the opposition parties in developing democracies, where the judiciary is not fully independent, choose to go to the court and judicialize politics?
One should note that democratic governance should not be considered as a prerequisite for the "judicialization of politics." Focusing on Egypt's authoritarian system, for instance, Moustafa (2003) argues that in authoritarian systems, the judiciary enjoys an expanded political role (e.g., protecting property rights and encouraging investment). Even in these types of regimes, opposition activists and human rights groups might use the higher court as a tool to challenge the regime on political issues. Moreover, in developing democracies where the judiciary is not fully independent, the courts might still rule against the incumbent government and protect the policy preferences of the opposition parties. Many studies showed that when the incumbent government is close to losing power (Helmke 2005;Iaryczower, Spiller, and Tomassi 2002) or when political power is highly fragmented (Chavez 2004;Ríos-Figueroa 2006;Scribner 2003), the judges are more willing to decide against government decrees. These studies seem to conceptualize the judicialization of politics as the judicial power to give anti-government decisions. However, it is important to note that the political conditions presented by these studies are just the enabling conditions for the judicialization of politics and provide an opportunity for taking policy claims to the courts. They do not account for why the political elites in the opposition would choose to go to the court instead of addressing the problematic legislation through traditional political forums, even when the judiciary is not independent.
Second, the "loser" argument and insurance theory imply that the opposition parties would choose to bring a contentious policy issue before the higher court whenever the governmental policies deviate from their preferences. In other words, if the opposition parties' interests and preferences were in line with the incumbent government's policy, then there would be no reason to challenge the constitutionality of that policy in front of the court. Both of these assumptions take the political litigants as naive policy-seekers whose decision to litigate is only affected by the distance between the governmental acts and their ideal policies (Magalhaes 2003). Assuming that the opposition parties' litigation is preference-based, these theories envision that the only concern of the main opposition party is to win the case in court.
Nevertheless, the opposition parties may go to the court not because they truly disagree with the legal standing of the bill but because they want to use litigation as a strategic tool to achieve certain political objectives. Focusing on the Israeli case, Dotan and Hofnung (2005) present the enhancement of political stature via media exposure as an important motivation for political litigation. According to these scholars, even if the opposition parties who go to court lose the case, they are still likely to enjoy greater coverage than if they had won by other means (Dotan and Hofnung 2005). Similarly, McCann (1994) argues that generating publicity can be enough for motivating the litigants. As a result, by contesting a policy in the court, it might be possible to secure a political victory without achieving a legal victory.
A number of scholars also suggest that the minority parties in the Parliament use litigation to obstruct the actions of the incumbent government (Stone-Sweet 2000; Tate and Vallinder 1995). In the context of developing democracies, due to high levels of mistrust between the political actors, the inefficiency of the institutional veto players, and the indifference to the preferences of opposition parties, the opposition parties themselves tend to adopt a more obstructionist pattern of behavior and oppose the proposals coming from the government more frequently (Mujica and Sanchez-Cuenca 2006). Nevertheless, the opposition parties' access to the courts or their desire to gain certain benefits from litigation does not mean that they will always have the necessary incentives to systematically refer every major piece of legislation for abstract judicial review. The existing literature fails to provide a systematic explanation of the variation in the frequency of the opposition parties' referrals to the court at certain time periods. By focusing on the choices that opposition parties make and the challenges they face, the next section of this article tries to answer why and when do the opposition political parties more frequently bring cases to the constitutional court and choose to judicialize politics.

The Logic and Theory of Strategic Litigation
So far, we have laid the groundwork for identifying litigation as a distinct form of behavior that a political opposition party adopts when faced with a problematic bill in the Parliament. The central argument of this study is that, as election time approaches, the opposition parties become more willing to go to court. However, the effect of the approaching election is hypothesized to be contingent on the opposition party's prediction of its post-election placing. I introduce this aspect as "strategic litigation." Since strategic litigation only makes sense if the opposition parties believe that litigation will improve their fate, the next step is to develop a unified theoretical framework.
Illustrating the logic of strategic litigation, Figure 1 portrays a sequence of actions where the majority in the Parliament (P1) moves first, the main political opposition party 4 (OP1) second, the court (C) third, the nature (N) fourth, the future majority in the Parliament (P2) fifth, and the future main opposition (OP2) party sixth.
In the first stage, the majority in the Parliament acts by promulgating a law. In the second stage, the main opposition party in the Parliament decides whether or not to refer that law to the constitutional court. In the third stage, the constitutional court decides whether or not to annul the law. In the fourth stage, Nature selects the incoming government. The main opposition party will be part of the incoming government, with a probability p. It will either hold a majority of the seats in Parliament and act as a single-party government or form a coalition government with another political party. With a probability 1 − p, it will be acting as the main opposition party, as a small opposition party, or it will be totally ousted from the Parliament. In the final stages, the future majority in the Parliament tries to promulgate another set of laws, and the main opposition party decides whether or not to refer that law to the constitutional court.
