The Yugoslavia Tribunal and the Selection of Defendants

After a historical overview of prosecutorial policies at Nuremberg and Tokyo tribunals, this article looks at the policy and practice of individual chief prosecutors of the International Criminal Tribunal for the former Yugoslavia regarding the selection of defendants.


Introduction
International courts typically deal with mass violence and their prosecutors must unavoidably be selective when selecting defendants.The selection of defendants lies at the heart of prosecutorial discretion.It amounts to the prosecutor's power to choose between two courses of action: to prosecute or not to prosecute.Ever since the post-war prosecutions of German and Japanese war crimes, the selection of defendants before international tribunals has been the subject of controversy.Commentators have criticised prosecutors for a variety of weaknesses: their lack of independence, opaqueness, bias, or support of misguided political goals.Such criticism must be taken seriously.It has been rightly stated that the extent to which "the prosecutor exercises his discretionary powers judiciously determines to a large degree the success or failure of international criminal tribunals". 1 The International Criminal Tribunal for the former Yugoslavia (ICTY) was considered partial by commentators in the former Yugoslavia.In the international arena it was felt that the prosecutor would frustrate the peace process when indicting high-level politicians.Sources more close to the ICTY complained about the lack of a clear prosecutorial strategy.
In this article I will analyse whether the criticism has any ground in reality.After a review of the selection processes at the Nuremberg and Tokyo tribunals, I will first examine the prosecutor's independence in establishing their policy.Secondly, I will ask whether the prosecutors developed and publicised an a priori policy for selecting defendants and what their guiding criteria were.Thirdly, my analysis focuses on whether the prosecutors articulated a clear policy in the public arena.Fourthly, I will examine whether the prosecutors had an overarching aim guiding their prosecutorial strategy.Lastly, I will look at the outcomes of the selection policies.This allows me to answer the question whether the prosecutorial practice matched the original mandate of the tribunal.

I. Selecting Defendants in Nuremberg and Tokyo
Formally, the prosecutorial policy of the post-Second World War tribunals in Nuremberg and Tokyo was the sole responsibility of the prosecutors.In practice, the policy was shaped by Allied governments, the founders of the Nuremberg and Tokyo tribunals, who had been parties in the war and victims of the very crimes the tribunals sought to prosecute.Such direct involvement had a significant impact on the selection process of the German and Japanese defendants.From the start, the Nuremberg trial was tainted by the Allies' desire for revenge, the need to oust the German leadership, the wish to educate the German people, and the aspiration to develop international law as an answer to acts of international aggression.2These larger aims influenced the debate on who would stand trial.At the 1943 Tehran Conference, Stalin proposed the execution of between 50 and 100 000 Germans, while Churchill envisioned the execution (without trial) of between 50 and 100 German war criminals. 3In 1944, U.S. Secretary of the Treasury Morgenthau Jr. suggested the deportation of millions of Germans as well as the execution of all Nazi party members. 4In the same year, General Eisenhower advised executing about 3 500 prominent Nazis.Stimson, the U.S. Secretary of War, developed plans for the prosecution of all members of the Gestapo and the senior leaders of the SS (Schutzstaffel). 5The Allied plans did not materialise.In the end, prosecutorial decisions were taken in a proper legal context, but not without heavy governmental involvement.
