1 Monday, 24 November 2008
2 [Open session]
3 [The accused entered court]
4 [The Accused Pusic not present in court]
5 --- Upon commencing at 2.18 p.m.
6 JUDGE ANTONETTI: [Interpretation] Registrar, kindly call the
7 case, please.
8 THE REGISTRAR: Good afternoon, Your Honours. Good afternoon
9 everyone in and around are the courtroom. This is case number
10 IT-04-74-T, the Prosecution versus Prlic et al., thank you, Your Honours.
11 JUDGE ANTONETTI: [Interpretation] [No interpretation] Mr. Pusic
12 is probably ill. I hope he gets better quickly. I would like to greet
13 the counsel, Mr. Stringer, as well as all his associates, women
14 associates since there are no men associates. I would like to greet them
15 all. The registrar and the usher as well as all the people assisting us
16 in the courtroom. Registrar has a few IC numbers to give us. This is
17 today the 24th of November. I would like to greet everyone starting off
18 with all the accused, interpreter's note.
19 THE REGISTRAR: Thank you, Your Honour. Some parties have
20 submitted lists of documents to be tendered through witness Tomic, Neven.
21 The list submitted by 1D shall be given Exhibit number IC 00889. The
22 list submitted by 2D shall be given Exhibit number IC 00890; the list
23 submitted by 3D shall be given Exhibit number IC 00891; the list
24 submitted by 4D shall be given Exhibit number IC 00892; and the list
25 submitted by the Prosecution shall be given Exhibit number IC 00893.
1 Thank you, Your Honours.
2 JUDGE ANTONETTI: [Interpretation] Thank you, registrar.
3 Last week the Trial Chamber handed down a decision relating to
4 the expert witness who is coming to testify today. The
5 examination-in-chief has been set to two hours and cross-examination by
6 other Defence counsel to -- I believe it's Mr. Praljak and Mr. Petkovic's
7 Defence counsel. In light of the scope of this testimony, has set the
8 cross-examination time to three hours. This witness is coming to testify
9 after Ewa Tabeau, whom we have already heard.
10 I see that Mr. Karnavas is ready. He has given us his file with
11 all the documents that will be presented. There are ten or so documents.
12 We shall bring the witness into the courtroom.
13 [The witness entered court]
14 WITNESS: SVETLANA RADOVANOVIC
15 [Witness answered through interpreter]
16 JUDGE ANTONETTI: [Interpretation] Good afternoon.
17 THE WITNESS: [Interpretation] Good afternoon.
18 JUDGE ANTONETTI: [Interpretation] Can you give us your first
19 name, last name, and date of birth, please.
20 THE WITNESS: [Interpretation] Svetlana Radovanovic, 14 November
22 JUDGE ANTONETTI: [Interpretation] What is your current
24 THE WITNESS: [Interpretation] I'm a full-time professor at
1 JUDGE ANTONETTI: [Interpretation] [Previous translation
2 continues] ... university.
3 THE WITNESS: [Interpretation] I have not received the
4 interpretation. The school of geography of the university in Belgrade.
5 JUDGE ANTONETTI: [Interpretation] Ma'am, have you already
6 testified before a court of law on those events that unfolded in the
7 former Yugoslavia or is it the first time you come to testify today?
8 THE WITNESS: [Interpretation] I've already testified.
9 JUDGE ANTONETTI: [Interpretation] You testified before which
10 tribunal and in which case?
11 THE WITNESS: [Interpretation] I've testified at this tribunal on
12 four cases: The Galic case; the Blagojevic case; the Beara, Nikolic, and
13 others case; and the Simic case.
14 JUDGE ANTONETTI: [Interpretation] Very well. So you have
15 testified in four cases. Were you a Prosecution witness or a Defence
17 THE WITNESS: [Interpretation] Defence witness. And I also
18 testified at the court in Sarajevo.
19 JUDGE ANTONETTI: [Interpretation] You testified in Sarajevo in
20 which case?
21 THE WITNESS: [Interpretation] I wouldn't be able to give you the
22 number of the case or the name of the accused. I could only give you the
23 name of the counsel, and the case had to do with genocide and Srebrenica.
24 I -- to be more specific, it was the Kravica case.
25 JUDGE ANTONETTI: [Interpretation] Very well. And were you a
1 Prosecution witness or a Defence witness?
2 THE WITNESS: [Interpretation] A Defence witness.
3 JUDGE ANTONETTI: [Interpretation] I would like you to read the
4 solemn declaration now.
5 THE WITNESS: [Interpretation] I solemnly declare that I will
6 speak the truth, the whole truth, and nothing but the truth.
7 JUDGE ANTONETTI: [Interpretation] Thank you, ma'am. You may sit
9 THE WITNESS: [Interpretation] Thank you.
10 JUDGE ANTONETTI: [Interpretation] Just some information I would
11 like to share with you so that this hearing unfolds in the best
12 conditions possible. Since this is the fifth time that you are
13 testifying before this tribunal, you know how things work out. You will
14 first answer questions that will be put to you by Mr. Karnavas, based
15 undoubtedly on the documents he will show you. After that, other -- the
16 other Defence counsel representing the other accused, since Mr. Karnavas
17 defends Mr. Prlic, well, the other Defence counsel can also put questions
18 to you as part of the cross-examination.
19 The Prosecutor, who is on your right, Mr. Stringer, will
20 cross-examine you when the time comes, when he has the floor.
21 The four Judges that make up the Bench will put questions to you.
22 This might be different from what happened in the other cases you
23 testified in. In this case the Bench does ask a lot of questions. This
24 is what you may feel is a bit different from what you have experienced in
25 the other testimonies you gave.
1 Please be as accurate as you can. Since you are a university
2 professor I'm not concerned about this at all, and I'm sure you'll be
3 able to sum up your answers in the best possible way. If, however, you
4 do not understand a question, and even if it is a Judge that puts a
5 question to you, sometimes the Judges don't put the questions as they
6 should, in that case do not hesitate and ask the person who put the
7 question to you to rephrase it, please.
8 We have a break every hour and a half, but if at some point in
9 time you don't feel comfortable or if you need a break, do not hesitate
10 to raise your hand and ask us to have a break. We are quite prepared to
11 do this.
12 This is what I wanted to share with you, ma'am, so that this
13 hearing can unfold in the best way possible in the interests of all of us
14 and in the interests of justice.
15 I'm very happy to give you the floor, Mr. Karnavas.
16 MR. STRINGER: [Previous translation continues] ... just to
17 inform the Trial Chamber and the witness, the cross-examination is goes
18 to be conducted by my colleague, Ms. West.
19 JUDGE ANTONETTI: [Interpretation] Very well. Thank you. We will
20 hear Ms. West with great pleasure also. We have already had the
21 opportunity to hear you. We regret, however, Mr. Stringer, that you are
22 not conducting the cross-examination yourself, but I'm sure we will have
23 the pleasure to hear you at a later stage.
24 Mr. Karnavas, you have the floor.
25 MR. KARNAVAS: Good afternoon, Mr. President; good afternoon,
1 Your Honours, good afternoon to everyone in and around the courtroom.
2 Examination by Mr. Karnavas:
3 Q. And good afternoon, Professor Radovanovic.
4 MR. KARNAVAS: Before I begin my first question, Your Honours,
5 there was a technical matter that was brought to my attention just prior
6 to coming into the courtroom, and that is Professor Radovanovic's report,
7 which is 1D 03110, was not added to the 65 ter list, and therefore at
8 this time I would be asking permission that it be added to the list so
9 then I could make reference to it during my direct examination.
10 JUDGE ANTONETTI: [Interpretation] This had been noticed by the
11 Bench. I shall confer with my colleagues, but I'm sure it's not a
13 [Trial Chamber confers]
14 JUDGE ANTONETTI: [Interpretation] The Trial Chamber fully agrees
15 to this. It is quite natural that this report be added onto your list.
16 MR. KARNAVAS: Thank you, Mr. President; thank you, Your Honours.
17 Q. Professor Radovanovic, in your report which we have, which is
18 1D 03110, attached to it there is an annex which has your CV, your
19 curriculum vitae. Is that an accurate one, madam?
20 A. Yes.
21 Q. Now, I don't want to go through all of your educational
22 background because we can all read it, but suffice it to say you hold a
23 Ph.D. from the faculty of geography, and as I understand it your
24 speciality is demographics; is that correct?
25 A. Yes. I do hold a Ph.D. in geography.
1 Q. So let's just briefly focus a little bit on your work experience.
2 If you could be so kind as to give us a brief overview of your work
3 experience specifically related to the field of demographics, because
4 after all that is what we are all interested in.
5 A. Since 1974 -- or, rather, from 1974 to 1992 I worked in the
6 National Bureau of Statistics of Serbia. As part of my job I worked
7 exclusively in the department for population statistics. In 1986 I
8 became the head of that department, and I participated in three
9 population censuses. The first one was in 1971, and then I participated
10 in the preparation, realisation, and processing of the 1981 census, and
11 in that, in 1981, I was a member of the expert team of the then SFRY for
12 the methodology of the census. In 1991 I also participated in the census
13 as a member of the expert team for methodology. The expert team was
14 composed of the representatives of all former states of Yugoslavia, and I
15 was also in charge of the population census of Serbia. I would like to
16 emphasise that the methodology --
17 Q. Let me stop you here. Two points. One, you may need to speak a
18 little bit slower, and, two, you may need to speak a little bit louder.
19 That may be helpful. Slower and louder. Thank you. And my apologies
20 for interrupting you.
21 A. All population censuses that were conducted in the former
22 Yugoslavia were conducted according to a uniform methodology which was
23 prescribed by the Federal Bureau of Statistics. In addition to the
24 census statistics I was also involved and I headed the services for vital
25 statistics, which means the services that were concerned with natality,
1 mortality, marriages, and from 1982 I was also the head of the service --
2 Q. You need to slow down. Otherwise, they can't -- you're going to
3 wear them out. So please slow down. Keep in mind that this is
4 simultaneous. Thank you.
5 A. I was part of the team that introduced the service for migration
7 In 1992, I joined the institute of social sciences, the centre
8 for demographic research, and that's where I worked on scientific
9 projects that this centre was involved in. They were funded by the
10 Ministry of Science and other institutions.
11 I also headed two projects myself. I also participated in the
12 following projects: The one of the European Centre for Peace and
13 Development of the United Nations that studied the population of one part
14 of Serbia known as the Raska Oblast region or Sandzak with a prevalent
15 Muslim population.
16 In 1992 -- or, rather, 1999, I apologise, I joined the faculty of
17 geography, and since then I've been a member of the faculty of geography.
18 I was the head of department for demographics, and for a while I was the
19 deputy dean for finances of the school of geography.
20 I teach three modules to regular students. One is introduction
21 to demographics, population statistics, and ethnodemographics.
22 Q. All right. Thank you.
23 A. I'm also --
24 Q. I'm sorry. I thought you were finished.
25 A. I also participated in a number of scientific and research
1 projects that were conducted and are still conducted at the school of
2 geography, and I have also participated in many projects that are
3 conducted by the Serbian Academy of Arts and Sciences under the
4 leadership of the Serbian Academy of Arts and Sciences. I've organised
5 and prepared the methodology for the interim census of the population of
6 one part of Serbia known as Silic [phoen] county, and that is actually in
7 the province of Kosovo and Metohija. I'm a member of the committee on
8 population of the Serbian Academy of Arts and Sciences and several other
9 associations such as the Association of Geography, the Statistical
10 Association, the Demographics Association and so on and so forth.
11 Q. Okay. Thank you. I think we have a pretty comprehensive
12 overview of your -- of your background.
13 Now, if I could turn your attention to this document, which is 1D
14 03110. That's the very first document in your binder. Do you recognise
15 this document, madam?
16 A. Yes.
17 Q. And what is it?
18 JUDGE ANTONETTI: [Interpretation] One moment. One moment.
19 Mr. Karnavas, I have a follow-up question to put concerning this
20 person's -- this witness's resume.
21 I listened carefully to everything you have done. With your
22 Croatian or Muslim counterparts did you ever work together on census
23 projects as part of a cooperation programme with all the other republics,
24 or did you only work on data concerning the Republic of Serbia?
25 THE WITNESS: [Interpretation] We worked together. The
1 methodology that was established by the Federal Bureau of Statistics was
2 our joint project which involved all the demographic statistics experts
3 which implies census and vital statistics from various republics. Before
4 you come up with an organisation and before you come up with a
5 methodology, first you have to organise a group of experts that will be
6 involved in the drafting of a methodology for the population census.
7 Then that methodology for population census is published by the Federal
8 Bureau of Statistics, and it becomes the only one and uniform for all the
9 republics of the former Yugoslavia, and then that methodology was applied
10 across the board. All of us had to carry out our censuses according to
11 that methodology.
12 I have to emphasise that the republics had the right to add some
13 questions that they deemed important or interesting for their particular
14 regions. The additional questions could differ, i.e., could only apply
15 to one republic. Everything else was applied across -- across the board
16 and had to be implemented in the same way, and this did not apply only to
17 the methodology. It also applied to the organisation, the processing,
18 the tabling programme and the programme of data publication.
19 JUDGE ANTONETTI: [Interpretation] Does this mean that today in
20 all the existing republics the methodology is exactly the same, nothing's
22 THE WITNESS: [Interpretation] According to the information that I
23 have from the republics, for example what happened in Croatia in 2001, as
24 well as in Slovenia and Serbia, most of it is the same methodology.
25 There are a number of questions which differ somewhat in Croatia from
1 what has been done in Serbia. However, the key elements of the
2 methodology have not changed, but I would like to add to that that these
3 are very complex things, because a methodology involves a lot of things.
4 It involves administrative and territorial division. Administrative and
5 territorial division has not remained the same in Croatia. It is not the
6 same in -- it was not the same in 2002 as it was in 2001. However, the
7 instruments, such as nomenclatures for the coding of ethnic groups, for
8 establishing various circles, professions, these have been changed.
9 However, the essence of the methodology is still in place. It is still
11 JUDGE ANTONETTI: [Interpretation] Very well. Thank you for your
12 very comprehensive answer.
13 Mr. Karnavas.
14 MR. KARNAVAS: Thank you, Mr. President.
15 Q. Now, looking at this document 1D 03110, this is your report, is
16 that right?
17 A. Yes.
18 Q. And could you please tell us how was this report created. Who
19 was involved in it? What was done in order to create it?
20 A. You contacted me, and you asked me whether I would be interested
21 in participating in that and whether I would like to look at the OTP
22 experts' report that was drafted for this case, and then you sent me the
23 report -- i.e., reports. We're talking about three reports, not just
24 one. I read the reports, and then we spoke again. I gave you my
25 opinion. I provided you my objections to those reports, and then you
1 instructed me that I could peruse the original materials that were used
2 by the OTP experts. You allowed me to come and inspect the materials,
3 and you gave me free hand to inspect any other source of data in drafting
4 my report for this case.
5 Q. All right. And as a result of your work, how many reports did
6 you produce? I know this is one, but there are how many parts to this
7 report? What does it address?
8 A. I produced a report in two parts. I could have come up with
9 three reports. However, a report on the wounded is something that I did
10 not produce because it presents the same problems as the previous two
11 reports. I studied that third report very carefully, and I'm absolutely
12 able to answer all your questions with this regard.
13 Q. All right. Now, you said that you came here to The Hague to look
14 at the data. Could you please tell us how long were you in The Hague and
15 with whom and what exactly you did?