An implicit assumption underpinning the central argument of this study is that the opposition parties might not care much about the final decision of the judges when deciding whether to refer a statute to the court. By contesting a policy in the court, it may be possible to secure a political victory without achieving a legal victory. Another simplified assumption of the strategic litigation theory is that an opposition party has complete and perfect information about the election results. Finally, the politicians' actions and their policy-making considerations are usually affected by the upcoming elections, meaning that they tend to make decisions that will positively affect the election results. Based on these assumptions, I present strategic litigation as a logical response used by the political opposition parties to achieve their objectives.
A closer analysis of the opposition parties' cost-benefit calculations in deciding whether to bring a public policy before the constitutional court would provide a theoretical explanation about when and why the opposition parties would more frequently go to the court and judicialize politics. By bringing legislative bills before the constitutional court, the opposition parties are not only able to challenge their constitutionality but also to reap certain benefits from these referrals. First, judicializing a bill makes it easier to highlight the policy differences between the opposition party and the incumbent government. For instance, an opposition party can convert a problematic piece of legislation into a public debate in which there is only one "good" side to be advocated. Freedom and human rights, democracy, political corruption, economic growth, or social justice are the issues that can be used in this sense (Stokes 1963). By accusing the ruling majority of violating constitutional rules and referring legislation for abstract review, the opposition parties can attempt to place themselves on the "right" side (and the majorities on the "wrong" one) of a debate regarding these issues (Magalhaes 2003). Second, voters are likely to punish a party at the polls if that party moves too far away from its policy preferences to compromise with the government (Vanberg 1998). Then, the opposition parties' frequent use of litigation will show that they do not compromise against the majority's policies. Third, the conventional wisdom as well as the literature posits that the electorate rewards the incumbent government if its performance is good, but the incumbent government is punished through reduced support if its performance is poor (Sanders promulgates a law litigates or not decides determines next promulgates a law litigates or not the law to the court for/against majority in the law to the court the parliament Figure 1. Strategic litigation game. 1996). Thus, obstructing the policies and actions of the incumbent government through litigation may be electorally advantageous for the opposition parties (Magalhaes 2003). As a result, by using litigation as a vote-seeking strategy, the opposition party may increase public awareness, foster public discussion, bring pressure for social and legislative change, cast doubt on governmental policies, or discredit them entirely. Although the opposition parties may gain certain benefits from going to the court, litigation is not cost-free. While an opposition party tries to decide whether to bring a public policy before the constitutional court, it should also take into consideration the corresponding costs. First, making a referral requires drafting a petition, which means using the opposition's legal experts' resources. Second, opposition parties are hardly perfectly unitary actors, and making a referral often requires overcoming some internal disagreements over which battles are worth fighting. In other words, not all opposition MPs are likely to agree on which bills are worth challenging and which are not. Thus, there might be some coordination costs involved in making referrals. Third, the opposition party might appear obstructionist due to making frequent challenges that might have negative consequences on the election results. Finally, there is a potential reputational cost to making frequent, frivolous referrals that are often rejected by the constitutional court. An opposition party that frequently loses in the court may look incompetent or weak.
The central argument of this study suggests that the effect of the approaching election on the main opposition party's use of litigation is conditional on its prediction about the general election results. This is because the cost-benefit calculations of an opposition party that expects to be part of the incoming government will be quite different from the calculations of an opposition party that expects to lose the general election and continue to act as an opposition party. If the main opposition party expects to lose the upcoming election, it either foresees that it will be acting as an opposition party in the post-election period or that it will be totally ousted from Parliament. Thus, as the election approaches, the main opposition party will try to use all available means to increase its vote share or protect its place in the Parliament. Due to the related costs, the main opposition party will not bring each and every piece of legislation to the constitutional court, but the benefits that it will reap from going to the court will be higher than the costs. By increasing its referrals to the court, the main opposition party shows that it does not compromise against the majority's policies and highlights the difference between its policy preferences and the policies enacted by the incumbent government. Accordingly, the first prediction is: Hypothesis 1: The main opposition party will increase its use of litigation as election time approaches once the prospect emerges that it will lose the general election and would therefore act as the main or a small opposition party in the new Parliament.
