Independent Judicial Review and the Reform of Judicial Nominations

Many constitutional courts and supreme courts possess judicial review powers allowing judges to annul government decisions and parliamentary legislation. Judicial review powers also allow judges to set new standards for government actions and to ignore existing precedents set by the courts (Tate Reference Tate1995). Independent judicial review allows judges to create de facto constitutional changes, which results in a variety of institutional rules (Kmiec Reference Kmiec2004). Thus, judicial review is a policy-making tool with a strong institutional impact that affects the government’s ability to govern (Hirschl Reference Hirschl2008) and potentially the extent of the effective practice of democracy (Epstein, Knight, and Shvetsova Reference Epstein and Shvetsova2001). The relevance and centrality of this topic stems from the fact that the independence of judges in their judicial review is crucial for democratic resilience (Magalhaes Reference Magalhaes1999; Levitsky and Way Reference Levitsky and Way2002; Sadurski Reference Sadurski2008; Corrales Reference Corrales, Diamond, Plattner and Walker2016).

Consequently, leaders seeking to limit human rights, the rule of law, and the liberal vision of democracy (Merkel Reference Merkel2004) have been targeting constitutional courts and supreme courts with constitutional review powers. Their goal is to undermine the scope and impact of the courts’ judicial review. If they want to maintain some commitment to democracy and the rule of law, they try to ensure that the governing elite’s nomenclature handles these issues, thereby making sure that the court does not veto their policies (Levitsky and Way Reference Levitsky and Way2010, 164–65).

For example, in Turkey, led by the then prime minister Recep Tayyip Erdogan from the AKP (Justice and Development Party; Özbudun Reference Özbudun2015), a 2010 reform increased the judiciary’s independence, which slowed down Erdogan’s quest to get rid of the rival Gulen faction and its supporters. Consequently, Erdogan and the AKP initiated a series of reforms in 2014 that limited the judiciary’s review powers and tried to change the judiciary’s composition to increase the government’s control over it (Özbudun Reference Özbudun2015). Similarly, in Poland, since 2017, the ruling PIS (Law and Justice Party) has been promoting reforms regarding judges’ selection committees, trying to fire judges, and increasing the accountability of others to the ruling party. Viktor Orbán’s government in Hungary has also taken similar steps (von Bogdandy et al. Reference von Bogdandy, Bogdanowicz, Canor, Taborowski and Schmidt2018).

A key ingredient in the manner in which governments affect the independence of judicial review is the institutional mechanism by which judges are selected: the judges’ selection committee (Garoupa and Ginsburg Reference Garoupa and Ginsburg2009). Governments differ in how they organize judges’ selection committees. Furthermore, judges’ selection committees vary in the agenda power they offer different branches of government over the selection of judges for courts (Volcansek Reference Volcansek2011). These differences result in variations in the preferences of the selected judges with regard to legal doctrines, political ideologies, and policy preferences (Dahl Reference Dahl1957; Segal et al. Reference Segal, Epstein, Cameron and Spaeth1995; Epstein and Knight Reference Epstein and Knight2000; Epstein et al. Reference Epstein, Martin, Segal and Westerland2007; Posner Reference Posner2010; Garoupa, Gomez-Pomar, and Grembi Reference Garoupa, Gomez-Pomar and Grembi2013; Weinshall, Sommer, and Ritov Reference Weinshall, Sommer and Ritov2018).

More agenda power in a judges’ selection committee offers politicians the option to nominate judges whose particular policy preferences coincide with theirs (Kelemen Reference Kelemen2012). Furthermore, politicians have the option to nominate judges who tend to intervene in government decisions or be more restrained in their rulings. Restrained judges refrain from invalidating government decisions and statutes, while interventionist judges are more likely to annul them (Tate Reference Tate1995; Kmiec Reference Kmiec2004; Cross and Lindquist Reference Cross and Lindquist2006; Segal Reference Segal2011). However, judges might vary in their probability of using either doctrine (Gillman Reference Gillman2001). Thus, for ruling political coalitions, one direct effect of the agenda power of judges’ selection committees should be the degree of court intervention that results from the judges they nominate. If politicians assume that a particular judge’s characteristic would imply a preference for restraint on particular policy issues, they will select judges who share that characteristic. Faced with the lack of such information on judges, what politicians can do is to seek judges whose legal theory shies away from intervention in the executive’s decisions (Whittington Reference Whittington2007).

