Clifford Carrubba, Matthew Gabel, and Charles Hankla. 2008. Judicial Behavior Under Political Constraints: Evidence from the European Court of Justice.

Clifford Carrubba, Matthew Gabel, and Charles Hankla. 2008. "Judicial Behavior Under Political Constraints: Evidence from the European Court of Justice." American Political Science Review 102(4): 435-52.

This paper investigates how two types of political constraints (threats of noncompliance and legislative override) influence decisions by the European Court of Justice (ECJ).

I. Background

• The ECJ has established two doctrines:
1. Direct effect - individual citizens of th EU can invoke the EU law, treaties and, to a lesser extent, secondary legislation as the bases for legal claims in national courts
2. Supremacy - if national law and EU law are incompatible, it is EU law that should be applied. 
• Constraints on judicial review in democracies:
1. Threats of noncompliance
2. Legislative override

II. Hypotheses

1. The more credible the threat of override by the European Council, the more likely the court is to rule in favor of the governments' favored position.
2. The more opposition a litigant government has from other member-state governments, the more likely the court is to rule against that litigant government because the ECJ might be able to count on third-party enforcement.
• Empirically determined by the opinion of governments for or against the plaintiff on the legal issue in the case, which were indicated by submitted briefs.
3. The relationship defined in hypothesis 2 is weaker in preliminary ruling proceedings (article 234 cases) that in direct actions. That is, the Court is less sensitive to threats of noncompliance for legal issues from preliminary reference cases (Article 234 cases) than for legal issues from other sources 

III. Data and coding

• Dataset of decisions on within-case legal issues rather than cases themselves.
• Each observation in the dataset is a legal issue disposed of by the ECJ when it decides a case. If a case deals with multiple legal  issues, there are recorded as separate observations.
• The data, as coded, allows the authors to examine both types of threats simultaneously and evaluate their relative influence on judicial behavior; allows an empirical comparison of the two constraints in the same judicial setting.
• Dependent variable in all the models:
• 1 if the ECJ rules in favor of the plaintiff; 0 if the ECJ rules against the plaintiff.

IV. Model Specifications

1. Test for the override threat:
• Ruling for Plaintiff_i = β₀ + β₁ Net Obs_i + ΓZ_i + e_i
2. Test for the override and noncompliance threats simultaneously:
• Ruling for Plaintiff_i = β₀ + β₁ Net Obs_i + β₂ GovLit_i + β₁ Net Obs_i*GovLit_i + ΓZ_i + e_i
3. To isolate legal issues where primarily noncompliance is at issue, the following model is estimated for all legal issues due to infringement proceedings under Article 226:
• Ruling for Plaintiff_i = β₀ + β₁ Obs for Plaintiff_i + β₂ Obs for Defendant_i + ΓZ_i + e_i
4. The model 3 is also used to isolate legal issues where primarily a threat of override is at issue by focusing on legal issues where neither the plaintiff not the litigant is a member-state government and where the legal issue pertains to a question brought under article 234, a request for a preliminary ruling form a national court. Noncompliance is not a concern in this case because there is no government litigant.
5. Test for hypothesis 3:
• Ruling for Plaintiff_i =  β₀ + β₁ Net Obs_i + β₂ Net Obs_i*GovLit_i + β₃ Net Obs*GovLit_i*Article234_i+ ΓZ_i + e_i
• The model specification is the same as in model 2, with the exception of the dummy variable Article234_i indicating whether a legal issue is from a preliminary ruling case. However, to test hypothesis 3, separate effects for net observations when a government is a litigant in a preliminary ruling case and when a government is litigant but not in a preliminary ruling case are estimated.

• i indexes legal issues, Net Obs is the net weighted observations/governments in favor of the plaintiff, GovLit is a dummy variable for the presence of national government as a litigant, Z is the vector of control variables, and e is the error term. The weights are the vote share of the member-states under qualified-majority voting in the Council of Ministers, which is highly correlated with the measures of member-state economic characteristics that reflect their ability to impose economic sanctions for noncompliance, such as national share of intra-EU trade and EU Gross Domestic Product.
• Obs for Plaintiff indicates the number of government observations for the defendant and Obs for Defendant indicates the number of government observations for the defendant, all weighted by national vote share in the Council of Ministers.

V. Empirical Results

• Model 1 is consistent with the first hypothesis since β₁ > 0.
• The probability of a ruling for the plaintiff increases with the net number of weighted observations for the plaintiff.
• As the likelihood rises that a coalition of member-states could form to override a ruling in favor of the defendant, the Court is more likely to rule for the plaintiff.
• Model 2 results comport with the second hypothesis since the conditional coefficient (β₁ + β₃) > β₁ > 0.
• The marginal effect of net weighted observations for the plaintiff is positive when a government is not a litigant and net weighted observations have an additional positive impact on ECJ rulings when a government is a litigant. 
• The additional effect of observations when a government is a litigant is generally greater than the effect of observations when a government is not a litigant.
• From Models 1 and 2, on average, threats of noncompliance have a greater impact on ECJ rulings than threats of override. This is not surprising because threats of noncompliance are easy to execute relative to threats of override.
• Model 3 results are consistent with the expectations: β₁ > 0, β₂ ≤ 0, and β₁ ≥ |β₂|.
• Weighted observations for the plaintiff have a positive and statistically significant effect on the probability the ECJ rules for the plaintiff.
• Weighted observations for the defendant have no statistically significant effect on ECJ rulings for the plaintiff.
• When a government is a litigant, net observations for the plaintiff in the postive range include infringement proceedings with observations for the plaintiff in the positive range include infringement proceedings iwth observations for the plaintiff, which are highly influential on the Court
• Infringement proceedings with the net observations for the defendant primarily involve observations for the defendant, which are not influential on the Court's rulings. 
• Model 4 results are consistent with the expectations: β₁ > 0, β₂ < 0, and β₁ + β₂ = 0.
• For legal issues from Article 234 cases with no government litigant, observations for the plaintiff and the defendant influence ECJ rulings in the expected directions and the effects are statistically significant. 
• The null hypothesis that the sum of the two effects previously mentioned is zero is cannot be rejected.
• Model 5 coefficients are inconsistent with expectations, which are (β₁ + β2) < (β₁ + β₂ + β₃), but standard errors are too large to determine significance for all coefficients.

VI. Conclusion

• Threats of noncompliance have a greater impact on ECJ rulings than threats of override.
• Results indicate that political constraints systematically shape ECJ rulings.
• The alternative interpretation: if legal precedent, the quality of the legal argument, or the content of relevant treaty articles favor a particular litigant, we would expect the Court, ceteris paribus, to favor that litigant. And, while government observations plausibly act as indicators of government preferences over legal issues, they may also signal the quality of each litigant's case on the merits. If so, then the better the case on the merits for a litigant, the more likely it is that governments will submit observations agreeing with that litigant's position and that the Court will rule for that litigant. So may be positive correlation between government observation and Court rulings even if the Court is not responding to government threats of override or non-compliance.