Observers say lessons urgently need to be learnt from prosecution failings.
Experts say a new investigation strategy due to be adopted by the International Criminal Court (ICC) needs to address longstanding issues about evidence-gathering that have left multiple trials on an uncertain footing.
Full details of the guidelines, sent to the ICC's 122 member states last week, have yet to be released, but they focus on ways to ensure that the Office of the Prosecutor (OTP) can present a watertight case at trial.
Key elements include plans for prosecutors to ensure that cases are ready at an earlier stage of proceedings, and for court investigators to corroborate evidence that is collected by third parties. The need to safeguard the security of both investigators and witnesses also remains a serious challenge.
In its first 11 years of operation, the ICC has often struggled to gather sufficiently convincing evidence against suspects.
The conviction of Congolese warlord Thomas Lubanga Dyila in 2012 - the prosecution's only success to date - was a milestone for the court. But judges on the case were scathing about the way the OTP handled the investigation, particularly its reliance on intermediaries and its failure to properly probe evidence that later turned out to be false.
The Lubanga judgement also highlighted a tendency for investigators to rely too much on third-party information such as reports from human rights groups and academics.
Similar flaws in the ICC's investigative procedures have been uncovered elsewhere.
When Callixte Mbarushimana, a senior figure in the Democratic Forces for the Liberation of Rwanda or FDLR, came before the ICC in 2011 for alleged human rights abuses in the eastern Democratic Republic of Congo (DRC), judges ruled there was not enough evidence to send the case to trial.
More recently, judges at the ICC declined to confirm charges against the former president of Ivory Coast, Laurent Gbagbo, after the OTP failed to present compelling evidence linking him to crimes on the ground.
Gbagbo remains in custody while the prosecution appeals against this decision.
In the court's investigations in Kenya, prosecutors were forced to drop their case against former civil service chief Francis Muthaura after it emerged that one key witness had lied to investigators.
Since Prosecutor Fatou Bensouda made that decision in March, it has emerged that several witnesses have withdrawn from the cases against Kenyan president Uhuru Kenyatta and deputy president William Ruto since charges were confirmed in January 2012.
Bensouda has repeatedly highlighted "unprecedented levels" of interference in her cases in Kenya. On October 2, judges unsealed an arrest warrant against a former journalist who is accused of bribing witnesses.
The succession of setbacks has prompted widespread concern over how the OTP conducts its investigations and why it has found it so hard to gather enough evidence to secure convictions.
RUSH TO CONFIRM CHARGES
One former investigator who spoke to IWPR on condition of anonymity said his colleagues were under a lot of pressure to get cases under way, and to come up with a piece of evidence for the OTP to build its case. This often meant they overlooked other information that might have been useful at a later date.
Montserrat Carboni, permanent representative to the ICC for the International Federation for Human Rights, told IWPR that the OTP had a tendency to gather just enough evidence to secure an arrest warrant. This would then be built on to confirm the charges, and then worked on again in order to clear the next hurdle in the case.
"If investigators are gathering evidence in stages, then there is a strong likelihood that they will overlook vital information," Carboni said.
The ICC acknowledges that this kind of approach is a concern. Michel De Smedt, the head of investigations at the court, said his office had addressed the problem in its new strategy.
De Smedt told IWPR that investigators were moving away from excessively focused investigations where they collected a limited amount of evidence, selecting incidents and perpetrators early on in the process in order to build a case around them.
Instead, he said, his office was moving towards "more open-ended, in-depth investigations, where we gather a broader range of evidence and don't seek a confirmation of charges until we are ready".
This is the reason why the court's investigation into crimes in Mali, which that country's government referred to the court in July 2012, has not yet generated requests for arrest warrants or led to evidence being placed before ICC judges.
The ICC launched a formal investigation into Mali at the start of this year at a time when large parts of the north of the country were still in the hands of Islamic extremists.
French troops regained control of the north's main town, Timbuktu, by the end of January, but other areas remained unstable. Despite this, investigators arrived in the country in early June. According to De Smedt, this was swift given the circumstances.