However, the results of institutional reforms stem from the complex nature of institutions: resilient behavioral structures that involve norms, habits, practices, and symbols whose influence could override the effects of formal rules and their change (March and Olsen Reference March and Olsen1983). A key factor in institutional change and its aftermath is institutional path dependence: the effect of past decisions on the current behaviors of the players within the institutions (March and Olsen Reference March and Olsen1983; North Reference North1990). We see no reason to assume that courts are any different in the manner in which they handle the share of policy inputs of which they are in charge.

Such outcomes should be mitigated by existing precedents regarding the topics the court reviews and the topics brought by petitioners to the court (Robinson Reference Robinson2013). One factor that should be considered is policy subsystems, meaning areas in which policy entrepreneurs, political parties, civil society activists, and government agencies handle a particular aspect of the policy process (Jones and Baumgartner Reference Jones and Baumgartner2005). Governments divide their activities across these subsystems. Thus, petitions brought before a constitutional court that reviews government activities also create a set of legal doctrines relevant to various subsystems (Robinson Reference Robinson2013). Consequently, if a government initiates a reform in judicial appointments in order to reverse court rulings on particular policy issues, having the court rule on such matters would still depend on petitioners seeking the court’s decision and existing precedents on these issues.

Within a public policy context, precedents are existing standard operating procedures that relate to particular aspects of public policies. Once decisions are made and accepted as precedents, they affect the manner in which judges relate to petitions and petitioners, at least with respect to the particular subsystems in which they were set. Hence, governments and judges may change, but the judges still need to relate to a particular policy subsystem in the same manner as they did before the reform. Thus, we posit:

Empirically, we expect that fewer interventionist judges on the court will lead to a reduction in the court’s intervention in government affairs. However, the study of institutional change implies:

Empirically, we expect that previous tendencies on the part of the judges to intervene in cases that relate to particular government subsystems will affect the degree of judicial interventionism.

We first use the Israeli HCJ case of Amitai and Others v. Minister of Science and Technology (HCJ 5769/18) to illustrate our theoretical hypotheses. We then test our hypotheses using our quantitative data set of more than 20 years of rulings by the HCJ.

Illustration: Amitai and Others v. Minister of Science and Technology (HCJ 5769/18)

This case involves a decision made by the minister of science and technology, Ofir Akunis, a member of the dominant right-wing Israeli Likud party, to void the nomination of Yael Amitai for the chancellors’ board of the German-Israeli Foundation for Scientific Research and Development (GIF). The GIF is sponsored and financed by the governments of Germany and Israel and allocates research grants for teams combining German and Israeli scientists (Maariv Online 2018).

A professional committee selected Amitai as the most suitable person for this post. However, in 2001, in the midst of the Second Intifada, Amitai signed a petition declaring that students who wished to refuse military service in the Occupied Territories would receive academic assistance from all of the academics who signed the petition (Roenl and Hovel 2019). Akunis claimed that he could not accept the nomination of someone who had deliberately encouraged citizens to defect from the army (Greentzeig Reference Greentzeig2019). Both Amitai and the Israeli Union of University Presidents petitioned the HCJ to overturn this decision (Greentzeig Reference Greentzeig2019).

The legal basis for Akunis’s decision stating that Amitai had engaged in a criminal activity was controversial (Ashkenazi Reference Ashkenazi2019). In fact, the ministry’s lawyers determined that Akunis’s interpretation of Amitai’s activities was incorrect and contradicted the rules that regulate the terms for GIF’s activities in Israel.Footnote ² The government’s attorney general decided that the minister’s decision to disqualify Amitai based on her political position was unreasonable. Consequently, the attorney general and his lawyers claimed that not only should Akunis’s decision be revoked but also Amitai should take her position on GIF’s board (Ronel and Hovel Reference Ronel and Hovel2019). Amitai’s petition against the minister claimed that his decision was based on incorrect assessments of her past activity and failed to take into account the consequences of that decision for Israel as a whole and the Ministry of Science in particular. Hence, Akunis’s decision was not reasonable and potentially disproportionally limited Amitai’s freedom of speech (Amitai).

