116 20210423 Ora 01 00 Bi
116 20210423 Ora 01 00 Bi
116 20210423 Ora 01 00 Bi
Corrected
CR 2021/9
ANNÉE 2021
Audience publique
________________
COMPTE RENDU
________________
YEAR 2021
Public sitting
VERBATIM RECORD
____________________
-2-
M. Gautier, greffier
⎯⎯⎯⎯⎯⎯
-3-
Registrar Gautier
⎯⎯⎯⎯⎯⎯
-4-
S. Exc. M. Bernard Takaishe Ngumbi, vice-premier ministre, ministre de la justice et garde des
sceaux a.i.,
comme agent ;
Mme Raphaëlle Nollez-Goldbach, directrice des études droit et administration publique à l’Ecole
normale supérieure de Paris, chargée de recherche au Centre national de la recherche scientifique
(CNRS),
M. Philippe Sands, QC, professeur de droit international au University College London et avocat,
Matrix Chambers (Londres),
H.E. Mr. Bernard Takaishe Ngumbi, Deputy Prime Minister, Minister of Justice, Keeper of the Seals
a.i.,
as Head of Delegation;
H.E. Mr. Paul-Crispin Kakhozi, Ambassador of the Democratic Republic of the Congo to the
Kingdom of Belgium, the Kingdom of the Netherlands, the Grand Duchy of Luxembourg and the
European Union,
as Agent;
Mr. Ivon Mingashang, member of the Brussels and Kinshasa/Gombe Bars, Professor and Head of
the Department of Public International Law and International Relations at the Faculty of Law,
University of Kinshasa,
Ms Monique Chemillier-Gendreau, Emeritus Professor of Public Law and Political Science at the
University Paris Diderot,
Mr. Mathias Forteau, Professor of Public Law at the University Paris Nanterre,
Mr. Pierre Bodeau-Livinec, Professor of Public Law at the University Paris Nanterre,
Ms Raphaëlle Nollez-Goldbach, Director of Studies in Law and Public Administration at the Ecole
normale supérieure, Paris, in charge of research at the French National Centre for Scientific
Research (CNRS),
Mr. Pierre Klein, Professor of International Law at the Université libre de Bruxelles,
Mr. Nicolas Angelet, member of the Brussels Bar and Professor of International Law at the
Université libre de Bruxelles,
Mr. Olivier Corten, Professor of International Law at the Université libre de Bruxelles,
Mr. Jean-Paul Segihobe Bigira, Professor of International Law at the University of Kinshasa and
member of the Kinshasa/Gombe Bar,
Mr. Philippe Sands, QC, Professor of International Law at the University College London and
Barrister, Matrix Chambers, London,
Mr. Jacques Mbokani Bateghana, Doctor of Law of the Université catholique de Louvain and
Professor of International Law at the University of Goma,
-6-
comme conseils ;
comme conseillers ;
M. Jean-Jacques Tshiamala wa Tshiamala, avocat au barreau du Kongo central (Eureka Law Firm
SCPA) et assistant en droit international au Centre de recherche en sciences humaines à Kinshasa,
Mme Blandine Merveille Mingashang, avocate au barreau de Kinshasa/Matete (Eureka Law Firm
SCPA) et assistante en droit international au Centre de recherche en sciences humaines à
Kinshasa,
M. Glodie Kinsemi Malambu, avocat au barreau du Kongo central et assistant en droit international
au Centre de recherche en sciences humaines à Kinshasa,
Mme Espérance Mujinga Mutombo, avocate au barreau de Kinshasa/Matete (Eureka Law Firm
SCPA) et assistante en droit international au Centre de recherche en sciences humaines à
Kinshasa,
M. Amani Cirimwami Ezéchiel, Research Fellow au Max Planck Institute Luxembourg for
Procedural Law et doctorant à l’Université catholique de Louvain et la Vrije Universiteit Brussel,
comme assistants.
comme agent ;
-7-
as Counsel;
Mr. Justin Okana Nsiawi Lebun, Professor of Economics at the University of Kinshasa,
Mr. Pierre Ebbe Monga, Legal Counsel at the Ministry of Foreign Affairs,
Ms Nicole Ntumba Bwatshia, Professor of International Law at the University of Kinshasa and
Principal Adviser to the President of the Republic in Legal and Administrative Matters,
as Advisers;
Mr. Sylvain Lumu Mbaya, PhD student in international law at the University of Bordeaux and the
University of Kinshasa, and member of the Kinshasa/Matete Bar (Eureka Law Firm SCPA),
Mr. Jean-Paul Mwanza Kambongo, Lecturer at the University of Kinshasa and member of the
Kinshasa/Gombe Bar (Eureka Law Firm SCPA),
Mr. Jean-Jacques Tshiamala wa Tshiamala, member of the Kongo Central Bar (Eureka Law Firm
SCPA) and Lecturer in International Law at the Centre de recherche en sciences humaines in
Kinshasa,
Ms Blandine Merveille Mingashang, member of the Kinshasa/Matete Bar (Eureka Law Firm SCPA)
and Lecturer in International Law at the Centre de recherche en sciences humaines in Kinshasa,
Mr. Glodie Kinsemi Malambu, member of the Kongo Central Bar and Lecturer in International Law
at the Centre de recherche en sciences humaines in Kinshasa,
Ms Espérance Mujinga Mutombo, member of the Kinshasa/Matete Bar (Eureka Law Firm SCPA)
and Lecturer in International Law at the Centre de recherche en sciences humaines in Kinshasa,
Mr. Trésor Lungungu Kidimba, PhD student in international law and Lecturer at the University of
Kinshasa and member of the Kinshasa/Gombe Bar,
Mr. Amani Cirimwami Ezéchiel, Research Fellow at the Max Planck Institute Luxembourg for
Procedural Law and PhD student at the Université catholique de Louvain and the Vrije
Universiteit Brussel,
as Assistants.
The Hon. William Byaruhanga, SC, Attorney General of the Republic of Uganda,
as Agent;
-8-
S. Exc. Mme Mirjam Blaak Sow, ambassadrice de la République de l’Ouganda auprès du Royaume
de Belgique,
M. John Bosco Rujagaata Suuza, chef du service des contrats et des négociations,
comme conseils ;
M. Dapo Akande, professeur de droit international public, Université d’Oxford, Essex Court
Chambers, membre du barreau d’Angleterre et du pays de Galles,
M. Lawrence H. Martin, avocat au cabinet Foley Hoag LLP, membre des barreaux de la Cour
suprême des Etats-Unis d’Amérique, du district de Columbia et du Commonwealth du
Massachusetts,
M. Yuri Parkhomenko, avocat au cabinet Foley Hoag LLP, membre du barreau du district de
Columbia,
Mme Rebecca Gerome, avocate au cabinet Foley Hoag LLP, membre du barreau du district de
Columbia et du barreau de New York,
M. Peter Tzeng, avocat au cabinet Foley Hoag LLP, membre du barreau du district de Columbia et
du barreau de New York,
M. Benjamin Salas Kantor, avocat au cabinet Foley Hoag LLP, membre du barreau de la Cour
suprême du Chili,
M. Ysam Soualhi, chercheur au Centre Jean Bodin (CJB) de l’Université d’Angers,
comme conseils ;
-9-
H.E. Ms Mirjam Blaak Sow, Ambassador of the Republic of Uganda to the Kingdom of Belgium,
as Deputy Agent;
as Counsel;
Mr. Dapo Akande, Professor of Public International Law, University of Oxford, Essex Court
Chambers, member of the Bar of England and Wales,
Mr. Pierre d’Argent, Professor of International Law at the Université catholique de Louvain, member
of the Institut de droit international, Foley Hoag LLP, member of the Brussels Bar,
Mr. Lawrence H. Martin, Attorney at Law, Foley Hoag LLP, member of the Bars of the United States
Supreme Court, the District of Columbia and the Commonwealth of Massachusetts,
Mr. Sean Murphy, Manatt/Ahn Professor of International Law, The George Washington University
Law School, member of the Bar of Virginia,
Mr. Yuri Parkhomenko, Attorney at Law, Foley Hoag LLP, member of the Bar of the District of
Columbia,
Mr. Alain Pellet, Emeritus Professor at the University Paris Nanterre, former Chairman of the
International Law Commission, member of the Institut de droit international,
Ms. Rebecca Gerome, Attorney at Law, Foley Hoag LLP, member of the Bars of the District of
Columbia and New York,
Mr. Peter Tzeng, Attorney at Law, Foley Hoag LLP, member of the Bars of the District of Columbia
and New York,
Mr. Benjamin Salas Kantor, Attorney at Law, Foley Hoag LLP, member of the Bar of the Supreme
Court of the Republic of Chile,
Mr. Ysam Soualhi, Researcher, Centre Jean Bodin (CJB), University of Angers,
as Counsel;
- 10 -
S. Exc. M. Arthur Sewankambo Kafeero, directeur a.i. des affaires régionales et internationales,
ministère des affaires étrangères,
comme conseillers.
