1. Introduction
Prospective litigants generally weigh the advantages and disadvantages of filing a civil lawsuit. To that end, they, inter alia, form expectations about the duration of court litigation. This concerns the proceedings at first instance, but a prospective litigant also has to take into account the likelihood and length of appellate proceedings and any other related litigation that could occur. This paper explains that prospective litigants are currently unable to make reasonable estimates. The data commonly used to measure the time taken for disposition are inadequate and underestimate the problem of delay in high value cases. Moreover, the spread of the time taken for disposition is often quite large, as a result of which it is difficult to make adequate predictions, even if proper data were to be available.
There has been a lot of comparative and empirical legal literature on delay in civil litigation.1 Many studies focus on mapping the (average) duration and undue delay of adjudication in one or several jurisdictions. Such empirical research has been conducted in, for example, Belgium, Germany, the United States, the Netherlands, Italy and Brazil.2 These studies show that court delay is common and widespread. Other studies have analysed attempts to combat delay by means of legal reforms, organizational changes or budget changes.3 Yet others have focused on the causes of court delay.4
Existing literature pays little if any attention to the basic question as to how the time taken for disposition is measured. Most authors use the available court data on the actual duration of court cases. These generally concern aggregated data that have been gathered by individual courts for internal case management purposes. The existing literature does not seek to measure the total time needed to resolve a dispute from the perspective of the litigants. This study fills this gap by focusing on the actual duration from the perspective of the litigants. This approach is consistent with legal requirements, such as the reasonable time requirement of Article 6 of the European Convention on Human Rights (ECHR). The ‘reasonable time’ requirement depends on numerous factors specific to the case. As a result, one cannot state that litigation should always be resolved within a set number of months or years.5 The European Court of Human Rights (ECtHR) takes all stages of the legal proceedings aimed at settling the dispute into account for the reasonable time requirement. This includes both interim and final decisions, appeal proceedings and proceedings subsequent to the judgment on the merits. The ECtHR generally regards as the starting point the moment when the action was instituted before the competent court, such as the issue of the writ. However, in exceptional cases the reasonable time may begin to run even before the issue of the writ. The end point covers proceedings concerning the final determination of costs and expenses and, if applicable, enforcement proceedings.6 Hence, the ECtHR does not only consider separate stages in the proceedings but takes the perspective of the litigants by looking at the total time needed to resolve the dispute.
This paper concerns the time taken for disposition in high value civil commercial cases in the Netherlands. The two main research questions we pose here are empirical: (1) How long do these disputes take to adjudicate from the perspective of litigants, taking into account all related procedures (i.e. what are the mean and median times taken for disposition)? (2) How much uncertainty is there with regards to the time taken for disposition (i.e. what is the spread or standard deviation of the time to disposition).
The paper is structured as follows. Section 2 discusses the measurement and presentation of the duration of court cases, as it is currently undertaken by courts and explains that aggregated figures are generally not useful for predicting the time to disposition in high value litigious cases. Section 3 sets out the research method and data. Section 4 addresses the (in)adequacy of court statistics for measuring the duration of court proceedings from the perspective of litigants. Section 5 discusses courts’ lack of methods for registering consecutive court proceedings pertaining to the same dispute. Section 6 presents the duration of distinctive phases of the proceedings and seeks to make a (conservative) estimate of the spread of the total time needed before a dispute is resolved. Section 7 concludes that existing court statistics and literature underestimate the problem of court delay from the perspective of litigants. Moreover, the existing literature fails to appreciate that the spread of the total time taken for disposition is far bigger than official statistics suggest. As a result, prospective litigants face more uncertainty than previously thought.
2. Available aggregate data on duration are not always useful
The measurement of disposition time is problematic in many jurisdictions at an elementary level. In these jurisdictions, case registration systems apparently do not generate precise data on the duration of individual cases. The European Commission for the Efficiency of Justice (CEPEJ) is part of the Council of Europe. It publishes extensive data on the judicial systems of the countries of Europe. CEPEJ observes that the average time needed for case resolution is still difficult to obtain ‘in most states or entities’. As a consequence, it estimates average duration by means of aggregate data. It defines disposition time as ‘the number of pending cases at the end of a particular period [divided] by the number of resolved cases within that period, multiplied by 365’.7 To illustrate, if the number of pending cases is larger than the number of resolved cases, the disposition time is longer than a year. This method of calculation provides a general insight into duration, but it does not lead to accurate estimates of average duration under all conditions. What is worse is that it does not give any information about the spread of duration (e.g. the standard deviation).
In addition, CEPEJ uses the concept of litigious cases to filter out cases in which, for instance, a default judgment is rendered.8 As such, this makes sense as uncontested cases are generally resolved in very little time. These cases are less interesting from the perspective of dispute resolution by the courts. However, this concept adds an element of subjectivity to the definition and makes it difficult to compare data with national statistics. The underlying issue is that cases differ in subject matter, size and complexity and in the procedural steps taken. Even if the average duration of all litigious cases is relatively short, it may well be that a substantial part of the cases takes a very long time to reach an end. The actual issues about duration may remain concealed. Also, empirical data generally fail to keep track of time if disputes lead to more than one procedure and/or judgments are appealed.
Even in jurisdictions where data on individual cases is available, typically only averages per main category of cases are published. Categories can be as broad as criminal and civil cases.9 It is exceptional that data on the distribution of duration is published. In the Netherlands, the courts have a comprehensive and detailed case registration system. The judiciary provides some information on distribution. According to these figures, duration longer than two years occurs in 10% of commercial cases at the civil departments at first instance and in 22% on appeal. Tables 1 and 2 show that there are large differences between small claims cases where the value in dispute is generally below EUR 25,000 and commercial cases in which the claim exceeds EUR 25,000.