In October 1943, the Allies, except the Soviet Union, established the United Nations War Crimes Commission (UNWCC).The commission was tasked with the investigation of war crimes, gathering evidence, and drawing up a list of German war criminals to be prosecuted and sentenced. 6The poorly funded commission relied almost entirely on the assistance of the member states, i.e. the Allied governments. 7The commission's National Bureaus conducted the preparatory investigations before passing on relevant information to the commission, which then examined the files to establish prima facie evidence of a war crime.If this was found, the suspect was added to the list. 8The commission had a hard time carrying out its tasks.Many cases submitted to the commission turned out to be unsuitable for prosecution.Files were incomplete and some cases merely involved 'trivial' crimes.The commission members, frustrated with the lack of progress, drew up guidelines to get more prominent suspects on the list, for example, by identifying those who had ordered, rather than executed, war crimes. 9The Allied governments failed to adopt the guidelines, however.The British and Americans especially feared repercussions against their fellow countrymen who were still stationed in Germany's occupied zones.Politicians wanted to make sure the commission's work remained low profile. 10he commission navigated between these external pressures and their own ambitious goals.In December 1944, the commission published a list of 712 suspects that had been proposed by the commission's governmental representatives.Of these suspects, 49 were deemed major war criminals.Various criteria were used to determine which suspects to put on the list.For one, victims of war crimes had to be citizens of an allied country.11American policymakers sought to prioritise the prosecution of crimes relating to waging an aggressive war. 12According to U.S. prosecutor Telford Taylor, the British were primarily influenced in their selection by the consideration that the general public should know the suspects. 13aylor thought the selection process demonstrated it proceeded rather randomly. 14The team of prosecutors felt extremely uneasy about the way the list of suspects had been drawn up. 15ccording to some, the American chief prosecutor Jackson, responsible for the list's compilation, was so ill-informed about the military and political power structure in the Third Reich that he was unable to determine who could be criminally charged.Taylor concluded that the selection of suspects that would eventually stand trial in Nuremberg had been achieved in a hasty, chaotic, and neglectful fashion. 16litical considerations also entered the debate around the Tokyo tribunal.According to Röling, the Dutch judge at the tribunal, the prosecutors in Tokyo depended on their governments even more than they had in Nuremberg.U.S. General MacArthur, supreme commander of the Allied Forces, played a dominant role in the procedures.Bix writes that MacArthur had the power "to reduce, approve, or alter any punishments meted out." 17 As in Nuremberg, the Allies' main objective was to punish the Japanese for what in their view was the most serious crime: waging an aggressive war as a crime against peace rather than punishing the Japanese for actual war crimes and crimes against civilians. 18The Americans, hoping the Japanese Emperor would play a central role in the political reconstruction process, instructed the committee that the supreme leader should not be prosecuted, despite his responsibility for waging an aggressive war.This decision, made by General MacArthur and supported by American President Truman, went against the chief prosecutor's wishes, who felt there were ample grounds to prosecute the Emperor. 19MacArthur even forbade the prosecutors to interrogate or summon the Emperor as a witness during the proceedings. 20It soon became clear that, as the Cold War was heating up, the Allies' search for a stable Japan and a speedy restoration of its political and economic institutions influenced the proceedings. 21 in Nuremberg, the prosecutors in Tokyo tried hard to keep a balance between legal and non-legal considerations.Finally, national teams of prosecutors selected the accused from a list of 250 persons. 22According to Pritchard, the selection was intended to be no more than a "representative cross-section of those whom the Allied powers collectively regarded as responsible for Japanese policy before and during the Pacific War." 23 Ultimately twenty-eight accused were indicted in Tokyo. 24Many of those closely involved with the proceedings disagreed with the selection.According to some, the fact that the Emperor remained exempt meant that others like former Prime Minister Tojo and a few army officials were blamed for the role the Emperor had played during the war. 25Much criticism was levelled at the decision not to prosecute a number of prominent politicians and businessmen who initially featured on the list of suspects on the grounds that they were projected to play a potentially useful role in the country's reconstruction efforts.In a 1985 lecture during the first congress ever held in Japan on the legacy of the tribunal, Röling concluded: "Tokyo judgment has suffered in Japan precisely because many people were not convinced that some of the statesmen found guilty were actually among those responsible [for leading Japan into the Pacific war]." 26 In sum, both in Nuremberg and Tokyo circumstances and politics influenced the selection of defendants.Though the suspects' responsibility did play a role, several factors undermined a consistent selection process.The lessons of the Nuremberg and Tokyo tribunals shaped the strategies of the prosecutors in the 1990s but the situation was very different.What had first been a prosecutorial project of the war's winners became the personal responsibility of an individual prosecutor.To what extent did the United Nations, in taking over the allies' role in international prosecution, succeed in avoiding the mistakes of the past?

II. Selection Policy
There were several differences between the prosecutors' mission at the allied-sponsored tribunals and the newly established ad hoc tribunals of the 1990s.The prosecutor acting both for the ICTY and the Rwanda tribunal no longer received instructions from governments.Instead, the international community, represented by the UN Security Council, henceforth commissioned the prosecutor's tasks.Moreover, the Security Councils' mission to establish the ad hoc tribunals also emerged out of its responsibility to maintain peace and security. 27he tribunals of the 1990s were to put an end to the crimes and contribute to the restoration and maintenance of peace.Security Council members had several additional goals in mind, such as justice for the victims, general deterrence, truth-seeking, and contributing to the historical record. 28This raises questions regarding the extent to which the prosecutor took such additional ambitions into account when developing his or her prosecutorial policies.The ICTY's mandate was "to prosecute persons responsible for serious violations of international humanitarian law committed in the territory of the former Yugoslavia since 1991 […]." 29 The Statue defined 'serious violations' as genocide, crimes against humanity, and war crimes. 30ly once did the Security Council issue a guideline regarding the prosecutor's selection strategy.In 2003, the Council narrowed down the prosecutor's mandate after the ICTY president proposed a completion strategy to conclude the investigations by the end of 2004 and all trial activities by the end of 2008. 31Subsequently, the Security Council's resolutions requested both tribunals to focus their activities on the highest-ranking political and military leaders. 32Significantly, the formulation of the Security Council's instructions only dealt with the level of responsibility and the position of the suspects as a threshold for prosecution.No criteria were included to indicate a threshold with respect to the gravity of the crimes.

