# I. introduction he judicial system is one of the foundations of a modern state. It is the guarantor of the application of laws and a factor of good governance and development impetus. Therefore, Morocco has become aware of the importance of promoting its judicial system to support the country's structural reforms. In fact, a reform committee was formed in 2012 to make a diagnosis and propose a reform plan. According to this diagnosis, the Moroccan judicial system is plagued by dysfunctions and weaknesses, as well as slowness and deep complexity, and even some perverse practices that have caused citizens to lose faith in the justice system. Following this diagnosis, the government adopted a national charter for judicial reform, comprising six strategic axes, and began the implementation in 2013. The state has committed significant human and financial resources to implement this reform. This includes increasing the number of judges and clerks, improving training quality, and improving infrastructure and justice services. This effort increased the share of the justice budget in the state budget to 1.6%, which raised questions about the effectiveness and efficiency of the actions implemented. Thus, our study aims to identify the efficiency determinants of Moroccan courts during the implementation of the judicial reform charter. For this, measuring the court's efficiency in the first stage and using an OLS ordinary least squares regression in the second stage to identify the variables that influence the court's efficiency. As with most previous studies, the number of judges, clerks, and staff costs were used as inputs for variable selection, while the number of cases resolved was used as output. Our primary source was data from the Department of Justice. To analyze the efficiency determinants, we identified the exogenous variables mentioned in the literature. Then we collected the data available in the reports of the Ministry of Justice and other state organizations such as the High Commission for Planning. As a result, this study contributes to a better understanding of the impact of judicial reform on court efficiency and the factors that influence it. This can assist decision-makers in making managerial decisions. Furthermore, to the best of our knowledge, our research is the first to examine judicial efficiency in Morocco and among developing countries. The paper is organized as follows: Section 2 presents a literature background of the study, Sect. 3 presents the methods and data, Sect. 4 presents and discusses the results obtained, Sect. 5 contains the main findings of the study and implications for future research. # II. # Literature Review According to the literature review, the DEA method has been widely used in studies aimed at measuring efficiency and productivity (Johnes, 2006a). This is enabled by the non-parametric method's ability to include multiple inputs and outputs in the measurement, which is useful for non-profit organizations. According to Voigt (2016), few studies have addressed the technical efficiency of justice systems, resulting in a lack of interest among researchers in cost-related studies in this sector. Moreover, according to Rosales-Lopez (2008), there are fewer DEA studies than those focusing on the quality of judicial decisions. This is more apparent in African countries where we have identified a single study (Elbialy and Garcia-Rubio, 2011) which deals with judicial efficiency. Researchers frequently mention the court size variable in studies that have addressed efficiency determinants. Indeed, Yeung and Azevedo (2011) indicated that the size of a court, represented by the number of judges and clerks, has a positive influence on the efficiency of Brazilian state courts. The same sense of relationship has been proven by Schwengber and Sousa (2005) concerning the Rio courts in Brazil and by Santos and Amado (2014) concerning the Portuguese courts. Beldowski et al. (2020) also indicated that an increase in the number of judges can significantly enhance the number of resolved cases. Moreover, they found that court efficiency is significantly associated with some auxiliary court staff members and variables capturing the economic development of court jurisdiction. Similarly, Finocchiaro Castro and Guccio (2014) demonstrate that citizens' high demand for justice and the presence of large courts are positively correlated with the presence of efficient courts. On the other hand, Finocchiaro Castro and Guccio (2014) pointed out that the number of pending cases is associated with low efficiency for Italian courts. The same result was reported by Ferro et al. (2018), Lewin et