Table 1
Duration commercial cases in days, Netherlands, 2018.
| SOURCE | TYPE OF DISPUTE | MEAN AND SPREAD | FIRST INSTANCE | APPEAL |
|---|---|---|---|---|
| CEPEJ | All litigious cases | Mean | 210 days | n.a. |
| Council for the judiciary | Small claims department (<25,000 euro) | Mean | 49 days | n.a. |
| Spread: % resolved within 6 months | 72% | |||
| Spread: % resolved within 1 year | 93% | |||
| Commercial department (>25,000 euro) | Mean | 266 days | 462 days | |
| Spread: % resolved within 1 year | 62% | 39% | ||
| Spread: % resolved within 2 years | 90% | 78% | ||
Source: CEPEJ 2022, p. 126 and Netherlands Council for the judiciary 2020, pp. 63–65.
Table 2
The Netherlands: percentage of civil and commercial cases at first instance that is resolved within the internal standard for the time to disposition in 2021.
| TYPE OF CASE | INTERNAL STANDARD | PERCENTAGE |
|---|---|---|
| Small claims, default cases | 14 days | 75% |
| Commercial claims, default cases | 42 days | 75% |
| Small claims, contested cases | 140 days | 43% |
| Commercial department, contested cases, single judge | 252 days | 38% |
| Commercial department, contested cases, panel of three judges | 280 days | 11% |
Source: 2021 Annual Report Judiciary (Jaarverslag Rechtspraak 2021, Table 10).
The Dutch judiciary recently started to publish data that clarify what percentage of all cases is resolved within its own internal standards for the time taken for disposition. Uncontested cases and low value disputes are often resolved swiftly. Complex contested claims present the biggest challenge. This is particularly so for commercial cases that are decided by a panel of three judges rather than a single judge. These are the cases that the judiciary itself considers of such complexity or importance that they warrant a decision by three judges (approximately 7% of all commercial cases). Such cases are rarely resolved within the internal standard for the time to disposition.
Table 3 provides an overview of all first instance and appellate commercial cases in the Netherlands in the period 2016–2018. These are not based on estimates, but on actual data on the time taken for disposition gathered by the judiciary for each individual case. The vast majority of cases at first instance courts concern small claims, and these cases, to a large extent, determine the overall average duration. These small claims include many debt collection cases in which the claim is not contested and in which the court renders a default judgment within a couple of weeks. The average time taken for disposition increases with the claim’s value. This is also the case with appeals, but less markedly than at first instance. The variation of the time taken for disposition is lower than at first instance, in particular because cases are almost always contested. In should be noted that the appeal rate is high for high value cases.10 In those cases, litigants are drawn into prolonged litigation.
Table 3
Volume and duration in days of commercial cases in the Netherlands’ courts, average per year of 2016, 2017 and 2018.
| AMOUNT OF THE CLAIM IN EUR | FIRST INSTANCE | APPEAL | ||
|---|---|---|---|---|
| VOLUME | TIME | VOLUME | TIME | |
| 0 < c < 1,000 | 268,500 | 20 | 33 | 386 |
| 1,000 ≤ c < 10,000 | 86,450 | 52 | 746 | 432 |
| 10,000 ≤ c < 100,000 | 17,360 | 166 | 1.510 | 501 |
| 100,000 ≤ c < 1,000,000 | 2,869 | 346 | 688 | 592 |
| 1,000,000 ≤ c < 5,000,000 | 388 | 513 | 141 | 661 |
| 5,000,000 ≤ c < 10,000,000 | 65 | 527 | 20 | 601 |
| 10,000,000 ≤ c < 20,000,000 | 27 | 478 | 12 | 557 |
| 20,000,000 ≤ c < 100,000,000 | 25 | 667 | 10 | 629 |
| c ≥ 100,000,000 | 11 | 731 | 2 | 764 |
| Claim value undetermined | 21,994 | 168 | 1,946 | 408 |
| Total | 397,688 | 45 | 5,107 | 474 |
Source: Costello et al. (2021), based on data provided by the Netherlands National Service Centre for the Judiciary.
The same data have been collected for four other countries, Norway, Lithuania, Italy and Ireland, in addition to the Netherlands.11 In these countries, the time taken for disposition also increases with the value of the claim, but in different degrees. That study also shows that country comparisons can go wrong. In Norway, for instance, fewer small cases go to court. The data underline the importance of differentiation.
Obviously, there is nothing wrong with the publication of average data on duration gathered by courts. On the contrary, it is essential that such data are available. From the perspective of a litigant who wants to know how long his/her specific case will take to adjudicate, averages for very broad categories of cases are, however, insufficient, even in jurisdictions in which reliable information is available. This is certainly true for cases that really matter in economic terms or that are considered to be of great significance by the courts themselves. In this paper, we will take a deep dive into the measurement of the duration of larger commercial cases.
3. Research method
To examine our research questions, we have constructed a dataset of 89 civil commercial disputes with a financial claim between 1 to 5 million EUR. These are moderately high value cases in which litigants are generally willing to invest resources. At the same time, the number of cases in this range is sufficiently high to conduct empirical research (see Table 3). Our dataset only includes cases that ended in a final judgment at first instance and which have led to an appeal before one of the appellate courts.
The case administration system used by the courts provides detailed information about all steps taken in a procedure. The Courts of Appeal register the case number at first instance. As a result, we were able to link the proceedings at first instance with those on appeal. For the cases in our sample, we could establish the total time taken for disposition, and detail each of the steps in the procedures and their respective durations.