II.1 Richard Goldstone
The first prosecutor of the ICTY, the South African Richard Goldstone, navigated in tumultuous political waters.For one, the wars in Bosnia and Croatia were still ongoing when he began his work.In addition, regional authorities were hostile to the new tribunal.Moreover, both the EU and the UN, involved in peace negotiations, failed to make cooperation with the tribunal their priority.Not surprisingly, perhaps, Goldstone refrained from publicly revealing much about his prosecutorial strategies as he was unsure about the effectiveness of his decisions.Neither did he disclose any information about the underlying aims of his strategies -if he had any -nor did he express the extent to which he felt bound by the Security Council's aims.He did not publish a policy document.Instead, we may reconstruct his strategy from passing remarks in indictments, a lecture, and interviews.An internal policy document, collated in his office in October 1995 and partially published in 2008, revealed a variety of prosecutorial aims.The document referred to advancing jurisprudence and promoting the Tribunal's reputation and effectiveness in the regions where crimes had taken place. 33Yet these succinctly formulated aims neither included further explanation, nor did they find their way into the prosecutor's public statements at the time.It is therefore hard to determine to what extent these aims influenced the prosecutor's prosecutorial policy or his decisions regarding the selection of defendants.Yet we may deduce some guiding principles from his actions.
Political circumstances influenced Goldstone's initial strategic decisions to a great extent.His first priority was to set the wheels of international prosecution in motion to secure the much-needed support from the UN.Because of the region's insecurity, on-site investigations were next to impossible.Practical and financial support was limited and the Office of the Prosecutor struggled with a shortage of resources.To make matters worse, the UN threatened to stop funding altogether if no indictments were issued by November 1994.
In order to get the ball rolling, Goldstone developed a strategy that heavily depended on external sources.He first decided to indict defendants who were already held in custody by other prosecuting authorities, or, in an exceptional case in Bosnia, detained by international forces. 34Second, Goldstone investigated mid-level and lower perpetrators, including physical perpetrators because he expected that investigation into the crimes by military and political top leaders would be too complicated and time-consuming. 35He called his approach a 'pyramidal' strategy expecting it would finally lead him to the top echelons: "In the former Yugoslavia there was no smoking gun, we had to build cases with witnesses.Who could the witnesses tell us about?They could tell about the people in the camps, the camp commanders.They had no evidence to give us about the orders higher up.So we had to build up the cases from the bottom up." 36Third, Goldstone decided to concentrate on events and crimes that other organisations, such as the UN Commission of Experts, had already investigated. 37The strategic choice bore the risk that, instead of taking responsibility for mapping and analyzing the violence in the former Yugoslavia on his own accord, he had to rely on reports that were not intended to expose political and military leaders' culpability.Moreover, some reports at the time soon turned out to be wildly inaccurate.In the heat of the war, numbers of victims were grossly overstated.It is now estimated that the actual figure of people killed in Bosnia-Herzegovina amounts to 97 000 instead of the more than 200 000 calculated at the time. 38Likewise, in 1992 the number of rape victims was believed to be over 50 000.Later the estimate had to be reduced to 20 000, leading current reports to refrain from giving any indication of the number of victims. 39At the time, however, Goldstone had no alternative but to follow external sources.His strategy was successful.It soon led to the prosecution of a large number of defendants.
