The weighting of cases is a mechanism used to assess the nature and complexity of a judicial proceeding. The most common way to assess the caseload in a particular court is by using the same unit and calibration to consider all cases before that court, even if each case has its own complexity.

Sometimes, in the litigation category, the methodology adopted by court managers is to distribute work between all the judges without particular attention to the complexity of the case. Nevertheless, it is apparent that a more structured and in depth approach would be useful in order to ensure an objective and effective evaluation of the cases, which should then be shared between all the stakeholders.

As to definitions, the term “workload” must be distinguished from “caseload”: “workload” relates to all the range of the activities of a judge (disposal of cases but also training, meetings, professional activities and any other duties); “caseload”, on the other hand, relates only about to the activity carried out by a court linked solely to the disposal of a case (mainly preparation, hearings, drafting, juridical researches). These definitions are very important as any ‘case weighting system’ could deal with both the “caseload” as well as the “workload”.

The weighting of cases necessarily requires that a clear set of ‘case categories’ be set up, within which sub-categories different cases may be posted for precise evaluation purposes, as without such categories, there would be no pertinent structure which may distinguish between the different type of cases being weighted accordingly (a payment order is simpler to solve than a case about medical responsibility due to the need of external expertise, for instance).

It is also important to understand what the objective of the evaluation of the weight of a case. According to the different experiences provided by different member States of the Council of Europe as part of this study, such systems are used to evaluate:

A. The proper and equitable distribution of cases among judges

B. The expected duration of a case and its hearings

C. The type and amount of resources to be allocated (especially human resources)

Apart from this, a properly set-up ‘case-weighting system’ could also allow the development of more refined statistics to build a more qualitative view of the caseload.

Nevertheless, on the basis of the information gathered by the SATURN Centre, it is clear that there exist certain differences in view between court managers and judges as to the merit or value of such an evaluation, particularly in an automated manner.

The SATURN Steering Group notes that it is very important that before commencing any exercise to introduce ‘case-weighting’, the objectives of this exercise must be clearly spelt out. This is being said since, though the ‘case weighting system’ may be used for the any of the purposes above-mentioned (such as case allocation, evaluation of duration of case and allocation of resources), it may have other uses and objectives, such as evaluation of the activity of judges and other court staff, which may, by their nature, be controversial, as a result of which, it is imperative that the objectives be clearly shared with all the stakeholders from the beginning of the process.

The objective of this SATURN study is to find an approach accepted by all professionals.

1.2 - Summary of case-weighting practices

As part of the study carried out by the SATURN Centre for judicial time management, various Member States provided information as to whether ‘case-weighting systems’ are in place in their State. Furthermore, information was also collected during the workshop with the Network of Pilot Courts (September 2017) and a court coaching activity (Malta, February 2018).

The goal is not to provide a comprehensive and extensive analysis of all the methodologies available, but a summary of relevant practices that reveal significant incentives, benefits and issues encountered by judicial systems.

In Austria (Annex VII), a system is in place which ensures a fair distribution of cases between judges. A survey dealing with ‘working hours’ is sent out to the judges. It asks the judges to log their total working hours for a period of six months. Once this period lapses and all the results are in, the system calculates the personnel requirements for the judiciary. The latter is done by dividing the time spent on a particular type of case by the number of cases that were processed. The resulting time value would later then bring about the number of cases pending to be processed. This data gathered would then be recorded in an excel sheet set up by a specialized Court Officer and with the results generated he analyses the new cases. This worksheet contains 40 case categories and carries out a calculation (‘case’ x ‘theoretical time / judges available’). This system was evaluated and assessed positively by a Swiss consultant, who obtained the same results by calculating the time actually consumed by a sample of judges. This measurement tool has now been in use for 10 years throughout the country in order to assess the forecast workload (new cases) and is adapted to calculate the actual workload (resolved cases). To date, the tool is only partially used for the allocation of staff resources. It is reviewed regularly to remain effective.

In Bosnia-Herzegovina (Annex I), a study (“Backlog Reduction Programmes and Weighted Caseload Methods for South East Europe, Two Comparative Inquiries”) has been conducted to identify the benefits and issues of the Delphi method[1]. This study also addresses experiences from Albania, Kosovo*, Montenegro, The Former Yugoslav Republic of Macedonia and Serbia. According to the authors of this study, the Delphi method seems less burdensome, less expensive, and can be completed in less time than the time study method. They discuss recent experiences in Bosnia and Herzegovina and Serbia, which “highlight the difficulties in developing and implementing a case weighting system via the time study approach. In contrast, experiences in Kosovo* demonstrate that credible case weighting systems can be developed, and implemented, based upon the Delphi method”. They underline that “A key to the development of any case weighting system, based on either the Delphi or time study method, is obtaining commitment and buy-in from key stakeholders, including judges, judicial council members, and the Ministry of Justice. This necessitates a clear articulation of the intended uses of the case weighting system and transparency in the process used to develop the case weights.” Now, in Albania and Kosovo* an ongoing CEPEJ project is also being implemented (see Annex III).

In Croatia (Annex I), there exists a clear and well established framework which is imposed on judges. A very precise categorization of cases (165) is provided and a definition of a minimum annual number of cases to be dealt with per judge is established, depending on the complexity of each case category. Between 2004 and today, the productivity results have improved considerable, as the overall number of pending cases has been reduced from 1.6 million to 400,000. Nevertheless, these standards set out per judge do have certain defects as, in certain case categories; the objectives are easily attainable while in other case categories, the objectives have never been attained. In Slovakia (Annex I), a working group of judges was set up in 2011 and produced a framework system comparable to that of Croatia.

In the Czech Republic (Annex IV), targets are also set for judges, but these are non-binding. These are based on the number of cases handled in the previous three years. In 2017, the estimated number of cases to be handled for 2018 was calculated by adding the number of cases handled in 2016 (number of cases solved x 0.5), 2015 (number of cases solved x 0.25) and 2014 (number of cases solved x 0.25). During a CEPEJ workshop in Malta, a scientific expert from the Czech Republic provided other possible methodologies, designed on an experimental basis.

In Denmark (Annex I), the existing system weighs cases not only for the courts but also for legal aid, as lawyers are paid according to the complexity of the procedure they participate in.

In Estonia (Annex II), an experiment was led by the President of the Administrative Court of Appeal. Here, the judges were requested to rank, during a reference year, the complexity of the cases they were dealing with from 0 to 40. After much discussion, a ranking system was created. The latter also took into account certain predetermined factors of complexity.

In Finland (Annex I), a questionnaire is regularly sent to judges to assess their workload.

In France (Annex II), a large consultation process was carried out in 2014 between all the stakeholders (MoJ, Inspection of justice, representatives of unions of judges, representatives of courts). The objective was identifying some relevant indicators and proposing a time measurement evaluation for each task/activity. Each indicator (per function of judge and type of litigation) was discussed after an exercise of quantitative measurement was carried out (number of cases disposed per FTE on a basis of 210 working days). A final vote between all the members of the working group led to the publication of a progress report which revealed discrepancies in the calculations being made. However, the work is not definitive and still seems to be ongoing.

In Germany[2], since the late 1990s, some Federal States entrusted the (former) accounting firm Arthur Andersen with the elaboration of a modern system for the evaluation of the performance of the judiciary. It was intended to enable the ministries of justice to implement effective personnel planning. The new system, called “Pebbsy”, was presented in 2002. It is now being implemented by the judicial administration in the 16 different Länder and has been adapted to their respective court structures. The system is based upon much factual research which was conducted in 7 Länder. When it was set up, about 1900 judges and prosecutors in more than 40 courts and prosecutors’ offices had to document the number of cases and the time spent to finish the cases. As a result, the study quantified the average working time of a judge in several categories. Under the old “Pensen”-system all civil proceedings of the local courts had been treated equally. Based on the study, the Pebbsy system distinguishes the following categories (considering simple, average and difficult cases as a whole):

- Neighbourhood disputes: average working time: 290 mins.

- Construction disputes: average working time: 260 mins.

- Traffic accident: average working time: 220 mins.

- Rental disputes between landlords and tenants: average working time: 170 mins.

- Other civil matters: average working time: 150 mins.

Based on this categorisation, each Länder calculates the annual workload of its judicial staff separately (Working conditions are different in each federal entity).. In addition, Pebbsy and the general introduction of information systems in the courts also have allowed broad data collection and, therefore, a precise evaluation of the efficiency of each court and even of each single judge. As a result, the transparency of the performance of the judiciary (“Erledigungszahlen”) has been largely improved. The data is provided by each single judicial unit and collected and processed in the ministries of justice. In the Federal State of Baden-Württemberg, the data is available for the presiding committees of the courts.

However, there is a large consensus that direct control of the operation of the judges is not to be allowed, because this control would infringe upon the independence of the judges. Accordingly, the system only serves as a point of reference for the personnel planning of the judiciary. However, it is also used as guidance for allocating the annual caseload within the courts (between the judges, chambers and senates).

There was also discussion on the application of bench marking techniques within the courts. In some Länder bench marking committees have been set up at the local and district court levels. These committees compare the handling of specific proceedings (i.e. assessment of costs) in order to find out best practices. Such techniques are currently applied to the judicial administrative staff. The general opinion is that a broad application of bench marking in the judiciary would pre-suppose a clear separation of the administration of the judiciary from the ministries of justice. Any supervision of justice should be effected by (independent) judicial councils. In general political discussions, however, these proposals are still met with reluctance.

In The Netherlands (Annex I), a case-weighting system was set up which categorises cases into four different categories, ranging from scale A (simple) to scale D (complex). To date, this mechanism is only effective in 3 of the Courts of First Instance and 3 Courts of Appeal. Its implementation makes it possible to determine the provisional hearing time for a case and to ensure a balanced assignment of cases between judges. However, the system is criticised for its lack of transparency and the absence of strong rules. The digitalisation of case processing (zero paper workflow, in progress) is likely to change this system.

In Norway (annex VIII), the Norwegian Courts Administration use a case management system called LOVISA. Data of this system is analysed by means of statistical methodologies set up to estimate the number of staff (judges and clerks) needed in each District Court and each Appeal Court. This information is then taken into consideration when discussing changes in the number of staff, if, for instance, a court applies for more resources. The model is based on a ‘regression analysis’, and evaluates the relationship between the number of cases of different categories, significant characteristics of the cases of each category (number of witnesses, use of interpreters, use of experts etc.) and the amount of time spent by judges and clerks to handle the cases. In the model, each case type is given a particular weight according to a stipulated number of witnesses, experts etc. The stipulation is based on a time study done in 2006. The input data used when carrying out the annual study in order to make the necessary calculations is based on the previous year’s statistics from the court, collected through the case management system.