In line with common practices in the HCJ, once an HCJ judge on duty related to the petition as justiciable, three supreme court judges were appointed to review this case (Salzberger Reference Salzberger and Russel2006): Judge Neal Hendel, Judge George Kara, and Judge Alex Stein. All three of them were nominated to the court by Israel’s judges’ selection committee after 2008, when the system of appointments changed in a manner that gave the politicians additional power in the judges’ selection process.Footnote ³ Thus, in line with hypothesis 1, we would expect that this panel of judges would be reluctant to overturn the minister’s decision.

The majority decided to overrule the minister’s decision while allowing him to reconsider the nomination after showing him that there was no legal basis for his objection. And the majority accepted the claim that ethical and political considerations can lead to denying someone a government position. However, using precedents set by the court, they concluded that the minister’s decision was indeed unreasonable (Amitai).Footnote ⁴ Thus, the majority’s decision was to partially accept the petition and make the minister reconsider his decision given the fact that the way he handled that process was wrong (Ashkenazi Reference Ashkenazi2019).

The majority was composed of two judges: Hendel and Stein. Judge Stein was nominated as a result of personal pressure exerted on the judges’ selection committee by the justice minister at the time of Stein’s appointment, Ayelet Shaked (Hovel Reference Hovel2018). Shaked was then a minister from the far right-wing party Ha’Bayit Ha’Yehudi (the Jewish Home). Her goal was to reduce the supreme court’s interventionism. Stein was supposed to play an active part in this process (Hovel Reference Hovel2018). In his decision, Stein included some reservations about the fact that Akunis did not receive proper representation before the court. In making that statement, Stein actually went against the strong stance taken by the HCJ in a variety of cases, according to which if the Israeli government’s attorney general decided not to represent a minister, the minister was responsible for his or her own representation (Greentzeig Reference Greentzeig2019).

Judge Kara (the minority opinion judge) claimed that the minister did not seek such support and agreed to represent himself when the attorney general did not agree to represent him. Both Hendel and Kara used decisions made by the HCJ before regarding the attorney general’s independence as an important issue in its own right.Footnote ⁵ Both also claimed that this position was the norm that the supreme court had adopted previously and should be upheld. These previous decisions do not have the weight of precedents but are accepted by the court as binding decisions.Footnote ⁶

Stein rejected that position, claiming that as long as a decision does not have the weight of a binding precedent, he could follow any set of other court decisions that coincided with his views on the matter under review (Amitai). Indeed, a crucial matter for judicial interventionism is whether the court can develop its own theory of law basing it not on precedents but on court conventions. If these conventions do not hold, then different judges with different views can decide on petitions without regard to court conventions, thereby affecting the level of the court’s interventionism (Lindquist and Cross Reference Lindquist and Cross2009, chap. 7). Thus, with more judges reluctant to intervene in government decisions, the level of the court’s interventionism would decline even if it went against the legal theory the court developed in the shape of legal conventions. In conclusion, the institutional reform was successful in introducing new preferences to the court. However, with accordance to hypothesis 2, that success was limited by existing court conventions and some judges seeking to uphold these conventions. We now turn to a more systematic study of our claims.

Israel’s High Court of Justice

Israel’s Supreme Court serves as the highest and final appeal instance in Israel’s legal system. In petitions against the government’s decisions and activities on either administrative or constitutional matters, the supreme court serves as the primary instance (Dotan Reference Dotan, Gavison and Kremnitzer2000) in its judicial role as the HCJ (Eisenberg, Fisher, and Rosen-Zvi Reference Eisenberg, Fisher and Rosen-Zvi2012). Israel does not have a formal written constitution or a constitutional court (Doron and Meydani Reference Doron and Meydani2007).Footnote ⁷ Consequently, Israel’s Knesset constructs its constitution through an incremental process of legislating basic laws. These basic laws define Israel’s institutional structures and the human rights to which it commits (Hofnung Reference Hofnung1996; Lerner Reference Lerner2011).