Despite the publicly professed lack of an overall strategy, Goldstone did present a few concisely yet ambiguously formulated principles of his prosecutorial policy for the selection of ICTY defendants. 40The ambiguity resulted, among others issues, from the need to navigate between the Security Council's 'institutional ideology of impartiality' in its role as peace negotiator on the one hand, and the NGOs' and UN rapporteurs' more factual-based reports that pointed at Serbs and Bosnian Serbs as the main aggressors. 41Goldstone's prosecutorial policy and strategy were based on three, rather clumsily formulated, principles.The first referred to the (political) independence of the Prosecutor: "Decisions with regard to indictments will be taken solely on a professional basis and without regard to political considerations or consequences." 42He failed to explain what he meant by professional basis in more detail, but Scharf argues that Goldstone meant the decision to prosecute should be based on sufficient evidence. 43In a later account, Goldstone further elaborated on what he meant by 'political considerations', explaining that a decision to prosecute should be made without taking the suspect's political or ethnic affiliation into consideration. 44The second principle set a priority in the selection of defendants: "Persons indicted will be those who appear to the Prosecutor to be most guilty and most culpable on the evidence available from time to time." 45 The 'most guilty' criterion seems to depend on the available evidence, although it is not entirely clear what the prosecutor meant by "the evidence available from time to time."Probably, Goldstone wanted to stress the fact that he could only prosecute if there was sufficient evidence and, given the circumstances in the region, this was clearly problematic.Goldstone submitted that the limited capacity of the tribunal to conduct trials forced the prosecutor to work selectively and only investigate the "most serious violations of international humanitarian law and those who may be ultimately responsible for them". 46Here, the prosecutor presented what appears to be a principled approach in terms of purely practical considerations.Goldstone became more specific in articulating the third principle where he considered the question of the suspect's responsibility: "With regard to the seriousness of the crimes, the most guilty are those who ordered them.At the same time, all efforts will be taken to ensure that those who executed such orders are also brought within the net of indictments." 47Goldstone immediately added an exception and kept open an opt-out, i.e. a focus on different levels of responsibility.Another policy decision was taken in response to early reports and criticism for failing to pay sufficient attention to rape in the ICTY's first trial against Duško Tadić.Goldstone decided that gender-related crimes should become a focus of his prosecutorial policy.The choice was "an important part of our mission to redefine and consolidate the place of these offences in humanitarian law." 48Here again, Goldstone received guidance from external forces.Given the circumstances, the ambiguity of Goldstone's policy is understandable.He realised he could only fulfil the tribunal's mandate with the help of its member states and local authorities.He had neither.In the public arena, however, his opaque policy formulation acquired the characteristics of an oracle.
Unbeknownst to the public, in the autumn of 1995, the office of the ICTY prosecutor compiled a list of criteria to rationalise the selection process. 49The criteria focused on the suspect personally as well as the gravity of the crime.Firstly, the criteria list the social sectors targeted for prosecution: politicians, the armed services and paramilitary groups as well as government officials at the local, provincial, and national levels.Secondly, the criteria refer to the defendant's nationality, his or her role and participation in the decision-and policymaking process, and the extent to which they controlled their subordinates, both formally and factually.Also, the suspect's personal or direct involvement in notorious atrocities was taken into account.Other criteria include more practical considerations such as the possibility of arresting the suspect and the availability of sufficient evidence.In defining the gravity of the crimes, the criteria point at the number of victims, the nature of the crime, the duration and geographical scope of the crime, and the nationality of the perpetrators and victims.Although the latter two criteria do not necessarily relate to the crime's gravity, Bergsmo feels that these criteria are relevant because of the 'representativity' between criminal victimisation and the scope of the prosecution: since prosecutorial activities must be representative of the criminal offences committed by the various parties in the conflicts. 50inally, the criteria deal with the suspect's potential legal defence and other anticipated legal obstacles like the potential theories regarding the "liability and legal framework of each potential suspect."Tactical, strategic, and policy considerations including the relevance for other investigations into higher-level suspects also are included. 51The prosecutorial office did not elaborate on these criteria in its documentation.Neither is it clear whether, or to what extent, the prosecutor applied them when selecting the suspects.A former prosecutor's assistant argued the list was no more than a 'catalogue' of relevant considerations rather than a selective, focused set of binding criteria.Neither did a ranking order exist. 52In fact, Goldstone never even publicly referred to the criteria in passing, but clarified his selection of defendants only in the most general terms.
Goldstone's tenure had mixed results.Goldstone's creative strategy resulted in an impressive record given the circumstances of the moment.He indicted a large portion of ICTY defendants over a relatively short period of two years.As we will see in discussing the composition of Goldstone's caseload, Goldstone fulfilled his stated ambitions and set the wheels of prosecution in high gear.He was responsible for making the new tribunal a visible asset among other international institutions.His strategy may have been ambiguously formulated, but he neither raised expectations nor made promises he could not fulfil.