In Serbia (annex IV), a case weighting methodology has been implemented in 2017. Here, data samples were used from 1000 case files, obtained from 10 First Instance and High pilot courts, which wasextracted from the electronic case management system. The methodology adopted classifies first instance cases from the abovementioned courts and it calculates the time used in the judicial process by examining different elements such as the number of witnesses, the duration of the proceedings etc. The result of the formula is used later to adjust case assignments and to reach an equitable balance workload among the judges. Data necessary to similarly classify second instances cases in the Higher Courts, however, was insufficient.

In Israel (CEPEJ Observer State, Annex VI), being an observer state, the implementation policies of a ‘case-weighting system’ are very relevant to analyse because they have adopted various case-weighting methodologies along the years, mostly inspired from practices imported from some courts in the United States.

One such methodology is the ‘Event-based case weighting methodology’, where the event frequency and event complexity are the key elements to take into consideration. The event frequency of the event assesses how often a specific event (a stage in the proceedings) is likely to occur, while the event complexity assesses the average amount of judicial time spent on a specific event.

Apart from this system, another methodology known as the ‘Delphi method’ is also used in Israel. When setting up this methodology, 51 judges were observed on an average panel of cases that were handled over a period of 5 years. The Time spent on judicial work was eventually estimated for each type of case. This Judicial time was measured from the first to the last event in the case, e.g. reading the case, preparing the hearing, conducting the trial and preparing and giving the judgment. The evaluation process considered the average of the whole Court for a specific category of cases, and not just a particular judge. The number of cases resolved by judgment and by settlement was also factored in. Measuring the case weightings for each court and eventually averaging the results based on the number of judges in the court helped “even out” the results and provide a more workable scenario.

Though set up, the process is an on-going one as the judges' perception of the reality of their work is sometimes different. Plans are now underway to extend the number of Judges observed to refine the evaluations..

1.3 - What are the Pros and Cons of the existing systems?

There appear to be two main approaches to setting up a case weighting system:

· A time based approach case weighting system

· A points based approach case weighting system

The time based approach aims to measure average time spent by a judge to deal with their tasks. This measurement was done in Israel with an anonymous questionnaire where a sample of judges had described their daily activities. This approach is also used in Austria (PAR System).

The utility of this approach is that it avoids identifying all the elements needed to assess the complexity of a case (see below) and it reaches its conclusions from the observations collected from a significant sample of time measurement. Nevertheless, this is a rather ‘subjective’ approach which could be subject to personal bias of the various contributors. For the Israeli experts, these biases could be reduced by coupling this exercise with a significant and detailed data collection.

The points based approach, on the other hand, could be a more ‘objective’ approach as it consists of allocating an index of complexity to a case based on quantifiable criteria. This index could be obtained in a generic manner, by agreement reached amongst the various stakeholders (In Estonia, for example, judges were requested to rank the complexity of their cases from 0 to 40 and, following much debate, a ranking was drawn up) or in a more precise manner where certain predetermined factors of complexity are taken into consideration.

The elements identified by the SATURN Centre to create such an index of complexity, though not exhaustive, would include the following:

a) Nature of the claim and type of proceeding

b) Claim monetary value

c) Number of claims/complexity of motions/line of defence

d) Number of parties/defendants/lawyers

e) Number of witnesses

f) Amount and range of expertise / Number of experts

g) Number of foreseen sittings predetermined audiences / hearings

h) Whether the case is an International / cross border case or not

i) Whether Interpretation / translation is needed

j) Whether the case raises a new issue of law not addressed in existing caselaw

The point based approach, though more objective at first sight, does however present difficulties in arriving at a consensus on the weighting to be applied to each of the factors aforementioned.

Nevertheless, it is clear that whatever approach is adopted, a constant evaluation and update seems necessary to ensure maximum compatibility to the reality of the evolution of the activity of the court.

Part 2: Proposal for a core of a weighting methodology for judicial cases

Following on from the experiences described in the first part of this document, the SATURN Centre has endeavoured to draw up a synthesis of the various practices to form the core of a common weighting system for European courts.

The following suggestions do not aim to eventually form part of a mandatory process to be implemented in all courts, but rather intend to provide a common flexible framework within a court or a cluster of courts.

After the definition of a possible methodology, the SATURN Centre will make some suggestion as to the deployment of this tool to reach an agreement between all the stakeholders.

2.1 - Different stages of the assessment of the weight of a case

According to the different objectives to be reached, the weighting of cases may be carried out at different stages of a case, in a defined instance (1^st instance, 2^nd instance or Supreme Court).

Main objectives of the case-weighting

Equitable distribution of cases between judges

Estimates the forthcoming caseload (to allocate proper means)

Calculation of caseload in real time[3]

Predict the evolution of the pending cases or the backlog (according to the human means allocated at a moment) and implement early warning signals in CMS

Calculation of the caseload handled by a court / judge

Could help to feed a database that contributes to the assessment in stages S1 and S2

When to operate case weight?

S1 - When the case is registered (incoming cases)

S2 - When the case is processed (continuous assessment)

S3 - When the case is disposed or resolved

2.2 - How to calculate the weight of a case?

2.2.1. Definition of case categories or of litigation categories

The necessity of setting up clear case categories is mandatory in order to adapt the method of calculation (see below) to each type of litigation (civil, criminal or administrative cases).

In order to provide a better description of each category, the definition of sub-categories may be set up in accordance to relevant articles of law or any other criteria specific to the matter concerned, this being able to provide a more precise description of a category. Indeed, a dispute about contractual liability in general (civil case) could be very simple or very complex according to the subject of the contract. A criminal case relating to homicide could also be very simple if the criminal confessed the facts and very complex if the facts are denied and there are a lot of witnesses to hear.

Regardlessof the approach taken, a synoptic table is a good tool for classifying cases according to their predictable complexity.

Type of litigation	Sub-category	Average/median Time or Complexity index
Civil	Simple contractual liability	1
Civil	Complex contractual liability	2,5
Criminal	Facts confessed	1
Criminal	Facts denied	2
…	…	…

2.2.2. Method of calculation

a) Event and time based measurement

The ‘event frequency’ and ‘event complexity’ are the key elements to take into consideration per each case category:

· Event frequency: how often a specific event is likely to occur

· Event complexity: average amount of judicial time spent on a specific event. Judicial time has to be calculated for all the different events in a case (e.g. reading the case, preparing the hearing, conducting the trial and preparing and giving the judgment).

One of the possible methods for the measurement process is to conduct an anonymous questionnaire with a first representative panel of judges to calculate the average (or median) time spent for the execution of each identified event within each category/subcategory. Depending on the applicable procedures, a distinction can also be made by mode of resolution (judgment, settlement for example). The addition of these average or median times of events by category / sub-category produces an estimate that can be used to classify new cases. Each judge can thus receive a balanced number of cases based not on the absolute number of cases (10 new cases per judge) but on the sum of the provisional times (90 working hours per judge).

The panel of judges involved in conducting the anonymous questionnaire can gradually be extended in order to regularly adapt the estimates to the reality of litigation processing and erase the subjectivity of individual time estimates.

b) Point based measurement

Two main approaches could be considered:

· A simple scale of complexity where some levels are set up, ranging for example from A to D, where ‘A’ contained the simple cases and ‘D’ contained the most complex cases;

· A more precise scoring system based upon some quantitative criteria, for instance:

o 1 point per party

o 1 point per claim x 1 to 5 (ranging from ‘1’ for simple legal problem to ‘5’ for a complex legal problem) x 1 to 5 (ranging from ‘1’ for a small financial amount to ‘5’ for a large financial amount)

o 2 points per expertise needed

o 0,5 per eventual pre-hearing x 1 to 5 (1=short predicable length for hearings to come / 5=long predicable length for hearings to come)

2.3 - How to drive change in a court to implement a case-weighting system?

[This part should be developed with the main findings of the workshop to be held with CEPEJ Pilot Courts (October 2018)]

· Managing the deployment in project mode: Create a steering committee inside the court to choose the method, to experiment and deploy the system incrementally, with an interdisciplinary team.

[…]

· Full transparency of the system: Description of the algorithm used to assess the caseload to the judges.

[…]

· Assess the results and refine them: Allow the system to evolve according to the criticism gathered.

[…]

2.4 - Possible future development by using data science

The use of the knowledge gathered with the real weight of the cases when they are solved is an opportunity to improve the accuracy. This information could be stored in a database and then used in previous stages described with a mathematical / statistical / probabilistic approach. When a new case is registered in the electronic case management system, the computer could search comparable resolved cases (number of parties, type of litigation etc.) and provide a more precise estimation.

The data science methodology (use of big data composed by the collection of all the information needed and processing with datamining / artificial intelligence) seems to be fully adapted to that need. In Latvia, such technologies are already used (for instance Microstrategy Business Intelligence Tool).

This approach also would eventually allow to build models to attempt to predict the evolution of pending cases / caseload and forecast the means needed (human resources, budget).

Part 3: Summary of the CEPEJ case weighting methodology

Phase 1: Setting up the system

1. DEFINITION OF CASE CATEGORY - A clear definition of the relevant case categories to be weighted is necessary :

a. Civil, commercial, administrative or criminal cases

b. Sub-categories of these, according to consistent clusters usually allocated to judges: for the civil sub-category, for example family law cases (litigious divorce and non-litigious), contracts, land property, voluntary jurisdiction, etc.

c. Specific proceedings should be considered and separated: short circuits for urgent matters for example, etc.

d. Classification could also be drafted per main category of litigation (e.g. litigious divorce, non-litigious divorce)

2. MEASUREMENT PER CASE CATEGORY

Option 1 – Event and time based measurement

The event frequency and event complexity are the key elements to take into consideration per each case category defined in Phase 1:

a. Event Frequency: how often a specific event (a stage in the proceedings) is likely to occur,

b. Event Complexity: average amount of judicial time spent on a specific event. Judicial time is measured from the first to the last event in the case, e.g. reading the case, preparing the hearing, conducting the trial and preparing and giving the judgment.

Initially, the case weight could be measured for each court by estimating the time needed for each type of case. This evaluation may be done through an anonymous questionnaire where a sample of judges describe their daily activities related to the cases. Then, once the results are collected, they are averaged (average or median) based on the number of judges in the court to standardise the results and allocate to each judge an appropriate number of cases. Once set up, the number of judges involved in the ‘anonymous questionnaire’ process should be increased as the judges' perception of the reality of their work may vary from person to person.

Option 2 – Points based measurement

Two main approaches could be considered:

· A simple scale of complexity, as adopted in the Netherlands, where four scales of complexity were set up, ranging from A to D, where ‘A’ contained the simple cases and ‘D’ contained the most complex cases; or

· A more precise scoring system based upon some quantitative criteria, for instance:

o 1 point per party

o 2 points per expertise needed

o 0,5 per eventual pre-hearing x 1 to 5 (1=short predicable length for hearings to come / 5=long predicable length for hearings to come)

When considering the complexity of a case, other considerations may be given attention when grading the points, such as:

a) Nature of the claim and type of proceeding

b) Claim monetary value

c) Number of claims/complexity of motions/line of defence

d) Number of parties/defendants/lawyers

e) Number of witnesses

f) Amount and range of expertise / Number of experts

g) Number of foreseen sittings predetermined audiences / hearings

h) Whether the case is an International / cross border case or not

i) Whether Interpretation / translation is needed

j) Whether the case raises a new issue of law not addressed in existing caselaw

Such scoring method allows calculating an index that could be considered as the “weight” of this case.