- 11 -
H.E. Mr. Arthur Sewankambo Kafeero, acting Director, Regional and International Affairs, Ministry
of Foreign Affairs,
as Advisers.
- 12 -
The PRESIDENT: Please be seated. The sitting is open. The Court meets this afternoon for
For reasons that he has duly made known to me, Judge Abraham is unable to be present on the
I shall first explain the procedure to be followed during the questioning of the experts. The
experts have followed the first round of hearing via web stream and will now be questioned via
WebEx video link. They will be called for questioning in the order in which their respective
contributions appear in the expert report. So, today, questions will first be put to Mr. Henrik Urdal
and then to Ms Debarati Guha-Sapir. On Monday, questions will be put to Mr. Geoffrey Senogles
and then to Mr. Michael Nest. Before being called for questioning, the relevant expert will have
joined the meeting by video link using the WebEx platform. The expert will then be invited to make
a solemn declaration provided for in Article 64 (b) of the Rules of Court. With respect to each expert,
the Democratic Republic of the Congo will be afforded 30 minutes to put questions. Thereafter,
Uganda will be afforded 30 minutes to put questions to that expert. And then, finally, there will be
20 minutes for any follow-up questions from individual Judges to that expert. If any aspect of the
questioning takes less time than has been allocated, the schedule for that day’s sitting will be
advanced accordingly. Between the questioning of the two experts, I expect that the Court will
The first expert to take questions this afternoon is Mr. Urdal, who has joined the meeting via
video link.
Good afternoon, Mr. Urdal. I call upon you to make the solemn declaration for experts as set
Mr. URDAL: I solemnly declare upon my honour and conscience that I will speak the truth,
the whole truth and nothing but the truth and that my statement will be in accordance with my sincere
belief.
The PRESIDENT: Thank you, Mr. Urdal. I now give the floor to Ms Michelle Butler who will
put questions to Mr. Urdal on behalf of the Democratic Republic of the Congo. You have the floor,
Madam.
- 13 -
appear before you today and to represent the Democratic Republic of the Congo in these proceedings.
Mr. Urdal, my name is Michelle Butler. I’m a barrister at Matrix Chambers in London and it’s
my role to ask you some questions on behalf of the DRC. I have only 30 minutes and I have quite a
lot to cover, so I’d be very grateful if you could try to keep your answers as succinct as possible. I’ll
Mr. URDAL: I’m the Director of the Peace Research Institute Oslo, and also a Research
Professor at the same institute. My area of specialty is demographic causes and consequences of
armed conflict.
Ms BUTLER: Thank you. You were appointed by the Court to estimate the number of civilian
deaths occurring during the DRC armed conflict between 1998 and 2003?
Mr. URDAL: That is correct, and specifically the direct deaths in civilians.
Ms BUTLER: Thank you. You and Professor Guha-Sapir were both appointed to opine on that
question. She was appointed to look at excess deaths. You were appointed to look at direct civilian
deaths. Have you previously come across each other in your work?
Mr. URDAL: We have not worked together, but I am well aware of Ms Guha-Sapir’s work.
Ms BUTLER: Is it right that you invited her to deliver the 2018 Annual Peace Address at your
Ms BUTLER: Thank you. I’d now like to ask you a few questions to assist the Court in
understanding how your and her approaches differ, if I may. So we’ve already touched slightly upon
this. Ms Guha-Sapir uses an epidemiological approach to calculating excess deaths; is that right?
- 14 -
Ms BUTLER: And you adopt a different approach. You apply a political science perspective?
Ms BUTLER: Your analysis relies on the Uppsala Conflict Data Program’s Georeferenced
Events Dataset (UCDP-GED). Do you mind explaining to the Court very briefly what that is?
Mr. URDAL: So the University of Uppsala collects global data on armed conflict and
specifically on armed conflict deaths. That includes three different types of conflict between armed
actors, where at least one of the actors is a State, and that’s what we define as State-based conflict.
Armed conflict between two parties of which neither is a State, and that’s defined as . . . I have to . . .
Mr. URDAL: No, that’s fine. Non-State conflict, but with armed actors. And then we have a
final category of one-sided violence, which is violence directed by an armed group, either a State or
Ms BUTLER: Thank you. So in short, the dataset measures battle-related deaths and deaths
Ms BUTLER: Right. So, in fact, there is no inherent inconsistency between your report and
Mr. URDAL: That is correct. And that is also explicitly stated in the report.
- 15 -
Ms BUTLER: Thank you. Now, I’ll just get some more detail from you, if I may, as to the
Uppsala dataset. As you mentioned, it’s with a university in Sweden. It’s independent. Would you
Mr. URDAL: It is a trusted dataset. It’s used both extensively in academic work and it’s also
used by policy ⎯ the policy community from the United Nations and two to governments.
Mr. URDAL: I would say that it is much used both because it’s clearly independent. It’s
following very strict definitions of what an armed conflict is, and it certainly is also open in a sense
that it is accessible to everyone, which also means that it is ⎯ the Uppsala Conflict Data Program is
also taking feedback systematically and revising their estimates when new evidence becomes
available.
Ms BUTLER: Thank you. And you explained just then that it’s got a very strict definition of
armed conflict. In relation to the inclusion of incidents within the database, does it adopt a fairly lax
Mr. URDAL: It is a very strict inclusion criteria for events in the so-called Events Dataset, and
it’s based on ⎯ in order to be included, it has to be known actors. They will have to be announcing
a name. They have to announce the reason for the political inconsistency. It has to pass at least
25 battle-related deaths per year, so it is certainly a very strict definition of armed conflict.
Ms BUTLER: Thank you. I’d like to ask you some questions now about whether certain deaths
are included in the dataset. Before I do so, I want to make clear that nothing I say is intended as a
criticism of you or your method. I’m simply seeking to assist the Court to understand what the dataset
does include and what the dataset doesn’t include. So does it include deaths from malnutrition caused
Ms BUTLER: Thank you. Does it include deaths from preventable diseases such as malaria or
population displacements?
Ms BUTLER: So the dataset’s strict inclusion criteria means it doesn’t measure indirect
Ms BUTLER: Thank you. Now, I’d like to turn now to some of the limitations of the dataset
that you actually touched upon in your report. Is it right to say that if a direct conflict death is not
Mr. URDAL: That is correct. It has to be a written source that is available to the coders that
Ms BUTLER: And can you briefly explain, is media or an urban bias a particular problem
Mr. URDAL: Here I need to speak on general terms because I have no specific knowledge
about the data collection on the DRC during this period, when it comes to the availability specifically
of data. But generally there is a known bias in datasets like this, that are based on news sources, that
there are biases both in terms of differences between countries, differences over time, as well as
differences then within countries, and an urban versus rural bias is certainly among those kind of
Ms BUTLER: Thank you. So you explain that there is a bias between countries. Would you
say that there can be a systematic under-reporting of African events versus non-African events in the
mainstream media?
Mr. URDAL: I would not be surprised if there is such a general bias, but I do not have any
solid evidence to say anything about what level of bias that might be. But as I said, the data is based
on news sources and news reporting, and to the extent that media presence is different between
different regions of the world and between different countries and between different regions within
Ms BUTLER: And that point there that you’ve just made about regions within countries,
would it be right that, say, deaths occurring in a very remote location within a country, or in a location
with very poor cell phone coverage, would be unlikely to be reported on by the media and therefore
Mr. URDAL: That is a likely assumption, but it depends entirely on media presence. And, of
course, if regions are of particular interest for one reason or another, it might see increased media
presence, and it’s the media presence and the presence of organizations, and especially international
non-governmental organizations, that are the sources of this information. So it will depend and will
Ms BUTLER: Now, media presence that you have just touched on there. So if we have an area
of, say, very dense fighting, would it be likely that the media would be present right in the thick of
that fighting and being able to report on deaths happening right there, or would it be more likely that
they would actually have to be a little bit back from the fighting for safety reasons?