II.2 Louise Arbour
Taking over from Goldstone in 1996, the Canadian Louise Arbour believed circumstances still did not allow her the "luxury of setting the prosecutor's agenda". 53Nevertheless, the situation in and outside the Yugoslavia tribunal had improved substantially.The wars in Bosnia and Croatia had ended.The tribunal was in full operation and had become an accepted partner in the international community.Members of the UN actively supported and provided the tribunal with assistance and resources.In the Office of the Prosecutor, experience and expertise had grown substantially.Lack of regional cooperation continued to be a major a challenge, however.Publicly, Arbour was more open than Goldstone about how she came to formulate her policy.Yet, she did not publish any policy document either.As she explained, "[...] the day to day life of the Prosecutor often consists almost exclusively of crisis management." 54This became apparent at the outbreak of the Kosovo crisis when "[…] events overrode our investigative plan […]." 55 Arbour organised extensive office debates in The Hague to articulate her policy aims.The debates concentrated on various options.Arbour expressed reluctance to justify international criminal law along instrumental parameters: "We do not undertake investigations to do historical surveys, to provide the public with general information, or to extend the bounds of international law, for its own sake." 56But she acknowledged the advantages of prosecution: "The recording by international investigators of irrefutable evidence of crimes prevents history from being falsified and the past from being distorted." 57According to Arbour, the prosecution of military and political leaders would also grant the victims a judicial status and prevent them from taking revenge. 58When the decision was made to start investigations in Kosovo for example, Arbour wondered if a potential prosecution could have a deterrent effect: "[…] could we play a part in deterring the worst?Should we not at least try to put deterrence theories to test?" 59 But the prosecutor felt: "We had a right to enter Kosovo […]." 60 And ultimately, Arbour was primarily concerned that: "Our investigations are directed exclusively to establishing personal responsibility for those individuals who are most culpable for the atrocities committed in former Yugoslavia." 61Such considerations of the different aims of prosecutorial policy showed that circumstances had improved significantly over the years.Arbour was in a position to develop a more articulate strategy thanks to the improved circumstances outside the office and the experience and expertise built up within the office.Incidents were now selected based on the prosecutor's own investigation instead of other organisations' reports.Suspects were no longer selected only on their availability.Arbour had established a system of sealed indictments and international assistance in the arrests of suspects had improved.A striking difference with her predecessor was Arbour's 'offencedriven' approach, which prioritised incidents based on the crimes' gravity in order to target the highest possible defendants, such as "Srebrenica, where there was massive loss of human life, and the promise of climbing up the chain of command to visit the responsibility of the highest echelons was greatest." 62Her strategy thus still resembled Goldstone's pyramidal approach; the role of the commanders could be investigated through the cases against direct perpetrators and those who had been physically involved in the perpetration of crimes.The pitfalls of such a strategy were discussed both within the office and publicly, as the debate on investigations into sexual crimes illustrates well: "Punishment would, in the case of sexual violence, be particularly difficult to visit on the commanders under the doctrine of command responsibility.It would require proving that the commanders either participated in the offenses, or knew that the offenses were being committed but failed to punish those responsible." 63Overall, under Arbour the contours of a balanced strategy became apparent.For one, the gravity of the crimes gained more prominence.
It was under Arbour that the so-called OTP Charging and indictment guidelines were drafted, setting out the methods to be applied for investigating and preparing an indictment.The internal document was not meant for publication, however.The Guidelines were largely based on the criteria set by Goldstone.The Guidelines included directions on how to conduct and prepare for investigations, formulated criteria for selecting 'targets,' and offered a framework for legal and procedural issues.The decision to pursue investigations was based on the suspect's position and role and the crime's gravity. 64The criteria refer to the suspect's formal position within the army, paramilitary units, police, political and government structure, the level of personal involvement in the crimes and participation in policy-making resulting in the crimes as well as the defendant's nationality and ethnicity.Other criteria consider the suspect's possible willingness to testify against their superiors, whether they had been prosecuted elsewhere, and whether they could be possibly arrested.The Guidelines' criteria implied a pyramidal strategy as they did not only target the highest officials but also mid-level authority and those "in leadership positions, who are accountable in their own right, and whose prosecution provides a foundation for investigation of their superiors." 65ven suspects without a particular position within the military, political or police structures could be indicted in case the defendant was "so notorious and has committed such heinous acts that it is appropriate to charge the person, irrespective of what authority he exercised or position held." 66Criteria concerning the gravity of crimes refer to the number of victims, the extent of destruction, the systematic nature, the duration and extent to which the crime was repeated, and the location and notoriety of the crime.Additional circumstances such as the relationship to other cases or the presence of witnesses and evidence also played a role.Finally, the policy document includes more general considerations such as the crime's symbolic and political significance and the victims' ethnic background.Single-incident targets were generally not indicted even if the suspect could provide potentially useful information in other cases. 67The prosecutor believed such a route was potentially risky and damaging for the Tribunal's reputation.Yet, she kept her options open to investigate and prosecute 'lowlevel perpetrators,' who were willing to appear as witnesses in cases against high-profile perpetrators.In such cases the prosecutor insisted on a guilty plea to sustain the credibility of the suspect as a witness.The procedure did not, however, imply that the witness would automatically be prosecuted by the Tribunal at a later stage. 68In short, the Guidelines reflect Arbour's public statements on selection issues and her preference for legal grounds as the main consideration for prosecution.