Although investigators moved quickly, De Smedt said great care had been taken to make sure that the evidence would stand up to scrutiny by ICC judges before arrest warrants were requested, and before any hearing to have charges confirmed.
The Mali case is expected to be ready by the end of this year or in early 2014.
While the OTP acknowledges that it has sometimes rushed to bring cases before the judges, there have been times when there has been good reason to press ahead before amassing all the necessary evidence.
When the indicted former vice-president of Congo, Jean-Pierre Bemba, visited Belgium in 2008, the authorities there were ready to cooperate with the court and the ICC knew that it had to move fast to secure his arrest.
At the time, the ICC lacked enough evidence to bring the case to trial, but the then prosecutor, Luis Moreno-Ocampo, made a decision to go with what they had and try to gather more at a later date.
Moreno-Ocampo also took the decision to move early on in his investigations in Kenya, arguing that in doing so he could isolate alleged perpetrators and prevent further violence at the next election.
The fact that the court is in its early days is also seen as a motivation for making arrests and getting charges confirmed quickly.
Alex Whiting, a law professor at Harvard Law School who recently left the ICC's investigation team, said such pressure was understandable within an institution that was set up in 2002 and that has been trying to establish itself as a force in the world.
"At all of the tribunals - the ad hoc tribunals and the ICC - it is normal that there is pressure in the early years to bring cases, to show that the court can be effective," Whiting said.
De Smedt recognises that there may be legitimate reasons for proceeding with a case before all the evidence is available, but he insists that the overarching consideration must be securing a conviction.
"The discussion has sometimes been framed as 'do we want to have an impact on the ground or an impact in court?'" De Smedt said. "The thing is, by achieving the second, we can often realise the first."
De Smedt did not rule out the option of proceeding with a case before all the evidence has been accumulated, but he added that his office would only now do so if it was going to be possible to gather more at a later date.
"If it is determined that we should move early, before we have all the evidence that we need, then the question we must be able to answer is 'do we have the prospect of gathering further evidence in the near future to allow us to be trial-ready'? If we do not, then we shouldn't proceed, even if we lose the possibility of an arrest," he said.
LENGTHY INVESTIGATIONS PROCESS
Another serious problem facing the OTP is the lengthy period that it takes to build a case and the impact that the passage of time has on the evidence presented to judges during trial.
In eastern DRC, the ICC opened its investigation in 2004, Lubanga was arrested in early 2006, and his trial started in January 2009.
Lord Justice Adrian Fulford, a member of the Court of Appeal in the UK who presided over the Lubanga trial, says cases have often suffered because witnesses testify years after they are first asked to provide evidence to the court's investigators.
"It makes the job of the prosecution and of the defence more difficult because the witnesses are giving their statements at one point in time and their accounts are not tested until a long time afterwards," he said. "Sometimes years go by before there is a proper process of evaluating whether or not what they say is reliable and credible."
Lord Justice Fulford described the conventional model of bringing witnesses to court in order to testify in front of judges as a sometimes "time-consuming and laborious means" of establishing the truth. He instead advocates a system where field testimony would be filmed in the presence of a judge or a trained, qualified legal officer. The prosecution and either the defence or a lawyer from the Office of Public Counsel for the Defence would then test the evidence, with the possibility of putting further questions to the witness on discrete issues later, for instance following full disclosure of their evidence and the complete instructions of the accused.
"It would enable cases to be built in an informed way, or a more informed way than they are now," Lord Justice Fulford said.
This approach would help bring greater efficiency and clarity to the case by ensuring that evidence remained essentially the same at all stages of the process.
"The OTP has tended to prepare and present one case at the pre-trial stage and then later put forward a different case at trial, particularly in terms of the witnesses that are relied on," Lord Justice Fulford said. "This way, if the prosecution has collected the evidence and it has been tested in a comprehensive way before the confirmation of charges, the OTP can then make an informed decision as to whether it is prima facie reliable and credible and fit to be presented at trial."
Lord Justice Fulford said that if evidence was collected in this way, the cases would in essence be ready for trial at the confirmation stage.