Mr. URDAL: This is difficult for me to be entirely precise about, because this again will differ
from situation to situation. I would say that it does not depend ⎯ this kind of data does not depend
on journalists being present exactly when the conflict events are taking place. It might be reported
based on reports that could come in a significant time after the actual event happened. So even though
security issues will, of course, determine, to some extent, whether or not journalists as well as other
- 18 -
observers can be present in specific regions, it is not necessary that the observers are present at the
Ms BUTLER: Thank you. So it is not necessary that the observers are present at the time of
the event. But, for example, for events where there are no surviving witnesses, it would be quite hard
for the media to report on that, because there would be nobody left to tell their stories to the media.
And so those kinds of incidents would not be reported in the database; is that right?
Ms BUTLER: And what about incidents where a person does not die immediately? So, for
example, let us say a woman surviving being raped initially but then some weeks later she succumbs
Mr. URDAL: That is dependent. Sometimes ⎯ so that will be a grey zone. And if the media
reporting is sufficiently detailed, that particular death can be, with some level of certainty, connected
to a specific event, then that death will be counted as a direct death and associated with that specific
event. But it could very well be that there are clear connections between violence happening and the
Ms BUTLER: So the key point then, from what you are telling me, is that there has got to be
this detailed media report of each death, and in the absence of that, it is not on the dataset?
Ms BUTLER: I would like to touch now on confidence intervals. This is something that you
dealt with in your report. As I understand it, the Uppsala dataset has three confidence intervals.
Mr. URDAL: Yes. The Uppsala conflict database has a minimum, a most likely, and a
It is important to underscore that the UCDP has a very conservative cautious approach so that
the minimum estimate is quite often identical to the best estimate. The best estimate is either often
- 19 -
identical to or just marginally higher than the minimum estimate. While the maximum estimate is an
estimate that is not the highest possible number but the highest reliable number.
And there might be a news report that, for instance, gives ranges, so between 50 and 70 people
were killed. And if that is then considered reliable information, the UCDP would normally call 50 as
the minimum estimate, and also the best estimate, and 70 as the maximum estimate.
Ms BUTLER: Thank you. That is very helpful. So, in effect, the best estimate is often the same
as the minimum. And we see that in your report with respect to the DRC where, I think, if we round
the figures off, we have approximately 28,000 in the low estimate, 29,000 in your best estimate, and
Mr. URDAL: That is correct. And that is all direct. That is in the whole of the DRC.
Ms BUTLER: Thank you. And so when you say the maximum is not the highest possible, it
is right, is it not, that you have described that as still being a conservative assessment because it does
still rely on that inclusion of the data through these news media reports and NGO reports?
Mr. URDAL: That is correct. There are a number of biases that are impossible to estimate with
Ms BUTLER: But you adopted the best estimate in your report rather than the high estimate,
even though even the high estimate is quite conservative. Are you able to explain why you adopted
that estimate?
Mr. URDAL: That is the estimate that is used in virtually all works as simply the best number
that we can, with some cautious certainty, say that we have evidence that existed. And it is important
also to say that when the UCDP is considering reliability, of course, it is not always possible to say
with any great certainty how we may best assess such reliability.
And this is also why I think it is important to underscore that in conflict settings, much
information is uncertain. The parties, obviously, have an interest in either inflating or reducing
numbers. And that is why the approach is taken to be cautious when estimating casualties and why
- 20 -
also the best estimate is the one that is used both in academic work as well as in most of the policy
work.
Ms BUTLER: Thank you. So this cautious approach, you were asked by the Court to estimate
Ms BUTLER: And if you had chosen to adopt the high estimate, could you have done that?
Are the military deaths and the civilian deaths separated in the high estimate dataset?
Ms BUTLER: Right. So in order to answer the Court’s question, in effect, you were sort of
shoe-horned into using the best estimate rather than using the high estimate, because otherwise you
Mr. URDAL: It was not possible to make the same kind of estimation as I did for my report
Ms BUTLER: Thank you. That leads me on to my next topic, which is the issue of the military
So looking at your best estimate numbers, of those your report indicates that there are
approximately 8,000 deaths which could not be counted in your final numbers because we could not
determine from the dataset whether they were civilian or military. There was not enough data to
Mr. URDAL: That is correct. It is 7,824 deaths that cannot be categorized as either military or
civilian.
Ms BUTLER: And is it likely that a substantial proportion of those 8,000 deaths were
civilians?
- 21 -
Mr. URDAL: I would say it is likely that a significant portion are civilians, but there is no
information available that could give us an indication of what proportion we would be talking about.
Ms BUTLER: Okay. And is it likely that certain demographics of victims, like young men, for
instance, might be wrongly recorded as military victims even though they were not taking part in
Mr. URDAL: It is a possibility. But, again, the UCDP is careful about the way that they are
coding, and the number of military victims among the 28,000 in total is relatively low. It is less than
Mr. URDAL: I would definitely agree that there is a likely undercount. This can be considered
Ms BUTLER: Yes, okay. So this minimum number that you have settled on using the best
estimate dataset — so we’ve got 11,227 targeted civilian deaths and 3,436 collateral civilian deaths.
In light of what you’ve very carefully and helpfully explained about the very strict inclusion criteria
of the dataset, what’s the likelihood that those numbers are a substantial underestimate of the true
Mr. URDAL: I would say that it is almost certainly an underestimate. The question is what the
margin of error is, and that is simply impossible to say anything about. But we know, given the
significant sources of biases, that this is likely to undercount the total number of casualties in total,
both, as I mentioned, because of the inclusion criteria as well as — sometimes ⎯ the difficulty in
approach. We know that, but it’s impossible to estimate the scale to which it’s an undercount; is that
right?
- 22 -
Ms BUTLER: Thank you. I’d just like to ask you one further set of questions, Professor Urdal,
in relation to the heavy reliance that was placed by Uganda yesterday in their submissions. I think
four or five times during the hearing, the counsel for Uganda kept on returning to this figure of 32.
They were saying that your report proves that there are 32 civilian deaths in the whole of the DRC,
in the whole of the war, that involved Ugandan troops. What’s your reaction to that suggestion?
Mr. URDAL: So this is the number specifically that is related to activities where we can say
demonstratively that the Government of Uganda was on one of the sides, and it is also a number
specifically that excludes both events that are undetermined — as to who the actors were and where
there might also be uncertainty. So, again, I would caution against saying that this is the number, but
it is a minimum number.
But I would also like to underscore that I was not asked to undertake a full analysis of deaths
Ms BUTLER: Yes, that’s right. So what you were saying there about uncertainty and
unknowns, is it right that if it is not known who were the actors involved in a particular death, that
Mr. URDAL: That is correct. If we cannot assess who the actor is, it is not included in the
dataset.
Ms BUTLER: And how likely do you think it is that Ugandan forces only killed 32 civilians
Mr. URDAL: That is beyond my competence to respond to. I have no particular area of
Ms BUTLER: And that is because the Uppsala dataset is not designed to determine the legal
The PRESIDENT: I thank you, Ms Butler. And I give the floor now to Mr. Lawrence Martin
to put questions to Mr. Urdal on behalf of Uganda. You have the floor, Mr. Martin.
Good afternoon, Mr. Urdal. I apologize for the somewhat awkward nature of this questioning,
since I cannot look directly at you while we are speaking together, and I hope you will be patient
with me.
I am going to cover some of the same ground that you just covered, but I will do my best not
to be duplicative. So, I think we have established that in your report you examined the Uppsala
conflict data programme or UCDP; is that correct? Did you hear me, Mr. Urdal?
Mr. MARTIN: Okay, thank you. Sorry, I did not hear your response.
Mr. URDAL: The UCDP data is considered to be the most, or at least one of the most,
authoritative sources. They have a very stringent definition of conflict and it makes it possible to
very clearly delineate what the methodology has been for the inclusion of the data.
It is also a global dataset. It has been collected for now a total of around 25 years, so it is a
dataset that has been over the years developed into a very trusted source.
Mr. MARTIN: In fact, in paragraph 14 of your report you referred to it as the authoritative
Mr. MARTIN: And you also characterized it at paragraph 17 as one of the most trusted sources
of armed conflict data in academia as well as the policy domain; is that correct?
Mr. MARTIN: And you said also that it is frequently referenced in key policy reports by
international organizations like the UNDP and the World Bank: is that correct?