16 A. With two assistants of mine, and they are IT experts, I spent
17 between 14 and 18 April in The Hague in the demographics department of
18 the OTP. I had at my disposal the main sources of data that had been
19 used by the OTP expert and some other sources of data that I wanted to
20 look at. I made sure that I was well informed with the contents of these
21 sources and to draw my conclusions from those sources. I obviously could
22 not peruse any other by the -- but the main sources of data. I could not
23 dwell upon any of them in any more detail. However, I leafed through
24 them in order to be able to spot the essence, to see how significant they
25 were, why they were applied and how, and so on and so forth.
1 From that exercise I was able to draw my own results, and in
2 addition to that, I obtained most of the expert reports, particularly
3 those that were relative to the ethnic structure. These were done by
4 Dr. Ewa Tabeau. I perused those. And I also obtained a number of
5 official publications that were published by the Federal Bureau of
6 Statistics of Bosnia and Herzegovina. These are mostly statistics
7 publications. And I have forgotten to say that I also obtained some
8 publications that were published by the ministry for refugees and
9 displaced persons.
10 I have listed all the references and sources that I used in my
12 Q. All right. We won't dwell on your report to any degree because
13 we can all read it, but I will be asking you questions that relate to
14 your report, but first it might be good just for housekeeping purposes to
15 go through some of the other documents that are in the folder that
16 apparently you used in addition to what was provided to you and what you
17 looked at from the OTP list. If we could look at 1D 03099. That's the
18 second document. If you could please tell us whether you recognise this
19 document and, if so, what is it is and why is it relevant? It's the
20 second document.
21 A. 1D 030 -- I have as the second document 1D 03099.
22 Q. That's correct. 1D 03099.
23 A. Yes, I do have that document.
24 Q. Okay. Would you please tell us what is this document then?
25 A. Well, it's a publication issued by the Federal Institute for
1 Statistics in Sarajevo, and it presents data canton by canton. There are
2 separate sets of statistical data for each canton from several aspects of
3 the statistics. What I was interested in was the data about the
4 population, in particular the table of the ethnic breakdown. So there is
5 the assessment of the overall number for the population with the ethnic
6 breakdown. And I used several brochures put out for various cantons,
7 because the municipalities that the Prosecution expert is dealing with
8 can be found in various cantons.
9 Q. All right. Well, why didn't you obtain this document, or did you
10 try to obtain this document from Ms. Ewa Tabeau while you were here in
11 The Hague examining the other material?
12 A. Well, I did write a request to you indicating what I would like
13 to have at my disposal when I come to the demographic unit. These
14 publications were listed in the request. However, I did not get those
15 publications, and I concluded that they did not have those publications
16 at their disposal.
17 Q. All right. 1D 03100, the next document. If you could look at it
18 and please tell us if you recognise it and, if so, why it's significant.
19 A. This is a publication from the ministry for human rights and
20 refugees, the sector for refugees of Bosnia and Herzegovina, and it
21 provides information about the situation in the field of the return of
22 the refugees from Bosnia and Herzegovina, displaced persons, and so on
23 and so forth. It was published in Sarajevo in 2004.
24 Q. Is this an official document or unofficial?
25 A. It's an official document.
1 Q. All right. If we look at the next document, 1D 03101. Do you
2 recognise this document and, if so, what it is and why it was relevant.
3 A. Well, it's one of the brochures published by the Federal
4 Institute for Statistics that I was talking about, and it pertains to the
5 cantons in Bosnia and Herzegovina. This quite specifically is the
6 Central Bosnia canton, and the title is "Central Bosnia canton in
7 figures," and the reason why I used it is the table for the assessment of
8 the number of the population with the ethnic background and because two
9 of the municipalities -- or rather just one of the municipalities that
10 are relevant that was analysed by the Prosecution expert, Gornji Vakuf,
11 is in this canton.
12 Q. If we go on to the next document 1D 03102. If you could tell us
13 what it is and why this is relevant.
14 A. This is again an official document from the Ministry for Human
15 Rights and Refugees. It was published in Sarajevo in 2003, and as you
16 can see, the title is "Bulletin 2003," and it contains comparative data
17 about refugees, displaced persons, and returnees, and so on.
18 Q. All right. But why was it necessary for you to look at this
19 particular document?
20 A. Well, this document speaks about the refugees and displaced
21 persons and provides information municipality by municipality. And since
22 the data pertained to the same census of the refugees and we have the
23 database in the demographic unit of the OTP, I consider this to be quite
24 important because it enabled me to follow this municipality by
25 municipality in terms of refugees, displaced persons, and returnees, and
1 so on. And I consider this to be the officially published data that are
2 backed by the State of Bosnia and Herzegovina and its ministry.
3 Q. Having reviewed Ewa Tabeau's reports, do you know whether she
4 makes reference to this particular bulletin that you indicate is, one,
5 official and, two, is relevant, particularly to the period of time that
6 we're interested in?
7 A. I did not notice her mention it anywhere at all.
8 Q. All right. If we look at 1D 03103. If you could please tell us
9 whether you recognise this document and, if so, what it is and why it's
11 A. Well, again this is the brochure for the Western Herzegovina
12 canton, the canton in figures. It is again important because it provides
13 the assessment of the number of the population by ethnic group, and there
14 are municipalities here that were necessary for -- in order for me to be
15 able to come to the total for the eight municipalities that Dr. Tabeau
17 Q. All right. Now, prior to coming here to The Hague we had made a
18 request for you to send some material by Federal Express at the request
19 of the OTP. Are these some of the documents that the Office of the
20 Prosecution was asking to look at, or were there other documents
21 unrelated to the ones that we've seen?
22 A. I'm not sure that I understood your question.
23 Q. Okay.
24 A. If I understood it right --
25 Q. Let me ask the question again to save time. Prior to arriving
1 here I had made a request for you to Federal Express certain material
2 that was being requested by the Office of the Prosecution, material you
3 had used and made reference to in your report. Do you recall that?
4 A. Of course I do recall that. I was very much opposed to this
5 request that you made because you asked me to send materials that had
6 been officially published, and I replied to you twice that I was not
7 asking the OTP to send the officially published materials to me, but then
8 when you raised your voice I sent those materials to you.
9 Q. All right. And are these some of the materials that were being
10 requested by the Office of the Prosecution, because one would assume that
11 if they had them, they wouldn't be making the request.
12 A. Well, I assume that the OTP does not have it since they're asking
13 me to provide them with the officially published materials.
14 Q. I'm going to ask you to listen to my question, and my question is
15 very precise. Are these the material or some of the material asked?
16 A. Yes.
17 Q. Okay. Thank you.
18 A. Not some but all of the materials that the OTP asked for, plus
19 the instructions on the voters, which is not here, but I also sent it to
20 the OTP because I used it in the drafting of my report.
21 Q. Okay. When you say the instructions on the voters, is that
22 something that was used during the election to which Ewa Tabeau makes
23 reference to or uses statistics from? Are we talking about that
25 A. We're talking about the instructions for the registration of
1 voters in 1997.
2 Q. Okay.
3 A. And Dr. Tabeau has the list of registered voters. I don't know
4 whether she has also the instructions.
5 Q. All right. In your -- in your field of work would it have been
6 useful, for instance, for Dr. Tabeau or this demographics department to
7 have had that sort of information all of these years since they've been
8 using that voter registration in virtually all of their reports?
9 MS. WEST: Objection, Your Honour.
10 MR. KARNAVAS: What's the grounds?
11 JUDGE ANTONETTI: [Interpretation] Ms. West.
12 MS. WEST: Thank you, Your Honour. Good afternoon,
13 Mr. President, the rest of Your Honours and everyone in and around the
14 courtroom. The question posed by Mr. Karnavas suggests the answer within
15 it and it's leading.
16 MR. KARNAVAS: Your Honour, we're dealing with an expert. The
17 question is would an expert rely on this sort of information. If the
18 answer is yes, then it poses the question at some other point how on
19 earth has Dr. Tabeau and the OTP demographics department been using these
20 statistics all these years without getting the instructions and why it
21 that Karnavas and his expert would now in 2008 have to provide that
22 information. If the answer's no, I move on, although again if it wasn't
23 relevant, then why would they ask for it?
24 JUDGE ANTONETTI: Well, madam, could you please answer the
1 THE WITNESS: [Interpretation] Could you please repeat the
2 question. I got a little bit lost.
3 MR. KARNAVAS:
4 Q. The question is: These instructions, is this something that
5 would be -- is this material that would be relevant to a demographer such
6 as Ewa Tabeau, especially since she's making reference to the electoral
7 results or those elections in various reports throughout the years?
8 A. Yes. Every instruction is an explanation of a part of the
9 methodology. This instruction tells us how the voters lists were made,
10 how voters were registered.
11 Q. All right.
12 A. Who can participate in the exercise, how to register, where to do
13 it, how to register the location, where one wants to vote.
14 Q. Very well. If I could now get you to look at -- just glance at
15 the next two documents. One is P 09836 and P 09837. Those are the last
16 two documents in your binder, and I am going to ask you a leading
17 question, which is: Are these the reports by Ewa Tabeau that you looked
18 at and analysed and wrote your report on?
19 A. Yes.
20 Q. All right. Now, before we discuss these reports, if you could
21 very briefly, recognising that the field of demography is very complex
22 and you're a professor, so we want a short, concise answer, not something
23 long-winded, if you could just please tell us a little bit about
24 demography, what it is and what are the tools that a demographer would
25 use in carrying out their -- their work?
1 A. A demographer has to have good quality sources for the data, must
2 be able to classify the data and also to process the relevant statistical
3 and demographic indicators and must be able to analyse the data and to
4 reach some causal conclusions. The instruments that you use are what is
5 called methodology.
6 Q. All right. Now, before we speak about methodology, if you
7 could -- and I'll try not to ask compound questions, especially if they
8 are lengthy because then you may be answering only one part of the
9 question. What is demography essentially?
10 A. To cut a long story short, it's the science about population. It
11 studies the size, the structure, the natural and territorial migrations
12 or changes. It is a multi-disciplinary science, because in addition to
13 its own methodology, it borrows methodologies from other sciences close
14 to it, and just as it lets other sciences use its methodology and its
16 Q. All right. Now, if you could go -- now we can talk a little bit
17 about what you indicated, methodology. What do you mean by that, so that
18 way we can identify what the terms are and use them consistently
19 throughout the afternoon.
20 A. Well, methodology is a complex set of procedures that you have to
21 device and implement if you want to research a phenomenon, whether it be
22 population or anything else in some other science. And this complex
23 system of steps, as far as demography is concerned, can be divided into
24 three steps, broadly speaking.
25 The first step is the selection of the sources for the data, and
1 this first step is actually the key step, the most significant one,
2 because if you don't select your data sources, if you don't assess them
3 and evaluate them properly, then your methodology may be perfect and you
4 may be fully versed in various techniques of statistics, if the first
5 step is not done properly, all the other steps that follow simply make no
7 The next step, the second step, is precisely this one: From your
8 data sources you extract the data. You group them based on some of your
9 features, the features that you're interested in, that you want to study,
10 and then you do your calculations in order to obtain statistical
11 indicators, statistical and demographic indicators, and then on the basis
12 of those you reach statistical and demographic conclusions which are part
13 of the analysis that you're conducting. And as you analyse this, you
14 establish things. You comment on certain phenomena that you encounter.
15 And, if possible, you apply the causal principle to see what the causes
16 are, what the consequences are, and so on.
17 Q. All right. And these three general or major steps that you've
18 indicated to us as part of the methodological processing in demographics,
19 are they recognised within the field of demography?
20 A. Yes.
21 Q. All right. Now if we can start with the first one, and I'm going
22 to ask you to be again short, concise, and very concrete. You indicated
23 that sources, that the sources are important. What do you mean by
24 sources? Could you give us an example, a definition and an example?
25 A. A data source is something that a demographer uses. It is the
1 basic -- the basis that you use in order to be able to produce some
2 results. A data source is not something that is constructed by a
3 demographer. A demographer usually knows where to go to get the best
4 information about the population. Various statistics regarding the
5 population, vital statistics, migration statistics. You can also study
6 various monographs and other documents. So a demographer should be able
7 to assess, to evaluate a data source that they intend to use.
8 Q. All right. And could you give us an example of a source? In
9 this particular case what sort of sources were used, say, by
10 Ms. Ewa Tabeau?
11 A. Dr. Tabeau uses the 1991 census in Bosnia and Herzegovina. She
12 used the voters list or electoral rolls from 1997 and 1998, and she uses
13 the database on the displaced persons and refugees.
14 Q. All right. Now, we'll discuss that in greater detail, but I just
15 want to go through the steps that you've laid out for us rather nicely,
16 and as I understand it also, and I believe you said this was the very
17 first step, you also used the word "key." Can you please explain what
18 you meant by that?
19 A. Well, that is the manner in which the OTP expert based on three
20 data sources that are incongruous in terms of time resolved some issues,
21 some problems she encountered in order to obtain her results. This is a
22 methodological key that is called the matching, if that's what you meant.
23 But if you want to say not the key but the key issue in methodology, the
24 key issue in methodology is the selection and the proper evaluation and
25 assessment of the data source. That's the key issue, because if you have
1 a bad data source, you cannot have good results no matter your expertise.
2 Demography and statistics of population have various keys, but there are
3 also key issues in methodology.
4 Q. All right. If we go to the next step. You said the next step is
5 grouping and extracting of data. How is that done? Can you give us a
6 brief overview?
7 A. Technically it is irrelevant what statistician or demographer
8 we're talking about. The extraction of data, that means that you have to
9 have a data source which enables you to establish the total number of a
10 certain element, and then you want to classify this according to some
11 features and analyse it. You can do that for each data source, and
12 that's how it's actually done if your data sources are incongruous. But
13 a different principle was applied here. Incongruous data sources were
14 used --
15 Q. Okay.
16 A. -- and on their basis the method that was used is well-known in
17 statistics and in everyday life. Attempts were made to extract and group
18 data from incongruous data sources as a product of a combination of those
19 data sources.
20 Q. All right. And again we're going to get to those reports. Right
21 now I just want to get the general so we have a general foundation, so
22 when we get to the reports we'll understand a little bit more precisely
23 what it is you're saying.
24 You did use the word "matching," and if you could please tell us
25 what you meant by that.
1 A. Matching or the identification method or the matching method is a
2 well-known method used not only in statistics and demography but in other
3 sciences. It is in fact used in everyday life. It boils down to this:
4 Through different data sources you can identify a single statistical
5 unit. In this person -- in this case we're talking about a person. A
6 very important thing when it comes to the matching method is to know what
7 the matching key is. What is it? It is a series of elements that --
8 something that we will use to identify a person from one source in
9 another source. For instance, the personal ID number, the unique
10 personal ID number, is an excellent matching key provided it is correct.
11 For a very simple reason, because it is unique, as its name tell us, and
12 it contains elements that will guarantee that if you match it, if you
13 match that key in two data sources, you will have the same person.
14 That matching key has 13 elements. The first seven elements are
15 the numbers indicating the day, the month, and the year of birth. The
16 following two digits indicate the state or the republic and the area
17 where the unique ID number was issued. The next three digits indicate
18 whether it is a male or a female. And the last digit is the control
19 digit, which makes it possible to prevent the duplication of the unique
20 personal ID number. It renders it impossible to have two persons have
21 the same ID number.
22 This is the most precise matching key that you can have, but the
23 sources that the Prosecution expert used did not have those ID numbers.
24 There are ID numbers in voters lists and in the census numbers, but they
25 are not correct. A large number of errors was noted, up to 40 per cent
1 in fact.
2 Q. All right. And again I must caution you, we're going to get that
3 particularly, but I first want to lay the foundation.