"The prosecution would be presenting witnesses who it was considered had been tried and tested, and worthy of being called," he said.
He argues that such a system would also be fairer to the defence.
"You reduce the risk of asking the judges to consider evidence that is, on examination, unreliable," he said.
Testing witnesses while their memories are still fresh and the events in question are reasonably contemporaneous is far better than asking them to be accurate about incidents that happened a long time ago. Supplementary questions can if necessary be put at a later stage, once issues have crystallised.
Lord Justice Fulford noted that collecting all or most of the testimony in the field, rather than doing so years later in the courtroom, would reduce the risk of witnesses being identified and intimidated. Once evidence has been gathered, tested and recorded, there is no longer any point in attempting to influence a witness.
STANDARD OF EVIDENCE TOO HIGH?
The burden on the prosecution to meet high standards of evidentiary proof is rightly an important part of safeguarding the rights of defendants brought before the ICC. But some legal experts say that compared with the practise at other international courts, judges at the ICC sometimes set the bar too high.
"In some cases, the ICC judges may have had unrealistic expectations about the kinds of evidence that will be available in these sorts of cases," Whiting said. "In the early years of the ad hoc tribunals, the judges were often more forgiving of the prosecution's efforts, and often gave the prosecution more time to develop their cases, even after [suspects] were charged and even after trial began, while in later years, the judges at the ad hoc tribunals became more demanding.
"But at the ICC, from the very start the judges strictly applied a more demanding set of procedures," Whiting said.
Judges dismissed the case against Mbarushimana because although the prosecution showed that crimes took place in the DRC, it did not clearly link the suspect to those crimes.
A similar stumbling-block has emerged in the Gbagbo case. Judges have declined to confirm the charges until the prosecution proves a link between the accused and crimes committed on the ground.
One of the grounds for the prosecution's appeal against the Gbagbo decision is the argument that there should be a distinction between proving a widespread and systematic pattern of abuse and proving that a specific incident took place.
The prosecution argues that judges looked at each piece of evidence in isolation rather than assessing the evidence cumulatively, which made it difficult to establish a link between actions taken by Gbagbo and crimes that were taking place in Ivory Coast.
BOOTS ON THE GROUND
Failures to meet the evidentiary threshold set by the judges has prompted criticism that ICC investigators place too much emphasis on third-party sources like NGO and media reports, rather than on collecting evidence.
War crimes expert Cherif Bassiouni, who has chaired five separate United Nations investigations into post-conflict situations, thinks this is a huge problem and wonders whether in some cases the ICC might be being too cautious.
"How can you build an effective case if you are sitting in The Hague and the only thing you have is a few witnesses with insufficient corroborating evidence because you have not been in the field?" he asked.
Bassouni's most recent investigation was into crimes allegedly committed during the 2011 uprising in Libya. He said that when he began the investigation, he contacted Prosecutor Moreno-Ocampo and invited him to send in some investigators, but the ICC's security advisors said that the situation was not safe enough.
The problem with relying on a third party's report is that proof of criminal guilt is based on the work of academics or NGO whose research did not have this purpose.
"The real skill in being an investigator is not just to collect information, but to make sure that this information can be used as evidence in a courtroom," said John Ralston, former head of investigations at the International Criminal Tribunal for the former Yugoslavia. "If the information has been collected by others - NGO reports, for example - then it is important for investigators to go and gather the evidence themselves, to verify its reliability and make sure it meets the evidentiary threshold of the court."
Ken Wafula, a human rights activist from Eldoret in the Rift Valley region of Kenya, complained that when the ICC opened its investigation into the country in 2010, investigators spent too long in Nairobi and not enough time in other areas where the violence took place.
"I think the ICC really missed an opportunity to engage directly with all the interlocutors in the [Eldoret] region - with the witnesses and the victims - which could have made [their] case very strong," Wafula said.
Wafula argued that if investigators had come to Eldoret earlier and engaged more with victims, this could have helped them build a stronger case. Wafula said that he did not meet investigators until early 2011.