al. (1982), and Castro and Guccio (2016) also mentioning the negative effect of workload. In the same framework, Fauvrelle and Tony C Almeida (2018) studied the efficiency change determinants of Brazilian State Courts between 2009 and 2014 and tested the influence of exogenous variables such as the proportion of criminal cases. The results indicated the non-existence of a significant relationship between the proportion of criminal cases and the court's efficiency. However, Garcia-Rubio and Elbialy (2011) asserted in their work that the complexity of civil cases negatively impacts the efficiency of Egyptian first instance courts. Regarding the influence of court human resources, Santos and Amado (2014) indicated that courts with a higher proportion of administrative staff are more efficient than those with a higher proportion of judges. On the other hand, some studies (Dimitrova-Grajzl et al., 2012;Schneider, 2005) highlight the academic level of judges and indicate that the higher a judge's level of education (for example, a doctorate), the more efficient the court. Other studies emphasize the significance of the court's management performance, which is directly related to the court president's profile. In this regard, Yeung and Azevedo (2011) confirmed, using management performance evaluation indicators, that efficiency is positively correlated with management performance. From another perspective, several studies have indicated the relationship between court efficiency and the age or seniority of judges. Thus, Ferro et al. (2018) showed through the study of first instance courts in Argentina during the period 2006-2010 that the age of judges negatively affects efficiency. Indeed, the court is inefficient as the average age of judges increases. Bhattacharya and Smyth (2001) reached the same conclusion when they studied the supreme courts in Australia and found that judges perform poorly as they get older. Furthermore, Ferro et al. (2018) and Dimitrova-Grajzl et al. (2012) found no significant relationship between a judge's service length or gender and court productivity. Elbialy and Garcia-Rubio (2011), on the other hand, confirmed that only high-ranking judges increase productivity and efficiency in Egyptian courts. Similar to judges, previous studies have attempted to investigate the possible link between the profile of a court's clerk and its efficiency. Thus, Ferro et al. (2018) indicated that the seniority of clerks positively affects efficiency, while there is no proven relationship between the gender of staff and efficiency. Dimitrova-Grajzl et al. (2012) underline that there is no relationship between the experience of clerks and the productivity of a court. They believe that the experience of judges and their specialization is more important than the seniority or experience of court clerks. In terms of economic activity, Fauvrelle and Tony C Almeida (2018) indicated that there is no correlation between GDP per capita and court productivity. Indeed, socioeconomic factors do not always have an impact on the efficiency of the court. According to Yeung and Azevedo (2011), the most efficient courts are those located in economically active regions, where cases are typically more complex. Another critical factor emphasized by Falavigna et al. ( 2015) is the court's geographic location. Indeed, according to this study, the efficiency of Italian courts increases from south to north west. The characteristics of the population were also treated as exogenous variables that could influence court efficiency. In this context, Gorman and Ruggiero (2009) assessed the efficiency of prosecutor offices in 26 US states, concluding that those in low-income counties with a minority population are less efficient. Furthermore, a 10% increase in the average income of the population results in a 3% increase in efficiency, whereas there is no significant relationship between the percentage of the population with a bachelor's degree and the efficiency of prosecutor offices. In general, the literature review reveals that many studies have examined the determinants of efficiency from various perspectives. However, the literature does not show similar results because the determinants vary across countries and judicial systems. # III. # Methods a) Data Envelopment Analysis Charnes et al. (1978) introduced the Data Envelopment Analysis (DEA) as a non-parametric method using linear programming to measure the efficiency of a set of units called Decision-making units (DMU) by constructing a border enveloping all the possible combinations of inputs and outputs for each DMU. Two basic models have been proposed as part of the DEA method. The