We selected cases that were withdrawn, settled or decided in 2017 or 2018 by the Courts of Appeal, by moving backwards from the last case that was adjudicated in 2018. We omitted cases if we were unable to obtain sufficient data to conclude our analysis. This included cases in which we were unable to trace detailed data on the proceedings at first instance. We also omitted cases in which we were unable to find at least one (public) court decision on the website of the judiciary. These published decisions enabled us to better understand the nature of the dispute and the precise course of the proceedings. The final data set consists of 89 cases. Of these cases, 13 were terminated at the level of the Courts of Appeal in the last month of 2017. The other 76 cases were terminated in 2018. Our dataset is composed of more than half of all appellate cases in this time period with a claim between 1 and 5 million EUR. Even though the selection is not random, we have found no indications of substantial selectivity.12
As the variation of the time taken for disposition is very large, our sample contains first instance procedures that started at very different moments in time, e.g. 2010 or 2016. Throughout the years, there may have been minor changes to local court rules or practices. Backlogs may have increased or decreased. The cases are, in that respect, not fully comparable. In practice, however, there do not seem to have been substantial changes to procedural law or the time taken for disposition in the last fifteen years.13
We examined the first and second instance procedures in all cases. We also examined which cases were subsequently appealed at the Supreme Court. We have further tried to identify other directly related procedures or remedies that concern the same dispute. We found related proceedings in 21 cases. The judiciary does not adequately register when distinct proceedings relate to the same dispute.14 Our detailed search for related cases has probably not been fully exhaustive. In at least two cases, litigation still seems to be ongoing.
4. Court statistics must be used with caution to understand the perspective of litigants
Before turning to the distribution of duration, we examine some measurement issues. In our dataset, the average time taken for disposition as registered by the courts amounts to 653 days at first instance and 709 days on appeal.15 The dataset contains cases that are settled on appeal. Appellate cases that result in a final judgment take on average 844 days to resolve. Does this mean that (prospective) litigants can assume that disputes in which appellate proceedings are initiated are on average resolved in 1,362 days or, if a final judgment is desired, 1,497 days (i.e. approximately four years)? The short answer is: no.
It must be borne in mind that the available court data are registered primarily for internal management purposes. This makes perfect sense: courts primarily keep track of the data to improve their own internal procedures. These data are not meant to measure the total time taken for disposition from the perspective of litigants. Here, we are interested in the overall time it takes to resolves a dispute, instead of the combined time of the court procedures as such. The combined statistics on the duration at first instance and on appeal do not suffice to measure the overall time that is needed to resolve a dispute. We provide three examples.
First, in the Netherlands, litigation is commenced by serving a writ of summons. The writ is served some time before the date on which the case is brought before the court. The claimant’s writ of summons identifies the precise date on which the case will be introduced before the court. The claimant must serve the writ at least one week before that date. The statutory time period that must be observed against many foreign defendants is much longer: at least four weeks or in some cases even three months.16 We know the precise date on which the writ of summons was served in 26 out of 89 cases. Our empirical findings show that the ‘waiting time’ between the service of the writ and the date on which the case is registered in the court system in these cases was on average 49 days. During this time period, the case is formally pending, but has not yet been registered at the court. The date of the writ of summons is not registered in the court system.17 Courts only measure duration from the moment the case is first registered in the court system. This makes sense as the courts cannot possibly exert influence on the proceedings before that. From the perspective of the litigants, the official registered data thus underestimate the time taken for disposition in high value cases by approximately 49 days.
Second, courts do not register the time between the judgment at first instance and the date on which the case is first registered at a Court of Appeal. In most civil cases, the term to file an appeal by serving notice of appeal is three months. The moment the notice of appeal is served marks the beginning of the appellate proceedings. As is the case at first instance, it takes some time before the case is subsequently brought before and registered with the Court of Appeal. Our empirical data show that the ‘waiting time’ between the judgment at first instance and the registering of the case on appeal is on average 162 days.18 The Court of Appeal only measures time taken for disposition from the moment the notice of appeal is registered. From a court management perspective, this makes sense; after all, these are matters that are not within its circle of influence. From the perspective of the parties, however, the combined time taken for disposition at first instance and on appeal based on the available court statistics, underestimates the time taken for disposition by another 162 days.
Third and more complicated, at least nine cases in our dataset were interrupted for a longer period of time. Reasons for a long adjournment of cases are numerous. These can, for instance, include a request by both parties to halt the procedure in order to reach a settlement through mediation. In such cases, it may very well be that the case will never be resumed at all. As a result, the court considers the case to be concluded. Once one of the parties wishes to resume the case, the case is sometimes resumed under the existing case number. In at least four cases, however, it was given a new case number and treated as a new case for purely administrative purposes. What basically happens in these cases is that one long court procedure is split into two shorter parts for court management purposes.
We illustrate the above by means of an example from our data set. An appellant files an appeal on 13 August 2013. During the proceedings, the appellant goes bankrupt. On 17 February 2015, the Court of Appeal renders an interim judgment in which it finds that the law prescribes that the case must be put on hold for an undefined period of time and will thus be removed from the list of pending cases.19 For administrative purposes, the case is thus considered to have been ‘resolved’. After all, a very large percentage of legal proceedings against bankrupt debtors are never resumed. On 25 August 2015, however, this case is resumed and entered into the administration system as a new case. A final judgment is rendered on 1 May 2018. In this matter, we are thus dealing with one prolonged appellate procedure that lasted for almost five years. The court administration system, however, registered two distinct proceedings with far shorter times to disposition. The administrative data do not consider the ‘waiting time’ of approximately six months between these two distinct (parts of the) proceedings.20
This issue of reregistering a pending procedure under a new case number may make perfect sense from the perspective of the courts. However, from the perspective of litigants, it means that the court data underestimates the ‘real’ time taken for disposition. In 9 out of 89 cases, the lower court or appellate court essentially ‘split’ the case into two parts. If one considers the two smaller parts as one single uninterrupted procedure, the average time taken for disposition increases significantly. For the entire dataset, the average time taken for disposition increases by 74 days. In the interrupted cases, the waiting time between the two distinct parts was on average 247 days. This waiting time by itself significantly affects the average time taken for disposition of the entire dataset by (yet another) 22 days. We have adjusted the dataset for these interrupted cases in the sense that we have considered the two distinct parts as a single procedure. The waiting time has been registered separately.