After the passage of two basic laws in 1992 that institutionalized Israel’s formal commitment to human rights and liberties, Israel’s Supreme Court decided to take a broad legal interpretation of these laws, thereby expanding its jurisdiction over constitutional review.Footnote ⁸ This jurisdiction includes the authority to void legislation and repeal policies if they contradict the human rights guaranteed in the basic laws (Barak Reference Barak2006).Footnote ⁹ The decision to use the basic laws as the constitutional basis for voiding legislation has endowed the supreme court with powers similar to those held by constitutional courts (Hofnung Reference Hofnung1998; Shetreet Reference Shetreet2002; Navot and Peled Reference Navot and Peled2009; Friedmann Reference Friedmann2016). Israel’s Supreme Court has been under political pressure due to its broad interpretation of its judicial review powers (Shetreet Reference Shetreet2002; Doron and Meydani Reference Doron and Meydani2007; Dotan Reference Dotan2014). This pressure has resulted in a series of institutional attempts to influence the procedures of judicial nominations and various political initiatives aimed at reducing the powers of the judiciary, particularly by influencing judges’ selection processes (Navot and Peled Reference Navot and Peled2009).

Two laws determine the legal background for selecting judges for Israel’s courts: basic laws Judiciary (Hashfita) and the Law of Courts. Basic law Judiciary states that judges are selected by a committee that includes nine members (it must not include fewer than seven members). The law requires that this committee must include the supreme court’s president, two other supreme court judges (selected by supreme court judges), the minister of justice (who is the committee’s chairperson), and another minister whom the government nominates. Two members of Knesset (elected by the Knesset) are also committee members along with two lawyers from Israel’s bar association (elected by bar members).Footnote ¹⁰ An amendment to the 1984 Law of Courts enacted in 2004 states that committee members do not represent the institution that placed them on the committee.Footnote ¹¹ At least formally before the amendment, factions on the committee (judges, politicians, and lawyers) selected candidates while taking instructions from their nominating institutions (the supreme court, the Knesset, and Israel’s bar association). The 2004 amendment meant that potentially there could be individual defections from the factions when voting on judges’ nominations.

In 2008, the government initiated and enacted a more critical amendment to the judges’ selection process. The amendment determined that in order to select a supreme court judge, there has to be a majority of seven of the nine committee members.Footnote ¹² When a simple majority was the selection rule, it gave the judges an advantage if they had even two supporters from the other committee factions (politicians or lawyers) or if these factions were split (Shetreet Reference Shetreet2003). In practical terms, this amendment meant that a coalition supporting a judge’s nomination has to include some of the politicians who are committee members (Friedman Reference Friedmann2016, chap. 33). Consequently, if the politicians act in a cohesive manner in the committee, they can be a major veto player on candidates and committee outcomes. Judges can still maneuver and block candidates the politicians are interested in against the judges’ will. However, the amendment does compel the judges to cooperate with the other factions on the committee because the reform created the theoretical possibility that the politicians and lawyers could force the nomination of a candidate of whom the judges do not approve.

Hence, we have a case of a judges’ selection committee whose judges were selected using methods devised more than a decade ago (2008) allowing a comparison of court behavior before and after the reform. We now turn to a quantitative examination of judicial review at the HCJ before and after the reform.

Data, Variables, and Measurement

Our data set begins with the landmark case of United Bank Hamizrahi Ltd v. Migdal, Cooperative Village in 1995 and includes cases up to December 2017. In United Bank Hamizrahi, which some describe as Israel’s Marbury (Weill Reference Weill2012), the supreme court declared that it could annul legislation based on the constitutional rights set by the basic laws passed in 1992 and amended in 1994. Given that HCJ judges are supreme court judges and that decision applied to the HCJ, we used it as our starting point. We collected the legal data about the cases from several sources, mainly from the Israeli legal database Nevo, which regularly publishes HCJ rulings and decisions.Footnote ¹³

The data set includes court rulings and decisions made by the HCJ on petitions against the prime minister, government ministers, and deputy ministers (all elected politicians). However, it excludes petitions primarily targeting the bureaucracy, the government ministries as a whole, the government as a whole, and the Knesset as a whole. We included petitions against these agents only if the petition targeted an elected politician in the Israeli executive branch. Our aim in constructing our data set was to examine decisions made by the HCJ with respect to Israel’s executive branch of government.