4 Let's go to the third step, which is the analysis, and in that
5 stage you indicated that you tried to draw some conclusions, determine
6 some cause and effects. So could you please explain that a little bit to
8 A. An analysis is the crown of any expertise, because it draws
9 certain conclusions as to what happened. It ties causes to effects and
10 explains what has been done. However, for any analysis to be good it has
11 to contain analytical tables.
12 Q. All right. You're not finished, go ahead and finish.
13 A. There is virtually no analytical tables in Ms. Tabeau's reports,
14 which speaks of the scarcity of the sources that she used.
15 Q. All right. Let's talk about those reports, and I won't go
16 through them in any great detail, but the first report, for instance,
17 that --
18 JUDGE ANTONETTI: [Interpretation] One moment. There's a
20 JUDGE MINDUA: [Interpretation] Mr. Karnavas, I'm sorry for
21 interrupting. There's something extremely important here.
22 Professor Radovanovic, in order to do the matching, you need to
23 have adequate sources, and you mentioned ID numbers. I heard this in
24 English. I don't know how it was translated into French. And you are
25 saying that the ID numbers which the Prosecution expert used were
1 erroneous. This is what I understood. And your ID numbers were correct.
2 Can you confirm this, please?
3 THE WITNESS: [Interpretation] I may have not explained it
4 properly and that's why you did not quite understand. The OTP expert
5 never mentioned having used ID -- citizens' numbers. I also did not use
6 them. I only said that this is the best identification key. However,
7 the problem of the sources used by the OTP expert lies in the fact that
8 there is a huge number of mistakes among the IDs. Their percentage is
9 over 40 per cent.
10 JUDGE MINDUA: [Interpretation] Thank you very much.
11 JUDGE ANTONETTI: [Interpretation] Mrs. Radovanovic, if I have
12 understood you correctly, in the former Yugoslavia, every citizen had a
13 registration number. That's what you call a JMBG. If I have understood
14 you correctly, this is a source which needs to be analysed when involved
15 in demography. From what I understood, Mrs. Tabeau did not use these
16 figures, this data. Is this right?
17 THE WITNESS: [Interpretation] Again I have not explained
18 properly. The fact is that every citizen of the former Yugoslavia did
19 have the so-called JMBG, because in the 1980s a law had been passed on
20 the JMBG. The JMBG was first collected in the 1991 census, and the
21 underlying idea was for the former Yugoslavia to come up or to compile
22 its population register. However, this unique personal ID number was not
23 collected in a way that might be conducive to that. Over 40 per cent of
24 people in towns and villages did not provide their ID number. That is
25 why in the census data there is no ID for every person, nor is it
1 consider data. It was not used for anything. It was never checked as
3 When I was speaking about the matching method, I tried to explain
4 and say that if the ID numbers were good and correct, then the matching
5 method would guarantee reliable data drawn from the sources that do
6 contain good IDs. The sources that were used for the analysis did not
7 contain or did not show the satisfactory quality of the IDs in them.
8 JUDGE ANTONETTI: [Interpretation] So if I have understood
9 correctly, in 1991 when there is this census, the authorities hand out, I
10 assume, a document to everybody, and this document should mention the
11 acronym JMBG. The -- each citizen then hands in his filled-in document
12 to the appropriate department. If I have understood you correctly, 40
13 per cent of the people at the time did not fill in the document properly
14 or did not give in the document at all. Is that right?
15 THE WITNESS: [Interpretation] No. The ID number is given to
16 every child who was born after the 1980s. When the birth certificate is
17 issued to a child, it contains the child's ID. People who were born
18 before the 1980s, before the ID was first introduced, they could not
19 obtain any documents such as passport before presenting at the police and
20 be provided with an ID.
21 In 1991, the statistics wanted to start preparations for the
22 introduction of the population registry. The preparation for the
23 population registry would entail the collection of IDs from the
24 population. And we did something that is not normally done in any
25 census. They asked for -- from the interviewer, because you have to know
1 that the censuses in the former Yugoslavia used the method of interview,
2 which means that I as the interviewer would come to your household as a
3 trained interviewer privy to the methodology, and I ask you to give me
4 your data, and I also ask you do you have an ID, could you give me your
5 personal ID documents so I could copy it -- from it? So it is the
6 interviewer that enters the ID. Many citizens did not want to provide
7 their IDs during census because they're not duty-bound to do so. I for
8 one don't do that. I just give them the first seven numbers which
9 reflect my date of birth. Some citizens are willing to provide their
10 full ID; some are not. When the interviewers come, some citizens are not
11 even there, so they could cannot provide their IDs. Some interviewers
12 are not conscious and careful and diligent and they enter erroneous IDs.
13 There was not an obligation. There is a law on census which
14 provides exactly what needs to be collected and the citizens comply, but
15 this was not one of those obligations.
16 When all the data of the census was collected it was established
17 in the statistics that ID numbers were collected from some 50 to 60 per
18 cent in their complete form. As for the rest, either these IDs were
19 missing the six last digits or they were erroneously recorded.
20 MR. KARNAVAS: Thank you.
21 JUDGE ANTONETTI: [Interpretation] Mr. Karnavas.
22 MR. KARNAVAS:
23 Q. Now, since we're speaking about the census in -- we got way ahead
24 of ourselves a little bit, if you could just tell us, in the census there
25 were also names. Now -- were there not? Names were collected. I'm
1 talking about the census in Bosnia-Herzegovina in 1991.
2 A. Names are collected, but they were never entered before 1991.
3 And that was done only in some republics.
4 Q. All right. Do you know why it was done in Bosnia-Herzegovina,
5 for instance?
6 A. I believe that the reason lies in the fact that the material was
7 scanned. The methodology of entering census data progressed with the
8 progress of technology. The material was scanned, and when you scan a
9 document you can automatically see the name. Serbia did things
10 differently. Only the numerical data were entered. There are ladies who
11 do that. So the first and the last names were never entered before the
12 year 1991, and when this was done it was done only in some republics.
13 The names were never entered onto the magnetic media. It did exist in
14 the questionnaires, and they were never considered as census data, and
15 indeed it is not one of the census data.
16 Q. All right. Now, while I was talking about sources, let's at
17 least at this point discuss the other two sources that were relied upon,
18 at least for the one -- for the one study. And the second source, as I
19 understand it, had to do with the voter registration. Could you please
20 give us an overview of that. How was that compiled? How reliable is
21 that as a source?
22 A. All the sources that were used by the OTP expert are incongruous.
23 The census collects data for the entire population, and you know exactly
24 the moment and how it is done. I'm trying to get to that. The registers
25 of voters use a completely different methodology of data collection. The
1 methodology that is used is on a voluntary basis. Every of-age citizen
2 volunteers to vote or not, and if they volunteer to vote, they provide
3 certain data, the first name, the last name. They can also provide their
4 place of residence before 1991, and they also choose their polling
5 station. They could register themselves or be registered by somebody
6 else. They could even vote in absence.
7 Registers of voters are not a source of data that to my knowledge
8 have ever been used in scientific or professional research. They are
9 considered unreliable not only in terms of their scope but also in terms
10 of the fact that something can be provided and other things don't have to
11 be provided. As soon as you have this arbitrariness the sources cannot
12 be reliable. I've never heard of a register of voters ever having been
13 used in a scientific or professional research as a source of data.
14 Q. All right. And what about the third source?
15 A. The third source of data originated from the State of Bosnia and
16 Herzegovina. It could be accepted as an official source of data because
17 it does contain -- it does contain a certain number of mistakes, but it
18 has a key disadvantage. For two-thirds of the refugees and displaced
19 persons it does not provide individual data, which is what Dr. Tabeau
20 used. The only data it provides is for the head of the household. We
21 know who the head of the household is and who the members of a household
23 These data were collected from 1996 to 1998 by the local
24 authorities and handed over to a certain place, and it was -- it was
25 processed in 2004, as far as I can remember.
1 JUDGE MINDUA: [Interpretation] One moment, Mr. Karnavas. I
2 believe we have a problem here.
3 On page 29, line 25 of the transcript, in English it says:
4 Registers of voters is a source of data. In French in the translation I
5 heard, at one point in time I heard that the voters roll is the source
6 referred to, and then I heard something else.
7 Could you give us an explanation for this, please? You mean the
8 voters roll is what you mean as not a reliable source?
9 THE WITNESS: [Interpretation] Well, yes. Lists of voters, voters
11 JUDGE ANTONETTI: [Interpretation] Ma'am, the difficulty I have,
12 maybe it's the same for my colleagues, we as Judges, we were subject to a
13 census in our countries, and we are on the list of voters, the voters
14 roll. If we apply our own system, we are very surprised by what you are
15 saying. It would be useful, therefore, for you to give us something with
16 which we could compare. If you know, for instance, in Germany the voters
17 roll is prepared in a particular way, and in the former Yugoslavia it is
18 different, so that we can understand the differences better.
19 The problem we have is that when we read the Prosecution expert,
20 we read it through the eyes of what we are accustomed to, i.e., voters
21 roll. You as a Defence expert, you say that this is not reliable and is
22 of no interest. That's all very well. Maybe you are right, but what we
23 would like to know is this: We would like you to give us some examples.
24 For instance, if the voters roll in Gornji Vakuf is not an
25 instrument that can be used to conduct a demographic analysis of the
1 population, does this mean that the municipality had established these
2 lists and this was on a voluntary basis and there was no obligation to
3 register on these lists and so on and so forth?
4 THE WITNESS: [Interpretation] Yes. There's no obligation to
5 register for voting. Entering your name on a voters list is your -- is
6 up to you. It's a voluntary thing. As far as I know, all over the
7 world, but I can't say that I know the entire world, I have never come
8 across a single scientific or expert research that relied on the data
9 obtained based on electoral rolls. What is absolutely possible is that
10 the data from the electoral roll may be used to estimate the voters for a
11 certain year, but the other way round, that the number of population is
12 estimated on the electoral roll, I've never heard of that, especially
13 when it comes to scientific and expert research. There may be some
14 lay-research, but I wouldn't be aware of that.
15 Electoral rolls, especially when it comes to this type of
16 analysis, have another key shortcoming, and this has to do with the
17 methodology used by Dr. Tabeau. They do not provide any data about the
18 father, and they don't provide data about the ethnicity of the voter.
19 How can they be used as comparative data for the calculation of an ethnic
20 structure, and how can they be a reliable source for the precise
21 identification of a certain person based just on the first and last
23 JUDGE ANTONETTI: [Interpretation] If I have understood you
24 correctly, as far as voters roll are concerned, if there was no
25 obligation or legal obligation to register -- to register on these rolls
1 in the former Yugoslavia, in other words only those people who wanted to
2 registered, and if you conduct surveys on the basis of these voters roll,
3 you run the risk of making quite a large number of errors.
4 THE WITNESS: [Interpretation] You're right. There's no legal
5 basis for that. There's no obligation to register. I know that there
6 are some states like, for example, Greece, where it is different and
7 everybody is obliged to register, but in the territory of the former
8 Yugoslavia in 1997 or 1998, I'm sure that this obligation did not exist,
9 and electoral rolls were established based on a voluntary registration of
11 In Serbia to this very day it is a voluntary exercise. You don't
12 have to register if you don't want to. And I believe that the situation
13 is the same in the other states of the former Yugoslavia.
14 JUDGE ANTONETTI: [Interpretation] What you have just said is
16 MR. KARNAVAS:
17 Q. Well, let me just follow up since you mentioned Greece, because
18 normally in Greece one is registered where they're -- where they're born.
19 What was the procedure in the former Yugoslavia as far as registering?
20 Where would they register at? Do you know?
21 MS. WEST: Objection, Your Honour. I appreciate the reference to
22 Greece, but I would just ask that counsel keep the facts coming from the
23 witness and not from himself.
24 MR. KARNAVAS: I'll rephrase, Your Honour.
25 Q. Do you know how the registration was done in the former
1 Yugoslavia, and did one have to register in a particular place and, if
2 so, where?
3 A. You mean for voters?
4 Q. Yes.
5 A. On the eve of the election you are informed that you can go to
6 your own municipality and check whether you are on electoral rolls. The
7 electoral rolls were established in the past without any obligation upon
8 you to go there and check. And then if you don't find your name, you
9 have to query the situation. There are some local organisations that
10 register the population, and censuses are very useful for that.
11 I see in the census how many people they have and how many voters
12 will vote. Otherwise there is no legal basis for somebody who is born to
13 automatically become a voter in the place where they were born.
14 Q. Were there any other requirements such as being a resident in a
15 particular place for a particular time?
16 A. If you're referring to Bosnia-Herzegovina.
17 Q. Yes, we're speaking about Bosnia. We're not speaking about
18 anywhere else.
19 A. If you're referring to 1997 and 1998, there was no requirement.
20 There was just a recommendation of -- of say to -- which said it would be
21 good for you to say where you lived in 1991. And as for now, you can
22 register at the place where you lived in 1991 or where you live now and
23 left before April 1992, or you can register where you resided in 1991 or
24 in the new municipality where you live now, or you can also register
25 where you lived in 1991 or the new municipality, but if it's split then
1 it's in its new part.
2 There are very many either/or. In practical terms you could --
3 you had a vast choice of places where you could register for voting. You
4 could register to vote in a place where you never resided or where you do
5 not reside at the moment.
6 JUDGE TRECHSEL: If I may. Ms. Radovanovic, is it possible to
7 register at more than one place?
8 THE WITNESS: [Interpretation] I can't answer that because I don't
9 known what kind of control was in place. Whether somebody could be
10 registered in Mostar, Sarajevo, I really don't know. I don't know what
11 kind of checks they exercised at the time.
12 JUDGE TRECHSEL: Thank you.
13 MR. KARNAVAS: Perhaps --
14 JUDGE ANTONETTI: [Interpretation] Ma'am, I am going to give you a
15 concrete example. Let's assume we take the case of a Serb who is born in
16 Gornji Vakuf. He is born to a mixed marriage. For instance, his father
17 is a Serb and his mother a Croat. When he is 20, this man goes to work
18 in Mostar and elections will be held. Can he register in Mostar, or does
19 he have to register in Gornji Vakuf?
20 THE WITNESS: [Interpretation] They can register in Mostar. In
21 1997 and in 1998 OSCE received voters registrations, which means I could
22 come to a certain constituency and I could say, "This is where I want to
23 register." And they gave you that possibility. They first tried to
24 check where you were in 1991, and then they check why you want to
25 register here. And then I could say, "Well, in April I -- 1991 I left
1 and I've been here for a year," or I can say, "My parents moved and I
2 want to join them in Mostar and that's the reason."
3 OSCE tried to come up or establish completely new electoral rolls
4 based on the population census, and these would reflect the population
5 status at a certain moment, but it also provided a possibility to all
6 people who were in Bosnia or who weren't because they were in foreign
7 states to register.
8 I personally know some people from Bosnia-Herzegovina who are
9 residents of Serbia and they reside in Serbia. However, they registered
10 to vote in Bosnia-Herzegovina at the time, in their municipalities or the
11 municipalities where their parents resided.
12 JUDGE ANTONETTI: [Interpretation] Let me take the last example
13 before the break. Let's picture somebody who was born in Prozor and who
14 at the age of 20 left to work in Germany, Frankfurt. But then during the
15 holidays he comes to visit his parents who have stayed in Prozor. His
16 parents tell him, "Well, there's an election coming up. It would be good
17 if you could vote for so-and-so because we like this person."
18 Now, in this case would -- could that person register on the list
19 in Prozor?
20 THE WITNESS: [Interpretation] Yes. His parents go to their
21 constituency. They say, "Our son is abroad. He is a citizen of Bosnia
22 and Herzegovina, and he has the right to vote," and they request that his
23 name be put on the electoral roll.