Above we gave three examples that, by themselves, imply that one must be careful to interpret the available data on time taken for disposition. These three examples already indicate that a litigant would have to make an upward adjustment of more than 20%. The three examples may, to some extent, concern peculiar aspects of the procedural system or administrative practices in the Netherlands. As the definition of duration is broadly the same, at least in European jurisdictions,21 we have every reason to assume, however, that these examples reflect a broad issue that is relevant for other jurisdictions as well. Court statistics serve primarily internal management purposes and, for obvious reasons, tend to disregard matters that fall outside the scope of control of a particular court.
5. Courts do not register consecutive court proceedings pertaining to the same dispute
5.1 Introduction: procedures concerning the same dispute
Section 4 showed that the average duration of court proceedings from the perspective of (prospective) litigants is higher than one would assume on the basis of the available statistics. This section discusses a more fundamental problem when interpreting court statistics. Courts do not register all consecutive court proceedings pertaining to the same dispute. There is little, if any, court management beyond the individual court. This means that it is difficult to ascertain whether one single dispute leads to various separate legal proceedings between the same parties. While the courts may see this as beyond their responsibility, the accumulation of procedures is of paramount importance for the parties.
The Dutch civil procedural framework provides parties with ample possibilities to bring several distinct procedures before the first instance courts that relate to the same dispute. Examples include summary proceedings to obtain a provisional or protective measure (kort geding) and proceedings to freeze assets by means of a pre-judgment attachment (conservatoir verhaalsbeslag). The rules of procedure also allow the claimant to divide the procedure in the sense that the procedure initially only concerns liability. Once liability is established and parties cannot agree on the damages, the claimant must start a new procedure to determine the exact damages (schadestaat procedure). It is not uncommon that, once a dispute is decided, new disputes arise that relate to the enforcement of the final decision (executiegeschillen).
Moreover, the legal system provides for several remedies. For present purposes, three are most relevant. First, parties are generally entitled to a de novo review on appeal. As a result, first instance decisions are often challenged on appeal, particularly in high value disputes.22 Second, in most civil cases, parties may, as a matter of right, challenge appellate judgments before the Dutch Supreme Court.23 The Supreme Court will however not fully re-examine the case. The Supreme Court only examines whether a lower court correctly applied the law. In high value cases, it is nevertheless not uncommon for parties to make an appeal to the Supreme Court. Third, revocation is an extraordinary remedy in the case of procedural fraud. A judgment can be revoked if it is later discovered that the judgment is based on deceit committed by the opposing party, or is based on forged documents, or if relevant documents have been withheld by the opposing party.24 The revocation procedure takes place at the court that delivered the judgment. This could be the District Court or the Court of Appeal. Once a judgment is rendered in revocation proceedings, one can file ordinary remedies against such a judgment.
As a result of the numerous possibilities to bring different actions and to file different remedies, there can be many different domestic proceedings that all concern the same dispute. We have searched for domestic proceedings that concern the same dispute. Our search could not possibly have been exhaustive as, in fact, it may well be that cases in our dataset will give rise to new proceedings that will only be filed tomorrow.
5.2 Separate proceedings on the quantum of damages
As noted above, the claimants can opt to bifurcate the proceedings in the sense that the procedures initially only consider liability. Once liability is established but parties cannot agree on the damages, the claimant will have to file a new action to determine the quantum of damages.
Our dataset includes four procedures on the quantum of damages that were preceded by procedures that concerned liability. For example, one straightforward dispute consisted of the following procedures: (i) liability procedure at the District Court (511 days), (ii) liability procedure before the Court of Appeal (581 days), (iii) damages procedure before the District Court (476 days) and (iv) damages procedure before the Court of Appeal (469 days). These figures are the registered time taken for disposition. Again, the waiting time between these separate procedures is not recorded. In this case, the time between final resolution of the liability procedure and the start of the procedure to determine the quantum of damages was very long: approximately three years. This period may have been used by the parties to negotiate a settlement. Obviously, this delay cannot reasonably be attributed to the courts. However, if one takes everything into account, the time needed to resolve the dispute, from the moment the case was first introduced at first instance until the moment the last appellate judgment was rendered, was approximately ten years.
As well as those four cases, our dataset includes two cases in which the operative part of the final judgment orders the defendant to pay damages, but in which the precise amount would have to be determined at a later point in time in separate damages proceedings. These damages procedures have not (yet) started. The cases may have been settled.
It may well be that there are relatively few cases in which there were separate proceedings on the quantum of damages (4 out of 89). In these cases, however, the delay that results from such separate proceedings is substantial.25 If one were to consider the separate proceedings in these four cases, this would require an upwards adjustment of the average registered time taken for disposition of the entire dataset by approximately 5%. The figure would be even higher if the waiting time between the distinct proceedings were to be considered. More important, however, is that these cases also significantly impact the ‘spread’ of duration (see Section 6).