On the other hand, if the main opposition party foresees itself winning the upcoming election, holding a majority of the seats in Parliament or becoming part of a coalition government, it will be less likely to use litigation as a vote-seeking mechanism. The reason behind this is that when the main opposition party is likely to win power, the costs of pursuing litigation will outweigh its marginal benefits. When the main opposition party envisions itself winning the upcoming election, it may decide to change the problematic bills itself, rather than losing time in the litigation process and waiting for the decisions of the court. In this regard, the second prediction is: Hypothesis 2: The main opposition party will not increase its use of litigation as election time approaches once the prospect emerges that it will have an electoral victory and either hold the majority of seats in the upcoming Parliament or become part of a coalition government.
In line with these hypotheses, the strategic litigation theory suggests that the changes in the opposition party's litigation behavior depend on the changes in the main opposition party's beliefs and expectations about the results of the upcoming general election. For that reason, evidence would be inconsistent with these hypotheses if the main opposition parties that were expecting to lose the upcoming general election did not significantly increase their use of litigation or if the main opposition parties that were expecting to win the upcoming general election significantly increased their referrals to the constitutional court as election time approached.

The Political and Judicial Context in Turkey
Before moving from hypothetical arguments to analyzing real-world events, it is important to understand the context in which the opposition parties decide to go to the court. In this case, it is crucial to explain whether all opposition parties in Turkey have faced a similar strategic environment. Accordingly, this part of the study draws on the Turkish political context, as well as the institutional design and behavior of the Turkish Constitutional Court (TCC).
Although smaller parties have been excluded from referring bills and decrees for abstract review, the largest parties in the opposition have always enjoyed the ability to litigate against laws passed by the Parliamentary majorities. 5 This is closely related to the centralization of the legislative process that is shaped by the number of veto players in the system. As in other European Parliamentary democracies, the stages of the legislative process in Turkey are relatively free from veto powers. Since Turkey is a unicameral system, it lacks a second chamber that might work as an institutional veto player in policy-making. This leaves the president of the Republic as the only potential veto power to consider. The president of the Turkish Republic has the authority to return any law to the Parliament for review before its promulgation. That being said, this is not a very strong veto power because if the Parliament enacts the same law without amendment, then the president must promulgate it. Since the unicameral Parliamentary system in Turkey allows the incumbent political party to dominate the executive and legislative branches, constitutional review remains the only veto point available to opposition parties in the Turkish Parliament (Hazama 2012).
Another political factor that leads to high levels of referrals by the opposition parties in Turkey is the lack of trust between the political parties. As a result of the strained relationship between the government and opposition parties, the governmental parties do not take the views of the opposition parties into account and this extremely impedes their ability to cooperate, even on matters of mutual interest (Turan 2003). Moreover, the confrontational nature of the government-opposition relations and the elected politicians' perception of politics in zero-sum terms prevent the development of culture of compromise in the Turkish Parliament (Kalaycıoğlu 1990). All of these political aspects provide an enabling framework for the opposition parties' use of litigation as a strategic tool. 6 On the contrary, to better evaluate the judicial context under which the opposition parties decide to go to court, first we have to look at the legal framework of the litigation process. The constitutionality of laws, decrees, and the rules of procedure of the Turkish Parliament are challenged directly before the TCC through an annulment action. Since the action for annulment is abstracted from any particular case, matters falling under these items can be brought before the court for abstract review. Abstract review procedures in Turkey are a posteriori. This means that the legislation in question must first be implemented before its constitutionality can be challenged. The right to apply for annulment directly to the constitutional court lapses sixty days after the contested law, decree, or rule of procedure has been published in the Official Gazette (The Constitution of the Republic of Turkey, Article 151). Furthermore, no allegation of unconstitutionality shall be made with regard to the same legal provision until ten years after its publication in the Official Gazette of the decision of the court dismissing the application on its merits (The Constitution of the Republic of Turkey, Article 152).
Another element of the judiciary that might affect the opposition leaders' decisions to go to the court is the likely outcomes of the judicial decisions. When there is a considerable ideological difference between the opposition party and the court, which suggests a high probability that the court will reject all of its referrals, then the opposition party might refrain from frequently filing challenges. In this regard, it is important to underline to what extent all opposition parties in Turkey face a similar judicial environment. First, the 1982 Constitution assigns no role to the Parliament in appointing the constitutional court judges. The president of the Turkish Republic, who should be politically independent from any political party, is the only person who appoints all judges. Second, neither the Parliament nor the incumbent government plays a role in the impeachment of the members of the TCC and no member of the court has been impeached since 1962 (Shambayati 2004). Third, the members of the court have very similar educational backgrounds and life experiences (Shambayati 2004). All of these aspects reveal that the Turkish judges are de jure insulated from any political influence, and in terms of its ideological preferences, the court has a homogeneous structure that does not change across time.