24 All citizens of Bosnia and Herzegovina, regardless of whether
25 they were ever registered as voters, have the right to register and to
2 JUDGE ANTONETTI: [Interpretation] Okay. Let me make it a little
3 more complicated. You've just said as he's a citizen, but let's say this
4 young man is in Frankfurt and he got married to a German lady and he has
5 become a German citizen through wedlock or because you've stayed enough
6 years in Germany. Let's also imagine that he also has a German passport,
7 but he also comes back to Bosnia because he has his family living there.
8 Can he still register on the electoral roll, on the voters roll?
9 THE WITNESS: [Interpretation] Yes. He can get all the documents
10 of the State of Bosnia and Herzegovina, including the ID card, passport,
11 and so on. The only condition is that he must not be at the same time a
12 citizen of the State of Germany.
13 In the former Yugoslavia it was impossible to renounce one's
14 citizenship and then become a citizen of some other state. The
15 regulations in the former Yugoslavia did not recognise dual citizenship,
16 so you could have -- be a citizen of one state because you're not forced
17 to renounce your citizenship in order to be a citizen of another state.
18 And if Germany does not have this requirement that you have to prove that
19 you renounced your BH citizenship in order to register as a -- to apply
20 for a German citizenship, that person can have two passports. There are
21 numerous examples. My granddaughter is a citizen of both Greece and
23 MR. KARNAVAS: Two corrections. One on page 36, line 1. The
24 example that was given is that her friend is a citizen of Serbia but
25 voted in Bosnia and Herzegovina. She can correct me if I'm wrong.
1 And on line 27, she indicated that one can be a citizen of
2 Germany as long as they don't tell the state that they're -- as long as
3 they don't tell Bosnia that they're a citizen of Germany they can still
5 Q. So perhaps the witness can confirm that.
6 A. Yes.
7 Q. Those questions.
8 A. Perhaps I would like to --
9 Q. We'll try to make a clear record.
10 A. Yes.
11 Q. So your friend is a citizen of Serbia but voted in
13 A. In 1997 -- in 1998 he voted in Bosnia-Herzegovina.
14 Q. We're just trying to be precise with the -- with the record,
15 that's why. Secondly -- it may have been mistranslated because you're --
16 you might be speaking too fast.
17 Somebody can be a citizen of Germany, have citizenship, a
18 passport. As long as he doesn't tell the authorities in
19 Bosnia-Herzegovina, he's capable of voting in Bosnia-Herzegovina. That's
20 what you were trying to -- that's what you indicated earlier, did you
22 A. Yes, but the authorities never ask you that. So he can vote for
23 as long as he doesn't say, "I don't want to vote because I am a citizen
24 of another state." For as long as he keeps quiet about his other
25 citizenship, he can vote in Bosnia-Herzegovina until the end of his life.
1 JUDGE ANTONETTI: [Interpretation] [Previous translation
2 continues]... clear. I think it's about time to take a break, so we'll
3 have a 20-minute break now.
4 --- Recess taken at 3.51 p.m.
5 --- On resuming at 4.15 p.m.
6 JUDGE ANTONETTI: [Interpretation] Mr. Karnavas, you have the
8 MR. KARNAVAS: Thank you, Mr. President, Your Honours.
9 Q. Professor, now if we could go to P 09836. Do you have that
10 document with you? P 09836. This is the first report on ethnic
11 composition and internally displaced persons, refugees. Do you have it,
13 A. Yes, I do have that document.
14 Q. Okay. We're not going to -- I'm not going to make any particular
15 references to this, but I want to direct your attention that for the next
16 half hour or so we'll discuss this, because I'll have another 30 minutes,
17 perhaps, to discuss the other report and any other issues. I have an
18 hour left of the two hours allocated for my direct examination.
19 Now, if you could briefly tell us whether you found any problems
20 concerning this particular report, keeping in mind, of course, that you
21 already spent some time telling us that -- or reminding us of the three
22 methodological steps that are important in the field of demography. So
23 if you could please start with that.
24 A. The problems in this report stem from the fact that errors are
25 made in all three methodological steps that were to be taken.
1 Q. Okay. Go ahead.
2 A. First of all --
3 Q. [Previous translation continues] ... more concrete.
4 A. First of all, the report is based on methodological incongruent
5 data sources that contain a vast number of errors. The error margin
6 ranges between 30 and 100 per cent.
7 Secondly, the standard statistical and demographic methods were
8 modified and adapted to the specific needs of this research.
9 Thirdly, statistical and demographic conclusions were made on the
10 basic -- on the basis of parameters which are not statistical and
11 demographic indicators at all, and that is why we cannot speak about this
12 as an expert and scientific report, as a report that would be acceptable
13 from the point of view of its expertise and scientific value.
14 Q. All right. If I could start first with the first problem. You
15 said that methodologically incongruent data sources. Why is that a
16 problem, or how is that a problem?
17 A. The census is an official data source. It is implemented by
18 official organs using specific methodology, and the reason why I say it
19 is a bad data source and why it is methodologically incongruent is
20 because the expert uses first and last names from the census data, and
21 they are not statistical data. They were not corrected in statistics,
22 only in the demographic department. And the census entails making a list
23 of all the population at the critical moment. The electoral rolls or
24 voters lists represent persons who are legally of age and who volunteer
25 to register.
1 I don't know whether any official controls were carried out
2 except for the fact that in the demographic unit of the OTP their names,
3 first names and last names, were corrected.
4 As I've already indicated, that's -- that covers the period of
5 1997, 1998.
6 And in addition to a large number of errors made in the first and
7 last names as they are listed there, I also mentioned that that is a
8 systematic -- systemic error in statistical and demographic terms because
9 there is no ethnic background listed there and no father's name. The
10 databases on the displaced persons also contain a large number of errors
11 when it comes to the first and last names, and there are no individual
12 data for all those persons who are considered to be refugees and
13 displaced persons. And in light of the fact that the Prosecution expert
14 based her methodology on comparing individual data, then I can say that
15 this database has an error margin of two-thirds. But as -- in terms of
16 aggregate data, I consider this data source to be a valid one.
17 Q. All right. Before we go to the second problem that you pointed
18 out to us, are there any questions from the Bench with respect to this?
19 JUDGE ANTONETTI: [Interpretation] Yes, ma'am. I have a question
20 about the margin of error.
21 In your science, that of demographics, what is the percentage at
22 which it is considered, the percentage of errors that it is estimated
23 that a demographic report is reliable or not? So what is the acceptable
24 error -- margin error which is agreeable by the international scientific
25 community in your field?
1 THE WITNESS: [Interpretation] Since demographics is based on the
2 statistics of population, the population statistics, then the error
3 margin is in line with statistics as a science and statistical analysis.
4 Up to 5 per cent is tolerated. A datum is considered to be valid if we
5 have the error margin up to 5 per cent.
6 JUDGE ANTONETTI: [Interpretation] You say 5 per cent in terms of
7 error margin. In Mrs. Tabeau's report, I understand that in your
8 assessment there is a 30 to 40 per cent error margin. Is that right?
9 THE WITNESS: [Interpretation] It is my assessment that it ranges
10 between 30 and 100 per cent, but I have to explain why this range between
11 30 and 100 per cent.
12 With the maximum degree of tolerance and given the great deal of
13 effort put into correcting the names, I think that the minimum -- at the
14 minimum, 30 per cent of the names couldn't be done properly. It is also
15 a question whether they did it at all. And I have to underline that it
16 is impossible to verify, to control this.
17 If you have the database for the displaced persons as the
18 aggregate data, then two-thirds of the persons for whom you don't have
19 individual data, you only have the aggregate for the heads of the
20 household and the members, you could not even verify that. But when I
21 say up to 100 per cent, if you have the data source that does not contain
22 the information that you're dealing with, for instance, the ethnic
23 composition, in statistics it is considered to be a systemic error.
24 Now, it is a whole different ball game if you start to construe
25 that, but you have a data source that in 100 per cent of the cases does
1 not contain the indicator that you use in your analysis later. The same
2 goes for the father's name as a major parameter for a more accurate
3 identification of a person.
4 JUDGE ANTONETTI: [Interpretation] Madam, I take it that you've
5 also said what you just said today in the four other trials to which you
6 were a witness and also for -- on the basis of Mrs. Tabeau's report.
7 THE WITNESS: [Interpretation] Yes. That's what I said. It was
8 not always about Dr. Tabeau's report. In one case it was about
9 Dr. Brunborg's report. In another case Dr. Brunborg -- or, rather,
10 Dr. Tabeau is a co-reference. She participated in it together with
11 Dr. Brunborg, and in the other two cases Dr. Tabeau is the sole author,
12 and that's what I said several times.
13 JUDGE ANTONETTI: [Interpretation] I did not have the time to look
14 again at the decisions. Of course I have read all of the decisions, but
15 I haven't had the time to look at it in depth, but as far as you know,
16 have you read the decisions in question about this discussion amongst
18 THE WITNESS: [Interpretation] No.
19 JUDGE ANTONETTI: [Interpretation] All right then.
20 JUDGE TRECHSEL: Yes. Ms. Radovanovic, there is something I have
21 not quite understood, and maybe it is not correctly recorded. I go back
22 for the others to page 41, line 4 to 7. You are recorded here as saying:
23 "And in light of the fact that the Prosecution expert based her
24 methodology on comparing individual data, then I can say that this
25 database has an error margin of two-thirds. But as in terms of aggregate
1 data, I consider this data source to be a valid one."
2 Frankly, I simply do not understand what you are telling us here
3 if that is what you said.
4 THE WITNESS: [Interpretation] Well, this is more or less what I
5 said. In essence it is correct. The database on the refugees and
6 displaced persons were set up in such a way that when those persons are
7 registered or recorded you take all the information about the head of the
8 household, if I may put it that way, and then you indicate the number of
9 family members. So for instance, if you register Svetlana Radovanovic, I
10 would provide all the personal details, the year of birth, ethnic
11 background, place where I come from, and my family members would be
12 registered together with me but you would not have the personal details
13 for them. For instance, you would not have the indication of their
14 ethnic background.
15 Now, this database has one-third of all the data that pertain to
16 the heads of the household, and for two-thirds you don't have the full
17 information. So the comparison, the method used by Dr. Tabeau, although
18 to be quite frank she resorted to using this database to the smallest
19 extent, gives her an overview of all the material.
20 As for the ethnic background of the refugees and displaced
21 persons registered in this manner for the household members was
22 determined by her on the basis of the ethnic background registered for
23 the head of the household.
24 JUDGE TRECHSEL: And you consider that to be correct?
25 THE WITNESS: [Interpretation] No. I believe that if you look at
1 the aggregate data in this document in the database that it is correct in
2 the aggregate, but if you apply the method that Dr. Tabeau applies, then
3 you cannot look at all of that, because Dr. Tabeau needs to have
4 individual data at her disposal, to look at the individual data. So I
5 think that at the aggregate level the database is correct, it's valid,
6 but it makes it more difficult to apply the methodology applied by
7 Dr. Tabeau, because in two-thirds of the material she cannot actually do
8 the matching.
9 JUDGE TRECHSEL: If I may --
10 MR. KARNAVAS: Just a line -- there's a correction which might
11 cause some confusion. On page 44, line 10, she indicated does not give
12 her an overview of the numbers. So maybe -- I don't know if that is one
13 of the things that you're focusing on, Judge Trechsel. It's not that it
14 gives her, but it does not give her. That's what was indicated.
15 JUDGE TRECHSEL: The --
16 THE WITNESS: [Interpretation] If I may add one more thing. The
17 database on the refugees and displaced persons was used to the least
18 extent by the Prosecution expert, but I considered it to be my duty to
19 let you know what its deficiencies were. On the basis of the data
20 extracted from the database on refugees and displaced persons, she
21 obtained a single result. The Prosecution expert obtained one single
22 result. All the other results regarding the refugees and displaced
23 persons are obtained through the census and through the voters lists.
24 JUDGE TRECHSEL: Thank you.
25 MR. KARNAVAS: Is that it?
1 Q. All right, going back, I mean Judge Antonetti asked you a
2 question which I was going to get at later on, but I'll just ask you now.
3 Within the scientific community, the method that was used by the OTP, and
4 they continue to use over the years this methodology, if we want to call
5 it that, is that acceptable within the scientific community in
7 A. No, but let me just say one thing. The methods used by the
8 Prosecution expert exist. The matching method exists, but the
9 Prosecution expert does not apply those methods in accordance with the
10 standards that apply in scientific research. They are modified, adapting
11 them to her research or his research.
12 Q. All right. And if I could be a little bit more concrete, because
13 on the first aspect, the first problem, you said that she used
14 methodologically incongruent data sources. Is this something that is
15 done within the field of demography, that you would take incongruent
16 sources and try to match it in order to come up with some sort of a
17 result that has meaning?
18 A. No. You don't take mutually incongruent data sources, but the
19 Prosecution expert modifies that and brings the data sources which are
20 not comparable at the aggregate level down to the individual level and
21 then matches them in a way that fits her purpose.
22 Q. All right. Well, could you give us an example so we can
23 understand. When you say she modifies that, how does she modify it? And
24 first let me ask -- perhaps I should ask: Is modification acceptable
25 within the field of demography? In other words, does the demographer
1 have the right to modify data within a particular source?
2 A. They do not have the right to modify the method that they use to
3 obtain the data. If you modified the method, you automatically get
4 modified data.
5 Q. Okay. Could you give us an example how that is done, because
6 there may be this presumption that since therefore the OTP, this is a UN
7 organisation, this must be an acceptable standard within the scientific
8 community. So please tell us where it is or how it is that you differ
9 with that opinion.
10 A. I said that they are methodologically incongruent data sources
11 and that it would be methodologically unacceptable to, for instance,
12 compare the age structure in 1991 and the age structure in 1997, 1998 at
13 the aggregate level obtained from the voters lists. So I'm talking about
14 the aggregate level. We have permanent residents. We know how they were
15 listed methodologically speaking, and on the other list we have people
16 who voluntarily registered. So this is the error, the deficiency there.
17 What the Prosecution expert is doing is applying the matching
18 method, and I explained what that is, through several data sources, in
19 this case the census and the voters list. In this case you identified
20 the same person in both sources. And I mentioned that if the JMBGs were
21 correct, it would not have been a problem, but those numbers do not
22 exist, and that forces the Prosecution expert to produce a matching key,
23 and this matching key, it would be only logical to do it in such a way
24 that you have at least a modicum of security that this is the same -- one
25 and the same person, the one that you declared is matched. The matching
1 key should, in addition from guaranteeing this minimum level of assurance
2 that the identified person is one and the same, should remain the same
3 throughout the matching process.
4 What is the matching key for the Prosecution expert is something
5 that we don't really know with that much accuracy, but in the demographic
6 unit of the OTP it is possible to apply 71 keys. In other words, you can
7 change one of the elements 71 times in order to identify a person. It is
8 only in one case that the expert notes that she used the first name, the
9 second name, the ID card number, and the date of birth, and under such
10 criteria, since the ID card number is not listed in any documents, no
11 matches could be obtained.
12 So we can take as our starting point that the expert -- the
13 Prosecution expert uses at the first step, when she says the first step,
14 that means that there is a second step in the matching, and that's an
15 alarm light. There are no steps in the matching process. If you have a
16 good matching key then you have a guarantee that there will be matches.
17 And let me just note that we use the matching method in our
18 everyday lives. You have the cash card that you use to extract money
19 from the cash machine, and you have the pin number, 1234, whatever, and
20 when you put the card in, it has a certain series of numbers written
21 there, and if you tap in your PIN
22 mistake in just one digit, no matching. You do not get the number. And
23 you don't have the possibility to change those digits 71 times. So the
24 only proper matching is when identification is obtained by matching all
25 the elements in a matching key.