Building on the work of her predecessor, Arbour was able to refine the selection strategy in the prosecutor's office even though she had stated she did not have the luxury to set her own agenda.That was probably too modest.As will be shown in section four, she was able to climb up the chain of command in order to target those ultimately responsible for what she considered the most serious crimes.Like Goldstone, Arbour's aim did not refer to UN ambitions but was limited to establishing legal accountability as leading principle of her selection strategy.

II.3 Carla Del Ponte
When the Swiss Carla Del Ponte succeeded Arbour in 1999, the violent conflicts in Bosnia Herzegovina, Croatia and Serbia had ceased.Only a limited conflict erupted in Macedonia in 2001 that came under her jurisdiction during her tenure.To a certain extent some governments in the region had turned into the prosecutor's allies, while others continued to obstruct cooperation with the tribunal by refusing to hand over documents or arrest indictees.In the prosecutor's office, expertise and knowledge of crimes and culprits were of the highest level and quality.Del Ponte considered the circumstances at the time most favourable not only to reflect on the prosecutorial policy's aims and criteria, but also to implement the tribunal's broader mandate. 69Nonetheless, like her predecessors, she chose not to publish a policy document.multi-ethnic society of the former Yugoslavia: "[…] by imprisoning those hard-line extremists whose continuing political and military involvement serves to hinder the creation of a lasting peace, it consequently improves the conditions for the rebuilding of a multiethnic society in the region." 71We also find such an instrumental approach in another element of her policy.Del Ponte emphasised that a lasting peace could only be reached if all parties to the conflict had been investigated: "My remaining investigations cover the region and concern all main parties to the conflict.By completing these investigations, ICTY will have proven that it worked impartially towards achieving justice, peace and reconciliation in the former Yugoslavia." 72The policy of even-handedness was also introduced after the end of the conflict in Kosovo when Del Ponte announced that, to achieve a lasting peace, she also wanted to prosecute crimes that were committed after the war.She referred in particular to the crimes Kosovar Albanians had committed against the Serbian minority. 73For this purpose the Tribunal's Statutes had to be amended, she argued: "We must ensure that the Tribunal's unique chance to bring justice to the populations of the former Yugoslavia does not pass into history as having been flawed and biased in favour of one ethnic group against another." 74Such statements certainly gave the impression that legal criteria were deemed less important.
In spite of improved circumstances and the availability of all necessary information in her office, Del Ponte maintained the multi-level approach of her predecessors, targeting not only the most senior officials but also mid-level defendants.She considered them a vital link to the highest levels and believed they played a crucial role in organising and committing crimes on the ground: "Such individuals often play a great role in setting the example and encourage, by their acts, speech and behaviour, the commission of other gruesome crimes." 75Del Ponte thought prosecution of these groups important because: "For the local people, the victims and the survivors, it was these people who brought their world to an end, not the remote governmental architects of the overall policy of genocide." 76She thought that if they were not brought to justice, "the ordinary population will not come to terms with the past, and the process of reconciliation and building a stable peace will suffer accordingly." 77After the issuance of the completion strategy, Del Ponte indicated she would concentrate on suspects at the highest institutional levels of the police, politics, and the military because she thought them responsible for allowing the atrocities to continue. 78At the same time, Del Ponte stated she would continue to investigate other categories of perpetrators, including direct or physical perpetrators of very serious offences, even if they had not held important positions. 79As we will see, Del Ponte refrained from prosecuting such defendants.Del Ponte's statements are a manifestation of the international prosecutor's dilemma of having multiple constituencies.Seeking to provide justice to victims at all levels and aiming to fulfil UN aims at the same time has not contributed to the transparency of her policy.Moreover, integrating aims of peace and reconciliation into her selection policy, Del Ponte took the risk of introducing political factors in her selection policy and thus marginalising other selection criteria, such as gravity of the crimes and the level of responsibility of the perpetrator.

III. Selection Practice
To what extent did the practice of the tribunal respond to the prosecutorial strategies?In what ways did prosecutorial practices match the lofty ideals the UN had articulated?Based on an analysis of the ICTY caseload, we get an insight into the difference between the stated prosecutorial policies and the actual practice.This analysis focuses on the indictments dealing with the two pillars of prosecutorial selection: criteria concerning the responsibility of the defendants and the gravity of the crimes. 80Between 1994 and 2004, ICTY prosecutors indicted 161 defendants. 81In short, the ICTY's legacy is a mixed bag of low, mid-level and high-level functionaries.As such, the tribunal's case load does reflect the prosecutors' stated strategies, but not the UN's agenda.