first is the CCR model (Charnes et al., 1978), which assumes that the units operate under constant returns to scale (CRS). The second is the BCC model (Banker et al., 1984) which assumes variable returns to scale (VRS). The basic DEA CCR model is appropriate when the units operate at the optimal size and in a perfect competition environment. It is focused on an input orientation, which means the minimization of inputs for a given level of outputs, and on the assumption of constant returns to scale (CRS). This model is unable to provide information on the extent to which the identified inefficiency may be due to technical or scale inefficiency. This is why Banker et al. (1984) proposed the BCC model to extend the initial CCR model by adopting the hypothesis of variable returns to scale (VRS). Thus, this allows the efficiency to be decomposed in a technical part due to the scale. The study considers a set of n DMUs that consume m input to produce s outputs. Based on the rating proposed by Johnes (2004), the technical efficiency of a DMU k, as defined by Charnes et al. (1978), is measured by the ratio between the weighted sum of the outputs and that of the inputs: The technical efficiency of each DMU is maximized under certain conditions (Johnes, 2004). Firstly, the weights of the outputs and inputs of the DMU k cannot generate an efficiency score greater than 1 (equation 3). Secondly, the weights applied to outputs and inputs are strictly positive (equation 4). For each DMU, the following linear programming problem has to be solved by maximizing the ratio TE_k such as: ???? ?? = ? ?? ?? ?? ???? ?? ??=1 ? ?? ?? ?? ???? ?? ??=1(1)MAX ? ?? ð??"ð??" ?? ð??"ð??"ð??"ð??" ð??"ð??" ð??"ð??"=?? ? ?? ?? ?? ??ð??"ð??" ?? ??=??(2) Under constraints: ? ?? ?? ?? ???? ?? ??=1 ? ?? ?? ?? ???? ?? ??=1 ? 1 ?? = 1, ? , ??(3) ?? ?? , ?? ?? > 0 ? ?? = 1, ? , ?? ; ?? = 1, ? , ??(4) Where: k: DMU for which efficiency is measured; j: The DMUs studied. Two approaches are possible to solve this linear programming problem. The input-oriented model, where the weighted sum of the inputs is minimized by keeping the outputs constant, and the output-oriented model, which will be the focus of this study, where the weighted sum of the outputs is maximized while maintaining constant the inputs. Thus, the primal equation for the outputoriented VRS model to be used in this research is presented below. It represents the multiplier form of the problem to be solved. # ?????? ? ?? ?? ?? ??=1 ?? ???? ? ?? ?? (5) Under constraints: ? ?? ?? ?? ??=1 ?? ???? ? ? ?? ?? ?? ??=1 ?? ???? ? ?? ?? ? 0 (6) ? ?? ?? ?? ??=1 ?? ???? = 1 (7) ?? ?? , ?? ?? > 0 ? ?? = 1, ? , ?? ; ?? = 1, ? , ??(8) Where v_(i ): The weighting coefficient of each input. ?????? ? ?? (9) Under constraints: ? ?? ?? ???? ? ? ?? ?? ?? ???? ? 0 ?? = 1, ? , ?? ?? ?? =1 (10) ?? ???? ? ? ?? ?? ?? ???? ? 0 ?? = 1, ? , ?? ?? ?? =1(11) ? ?? ?? = 1 ?? ?? =1(12)?? ?? ? 0 ??? = 1, ? , ??(13) Where 1/?_k: The technical efficiency score; ?_j? The weighting coefficients which measure the capacity of each DMU (j) to constitute the benchmark. The VRS model also helps to identify the nature of returns to scale (increasing or decreasing returns to scale) and the number of outputs that can be produced by each DMU using the same level of inputs. Thus, based on the DEA model just described, we used the DEA method in the first stage to assess the efficiency of 109 Moroccan courts during the period of implementation of the judicial reform charter between 2013 and 2018. The choice of this method is justified by its many advantages, including the fact that it is The duality rule in linear programming can be used to rewrite an equivalent form called "wrapped form". This is generally preferred since it contains only s+m constraints instead of n+1 of the multiplier form. Thus, the dual formula of the output-oriented VRS model is written: appropriate for public organizations and does not require price information, as well as the relaxed assumptions on input-output data distribution and the non-specification of the technological frontier (Cooper et al., 2001). Our research covers nearly all Moroccan courts, including 21 appeal courts, 68 first instance courts, 9 administrative courts, and 11 commercial courts. We excluded the Supreme Court because it is a law court that cannot be compared to other courts, as well as one first instance court whose input and output values are extreme and cannot be compared to other courts. We used the DEA under output orientation method during the first stage analysis because Moroccan courts are plagued by issues related to a large stock of pending cases and slow processing. The