3. MODEL - A table should summarize all the data collected in order to obtain an average or median weight of each precise case category.

Phase 2: Using the system

4. INITIAL ASSESSMENT - The calculation of the case weight can be utilised in order to better assign the cases at their initial stage amongst the Judges available in order to ensure a proper and equitable distribution of work whilst also enabling the managers to better assess the relevance of such system to assess the resources required in a court.

5. REAL TIME ASSESSMENT[4] - The possibility should be considered of calculating the case’ weight for the entirety of the cases’ duration, in accordance with the projected events of this case and it should continue to evolve until the end.

6. FINAL ASSESSMENT - To improve the accuracy of the system, the weight of the case may be assessed very precisely once it is completed, so as to evaluate whether the evaluation set out initially were correct and improve the forecast for future cases.

The methodology of the final assessment in should follow some simple rules to prevent bias, such as the presence of two evaluator minimum (judge / court administrator).

Phase 3: Continuously update the system

7. REGULAR UPDATE - The model of calculation should be regularly revised so that it fits the evolution of the proceedings.

8. LEARN FROM EXPERIENCE - The system should learn from experience. The estimation of the weight of a case when it is solved should feed a database to allow for latter calculation of new cases.

Working process

- Framing of the subject during the 12^th Pilot Courts meeting held in Kristiansand (Norway) on 6 September 2017

- Exchange of experience during a court coaching activity held in Valletta (Malta) on 12 February 2018

- 1^st discussion during the 23^rd SATURN Centre meeting held in Strasbourg (France) on 5 April 2018

- Presentation of the subject during the 30^th plenary meeting of the CEPEJ held in Strasbourg (France) on 27-29 June 2018

- 2^nd discussion during the 24^th SATURN Centre meeting held in Strasbourg (France) on 9-11 October 2018

- Workshop during the 13^th Pilot Courts meeting held in Strasbourg (France) on 10 October 2018

- Presentation and discussion during the XX plenary meeting of the CEPEJ held in Strasbourg (France) on XX/XX/XXXX

Preparatory documents summarised by the CEPEJ Secretariat

- Report of the workshop held in Kristiansand (Norway) on 6 Sept.2017 with CEPEJ pilot courts provided by Noel Rubotham (SATURN member, Ireland)

- Report of the workshop held in Valletta (Malta) on 12 Feb.2018 and chaired by Francesco DEPASQUALE (SATURN member, Malta), with Villem Lapimaa (CEPEJ Expert, Estonia), Jonas Vaclav (CEPEJ Expert, Czech Republic), Gali Aviv (CEPEJ Observer, Israel), Shanee Benkin (CEPEJ Observer, Israel) and with contributions from Joanne BATTISTINO (Ministry of Justice, Malta) and representatives of Maltese Judiciary

- Feedback from cooperation activities in Albania and Kosovo* provided by Jacques Buhler (CEPEJ Expert, Switzerland)

- Presentation of case-weighting approach in Czech Republic provided by Jonas Vaslav (CEPEJ Expert, Czech republic)

- Presentation of case-weighting approach in Serbia provided by Ivana Nincic (SATURN Deputy member, Serbia)

- Presentation of case-weighting methodology by Israel Supreme court

- Presentation of case-weighting approach in Austria provided by Georg Stawa (President of the CEPEJ, Austria)

- Presentation of case-weighting approach in Norway by Mr. Frank Egil Holm and Mr Audun Hognes Berg (CEPEJ Member, Norway)

Bibliography

- Regional Cooperation Council, Backlog Reduction Programmes and Weighted Caseload Methods for South East Europe,Two Comparative Inquiries, 2016.

- CEPEJ, Practical ways of combating delays in the justice system, excessive workloads of judges and case backlogs, 2005.

[TO BE COMPLETED]

List of annexes

Annex I – Summary of the workshop held in Kristiansand (Norway)

Annex II – Summary of the workshop held in Valletta (Malta)

Annex III – Examples of case-weighting provided for CEPEJ Activities (Albania and Kosovo*)

Annex IV – Draft document provided by Mr Jonas Vaclav, CEPEJ Expert (Czech Republic)

Annex V – Presentation of the methodology used in Serbia

Annex VI – Presentation of the methodology used in Israel

Annex VII – Presentation of the methodology used in Austria

Annex VIII – Presentation of the methodology used in Norway

Annex I – Summary of the workshop held in Kristiansand (Norway)

CEPEJ SATURN Working Group and Pilot Courts Network meeting

Kristiansand (Norway), 6^th September 2017

Workshop 2: How to weight a case? Sharing of good practices

Drafted by Noel Rubotham (Ireland, member of the SATURN Centre) and the CEPEJ Secretariat

Francesco Depasquale (FD), moderator, introduced the concept of case weighting of court caseloads, noting that a range of different types of case weighting methods were employed in court systems internationally.

The object of the workshop was to ascertain from participants whether their jurisdictions operated case weighting systems and, if so, how these systems worked, with a view to assisting SATURN in formulating case weighting guidelines.

FD then gave a Powerpoint presentation on case weighting. FD noted that CEPEJ's Compendium of “best practices” on time management of judicial proceedings had identified that caseload weighting as a tool in forecasting and monitoring caseload and workload capacity of the courts: court caseload changes over time, and should be monitored. That process of forecasting and monitoring can assist in

determining the appropriate complement of judges required to handle caseload
redrawing the judicial map
determining levels of productivity of judges and
determining the appropriate caseload to be assigned to judges, which in turn impacts on the length of proceedings.

Case weighting methodology, typically, will establish a set of case types or categories which enable data collection to be grouped logically, and will identify steps or activities generated in relation to a case, such as the number of court sessions to be held, the extent to which reliance on witness evidence will be needed, the level of documentation requiring to be reviewed and incidence of written decisions/judgments.

Case weighting establishes the “judicial effort” required for a given caseload, which effort should be measurable in minutes. Case weighting involves the gathering of data on the number of cases handled within a given time period and the number of events of steps in individual cases, and may include complexity indicators.

Among the commonly used techniques for case weighting are

Time and motion caseload studies involving (a) the maintenance of timesheets over a significant period of time, (b) the logging by judges of time spent on a case type or stages in a case type; (c) surveys of judges; and (d) case data analysis (by reference to case files, case registries, case types, case activity and transactions);
Expert evaluation by means of the Delphi method, involving an evaluation conducted by experts based on the structured collection of feedback from judges.

As to the Delphi method, reference was made to the advantages and disadvantages of this method as set out in a report for the Regional Co-operation Council of Bosnia-Herzegovina entitled “Backlog Reduction Programmes and Weighted Caseload Methods for South East Europe, Two Comparative Inquiries”. The advantages of the Delphi approach were in that report identified as-

1) its use of expert opinion,

2) it is consensus based,

3) it is less burdensome than a large-scale data collection effort,

4) it can be completed relatively quickly, and

5) is less expensive than a time study based approach.

The disadvantages of the Delphi method are -

1) its validity depends on the quality of the questions posed to respondents,

2) it is reliant on human interpretation/perceptions,

3) it is ultimately based on personal estimates, which may not be accurate.

Participants were invited to offer instances of case weighting and allocation methods employed in their courts. FD indicated that assignment of cases in Malta was done initially by specialisation, and cases were allocated to individual judges on a random basis.

A Netherlands participant indicated that there is no mandatory requirement to employ case weighting methodology, and only a few courts have done so. In the 1st instance courts, a case weighting system exists for consumer cases having a value of up to €25,000. The weighting is done using a 4 part scale – A, B, C and D - with A representing a simple case and D representing a complex case.

The case weighting system is used to determine (a) how many minutes a case will take to be heard and (b) how the case should be assigned within the court. Assignment of cases to judges is done based on a number of criteria, viz. the degree of complexity of the claim, the public/media profile of the case, the level of experience of a judge and the level of judicial expertise required for the case.

There are no published rules as to assignment, and the method is not very transparent, and the absence of rules for assignment has been the subject of query in the media. It is considered that the system will have to change in order to meet European standards. Any “paperless” approach to assignment of cases will require a case weighting system to underpin it if it is to work.

An Austrian participant indicated that the driver of allocation of cases within the courts in Austria was a fair distribution of caseload to judges, based on an analysis of incoming and disposed of cases. In the 1990s Swiss consultants had calculated the number of minutes it took for a judge to dispose of a case in 90% of the case statistical categories. This was done by judges completing a questionnaire as to how much time they expended on a case. The survey was conducted nationwide and tested for validation.

40 categories of case were employed and judicial resources needed were calculated based on

the no. of cases in a case category
the time required for case disposition in that category and
the number of hours a judge was available in a year.

The system has provided effective results. If the workload of a court is shifting/changing, there is clarity as to how human resources may need to be reallocated. However, the operation of the system does not guarantee that resources are in fact allocated to meet the workload change. In Austria, the experience has been that additional resources to those based on the allocation have been needed. Furthermore, the system enables allocation of resources to a court but does not prescribe how the court in turn allocates those resources among the judges within the court. A further weakness is that the system focusses on incoming cases – consideration is being given to determining distribution according to cases disposed of. Even if a precise mathematical model is employed, the system needs to be refined to accommodate anomalies.

The courts in Israel are in process of revising their case weighting system, which had been arrived at based on caseload data and employment of the Delphi method. Case weights related to different case events within a case were assigned to 101 case types (85% of case categories). Judicial time is measured from the first to the last event in the case, e.g. reading the case, preparing the hearing, conducting the trial and preparing and giving the judgment. The number of hearings in the case and not just the trial hearing is being measured. The number of cases resolved by judgment and by settlement is also factored in. Measuring the case weightings for each court and then averaging the results based on the number of judges in the court helps “even out” the results.

Under the revised scheme, judges will be asked to complete a time log for a month-long period. For each case type, the system will calculate the number of events per case type, or components of case type. Different types of event type closing a case will be given a proportionate percentage weighting. Instead of a Delphi – opinion-based – group evaluation approach, the preference is for a time log.

The Netherlands participant indicated that court support staff resourcing needed to be considered in a case weighting system.

A participant from Kosovo indicated that a pilot project on case weighting was being conducted.

The Croatian participant indicated that framework criteria had been introduced in Croatia for allocation of cases to judges - using 165 case types selected by a consultative committee - which in effect set norms for judges, viz. how many judgements a judge should issue for a given case type on an annual basis. The system is used to evaluate judges.