When we focus on the issue of the responsibility of the defendants, the ICTY caseload reveals a clear increase in leadership cases over the years.Nevertheless, over 80 percent of all ICTY indictees occupied mid-level or lower level positions.Only 20 percent of the defendants were senior officials active at state or federal level.This is partly explained by the fact that Goldstone's share of the ICTY caseload was relatively high.He issued indictments against seventy suspects.The early indictments resulted from his strategy that what counted most of all was to set the wheels of prosecution in motion.Indeed, he based his first indictment merely on whether sufficient evidence against the suspect was available.As Goldstone admitted, the suspect was a "comparatively low-level member of the Bosnian Serb Forces," and "hardly an appropriate defendant for the first indictment issued by the first ever international war crimes tribunal." 82Goldstone's first trial in The Hague, against another Bosnian Serb, resulted from his decision to request the transfer of perpetrators already prosecuted or investigated elsewhere.The defendant, detained by German judicial authorities, had not been indicted by the ICTY at the time of Goldstone's 1994 request.Three other defendants were indicted in the same manner.In May 1995, Goldstone demanded the judicial authorities of Bosnia Herzegovina defer their cases and the suspects were indicted three months later. 83Likewise, Goldstone requested the deferral of Bosnian investigations into the Bosnian Croats' crimes in the Lasva Valley in Central Bosnia and went on to indict eight suspects. 84Another substantial number of indictments grew out of Goldstone's decision to take up cases other organisations had investigated and reported upon.The indictments dealt with crimes committed in the Bosnian district of Prijedor and the Bosnian cities of Foča and Brčko, regions the Commission of Experts and other organisations had examined extensively.The majority of Goldstone's impressive number of 70 indictments during his tenure resulted from his pyramidal strategy.Most defendants were low profile -a reason why, when she took over, Arbour rightly decided to withdraw indictments against 20 of them.Still, of the remaining 50 cases, over 80 percent involved low-level perpetrators.Some defendants had carried out tasks in one of Bosnia Herzegovina's many detention camps, others had been active at the village level, or held no formal position at all.The rest (20 percent) of the Goldstone cases dealt with mid-level suspects, who had operated in provincial state or military institutions.Only a small number of his cases dealt with high-level suspects in national positions such as the Bosnian Serb politician Radovan Karadžić and military leader Ratko Mladić.Goldstone did not indict anyone from the Serbian political or military leadership.This outcome is in line with his belief that it would take time to investigate their involvement in crimes on Bosnian and Croatian territory.
Arbour indicted 31 ICTY defendants who had been active at different levels.Many indictments were based on Goldstone's investigations.It explains why former functionaries at detention camps still formed more than a quarter of 'her' defendants.Mid-level defendants made up one third of Arbour's caseload.New investigations into the crimes in Kosovo did lead to the first indictments against top Serbian politicians and military, among them Serb president Slobodan Milošević.Interestingly, Arbour did not need to pursue a pyramidal strategy with respect to the Kosovo crimes.No lower or mid-level suspects were prosecuted in the context of these investigations.Del Ponte issued indictments against a total of 53 defendants. 85She indicted more senior officials at the highest governmental and military levels than her predecessors.One third of her caseload dealt with defendants in national or federal positions.Nevertheless, almost half of the defendants Del Ponte indicted were mid-level provincial officials and the remaining defendants came from a lower rank, operating at the community or village level.Even though the overall outcome at the ICTY may be logical given the circumstances and the decisions taken therein, it is not clear why so many low-level perpetrators were needed to make the pyramidal strategy effective.Goldstone's defendants, for example, had operated in only a limited number of crime scenes.Arbour's and Del Ponte's lower-level cases may be explained in part by the fact that investigations were already going on when they took over, but they also resulted from their own policies.The broadly based pyramidal caseload is thus the clear result of a combined prosecutors' strategy.
