3rd Economic Analysis of Litigation Workshop June 27-28 2016 How and why Russian judges decide on appeals of the decisions of administrative authorities (on the example of antitrust law) 3rd Economic Analysis of Litigation Workshop June 27-28 2016 Svetlana Avdasheva, Svetlana Golovanova, Elena Sidorova Higher School of Economics , Moscow, 2016 www.hse.ru

? To rely on regulators in contrast to judges Institutional reform in transition economies To improve legal rules and protect the independence of judges To rely on regulators in contrast to judges (Shleifer, 2012 ) photo ? photo photo Higher School of Economics , Moscow, 2016

Future of the rule of law in transition economies Still developing legislation Increasing impact of court decisions Evidence of decisions that seem ‘wrong’ Judicial errors? Prosecutorial bias? Corruption? Evidence on how judges behave is important: to answer the questions on the future of institutional development in transition to assess the legal reforms (for instance, Russian on-going court reform, convergence of commercial courts with the courts of general jurisdiction) photo photo photo Higher School of Economics , Moscow, 2016

Scarce empirical evidence on the judicial behavior in transition ex-socialist economies Slovenia (Dimitrova-Graizl et al, 2010, 2011…) Determinants of judicial productivity and quantity-quality tradeoffs – including caseload, gender, education, promotion concerns, salary, specialization Russia (Schults et al, 2014) Prosecutorial bias in criminal courts Russia (Volkov et al, 2014) ‘Claimant-bias’ in commercial courts Impact of personal characteristics of judges on outcome Russia (Trochev, 2012, 2014) Explains paradox of successful suing executive authorities in Russian courts Complex strategies to protect judicial independence photo photo photo Higher School of Economics , Moscow, 2016

Our focus Presumed prosecutorial bias as bias in favor of executive authorities (‘failure of judges’) Judges in the courts specialized on civil cases (*) In the claims which result from completely new legislation for ex-soviet country Are decisions predicted by the ‘specific ex-socialist considerations’ ? If yes, which are these ‘specific ex-socialist considerations’? photo photo photo Higher School of Economics , Moscow, 2016

Russian context-1: commercial courts Until recently: two relatively isolated branches of court systems to decide criminal and civil cases between physical persons (general jurisdiction courts) to decide cases between entrepreneurs as legal persons and between entrepreneurs and executive authorities (commercial courts) organized by regions – 83 regional courts of first instance Specialized appellation and cassation instances in every branch Since 2014 under the process of unification (subordination of commercial courts system to general jurisdiction court system) Access to court system is extremely easy in Russia – negligible fees, no restrictions of representation, motivation on no backlog Recenty from 1,2 to 1,4 mln cases annually, about 5 mln against administrative decisions of executive authorities photo photo photo Higher School of Economics , Moscow, 2016

Russian context-2: judges and their motivation Legislation on the judges’ carrier in commercial courts of first instance: First three years appointment Then life tenure (conditional on the records in first three years) Motivation Motivation on no backlog Motivation on no appeals : if appellant is not satisfied with the decision of the first instance + appeal + cassation and Supreme Court considers decision does not comply with substantive or procedural rules – the case returns to the first instance Executive authorities almost automatically appeal unfavorable court decision, more often than companies (in case of company vs authority) photo photo photo Higher School of Economics , Moscow, 2016

Russian context-3: antitrust Among the clams to executive authorities (under administrative legislation enforcement) claims to annul infringement decisions of competition authority (on antitrust legislation) are not large but important part High threshold of fines on antitrust violation since 2007 In comparison to EU, under similar substantive rules antitrust legislation is enforced much more often Distributional considerations are extremely important for the authorities Abuse of dominance: ‘exploitative’ conduct Large number of appeals on infringement decisions of competition authorities: roughly about every 3rd decision is appealed, from 3 thousand decisions (followed from 10 thousand investigations) annually photo photo photo Higher School of Economics , Moscow, 2016

Specific context provides opportunities to test hypotheses On the bias towards executive authorities: If there is a bias towards executive authorities it should be stronger for the judges on their first three-years appointment It should be stronger in the regions with poor institutions On the role of experience and education of judges Since large part of antitrust cases considered in the courts should be enforced according to effect-based standard Comparison of expected welfare effects depends on the ability to evaluate positive efficiencies specific to the case Better educated judge with higher likelihood makes decision in favor of company More experienced judges can do the same On the impact of regional institutions In the regional with better institutional environment judges should be less biased towards executives photo photo photo Higher School of Economics , Moscow, 2016

Data and hypotheses of empirical analysis 3682 decisions of the first instance courts on the claims to annul the infringement decisions of Russian competition authorities on art. 10 (abuse of dominance) and art. 11 (agreements and concerted practices) during from 2008 (first year of enforcement after the introduction of high penalties standards) to 2012 Fast decisions: 10 months to the final decision in average 2/3 judges in the sample are female 22% in their first three year appointment 20% obtain their degrees after the breakdown of Soviet Union Annulment rate decreases from 51% in 2008 to 34% in 2012 Evidence in favor of increasing bias towards the administrative authorities? photo photo photo Higher School of Economics , Moscow, 2016