DEA BCC model was used with the VRS assumption. This model is more appropriate for the courts' case because it is a monopoly system that is difficult to operate on an optimal scale. # b) Two stage regression model The literature refers to second-phase or "twostep" studies that link efficiency scores to explanatory variables. These variables are generally considered as exogenous to the entity's production technology and are not directly related to the inputs and outputs used to calculate efficiency. Therefore, in the second stage of analysis, we identified a set of exogenous variables based on previous research findings, then classified these variables into three categories: those related to judicial activity, those related to judge's clerk's profiles, and finally variables related to socioeconomic and demographic factors. There is no agreement in the literature on the best regression model to use in this case. Indeed, McCarty and Yaisawarng (1993) argue that a Tobit model is more appropriate because the efficiency scores are capped at 1, whereas Hoff (2007) believes that an OLS regression is sufficient. McDonald (2009) considers that the Tobit regression may be inappropriate and that the OLS regression gives more consistent results. More recently, Banker and Natarajan (2008) proposed a model with a less restrictive formbased mainly on the use of the ordinary least squares (OLS) method. Thus, we performed the statistical tests for residuals normality and heteroskedasticity, then we opted for ordinary least squares regression to identify the variables that influence the court's efficiency. Three models are used, one with the technical efficiency CRS as the dependent variable, another with the efficiency under the VRS assumption, and the last with the scale efficiency variable, while all 20 explanatory variables are used in all three models. This decision is motivated by our intention to find disparities in the influence of the explanatory variables on each of the efficiency types examined. # c) Data All data about the judicial system came from the Moroccan Ministry of Justice, while data about exogenous variables came from official state reports and documents. There is no consensus in the literature regarding an ideal model for selecting the input and output variables used in the DEA model (Johnes, 2006b;Avkiran, 2001). Therefore, we used inputs and outputs that are frequently used in literature and that appear to us to be the most representative of the Moroccan court production process. # ? The inputs: ? Nb_judge: Number of judges; ? Nb_clerk: Number of clerks; ? F_cost: Court operating expenses. ? The outputs: ? Cases resolved: Number of cases resolved or number of judgments rendered. For the selection of exogenous variables, we consider the literature and the availability of data to integrate 20 independent variables into our model as follows: ? Dependent variables: ? EFF_CRS: Efficiency under CRS assumption; ? EFF_VRS: Efficiency under VRS assumption; ? EFF_S: Scale efficiency. # ? Independent variables: Variables related to judicial activity ? FEM_CLERK proportion of female clerks. # Variables linked to socio-economic and demographic factors: ? LOGPOP Log of the number of inhabitants in the perimeter of the court; ? GDP_H GDP per capita in the city where the court is located; ? D_SDR The kilometer distance between the court and the regional administrative entity in charge of managing the administrative, logistical, and human resources of the courts; ? North, South, Center, or East: location of the court in the country; ? DIP_POP proportion of the population with a higher education level; ? INC_POP Income of the region's population. Table 1 shows the descriptive statistics for the input and output variables, the efficiency scores obtained in the first stage of analysis, and the statistics for the exogenous variables, which are the independent variables in our model. The period of implementation of the judicial reform charter has seen an upward trend in the means and productivity of the courts. Thus, in 2018, the data shows an average of 35 judges and 109 clerks per court. On the other hand, operating expenses vary significantly, with a minimum of 6.7 million dirhams and a maximum amount of 112.8 million dirhams. On the other hand, the statistics of exogenous variables, show a wide range of court sizes, whether in terms of cases (from 1675 to 188936) or staff (30 to 467). The same observation applies to pending cases. The average is around 5,245 cases, with a maximum of 35,972 pending cases. Our model also includes seven dummy variables with mean values that are frequently less than 0.5. IV. # Results