1 So the expert uses the first name, the last name, and the date of
2 birth. In light of the problems that the expert speaks herself about the
3 errors in the first and last names and problems with the dates of birth,
4 in the first step the expert is able to achieve a relatively small number
5 of matches or, for instance, to get 25 per cent of matches. The 20 per
6 cent -- we're talking about 20 per cent of matches with a minimum degree
7 of assurance that matches are proper, but the -- this does not fit the
8 expert's purpose. And then she cuts short the matching key and then she
9 says, "I'm not going to use the first name, last name, and date of birth
10 but only the year of birth," and then she achieves the level of matching
11 of, let's say, 50 to 60 per cent. And then the expert can be unhappy
12 with it, and then they may decide to use only the initial of the first
13 name. And you can then apply 71 criteria in this manner.
14 I'm not saying that Dr. Tabeau is using 71 criteria, but we don't
15 know how many criteria she's actually used, and we know that she does
16 have two steps in the matching process. I assume that she's using the
17 first name, the last name, and the year of birth, because as I
18 verified -- or as I controlled the materials I used those identification
19 details and obtained results that were more or less the same as
20 Dr. Ewa Tabeau's.
21 Let me stress that just the first name, just the last name, and
22 just the date of birth or just the year of birth are not minimally
23 reliable information for the matching process.
24 Let me give you an example. This is something that I got from
25 the demographic unit of the OTP. For instance, in the voters list there
1 are 69 Mirsad Halilovics, and in the census there are 87 persons by the
2 name of Mirsad Halilovic, and of them in the voters list 8 are born --
3 were born in 1965, and in the census 14 of them were born in that year.
4 So you can match a minimum number of 14, but that multiplies, Mirsad
5 Halilovics born in 1965, and then the Prosecution experts call this the
6 visual method, or they like to use the term "conservative method."
7 So you use some other data and you can say well, this Mirsad is
8 from Sarajevo, this Mirsad is from Mostar. So now you have an
9 opportunity to decide whom to accept and whom to consider matched or
11 The moment in the matching process when you decide what to do
12 gives rise to alarm, because regardless of your fairness, your honesty,
13 your striving to obtain scientific truth, that gives you an opportunity
14 to achieve targeted results. It gives you an opportunity to fit the
15 statistics to the framework of your specific research. So -- yeah, I do
17 JUDGE ANTONETTI: [Interpretation] Witness, I shall give you a
18 concrete example. These are very technical issues we are talking about
19 since something is in dispute on the basis of the various expert reports,
20 and in dispute is the expert report provided by the Prosecution. So
21 let's take the case of a Croatian who comes from Kakanj. He has left
22 Kakanj, and he finds himself in Zagreb. He is then registered with the
23 registry office for refugees in Zagreb. I assume that when he registers
24 he will give his JMBG number. This is a Croatian from Kakanj who is a
25 refugee and who has been registered in Zagreb.
1 The expert that is going to look into his case will have - if I'm
2 making a mistake, please correct me - will have the registration that was
3 registered in the registry office, can check whether -- what his age was
4 in 1992, 1993. Let's say he was 30. And he can check that against the
5 electoral roll in Kakanj. He can also check by checking this data with
6 the census list of 1991 in Kakanj if at the time he resided there. Let's
7 assume that he lived there.
8 As this Croatian was born in Kakanj, he is perhaps a Catholic.
9 The expert can then check whether in the registry of the church which the
10 parish priest has, the expert can check and go to the municipality and
11 check whether this person was registered when this person was born in
13 In this case does the expert not have a whole series of files and
14 data whereby he or she could cross-reference the data? And the expert
15 could then draw a conclusion with a hundred per cent certainty based on
16 the example I've just given you.
17 THE WITNESS: [Interpretation] There is a number of available data
18 sources, but they're not used. I only use what I can find in the census,
19 what has been registered, and I use only what exists on the voters list,
20 and I'm not interested in anything else.
21 JUDGE ANTONETTI: [Interpretation] Witness, I'm not here to
22 separate the good experts from the bad. I am here to understand an
24 Let's take the case of a Swedish demographer. He is being asked
25 to go and check the fate of a particular refugee who comes from Kakanj
1 and who was registered in Zagreb as a refugee.
2 This expert, would he not do what I have just described, i.e., go
3 to the church registry of Kakanj, go to the municipality, go to the
4 registry office of the municipality, check the electoral roll and check
5 the census data? Is that not what an expert would do to avoid any
7 THE WITNESS: [Interpretation] This would be the best thing to do,
8 the most correct thing to do. Whether they would do it or not, I don't
9 know. I must repeat this is a very complex tasks which requires a lot of
10 time, but in many cases -- in most of the cases this can be done. It's
12 JUDGE ANTONETTI: [Interpretation] So what I have described is
13 what would happen in an ideal world, but if I understand you correctly,
14 Mrs. Tabeau did not do this.
15 THE WITNESS: [Interpretation] Yes. This would happen in an ideal
16 world and, yes, Mrs. Tabeau did not do it.
17 MR. KARNAVAS: Thank you, Mr. President.
18 Q. Now, I take it, Professor Radovanovic, that part of your previous
19 lengthy answer dealt also with the second problem that you -- you
20 outlined. You indicated that there were three problems in Ewa Tabeau's
21 report. Do you recall what the second problem was? The first one thing
22 the methodological incongruent data sources. Then you talked about the
23 second --
24 A. The second problem is the modification of matching method which I
25 tried to explain. And the third problem is -- is calculation of
1 non-demographic parameters and proposing the non-demographic parameters
2 as something on which one can draw conclusions on statistical and
3 demographic phenomena.
4 When I say non-demographic parameters I'm saying the following:
5 Dr. Tabeau, for example, discusses the changes in the ethnic structure
6 and mentions certain figures or relative figures. I could open her
7 report and read from it, but let me try and give them an approximation in
8 order to explain.
9 For example, she says that a number of Croats in the territory of
10 the eight municipalities that she considers parts of Herceg-Bosna
11 increased -- was increased by 22 per cent during the period from 1991 to
12 1997 or 1998, and she derives that information based on the conclusion of
13 the percentage -- changes of the percentage that was calculated in the
14 following way: If the share of Croats in 1991 was, for example, 44 per
15 cent and the share of Croats in 1997 or 1998 according to her estimate of
16 the number of population of that ethnic group is 50-something per cent,
17 Dr. Tabeau derives her conclusion according to which the number of 52.4
18 per cent in comparison with the percentage of 44.2 is 20 per cent higher.
19 And the calculation is correct. However, it has nothing to do with the
20 increase in the shares of certain ethnicities in the overall population.
21 If there were 44 per cent at one point and now it is 54 per cent,
22 then the increase is only 10 per cent, not more, not 20 or 22 per cent.
23 These are the parameters which are not statistical or
24 demographic. And based on these parameters the OTP experts highlights
25 some things and makes her conclusion on those parameters.
1 To be truthful, the expert also provides a number of tables
2 containing true and correct parameters but she does not provide any
3 comments to them or practically none.
4 There are other problems, for example, comparing apples and
5 oranges within --
6 JUDGE TRECHSEL: Excuse me. Before it gets lost, I -- I could
7 not quite follow your mathematics. You say if the share, and this is
8 page 52, line 25, if the share of Croats in 1991 was, for example, 44 per
9 cent, and the share of Croats in 1997 or 1998 is 50-something, 54 I think
10 it was first, yes, 54, 44 and then 54 per cent, you say the increase was
11 only 10 per cent. For me the increase is 20 per cent. I don't know how
12 you calculate it, because the increase in -- in numbers is 10, and 10 to
13 44 is almost 25. Certainly 20 and not 10 per cent, but maybe I'm
14 completely wrong and then I stand to be corrected.
15 THE WITNESS: [Interpretation] I'll do my best. A share is the
16 share of Croats in the overall population in 1991, and this is a
17 statistical and demographic parameter. And now we say in 1991, of all
18 the population that resided in the area there were 44.4 Croats, for
19 example. And now we have another statistical parameter that applies to
20 the year 1997 or 1998. We do our calculation and we say that of all the
21 population in 1997 or 1998 resided in the relevant territory, the share
22 of Croats was 54.4 per cent. The share, in comparison with the year
23 1991, the share in 1997 went up about 10 per cent, and this relative
24 figure at this moment has nothing whatsoever with the total number of
25 population. These are just relative parameters which demonstrate the
1 quality. However --
2 JUDGE TRECHSEL: You do not have to go on. The passage which I
3 quoted, you spoke of increase, or it is written "increase," and now you
4 take the percentage, the total percentage, not the increase, and that's
5 of course a different matter. So we must probably correct one or the
6 other. The increase is based on the lower figure, and then there is a
7 certain percentage of increase, and if it's 25 per cent increase it's 25
8 per cent increase.
9 You say -- or you speak of the increase between 54 up to -- 44 up
10 to 54 of the whole, and that is a different calculation. Then it's 10
11 per cent. I understand now. Thank you.
12 JUDGE ANTONETTI: [Interpretation] Witness, now as far as the
13 ratio is concerned, I listened very carefully to what you have just told
14 us, which was clarified by the question put by my colleague, but here
15 again I think there is a problem which I shall explain to you.
16 If the ratio represents the total number of Croats in the
17 population at the time, you say 42.4 per cent. Let's assume that that is
18 the correct figure. But in 1998 the ratio, i.e., the number of Croatians
19 accounted for in the entire population since there have been a lot of
20 displaced people, a lot of people who have left, who have gone abroad,
21 who have been displaced, some remained in Croatia and so on, in that case
22 I think we shall be comparing apples and oranges, because the basis for
23 the comparison is no longer the same.
24 THE WITNESS: [Interpretation] To assist both myself and you,
25 please could you look at Table 1 in my report. I don't want to
1 improvise, and I would like to explain what a share is.
2 Technically we're comparing apples and oranges, but this is
3 another methodological issue, and Dr. Tabeau derives another group based
4 on ethnicity and we're again talking about matching. And now we're
5 talking about the meaning of that statistical group.
6 If we accept or if we agree to the fact that by matching data,
7 irrespective of their quality and reliability, Dr. Tabeau arrived at
8 118-something thousand people in 1987 --
9 JUDGE ANTONETTI: [Interpretation] One moment, please. On what
10 page of your report is your table?
11 THE WITNESS: [Interpretation] In B/C/S it is page 8. I don't
12 know what page would that be in English, I'm afraid.
13 JUDGE ANTONETTI: [Interpretation] Let's look at the B/C/S
14 version. Yes. We are very familiar with this language now. We'll be
15 able to work on this document in B/C/S.
16 MR. KARNAVAS: It's 9 in English, Your Honours.
17 JUDGE ANTONETTI: [Interpretation] So please go ahead.
18 THE WITNESS: [Interpretation] This data was taken from
19 Dr. Tabeau's report, and this data shows the collective ethnic structure
20 in the eight municipalities of Herceg-Bosna.
21 In 1991, we can see the total number of adults who were
22 distributed across ethnic groups, and the total number is 231.610. And
23 these are the shares: In the total population Croats accounted for 44.4,
24 Muslims accounted for 34.6, and then follows Serbs and others.
25 The second part refers to the data from 1997, and this is the
1 fruit of Dr. Tabeau's methodology. She compared the census which
2 contains data on ethnicity and electoral rolls which do not contain data
3 on ethnicity.
4 Dr. Tabeau says, "I came up with 118.798 citizens, and when I
5 take their ethnicity as they declared themselves in 1991, this is what
6 their ethnic distribution looks like, so many Croats, so many Muslims, so
7 many Serbs and others." This means she took over the ethnic affiliation
8 from 1991 by the exercise of modify matching in 1997 and 1998, this is
9 the picture that we end up with.
10 Let's not say that the comparison between 1991 and 1997 is good
11 only if they are properly matched. If we calculate all the demographic
12 indicators, we deal with a methodology which says a share of population
13 in the overall population is expressed in percentages and is calculated
14 in the following way, this or the other. And then we see that the share
15 population by ethnicity in 1991 was what it was, whereas in 1997 it was
16 again what it was in 1997.
17 And if we take, for example, Croats for easier calculation, then
18 we can see that the number of Croats who made up the total population in
19 1991 was 44.4 per cent, which means a share in the total mass of Croats
20 was 44.4 per cent. In 1997, of the total mass of the population that we
21 had at the time, Croats accounted for 55.4 per cent.
22 The conclusion would be that a share of Croats in 1997 and 1998
23 in comparison with 1991 was up by 10 per cent.
24 The relative indicators give us an opportunity to compare things
25 without going into any absolute figures. What does that mean? Obviously
1 the number is decreasing, the absolute number. However, in order for you
2 to be able to provide a quality explanation of certain things, you use a
3 relative indicator. The fact is that the numbers went down, and nobody's
4 denying that fact. However, if you look at the statistical figures and
5 masses, they do not have the quality changes. They only show quantity
6 changes that apply across the board to all ethnic groups. However, the
7 quality in ethnic terms of that share of population is what is here, and
8 that's why we have an index and a rank here. Indices show how much the
9 share of population went down in a certain period, whereas rank --
10 JUDGE PRANDLER: Excuse me, Professor Radovanovic. I would like
11 to ask you, you have said the beginning of at least a few minutes ago
12 that the relative indicators give us an opportunity to compare things
13 without going into any absolute figures. What does that mean?
14 "Obviously the number is decreasing, the absolute number," and here I
15 stop at the end of quotation.
16 Now, I really do not see to which number do you refer to when you
17 speak about -- when you mentioned that "obviously the number is
18 decreasing." Do you mean the number of Croats or the number of Muslims
19 or number of whoever? And I would like you to explain, if you would be
20 so kind, to which number and to which population you are referring to.
21 THE WITNESS: [Interpretation] Oh, the total number of population
22 is down as well as each ethnic group which constitutes the total number
23 of the population. However, in order to be even more clear, when I said
24 that it was important to put things in perspective and look at relative
25 figures, it is important but it is also important to look at absolute
1 figures. However, when we're talking about the quality of something, if
2 we only have absolute figures then we cannot exercise any quality
3 comparison in matching.
4 Maybe I would be more clear if I said this: In
5 Bosnia-Herzegovina, the birth figure is about 50.000 babies every year.
6 In China the figure is about 7.5 million. Is the birth rate the same in
7 Bosnia and China? What would you say? At a first glance it's absolutely
8 impossible if you compare 50.000 with 7.5 million. That is why we have
9 the relevant statistical and demographic indicators which give us the
10 birth rate.
11 I'm again speaking off my head. I'm saying that the birth rate
12 in Bosnia is, for example, 13 per mil, whereas the birth rate in China is
13 also 13 per mil. And now I have a quality indicator which says that the
14 birth rate of China and Bosnia do not differ at all. However, I know the
15 quality and the frequency of birth, the rate tells us that 13 new babies
16 are born per 1.000 inhabitants. Bosnia has population that it has, China
17 has the number that it has, and per each thousand new -- 13 new babies
18 are born. This gives me the frequency and the quality of something. And
19 in those terms the relative indicators or shares speak about the quality,
20 not about the quantity of something. The quantity changed, but there is
21 no quality changes in the structure of the population when it comes to
22 Croats and Muslims. There is quality change which applies across the
23 board. It applies to Croats, Muslims, Serbs, and others equally.