Although the TCC has a relatively apolitical structure on paper, it is well established that the constitutional court judges are administrative agents who defend the values and interests of the state elites (Belge 2006;Shambayati and Kirdiş 2009). If this is the case, then the court should favor the state elite opposition parties over non-state elite opposition parties. 7 However, in his empirical analysis of the TCC decisions, Hazama (2012, 435) shows that the court does not respond differently to unconstitutionality claims by the state elite and non-state elite opposition parties. In other words, the probability of the court's decision to nullify the objection is not significantly affected by an opposition party's state elite versus non-state elite status. More importantly, Hazama (2012, 432) finds that between state elite and non-state elite main opposition parties, the mean frequency of referrals per year was not significantly different. Given all of these factors, we can say that all opposition parties face a similar strategic environment in terms of how readily the court is likely to agree with them.

Data, Measurement, and Model
The aim of this study is to analyze whether the willingness of the political opposition parties to go to the constitutional court changes in accordance with their expectations about the upcoming general election results. In other words, does an opposition party who expects to win or lose the general election decrease or increase its referrals to the court, respectively?

Data
As to the actual use of litigation by the opposition parties, Figure 2 shows that there is a considerable variation in the frequency of litigation use over the years and across different governmental periods. Showing trends in the laws sent by the main opposition parties to the TCC by year, Figure 2 clearly fits with the patterns predicted by the hypotheses of the current study.
To contextualize the findings presented in Figure 2, one should also take into consideration the incumbent governments and the main opposition parties in Turkey between the years 1984 and 2011 (see Table 1). Accordingly, we see that under the first ANAP government, between the years 1984 and 1987, the main opposition party, the SHP, sent an average of 6 percent of the laws to the constitutional court. However, one year prior to the election, the opposition party increased the number of laws referred to the court up to 10 percent. Similarly, while under the first and the second AKP governments, the main opposition party sent an average of 12 and 15 percent of laws to the constitutional court; one year prior to the election, the percentages of laws sent to the court by the main opposition party, the CHP, peaked at 15 and 16 percent, respectively. This trend underscores the intuition that once the main opposition party becomes certain that it will lose the upcoming election, it goes more frequently to the court as the election approaches.
In contrast, under the second ANAP government (1987)(1988)(1989)(1990)(1991) and the DYP-SHP government (1991)(1992)(1993)(1994)(1995), there was no increase in the number of laws sent to the constitutional court by the main opposition parties at that time. Instead, consistent with the second hypothesis, the trend in the laws sent to the court shows that as it became more likely that the main opposition party would be part of the incoming government, the percentage of the laws sent to the court declined or stayed the same as the election approached. For instance, under the second ANAP government, the main opposition party, the SHP, sent an average of 6 percent of the laws to the constitutional court. Yet one year prior to the election, in 1991, it sent only 3 percent of the laws to the court. On the other hand, under the DYP-SHP coalition government, the main opposition party, the ANAP, sent an average of 10 percent of the laws to the court, whereas during the final year of the outgoing government, it sent only 6 percent of the laws to the court. This trend underscores the intuition that once the main opposition party becomes certain that it will win the upcoming election it decreases its number of referrals to the court as the election approaches.
To better test the strategic litigation theory, we need a political framework where the life of an opposition party shows some continuity. In the Turkish case, however, the periods of 1995-1999 and 1999-2002 do not provide an appropriate political framework to test the strategic litigation theory, since one year prior to the general elections held in 1999 and 2002, the main opposition parties (the RP and the FP) were legally banned from politics and were closed down by the TCC. For this reason, during the 1995-1999 and 1999-2002 periods, Turkey did not have a single main opposition party that could formulate a strategic policy for the upcoming elections.
In this regard, the data set includes laws and decreelaws passed by the Parliament during the following five governmental periods: the first ANAP government (January 1984-November 1987, the second ANAP government (November 1987-October 1991, the DYP-SHP government (October 1991-December 1995, the first AKP government (November 2002-July 2007, and the second AKP government (July 2007-June 2011). The starting and ending dates of each governmental period refer to a general election. The total number of laws and decree-laws passed by the Parliament during these five governmental periods is 2,933. Each law and decree-law is coded dichotomously according to whether or not the main opposition party sent the regarding law to the TCC. For that reason, all cases filed by the main opposition parties to the TCC during the given periods are analyzed. This corresponds to 281 cases. 8

Model and Measurement
The dependent variable of this study, litigation, is a binary variable that measures whether the main opposition party brings the law/decree in question, before the constitutional court for abstract review. Litigation is coded 1 if the main opposition party brings that law/decree before the constitutional court and 0 if it does not. Given that the dependent variable is dichotomous, binary logistic regression is the appropriate estimation technique.