and Discussion The use of a two-stage approach allowed in a first stage to measure the court's efficiency between 2013 and 2018 in order to assess the impact generated by the implementation of the judicial reform on productivity. Secondly, the use of the OLS regression allowed to test the influence of exogenous variables on the court's efficiency in 2018, and thus to identify the determinants that allow courts to become more efficient in the context of reform. We present in Table 2 the summary statistics of the efficiency results under the assumption of constant and variable returns to scale. The results by the court groups are presented in Appendix 1. The results have shown that the average efficiency level is relatively low in 2018, with 56.2% under the CRS assumption and 72.4% under the VRS assumption. Thus, the Moroccan courts have a wide margin to improve their efficiency. Regarding the results by court type, first instance courts are the least technically efficient, and the Appeal courts are the most efficient. Our results also show that 19 courts had CRS efficiency scores between 9.9% and 50%, compared to only six fully efficient courts. The same results allow us to assess the nature of returns to scale. Thus, most courts operate under increasing scales, which means that they can further increase their productivity by taking advantage of economies of scale. Administrative courts are the exception, with a majority under decreasing returns to scale, implying the need to divide these courts into several entities. For the second stage, we present in Table 3 the results of the estimation of the analysis models, and in Appendix 2 the details results. The results show that the value of the statistic R2 for the three models relating to the CRS efficiency, the VRS efficiency, and the scale efficiency is 0.6402, 0.4337, and 0.6501, respectively, indicating a good quality of the econometric model since these values tend towards unity, except for the value of the model relating to the VRS efficiency, which is relatively low. Nonetheless, we can confirm that there is at least one independent variable that contributes to the variation in Moroccan court efficiency. According to the study's findings, CRS efficiency is higher in courts that serve densely populated areas, as well as in courts with the greatest number of experienced judges and pending cases. The least efficient courts, on the other hand, are distinguished by a large number of pending cases and an overabundance of judicial personnel. On the other hand, the VRS technical efficiency of courts increases for Appeal courts and when the number of cases pending is large and decreases when the number of judicial staff is excessive. Finally, the scale efficiency associated with the evolution of the court in an adequate production scale increases with the number of cases in process and the population covered by the court's perimeter and decreases for Appeal courts and specialized ones. These findings have several managerial implications for Moroccan court administrators. First, the fact that an increase in the number of cases in progress improves court efficiency demonstrates that the increased workload induced by the cases allows the court to benefit from economies of scale to resolve more cases with fewer resources. This implies that the courts must maintain a high level of activity and not be divided or replaced by new courts. However, it should be noted that we must continue to monitor returns to scale. Indeed, once a court is subject to decreasing returns to scale, we can no longer improve its efficiency by increasing the number of cases handled. Furthermore, the positive effect induced by the presence of a high proportion of experienced judges demonstrates a long-standing mistake made by Moroccan court managers, who tend to cover the needs of overdue courts with recruits. Indeed, it has been observed that the more experienced a judge becomes, the more he gains stability by only practicing in courts of the same judicial district. As a result, whenever a court experiences problems related to overwork or an increase in judgment delays, it becomes a priority when assigning new judges recruited. As a result of our findings, there is a need to change the type of decisions made by assigning more experienced judges to overdue courts, either as part of a temporary delegation of 3 or 6 months or by allowing judges to transfer from other courts. The negative effect on efficiency caused by the number of judges and clerks, on the other hand, implies that all courts must consider the practices of their benchmark within the framework of the reference groups identified