Mr. MARTIN: You also said, I believe at paragraph 25, that it meets the highest academic
Mr. MARTIN: Thank you. Thank you. Are you aware of the Armed Conflict Location & Event
Mr. URDAL: Yes, I am. It was originating at my own institution, the Peace Research Institute
Oslo.
Mr. MARTIN: And is there is a reason you chose to use the UCDP over the ACLED dataset?
Mr. URDAL: These are comparable datasets. The UCDP has a slightly different, sort of,
definition than ACLED. And when it comes to defining clearly the difference between actors, the
structure of the UCDP datasets make for an easier comparison. But I would like to underscore that
also the ACLED dataset is a very trusted and much used source with a lot of the same information
Now, I would like to discuss the results of your analysis of the dataset for a moment, if I may.
You went over some of this ground with counsel for the DRC. But, in your report you chose numbers
Mr. MARTIN: And in your report at paragraph 21 you characterize a best estimate as being
considered to be the most reliable estimate of deaths identified in the sources: is that right?
- 25 -
Mr. MARTIN: If I understood correctly in your responses just a few moments ago, you also
described it as the most likely number and the most often used number: is that correct?
Mr. URDAL: I would caution to say “the most likely” because it is the most reliable, given
the sources that we have. And as I did describe, there are biases that, in many cases, are in reality
unknown. So, we do not know what the margin of error is. But we do not have any better and more
Mr. MARTIN: Okay. So, you stand by your observation that that is the most reliable estimate?
Mr. MARTIN: Thank you. And the result of your analysis suggested that the best estimate was
that, throughout the entire territory of the DRC, there were 14,663 civilian deaths from direct
Mr. URDAL: That is correct. That is for the period of 1 August 1998 through 2 June 2003.
Mr. MARTIN: And again, that is throughout the entire territory of the DRC?
Mr. MARTIN: Now, there were some discussion earlier of the fact that for some of those
deaths it was unknown whether they were civilian deaths or military deaths; is that correct?
Mr. MARTIN: Okay. In your report, you stated that there were 23 events across the DRC
recorded as having involved the troops of Uganda as one of the actors of the dataset; is that correct?
Mr. MARTIN: And most of those events took place in North Kivu; is that correct?
- 26 -
Mr. URDAL: I have to admit that my detailed knowledge of geography of the DRC is limited.
But, according to my notes, yes, the majority of the events ⎯ exactly 20 ⎯ took place in the province
Mr. MARTIN: Thank you. And your results indicate that there was a total of 211 people
reported killed, of which 179 were military deaths and 32 were civilians; right?
Mr. URDAL: That is correct. And that is specifically events that involved the Government of
Mr. MARTIN: Thank you. And in this instance, in events involving troops of Uganda, there
were zero unknown deaths. That is, zero deaths that could not be attributed to as either military or
Mr. MARTIN: So according to your analysis then, 32 of 14,663 civilian deaths took place
during events that involved the Government of Uganda as one of the actors; is that right?
Mr. MARTIN: And if I told you that that represents 0.22 per cent of the direct civilian deaths,
Mr. URDAL: It does not surprise me in the sense that this is mathematically entirely correct,
yes.
Mr. MARTIN: The DRC asked you a number of questions about limitations in the dataset and
the fact that that may produce conservative numbers. Do you recall those questions?
Mr. MARTIN: At paragraph 29 of your report, you wrote: “I have no data allowing for an
assessment of what this potential underestimation of direct conflict deaths may amount to.” Is that
correct?
Mr. MARTIN: And in your response to the Parties’ observations on your report, you indicated:
“[I]t is not possible to provide a sound assessment of how significant this underestimation is.” Did
you not?
Mr. MARTIN: So, in other words, any effort to quantify the underestimation would be
Mr. URDAL: Yes. I mean, there are ways that one could do it, including comparing the
different data sources that we discussed with ACLED, but there will be significant uncertainties. And
it will be difficult to establish with any level of certainty what the actual number would be. That is
correct.
Mr. MARTIN: Thank you. Now in your report, you indicated that one of the limitations that
can lead to underestimation of the numbers is that it “only includes events associated with armed
conflicts in which at least 25 battle deaths have occurred within a calendar year”; is that right?
Mr. URDAL: Yes, that is correct. So, events of violence happening between actors that do not
Mr. MARTIN: But that limitation is not relevant to the DRC, is it?
Mr. URDAL: It is relevant to the DRC, but we do not know, because there might be violent
organizations that engaged in a number of armed groups conflicts that were active in the DRC, also
that included in the Ituri province, that may have been clashing between each other. So it is a pair of
actors, which means that if some pairs of these actors had violent clashes leading to deaths that did
- 28 -
not pass the 25 deaths in a year, those deaths belonging to that specific dyad, that specific pair of
Mr. MARTIN: I see. I understood your report to mean that 25 battle-related deaths in a country
Mr. URDAL: No, that was not the case. It is 25 battle deaths between pairs of actors.
Mr. MARTIN: Very good. Another limitation you mentioned concerned changing reporting
practices; right?
Mr. MARTIN: But in your report at paragraph 26, you said that this is “considered less of a
problem when studying one country during a limited [period of time], such as the DRC over the
Mr. MARTIN: And in response to questions from counsel for the DRC about whether the fact
that a location being rural and remote might result in it receiving less press attention, you responded
Mr. URDAL: I have no direct knowledge about this issue for the case of the DRC, no.
Mr. MARTIN: Thank you. At paragraph 40 of your report, you listed a total estimated value
Mr. MARTIN: But all you did there was multiply the total number of civilian deaths
throughout the DRC during the relevant period ⎯ multiply those figures ⎯ by the figures suggested
Mr. MARTIN: So you are not actually recommending that Uganda pay that amount?
Mr. MARTIN: And, in fact, in your response to the observations of the Parties, you
specifically said: “No attribution to Uganda of a particular number of deaths has been made”; is that
correct?
Mr. URDAL: Yes, that’s correct. I am making no judgment about the attribution of any deaths.
Mr. MARTIN: Very good. Thank you. Those are all the questions I have, Dr. Urdal. Thank
Madam President, if I may, just for the record: counsel for the DRC characterized our argument
yesterday concerning Mr. Urdal’s report. This is not a moment to really address that, but I just want
The PRESIDENT: Thank you, Mr. Martin. Let me just take a minute to see whether there are
any questions from the judges for this expert. I have two methods for receiving them. One method
from the people who are in the Great Hall, one from the people onscreen.
I am not seeing any questions from the judges for you, Mr. Urdal. So that brings . . . So that
was not the method I had in mind, but it worked, and so if I understand correctly it’s Judge Sebutinde
Mr. Urdal, thank you for your testimony. I am going to ask you a question arising out of your
In paragraph 14, you refer to an estimated total value of civilian lives lost in the amount of
US$388,350,000. You also refer to the total number of civilian deaths as 14,663 deaths. That’s in
that paragraph. We then come to the table shown as table 1.3, I think towards the middle, just before
the appendices. And in that, you repeat these figures, the estimated value remains US$388,350,000.
Now, my question is: is this figure both the total number of civilians and the figure of the
value, does this relate to civilian deaths in the whole of the DRC in the period in question, or just
Mr. URDAL: Yes, the number of civilian or the number of deaths mentioned in paragraph 14
totalling 28,981 deaths, are military as well as civilian as well as unknown deaths in the whole of the
The numbers in table 1.3 are the numbers specifically for the Ituri province, and that has a
lower total number rather than 28,981, the total number of deaths in table 1.3 is 10,398. And it is
distributed across the different categories: civilian deaths, military deaths and unknown deaths.
The total amount that is provided in paragraph 14 ⎯ of US$388,350,000 ⎯ is the amount for
all civilian deaths. That is both: deaths that are deliberate as well as collateral damage according to
So the total amount, again, is the amount for all civilian deaths for the whole of the DRC, as
The PRESIDENT: Do you have any further questions, Judge Sebutinde? Are there any other
questions from someone who is sitting at the Bench in the Great Hall?
Very well. In that case, this brings to an end the questioning of Mr. Urdal. We thank you very
much, Mr. Urdal, for appearing before us, for providing your report.
The PRESIDENT: You may now sign out of the WebEx platform and continue to follow the
remainder of the hearing via webstream if you wish. The Court will now observe a 10 minute break,
after which it will resume for the questioning of Ms Guha-Sapir. Thank you.