Since the main argument of this study posits that the impact of the approaching election on the litigation frequency is conditional on the main opposition party's prediction of its upcoming general election results, we need a measurement of the political opposition parties' expectations about these election results. Because it is impossible to directly measure these expectations, I create proxy measures by drawing on local election results. I suggest that local elections provide important information for the opposition parties about their future electoral success and serve as the earliest source of information about the results of upcoming general elections.
In their influential study, Anderson and Ward (1996) argue that local elections are "barometer elections," which are used by the citizens to send signals to key political actors regarding the incumbent government. For that reason, local elections are believed to reflect changes in public attitudes toward the incumbent government as a response to political and economic conditions. Similar to the British by-elections and German land elections that have been defined by Anderson and Ward (1996) as barometer elections, Turkish local elections also provide some clues as to future voting trends in general elections (Çarkoğlu 2009).
In the Turkish case, when we compare the results of each local election and the subsequent general election, we see that local elections tend to successfully predict the following general election results. For instance, under the second ANAP government, the changing mood of the electorate became evident in the local elections of March 1989 when the ANAP suffered its first major loss. On the other hand, winning 28 percent of the votes, the main opposition party, the SHP, came out as the winner of this local election. This event served as an early sign of the SHP's success in the general election that would be held in October 1991. Similarly, the local election conducted under the DYP-SHP government in March 1994 was an early indication that the ANAP (the main opposition party at the time) would be able to become part of the incoming government. Suggesting that local elections signal certain information to the main opposition party about the general election results, I generate a series of dichotomous timing variables. 9 Since, on average, the local elections were conducted two years prior to general elections, I take the last eighteen months before the general election as the earliest point of information about the upcoming election results. To examine the effects of changes on the main opposition party's behavior under varying relative amounts of certainty and information, the last eighteen, twelve, and six months prior to the general elections are taken as cutoff points. In line with these cutoff points, I create the following independent variables: last-18months, last-12-months, and last-6-months. All laws falling within the mentioned period of the approaching election are coded as 1 and all other laws are coded as 0. This group of dummy variables helps me to compare the behavior of the main opposition party under periods of relative certainty with their behavior under periods of relative uncertainty. Finally, one should bear in mind that the variation in certain political aspects might lead to variation in the use of litigation by different opposition parties. For instance, political fractionalization (single-party government or coalition government), the durability of the incumbent party in office, or ideological differences between the main opposition party and ruling party(ies) might account for the variation in the opposition parties' referrals across different governmental periods. Thus, while testing the core hypotheses of this study, we need an appropriate model that will control for these potential political determinants. In this regard, I used separate binary logistic regressions for each governmental period.

An Analysis of Strategic Litigation and the Empirical Results
The key hypothesis of this study predicts that when the opposition party thinks that it will lose power in the upcoming general election and not be part of the incoming government, it will send laws passed by the Parliament more frequently to the constitutional court as the election time approaches. The second hypothesis, on the other hand, envisages that once the main opposition party expects to be part of the incoming government, it may decrease (or at least not significantly increase) the number of litigations sent to the court as the election approaches.
Given the evidence that the main opposition parties under the first ANAP government, as well as the first and second AKP governments, lost the general elections and continued to act as the main opposition parties, the specific prediction is that each of these main opposition parties should have increased their willingness to go to the court in the final years of each governmental period. In contrast, since the main opposition parties under the second ANAP government and the DYP-SHP government won the upcoming general elections and succeeded in becoming part of the incoming governments, I suggest that these main opposition parties lacked similar incentives to use litigation as a strategic tool at the end of these two governmental periods. Thus, evidence would be consistent with both strategic litigation hypotheses if the main opposition parties' referrals to the constitutional court increased at the end of the first ANAP government, as well as the first and second AKP governments, but decreased or did not change at the end of the second ANAP government and the DYP-SHP government.
The results for each governmental period are presented in Table 2. In this table, we see three sets of models for each governmental period. In the first group of models (1, 4, 7, 10, and 13), the key variable is last-18months, and it is calculated by assigning a value of 1 to all laws passed by the Parliament within 18 months prior to the general elections and 0 for other laws. In the second (2, 5, 8, 11, and 14) and third groups of models (3, 6, 9, 12, and 15), the key independent variables are last-12months and last-6-months. These variables are coded as 1 if a law passed the Parliament within the last twelve and six months prior to the general elections, respectively. All other laws take a value of 0.