during the first stage analysis. Although the staff is not always interchangeable, using the same number of judges and clerks as the reference group's best practice court will almost certainly allow for an increase in the number of cases resolved. In the same context, the result regarding the negative impact of pending cases on court efficiency, which is consistent with previous studies (Marselli and Vannini, 2004;Ferro et al., 2018;Lewin et al., 1982;Schneider, 2005;Castro and Guccio, 2016), demonstrates the need for Moroccan courts to act on these cases by gradually reducing them, rather than using this as an argument for an excessive increase in the resources mobilized. Finally, it was found that court specialization (administrative or commercial) is a determinant of inefficiency. Even though this result contradicts some previous study findings (Gorman & Ruggiero, 2009), it illustrates that in the Moroccan context, specialization leads to the use of more means to resolve fewer cases than general courts. This implies that the Kingdom should stop establishing new commercial or administrative courts. The appropriate managerial solution would thus be to create specialized administrative structures at the level of the already existing appeal courts and first instance courts, rather than to build new buildings, which would incur additional costs. Our findings on efficiency determinants provide us with guidelines to follow in order to improve court efficiency. This primarily entails considering our recommendations when developing the judicial map, determining the courts' human and logistical resources, and allocating these resources. V. # Conclusion This study aimed to identify the efficiency determinants of Moroccan courts during the implementation of the judicial reform charter. Thus, we applied the analysis approach in two stages. First, we assessed the court's efficiency. Second, we investigated the impact of exogenous variables on the obtained efficiency scores. This study filled a gap identified in our literature review since very few studies have attempted to evaluate judicial efficiency and identify its determinants in African countries. So the results provide managerial implications for court managers and can provide recommendations for decision-making. In this context, the results showed that the average efficiency of Moroccan courts is relatively low despite an improvement during the period of implementation of the reform. In addition, the majority of courts can use economies of scale to resolve a larger number of cases, reducing the stock of pending cases and adjudication delays. On the other hand, the results of the second stage analysis clearly show that the efficiency of Moroccan courts is positively influenced by the size of the court in terms of cases, by the presence of experienced judges and it increases in courts located in the most populous cities. On the other hand, efficiency decreases when the number of pending cases is excessive, as well as when the number of judicial staff is excessive. These findings are important for understanding judicial efficiency in Morocco and assisting managers in making decisions. However, we believe that our research has some limitations that require further research. Among these limitations are the difficulty of selecting all of the courts' inputs and outputs in measuring efficiency, as well as the non-integration of a set of exogenous variables due to the lack of certain data. As a result, we believe that it would be interesting to include other outputs in future research, such as prejudicial cases resolved in court using alternative methods. We also recommend experimenting with different methods of measuring efficiency, as well as different regression models, such as Tobit's truncated regression. It is also possible to include other exogenous variables in the regression model, such as remuneration, employee motivation, or the rate of court computerization. 1MeanStd devMinMax 2MeanStd devMinMaxEFF_CRS (CCR Model)0,5620,2250,0991,000EFF_VRS (BCC model)0,7240,2230,2061,000EFF_S Scale0,7900,2170,1921,000 3VARIABLESEff_CRSEff_VRSEff_SSIZE_CASES5.68e-06***5.08e-06***1.39e-06*SIZE_STAFF-0.000992***-0.00121***-8.61e-05PEND_CASE-1.10e-05**-6.79e-06-6.13e-06PRP_JUDGE0.6010.7220.0590PRP_CRIM0.1410.1570.0883SPECIA-0.433***-0.0643-0.488***TYPE-0.06750.198**-0.314***EXP_JUDGE0.323***0.2080.154FEM_JUDGE-0.120-0.1590.0966DEG_CLERK-0.06550.179-0.224EXP_CLERK0.000163-0.0009260.00673FEM_CLERK0.337-0.1280.480**lOGPOP0.323***0.1170.345***GDP_H1.04e-07-2.52e-061.35e-06D_SDR0.0003830.0003443.25e-05North0.0659-0.1100.163South-0.0145-0.0352-0.0295Center0.0590-0.08630.121 © 2021 