The PRESIDENT: Please be seated. The sitting is resumed. The Parties and the Court will
now have an opportunity to put questions to Ms Debarati Guha-Sapir, who has joined the meeting
by video link. Good afternoon, Ms Guha-Sapir. I call upon you to make the solemn declaration for
experts, as set down in Article 64, subparagraph (b), of the Rules of Court.
Ms GUHA-SAPIR: I solemnly declare upon my honour and conscience that I will speak the
truth, the whole truth, and nothing but the truth, and that my statement will be in accordance with
The PRESIDENT: Thank you. I now give the floor to Professor Philippe Sands to put
questions to Ms Guha-Sapir on behalf of the Democratic Republic of the Congo. You have the floor,
Professor Sands.
Mr. SANDS: Thank you, Madam President, and Members of the Court. It’s an honour for me
to appear before you once again on behalf of the Democratic Republic of the Congo after a short
17-year interlude.
Matrix Chambers. I have about 30 minutes with you and a number of questions to address, so I would
be hugely grateful if you could keep your responses as tight and short as possible. I wonder if you
could begin by telling us your professional title and your area of expertise.
Mr. SANDS: Thank you. And you are originally from India; is that right?
Mr. SANDS: Very good. And you studied a Master’s at Johns Hopkins University?
Ms GUHA-SAPIR: Yes.
Mr. SANDS: Thank you so much. And you are appointed by the Court, are you, as an
independent expert?
Mr. SANDS: So, your area of expertise causes you to study, does it, disasters and conflicts by
Ms GUHA-SAPIR: Yes.
Mr. SANDS: Could you briefly explain what that epidemiological approach is?
Ms GUHA-SAPIR: So, the study of disasters and armed conflicts for epidemiologists is really
focusing on risk factors. So that’s the essence of epidemiology; what factors increase your risk of
mortality, or morbidity, in certain settings. And in the settings that I work in, it’s disasters and civil
conflicts, and therefore, for example, my studies would involve looking at what risk factors increase
Mr. SANDS: And would you say that one of your particular areas of specialism is, indeed, in
Mr. SANDS: And that work, has it included assisting with other courts and tribunals, for
Ms GUHA-SAPIR: Yes, I have served as an expert testimony witness for the case on Darfur,
Mr. SANDS: So you’re familiar with the process that we are going through right now, I take
it?
- 33 -
Mr. SANDS: You have been appointed by the Court in this case to assist the judges in
providing a global estimate of lives lost amongst the civilian population due to the armed conflict in
the Democratic Republic of the Congo between 1998 and 2003; is that correct?
Mr. SANDS: And you have prepared for the Court and the judges two reports, one from
Ms GUHA-SAPIR: Yes.
Mr. SANDS: And is it the case that the second report was prepared in response to observations
you received which had been prepared by Uganda and by my client, the Democratic Republic of the
Ms GUHA-SAPIR: Yes.
Mr. SANDS: And is it right that your overall conclusion is that excess mortality in the
Democratic Republic of the Congo between 1998 and 2003 was around 4.98 million human beings
for the country as a whole, of which 401,057 are in Ituri and 3,713,065 are in the eastern region?
Ms GUHA-SAPIR: Yes.
Mr. SANDS: Thank you. Let’s turn now to some of the criticisms made by Uganda in relation
to your methodology.
In its observations, Uganda has criticized you for not basing your mortality estimations on
death certificates, civil registration logs, or mortuary records. Could you tell us why you did not rely
Ms GUHA-SAPIR: Well, anybody who has worked in an African country in those kinds of
settings. Not just in an African country, but in an area where there is a conflict going on and security
is very high, death certificates are often a pipe dream. And the institutions of civil registration,
- 34 -
certification of deaths, are essentially inexistent. And understandably so when you think of some of
the towns in the eastern DRC have paved roads which are about 300 kilometres from the village
itself. So there is no incentive for anybody to go and report a death, because it brings no benefit to
the family and there is, indeed, almost no institutional infrastructure to record those deaths. So I think
it goes without saying that death certificates or civil registration is not going to be a source of any
data.
Mr. SANDS: Sadly, I am familiar with the kinds of accounts that you are giving in that
situation. So in the absence of such records, as you have just described and explained, to reach your
figure you applied an excess mortality methodology. Could you briefly explain to us what that
methodology is?
for deaths which are hard to pin down to a specific cause. So, like, heatwaves in Europe, excess
What does it do? In fact, all it does is it looks at how many deaths would be expected in a
population had nothing ⎯ counterfactual, in fact ⎯ so if nothing had happened, then how many
deaths would you expect; and then you look at the observed number of deaths. So you get the number
of deaths that actually happened, and the difference between what is expected and what is observed
Mr. SANDS: Thank you very much. In fact, just on a personal note, I’m speaking to you from
London, the United Kingdom, and we are, of course, very familiar with this methodology in relation
Ms GUHA-SAPIR: Yes.
Mr. SANDS: We almost have daily figures. Is that the same approach that is being taken in
Ms GUHA-SAPIR: Yes, the basic idea is identical. It’s what would you expect, how many
people would you expect to die if nothing had happened. So you use a baseline, and then you look at
- 35 -
the difference between what you would expect and what you actually observe. And that’s exactly
Mr. SANDS: And so that is a methodology that is tried and tested around the world, is it?
Mr. SANDS: And would you say then that it is a reasonable approach to use in the
Mr. SANDS: Thank you. Let’s turn to a second Ugandan criticism, and that is the choice of
Uganda has criticized your choice of the baseline. You, in your report, have relied on a
UNICEF crude death rate that was carried out in 1997 and published in 1999: is that correct? And if
Ms GUHA-SAPIR: Two reasons. One is we wanted to have a baseline which was drawn from
a respectable source, so the United Nations. We wanted that kind of respectability on the estimate of
the baseline. Secondly, it was a baseline which was in a very appropriate year. So, it was 1997. It
was a year just before the period we are supposed to study, we were looking at, for the impact of the
war, and it was the closest to that period, and it was from the UNICEF, which is a respectable and
reliable source. That was essentially the reason for using it, yes.
Mr. SANDS: Thank you. UNICEF was a Member of the United Nations family.
What the Ugandan Government says is that you’ve used the wrong CDR. What you should have used
was a 2019 calculation that was included in the UN Population Division publication called World
Population Prospects. So, could you explain why you chose to use the UNICEF statistics rather than
the latest statistics? It may be that you have already answered this, but since the critique is quite
robust by our Ugandan friends on that issue, it would be helpful if you could explain why the 2019
there, it is not really statistics drawn from data from the crowd. It is a projection. And if my memory
serves me, it did not provide, say, any kind of confidence intervals in which we could situate the
UN Population Division estimate. But I think the main reason we did not use it was, they are
projections. They kind of project backwards or they project forward, depending on where they are,
and the UNICEF estimate was a much more defensible estimate. It was from the right year. It was
from the year just before the war began, before the conflict stepped up, and the methodology of the
UNICEF and the estimate was the most reliable one that we could choose, yes.
Mr. SANDS: Would it be fair to say that it is more reliable because it offered a closer reflection
Ms GUHA-SAPIR: Yes, absolutely. Absolutely. It was drawn from the region. It was drawn
from the country. Clearly, it was the best baseline to be used. There was nothing else that could really
face the kind of scrutiny that would be required for a baseline, defensible baseline, no.
Mr. SANDS: And I think you say in your second report at paragraph 26, that even if you had
used that 2019 baseline figure, it would have reduced the figure by a very large number of people,
288,640, but only 5.8 per cent of the figure you had used: is that correct?
Ms GUHA-SAPIR: Yes, yes, that’s right. It wouldn’t have changed a great deal in that, in fact.
Mr. SANDS: Very good. Can we come to a third area of criticism from the Ugandan side ⎯
and one that relates to the use of small-scale surveys? Uganda has criticized you for using these
small-scale surveys. And I wonder if I could just ask you a couple of questions on that, rather than
the crude death rates of the 2019 report, which you have just explained why you did not want to use
that.
Are the crude death rates for 1998-2003 contained in the 2019 document based on death
Mr. SANDS: Yes, of course. I’m so sorry. Are the crude death rates for 1998 to 2003, which
are included in the 2019 worldwide population prospects document, are those figures based on death
Ms GUHA-SAPIR: I wouldn’t put my hand on fire about that, but I would be extremely
surprised. I would be really, really amazed if they were dependent on death certificates.