I discuss the results presented in Table 2 in conjunction with Table 3, which shows the probability of litigation by the main opposition party for three specific time periods: eighteen, twelve, and six months prior to the general election. Table 3 also presents the baseline probability of litigation for each governmental period. The baseline probability is calculated by holding the last-18months, last-12-months, and last-6-months variables constant at 0. This allows one to see how the predicted probability of litigation looks compared to the baseline probability.
Starting with the first ANAP government, the results for models (1-3) presented in Table 2, overwhelmingly support the first hypothesis, which suggests that when the main opposition party expects to lose the upcoming election, it will increasingly go to the court as the election approaches. Consistent with this prediction, for the first ANAP government, every time-to-election coefficient is positive and statistically significant. Table 3 shows that eighteen, twelve, and six months prior to the election, the probability of the main opposition party going to the court increased by 4, 6, and 7 percent, respectively.
Repeating the analysis for the first AKP government, we get equally impressive results. Table 2 displays that the time-to-election coefficients in models (10-12) are positive and statistically significant. Similarly supporting this finding, Table 3 shows that at eighteen, twelve, and six months prior to the election, the probability of the  1984-1987Second ANAP Government 1987-1991DYP-SHP Government 1991First AKP Government 2002-2007 Second AKP Government [2007][2008][2009][2010][2011] (1) (3) (8) (10) The unit of analysis is a law promulgated by the Parliament, coded 1 = if that law was sent to the constitutional court by the main opposition party, 0 = if that law was not sent to the constitutional court by the main opposition party. Under each governmental period each column contains results for different timing variables. Cell entries are binary logistic regression coefficients. Standard errors are in parentheses. *significant at the 0.05 level. **significant at the 0.01 level. ***significant at the 0.001 level.
main opposition party going to the court increased by 6, 10, and 10 percent, respectively. For the second AKP government, the results are less convincing. Although the coefficients for the last-18months and last-12-months (13-14) are positive, they fall short of statistical significance. However, model (15) shows that in the last-6-months prior to the election, the main opposition party significantly increased its referrals to the constitutional court. According to the predicted probabilities presented in Table 3, in the last six months prior to the general election, the main opposition party was more likely to go to the constitutional court by 6 percent.
To test the second hypothesis of this study, I repeated the same analysis for the second ANAP government and the DYP-SHP government. The empirical results presented in Table 2 strongly support the given hypothesis that suggests the main opposition party will not increase its referrals to the court as the election approached when it expects to win the upcoming general election. Every time-to-election coefficient for the second ANAP government and the DYP-SHP government in models (4-9) is negative. For that reason, one might even argue that when the opposition party expects to be part of the incoming government, as election time approaches, the main opposition party would less frequently go to the court. However, since the time-to-election coefficients are not statistically significant, models (4-9) at least confirm the expectation that when the main opposition party expects to win the upcoming election, its referrals to the court would not increase as the election time approaches.
Although the patterns identified in Tables 2 and 3 empirically support the existence of strategic litigation in the Turkish context, showing how this strategy works in practice is also important. The main opposition parties' talks during their election campaigns reveal the practical implications of our theory. In this study, I have suggested that a main opposition party uses litigation as a tool to differentiate its policy preferences from the incumbent government's policies. In addition, I have proposed that when the main opposition party expects to lose the upcoming general election, it will more frequently go to the court to underline these political differences. In line with this argument, we see that one month prior to the 2011 general election, the main opposition party, the CHP, made some referrals to one of the cases that it brought before the constitutional court. Regarding the land registry law that had been challenged by the CHP, in his talk to the people, the leader of the CHP differentiated between the incumbent government's policy and their future policy on this issue. He said, "This is the difference between us and the AKP. We are on the side of all citizens, whereas they are on the side of their supporters" (Habertürk Newspaper 2011). This statement clearly reveals how the strategic litigation works in practice.

Robustness Check and Alternative Interpretations
An important concern for our empirical analysis is that the results may be biased against the governing majority's legislative behavior. In other words, an opposition party's increasing referrals to the court might be a result of the governing majority's legislative behavior instead of the opposition party's strategic calculations. The number, subject, and controversial nature of the laws passed by the legislature may vary with the approaching election time and therefore increase the opposition parties' referrals to the court. Thus, in this part of the study, we check the robustness of our empirical results by controlling for the explanatory power of these possible determinants.