Global JournalsDeterminants of Judicial Efficiency in Morocco * Investigating technical and scale efficiencies of Australian universities through data envelopment analysis NAvkiran Socio-Economic Planning Sciences 35 1 2001 * Some models for estimating technical and scale inefficiencies in Data Envelopment Analysis RBanker ACharnes Cooper W Management Science 30 9 1984 * Evaluating Contextual Variables Affecting Productivity Using Data Envelopment Analysis RDBanker RNatarajan Operations Research 56 1 2008 * Judges and court performance: a case study of district commercial courts in poland JBeldowski LDabros WWojciechowski European Journal of Law and Economics 50 2020 * Aging and Productivity Among Judges: Some Empirical Evidence from the High Court of Australia MBhattacharya RSmyth Australian Economic Papers 40 2 2001 * Measuring Potential Efficiency Gains from Mergers of Italian First Instance Courts through Nonparametric Model MFCastro CGuccio Public Finance Review 46 1 2016 * Measuring the efficiency of decision making units ACharnes WCooper ERhodes European Journal of Operational Research 2 1978 * Data envelopment analysis: A comprehensive text with models, applications, references and deasolver software WWCooper LMSeiford KTone Journal-operational Research Society 52 12 2001 * Judicial Incentives and Performance at Lower Courts: Evidence from Slovenian Judge-Level Data VDimitrova-Grajzl PGrajzl KZajc JSustersic Review of Law & Economics 8 1 2012 * Assessing judicial efficiency of Egyptian first instance courts: A DEA analysis NElbialy MAGarcia-Rubio Joint Discussion Paper Series in Economics 30 2011 * Judicial productivity, delay and efficiency: A Directional Distance Function (DDF) approach GFalavigna RIpi AManello GBRamello European Journal of Operational Research 240 2 2015 * Determinants of Judicial Efficiency Change: Evidence from Brazil TAFauvrelle ATony C Almeida Review of Law & Economics 14 1 2018 * Efficient courts? a frontier performance assessment GFerro CARomero ERomero-Gomez Benchmarking: An International Journal 25 9 2018 * Searching for the source of technical inefficiency in Italian judicial districts: an empirical investigation FinocchiaroCastro MGuccio C European Journal of Law and Economics 38 3 2014 * Assessing judicial efficiency of Egyptian first instance courts: A DEA analysis MAGarcia-Rubio NElbialy Joint Discussion Paper Series in Economics 30 2011 * Evaluating U.S. judicial district prosecutor performance using DEA: are disadvantaged counties more inefficient? MFGorman JRuggiero European Journal of Law and Economics 27 3 2009 * Second stage DEA: Comparison of approaches for modelling the DEA score AHoff European Journal of Operational Research 181 1 2007 * Efficiency measurement. In International Handbook on the Economics of Education JJohnes 2004 Edward Elgar Publishing Ltd Cheltenham * Data envelopment analysis and its application to the measurement of efficiency in higher education JJohnes Economics of Education Review 25 3 2006a * Measuring teaching efficiency in higher education: An application of data envelopment analysis to economics graduates from uk universities 1993 JJohnes European Journal of Operational Research 174 1 2006b * Evaluating the administrative efficiency of courts AYLewin RCMorey TJCook Omega 10 4 1982 * RMarselli MAndvannini 2004 * L'efficienzatechnica Dea DistrettiDi Corte D'appelloItaliani 2004/09 Aspetti Metodologici 42 Benchmarking E Arretrato Smaltibile. CRENOS Working Paper * Technical Efficiency in New Jersey School Districts TMccarty SYaisawarng the Measurement of Productive Efficiency: Techniques and Applications SSHo Schmidt 1993 * Using least squares and tobit in second stage dea efficiency analyses JMcdonald European journal of operational research 197 2 2009 * Economics of court performance: an empirical analysis VRosales-Lopez European Journal of Law and Economics 25 3 2008 * On the need for reform of the Portuguese judicial system Does Data Envelopment Analysis assessment support it? SPSantos CAAmado Omega 47 2014 * Judicial career incentives and court performance: an empirical study of the German labour courts of appeal MRSchneider European Journal of Law and Economics 20 2 2005 * Efficiency Estimates For Judicial Services In Brazil: Nonparametric FDH (Free Disposal Hull) And The Expected Order-M Efficiency Scores For Rio Grande Do Sul Courts. Encontro da anpec SBSchwengber MSousa 2005 33 * Determinants of judicial efficiency: a survey SVoigt European Journal of Law and Economics 42 2 2016 * Measuring efficiency of Brazilian courts with data envelopment analysis (DEA) LLYeung PFAzevedo IMA Journal of Management Mathematics 22 4 2011