To develop a data set of final decisions related to the executive, we asked the coders to remove any petition that was rejected without certiorari by the court. Procedurally, the HCJ has a first review of petitions, and the judge on duty can decide to reject a petition without certiorari. Moreover, even petitions approved by the judge on duty can be rejected without certiorari by the panel of three judges that reviews approved petitions. When the coders identified such a decision in a case docket, they did not include it in the data set. We asked the coders to do the same with petitions withdrawn by petitioners or that became redundant even after passing the first two phases of elimination. Consequently, the data set includes final decisions on petitions aimed at the executive that were justiciable, concrete, and ripe for a court hearing.

There were about 20,000 petitions filed with the Israeli HCJ during this period of more than 20 years that we studied (e.g., Dotan Reference Dotan2014, 25). Using our search terms and keywords narrowed down the potential set to 2,960 court rulings.Footnote ¹⁴ For each case, we had the full information regarding the way in which the judges ruled on each case and who the judges on the panel were. We divided these court decisions into 9,806 judicial decisions taken by individual judges who belonged to the court panels that made the 2,960 rulings.Footnote ¹⁵ Thus, our unit of analysis is a judge sitting on an HCJ panel making final decisions regarding petitions against politicians in the executive branch of the Israeli government.Footnote ¹⁶

The data set includes the date of the petition, the date of the ruling, the main issues raised in the petition, the petitioning parties, the petition’s respondents, the court ruling, and the particular ruling made by each judge sitting on the case. This data set includes 41 judges dealing with a variety of constitutional and administrative matters the court reviewed, interacting with about eight different parliaments and 10 different governments in times of varied intensity regarding Israel’s internal and external conflicts. Consequently, this data set fulfills the main requirement of a case study used to test theoretical hypotheses: the ability to maximize variance in the studied variables (Gerring Reference Gerring2004).

Data Analysis

Table 1 shows the results of our analyses of the variables in various periods: the entire data set (model 1), the HCJ before the 2008 reform (model 2), and the HCJ after the reform (models 3 and 4). We estimated the postreform court twice: without the judges’ functions on the panels (model 3) and with them (model 4). The independent variables included four characteristics of the judges: their nationality (Arab or Jewish), whether the judge had been on a court of lower instance, the judges’ religiosity, and whether the judge was nominated after the 2008 reform. We then estimated the judges’ functions on the panels, first for the effect of the various judges who headed the panels and then for the effect of the panel’s head judge being nominated for the supreme court after 2008. Our next step was to explore the policy subsystems that were not collinear—security, leisure, social welfare, and religious affairs—as well as the interactive effect between the judges’ religiosity and whether the petition related to religious affairs. We added this interaction term due to the potential influence of judges’ religiosity on their decision-making on matters related to religion, which are very relevant to the Israeli case as a whole (Dowty Reference Dowty1999) and HCJ judicial review in particular (Sommer Reference Sommer2010; Weinshall-Margel Reference Weinshall-Margel2011). We modeled the petitions’ political context as a grouping variable, which reflects the external effect of the political and administrative environment on the judges and the petitions.

Table 1.Regression Models

Note. Robust standard errors in parentheses. RE = random effects.

+. p < .1.

* p < .05.

** p < .01.

*** p < .001.

Our modeling strategy in compiling four multilevel models was aiming at testing our hypotheses while monitoring various possibilities for causality we related to in the theoretical section and in our review of the Israeli case. In terms of hypothesis 1, we expected that a judge and even more so a judge who heads a panel nominated after 2008 would decrease the likelihood for a petition’s acceptance (a negative sign). However, following hypothesis 2, we also expected that the various policy subsystems would have their particular and contradicting effects on a petition’s acceptance. In order to explicitly monitor these complex effects, we estimated our model in four separate phases, nesting the model in its political context and monitoring for the time effect and judges’ institutional positions (i.e., whether the judge headed a panel) on judicial review.

Each table cell includes the parameter coefficient and robust standard error in parentheses. We also calculated the models without robust standard errors and verified that indeed the multilevel models are more efficient than standard logistic models. The model estimates included the Akaike information criterion, Bayesian information criterion, and log pseudolikelihood. The rule of thumb for these measures is that the smaller they are, the more efficient the model they estimate becomes. The most efficient model is the one that relates to the period after the reform and tests the post-2008 effects on the judges. In terms of our grouping variable, there is a decrease in the variance in the data and standard deviations stemming from whether the petition had a political content, in the post-2008 models ( $\mathrm{constant}=.273$ and .222), in comparison to the general model ( $\mathrm{constant}=.517$ ) and the pre-2008 model ( $\mathrm{constant}=.856$ ). In other words, in the prereform period, the fact that a petition had an ideological political content had a much stronger influence on the manner in which the judges related to that petition than after the reform.