5.3 Appeals to the Dutch Supreme Court
Appeals to the Dutch Supreme Court are very rare. However, such appeals are more common in high value litigation, such as the cases in our dataset. Appeals to the Dutch Supreme Court were made in 12 out of the 89 cases in our dataset (13%). If we only consider cases that resulted in a final judgment on appeal, it concerns 12 out of 62 cases (20%). This implies that the odds that high value appellate judgments will indeed be appealed are substantial.
The Supreme Court itself works relatively quickly and rarely takes more than two years to resolve a dispute. The Supreme Court upheld the appellate judgment in seven cases (58%). That means the appellate judgment becomes final and binding at a later point in time. On average, the appeal in these cases took 568 days. This covers the total time measured from the moment the appellate judgment is rendered until the judgment of the Supreme Court.
The Supreme Court can also annul an appellate judgment and refer the case back to the same or another Court of Appeal. The case will be retried, taking into account the judgment of the Supreme Court. This happened in five cases in our database. In such cases, a long delay seems inevitable. These Supreme Court proceedings took on average 580 days. The waiting time between the Supreme Court judgment and the commencement of the referral proceedings can be quite substantial. It could be measured in two out of five cases. In these cases, it amounted to an average of 218 days. The referral proceedings themselves take on average 410 days. In theory, the appellate judgment in referral proceedings can yet again be challenged before the Supreme Court. The figures that we present concern conservative minimum estimates. In some cases where we were able to find administrative data on the referral procedures, the referral proceedings seem to continue to this present day.
Table 4
Recourse to the Supreme Court.
| CASSATION PROCEDURES | n | MEAN DURATION IN DAYS | |||
|---|---|---|---|---|---|
| CASSATION | REFERRAL | WAITING TIME | ADDITIONAL TIME | ||
| Appeal to Supreme Court without a referral | 7 | 568 | 568 | ||
| Appeal to Supreme Court, and referral to Court of Appeal | 5 | 580 | 410 | 218 | 1,208 |
In the twelve cases discussed above, the appeal stretches the total time needed to resolve the case. If this were to be accounted for, it leads to an increase in the average time taken for disposition in the entire dataset by approximately 6% to 7%. Supreme Court litigation also significantly impacts the ‘spread’ (see Section 6).
5.4 Revocation procedures
As indicated above, parties may institute proceedings for the revocation of one or more judgment(s). Such proceedings are rare but appear to be more common where the financial stakes are large. One example in our database concerned a case in which the registered time taken for disposition at first instance and on appeal was very time consuming, namely 2,968 days. The appellate judgment that was ultimately rendered became final and binding. Revocation proceedings can be brought against such final decisions if deceit or procedural fraud is discovered at a later point in time. In this case, the time between the moment when the final judgment was rendered and the procedural fraud was allegedly discovered was very long. The revocation proceedings commenced 1,266 days after the appellate judgment was rendered. The revocation proceedings themselves took 532 days. From the perspective of the litigants, the dispute thus suddenly reemerged long after it was allegedly resolved.
Revocation procedures have been filed in at least 3 out of the 89 cases in our dataset. This figure may be higher in practice. We may have missed revocation proceedings. It is also conceivable that new revocation proceedings will be filed in the future against final appellate judgments in our dataset.
Be that as it may, revocation proceedings remain scarce events, even in high value cases. At the same time, once these proceedings are filed, they can cause significant delay. This is particularly so if remedies are filed against a judgment rendered in such revocation proceedings. As a result, if one were to take revocation proceedings into account, this would impact both the average time taken for disposition as well as the spread (i.e. standard deviation) of the entire dataset. This is even more the case if one were to account for the lengthy ‘waiting time’ between the final judgment and the moment the revocation proceedings are commenced.
5.5 Foreign proceedings relating to the same dispute
We emphasize that the data that we have at our disposal are very rich, detailed, and accurate. Nevertheless, we are unable to properly search for all procedures that relate to the same dispute. This is the case for domestic procedures, but even more so with regard to proceedings that have been filed abroad.
It is well known that some legal disputes are brought before the courts in numerous jurisdictions. Sometimes parties dispute the international jurisdiction of a court.26 One Dutch example concerns a case that commenced at a Dutch District Court in 2006 and was ultimately resolved by the Dutch Supreme Court in 2017. The Supreme Court determined that the case should not have been brought before courts in the Netherlands.27 In such a case, the claimant should thus commence an action before another court; in this matter it was the competent court in Germany. The Dutch Supreme Court’s administration system quite understandably considered the case closed and resolved in 2017. The case may have been brought elsewhere. Dutch courts, however, cannot and do not keep track of potential follow-up procedures in other jurisdictions.
At other times, a claimant may have first prevailed in litigation elsewhere and may have obtained a favourable judgment. If the defendant fails to pay the amount due and payable, the claimant may seek to enforce the judgment, for example in the Netherlands. Occasionally, this leads to lengthy proceedings on the recognition and enforcement of foreign judgments.28 Dutch courts may, to some extent, be aware of earlier foreign proceedings, but they do not and in most instances cannot store and register information that pertains to these foreign proceedings in their administrative systems.
It is clear that many cases in our dataset are not purely domestic disputes. In many cases, for example, parties to the dispute are foreign entities. Occasionally, it is also possible to infer from the public decisions that the case has a clear international dimension. One case in our database concerned a shipping accident in the White Sea involving vessels from Norway, Liberia and Russia. The court decisions make it clear that there had been earlier litigation abroad concerning the very same accident.29
Our data do not enable us to draw reliable conclusions on related foreign litigation. It may well be that judgments rendered in cases in our dataset have led to enforcement litigation in other jurisdictions. Data on such follow-up procedures are, however, not readily available. It is clear, however, that foreign litigation can prolong the time taken to resolve disputes from the perspective of litigants. As increasingly disputes have an international dimension, national statistics capture a decreasing part of dispute resolution.