24 JUDGE PRANDLER: Thank you for this explanation. On the other
25 hand, I have to confess that I'm still not very much convinced. Then you
1 say, "The quantity changes but there is no quality changes in the
2 structure of the population when this comes to Croats and Muslims," it is
3 a quotation what you have just said. Then I really do not see the basis
4 of this statement because there were at least 10 per cent increase of the
5 numbers of Croats, if I'm not mistaken, from 44 to 54-something. So it
6 is what I am trying to -- getting at, that what is the -- what is your
7 figures about the changes in general among the three major contingents of
8 the population in -- at least in Herzegovina at that time, that is the
9 Croats, the Bosniaks, and the Serbs? So it is what I would like you to
10 point out. And frankly, I'm not very much convinced with the comparison
11 with the Chinese birth rate, because you know that it is a quite other
12 issue. And I would like you to -- again to come back to the original
13 question, and that is the changes in the overall picture of population in
14 that particular area, that is among the three major groups. Thank you.
15 THE WITNESS: [Interpretation] The data we are looking at is not
16 my data. This is the OTP expert's data. The OTP expert claims that the
17 number of Croats in 1997, in comparison with the year 1991, went up by --
18 actually the percentage of Croats went up by 22 per cent. I'm saying no
19 to this. Taking into account the OTP Prosecutor's data and applying the
20 correct methodology of shares and not the percentual [as interpreted]
21 estimate of percentages, and this gives me a different picture, a picture
22 that shows that the figure -- the number did change but the percentage
23 did not go up by 20 -- the share of Croats did not go up by 22 per cent.
24 It only went up by less than 10 per cent.
25 I have tried, but I don't seem to have been clear, to explain how
1 the non-statistical demographic indicators were used by the expert to
2 claim something that is incorrect and how it would look if the expert
3 used a proper statistical and demographic indicator. And I'm quoting
4 this, and I'm providing my explanation for that.
5 JUDGE ANTONETTI: [Interpretation] As I look at the table, one
6 notices that there is a share of 44.4 moving to 54.2, and Muslims move
7 from 44.6 to 37.4, and therefore a logical mind would deduce that now --
8 by now, by 1997 there were many more Croats than there were in 1991.
9 Well, that is an assessment or a conclusion that one can draw.
10 Now, if I make the following analysis, as you were speaking I did
11 the subtraction between the number of Croats in 1997 and the Croats from
12 1991, and when you do that subtraction, you realise that there's a
13 deficit of 38.499 Croats. As I subtract the number of Muslims from 1997
14 to 1991 -- or 1991 to 1997, rather, there's about 36.000 Croats -- sorry,
15 Muslims -- fewer Muslims.
16 Now, that difference, if I look at the population from 1991, 100
17 and so many thousands and 80-something thousand for the Muslims. Now if
18 I for the number of Croats have gone down by 38.499 down from 102.866,
19 it's about 44 per cent and if I do the same calculation with the Muslims,
20 I also end up at 44 per cent, and therefore a logical mind may deduce
21 that there hasn't been a significant change. What do you think about
22 that, since I have 44 per cent in both cases?
23 THE WITNESS: [Interpretation] Under the assumption that you are
24 happy and that you accept the figure that Dr. Tabeau claims as the number
25 of identified person there, if we accept that identification, and we will
1 see that she first says there was a certain number and then she changes
2 that to 80.000. She operates with three different figures. And if we
3 are to deduce apples from oranges we can end up with that figure, but I
4 simply deny the methodology used by Dr. Tabeau, and I believe that this
5 is not a good number. It could be either higher or lower, but Dr. Tabeau
6 also is not certain. She says one thing at one time and another thing at
7 a different time.
8 The 170.000 [as interpreted] inhabitants in a table of this sort
9 in this type of calculation represents the constant population of the
10 eight municipalities at the moment when they voted, and that's the only
11 way they could be compared with 230.000. And although she provides that
12 in the first table, Dr. Tabeau goes on to claim that the number of
13 permanent inhabitants who resided there in 1991 and voted in 1997 were no
14 more than 80.000. And then when we come to the matching exercise, she
15 reduces the 80.000 and ends up with 64.000 refugees. And if we accept
16 all the figures provided by Dr. Tabeau, then your conclusion is
17 absolutely correct.
18 This is nothing more than a simple mathematical operation which
19 demonstrates precisely what you have just told us.
20 If we accept statistical and demographic indicators, however,
21 which were derived based on the same data, they also show that there is
22 no radical change at all. The ranks of the representation of certain
23 ethnic groups, especially Muslims and Croats, have remained the same.
24 MR. KARNAVAS: All right.
25 JUDGE TRECHSEL: I'm sorry, Mr. Karnavas. It's just a matter of
1 probably of either translation or recording. I'm referring to two lines,
2 the last of page 56, and 57 the first two lines. I read what the expert
3 is referred as having said: "And then we see that the share population
4 by ethnicity in 1991 was what it was, whereas in 1997 it was again what
5 it was in 1997." To me this is a total tautology and I don't know
6 whether you said this or whether you wanted to say something with some
8 THE WITNESS: [Interpretation] The share in 1997 is as is
9 indicated in the table. That's what I wanted to say, because I didn't
10 want to go through the exercise of reading each and every percentage.
11 JUDGE TRECHSEL: Well, thank you. That's different. If you
12 wanted to say it's what's in the table, that's a different matter, but
13 when you say 1997 it was what it was in 1997, that is a zero sentence.
14 Thank you.
15 MR. KARNAVAS:
16 Q. Maybe I could -- I don't want to dwell on this table because we
17 have some other matters to cover, but these figures that are listed in
18 1991 column and 1997 column, okay, the total and then the breakdown,
19 Croats, Muslim, Serb and others, how reliable are these figures for 1991?
20 Let's start with that. How reliable are they?
21 A. The data were taken over from the census, but the Prosecution
22 expert took persons in certain age brackets as adults in 1997 in order to
23 have comparable data. It's relatively reliable. Why do I say relatively
24 reliable? It's not because I doubt in the population figures as
25 presented in the census, but purely methodologically there is a method
1 for putting together the age brackets.
2 Dr. Tabeau said "all those who were born in 1980 and before, I
3 declare them to be legal adults because that means that in 1998 they
4 would be eligible to vote." And generally speaking now, I don't want to
5 go into that, but Dr. Tabeau does not say whether this is the calendar
6 year, the year of birth or whatever, but let us now agree to accept that.
7 There is another problem that is rather controversial here. If
8 you put together this set that you want to compare with 1997, and we're
9 not going into the quality now, the datum from 1991, if you aged, if I
10 may say so, a certain percentage of the people, don't you take into
11 account that some of them may have died in the meantime or moved out. In
12 1997 -- well, I can't tell you now. I have this information here in the
13 hotel. I think there were some 10.000 people for all those
14 municipalities that were above 75 years of age. Technically speaking,
15 they are potential voters.
16 You have the vital statistics indicating that in the eight
17 municipalities, 1996, 1997, 1998, over 5.000 people died. What about
18 those who died in 1992, 1993, 1994, 1995? Official statistics do not
19 provide us with the data.
20 The demographic unit does have, on the special order of the
21 Tribunal in The Hague, the statistics of the federation of
22 Bosnia-Herzegovina and of Republika Srpska, the institutes or the bureaus
23 did a special research or study into the mortality between 1992 and 1995.
24 Those databases exist in the demographic unit of the Prosecution.
25 If it is possible to match those with the voters list in order to
1 take the ethnic background of a person and glue it onto the -- onto a
2 voter, then it would be only fair to clean up the 1991 census using the
3 same methodology and to trim it down to a more acceptable level.
4 Q. All right.
5 MS. WEST: Excuse me, Your Honour. I'm sorry to interrupt. It's
6 just if I could have a page cite on this issue of deaths from the report.
7 MR. KARNAVAS: That can be covered on cross-examination at this
8 point in time, unless we want to take a break and the expert can look at
9 it at this point in time, but as I understand it my colleague is not
10 going to be cross-examining for a while.
11 JUDGE ANTONETTI: [Interpretation] Madam, I'm looking at your
12 table, and that will be the last question before the break, table number
13 1. I'm looking at the situation regarding Serbs. 30.495 in 1991. In
14 1997 there were only 3.281 left. So, you know, all -- most of them have
15 disappeared. And in terms of the rank, actually, they ranked number 4.
16 So when we talk about ethnic cleansing at the time is the first ethnic
17 group concerned, wouldn't it be the Serbians? The Serbs, sorry.
18 THE WITNESS: [Interpretation] I'm a demographer. I don't know
19 what ethnic cleansing is. I know what changes in the ethnic structure
21 These data indicate that the most radical changes within the
22 ethnic groups that we considered were experienced by others, and they
23 include Yugoslavs and some other ethnic communities and by Serbs. But I
24 stress once again this -- these are not my data. These are data taken
25 from Dr. Ewa Tabeau's report, and I indicate the source here.
1 JUDGE ANTONETTI: [Interpretation] Right. And so if we understand
2 things correctly, in the index the Croats and Muslims retain the same
3 rank as indicated by Mrs. Tabeau.
4 THE WITNESS: [Interpretation] Yes. According to these data, they
5 have the same rank, and qualitatively speaking, there are no major
6 changes. Serbs and others experience radical downsizing, and this is
7 what can be seen from Dr. Ewa Tabeau's table. But this is not the
8 conclusion that she draws. Dr. Tabeau reaches a completely different
9 conclusion. She talks about Croats specifically, Muslims specifically.
10 She uses a term that is unheard-of in demographics, and that's
12 I may understand that it might be used in some areas but not in
13 demographics. What is a non-Croat? All those who are not Croats. And
14 who is a non-Serb? So Dr. Tabeau does not provide us with an analytical
15 table which would then, regardless of whether I agree or disagree with
16 her data, even the data obtained in this manner would then show that in
17 relation to Croats and Muslims there are no radical changes. There are
18 no radical changes of the ethnic structure.
19 Dr. Tabeau provides us with a huge pile of simple tables, and
20 there is a major difference between analytical tables and simple tables,
21 where she shows each ethnic group separately. It is very difficult to
22 find one's way around all that.
23 I wanted to say that the results obtained by Dr. Tabeau and her
24 conclusions are not consistent when we're talking about ethnic structure.
25 JUDGE ANTONETTI: [Interpretation] Well, ma'am, on the table we
1 have -- I think that it would have been useful to have the previous
2 census figures dating back from 1981. Do you have the figures of the
3 total population from 1981 with a breakdown of Croats, Muslims, and Serbs
4 to see whether from 1981 until 1991 there was or wasn't a decrease due to
5 the fact that some have moved to the West or have gone on to work in
6 Germany, et cetera? Do you have these figures with you or don't you have
8 THE WITNESS: [Interpretation] I do have the figures. Not here
9 but in the witness room. I have a book published by the Federal
10 Institute, by ethnic composition and it provides the ethnic structure at
11 the level of Bosnia-Herzegovina from the 1948 census up to 1991 census.
12 I wanted to draw your attention to this fact. The Federal
13 Institute publishes data for the overall population and Dr. Tabeau here
14 represents only the adult population. But if we look at the overall
15 population for that level we can see that there are no radical changes.
16 JUDGE ANTONETTI: [Interpretation] Well, if you can have a look at
17 this book during the break and tell me how many was the total population
18 in 1981, how many Croats and Muslims, I think it could be useful.
19 Okay. We're going to take a 20-minute break.
20 --- Recess taken at 5.29 p.m.
21 --- On resuming at 5.51 p.m.
22 JUDGE ANTONETTI: [Interpretation] The court is back in session.
23 Witness, during the break you were able to have a look at your
24 book. Can you tell us what the population figures were in 1981 and what
25 the breakdown was between the Croats, the Serbs, and the Muslims?
1 THE WITNESS: [Interpretation] This is the book published by the
2 Federal Institute For Population Statistics. It was done in Zagreb, this
3 book, but it pertains to the 1991 census. At page 1 of this book we have
4 a comparative overview of population in Bosnia and Herzegovina for the
5 censuses between 1990 -- 1948 to 1991 providing a broader range of ethnic
7 In 1981, there were 4.124.256 inhabitants. Out of this figure
8 758.140 inhabitants were Croats. I have to consult this piece of paper.
9 MR. KARNAVAS: Perhaps we could put it on the ELMO so that way,
10 everyone can see it.
11 JUDGE ANTONETTI: [Interpretation] Yes, please.
12 THE WITNESS: [Interpretation] So 758.146 Croats; 1.630.033
13 Muslims; 1.320.738 Serbs; and here we have all the other ethnic groups
14 where the Yugoslavs were the largest group, 326.316 inhabitants.
15 JUDGE ANTONETTI: [Interpretation] Witness, in the table at the
16 bottom we see the percentage figures in 1991, 17.4 per cent of Croats,
17 17.4, and 43.5 of Muslims, 30.1 -- 31.2 Serbs. Is that correct?
18 THE WITNESS: [Interpretation] Yes.
19 JUDGE ANTONETTI: [Interpretation] We don't have the table,
20 however, for the eight municipalities.
21 Mr. Karnavas.
22 MR. KARNAVAS:
23 Q. All right. We --
24 A. In this book we don't have the tables for the eight localities,
25 but we have something else. We have the cantons in Bosnia projecting the
1 figures for 2002, and here we can have a breakdown by municipality. We
2 can have the -- it's the assessment for the number of inhabitants in
3 Bosnia and Herzegovina ethnicity by ethnicity.
4 MR. KARNAVAS: Did Your Honour want to look at -- see that?
5 JUDGE ANTONETTI: [Interpretation] Mr. Karnavas, if you feel it is
6 useful to have the figures for 2002 for the eight cantons --
7 MR. KARNAVAS: I was just asking, Your Honours, given that my
8 time is limited and --
9 Q. Let me go back to one of the -- one of the things that you said,
10 because there were lots of questions concerning these numbers in the
11 table that you showed, and I did ask you a question about how reliable
12 they were, but at one point you began by saying that the ethnic
13 affiliation by modified matching, Ewa Tabeau used this method, ethnic --
14 I take it identifying ethnic affiliation by modified matching. Can you
15 please explain what you meant by that?
16 A. In voters lists there is no information about the ethnic
17 background or ethnic affiliation. In order to obtain the information
18 about the ethnic affiliation of the persons who registered as voters, the
19 Prosecution expert did the matching. And I've already told you how she
20 modified the key, first name, last name, year of birth. And when she
21 matched a person, she states she is able to obtain the information about
22 the ethnic background from the census whether this person is a Muslim,
23 Croat, Serb, Yugoslav. And then she adopts the ethnic background and
24 glues it onto the information from the voters list and says, "In
25 accordance with my matched data, these voters are of this ethnic
1 background." And she presents this as two completely separate and
2 independent data sources, and now you compare what the situation was like
3 in 1991 with what was going on in 1997. This is not correct.
4 Technically and practically you're comparing what was going on in 1991
5 with again what was going on in 1991 for the persons that you managed to
7 What is the big problem here? In all the censuses in the former
8 Yugoslavia, the criterion for the declaration of one's ethnic background
9 was subjective. It's a subjective criterion, and that means that I can
10 decide at any given moment, based on how I'm feeling, having assessed the
11 political situation, having taken a dislike to the census-taker, I can
12 decide at one census to declare myself as a Serb, at the second census to
13 declare myself as a Croat or as a Yugoslav, and I also had the right not
14 to declare my ethnic background. So it is a subjective criterion which
15 gives you an opportunity to change your declaration of your ethnic
17 And I am sure that there were cases of this kind, and in some
18 ethnic groups in some ethnic communities which are smaller, and if you
19 have a smaller mass in statistics it is more sensitive and all the
20 problems are more visible. In smaller ethnic groups you can see the huge
21 irregularities cropping up from one census to the next. And these
22 irregularities cannot be explained away in demographic terms. They are
23 simply a consequence of people declaring themselves differently.