De Smedt argues that there is a legitimate purpose in using secondary sources, particularly when prosecutors need to demonstrate the background to the crimes charged.
"Next to forensic evidence, we will always have to also rely on witness testimony or first-hand sources to prove that crimes took place in the first place," said De Smedt. "But in order to demonstrate a pattern behind the crimes, then we should be allowed to rely on other evidence, such as third-party reports if they are based on a proper methodology and independent from each other."
The OTP is often forced to rely on evidence gathered by a third party because of the need to protect witnesses, as well as providing security for its own staff,.
For De Smedt, witness safety is a far bigger worry than staff security, and is the key reason that the ICC has chosen to tread softly in some areas.
"We can manage staff security, and there have been very few serious threats against [ICC] staff," said De Smedt. "Witness security, on the other hand, is far harder to deal with. We have a duty to protect witnesses under the [ICC's treaty] Rome Statute. Investigators can always leave, but witnesses and their families are caught up in the country. We shouldn't risk the lives of people to build a case. It is much better to lose a case than the life of a witness."
Witness protection is also a matter of credibility for the court.
"If witnesses are put in severe danger in one case, then we may face difficulties in getting witnesses to testify for another one," De Smedt said.
Bassiouni accepts such concerns, but underscores the importance of getting investigators on the ground as early as possible, since key pieces of evidence can disappear over time.
"Injuries may heal or they may take on a different form," he said. "If people have been put in a mass grave, then the bodies might decompose and the trace of the bullet may not be so evident."
In response to such concerns, the ICC is looking at establishing a permanent field presence in certain countries. But it does not have the luxury of being welcomed everywhere it needs to work.
"Of course this depends on two crucial things - security and state cooperation - so wouldn't be possible everywhere," De Smedt said. "In Mali, I am mindful that we are dealing with terrorist organisations which could be a real threat to investigators. In Sudan, it seems unlikely that the government will ever let us have an office in Darfur."
NEED FOR INTERNATIONAL ASSISTANCE
Experts are generally in agreement that an effective investigation strategy does not depend on the court alone; the international community also has a huge role to play. Evidence from other international tribunals would seem to support this.
"Remember that the ICTY succeeded in part because the United States and the European Union conditioned aid and acceptance into the EU on cooperation with the court, which allowed the ICTY to have access to witnesses, documents, and accused persons," Whiting said. "Unfortunately, the ICC will have to be more realistic about whether countries - situation countries or other influential countries - will really help the court when there is a case. Sometimes they will, but sometimes they won't."
In countries where the ICC launches investigations, there are often few political incentives to ensure that required level of cooperation is forthcoming. This is an area where states committed to international justice can provide meaningful support.
"Governments need to provide real and practical assistance that isn't loaded in favour of or against particular individuals, groups or factions," Lord Justice Fulford said. "Witnesses frequently require relocation and protection, and generally substantive help is necessary for the process of gathering evidence in complex and sometimes dangerous situations.
"This inevitably involves money, manpower and diplomacy. These investigations are so vast and complicated that it is very difficult for the court to do it all by itself."
As well as cooperation and diplomatic backing, the ICC also needs resources. The OTP says a larger budget from member states would help it collect evidence more quickly and develop the court's forensic capacity. De Smedt says that there are plans to double the number of in-house forensic experts from four to eight.
Member states will meet in The Hague in November to discuss whether to increase the court's budget. The proposal currently on the table is to increase the budget from 115 million to 126 million euro, with an additional 7.5 million going to the OTP. This would put the OTP's budget at just under 36 million euro.
The international community has often been accused of failing to back the ICC fully, both with funding and on broader issues. De Smedt accepts that the actions of member states are often governed by national considerations rather than their obligations under the Rome Statute. But he believes their full support is essential if the court is to have a meaningful and lasting impact.
"If the international community is not willing to stand up for justice, then justice will not happen," he said.
Blake Evans-Pritchard is an IWPR contributor in The Hague. Simon Jennings is IWPR's Africa Editor in London.