In Finland, a judge is required a number of times each year to report via a questionnaire on the number and types of case processed by the judge.

The Czech Republic has no case weighting system.

Azerbaijan initially attempted to employ a random case allocation system but judges were unhappy with it as it did not take into account differences in case profiles. For 4 years, an allocation system has now been in place which distinguishes between easy, medium and complex cases. The system now ascertains whether a judge has a sufficient case load of complex cases and can reallocate cases accordingly. In deciding what cases were simple, medium and complex, reliance was placed on a survey among judges and the case management system also recorded how much sittings time was needed for a case type. The survey results were compared with the results from the case management system to validate the differentiation. Simple cases were categorised as those taking 3 months; medium - up to 12 months; and complex - in excess of 12 months.

Work is now being undertaken to achieve an allocation based on caseload, personnel resources and budget needed.

Spain does not have a system of classification of cases for case allocation purposes except for the Commercial Court. A recent legislative change enables commercial cases to be assigned for trial in one of three designated cities.

In Denmark, the existing system weights cases not only for the courts but also for legal aid, since lawyers are paid according to the complexity of the procedure. In Slovakia, a working group of judges was set up in 2011 and produced a framework system comparable to that of Croatia.

In summing up, it was indicated that attention needed to be given to:

the data collection process required to underpin a case weighting system
the elements which should be seen as core to a case weighting system and
how to drive change in courts systems in implementing case weighting.

G.N.R and the CEPEJ Secretariat

Annex II – Summary of the workshop held in Valletta (Malta)

CEPEJ SATURN Working Group

Court coaching activity – Valletta (Malta), 12 February 2018

Venue: Valletta Courthouse, training room

Drafted by CEPEJ Secretariat

Agenda of the Court coaching

Morning

1. Introduction by Mag. Francesco DEPASQUALE (SATURN member) and the CEPEJ Secretariat and presentation of the team of experts

2. Presentation of the objectives of the meeting and of the framework note by Mag. Francesco DEPASQUALE and the Secretariat

3. Analytical review of the framework note and comments by the experts, summary of the first findings

Afternoon

4. Presentation of the first findings and discussion with members of Malta’s judiciary

5. Summary of the discussion by the experts

6. Definition of a working schedule to fine-tune and finalise the document

List of participants

Morning and afternoon

- Francesco DEPASQUALE, Judge, CEPEJ member and representative of the SATURN Group, Malta

- Joanne BATTISTINO, Research and Special Projects Officer, CEPEJ national correspondent, Malta

- Jonas VACLAV, Statistician, ministry of justice, Czech Republic

- Villem LAPIMAA, Judge, Head of appeal court, Estonia

- Dr Gali AVIV, Director of the Israeli Courts Research Division, Supreme court, Israel

- Shanee BENKIN, Judicial Researcher, Supreme court, Israel

- Yannick MENECEUR, Administrator, CEPEJ Secretariat

Afternoon

- 12 representatives of the Maltese Judiciary (judges, magistrates, registrars and assistants)

Main outcomes

In the framework of the SATURN activities

1. After the discussion with representative of the Maltese Ministry of Justice, Culture and Local Government and members of Malta’s judiciary, the group of experts agrees on the need to set clear objectives to the case weighting indicator. As a first step, one key of success could be building, as in Estonia, a system for judges by judges to allow a better case management.

2. The group of experts agrees on the use of all the information collected in Malta (presentation of experts, discussion with members of Maltese Judiciary) in order to enrich the final SATURN document about case weighting, especially the list of criteria that could have an impact on the complexity of a case.

3. The final SATURN document about the case weighting should include an Executive Summary, handy to understand for policy makers.

4. All the presentations already gathered could be implemented in a specific part of the CEPEJ Website, as scientific references, after agreement by their author.

5. A collection of different categorisations of cases in some European countries could be added to the final document and could be used as well as for other exercises being carried out by the SATURN Group.

In the framework of the Court Coaching activity

6. The members of Malta’s judiciary that attended the afternoon meeting have found extremely relevant exchange of views and sharing of practices from other European countries. They agree on the principle to work on a case weighting method within their court.

7. For the purpose of developing in Malta an adapted case weighting system, it has been requested by the members of the Maltese judiciary that they be provided with different categorisation of cases in some European countries. It was considered that attaining information from the various countries on the different case categories adopted in their countries would be of assistance, at this stage, and a starting point for further discussions.

8. On that basis, an internal meeting, organised by the CEPEJ member, could be held. Further bilateral cooperation between Malta and Israel is being considered.

Report

The CEPEJ has been invited by the Ministry of Justice, Culture and Local Government through the CEPEJ representative, to assist them in the setup of a case weighting methodology. This request was justified by the ongoing work of the SATURN Group about this topic and the content exchanges during the last pilot court meeting (6 Sept.2017, in Kristiansand, Norway), especially the presentation of observers from Israel that already used such a system in Supreme Court.

Francesco DEPASQUALE chaired the workshop with the assistance of the CEPEJ Secretariat. He indicated three main objectives which the workshop could discuss:

B. What is the scope of the current case weighting systems? What are the Pros and Cons of the existing systems?

C. How to convince the judges that the implementation of a proper case weighting system is important for an efficient management of the Court?

D. How to build a practical system?

The Secretariat shared with the experts the content of a document drawn up by the CEPEJ Secretariat which included the contributions of 12 Member States plus 1 observer State (Israel) and summarized the way CEPEJ preferred to approach the issue: the goal is to build an effective document for policy makers and practitioners, after a comprehensive review of the existing systems.

1. What is the scope of the current case weighting systems? What are the Pros and Cons of the existing systems?

As definition, the term “workload” must be distinguished from “caseload”: “workload” is about all the range of the activities of a judge (disposal of cases but also training, meetings, professional activities and any other duties); “caseload” is only about the activity linked to the disposal of a case (mainly preparation, hearings, drafting, juridical researches). This definition was important as any ‘case weighting system’ could deal with both the “caseload” as well as the “workload”.

Two main types of case weighting systems could be identified as:

E. A time based approach case weighting system

F. A points based approach case weighting system

These two methodologies both ~~need~~ require a proper case categorisation assessment/evaluation before any other development may take place.

The goals which these systems may achieve are mainly two-fold:

G. The proper and equal distribution of work between judges

H. A useful tool to assess the proper means and allocation of resources according to each case category

In further developments, these systems could also help to build a new statistical approach in courts, (by adding the weight of pending cases for example to the current numbering of these cases where 1 case = 1 unit).

The time based approach aims to measure average time spent by a judge to deal with their tasks. For instance, this measurement was done in Israel with an anonymous questionnaire where a sample of judges had described their daily activities. This approach is also used in Austria (PAR System) for instance.

The utility of this approach is that it avoids identifying all the elements needed to assess the complexity of a case (see below) and it reaches its conclusions from the observations collected from a significant sample of time measurement. Nevertheless, this is a rather subjective approach which could be subject to personal bias of the various contributors. For the Israeli experts, these biases could be reduced by coupling this exercise with a significant and detailed data collection.

The points based approach could be a more objective approach as it consists of allocating an index of complexity to a case based on studies carried out. This index could be obtained in a generic manner, by agreement reached amongst the various stakeholders (In Estonia, for example, judges were requested to rank the complexity of their cases from 0 to 40 and, following much debate, a ranking was drawn up) or in a more precise manner where certain predetermined factors of complexity are taken into consideration.

Some of the elements identified by the experts, even referring to ECHR decisions, could be the following:

a) Legal basis of the claim and type of proceeding

b) Claim monetary value

c) Number of claims/complexity of motions/line of defence

d) Number of parties/defendants/lawyers

e) Number of witnesses

f) Amount and range of expertise / Number of experts

g) Number of foreseen sittings predetermined audiences / hearings

h) Whether the case is an International / cross border case or not

i) Whether Interpretation / translation is needed

j) whether the case relates to a new case category or whether the subject matter is known to the judiciary (reference was made to the difficulties of presiding over cases that are new to the court, such as corruption and money-laundering cases)

This points approach, though more objective at first glance, seems nevertheless more difficult to setup due to the difficulties one may encounter to agree on each factor of complexity and their own weight. Moreover, defining each factor the weight each factor should be allocated is still prone to a subjective interpretation by those involved in the process.

Whatever is the approach, a constant evaluation and update seems mandatory to ensure maximum compatibility to the reality of the activity of the court.

2. How to convince judges that a proper case weighting system is important?

The discussion with representatives of Judges, Magistrates and Court Registrars of the Maltese Courts, following a short presentation provided by a representative of the Maltese Ministry of Justice, highlighted the difficulties that implementing such a system may encounter at the moment and the high level of reservation expressed by the various stake holders. These reservations and suspicions should not be limited to the specific context of Malta, there being a small population of legal professionals involved, as such reservations arise due to the particular circumstances of the judicial activity in any particular country.

The main arguments raised by the audience of professionals were the following:

- The particular circumstances of their own work when compared to other European countries (thus leading to their discarding the utility of any comparison process being carried out)

- The particularity of each case category (and within a category of each case) and lack of relevance of any global approach of measurement

- The existing excessive workload in Courts may prevent the effective introduction of a new management tool unless more resources are allocated to provide for such a new tool

- Negative consequences resulting from the availability of such a tool:

* Undue pressure from external authorities on Judiciary (Ministry of Justice or Judicial Council) placing more emphasis on the Judiciary to solve more cases whilst not allocating any further resources

* Inequality of the system: Judges that solve cases in a shorter period of time may be requested to deal with the cases still pending before other colleagues of theirs who are not as efficient as them.

At face value, these classical recurrent arguments may reveal a culture of individualism and/or a lack of management culture amongst the Judiciary. However, this is incorrect as the court professionals at the meeting showed very clearly that they are very sensitive to the imperative need to deliver a proper and qualitative decision to the parties within a reasonable timeframe. Their main concern, however, was based on their assertion that more resources were needed, which need could not be verified as appropriate or not without an objective assessment of the situation.

In view of the above, the college of experts agree that all stakeholders should be made aware that the development of such methodological tools to help them should be carried out by the Judiciary themselves, so as to empower the judiciary as well as give them the opportunity to be more accountable and efficient. In consequence, the experts insist that, in accordance with the particular individual elements of each judicial system, the development and ownership of such a case weighting system should be entrusted, at least in its inception stage, to the judiciary itself and only once it is clearly defined by the Judiciary should it be shared with other stakeholders. There could be assistance of course of other authorities/bodies to develop that. This approach is in line with the CEPEJ SATURN guidelines.

3. How to build a practical system?

Various countries have introduced and implemented case weighing tools and it is therefore relevant, at this stage, to have a cursory look at their implementation.

In Czech Republic, the current case weighting system was established in 2007 by a Commission made up of representatives of the Ministry of Justice and Judges from various courts. The Commission has set up the whole system (methodology, utilization, all case weights etc.). Since then, the system has not been changed (with the exception of weighting of particular cases). The weighting of the cases was agreed to on the basis of agreement between the members of the commission. No statistical calculation was used.