The complete model shows that Jewish judges who served on lower-level courts, headed a panel (especially if they were nominated after 2008), and dealt with petitions against ministries handling security affairs tended to rule against the petitioners. In the prereform court (model 2), judges who were Jewish, had served on a lower court, were religious, headed the panel, and dealt with security issues tended to reject petitions. In addition, judges during that period tended to accept petitions on social welfare issues. Regarding religious affairs, as a general trend (model 1), the HCJ tends to accept petitions on this matter. However, the interaction term implies that this trend depended on the judges’ background. In the postreform court, judges nominated after the reform, especially those who headed a panel and dealt with security matters, tended to reject petitions.

In conclusion, following the 2008 reform, we definitely have judges who are less interventionist than those who were nominated before them and served with them. Hence, the strategy of affecting judicial nominations to the supreme court (and consequently the HCJ) was a success for those whose goal was to limit judicial review. However, the general trends of the court’s decision-making from before the reform persisted for at least some policy subsystems.

Concluding Discussion

There have been various attempts in democracies worldwide to influence the composition of the courts in order to reduce judicial interventionism. Our results enrich the understanding of troubled democracies that engage in such practices (Merkel Reference Merkel2014; Levitsky and Way Reference Levitsky and Way2015). The first lesson this study offers is a well-known one: institutions matter, and tampering with institutional rules may produce alternative policy outcomes (Hammond and Butler Reference Hammond and Butler2003). In this case, the judges reacted to their political environment and the attempt by the executive to exert pressure on them in the form of institutional reforms designed to limit the judiciary’s powers. Even though the new judges nominated were less interventionist, the court as a whole reduced its intervention in government decisions. In strategic terms, the judicial nominations reform was a signal that the court should refrain from intervening in government decisions (Epstein and Knight Reference Epstein and Knight2000; Epstein, Knight, and Shvetsova Reference Epstein and Shvetsova2001). In terms of democratic theory, this is a setback and a real threat to democracies and their resilience (Levitsky and Way Reference Levitsky and Way2002). Moreover, our findings might support the claim made by theorists of minimalist democracy, according to which courts are too fragile to lean on in battles to maintain the resilience of a democracy (Shapiro Reference Shapiro2016, 70–72).

However, our findings also indicate that, due to institutional path dependence, reformers who seek to change judges’ behavior yet still maintain the democratic rules of the game should not anticipate a dramatic change. Specifically, in the Israeli case, even though it is a defective democracy (Merkel Reference Merkel and Herman2012) facing real challenges to its democratic prospects (Mordechay and Roznai Reference Mordechay and Roznai2017), Israel has developed institutionalized legal norms of judicial involvement in government decisions. Hence, any attempt to reform this court still faces institutional precedents and norms that curb the influence of such reforms. This shows that the effect of institutional reforms on policy outputs in democracies is not necessarily a direct outcome of formal rules (Scheppele Reference Scheppele2013).

The reform we discussed here is not too different in political purposes from those recently initiated in parts of Europe. Changes in the judiciary’s composition have been promoted in Poland and Hungary, with some European states aiming to follow suit, namely, Romania and Bulgaria (Kelemen and Pech Reference Kelemen and Pech2019). These governments claim that by limiting the court’s power and its judicial intervention in state affairs, they are expressing their own national vision, which is different from the liberal democratic version of the rule of law (Kelemen and Pech Reference Kelemen and Pech2019). Soon thereafter, the European Court of Justice disagreed with this idea and its consequences (Rosas Reference Rosas2019).

We analyzed the consequences of reforming the judicial nomination process for political reasons for the independence of the judiciary. Adopting this approach results in a hesitant judiciary that shies away from interfering in matters that do not coincide with existing legal doctrines. In places where such doctrines are not already in place (Matczak, Bencze, and Kühn Reference Matczak, Bencze and Kühn2010), the outcomes should be even more direct than they were in Israel. Thus, reforms in new democracies designed to limit the power of constitutional courts through institutional reforms could achieve their goals, thereby harming democracy and the rule of law.