Mr. SANDS: And then a related question: To the best of your knowledge, are the figures in
the 2019 documents based on annual surveys which were conducted each year between 1998 and
Ms GUHA-SAPIR: I am not aware of any annual surveys that have been conducted by the
United Nations in the Democratic Republic of Congo. The only nation-wide surveys that are
conducted are the demographic and health surveys, and that was conducted many, many, many years
ago, so I would, again, be very surprised if there are any surveys on which they are based. These are
A final point on this. Uganda proposes that it is possible to accurately measure excess death
from an armed conflict using crude death rates which are derived from data taken at six-year
Ms GUHA-SAPIR: To be quite frank, I have not seen that proposal. It seems — I cannot get
my head around these six-year intervals, how we could get these excess death rates. This has to be
Mr. SANDS: In your experience, and in your practice over your career, have you ever used
such an approach?
Ms GUHA-SAPIR: No.
Mr. SANDS: Right. And would you say that the approach on the basis of your knowledge and
expertise that you are using is a more accurate approach than the one that Uganda proposes?
- 38 -
by most epidemiologists for this purpose. The approach of excess deaths using a baseline and
observed values has been the way that these calculations are made. I am not even aware of a great
deal of other alternative methods. So, yes, I think the approach that we have put forward is, in fact,
Mr. SANDS: Thank you, Professor Guha-Sapir. It seems on some readings, applying the crude
death rates from the United Nations’ 2019 data as Uganda proposes might show that there were no
excess deaths during the 1998 to 2003 war. Does that strike you as a reasonable conclusion?
Ms GUHA-SAPIR: Frankly, no. You know, I cannot embellish that any more, but the answer
to that is, it does not seem to me like a reasonable conclusion. I think, in the questions, they did
mention the questions raised by Uganda on this, and I think there was also a question of whether or
not the quality of life, in fact, was improved. But, no, it does not seem to me like a reasonable —
Mr. SANDS: Thank you. The surveys that you have placed reliance upon have been criticized
by Uganda on the basis, in part, that they were conducted by advocacy groups, as they call them,
whose independence cannot be relied upon. Do you agree with that critique?
Ms GUHA-SAPIR: The short answer is no. Most of the surveys that I have used ⎯ about
38 separate surveys ⎯ have been conducted by Coghlan, who is a professor in Melbourne, I think,
right now — he’s an Australian professor of some renown and some reputation; by Professor Les
Roberts, who is a professor in Columbia School of Public Health, by the CDC, the United States
And I think one could safely say that all of these sources are respectable sources done by
people who have a reputation to protect and it is very unlikely they have simply put together a survey
Mr. SANDS: Indeed. In fact, I just had a look earlier today at the CDC website and noticed
that precisely on COVID-19 they use a methodology on excess mortality that is very much akin to
the one that you use, so that does seem to endorse your approach.
- 39 -
Can I turn to another subject, as we reach a close? And that is the appropriateness of the
application by the Democratic Republic of Congo of coefficients. And I appreciate that you are not
asked in your expert report to attribute or to come up with specific numbers, but you do say a little
And if you are comfortable ⎯ if you prefer to have that in front of you, that is absolutely fine,
of course. At paragraph 47 of your first report, you describe the concerns of the Democratic Republic
of the Congo on challenges in estimating numbers of deaths, and you characterize them as legitimate
and widely acknowledged by the academic and research community. Is that right? That is your view
Ms GUHA-SAPIR: The legitimacy applies to, really ⎯ pertains to the estimation of deaths in
these settings, and that they are a challenge in any of these settings is absolutely legitimate and is
widely acknowledged, which is the reason for which there is always a great deal of debate in
academia about these things, as you probably are aware, of Iraq, of Darfur and now in Congo.
I am less able to say anything, really, intelligent or even instructed, on the coefficients. I
really ⎯ it is hard for me to make a pronouncement on that. In the coefficients that the DRC has
used, that’s 10 per cent ⎯ 45 per cent of 10 per cent, is what it is. And I have simply applied it to
show what it would ⎯ how my estimates would compare to the Congo’s estimates.
Mr. SANDS: Understood. I mean, we are not for a moment putting to you that you have
endorsed these in any way. That is very clear and very helpful and, if I may say, very fair in your
report. I am simply making the point, and please correct me if I am wrong, that you are, in effect,
saying that in the circumstances in which the Democratic Republic of Congo found itself, it was
legitimate for it to find an approach to work out what the numbers might be?
here, yes.
Mr. SANDS: And you do also say, I think, at paragraph 47 in relation to the 10 per cent, which
leads to the Democratic Republic of Congo’s figure of an estimated 400,000 deaths, you describe
that as, using your word in quotes, as “conservative”. Do you stick by that?
- 40 -
Ms GUHA-SAPIR: Given the condition of Ituri and eastern Congo at that time, in those five
years, and the complete breakdown ⎯ I mean, I am just ⎯ the complete breakdown of any primary
healthcare, any vaccination programme, any child support, I think, honestly, I think that 10 per cent
Mr. SANDS: I have understood you very clearly, Professor Guha-Sapir, on this, and I am
certainly not pushing you any further than that. I do note that applying those coefficients, the DRC’s
figure ended up being 182,000 human beings, and you ended up on those coefficients at 244,500. So,
Ms GUHA-SAPIR: It is.
Ms GUHA-SAPIR: No, it’s not. It is not. It is not a huge difference between the Congo’s
estimates. Ours is higher. Mine is higher, somewhat higher, yes. But it is not a huge difference, that
is correct, yes.
Mr. SANDS: I have just got two or three concluding points, very briefly. You may have heard
before, when my colleague Ms Butler asked Professor Urdal about the relationship between his
figures and your figures, and he explained that there was no inconsistency between the two sets of
figures. Is that also your view, that you and he have engaged in a different exercise and that the
figures he has come up with are, in effect, subsumed within your larger figure? Is that a fair way of
putting it?
Ms GUHA-SAPIR: Yes, it is. I think Mr. Urdal explained it very well. I think I have been
looking at excess deaths which are indirect deaths. They do subsume the UCDP estimates, because
essentially an excess death, if there was no war, the soldiers would not have died or the
combat-related deaths would not have occurred. So it does include Urdal’s estimates, and we were
Mr. SANDS: Yesterday, as you may have heard, counsel for Uganda, Mr. Martin, criticized
you on a number of occasions, and I am going to be very precise about what he said. He described
your estimates as “unreliable” and he described your methodology as “unreliable”. That is his word.
Ms GUHA-SAPIR: You know, I would be interested to hear what would be a more reliable
methodology. That would interest me as an academic, and I am always open for, you know, to stand
corrected. But at this time, I think what I did comes closest to a reliable estimate for excess deaths
using the best methodical approaches that are prevailing at this moment, yes.
Mr. SANDS: Can I take it from your answer that, notwithstanding the written and oral
challenges to the reliability of your reports, your position is that you maintain the positions you have
Ms GUHA-SAPIR: Yes, I do maintain the position. I am confident about we have done, what
Mr. SANDS: And, finally, one last question: is there any reason you might be aware of, why
Ms GUHA-SAPIR: No, I think that ⎯ I mean, I don’t see any reason not to make it public.
Mr. SANDS: We have had this exchange of questions which are certainly very public, and
Madam President, that brings the Democratic Republic of Congo to the end of its questions
for Professor Guha-Sapir, and we thank her for her report and her time today.
The PRESIDENT: Thank you, Professor Sands. I now give the floor to Mr. Lawrence Martin
to put questions to Ms Guha-Sapir on behalf of Uganda. You have the floor, Mr. Martin.
Mr. MARTIN: Thank you again, Madam President, distinguished Members of the Court. And
good afternoon, Dr. Guha-Sapir. I will apologize to you, as I did to Dr. Urdal before you, for the
- 42 -
somewhat awkward nature of this conversation we are about to have. I am looking at you on a small
piece of a video screen off to my right, and I am not quite sure what you are seeing. So I apologize
I’d like to begin with a point that you’ve already discussed with Professor Sands, but just so
we are proceeding from a common platform. As you described it, your methodology for estimating
the number of excess deaths attributable to the conflict involves comparing the hypothetical death
rate had the conflicts not occurred ⎯ that is the baseline crude death rate ⎯ with the nominal death
rate during the conflicts, the posterior crude death rate: is that correct?
Ms GUHA-SAPIR: Yes.