Generally speaking, the higher the defendants' position, the more serious the crimes with which they were charged are.The gravity of the crimes in the indictments has been analysed on the basis of the numbers of victims involved.Almost all victims were civilians.Half of all the crimes in the indictments concern a large number of victims: from hundreds to hundreds of thousands.Again, there is a difference between the early indictments issued by Goldstone with more crimes with relatively low numbers of victims, and those issued at a later stage with more crimes charged with high numbers of victims.Del Ponte, however, also charged more crimes with few victims than her predecessor.This was a consequence of her policy of indicting all parties that had taken part in the conflicts.While Goldstone indicted members of the three Bosnian groups, Croats, Serbs and Muslims and Arbour prosecuted Bosnian Serbs, Bosnian Croats, and Serbs, these defendants were not selected for their ethnicity.Del Ponte indicted suspects from eight national or ethnic groups in four different countries. 86To be sure, the geographical spread of the region's conflicts in Kosovo and Macedonia during her tenure partly explains such an ethnic diversity, but her policy to target all sides of the conflicts also accounts for the greater variety.All but one indictment that Del Ponte levelled against Bosnian Muslims, Macedonians, and Kosovar Albanians concerned crimes involving fewer victims in comparison to the indictments regarding other groups.Looking at the relative gravity of the crimes in the indictments, her strategy of even-handedness implied that a number of defendants was selected based on membership of a particular national or ethnic group rather than the gravity of their crimes.As most of these defendants were senior officials, part of the Del Ponte caseload is an exception to the rule that the seniority of defendants is equal to the gravity of their crimes.
The overall composition of the ICTY caseload reflects the inability to indict the most senior responsible officials during the early years.Comparing the gravity of the crimes in the indictments, actual differences between the national and ethnic groups and their share in the crimes become visible.Bosnian Serbs and Serbs were indicted for the ICTY Statute's three crime categories (genocide, crimes against humanity, and war crimes).They were also charged for crimes involving the largest number of victims-running in the hundreds of thousands.The Bosnian Croats (16 percent of all defendants), Serbian Croats (4 percent), and Croats (5 percent), and Kosovar Albanians (over four percent) were accused of war crimes and crimes against humanity. 87The Bosnian Muslim defendants (over 6 percent) and Macedonians (1.5 percent) stood trial for the single category of war crimes.The policy of even-handedness that was employed during the last phase of issuing indictments shows that gravity was not always the guiding principle in selecting defendants.

Conclusion
Prosecutorial strategies shifted from a governmental responsibility during the post Second World War tribunals to an individual prosecutor's task at the current ad hoc tribunals.This one man or woman's task included elements of governance as the UN required the prosecutor to consider the local effects of their decision.The tribunal's history shows that, for a large part, the prosecutors developed their selection policy independently.The prosecutors decided to what extent they integrated UN goals into their policy.Only once did the UN Security Council give additional instructions about the prosecutor's mandate, in the context of the completion strategy.There is no evidence the instructions influenced the substance of the prosecutors' policy.The relative independence of the prosecutors did not mean that the prosecutors were always able to put their policies into practice.Circumstances in and outside the prosecutor's offices, in particular the assistance from powerful nations, shaped the prosecutor's ability to put strategy into practice.
Prosecutorial policies at the ICTY lacked transparency.Grounds for selecting defendants were often opaque.Understandably, during the early years, the prosecutor shied away from publishing information on how he selected particular targets because he was unsure whether he could get enough assistance and resources.Over the years, especially after the completion strategy was issued, the prosecutor became more open about considerations and choices, although no official policy document was ever published.Complete openness may not always be feasible in the highly politicised context of international tribunals.To avoid perceived bias or politicisation of the selection process, however, prosecutors should be as clear as possible about the grounds for prosecution.
In most cases, grounds for selection focused on the level of the suspects' responsibility and the gravity of the crimes.Generally, the two criteria are related: the more senior the defendant, the more serious the alleged crimes.In principle, the prosecutors in their formulated policies followed the UN's focus on senior officials and top-level leaders even though strategies to achieve this aim differed.The prosecutors formulated an exception to the rule, either for strategic or principled reasons.The prosecutors opted for the possibility to prosecute 'notorious' perpetrators, including those who had not held any formal position.There may be legitimate grounds for such decisions, but they should be avoided because they often resulted from secondary considerations.
Most leaders were indicted for very serious crimes involving large numbers of victims.Some senior defendants, however, were not indicted for the most serious crimes.This outcome resulted from a policy that focused on factors other than the responsibility of the suspect or the gravity of the crime.The policy of even-handedness to prosecute all parties for its potential-yet never established-effect on reconciliation serves as the most telling example.The policy of even-handedness implied that in some cases indictments were issued on the basis of membership of certain groups rather than on the gravity of crimes committed.In 1999, former UN Secretary General Kofi Annan correctly proposed to abolish the "institutional ideology of impartiality". 88The lawyers in the international prosecutor offices should follow suit.If not, they will go beyond their mandate and become diplomats, threatening the legitimacy of present and future courts. 88A/54/35 (1999), § 505.