In Estonia, the system was built in two stages (the second stage is still ongoing):

- Stage 1: A pilot system was drawn up in the Tallinn Administrative Court, agreed to by all the 18 judges of the Court. A team leader among the judges co-ordinated all the project and information gathered from all his colleagues. It was agreed that it was essential to set up a reference year, this being necessary to be able to gauge their weighting accordingly, and the judges all agreed that 2016 could be considered as a reference period, due to what they consider a good balance of the workload. All the judges then proceeded to allocate, on their own initiative and experience, from 1 to 40 points to the each of the case categories which they preside over, 1 point being equivalent to 8 hours of working time. Despite initial difficulties such as managing all the data, with much input being overestimated by the various judges, eventually, following various meetings and workshops, the weighting was finally established on a reasonable basis. Eventually, the total number of points they obtained per judge calculating the cases they had was divided by 200 days, this being number of working days each judge was considered to have to adjudicate within one calendar year, this calculation led to attaining an estimation of its caseload.

- Stage 2 (ongoing): This stage aims at expanding the system on a nationwide basis. Nevertheless, the Estonian Ministry of Justice which is now piloting this project, is aiming to utilise this system to calculate the resources which need to be allocated to each Court and Judge, this goal is facing opposition and difficulties from the Court and Judges who have not yet implemented the system.

In France, a large consultation process was carried out in 2014 between all the stakeholders (MoJ, Inspection of justice, representatives of unions of judges, representatives of courts). The objective was identifying some relevant indicators and proposing a time measurement evaluation for each task/activity. Each indicator (per function of judge and type of litigation) was discussed after quantitative measurement (number of cases disposed per FTE on a basis of 210 working days). A final vote between all the members of the working group led to the publication of a progress report which revealed discrepancies in the calculations being made. Nevertheless, at this moment, no public information seems available about the application of these calculations in the courts.

In Israel, in the Supreme Court, a system was built based on anonymous declarations of time spent per activity by each judge, such declarations being available on an IT systems designed for this purpose. The average was calculated upon this basis to be able to assess the resources needed for each case category.

The observers from Israel underlined the fact that it was important to be aware of limitations in the case weighting system. First of all, since case weights are attained from the average time an average judge invests in a case, they are subject to the “Law of Large Numbers”. This means that comparing workloads should be done only for masses of cases, and the more cases are evaluated in detail, the more accurate the information attained will be. This also means that applying the weights at the single-judge level may be tricky, and should be approached with caution. Lastly, the weights established are only a static snapshot of the current situation. They reflect a certain reality, and need to be updated regularly in order to remain relevant and effective.

4. Follow-up of the activity

For the purpose of the SATURN Working group, the group of experts agrees on the use of the information collected (presentation of experts, discussion with members of Maltese Judiciary) in order to enrich the final document about case weighting.

Francesco DEPASQUALE underlined the need to draft a short document, being an Executive Summary, handy to understand for policy makers. All the presentations gathered could be implemented in a specific part of the CEPEJ Website, as scientific references, after agreement by their author.

For the purpose of developing in Malta an adapted case weighting system, it has been requested by the members of the Maltese judiciary that they be provided with different categorisation of cases in some European countries.

At a workshop level, it was considered that attaining information from the various countries on the different case categories adopted in their countries would be of assistance, at this stage, as this could be of relevance both for this exercise as well as for other exercises being carried out by SATURN.

On that basis, an internal meeting could be held and further bilateral cooperation with Israel is being considered.

Annex III – Examples of case-weighting provided for CEPEJ Activities (Albania and Kosovo*)

1^st example drafted by the CEPEJ Secretariat

The Cooperation unit of the CEPEJ has a request of the project team who is building the new case management system (CMIS) in Kosovo*. They need to find a calculation method to allocate automatically cases to judges, taking into account the complexity of these cases.

You will find below the very first findings of the Secretariat based upon the experience of ECtHR and a more précised document drafter by the CEPEJ Expert Jacques Bülher.

Extract of the report

To allow automatic case allocation of a case to a judge, the CMIS Team wishes to prevent to allocate a case to a judge that has too much workload. The calculation of such workload supposes, first, to case weight the new file.

Some objective elements have already been identified by the CMIS Team to allow case weighting:

- The number of parties

- The number of claims in a case

These first elements are according to the definition of a complex case given by the European court of Human Rights, but some more elements could be added:

- The complexity of the matter (for example, land property cases could be more complex than a repetitive payment orders)

- The stakes of the procedure (for example financial amount, important issues of general interest)

- According to the matter, the predicable need to designate judicial expert

All among the life cycle of the case, this case weighting could also evolve according to:

- The predicable length of the hearings, by taking into account eventual previous length

- The number of preparatory hearings

A “scoring” methodology should be applied for all these criteria by the computer system. A theoretical example should be provided:

- 1 point per party

- 1 point per claim x 1 to 5 (1=claim based a simple legal problem / 5=claim based on a complex legal problem) x 1 to 5 (1=small financial amount / 5=high financial amount)

- 2 points per expertise needed

- 0,5 per eventual pre-hearing x 1 to 5 (1=short predicable length for hearings to come / 5=long predicable length for hearings to come)

These first recommendations could be challenged by judges and registrars, per category of case (in civil matter: urgent cases, repetitive cases, complex civil cases, etc / in criminal matter: cases with provisional detention, complex criminal cases, minor offenses, etc).

* This designation is without prejudice to positions on status, and is in line with UNSCR 1244 and the ICJ Opinion on the Kosovo Declaration of Independence.

2^nd example drafted by Jacques BUHLER (Switzerland, CEPEJ Expert)

Strengthening the Quality and Efficiency of Justice in Kosovo* (KoSEJ Action)

Inputs for a lottery application for courts

1. Introduction
A lottery application within court can help

a. to improve the impression of independence and impartiality of a court

b. to achieve an equal distribution (workload) between judges

c. to improve the efficiency of the distribution of cases and the designation of judges for cases judged by a panel.

A lottery application has to be linked with a judicial case management system (CMIS). Below you can see a possible IT architecture in courts. In this case the case allocation system (Lottery) is a separate system which communicates with the case management system. It can also be integrated.

For its functioning, a case allocation system needs

· information in relation with the profile of the judges

· information related to the concrete case

· to integrate the general strategy of the concerned court for the distribution of cases between judges and for the weighting of the cases.

The workflow of an automatic allocation of cases contains generally the following steps:

2. Targets of a lottery application for courts

a. No human intervention for the repartition of cases between judges

A lottery system has to minimise the intervention of a judge or of a member of the court’s staff for the repartition of the cases between the judges. Manual interventions have to be well motivated exceptions. So it makes possible to improve for the court users the impression of independence and impartiality of a court.

b. Designation of the judge(s) in charge of a case
When cases are arriving in a court, a magistrate or a member of the court’s staff has to attribute the incoming cases to a judge or to designate the judges of the panel of judges who have to adjudicate the cases. A case allocation system contributes to do it with impartiality.

c. Distribution of an approximately equivalent workload to all judges of a court

A lottery system has to divide equitably the incoming cases between the judges of a court or of a section of a court. At the end of the specified period (generally a year), the received workload of all judges has to be approximately equivalent (in number of cases or in number of points in case of using a case weighting system).

d. Improvement of the efficiency of the court

For an efficient attribution of the cases, a lottery system can contain a complex algorithm for an equitable distribution of cases between judges and for taking into consideration a lot of parameters regarding the characteristics of the cases and the profile of the judges.

3. Profile of the judges

The main elements of the profile of a judge which can be useful for a lottery system are:

· Name, surname

· Gender

· Languages

· Specialisation

· Court and section of the court

· Function (President, Vice-President) within the court

· Function within the section of the court (President, Vice-President)

· Percentage of time dedicated to solving cases (part time judge, or starting its activity during the year, or judge with other functions than only solving cases)

· Absences (long duration only[5])

· Permanent causes of challenging[6]

· Etc.

4. Possible parameters for a lottery system

To achieve the targets mentioned above and depending on the organisation of the court, the following parameters can be useful:

a. Parameters for the calculation of the workload

i. Period of time for the calculation
Generally the period of time has to be long enough, for example one year. So it makes possible to integrate “normal” absences like holidays, continuous training, etc. without special compensation mechanism.

ii. Type of cases

The effort to instruct and solve a case is not the same for each type of cases. A complex case needs more time and efforts than a simple case. Weighting the types of cases could help to estimate the workload done by each judge. Weighting of cases is not necessary when each judges receives approximately the same number of each types of cases during the year.

iii. Number of cases or number of points per case

The number of cases or the number of points of all cases adjudicated during the considered period of time has to be taken into consideration for the calculation of an equivalent workload between each judge.

iv. Distinction between the activity as “reporting judge” and the simple participation as judge within the panel

Often the activity of “reporting judge” within a case needs more efforts and time than the simple participation within the panel of judges. When it is so, it could be useful to separate the calculation of the workload for the cases attributed as “reporting judge” and of the workload for the cases in which the judges are involved “only” within the panel (see below fig. 7).

v. Percentage at disposal for solving cases

If there are differences between judges regarding the time they can spend for solving cases, the percentage at disposal for solving cases has to be taken into consideration for the calculation of the workload. For example in some courts, the court’s president can only spend 70% of its time to solve cases.

vi. Attribution of new cases considering the workload

The target of a case allocation system for the judges is that at the end of the year they had all an approximately equivalent workload (in number of cases or in number of points). A second target is that the case allocation system contributes to solve the cases as soon as possible. There is an apparent contradiction between the two targets.

The algorithm of a case allocation system can be:

When a new case arrives:

a) thesystem calculates how many cases or points all judges of the concerned court or section of the court:

- have solved since the beginning of the year

- are pending on the date of the case allocation for each judge.

b) The system calculates a ranking of the judges based first on the number of solved cases and, if two judges have the same number of cases on the number of pending cases.

b. Parameter of the specialisation of the judges

This parameter can be taken into consideration to designate the judge.
For example for each tax case:

i. one of the judges of the panel has to be specialised in tax law or

ii. the single judge or the reporting judge has to be specialised in tax law.

c. Parameter of the gender of the judges of the panel

For certain types of cases, it is mandatory (depending on the legislation) or simply better to have judges of both genders within the panel. Examples: sexual delicts or cases regarding the equality between man and woman.

d. Parameter of the absence

In a court, some cases are urgent and have to be solved within a short period of time. For this type of cases, it can be useful to link the lottery system with a list of the absences of the judges, which contains all the absences known in advance like holidays, continuous training, etc.

i. A lottery system can use the information if a case is urgent or not. The urgency can be transmitted automatically by the case management system, depending on the type of cases, or inserted manually by the operator of the lottery system before the designation of the judge(s) (within the lottery system). In urgent cases, judges can be designated only they are not absent.

ii. If a judge is absent, for example, for less than 5 working days, he can be designated by the lottery system. But if he will be absent for more than 5 working days, the lottery system can taking this absence into consideration. He will not receive a case the last one or two days before he is absent and he will receive new cases only one working day before he comes back to the court.

iii. If a judge is absent for an extraordinary reason for a very long time (accident, etc.), this absence is integrated in his profile and it changes the number of cases he has to solve during the year (a less number than its colleagues who had no long absence – incidence on the calculation of the workload). In this case he receives no cases during its long time absence.

e. Parameter of challenging

When it is known in advance when a judge has to decline to give its opinion, it is possible to integrate this information into the case allocation system to avoid that this judge will be designated when a certain person is involved in a specific case.

f. Parameter of having the same judge or panel than in a similar case

When it is useful to give all very similar cases to the same judge(s) for efficiency reasons, it should be possible to link this cases before the using of the lottery system to obtain for all the linked cases the same result (single judge or panel).