24 And it is quite interesting for us to look at this category
25 Yugoslavs here. Given the political climate at the time when the census
1 is taken in 1991, the political unrest, the imminent break-up of
2 Yugoslavia, you can see that the number of Yugoslavs has gone down
3 substantially. I think that a rough calculation would be a decrease of
4 about 100.000 Yugoslavs less in the ten years between 1981 and 1991. It
5 cannot be explained by demographics, because Yugoslavs did not move out
6 in such great numbers, and they did not die in such great numbers. So
7 the only way in which you can explain that is that they declared
8 themselves differently.
9 And since we now have in front of us the information ranging from
10 1948 to 1991, it is interesting to look at the number of Muslims, and we
11 can see a growth that cannot be explained in terms of demographics. You
12 can see that in 1961, you can see that there were 842.248 Muslims, and we
13 can see that in 1991 the number was 1.484.430 Muslims. It is a
14 substantial increase in the number of people who declare themselves as
16 It was not a case of 500.000 Muslims moving in from other areas
17 or the population growth being that high. So such drastic changes can be
18 explained by the fact that the people declared themselves differently.
19 I am giving you Muslims as an example here. This is quite
20 complex in light of the fact that from one census to the next the options
21 for declaring ethnic affiliation differed, and this is a good example
22 because it is rather drastic.
23 Q. All right. But --
24 JUDGE TRECHSEL: Just a question. Ms. Radovanovic, is there any
25 other way than self-declaration to fix the ethnic identity of a person?
1 THE WITNESS: [Interpretation] In the census.
2 JUDGE TRECHSEL: And is the census based on self-declaration?
3 THE WITNESS: [Interpretation] Yes, exclusively on that, on
5 JUDGE TRECHSEL: So in fact there is no other way than asking a
6 person to know whether that person is a Muslim or a Croat.
7 THE WITNESS: [Interpretation] Correct.
8 JUDGE TRECHSEL: Thank you.
9 MR. KARNAVAS:
10 Q. What about the voter registration list? Didn't they list the
11 national make-up of the individual?
12 A. No. Nobody asked people to declare their ethnicity. At that
13 moment this was not considered significant. Your ethnicity did not
14 matter. I repeat, the voter registration list does not contain the
15 question on ethnic background.
16 Q. I'm going to ask --
17 JUDGE ANTONETTI: [Interpretation] The increase between 1961 and
18 1991 in the number of Muslims, 500.000, that is a very high figure. What
19 is that due to? Is it due to the birth rate or is it due to other
20 factors, i.e., that the Muslims in 1961 felt they had no identity, were
21 not recognised as such and it is only in 1991 when one talks about a
22 possible independence that there is an awareness about this? How can you
23 explain such a differential over a period of 30 years?
24 THE WITNESS: [Interpretation] This is a problem in statistical
25 terms, and it is as follows: In the 1948 census, there was a possibility
1 for anybody to declare themselves as a person without ethnicity or a
2 Muslim Serb, a Muslim Croat, a Muslim Macedonian and so on and so forth.
3 In the 1953 census, the statistical nomenclature included a determinant
4 which says a Muslim in ethnic rather than in religious sense.
5 The 1961 census says Muslim as nationality, i.e., national
6 affiliation. That's how it was described.
7 The 1971 census says Muslim with a capital M, which denotes
8 affiliation with a certain people. And that continued in 1981, 1991.
9 And most probably it will be the same in 2011 in Bosnia. And I suppose
10 this will change into Bosniak.
11 In statistical and demographic sense this data show a broad range
12 not only because of the natural population growth. I believe that this
13 includes one part of natural growth. Primarily it is due to the ethnic
14 consolidation, i.e., the way people declared themselves and the way they
15 were instructed to declare themselves from one census to the next.
16 JUDGE ANTONETTI: [Interpretation] Do you think the religious
17 factor may have played a part?
18 THE WITNESS: [Interpretation] I'm not sure. I would not be able
19 to define it that way. There is a confusion here that was created by the
20 State of Yugoslavia. Nowhere in the world will you find the determinant
21 Muslim as national or ethnic affiliation. This is a religious
22 determinant, a confessional determinant. I don't believe that the ethnic
23 factor played a decisive role, but I never studied that, and I would not
24 be able to tell you whether it did play a role or not in all that.
25 JUDGE ANTONETTI: [Interpretation] Now, this increase, 500.000
1 people, was this not due to the fact that the central -- the federal
2 government did not play a part in this and let people express their
3 faiths, which means that some people claimed to belong to the Muslim
4 faith as a result of this, because the Communist regime was lagging
5 behind in these matters, or was it due to something else?
6 THE WITNESS: [Interpretation] We lived in a very specific system
7 which believed that it had solved the national issue in the best possible
8 way, and I believe that this could have reflected the freedoms that
9 existed. You could declare yourself to be whatever you wanted to be.
10 JUDGE MINDUA: [Interpretation] Witness, I'd like to get back to
11 this issue of the increase in the number of Muslims between 1948 and 1991
12 through the various census, and you said that each person declared he or
13 she belonged to a particular ethnicity. I was a little bit surprised by
14 that, because in a socialist [Realtime transcript read in error
15 nationalist"] Yugoslavia this system along ethnic lines had been designed
16 to balance things out.
17 Now, if people could change ethnicity from one census to another,
18 how could the State clearly establish the nationality of a particular
19 person and then conflicts could erupt in the case, for instance, when
20 positions needed to be shared out between the various ethnicities. When
21 some positions had been earmarked for Serbs or Croats and then one person
22 could just change his or her nationality, his or her ethnicity and state
23 that he or she had a different nationality as opposed to what had been
24 stated in the previous census, what happened then?
25 I'm sorry. Judge Prandler has remarked that on page 73, line 20
1 it says nationalist Yugoslavia. No. I talked about socialist
3 Now, could you answer my question, please?
4 THE WITNESS: [Interpretation] Your question is rather complex and
5 entails a few things that I may not be an expert in. In a political
6 sense we lived in a state where there was fraternity and unity, and
7 nobody had the need to emphasise their nationality. On the other hand,
8 yes, there was a key when positions were allocated to have all the
9 nationalities represented. However, a census allowed you to declare
10 yourself the way you wanted. The -- a census was not a benchmark for
11 somebody to look at the census and see whether Svetlana Radovanovic
12 declared herself a Serb in 1981 and then as a Croat in 1991. The
13 political circles had their own keys to do stuff, and that's as much as I
15 You even have a clause for the 1991, 1981, and 1971 census which
16 says under the constitution anybody has the right not to declare their
17 ethnic background. So nobody was forced to declare themselves in those
18 terms. However, the atmosphere, primarily political atmosphere of a
19 census, in an ethnically heterogeneous midst and the interviewer whose
20 influence can also be felt and can be of a decisive role, all this
21 contributed to people not answering questions which were not precisely
22 defined, which could be considered as subjective. Those issues were
23 disputable. Those questions were disputable.
24 When I say disputable, I'm not trying to deny the statistical
25 official data of the Bosnia-Herzegovina statistics. What I'm saying is
1 that an expert, a demographer who is looking at the national breakdown
2 can use official data to do so, but they have to have good knowledge of
3 methodology and the times and the ways the data was collected. And also
4 such an expert has to be very cautious. They have to tread with caution,
5 because we're talking about the way people declared themselves.
6 The way I declare my national position, my ethnic position in
7 technical terms, does not have to do any -- anything to do with my
8 national identity proper. In the census I can say that I hail from
9 Papua, new -- that's my right. In cultural terms my national identity
10 can also be my private matter, and I don't have to declare myself as
11 anything should I not wish to do so.
12 JUDGE MINDUA: [Interpretation] Thank you very much.
13 MR. KARNAVAS:
14 Q. All right. Well, let me -- let me just ask you a simple
15 question, and I'm going to ask you to give me shorter answers. I don't
16 have unlimited time, regrettably, so -- but essentially the question is
17 this: How reliable is this report that we have been speaking of? How
18 reliable are the data, the statistics? To what extent within the
19 scientific community of demography would this sort of report, the methods
20 that were used, the data that was collected, the manner in it was
21 processed, how reliable is it?
22 A. This report does not meet even the minimum requirement of
23 reliability. It is a product of methodologically incongruent sources of
24 data which abound in mistakes and by the application of modified standard
25 methods that were used in statistics and demographics. And in those
1 terms the report cannot be qualified either as either an expert report or
2 a scientific report.
3 Q. All right. Well, in her testimony Dr. Tabeau at some point
4 indicated that this was sort of a unique technique or a unique method
5 being use here in demography. And perhaps -- and I know that you
6 listened -- you listened to her testimony. Can you please explain to us
7 whether this methodology is unique and, if so, where -- how reliable is
8 it as a methodology within the demographic field?
9 A. This is not a scientific or professional methodology. It is
10 unique, and it is applied and used only in the demographic department of
11 The Hague department and nowhere else in the world as far as I know. And
12 I believe that I'm familiar with quite a lot of literature produced all
13 over the world and in the former Yugoslavia. There is no single expert
14 or scientific research that has used a method modified in this way or
15 similar data sources.
16 Q. All right. Now we're going to speak briefly about the second
17 report, but perhaps given your answer, it begs the question how then,
18 given the way you have criticised Brunborg and Tabeau, how is it that
19 they are using this methodology, if we want to call it that? Is it that
20 they don't know? Is it that they're on the cutting edge? Is it that you
21 might be wrong?
22 JUDGE TRECHSEL: Mr. Karnavas, doesn't that really call for
23 speculation? How is the expert to know what was in the minds of
24 Dr. Tabeau?
25 MR. KARNAVAS: Here's -- I can answer that, and this is why I
1 don't think it's speculation, because we're dealing with an expert and so
2 if an expert, for instance, manipulates data, manipulates sources, and it
3 would appear from the testimony, and I'll put the question if this is
4 manipulation going on, then obviously we may be getting closer to the
6 THE WITNESS: [Interpretation] Are you expecting an answer from
8 MR. KARNAVAS:
9 Q. I'll back up. I'll rephrase for the benefit of the Bench. Is
10 manipulation going on and, if so, please inform the Trial Chamber how is
11 it that these experts would be manipulating data sources, methodology?
12 A. This is manipulation. Why do I dare qualify it as such? I do,
13 because I believe that the OTP expert knows exactly what correct methods
14 are, how they are applied, what are good sources of data. If I had never
15 participated in drafting any report, the expert did not have to read any
16 of my reports, she would still know what I'm talking about. She's a
17 demographer. She's qualified. She has educational -- certain
18 educational background, and she has certain knowledge in demographics.
19 If I have read six reports drafted by the OTP expert and all the
20 six refer to different areas, to different municipalities, to different
21 ethnic groups and different accused, and all six of them are based on the
22 same sources of data, which was the 1991 census and the lists of voters,
23 they all use the same modified matching method. They all match
24 non-existent ethnic structures. They all deal with refugees and
25 displaced persons by simple deduction of the number of population in 1991
1 and what the expert matched. They all say we have a statistical
2 definition of refugees and displaced person. There is no such
3 definition, and any demographer is aware of that -- aware of that.
4 What I'm saying is that if they're absolutely identical, if they
5 use the same sources, if they use the same methodology, then I would be
6 confident to say that they are all a product of manipulation.
7 Q. Do you see -- do you see any efforts in -- as far as the numbers,
8 generating numbers, percentages or high numbers as part of this
9 manipulative process and, if so, could you please explain how such
10 manipulation may or indeed be going on by this particular unit of the
12 A. I believe that the expert strived to augment the numbers of
13 certain groups of population. If she needed more refugees and displaced
14 persons, she could manipulate and say, "This is as many as I matched."
15 There was always an evident strive towards augmentation, and this was
16 based on what the expert's brief was as to what she was supposed to find.
17 And as for matching, for example, if you use in one case 71 criteria to
18 match one thing with another, this is already dubious, but in this
19 particular case and in these reports, the expert says, "I'm using a
20 matching method." She says, "I have two steps to do so." And she does
21 not give us the statistics of the steps. This means that the expert is
22 covering up and prevents us from looking at the degree of combinations
23 that she's using.
24 When I say statistics, this is what I mean: The witness says,
25 "I've matched 140.000 people from the voting lists to the census," but
1 she doesn't say that she did that according to the criteria of name,
2 family name, the name of the father, the date of birth, 20 per cent;
3 according to the criteria of name, the family name, the date of birth, 50
4 per cent, and so on and so forth. Another criteria would be just the
5 initials without the whole name and the figure would be 80 per cent. And
6 this matching statistics would give us a true representation of the
7 combinations that the expert was using, and we would also have the right,
8 the true picture of the group of data that had a certain degree of
10 The expert knows how to do the statistics. She's the one that
11 does the data. In no single report have I seen the matching statistics
12 that I'm talking about.
13 Q. Okay. If we could go on to the second report. This is P 09837.
14 This is the report that you had an opportunity to also look at; is that
15 correct? It's at the very end.
16 A. Yes.
17 Q. And here we can see from the title, "Killed persons related to
18 the siege of Mostar." In demographics could you please define what
19 "siege" means in your field, in your science of demography?
20 A. I can't define this because the term "siege" does not exist in
22 Q. Is it the demographer's job to characterise an event as a siege,
23 for instance?
24 A. This is not a task of a demographer, because they don't have any
25 expertise to do that. I believe that it would be the military, the
1 politics, or some other area that would deal with that. But a
2 demographer might venture to do so. However, if they want to venture
3 into somebody else's area in scientific terms, then it would be
4 professionally fair and correct for them to put a footnote and explain
5 their understanding of the term "siege."
6 Q. All right. Very briefly if you could tell us, having analysed
7 this report, did you find any problems with it and, if so, if you could
8 just briefly describe them to us.
9 A. Well, problems are primarily methodological in nature, and they
10 boil down to the data sources and to the manner and method applied by the
11 Prosecution expert. Manipulating the figures are also quite obvious, but
12 something that really astonished me is the construction of two new
13 methods for the determination of the ethnic background. I have never
14 seen this anywhere except in the demographic unit of the Prosecution.
15 Q. Okay. Could you please explain that, and I will ask you to speak
16 slower, okay, because you're using some technical terms. It's getting
17 late, and this is obviously rather important since we seem to be sort of
18 ploughing through uncharted territory.
19 A. The expert uses two data sources. The first data source is the
20 registry books, the register of deaths. The second data source are the
21 registry books of the War Hospital. The first data source --
22 JUDGE ANTONETTI: [Interpretation] Mr. Registrar, can you please
23 tell us how much time we have left.
24 Please proceed, ma'am.
25 MR. KARNAVAS: I did, Your Honour, have approximately 48 minutes
1 coming into the court after the break.
2 Q. If you could pick it up from the top.
3 JUDGE ANTONETTI: [Interpretation] Yes, that's correct. The
4 Registrar's going to let us know.
5 MR. KARNAVAS:
6 Q. Go ahead, because you're eating into my time now because of that
7 interruption, if you could go ahead and please answer the question,
8 specifically about the manipulation in this particular report.
9 A. The first data source are the registers of birth
10 [as interpreted], which are a statistical source of data, but they do not
11 contain information about the ethnic background; yet the expert draws
12 conclusions about the ethnic background from this source.
13 The second source is the registry book of the War Hospital. It
14 is an amateurish source, not done in accordance with any statistical
15 methodology. It does not contain any information about the ethnic
16 background but the Prosecution expert has a way of dealing with that.