As an election approaches, the incumbent government may tend to pass laws dealing with certain issues to increase its vote share or to secure its interests/policies after it leaves office. If the subject of the laws passed by the legislature change as an election approaches, then how can one know whether the increase in the number of cases referred by the opposition party to the court was driven by strategic litigation or simply by the change in the nature of the laws passed by the Parliament? Thus, it is not the main opposition party's electoral expectations but rather the subject of the laws passed by the legislature that might account for the increase in the probability of sending laws to the court for abstract review. To check the robustness of our empirical results against this alternative explanation, I ran binary logistic regressions for the first ANAP government, as well as the first and second AKP governments, by including in the model the subjects of the laws as control variables. The results show that the subject of the laws passed by the legislature does not erase the effect of the approaching election time. 10 In the existing literature, it has also been stated that the passing of bills is accelerated as the general election approaches (Borghetto and Giuliani 2012;Lagona and Padovano 2008). Thus, as an election approaches, the opposition party's increasing level of referrals might be explained by the increasing number of laws passed by the legislature. To test for this alternative explanation, I conducted an additional empirical analysis where I controlled for the number of laws passed by the Parliament. Analyzing the combined data set for the first ANAP government, as well as the first and second AKP governments, I find that even after controlling for the number of laws passed by the legislature, the coefficients of the key variables preserves the same sign and remain statistically significant. 11 This finding strongly supports the hypothesis that when the main opposition party thinks that it will lose the upcoming general election, it increases its referrals to the constitutional court as the election time approaches. Thus, the legislative behavior, which is operationalized as the number of laws passed by the legislature, does not appear to affect the significance of the strategic calculations of the opposition party. In this empirical analysis, I also showed that our previous results are robust to using an alternative measure of the dependent variable.
Finally, one might argue that as an election approaches, the legislature is more likely to pass politically controversial bills, which in turn would lead to increasing referrals to the court. Yet this alternative explanation cannot jeopardize the confidence in our results because there are neither theoretical nor empirical studies supporting that assumption. On the contrary, the existing studies argue that the incumbents usually abstain from pushing through controversial laws prior to elections. Focusing on the German case, for instance, Manow and Burkhart (2007) argue that at the end of the Parliamentary term, the majority in the Bundestag abstains from introducing controversial bills. Similarly, Martin (2004) argues that issues that divide the government from the opposition are significantly more likely to be prioritized on the policy agenda.
Regarding the controversial nature of the laws, an important concern might be the governing majority's perception of the likelihood of victory or defeat in the upcoming elections. 12 In other words, incumbents who are on their way out of office might scale back their legislative output considerably and perhaps refrain from proposing controversial bills that may lead to no significant increase in the opposition party's referrals to the court as the election approaches. According to the very same logic, an incumbent party that expects to win the election will increase the controversial bills before the election, which will lead to increased levels of referrals to the court by the opposition party. A proper consideration of this issue would involve a full theoretical and empirical account of legislative behavior, which might clarify whether or not the incumbents' decision to pass controversial bills as an election approaches changes according to the incumbents' perception of victory or defeat in the upcoming election. Such an effort is unfortunately beyond the scope of this work.
However, I present two pieces of evidence that explain why our results cannot be driven by this suggestion. First, as stated above, the existing literature suggests that regardless of the governing majority's perception of the likelihood of victory or defeat in the upcoming elections, the incumbents usually abstain from pushing through controversial laws prior to elections. This means that regardless of the opposition parties' perceptions of the election results, we should observe decreasing (or at least not increasing) levels of litigation. However, this is in complete contrast with our first hypothesis and empirical findings, which suggest increasing levels of referrals by the opposition party when the incumbent government expects to stay in power. Second, if there are certain costs to pushing through controversial bills, it is far from obvious why we should expect an incoming government to endanger its vote share by pushing through controversial laws. An incumbent government that expects to stay in power after the elections may easily push through these laws after the election.
As a result, even after controlling for alternative determinants, we see that the empirical results are consistent with the predictions of the hypotheses and previous findings. All of these aspects increase our confidence that our empirical models are accurate and our findings are not biased.

Discussion and Concluding Remarks
While studying judicial review and the judicialization of politics, scholars have long focused on judicial behavior and the political context within which the judges operate, while usually ignoring the role and motivations of the actors who stimulate the judicial intervention in the first place. Since cases do not come to the courts at random, it is crucial to understand why and when political issues are judicialized rather than being addressed and resolved through traditional political forums in developing democracies where the judiciary is not fully independent. I use the Turkish case to develop a framework for understanding how courts are drawn into the policy game and how policy players use courts to advance their strategic political objectives.
The main argument of this study is that the effect of the approaching election on the opposition party's use of litigation is conditional on its prediction about the general election results. This is because the cost-benefit calculations of an opposition party that expects to be part of the incoming government will be quite different from the calculations of an opposition party that expects to lose the general election and continue to act in its oppositional role. The empirical results show that once the main opposition party believes that it will lose the general election, it increases its referrals to the court as the election time approaches. On the other hand, when it foresees itself winning the upcoming election, the number of its referrals does not show a significant change.