5. Possible functioning of a case allocation system

A case allocation system can foresee the following main steps: Selection of a file (when needed), allocation of the case to a single judge or to a panel (including designation of the reporting judge and the “simple” members of the panel and, exceptionally, modification of one designated judge.

a. Selection of the file

When the incoming cases are not allocated automatically when they are registered, it is possible to enter the file number of the case that has to be allocated to a judge or a panel.

b. Adaptation of the default parameters for a specific case

The default parameters can be adapted to the specificity of the case, for example

i. Number of judges: When the default parametrisation is a single judge, it has to be possible to increase the number of judges to 3 (panel)

ii. Gender: If it could be useful or foreseen by law to have in a panel judges from both genders, it should be possible to ask the system to designate judges of both genders for the panel.

iii. Specialisation: If it could be useful for solving the case that a specialised judge is part of the panel, it should be possible to ask for the designation of a minimum of one judge who is specialised in the topic of the case.

iv. Priority: If the case is urgent, for example when the applicant is in detention, it should be possible that the system allocates the case only to a present judge or select for the panel only judges who are not absent.

v. Language: The system should be parametrised that the cases are allocated to a judge which knows the language of the procedure.

c. Modifications of a designated judge

Modifications of a judge who was designated by the case allocation system have to be an exception and done for a valuable reasons. It should be impossible, within the system, to validate a manual modification if the field of the reason of the modification is not completed.

Fig. 6: The user of the system has only to isolate the judge who has to be replaced. The algorithm of the system replaces automatically the judge who has to be replaced by another one.

6. Control of the lottery system

The integration of the parameters mentioned above allows an equivalent repartition of the workload between judges and minimising the manual intervention into the automatic designation of judges or of the panels of judges made by the lottery system. Sometimes manual modifications are necessary. It is necessary to control both: the functioning of the case allocation algorithm and the manual modifications.

a. Control of the functioning of the repartition algorithm of the workload

When one of the target of the case allocation system is that at the end of the year all judges have received a equivalent workload (number of cases or number of points, it is useful to control the functioning of the algorithm.

Fig. 7: Example of a controlling list that give an overview of the number of cases (or points) allocated to each judge as reporting judge and as member of a panel. It appears for example that all judges were involved approximately in the same number of cases as member of a panel during 2016. This was a target of the algorithm.

b. Controlling of the reasons of manual modifications

A control list of all the modifications containing

i. the file number,

ii. the name of the parties,

iii. the date and hour of the initial designation made by the lottery system,

iv. the name(s) of the judge or of the panel designated by the lottery system

v. the date and hour of the final manual designation of the judge or of the panel

vi. the name(s) of the final manual designation of the judge or of the panel

vii. the reason of the manual modification

viii. the name of the operator with has made the manual intervention into the lottery system.

c. Controlling of the frequency and reasons of manual modifications

It is useful to have a control statistic which indicates the percentage of manual modifications and the reasons for the manual interventions. The target is to have a percentage of manual modifications less than 10%.

Strasbourg, 16 October 2017 / V.2 / BrJ

____________________________________

Annex IV – Draft document provided by Mr Jonas Vaclav, CEPEJ Expert (Czech Republic)

Case weights in the Czech Republic

Disclaimer: The analysis provided in this document reflects the work of the expert and his own understanding of the case-weighting methodology in Czech Republic.

Setting the case weights

There is only one “type” of case weights used. Current case weights system has been established in 2007 by the commission composed of representatives of the Ministry of Justice and judges from various courts. The commission has set up the whole system (methodology, utilization, all case weights etc.). Since then, the system has not been changed (with the exception of height[7] of particular case weights). The height of the case weights has been set up on the basis of agreement between the members of the commission. Unfortunately, no statistics were involved.

Basic description of the case weights

Each type of case has assigned its weight, which is expressed as the number of cases judge is supposed to resolve per year assuming that he is working only on that particular case “type”. The case weight is supposed to express how difficult and time-consuming is given type of case. District courts and regional courts have different case weights. High courts have their own case weight as well, but currently they are not being utilized.

Here are some examples of case weights:

· District courts, criminal cases: 282 cases per year

· Regional courts, first instance criminal cases: 28

· Regional courts, second instance criminal cases: 147 or 291 (it depends on the type of particular second instance criminal case)

· Regional courts, administrative law cases – financial matters: 88

· Regional courts, administrative law cases – general cases: 88 or 120 (it depends whether case is being resolved by single judge or by senate).

· District courts, electronic payment order: 2933

· District courts, inheritance procedure: 880

· District courts, litigious civil cases – general: 352

There are 29 case types for district courts and 30 case types for regional courts.

There might be a problem mainly with civil litigious cases and criminal cases. Basically, there is only one (three for civil litigious cases) defined case type for the whole agenda, which is not sufficient. Thus there is no distinction between various types of criminal cases (e.g. murder, theft and tax evasion) and there is also no distinction between various civil litigious cases.

Utilization of case weight – appointment of judges

Case weights are used only for one purpose – to determine the need of judges for all courts. Basically, it is done on the basis of weighted case fillings. There might be the second issue. The calculation does not take into account pending cases or backlog. Thus the pending cases and backlog have no effect on the calculated need of judges.

Firstly, the number of filed cases for all case types is estimated for the following year for each court. Here, it is possible to identify another potential problem, because the estimation method is very simple –weighted mean of the number of cases filed in three previous years is used. In 2017, the number of case filings for year 2018 was estimated as follows:

Where CF is the number of cases filed and ECF is estimated number of cases filed.

For each case type, the number of judges required to resolved estimated number cases filed is calculated. Then, the numbers are summed up.

For example: There is a district court A. The number of case filings for this court for year 2018 has been estimated as follows: 2000 litigious civil cases – general, 1000 criminal cases and 1500 inheritance procedures. Thus the need for judges is calculated as: 2000/352 + 1000/282 + 1500/880 ≐ 10.93 ≐ 11. Thus the need for judges for court A is approximately 11 judges.

Finally, few important points:

· The system of appointment of judges is slightly more complicated in reality (there are some exceptions, sometimes judges are appointed to courts with high caseload and high case resolution time etc.)

· The judges are usually being appointed three times a year, usually 30-35 judges are being appointed. The judges are appointed on the basis of the difference between need for judges and actual number of judges.

· Calculation of need of other court personnel (e.g. Rechtspflegers) is very similar.

Other possibilities to operate case weights

Scientifically, other calculations are possible. For instance:

· We find “the number of cases judge is supposed to resolve per year” to be rather clumsy and hard to work with. Thus we rather transfer it to the case weights. For example case weight for District courts, litigious civil cases is 352/352 = 1, case weight for District courts, criminal cases is 352/282 ≐ 1.248, etc. Afterwards we multiply the number of cases of all case type by its weights and sum all cases to get weighted number of cases.

· We utilize case weights not only for the number of cases filed, but also for all incoming cases, resolved cases, pending cases, caseload and backlog.

· We experiment with the utilization of more sophisticated methods of forecasting (e.g. ARIMA). In the predictions, we mainly work with time series of weighted caseload.

· We often work with huge datasets that contains data for finished cases of main case types (criminal cases, litigious civil cases, cases of upbringing and maintenance of a minor). One observation equals one finished case (or person). This allow us to crate second set of case weights – case weights for criminal cases and also for litigious civil cases. Here are some facts about second type of case weights:

o In the data, we distinguish more than 400 case types of litigious civil cases and all sections (paragraphs) of criminal code of the Czech Republic for criminal cases. This data allow us to experiment with determining theoretical difficulty of more case types and thus to determine theoretical difficulty of the caseload of all courts individually.

o We can determine the difficulty of various groups of case types or even individual cases types. Of course, we are limited by the information contained in the dataset.

o As a measure of difficulty for given case type we typically use mean number of hearings or mean (median) case length. We would like to use mean number of defendants too, but it is tricky to get this information from the data. We also find comparison of the reversal rates or the appeal rates between case types to be quite interesting.

o For example, we use following methodology: let’s use the number of hearings as an approximation for the criminal cases difficulty. We calculate the mean number of hearings in the whole dataset (MH) and also mean number of hearings for each section (paragraph) of the criminal code (MH_j), where j is the number of the section of the criminal code.
Afterwards we calculated the case difficulty coefficient (DC_j) for each section (paragraph) of the criminal code as follows:
Thus difficulty of cases with DC_j < 1 is below average and the difficulty of cases with DC_j > 1 is above average.
Finally, we calculate the difficulty coefficient for criminal cases of each court as weighted mean of case difficulty coefficients, where the weights are numbers of cases (of each case type) resolved by the court. The coefficients for courts vary between 0.7 and 1.3.
We then multiply the number of cases (resolved, incoming, caseload, pending and backlog) by the coefficients. Thus, we somehow control for the case difficulty.

o Again, the description of unofficial case weighting is simplified here. Of course, we utilize and experiment with other methods (e.g. regression analysis).

Annex V – Presentation of the methodology used in Serbia

Annex VI – Presentation of the methodology used in Israel

Annex VII – Presentation of the methodology used in Austria

Preliminary Remarks

At the core of the project "creating a personnel demand calculation" (now called PAR II to distinguish it from the previous PAR) supported by the Federal Ministry of Justice and by the professional and employees' representatives is projecting the demand for judges and public prosecutors on the basis of an empirical survey of the working hours spent by a representative sample of courts and public prosecution offices over a limited period of time.

The very methodology of average calculation based on samples necessarily follows that the results must also be average results, and with this in mind only consideration of the whole (all courts, all public prosecution offices) or of again a representative sample thereof can be "valid", i.e. representative. It is the essence of any average value that the partial results the average is comprised of may vary considerably. Thus, the smaller the specifically examined sub-segment, the more it must be expected that it indeed will deviate from the whole, and this is why the findings from the nation-wide average calculation of PAR II may not be indiscriminately implemented locally. As before, such finding may serve as indicators and further source of information for distributing business or issues of equal capacity utilisation within an office, but they can in no way replace the required perspicacity when assessing the actual condition on site. Everybody knows from own experience that a limited quantity of specific proceedings may require much more or much less time than the average.