17 The third important methodological problem is that the expert does not
18 know what East Mostar is; yet all the time she keeps talking about East
19 Mostar she is unable to define the area that she's studying. And in
20 conceptual -- the conceptual, temporal, and spatial coverage of what
21 you're studying is an important methodological element. So that would be
22 it in a nutshell.
23 Q. All right. Let's go back. If we can go to the second one, at
24 least, where you indicated that she has a way of dealing with the -- I
25 believe it was the ethnic background, and that dealt with the -- with the
1 book. You indicated that there were three problems, madam, so I'm
2 looking at the second problem that you indicated, and you said that Ewa
3 Tabeau has no problem in making determinations on the ethnic background.
4 Could you please explain what you mean.
5 A. Yes. The expert uses two methods, and I have to say "methods."
6 One has to do with the determination of the ethnic structure based on the
7 register of deaths, and the second has to do with the determination of
8 the ethnic structure based on the books from the War Hospital. The
9 ethnic structure, on the basis of the books from the War Hospital, is
10 based -- is done with the assistance of experts. She does not state what
11 educational background those experts have, who they are. On the basis of
12 the first name and the last name. And as for the ethnic structure, based
13 on the register of deaths she determines it on the basis of the frequency
14 of the appearance of a certain name in the censuses, and on the basis of
15 the frequency she determines what ethnic background the dead person has.
16 What this means is if the name Svetlana has a higher frequency
17 among Croats than among Serbs, then Svetlana is a Croat. If the name
18 Svetlana appears with greater frequency among Serbs than among Croats or
19 Muslims, then Svetlana is a Serb.
20 As far as I'm concerned, that was rather astonishing, and I have
21 yet to see or hear of another case when it was done in this manner.
22 Q. How reliable are the results in this particular report? And when
23 I say "reliable," I'm speaking about it within the scientific community
24 of demography.
25 A. They are absolutely unreliable. If we take into account only the
1 way in which the ethnic structure was determined, if we look at the
2 quality of the sources, then we cannot accept them. And in particular if
3 we take into account the fact that you don't know what area is covered,
4 if you cannot define something, then how can we talk about any
5 reliability? It is absolutely unreliable. And the fact that speaks to
6 it in particular is the fact that there is not a single piece of
7 information that is vital for demographers telling us what segment of the
8 population are we talking about when we talk about those registered as
9 dead. Is it 0.5 per cent, 15 per cent, 50 per cent? It is absolutely
10 impossible to divine that, because we don't know area is covered and what
11 population is covered.
12 And this analysis is not an analysis at all, because it is not
13 labelled an analysis.
14 Q. All right.
15 A. It is labelled the siege of Mostar. Could you please assist me.
16 I don't have it in front of me. "Persons killed in the siege of Mostar,
17 statistical analysis of the books of the Mostar War Hospital and the
18 Mostar registers of death."
19 Q. All right. Let me ask you this: Have you ever seen this
20 methodology applied within the field of demographics?
21 A. No, never.
22 Q. All right. Did you see any manipulation in this particular
23 report in order to augment a particular ethnic group and, if so, how?
24 A. I did see instances of that, and it would be simpler to actually
25 show where. Although Dr. Ewa Tabeau does not know what East Mostar is,
1 she has table 12 in her report where she presents the criteria --
2 MR. KARNAVAS: It's page 15, Your Honour.
3 Q. Okay, continue.
4 A. Just a moment. Let me find it. That's table 12, I believe. I
5 can't find it now. Yes. Table 12. "An overview of the criteria for the
6 selection of data for the analysis of deaths related to the siege of
8 And I would like you to pay attention to line 1. In the last
9 column we see the figure 539. There is the result that the Prosecution
10 exert --
11 MS. WEST: Your Honour, excuse me. I'm sorry to interrupt. It's
12 table 12 of which report?
13 JUDGE TRECHSEL: It's page 15 in the English version, page 15.
14 MR. KARNAVAS: P 098 --
15 THE INTERPRETER: Microphone, please, for the counsel.
16 MR. KARNAVAS: It's P 09837. That's the second report that we've
17 been discussing for the last 20 minutes, and we're on page 15. This is
18 table 12 under paragraph number 4.1, the selection procedure summarising
19 table 12. That's --
20 Q. So continue, Professor.
21 A. Here we have 539 cases, and here you have the criteria, how this
22 was obtained.
23 In the first line it says number, validity, duplicate, what
24 material is it. And then it says "East Mostar." Whether it is or it is
25 not East Mostar, and it says "No." And then whether the year
1 corresponds, yes, whether this was war-related cause, yes, and then
2 whether you have surname, first name, and so on.
3 Dr. Tabeau also notes this in the text, not only in the table,
4 and she says -- should I now find it to quote, because I don't want to
6 At page 6 of the Serbian version, the version in B/C/S --
7 Q. Could you give us the -- the number of the paragraph? That would
8 help us.
9 A. I don't really know.
10 Q. They should be numbered, like 2.2, 2.3.
11 A. I don't have it.
12 Q. Or page 4, page 4 in English. Okay. It's page 4.
13 A. No, I can't find it. Okay. Okay. Dr. Tabeau says this:
14 "Although the category 'Not East Mostar' mainly refers to the records of
15 deaths from outside East Mostar, the records from East Mostar which for
16 various reasons have not been marked as East Mostar are contained under
17 this category as well."
18 Q. All right. Anything else that you wish to comment in particular
19 to this -- this report as far as the problems, manipulations, anything
20 that would assist the Trial Chamber in its evaluation of it? You
21 don't --
22 A. Well, something that is called a distribution method that
23 Dr. Tabeau uses, she uses in both reports, but in this report this is
24 particularly pronounced. The data distribution method exists in
25 statistics, and this is not something that is controversial, but it is a
1 recommendation, at least when it comes to official statistics is not to
2 use it in statistical research. It means that you have a set of
3 statistical units, and one part possesses a certain feature and the other
4 part does not.
5 In this specific case it means that out of, let's say, 400 cases,
6 4 -- or, rather, out of 439 cases that Dr. Tabeau lists as persons killed
7 because of the siege of Mostar -- and I would like to ask you to look at
8 table 2 from my report, the data were taken from Dr. Tabeau. No, that's
9 the cause of death. No. I didn't give you the right reference.
10 So this relates to 539 cases that are classified in according --
11 in accordance with the cause of death. In all of those cases you have
12 the cause of death. And you will see that in most of those cases the
13 cause of death is not listed.
14 THE INTERPRETER: Interpreter's note: Could the witness please
15 repeat the number of cases for which the cause of death was determined.
16 MR. KARNAVAS:
17 Q. Slow down. You'll need to repeat yourself, and again, which
18 table are you referring to?
19 A. I'm talking about table 2 --
20 Q. Okay.
21 A. -- from my report or, rather, this is a table that is contained
22 in Dr. Tabeau's report. Maybe it would be clearer if I were to refer to
23 that, the point.
24 So what is the point? So for the overall mass you don't have the
25 cause of death. You have the cause of death for only one part of the
1 mass, and that's 135 cases. And now you classify them. Total shelling,
2 76 cases killed; 34 cases gunshot wounds and so on. I'm not going into
3 that, whether the cause of death was determined properly or not.
4 Now, in order to be able to show all the cases --
5 JUDGE TRECHSEL: Sorry, the expert is referring to table 16 on
6 page 17 of the Tabeau report.
7 MR. KARNAVAS:
8 Q. And I'm going to have to ask you, madam, because we're trying to
9 make a clear record, you need to make reference to Tabeau's report so we
10 know. Now we know that we're speaking about table 16 in P0 9837, which
11 is on page 17 on the English.
12 Okay. So if you could continue. Just go ahead with your
13 narrative so I can conclude today. This is the distribution of the
14 victims in the siege according to cause of death, this is the table, and
15 it's under 16 in -- it's on 28, page 28 -- okay.
16 A. Table 16, distribution of the victims of the siege according to
17 cause of death. You can see here that you have a total of 539 persons
18 that are classified according to their cause of death. Out of that, 404
19 persons do not have the cause of death determined.
20 Now Dr. Tabeau makes an estimate. You can look at column number
21 3, estimated cause of death. You see shelling was 76 in the original
22 data sources, and now she extrapolates and estimates that there are 303.
23 How does she make that estimate? For the known group of persons in the
24 known death category, and then we're talking about 76 plus 34, plus 18,
25 plus 6, plus 1. That's 135. Now she says 135 is 100 per cent. 76 of
1 the 135, that would be about 50 per cent. The 50 per cent, now I
2 calculate the 50 per cent of the 404, and it's 200 and something or
3 other. 200 and something plus 76, that's 303 cases. So I estimate that
4 their cause of death was shelling. And she does that for all causes of
6 So this method, the distribution method, means that the
7 proportion in a smaller mass is transferred to a larger mass, and you
8 know in advance that you will get bigger figures, because 30 percent of
9 100 and 30 per cent of 400, it's a huge difference. So this method, the
10 proportion method, it's a well-known method, but applied in this manner
11 to such a small mass it is particularly sensitive because you can see
12 that we get an increase of by three or four times. This method applied
13 to a large statistical mass was used in the census, and let me tell you
14 when it was. A certain number of people did not have their age group
15 determined, and if you have 4.4 million inhabitants of Bosnia and
16 Herzegovina, you did not know there the age of maybe 1 or 3, up to 3 or 4
17 per cent of them. So in order not to have this information as unknown,
18 then you make a proportionate calculation. But if you have such a small
19 mass you increase the figures by 4 times across the board. So from this
20 point of view, using this method constitutes abuse, because a demographer
21 knows that if you apply it to small statistical mass, your data end up
22 being deformed.
23 Q. I guess I have one final question and that would be with respect
24 to the third report by Ewa Tabeau, recognising that you did not provide a
25 written analysis of that, can you just in one brief sentence tell us
1 whether the same methods are used in that report dealing with the wounded
2 as the ones that you just explained dealing with the dead?
3 MS. WEST: Objection, Your Honour. As Mr. Karnavas has just
4 voiced, there's no written report on this. This is not something that
5 we're prepared to deal with and it should not be asked of this witness,
6 particularly if she has not prepared an expert report about it.
7 JUDGE ANTONETTI: [Interpretation] Yes, but on the tables it
8 mentions the injured, so it might be interesting to see how all this was
10 MR. KARNAVAS:
11 Q. Can you tell us how the injured were calculated based on the
12 President's observation? How does Ewa Tabeau tabulate the injured as
13 opposed to the dead?
14 A. I'm sorry, what table? What table? I'm sorry.
15 JUDGE TRECHSEL: Mr. Karnavas, should the witness not at least be
16 asked whether she has studied this report at all, because that is not --
17 that should be the foundation for any further question.
18 MR. KARNAVAS: Well, the witness has indicated that -- she said
19 this early on, Your Honours, at the very, very beginning, if you go to
20 the beginning of my -- I'll be more than happy to get the page number.
21 Q. Did you have an opportunity to look at the third report?
22 A. Yes. And in my second report I mentioned the fact that within
23 the War Hospital there were 6.000 wounded, but I studied in great detail
24 the third report, too, and I didn't say --
25 THE INTERPRETER: Interpreter's note: Could all the microphones
1 please be switched off that are not in use.
2 THE WITNESS: [Interpretation] I did not put it in my report first
3 of all because I didn't have enough time, but the second reason was that
4 would constitute repetition, pure and simple, because all the problems
5 are the same. There's only one data source, the War Hospital books. The
6 area again is unknown, the territory is unknown. The only thing we know
7 is that the War Hospital was on the left bank of the Neretva in eastern
8 part of Mostar. And then we can go on to the way in which various things
9 were calculated, statistics that are her own. And this is again not an
10 analysis. This is simply a comment on the data source from which you
11 extract some statistical data, and if you cannot compare it to anything,
12 then what analysis are we talking about?
13 So if I extract my own data and I present my data, there's no
14 comparation. I'm not putting them in relation to anything. It's not --
15 it's no analysis.
16 Q. Thank you. I have no further questions.
17 MR. KARNAVAS: Is there a problem, Judge Trechsel? I mean, we
18 can --
19 JUDGE TRECHSEL: It's okay.
20 JUDGE ANTONETTI: [Interpretation] No. We wanted -- we wanted to
21 ask the Defence as far as tomorrow is concerned. Who is going to be
22 taking the floor tomorrow? From what I understood, Praljak Defence
23 counsel and Petkovic Defence counsel would like to take the floor.
24 Mr. Praljak.
25 THE ACCUSED PRALJAK: [Interpretation] Your Honours -- Your
1 Honour, I would just like to ask one thing, since to my mind some things
2 that I believe I should be able to clear up with the madam still exist.
3 Is it possible that at the expense of the time that I'm allotted for the
4 cross-examination that we get some more time since the topic is a highly
5 complex one, very important, and it begs for a full clarification so that
6 we can understand what we're talking about. I believe that I can do a
7 good job, and I would like to ask for some time that I have not been
8 allotted in accordance with the rules, an hour and whatever it is that
9 we're -- we get, because I believe that Madam Alaburic will be
10 cross-examining. If I can have some time that then will be deducted from
11 the time that I will be given later. I beg for your understanding,
12 because in light of the incrimination or -- that Mostar was subjected to
13 on selective shelling and that it was under siege, it is very important
14 for me and for General Petkovic to clarify every single detail.
15 JUDGE ANTONETTI: [Interpretation] Ms. Alaburic.
16 MS. ALABURIC: [Interpretation] Good afternoon to everyone in the
17 courtroom. General Praljak and I tried to divide the time allotted to
18 us, and it turned out that an hour and a half was not enough to him, and
19 I would like to support his motion that we be allowed to examine this
20 witness slightly longer. We don't believe that it will be longer than an
21 hour and a half so that General Praljak gets about an hour and -- an hour
22 and 10, and I as a modest woman will be happy with 20 minutes.
23 JUDGE ANTONETTI: [Interpretation] The Trial Chamber had decided
24 it would be an hour, but we shall discuss it again.
25 I have understood that there will be no questions from Stojic and
1 Pusic Defence counsel, from Mr. Coric Defence counsel either; is that
3 MR. KHAN: Good evening, Mr. President, Your Honours. I can
4 confirm that as matters stand at the moment, there is no intention on
5 behalf of Mr. Stojic to ask this witness any questions at all.
6 JUDGE ANTONETTI: [Interpretation] Very well. Mr. Coric's Defence
8 MR. PLAVEC: [Interpretation] Your Honours, we won't have any
9 questions for this witness.
10 JUDGE ANTONETTI: [Interpretation] Mr. Ibrisimovic.
11 MR. IBRISIMOVIC: [Interpretation] Thank you very much,
12 Mr. President, you're right. We will not be cross-examining this
14 JUDGE ANTONETTI: [Interpretation] I shall consult my colleagues.
15 [Trial Chamber confers]
16 JUDGE ANTONETTI: [Interpretation] The Trial Chamber, after having
17 deliberated, decides that for the Petkovic Defence counsel and Praljak
18 Defence counsel will have an hour and a half all in all. You can share
19 the time out as you deem fit. I believe Praljak Defence will need one
20 hour, and Mrs. Alaburic half an hour, but anyway, it's for you to sort it
22 Last recommendation, Witness, tomorrow we shall resume at 9.00.
23 Your testimony will be very short tomorrow. Since you are a witness of
24 the Court, you must not contact Mr. Karnavas in any way. This is
25 something I am sure you have been advised of already. You will return
1 tomorrow, and the hearing will start at 9.00. I wish you all a very
2 pleasant evening.
3 --- Whereupon the hearing adjourned at 7.00 p.m,
4 to be reconvened on Tuesday, the 25th day
5 of November, 2008, at 9.00 a.m.