Data and hypotheses of empirical analysis Variables Expected impact on the likelihood of annulment H1: On bias towards executives First three year appointment (Yes = 1) Decrease H2: On the experience/education of judge Number of antitrust cases resolved (log number of cases) Increase Judge educated in post-socialist period(Yes =1) Judge with economic education (Yes =1) Н3: On the impact of regional institutional environment Investment risk (as negatively correlated with quality of institutional environment) (divided in quartiles) Control variables: claimant Amount of fines (log_ RUR) Conduct remedies are applied Control variables: competition authorities Experience (number of appeals considered before judge – log number) Workload (log cases per capita) Prosecutorial bias (ratio of infringement to investigation) photo photo photo Higher School of Economics , Moscow, 2016

Data and hypotheses of empirical analysis Variables Expected impact on the likelihood of annulment Control variables: judges Gender (male = 1) Increase Experience in deciding antitrust cases – log number of cases Experience in commercial courts (general experience) – log number of year ? PhD in Law (general experience) (Yes = 1) Control variables: conduct Duration of litigation (log months of litigation) Abuse of dominance in question (Yes = 1) Decrease Control variables: region Log Regional GDP Institutional quality (investment risk rank) Control variables: industry Share of industry in GDP photo photo photo Higher School of Economics , Moscow, 2016

Main empirical results Variables (1) (2) (3) (4) (5) Complexity of the case Art_11 (agrts) 0.07*** (0.02) 0.08*** (0.02)   Duration 0.00*** (0.00) 0.01*** (0.00) 0.01***(0.00) Expertise 0.06** (0.02) 0.06** (0.03) Characteristics of the contested sanctions Conduct Remedies 0.05*** (0.02) 0.05** (0.02) Monetary Fines 0.07*** (0.03) Ln_Fines 0.02*** (0.00) Characteristics of CA subdivision Experience_CA -0.04*** (0.01) -0.06*** (0.01) -0.05*** (0.01) Characteristics of the judge Mage judge 0.03* (0.02) Econ Education 0.12* (0.07) 0.10 (0.07) Post-Soviet Degree Experience Ant-t 0.04*** (0.01) 0.03*** (0.01) Year YES Regional variables Number of obs 3682 3673 3630 3130 3132 LR chi2 80*** 111*** 126*** 139*** 142*** Pseudo R2 0.016 0.023 0.026 0.033 0.034 photo photo photo Higher School of Economics , Moscow, 2016

Empirical results-1 (from probit regression) On the main hypotheses: Decisions of the judges on first appointment do not differ from those of tenured judges Institutional quality does not influence the probability of annulment Judges with economic education annul infringement decisions more often Judges with post-socialist education annul infringement decisions more often The higher the experience of judge in resolving antitrust cases the higher the probability of annulment of infringement decision. General experience (either measured) does not provide that impact Other interesting results: Women judges confirm the decisions of competition authorities more often Case complexity increases the probability of infringement decision annulment Experience of competition authorities decreases the probability of annulment Larger share of industry in GDP increases the probability of infringement decision annulment photo photo photo Higher School of Economics , Moscow, 2016

Empirical results-2: why does annulment rate decrease? In the baseline regression years are significant… They become insignificant when we take into account: Increasing experience of competition authorities (from 5 cases being presented before judge in 2008 to 72 cases in 2012 cumulative for regional subdivision of competition authority) This effect overweighs the effects of increasing experience of the judges (from 11 to 64 cases decided in average before particular case) of increasing amounts of fine (from 1,1 mln RUR to 2.5 mln RUR in average) Therefore there is not independent decreasing trend in annulment rates… photo photo photo Higher School of Economics , Moscow, 2016

Discussions and implications Not all the hypotheses are confirmed, and even not control variables work in expected direction In spite of all that, contribution of the results to discussion: On the comparative analysis of generalist vs specialized judges. As in Baye and Wright (2011) specialization (which is experience in our specification) matters for the litigation outcome On the direction and scale of bias in the decisions of Russian judges. Results supports the conclusions of Volkov et al (2014) and Trochev (2012, 2014) on the sharp difference between judges in commercial courts and courts of general jurisdictiolns On the expected outcome of the on-going court reform in Russia. Possible unfavorable outcomes of subordination of commercial courts to the higher courts of general jurisdiction photo photo photo Higher School of Economics , Moscow, 2016

Conclusion There is no doubt that transition matters for the decision of the judges: However there is no evidence that judges in Russian commercial courts are biased towards executive authorities There is no evidence that overall quality of institutions affects judges Therefore rule of reason in the Russian commercial courts is not completely ‘failed’ At the same time education and experience matter a lot Under enforcement of administrative legislation courts with better educated and more experienced judges effectively correct Type I errors (wrongful convictions) photo photo photo Higher School of Economics , Moscow, 2016

20, Myasnitskaya str., Moscow, Russia, 101000 Tel.: +7 (495) 628-8829, Fax: +7 (495) 628-7931 www.hse.ru