This study speaks to broader debates surrounding the politics of constitutional review in developing democracies. The existing literature on judicial politics in developing democracies suggests that when an incumbent government is close to losing power, judges will shift their decisions to favor the incoming government (Helmke 2005). Based on this theory, one might argue that when the incumbent government is expected to win the elections, as election time approaches, the opposition party will decrease (or at least not increase) its referrals to the court because it will know that the judiciary will not protect its policy preferences. Despite that logical assumption, the empirical results of this study show that the case is just the opposite; when the opposition party expects to lose the election, it will increase its referrals to the court.
Moreover, based on studies suggesting that when the government is divided, the court is more willing to make anti-government decisions (e.g., Iaryczower, Spiller, and Tomassi 2002;Ríos-Figueroa 2006;Scribner 2003), one might argue that an opposition party under a coalition government should more frequently go to the court than an opposition party under a single-party government. In other words, the behavior of an opposition party under a single-party government must differ from the behavior of an opposition party under a coalition government. Nevertheless, our findings show that an opposition party's behavior under single-party government (e.g., SHP under the second ANAP government) might be very similar to the behavior of an opposition party under a coalition government (e.g., ANAP under the DYP-SHP coalition government). However, the opposition parties under single-party governments (e.g., SHP under the second ANAP government and CHP under the first AKP government) might display totally contradictory behavioral patterns.
All of these findings indicate that the opposition parties' decision to go to the constitutional court is not greatly affected by the court's possibility of making antigovernment decisions or the independence of the judiciary. This shows that the opposition parties' decision to go to the court is more than a simple move to win the case and change the specific law/policy under consideration.
Instead, it is a strategic tool that is only used after certain cost-benefit calculations are made.

Acknowledgment
The author would like to thank to Jeffery K. Staton, the editors and anonymous reviewers of Political Research Quarterly for their valuable comments and suggestions on the previous drafts of this article.

Declaration of Conflicting Interests
The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding
The author received no financial support for the research, authorship, and/or publication of this article.

Notes
1. In Bulgaria, Croatia, Czech Republic, Lithuania, and Slovakia, the minimum of one-fifth of the Parliamentary members is needed to initiate the abstract judicial review of legislation. On the other hand, while in Germany and Austria the support of a third of the lower house members is needed, in Poland, Portugal, and Spain, the support of approximately 10 percent of deputies is required for abstract review referrals (Magalhaes 2003, 140-41). 2. The Global Competitiveness Report (2008)(2009) shows that Turkey is ranked sixty-fourth (out of 134 countries) for its judicial independence. Its score is a 4 on the scale that ranges between 1 (not independent judiciary) and 7 (entirely independent judiciary). 3. Gauri and Brinks (2008, 4), for instance, deconstruct the judicialization process into four stages: (a) the placing of cases on the court's docket; (b) the judicial decision; (c) a bureaucratic, political response; and (d) the follow-up litigation. 4. The main opposition party refers to the opposition party with the largest number of seats in the Parliament. 5. The constitutional validity of laws may be challenged directly before the Turkish Constitutional Court (TCC) through an annulment action by the president of the Republic, the Parliamentary group of the main party in government, the main opposition party, or a minimum of one-fifth of the total number of members of the unicameral Parliament (The Constitution of the Republic of Turkey, Article 150). 6. See the online appendix at http://prq.sagepub.com/supplemental/ for the detailed list of the incumbent governments and opposition parties in Turkey between the years 1984 and 2011. 7. The Republican People's Party (CHP), the Social Democratic Populist Party (SHP), and the Democratic Left Party (DSP) are considered as state elite parties. All other political parties in the Parliament are considered as nonstate elite parties (Hazama 2012). 8. A total of 281 laws and decree-laws were sent to the TCC by the political opposition parties during January 1984-December 1995 and November 2002-June 2011 periods. Twenty-three laws were sent to the court under the first ANAP government (January 1984-November 1987, thirty-five laws were sent under the second ANAP government (November 1987-October 1991, forty-eight laws were sent under the DYP-SHP government (October 1991-December 1995, one hundred laws were sent under the first AKP government (November 2002-July 2007, and seventy-five laws were sent under the second AKP government (July 2007-June 2011). 9. The major assumption underlying strategic litigation theory is that the opposition party's expectation-that the opposition party will not be part of the incoming government-increases the probability that a law will be sent to the court. For this reason, I treat timing variables as dichotomous but not as a continuous variable that measures the elapsing time per se. 10. For the detailed presentation of the results and coding of the subject of the laws variables, see the online appendix. 11. For the detailed presentation of the empirical results see the online appendix. 12. I thank an anonymous referee for bringing up this point.