1. The Survey

In order to obtain the required data for the projecting process, judges at representatively selected 30 district courts and 11 regional courts, as well as five public prosecution offices were requested to log their total working hours for a period of six months (from October 2008 to March 2009), while complying with a set of logging rules. Obviously, such statistics are the more precise, the more individual data they are based on. On the other hand, any survey reaches boundaries with respect to expediency and financing. The assigned management consultants advised to aim at obtaining processing times for at least 100 each of commenced and concluded proceedings of each class of business. In view of the survey matter and the long survey period, the only way of obtaining the required data was to ask judges and public prosecutors at the surveyed offices to concomitantly log their personal working hours.

For this purpose, working groups were established in advance as from summer 2007, only interrupted by the introductory phase of the amended Code of Criminal Procedure, representing practitioners, the professional association, the Federal Computing Centre and the Ministry of Justice under the leadership of a renowned management consulting firm. This consulting firm had previously supervised the German PEBB§Y-Project (Personalbedarfsberechnungssystem), which was in turn based on the first Austrian personnel demand calculation system PAR.

Logging of working hours was performed on a file basis, i.e. on so-called "process sheet", associated with individual proceedings. Individual proceedings were in turn systematically categorised in business classes, pooling different file keys or sometimes only single case codes in a file key register (class catalogue). If any proceedings were already checked on the register, the time log was recorded on so-called "extension sheets" also allocated to the individual business classes. Only process sheets counted as quantity, extension sheets counted only with the time logged, as the checked proceedings as such had already been counted once.

To a smaller extent, logs were recorded only in a summary manner on period-related sheets ("half month sheets"), on which the times spent for certain kinds of activities (e.g. further education) were logged without linking them to individual proceedings.

Almost all judges and public prosecutors of the participating offices took active part in the survey (e.g. all judges working at 26 of the 30 participating district courts took part). More than 352,000 sheets were returned during the survey period. 99.92% thereof could be used by processing them in a data base.

2. The Personnel Demand Calculation System PAR II

2.1 Components

As before, the personnel demand calculation system breaks down into two components, each forming a specific segment. The major part consists of the so-called calculatory variable processing times ("judicial times"), i.e. those working times that are logged in relation to the number of files accumulated. The term "calculatory" denotes that this is a parameter (average value) the actual value for specific proceedings may differ from. The much smaller part of about 10% of the total time spent are the calculatory infrastructural times, which are calculated unrelated to the business volume, i.e. for three areas:

(1) Justice administration, own further education, education of third parties (trainees, candidate judges), other basic tasks such as business and staff meetings, statistical work, reports, etc.,

(2) Travelling and administrative times in case of dual assignments, travelling and waiting times for circuit court days,

(3) Functions in a staff and professional representative body, in staff panels and staff commissions, as compliance officer and in the European Judicial Training Network EJTN.

2.2. Obtaining time values (base numbers) for calculatory variable processing times

The PAR system for calculating personnel requirements in the judiciary (calculatory variable times) consists of dividing the time spent for processing certain cases (PAR II classes) obtained from the surveyed offices during the survey period by the number of cases processed, and projecting the total processing time (for all offices and for a whole year) from the resulting time value ("base number") and the total number of cases to be processed per year by all offices. So the key of the formula are the respective quantities, which are first divided by the time spent during the limited survey period at the surveyed offices, and which subsequently must be multiplied by the time values (base numbers) as drivers to calculate total time requirement.

*total time requirement =*	*surveyed time requirement*	x *total quantity*
	*surveyed quantity*

During the survey period, the survey quantity is a result – simply speaking – of the number of logged process sheets. More in detail, to arrive at the survey quantity, proceedings commenced and concluded during the survey period are quantified with a factor 1, only commenced or only concluded proceedings are quantified with a factor 0.5, and proceedings commenced prior to the survey period and only concluded after the survey period are quantified with the factor 0. Regardless of whether proceedings have been calculated as a whole, half or not at all to arrive at the (calculatory) surveyed quantity, all times used and logged for such proceedings were calculated wholly and unabridged to arrive at the surveyed time requirement, for that matter. Even the time requirements of proceedings neither commenced nor concluded during the surveyed period are included as "supplement" in the times of the proceedings counted for the quantity, so that also the time requirement for proceedings of long duration will be included in the calculation pro rata.

In the experience of the consulting firm, a time value obtained (base number) only becomes empirically valid, if the average value obtained is based on at least 100 proceedings surveyed. This threshold has been considerably exceeded in all major business classes, with the rule-of-thumb assumption, that due to the six months duration of the survey covering 1/3 of district courts (in terms of capacity) about 1/6 of the total annual case volume of district courts, and due to the survey covering half the regional courts and public prosecution offices about 1/4 of the annual case volume of regional courts were included in the survey. Only in some very small business classes (B06, B15, B22, L01, L05, L29), the quantity of actually surveyed proceedings were only around or below the threshold, resulting in the base number from these classes having only minor empirical validity.

As obviously process sheets from outside the surveyed offices and the survey period as an indication of the number of proceedings are not available, any real time operation must fall back on the quantities of proceedings logged in the registers. Therefore, the survey must be retranslated into the systematology of the Automated Process Register (VJ).

The transfer of survey results into real time operation occurs in two stages. The first stage relates to the retranslation of VJ mapping class catalogues into VJ classes, i.e. the correct bundling of case codes and process characteristics with the related case volumes of those groups (classes) as defined in the survey. Once such retranslation is concluded, in a second stage the quantities missing in the survey (i.e. quantities from other offices and from periods not surveyed, i.e. mainly periods in the future) may be obtained from the registers (from VJ).

But considering that almost all judges at the surveyed courts had participated in the survey, it made sense to compare the results of the translation and retranslation process for the surveyed period and the surveyed courts with the surveyed quantities, to find out whether the registers (VJ) for the same period and the same offices would yield quantities in correlation with the survey quantities (i.e. the count by the survey participants). If e.g. the surveyed courts effectively used and logged 100,000 minutes for 1,000 of one class of proceedings (based on the number of process sheets) (resulting in a time value or base number of 100 minutes), it ought to be expected – provided the entire survey methodology including the translation processes has functioned properly, that also VJ yields a quantity of about 1,000 proceedings again, resulting in total working hours of 100,000 minutes when recalculated. Would VJ only yield a quantity of 700 proceedings instead, only 70,000 minutes of the actually spent time would be included in the calculation, but would the result be 1,300 proceedings, total working hours of 130,000 minutes would be calculated, although actually only 100,000 minutes were spent. In both cases, the mathematical calculation of working hours spent would deviate from reality in the personnel demand calculation.

Therefore, upon completion of the survey, the survey quantities (quantities of sheets) were matched with the VJ quantities (volume according to VJ) and the definitive time values (base numbers) were calculated from the working hours collected (rounded up or down in a commercial manner to show whole minutes), taking into account the rate of participation by individual branches.

2.3. Judges as Rechtspfleger

Also the judicial activities of judges in the function of Rechtspfleger, or in proceedings entrusted to Rechtspfleger, differentiated by variable rules, were included in the survey. If judges had to take specific measures in proceedings entrusted to Rechtspfleger (activities reserved to judges according to the Rechtspfleger file) such times spent were added by period to the respective basic classes of the judges, and by this clever move they are attributed to the judges active in these business classes. Moreover, the activity of "judges as Rechtspfleger" in default actions was included in PAR II as separate class (B30 or L40). But those activities performed by judges standing in for (temporarily or for the longer term) absent Rechtspfleger – a quantity nation-wide of two full-time equivalents at the district courts and about 0.5 full-time equivalents at the regional courts can be assumed from projecting the survey results – could not be included in PAR II, because no detailed data are available from the automated file register (VJ) for such cases, where and to which extent such substitute activity has occurred.

2.4. Judicial Non File-Based Activities

Judicial activities of judges and public prosecutors not specifically file-based (legal advice days, telephone advice service, general meeting concerning issues of jurisdiction or with other authorities, etc.) were included separately, broken down into classes (in the old PAR only legal advice and circuit court days were included without any breakdown). The total working hours obtained were linked to calculation indicators ("drivers") (as a rule the total working hours in this branch), so that fictitious base numbers could be calculated. It must be pointed out that the driver per se may be defined freely and is not directly related to the frequency of such activities. The resulting time value (base number) has no relationship to the actual time spent with a single non file-based activity, only the result of multiplying the driver selected and the base number provides information about the volume of non file-based time spent which has actually accrued in this branch.

2.5. Measuring Calculatory Infrastructural Times

Three groups of tasks count among calculatory infrastructural times:

(1) Justice administration, own further education, education of third parties (trainees, candidate judges), other basic tasks such as business and staff meetings, statistical work, reports, etc.,

(2) Driving and administrating times in case of dual assignments, driving and waiting times for circuit court days,

(3) Function in a staff and professional representative body, in staff panels and staff commissions, as compliance officer and in the European Judicial Training Network EJTN.

Based on the survey results and the existing definitions agreed with the professional representatives the following percentages of staff employment (systematic permanent posts) were defined for group (1):

District Courts:

· Measuring administrative/management tasks on the basis of the same formula as before, i.e. 5% of the first 3 permanent judge posts, but at least 0.05 full time equivalent, 4% of the next 7 permanent judge posts, 3% of the next 10 permanent judge posts and 2% of all permanent posts above 20;

· Measuring "other basic tasks" at 3% of work capacity (permanent posts, corresponding to about 75 minutes per permanent post and each of the statistical 42 work weeks);

Regional Courts:

· Measuring administrative/management tasks pursuant to Section 31 Court Organisation Act at 2.5% of the permanent posts of the court of law and the subordinated district courts, limited to 3.5 permanent posts or 3% at the two regional criminal courts;

· Measuring "other basic tasks" at 2% of work capacity (corresponding to about 50 minutes per permanent post and each of the statistical 42 work weeks);

Public Prosecution Offices:

· Measuring administrative/management tasks by applying the formula 8% of the first 5 permanent posts, 3.5% of the next 10 permanent posts, 3% of the next 10 permanent posts and 2% of all permanent posts above 25;

· Measuring "other basic tasks" at 2% of work capacity (corresponding to about 50 minutes per permanent post and each of the statistical 42 work weeks).

In a uniform manner, the tasks "own further education and education of third parties" are measured at 4.5% of permanent posts for all judges and public prosecutors.

Activities of groups (2) and (3) only occur at specific courts or public prosecution offices.

For travelling and waiting times in connection with circuit court days, 130 minutes (district courts) respectively 162 minutes (regional courts) actual time values were obtained, each of which will be multiplied by the actual number of circuit court days held by the specific court. As before, the complication of a dual permanent post established is taken into account at 2% of work capacity at the courts involved.