5.6 Implications for prospective litigants’ expectations
The data show that, in high value cases, it certainly is not exceptional that there are several different court procedures that prolong the resolution of the dispute from the perspective of litigants. In 21 out of the 89 cases in our dataset, there have been other domestic proceedings apart from the proceedings at first instance and on appeal. If one were to consider the time taken for disposition of these additional proceedings, the average time taken for disposition of the entire dataset would increase by approximately 239 days. That corresponds to an increase of the average registered time taken for disposition of approximately 17%. This does not include additional unregistered waiting time between these separate proceedings.30 These figures constitute conservative estimates as we have probably not been able to identify all related domestic proceedings, nor have we corrected for any related foreign proceedings.
The above means that the available court administration data may be very reliable and accurate, but have inherent limitations that make them less suitable to estimate the time needed to resolve the entire dispute. In Section 6 we will present standard measures of duration and the spread of duration for the whole data base of 89 cases.
6. The spread of the time taken for disposition: does uncertainty prevail?
6.1 Introduction
This section discusses the spread of the time taken for disposition of large commercial court cases in one instance, as well as taking into account all instances and proceedings. It first addresses the time involved in individual steps in a procedure in a single instance (Section 6.2). It is clear that the spread of the duration of separate stages in the proceedings is relatively large. It should be noted that a large spread does not always imply that the time taken for disposition is unpredictable. After all the parties can exert at least some control over the course of the proceedings (Section 6.3). When taking all litigation relating to the dispute into account, the overall spread is very substantial (Section 6.4). A significant number of disputes are not resolved in four or even six years. In many cases, it will be difficult for prospective litigants to make a proper estimate of the duration of future litigation.
6.2 Uncertainty and the different phases of a procedure
Based on an in-depth analysis of the administrative court data, a breakdown of the corrected duration of a procedure in one instance can be made. Table 5 provides the time needed for the most important phases of the procedure in the 89 cases in our database. The average duration is predominantly determined by the procedural steps that occur in most cases: (i) the initial exchange of written submissions, (ii) waiting for the hearing and the hearing itself, and (iii) the preparation of (interim) judgment(s). Possible side steps include the taking of evidence and interim motions. These interim motions mainly concern procedural matters, such as the joinder of cases or the discovery of documents. These other steps usually take considerable time, but do not occur often enough to dominate mean processing time. The category ‘other’ primarily concerns instances in which the procedure is temporarily paused, for instance due to bankruptcy of one of the parties or settlement attempts by the parties.
Table 5
Breakdown of the adjusted32 disposition time in commercial cases.
| PHASE | AVERAGE (89 CASES) | |
|---|---|---|
| FIRST INSTANCE | COURTS OF APPEAL | |
| Written arguments | 185 days | 216 days |
| Oral hearing(s) | 137 days | 225 days |
| Interim motions | 40 days | 28 days |
| Evidence | 54 days | 36 days |
| Other | 67 days | 47 days |
| Voluntary dismissal | Not applicable | 29 days |
| Judgment(s) | 178 days | 193 days |
| Total | 661 days | 775 days |
Table 6 shows the average and median processing time for each phase, if this phase occurs.31 Evidentiary activities require a large amount of time and are an important reason why the duration of cases shows a huge variation. The duration at first instance in a simple procedure, consisting only of written arguments, an oral hearing and judgment, would take approximately 526 days at first instance and 838 days on appeal. If the procedure is more complicated and involves e.g. the examination of witnesses, the proceedings will be prolonged by approximately one year.32
Table 6
Duration in days of a procedural phase, if this phase actually occurs.
| PHASE | NUMBER | MEAN | MEDIAN | ST. DEV. |
|---|---|---|---|---|
| First instance | ||||
| Written arguments | 88 cases | 187 days | 154 days | 137 days |
| Oral hearing | 76 cases | 161 days | 143 days | 90 days |
| Interim motion | 35 cases | 135 days | 98 days | 136 days |
| Evidence | 13 cases | 369 days | 307 days | 272 days |
| Other | 18 cases | 330 days | 134 days | 559 days |
| Judgment(s) | 89 cases | 178 days | 127 days | 174 days |
| Appeal | ||||
| Written arguments | 77 cases | 250 days | 196 days | 212 days |
| Oral hearing | 61 cases | 328 days | 294 days | 177 days |
| Interim motion | 13 cases | 199 days | 168 days | 131 days |
| Evidence | 9 cases | 358 days | 301 days | 251 days |
| Other | 14 cases | 298 days | 236 days | 246 days |
| Voluntary dismissal | 27 cases | 97 days | 70 days | 125 days |
| Judgment(s) | 62 cases | 260 days | 197 days | 209 days |
It is important to note that the standard deviations in Table 6 are relatively large. This indicates that the duration of each phase shows a wide variety. This makes it difficult for parties to make an accurate prediction at the outset of how long each step in the proceedings will take, if it occurs. As will be discussed below, however, parties can seek to influence the course of the proceedings and thus reduce uncertainty.