Mr. MARTIN: Thank you. And in your opinion, since you calculate the baseline crude death
rate as being lower than the posterior crude death rate, there were a large number of excess deaths
Ms GUHA-SAPIR: Yes, “large” is a qualitative term. So, yes, there was a number of excess
deaths, yes.
Mr. MARTIN: And in your opinion you have estimated the baseline crude death rate by
reference to UN data for 1997, the year before the conflict: correct?
Mr. MARTIN: And that number was 14 deaths per 1,000 people per year ⎯ or 1.2 per 1,000
Mr. MARTIN: And you got that number from a 1999 UNICEF publication: correct?
Mr. MARTIN: In fact, I believe in your questions and answers with Professor Sands, you
Ms GUHA-SAPIR: Yes.
Mr. MARTIN: Are you aware that the number itself actually came from the UN Population
Division?
Ms GUHA-SAPIR: Yes, yes, absolutely. I think they use each other’s numbers and they adjust
So, UNICEF typically does use the UN Population Division’s numbers. But they are not
identical, as you may have noticed as well. The figures on the UNICEF report is not identical to the
Ms GUHA-SAPIR: Yes, so, UNICEF does use the UNDP numbers, but they adjust it to their
own using their own methodology. They have a special working group, a task force, a mortality task
In this particular case, I do not know what adjustments were made but the numbers are not the
same.
Mr. MARTIN: In your report you used a different method to estimate that posterior crude
Ms GUHA-SAPIR: Yes.
Mr. MARTIN: Okay. And the number you came up with was 2.929 per 1,000 people per
Mr. MARTIN: I am afraid I do not have the paragraph in front of me. It is in one of the tables.
Mr. MARTIN: I am referring to the number that you calculated for the posterior crude death.
- 44 -
Mr. MARTIN: And it is 2.929 per 1,000 per month; is that correct?
Ms GUHA-SAPIR: Not quite. Is it? It is ⎯ from what I see, sorry. Yes, I am sorry. I am sorry,
Mr. . . .
Ms GUHA-SAPIR: Martin, sorry. Yes, you are right. You are right.
Mr. MARTIN: Okay. And are you aware that the most recent 2019 revision of the World
Population Prospects published by the UN Population Division indicates that the crude death rate in
the DRC in 1997 was 1.386 per 1,000 people per month?
Ms GUHA-SAPIR: No, I cannot say I am familiar with that. The posterior crude death rates
that we have calculated are based on surveys in the region that have been done in the Congo, and our
position, my position, on this is that using surveys from the country itself, from the study site has a
much more reliable estimate of what is going on in the country than the projections made by the
United Nations.
Mr. MARTIN: I understand that, Dr. Guha-Sapir. But for the moment, I am just asking you
Ms GUHA-SAPIR: Yes.
Mr. MARTIN: And if I told you that the crude death rate, according to the UN Population
Division in 1997, was 1.386 deaths per 1,000 people, would you have any reason to disbelieve that?
difference between the methodologies that have been used to come to 1.386 or come to 2.929. It is a
difference in methodologies. And you can argue about whether or not the methodology used by
UN Population Division . . . We spoke to them, I consulted them. I have very good colleagues in
there.
- 45 -
Most of their work is modelled work, is modelled projections, and so it is a question. It is not
so much a disbelief as much as whether or not the methodology of one, how the methodology of one,
Mr. MARTIN: I very much appreciate that, Dr. Guha-Sapir. And I apologize. But just because
our time is limited, if we could keep the conversation as focused as possible on the specific questions,
that would be helpful. Of course, I don’t want to stop you from adding anything you consider
So, the UN Population Division shows that the crude death rate in the DRC in 1997 was 1.386.
This is higher than the figure of 1.2 that you used; is it not?
Ms GUHA-SAPIR: Yes.
Mr. MARTIN: Are you also aware that the most recent UN Population Division data for the
years 1998 through 2003 indicates that in each and every one of those years, the crude death rate in
the DRC was actually lower than the crude death rate in 1997?
In that respect, I would refer the Court and you to paragraph 38 of Uganda’s observations on
the Experts’ Report. And if you have that in front of you, I would be happy to run through those
Ms GUHA-SAPIR: I have my own response, but I may have addressed paragraph 38 and ⎯
no, that is paragraph 38 of my own. I don’t have it in front of me in hard copy. So, if you really want
Mr. MARTIN: No, that is fine. As I say, Dr. Guha-Sapir, they are stated in our observations
at paragraph 38. But according to the United Nations data, the most recent United Nations data, the
crude death rate in 1998 — this is stated in per 1,000 individuals per month — is 1.376.
Ms GUHA-SAPIR: Yes.
Mr. MARTIN: In 1999: 1.357; 2000: 1.329; 2001: 1.294; 2002: 1.254; 2003: 1.212.
Now, all of these figures are substantially less than the 2.929 figure that you used; are they
not?
Mr. MARTIN: Okay. And since all of these figures are lower than the 1997 figure of 1.386 per
1,000 individuals per month, if — and I understand if you would not agree, but if you were to use
the most recent UN Population Division numbers in your calculation, the result would be that there
would be no detectible excess deaths in the DRC during the relevant period; right?
In your response to Ugandan’s observations on your report, you said the 1999 estimate of the
baseline crude death rate “more closely reflects the reality of the situation compared to projections
Ms GUHA-SAPIR: Yes.
Mr. MARTIN: Are you aware that in discussing its methodology, the UN Population Division
states that:
“The preparation of each new revision of the official population estimates and
projections of the United Nations involves the incorporation of new information about
the demography of each country or area of the world involving, in some cases, a
reassessment of the past.”
Are you familiar with that?
Ms GUHA-SAPIR: Yes.
- 47 -
Mr. MARTIN: Now, in some of your writings you have identified difficulties with trying to
Ms GUHA-SAPIR: Yes.
Mr. MARTIN: In particular, in this 2018 article in The British Medical Journal, of which you
were the lead author, you wrote that death estimates for the DRC are “highly controversial” and
Ms GUHA-SAPIR: Yes.
Mr. MARTIN: And in the same article, you emphasize the need for “verifiable body counts
from mortuaries and vital registration records” in producing reliable death estimates. Is that also
correct?
Ms GUHA-SAPIR: Yes, correct in the sense, yes, I have written that in that paper.
Mr. MARTIN: Thank you. As we discussed, you relied exclusively on retrospective sample
Ms GUHA-SAPIR: Yes.
Mr. MARTIN: Do you recall that in the same 2018 article I just mentioned, you stated that
sample surveys have “inevitably large confidence intervals and all the usual sampling biases that
Ms GUHA-SAPIR: Yes.
Mr. MARTIN: In any case, you relied on 38 surveys, I believe; 33 by the International Rescue
Ms GUHA-SAPIR: I don’t know whether that is ⎯ so, let us see. So there were 38 surveys of
which 20 were undertaken by the CDC, six by Les Roberts and the IRC, five by the IRC, five by
Michel Van Herp of MSF, and one by Ogle. That’s the distribution.
- 48 -
Mr. MARTIN: Thank you. So, in any event, you incorporated at least some of the IRC surveys
in your study?
Ms GUHA-SAPIR: Yes.
Mr. MARTIN: In a 2009 co-authored article in the journal Conflict and Health, you stated that
the IRC surveys in the DRC “have faced security and logistical obstacles that intensify the
methodological limitations in risk to personnel, hinder the implementation of best practices and
ensure that difficult choices must be made throughout the data collection period”; correct?
Ms GUHA-SAPIR: Yes.
Mr. MARTIN: Thank you. Are you also aware that Mr. Les Roberts, the lead investigator for
31 of the 33 IRC surveys, in a 2001 report said that there was “no follow-up or confirmation of the
information provided by interviewees”, and he acknowledged that “people may have lied to
interviewers”?
Mr. MARTIN: And are you aware that in another article co-authored by Mr. Roberts, that
article states:
“Survey implementation is often haphazard and fraught with biases, and surveys
conducted during complex humanitarian emergencies are prone to several
methodological limitations.”
Ms GUHA-SAPIR: Yes, I can see how Professor Roberts could have said that.
Mr. MARTIN: Given these problems, why do you use United Nations data for the baseline
crude death rate but a different source, sample surveys, for the posterior crude death rate?
which is a UNICEF estimation, and I depend on UNICEF to provide me with a reliable estimation.