2.6. Assumed Work Capacity, District Judges and Public Prosecutors

The personnel requisition (staff requirement or more precisely staff employment) results from dividing the calculated total time requirement by the defined work capacity of 1,720 annual net working hours of judges and public prosecutors (comp. the note published in relation to the regulation adopted by the Federal Ministry of Finance concerning guidelines of quantifying and displaying financial impact of new legislative measures, FLG II N^o 50/1999, concerning reference values for calculating average staff expenditure/costs, average rental costs and the calculatory interest rate; comp. FLG II N^o 126/2010).

These calculatory 1,720 hours of net annual work performance already proportionally include all average cases occurring of absence from work not replaceable under the staff plan (holidays, sick leave, therapy, leave of absence for nursing purposes, work bans); this means, that on a nation-wide average, each judge and public prosecutorworks a minimum of 1.720 net hours (in specific cases a different value may accrue, if e.g. the single judge at a small district court is on sick leave for six months, which must be disregarded though, when calculating averages).

On the district level of higher regional courts and higher regional public prosecution offices, in addition to the work capacities of the individual offices, also the capacities of the district judges and district public prosecutors must be taken into account as constituent part of work capacity available.

2.7. Public Prosecutor's Office against Corruption, Compensating for Large Proceedings

Calculations concerning the Public Prosecution Office against Corruption (KStA) established only during the PAR II survey as special public prosecution office only for specific criminal cases (comp. Section 20a Code of Criminal Procedures), considered complicated, but impossible to break down into classes, are quite difficult, while on the other hand a considerable part thereof may possibly be (re-)assigned to other public prosecutors for further processing. The calculation of national averages is based on a mix of overall cases and a specific relationship of different transaction forms, with a much smaller percentage of miscellaneous transactions and a much higher percentage of charges and indictments brought in; in both criteria the KStA significantly deviates from the average and precludes any application of averages per se. It must be considered as a coincidence in this connection, that the inclusion of KStA in the general calculation of nationwide averages calculation for the year 2009 at least produced an acceptable result, as the requirement calculated in this manner actually came close to actual staff employment numbers.

The last feature to be mentioned is the "compensation for large proceedings" valid for PAR II public prosecution offices, based on a proposal agreed between senior higher public prosecutors. This proposal provides that the total available work capacity of 17 full time equivalents at the public prosecution offices shall not be allocated in proportion to the requirement obtained from the PAR II calculation, but rather at the ratio 12:2:2:1 to the senior public prosecution districts Vienna, Graz, Linz and Innsbruck. This is supposed to account for the fact that large (investigative) proceedings preponderantly occur in the Vienna area, for which no adequate allowances would be made by PAR when calculating nationwide averages.

3. Quantification

The work groups appointed have defined class catalogues for PAR II purposes, pooling individual VJ case codes into business classes. Not all instances of PAR II classes correspond to those in the class catalogue of the first PAR. There are deviations from the presently valid breakdown of the statistical summary, so that it is not yet possible to perform a complete PAR II calculation for a period of less than a calendar year or for a unit below court level. In the introduction, mention was already made of the general problem of applying average time values to small units or single proceedings or shorter periods.

Annex VIII – Presentation of the methodology used in Norway

MAIN FOCUS ON TIMEFRAMES

The main focus on the activity in Norwegian courts is related to timeframes.Targets set by the Parliament (Stortinget) early in the 1990ies.

- Civil cases (district courts and appeal courts) should be handled within 6 months

- Criminal cases (district courts and appeal courts) should be handled within 3 months

- Single judge criminal cases (remand, voluntary guilty plea etc) should be handled within 1 month

THE PART OF THE NATIONAL BUDGET ALLOCATED TO THE COURTS IS ADMINISTERED BY THE NORWEGIAN COURTS ADMINISTRATION

- More than 9/10 of the budget is related to wages and rents (no courts own their own buildings)

- The cost of higher wages and increased rents are normally compensated/adjusted in the budget for the Judiciary.

- Main ICT-investments are covered by the ICT-budget at NCA.

* With the exception of the smallest courts, the court buildings are owned and governed by a public sector administration company called Statsbygg, and the Norwegian Courts Administration rents the court buildings from this company.

MONITORING THE ACTIVITY IN THE COURTS

Main activity in the courts: solve cases!

NCA is monitoring the activity in the courts, and the main input for our work is the case handling system, LOVISA. All activities regarding the cases that are handled by the district courts and the appeal courts are registered in LOVISA. It is an important system for the judges and clerks, to secure a good handling of each case that enters the courts.

LOVISA also provides us with data that could be used to say something about:

- The activity in the courts

- Trends in what types of cases that has to the handled by the courts

- LOVISA also provides us with information about the resources needed by each court the solve their portfolio of cases. And by resources we mean judges, clerks, courtrooms and tools to work on the cases.

For NCA

- Monitoring the caseload and flow of cases

- Yearly statistics showing the activity in each court

- Yearly statistics showing the activity in a few important case types

- Steering and control

- Important part of the dialogue with each court

- Discussions regarding workload assessment – what resources do the court need to be able to solve their cases.

- The data also provides an important knowledge base for processes concerning budgets, hearings, evaluations, revision of laws/procedures

We also provide external actors and institution with material: As general information to the public, as knowledge about the court system to other public agencies and units. We are providing facts and information to the media (almost daily). We are providing facts, data and analysis to researchers, students (weekly). We are providing facts and data to international partners (for example CEPEJ-reporting to the Council of Europe)

WORKLOAD ASSESSMENT

One important use of this statistical material is to assess the workload in each court, and to assign resources accordingly.

If we look at different member states in the CEPEJ-cooperation – this is done in several different ways.

In 2006-2007 we developed a model, based on detailed time studies in the courts, on how much time a judge and a clerk need to use in each specific phase of the resolving of a case. This gave us a model on how complex different case types were, and if different characteristics of each case or court would influence the time needed to solve a case.

So, to the main point in this presentation:

THE RESOURCE ALLOCATION MODEL – and CASE WEIGHTING

The Norwegian Courts Administration uses a model that can estimate the number of staff (judges and clerks) needed in each District Court and each Appeal Court. Results frommodel calculation constitute part of the basis for considerations with regards to changes in the number of staff, if, for instance, a court applies for more resources.

The model is based on regression analysis, and shows the relations between the number ofcases of various types, significant characteristics of the cases of each type (number of witnesses,use of interpreters, use of experts etc.) and the amount of time spent by judges andclerks to handle the cases. In the model each case type is given a particular weight accordingto a stipulated number of witnesses, experts et cetera. The stipulation is based on a timestudy done in 2006. The input data for making the model calculations are based on last yearsstatistics from the court, collected from the case management system.

THE TOTAL MANPOWER IN A COURT IS USED ON DIFFERENT ACTIVITIES

- Time studies showed that: 60% of work in courts were case-related

- 40% went to administration and other activities

AND THE CASE-RELATED PART OF THE WORK?

Of the case related work (total 62%):

• 19% in the courtroom

• 25% to preparation and writing of verdicts

• 18% to other cases (eg. Notary Public, Notarial Acts, debt settlement, bankruptcy, probate cases and so on)

Explanation: A Civil case is equivalent to two ordinary criminal cases, 8 single judge cases, et cetera. The high weight of a civil case reflects especially the out of court settlement mechanisms, and in particular the municipal conciliation boards settling approx. 140 000 civil disputes annually. It would be extremely interesting to see how civil/commercial cases are weighted compared to criminal cases in other CoE countries.

A FEW SELECTED CASE WEIGHTS:

One judge will normally produce around 1000 case related workhours each year. A civil case normally takes just over 30 hours.

If a court is given one extra judge – we should expect an increase in production of approx 35 civil cases, or 70 criminal cases, or 280 single judge cases

PRACTICAL USE OF THE MODEL:

CALCULATIONS

The model is complex, but a brief presentation would be like this: The input to the model is the number of cases incoming to each court, and the registration of a number of characteristics in Lovisa (that could be the average number of hours in court, the average number of witnesses, average number of cases with translation and so on). The time studies showed that a few of these variables would influence the time needed to solve a case.

Each year we calculate and present the calculations to the courts: based on the expected inflow of cases: how many clerks and judges do they theoretically need to solve their portfolio of cases.

Resources are not automatically distributed based on the model calculations, but it is an objective foundation for good discussions about the needs of each court.

Resources are not allocated directly on basis of the information in the case management system. To a large extent, the budget allocation is similar to previous years. However, The National Court Administration has a model that can estimate the number of staff (judges and clerks) needed in each District Court and each Appeal Court. Results from model calculation constitute part of the decision basis whenever changes in the number of staff are considered, for instance in the budget process if a court applies for more resources.

After 7-8 years, the resource allocation model is in the need for a revision, and we are looking for input on other approaches that could improve the model.

We are often met with the claim that many cases are becoming increasingly complex, and that the time estimates from 2006-2007 is no longer relevant.

Within in a few years we will probably still report on case handling time in the courts. But, in addition we will probably also look into more detail about productivity and efficiency in the courts.

PRESENTATION OF THE MODEL RESULTS

Every April we communicate this table to the courts.

- It shows the actual manpower in columns 2-4, total and divided between judges (D) and administrative personnel (S)

- It also shows the model calculations in columns 5-7. The theoretical need of manpower.

Example: Aust-Agder district court: Should, according to the calculation, have had two extra personnel (20,6/22,8) – one extra judge and one extra clerk.

[1] Regional Cooperation Council (RCC), Backlog Reduction Programmes and Weighted Caseload Methods for South East Europe, Two Comparative Inquiries, 2016. Available on: https://www.rcc.int/download/docs/Court-Backlog-Study_FINAL_za%20web.pdf/f2bdb2ae4d27f8588034538cb54b6011.pdf

[2] CEPEJ, Practical ways of combating delays in the justice system, excessive workloads of judges and case backlogs, 2005. https://rm.coe.int/1680747dc2#_ftn35

[3] The SATURN Centre emphasize that this objective is not yet developed, nevertheless, in view of the ever-improving data collection systems adopted by States, this potential objective may be considered in more sophisticated and advanced case management systems. It seems hardly reachable if there is only basic electronic case management system available.

[4] The SATURN Centre emphasize that this objective does not seem reachable in almost all judicial systems, nevertheless, in view of the ever-improving data collection systems adopted by States, this idea may be considered in more sophisticated and advanced electronic case management systems. It seems hardly reachable if there is only basic electronic case management system available.

* This designation is without prejudice to positions on status, and is in line with UNSCR 1244 and the ICJ Opinion on the Kosovo Declaration of Independence.

[5] For shorter absences, see further.

[6] For example, when the wife or the husband of a judge is an active lawyer, the case allocation system can register that in the cases where this specific lawyer is involved, the concerned judge doesn’t receive the case of is not involved in the panel.

[7] This probably isn’t the proper word, but I could not think of better expression