6.3 Can parties influence the time to disposition and reduce uncertainty?
To what extent can courts and individual parties influence the duration of cases? The responsibility for the length of procedures differs according to the phase of the procedure. The time taken to exchange written submissions is primarily the responsibility of the parties and their lawyers. The last phase, delivering judgment, is the responsibility of the court. Other stages in the proceedings are a mixed responsibility. For instance, the waiting time for the hearing(s) is determined by several factors, including the capacity of the courts, parties’ availability for a hearing and their (often joint) requests for postponement. Scheduling is more complicated when there are more parties. Evidentiary issues and interim motions are generally initiated by the parties and their lawyers. However, the speed with which procedural decisions are taken or experts are appointed and how they are monitored depends largely on the courts. ‘Other’ complications arise if parties ask for time to negotiate a settlement or if one of the parties goes bankrupt. Such developments are generally beyond the influence of the courts. Moreover, the parties decide whether to go on appeal or to accept the judgment.
Table 7 simplifies matters by allocating stages in the proceedings to either the parties, the court or by dividing them. We assume that the courts are largely responsible for the oral hearing (75%) and that parties and courts are equally responsible for interim motions and evidence (50%). The figures below are an oversimplification of reality, but do show that the division of responsibility for the long duration of cases is more or less equally split between the parties and the court.
Table 7
Contribution by parties and courts to duration in days of cases that end in a final judgment on appeal (62 cases).
| PHASE | MEAN FIRST INSTANCE | MEAN APPEAL | ||
|---|---|---|---|---|
| PARTIES | COURT | PARTIES | COURT | |
| Written arguments | 170 days | 253 days | ||
| Oral hearing | 33 days | 98 days | 73 days | 218 days |
| Interim motions | 23 days | 23 days | 17 days | 17 days |
| Evidence | 14 days | 14 days | 25 days | 25 days |
| Other | 80 days | 43 days | ||
| Judgment | 159 days | 260 days | ||
| Total | 320 days | 294 days | 411 days | 520 days |
| Percentage of total duration | 52% | 48% | 44% | 56% |
Thus, both parties with their lawyers and the courts share responsibility for duration. The influence of the parties goes beyond the duration of the individual stages of the proceedings, as they decide to file interim motions or requests regarding evidence. Beyond their influence on individual procedures, the parties decide whether they will settle the case or will initiate (follow-up) litigation. It is the parties that decide whether they will file remedies. In the Netherlands, parties are free to make a binding agreement to skip the court at first instance and raise a dispute directly before the competent Court of Appeal.33 Parties can also mutually agree to waive their right to file remedies: they can waive the right to make an appeal to the Court of Appeal and/or the Supreme Court.34 If the parties cooperate, they should in principle be able to avoid undue delay.
However, as parties generally have conflicting interests, it is conceivable that only one of them will seek a swift resolution of the dispute. Defendants may have little reason to avoid delay. They might even have an incentive to cause delay. A party can by itself exert influence, but it cannot unilaterally ensure a swift resolution of the dispute. If the parties cannot reach a joint agreement on the swift resolution of the dispute, uncertainty prevails. To narrow down its prediction of the time to disposition, a party will have to guess which litigation strategy the other party will use.
Of course, courts can intervene at a higher level by changing the procedural rules and by strictly applying them. They can largely determine the time granted for the exchange of written arguments and they have case management powers to further ensure a swift resolution of interim motions and the taking of evidence. In principle, they can reduce the effectiveness and thereby the attractiveness of delay strategies for the parties. Also, by the quality of the material judgments they affect the incentives to lodge appeals. Nevertheless, overall duration is highly dependent on the litigation strategies of the parties.
6.4 Uncertainty: combined spread of duration at first instance and on appeal
The combined time taken for disposition of the cases in our dataset, as registered by the courts, equals 653 days at first instance and 709 days on appeal (see Section 4 above). Histogram 1 below shows the spread of the total registered time taken for disposition measured in years. Table 8 gives the statistics for the combined duration at first instance and on appeal. See Annex 1 for the duration per case. According to these figures, 59 out of 89 cases are resolved within four years. This is an important benchmark: many legal scholars agree that more than two years of litigation per instance will generally violate the ‘reasonable time’ requirement in Article 6 ECHR.35
Registered data spread of duration, number of cases resolved in less than x years.
Table 8
Mean and standard deviation of time taken for disposition in all cases and cases resolved by means of a final judgment on appeal (n = 89 resp. 62).
| ALL CASES (n = 89) | CASES RESOLVED BY FINAL JUDGMENT (n = 62) | |||
|---|---|---|---|---|
| MEAN | STAND. DEV. | MEAN | STAND. DEV. | |
| Registered time taken for disp. | 1,363 days | 695 days | 1,451 days | 622 days |
| Adjusted time taken for disp. | 1,436 days | 736 days | 1,547 days | 668 days |
| Adj. time incl. other proc. | 1,675 days | 909 days | 1,849 days | 866 days |
| Adj. time incl. other proc. and waiting time | 1,938 days | 968 days | 2,101 days | 948 days |
As explained above, we have made adjustments with regard to nine cases that had been ‘split’ into two parts for administrative reasons. These adjustments led to an increase in both the average time taken for disposition and the standard deviation (respectively 1,436 days and 736 days). Histogram 2 shows the adjusted data. The number of cases that are resolved within four years drops from 59 to 54.
Adjusted data spread of duration, number of cases resolved in less than x years.
As discussed above, we have been able to identify related domestic proceedings, other than the proceedings at first instance and on appeal, in 21 of the 89 cases in our dataset. If we take the adjusted data discussed above and include the registered time in these related proceedings, the average time taken for disposition increases to 1,675 days. The standard deviation increases to 909 days. Histogram 3 shows the spread. The number of cases resolved within four years drops to 44 out of 89 cases. The ‘tail’ thus gets ‘fatter’.
Adjusted data incl. related litigation spread of duration, number of cases resolved in less than x years.