For the conflict period mortality, I do not have any other sources except the 38 surveys that we have
used.
- 49 -
Now, this is the only source of data on mortality from the ground that is, in fact, available,
with its pros and its cons. And debate in the academic circles is always fraught with tensions and
stresses. That is the whole nature of debate. That is what happens in academics.
Nonetheless, the 38 surveys that have been used for this particular estimation, for the work for
this Court, have been probably the most ⎯ the most, I am not saying the reliable, the most reliable ⎯
sources of data from the ground on mortality that were available. And that is what I have done, and
I think that is the best possible estimation that one could have come up with at this time and with the
Mr. MARTIN: Thank you, Doctor. But data is also available from the United Nations
Population Division. And I think we have agreed if, and I understand that it is a very big if in your
mind, we were to use that data to perform the same analysis, the result would not show any detectible
Mr. MARTIN: Thank you. Now, in reaching your expert opinion, you made no determination
of any kind as to any deaths for which Uganda may be directly responsible, is that right?
Mr. MARTIN: In fact, in your response to the Parties’ observations, in paragraph 38, you said
“[m]y Terms of Reference do not ask me to assign indirect deaths to specific perpetrators”. Is that
correct?
Ms GUHA-SAPIR: Yes.
Mr. MARTIN: Thank you, Dr. Guha-Sapir. That is all the questions I have. I very much
The PRESIDENT: I thank Mr. Martin. And let me now check and see whether any of my
colleagues on the Bench have a question for this expert. I am not seeing any here. I am not seeing
Ms Guha-Sapir.
So when we talk about excess mortality, which Professor Sands noted has gotten more
attention in the popular press lately due to the COVID-19 pandemic, I understand we are trying to
appreciate how one specific event has changed mortality over time. And what I do not understand is
the baseline year was a year, and various things were taking place during that year. And then the
subsequent years also are individual years where various events occurred. So, for example, in the
baseline, there may or may not have been an epidemic, or locusts, or extreme weather, or a volcano,
and then in year two, year three, year four, there may or may not also have been those other variables.
So I am just trying to understand how this methodology takes into account the fact that, as I
understand it, the methodology attempts to indicate the excess mortality as a result of one cause, in
this case, the conflict, but in the meantime there are changes year over year of all sorts. So if you
could explain that, that would be very helpful to me. Thank you.
Ms GUHA-SAPIR: Yes, I think you have raised a very good point. And you are absolutely
right. What it is saying is how do we discount for everything else that is occurring that may have also
That is essentially what. And why is it that we are saying that the excess deaths are actually
entirely attributable to this one particular cause. I think that’s a valid question and I think that requires
a much more detailed analysis of whether or not there were other significant events in those years
which would have actually changed the mortality profile of the country in question, which needed to
What this is in the Congo, in particular, in these five years, is really an expression of the people
who have died in those five years that would not have died in the previous year. That is it. And so
after that to attribute it to a very particular cause is not possible, no. These are excess deaths. That is
- 51 -
all it is. These are these many people who have died who should not have died had the prevailing
So the prevailing baseline rates of 1997, if they were applied to the five years that followed,
these people should not have died. But you are absolutely right in the sense that it is very difficult to
attribute it to a particular cause that all of these deaths were actually attributable to a particular cause.
The PRESIDENT: Thank you. And now I have a question from my colleague,
The PRESIDENT: Judge Robinson, please. You have the floor. You will have to do something
The PRESIDENT: Yes, we can hear you now, Judge Robinson. Please go ahead. Thank you
Judge ROBINSON: And I too would wish to thank her for her patience and for the information
She stated in answer to Mr. Martin that the UN Population Division has modelled projections,
and so I wanted to ask her what does that mean and what impact, if any, the approach of model
Ms GUHA-SAPIR: So, in many cases, you know, this ⎯ I am not making pronouncement on
the entire UN Population Division and the demographic models that they use.
In general, for countries where statistical reporting is very weak ⎯ which is the case of the
Democratic Republic of Congo ⎯ most of the times, in my consultations with them in preparation
for this report, most of the times they essentially look at the past ten years and they put in some
adjustment factors and they project it for the next five years. What those adjustment factors are, I
cannot tell. They differ. They use factors that they have got from somewhere. I am not exactly
100 per cent sure what they are, but that is essentially what is done, because there is no statistical
- 52 -
reporting. Or they only get the results from civil registration, such as it is, remembering that the
statistical division of the United Nations has, in one of its papers, reported that about 13 per cent of
the total mortalities reported by civil registration. 13 per cent. So the majority of the deaths are not
But nonetheless, the UN Population Division may use the civil registration data to adjust their
figures from the past years to make projections for the next years. That is more or less the way I
Judge ROBINSON: Am I right in thinking that you do have some doubts about the soundness
of this method?
Ms GUHA-SAPIR: So, yes, I think ⎯ I do not know about the soundness. But I think the
UN Population Division’s projections, if you look at them, especially from countries where statistical
reporting is weak, they are more or less ⎯ they stay very stable. Why do they stay stable? It is
because they use the previous years and they adjust it a tiny little bit, of which they are absolutely
sure, and they project it to the next year. And I would say ⎯ being a field person myself and being
a field epidemiologist ⎯ I would be hesitant to use those figures for any kind of denominators or
Judge ROBINSON: Just one other more an observation or a request for a clarification than a
question.
Mr. Martin had you confirm that the estimates of deaths is based on mortuary reports and some
other factors that I cannot recall. But in your answer, I was observing your demeanour and I had the
impression that you might have wanted to say something else by way of answer, apart from
confirming what he asked. And I just wanted to find out whether you did want to say something else,
something additional, or whether that was, indeed, the totality of your answer.
- 53 -
Ms GUHA-SAPIR: Honourable Judge, to ask for mortuary reports from places like Ituri and
ingenuous, really. It really is. Not to mention unfair. And I think mortuary reports is not a realistic
scenario. There is no possibility. There is no institution in that region of the Congo that would even
register the death of a five-year-old child who could not be taken to the hospital because there was a
war going on. There was not even an institution. If you want the parents to carry that dead body of
the child to an institution to register her death, there is no incentive for the family, except taking the
risk of their own lives on the way to the hospital, to the institution. So to ask for death certificates
I cannot see ⎯ I cannot see the basis for such a . . . such a position, yes.
The PRESIDENT: I thank Judge Robinson for his question. I am not seeing any other
questions here in the Great Hall. I am not seeing any other questions from those who are joining
remotely.
So that then brings to the an end the questioning of Ms Guha-Sapir. I thank you very much
for appearing before us and for contributing to the report. You may now go ahead and sign out of the
WebEx platform and continue to follow the remainder of the hearing via webstream if you wish.
The PRESIDENT: Now I shall give the floor to Judge Tomka, who has a question directed to
en séance publique son ordonnance de réparation aux victimes en vertu de l’article 75 du Statut de
Rome dans l’affaire Le Procureur c. Bosco Ntaganda s’agissant de certains crimes commis en Ituri.
- 54 -
Pour rendre sa décision, la Chambre s’est appuyée notamment sur deux rapports d’experts. Ces
rapports ont pris en compte quelque 50 décisions rendues par les juridictions congolaises.
responsabilité pénale individuelle pour les crimes de guerre et les crimes contre l’humanité, les
Parties pourraient-elles indiquer quel est, selon elles, le poids à donner à l’approche suivie par la
Ntaganda ? Je fais tout particulièrement référence aux paragraphes 237 et suivants de l’ordonnance
de réparation.
On 8 March 2021, Trial Chamber VI of the International Criminal Court delivered in a public
sitting its Order on Reparations to victims under article 75 of the Rome Statute in the case of The
Prosecutor v. Bosco Ntaganda for certain crimes committed in Ituri. In coming to its decision, the
Chamber relied inter alia on two experts reports. These reports have taken into account some
Bearing in mind that the International Criminal Court deals with matters relating to individual
criminal responsibility for war crimes and crimes against humanity, could the Parties provide their
views as to the weight to be given to the approach adopted by Trial Chamber VI in the Ntaganda
case in relation to the valuation of the harm suffered? I refer them particularly to paragraphs 237 and
The PRESIDENT: I thank Judge Tomka. The written text of this question will be
communicated to the Parties as soon as possible. The Parties are requested to respond orally to the
question during the second round of oral pleadings. The Court will meet again on Monday
26 April 2021 at 3 p.m. for the second and final session of the questioning of Court-appointed
___________