Finally, as previously discussed, courts only register cases from the moment these are registered at the court until the moment a judgment is rendered or the case is otherwise terminated. Courts do not register the time between procedures. The unregistered waiting time concerns (i) the time between the service of the writ of summons and the registration at the court at first instance, (ii) the time between the judgment at first instance and the registration at the appellate court, (iii) the time between different parts of a procedure that has been split into two parts for administrative purposes and (iv) the time between different proceedings that concern the same dispute. If we take the adjusted data and include related litigation and waiting time, the average time taken for disposition increases to 1,938 days and the standard deviation increases to 968 days. Histogram 4 depicts the spread. The number of cases resolved within four years drops to 36 out of 89 cases. If we consider waiting time, the ‘tail’ gets ‘fatter’. In 33 out of 89 cases it will take more than six years to resolve the dispute.
Adjusted data incl. related litigation and waiting time spread of duration, number of cases resolved in less than x years.
We note that the data above include all cases in our dataset. These include cases that have been settled or otherwise terminated before a judgment was rendered on appeal. Cases in which a final judgment is rendered take longer to resolve and have a slightly smaller spread, as is shown in the table below.
7. Discussion and conclusions: duration and spread of duration
Prospective litigants generally weigh the advantages and disadvantages of filing a civil lawsuit. To that end, they need to form expectations about the duration of court litigation. Existing literature pays little if any attention to the basic question as to how the time taken for disposition is measured. Most studies use the available court statistics to calculate or estimate the average time taken for disposition at a highly aggregated level. We conclude that these data are inadequate for prospective litigants to make an accurate prediction of the time taken to resolve a dispute. Our reasons for this conclusion are the following.
First, available court statistics tend to utilize aggregate data on all civil cases. As small and relatively low value cases are numerous, these tend to have a large impact on the registered median and mean time taken for disposition (see Section 2). As a result, court statistics might be inadequate for understanding the duration of complex high value cases.
Second, available administrative systems register the time taken for disposition for internal management purposes. Such data are helpful to monitor individual cases or the performance of the courts over time. For a number of reasons, however, the courts’ registration does not provide a realistic estimate of the time taken for disposition from the perspective of the parties. Courts, for example, split procedures for administrative reasons and they do not measure the time between the procedure at first instance and on appeal. As a result, the registered data underestimate the time needed to resolve a dispute (see Section 4).
Third, courts do not and sometimes cannot systematically keep track of distinct court procedures that relate to the same dispute. Our research shows that it is quite common for there to be several proceedings with the same parties that concern the same subject matter. We have related domestic proceedings in 21 cases out of the 89 cases in our database. Related litigation has a large effect both on the median and average time taken for disposition as well as the spread thereof (see Section 5). From the perspective of the functioning of the judicial system, the question may be raised whether the possibilities for the initiation of appeal and/or other follow-up procedures should be reduced to speed up dispute resolution and to improve the predictability of duration.
Fourth, if one uses available court data to measure the time taken to resolve disputes from the perspective of litigants, one would have to make significant upward amendments. If one were to make adjustments and also consider related litigation as well as waiting time, the average time taken for disposition increases by more than 40% (see Section 6).
Fifth, our research shows that the spread of the time taken to resolve a dispute is large in the type of case which we examined. An individual litigant certainly can exert influence over the time taken for disposition. The time taken for disposition, however, depends on numerous other factors beyond an individual litigant’s control, such as (i) the conduct of the opposing party, (ii) the conduct of the judges involved, (iii) the judicial organization and (iv) the legal framework. As a result, litigation can be uncertain. It is, therefore, very difficult for a prospective litigant to predict whether litigation will take one, two, four or eight years. To narrow down uncertainty, litigants can confer and agree on the resolution of their dispute. If that is impossible, they can seek to predict the litigation strategy of the other party. In some cases, this will be relatively easy based on earlier conduct, reputation and resources of the parties, but in many cases this will be impossible. In all cases, the trial dynamics may lead parties to change their strategies.
Our data concern the legal system in the Netherlands. The Dutch civil procedural system is rooted in the civil law tradition and has many similarities with the systems that are in place in e.g. France and Germany.36 The rules of procedure are primarily codified in the Dutch Code of Civil Procedure.37 One must be careful before assuming that conclusions that pertain to one jurisdiction would be equally valid in other jurisdictions. Some of the specific problems described above may be related to peculiarities of the legal system or the court administration system in the Netherlands. Nevertheless, we do believe that the more general points that have been made in this contribution are equally valid elsewhere. Court statistics – if available at all – are likely to underestimate the problem of undue delay in important high value civil and commercial cases.
If parties and courts would like to address undue delay in high value cases, an important first step would be to identify the regular occurrence of accumulation of procedures that pertain to the same dispute. Our first attempt to analyse this problem for a limited number of procedures shows that the accumulation of different procedures pertaining to the very same dispute deserves more attention. There seems to be limited case management beyond the individual case. It would be important to understand how often the same parties litigate what is, in essence, from their perspective one single dispute and how long this takes.
We can conclude, however, that those that wish to rely on court data to predict the time needed to resolve disputes will significantly underestimate the problem of undue delay. Undue delay is most prevalent in complex high value cases. In a fair share of these cases litigation drags on for four, six, eight or ten years. As the spread of the time taken for disposition is quite large, it means that prospective litigants cannot always make an adequate prediction of the road ahead. Those that are risk averse probably have little choice but to avoid the justice system, for example by accepting an unfair and unequitable settlement offer. To combat undue delay, strong time management of the whole process beyond individual courts and procedures might be explored. Parties and judges would need to seek mutual agreements and/or procedural directions that avoid an accumulation of different procedures before different courts pertaining to the same dispute.38