²⁷⁰⁶⁴

Tuesday, 7 October 2003

[Open session]

[The accused entered court]

--- Upon commencing at 9.02 a.m.

JUDGE MAY: Mr. Nice, you have Ewa Tabeau to call, a fairly short witness in terms of evidence in chief, I would guess.

MR. NICE: Yes. Mr. Groome will be taking her, but certainly can be dealt with very compactly.

JUDGE MAY: And what will the position be then, if you can help us?

MR. NICE: We're in a very unfortunate position not of our creation. The following witness was to be a Mr. Rousseau, formerly a colonel in the Canadian Armed Services and an ECMM monitor or something to that effect, and he was a witness who covered some discrete matters. We made arrangements I think on three occasions to suit his convenience for him to come here. Cancellations -- I can't remember the details of the cancellations, but nevertheless three, possibly two, but I think three occasions for him to be here. Tickets were provided for him to fly over yesterday or the day before. People were here at the airport to collect him, and he simply decided of his own volition not to get on the plane, not to tell us about it, and to refuse to come.

I fear that probably doesn't constitute contempt, but it seems to me it's going to constitute enough material for us to seek to subpoena him if we judge, in light of this extraordinary cost of all this exercise and his potential value to the case, that that's worth it. We're very sorry ²⁷⁰⁶⁵ to say that there is, therefore, an unexpected hole in the witness list. The following witness, C-255 - or was it 225? I've forgotten now I haven't got my list in front of me - is fully ready to go but he's only flying in this morning on the early flight. He's coming straight to the building and if we finish Ewa Tabeau, he should be ready to give evidence immediately.

He is, of course, the subject of an application, a general and particular application that evidence in chief should be taken in writing. This is, I know, an application that my learned friend Mr. Kay has not yet seen.

JUDGE MAY: It's an application arising out of the latest ruling by the Appeals Chamber in relation to 89(F), and since it's -- since it is partly a matter of principle and certainly going to be a matter of practice, I think that we will have to spend a little time on that this morning.

What I can say in general terms is that the view of the Trial Chamber is that these applications will have to be dealt with singularly and not a matter of course because that's what the ruling says, although, of course, they can be dealt with expeditiously and orally, I should have thought myself.

MR. NICE: Your Honour, may we perhaps address the position of that witness and possibly the position more generally before or after one of the breaks when we see how Ewa Tabeau is going.

JUDGE MAY: Yes. Certainly.

MR. NICE: Sorry. I have only one other, perhaps for the moment, ²⁷⁰⁶⁶ particular administrative matter to raise with you and it's this: There's a particular witness ordered by the Court to come on the 3rd and 4th, a Monday and Tuesday. The Court's ruling is that in general court sittings will be Tuesday, Wednesday, and Thursday. On the week of the 3rd and 4th, there is a witness due to come back on the Thursday the 6th, and it's the only day upon which he could come back, so that there's a little bit of a conflict there; Monday, Tuesday and Thursday. We are not approaching the witness ourselves directly about the timetable because he's been ordered to attend by the Court. We will approach him if the Court would like us to see if he can move to the Tuesday and Wednesday if that would be more helpful.

JUDGE MAY: We had in mind in fact to order the sittings that week to Monday to Wednesday --

MR. NICE: It's Mr. Harland.

JUDGE MAY: -- to accommodate that witness.

MR. NICE: It's Mr. Harland, who you'll remember has to come back to conclude his cross-examination, who apparently can only come on the Thursday and has been booked for the Thursday. I will, of course, make inquiries to see if he is flexible.

JUDGE MAY: If you would make those inquiries, because that's a week we'll have to -- we'll have to look at rather carefully. I'm going to deal, before the witness comes in, with one other matter, which is to give an oral ruling, a short oral ruling in relation to the Prosecution motion of the 23rd of September, which was entitled "On the Implications of the Recurring Ill Health of the Accused" and on which ²⁷⁰⁶⁷ some argument was heard last week.

In relation to the methods suggested of taking evidence in chief in the absence of the accused through sickness, the Trial Chamber can see no purpose to be served in extending Rule 92 bis together with a video recording as a means of obtaining evidence in chief. However, the Trial Chamber will keep under review the proposal to use deposition evidence and rule on it if the relevant circumstances arise.

The Trial Chamber has extended the role of the amici and will retain under consideration the possibility of assigning Defence counsel. The Trial Chamber were going to order a sitting on the week of the 3rd to the 5th of November, Monday to Wednesday. We will keep that under consideration in the light of what we've just been told about Mr. Harland, but I should point out that there may be other occasions when changes to the schedule for Tuesday to Thursday are necessary.

[Trial Chamber confers]

JUDGE MAY: Mr. Milosevic, you will have the opportunity. You see the Prosecution have returned to the argument that Defence counsel should be assigned, and you've heard me say that we will keep it under consideration. If you want to say anything about that suggestion which they have returned to then you can do so when the opportunity arises. Yes. Well, we'll need to find time for administrative -- an administrative hearing in any event, and that can be one of the matters which is considered.

Yes, Mr. Groome.

MR. GROOME: Your Honour, the Prosecution calls its next witness, ²⁷⁰⁶⁸ Ms. Ewa Tabeau

[The witness entered court]

JUDGE MAY: Yes. If you would take the -- yes. If you would take the declaration, please.

THE WITNESS: I solemnly declare that I will speak the truth, the whole truth, and nothing but the truth.

JUDGE MAY: Thank you very much. If you'd like to take a seat.

THE WITNESS: Thank you.

WITNESS: EWA TABEAU

JUDGE MAY: Yes, Mr. Groome.

MR. GROOME: Your Honour, the Prosecution will be tendering one binder with 12 different exhibits. Could I ask that a number now be assigned to that binder.

While we're waiting for that, Your Honour. Examined by Mr. Groome:

Q. Ms. Tabeau, in one of the exhibits that will be tendered here this morning there is a summary of your educational background and professional experience on page 235 of tab 2. Can I ask you just to supplement that and ask you for the record, are you an employee of the Office of the Prosecutor?

A. Yes, I am.

Q. And how long have you been such?

A. Since September 2000.

THE REGISTRAR: The exhibit number will be 548.

MR. GROOME: Thank you. ²⁷⁰⁶⁹

Q. I'd ask that we begin your testimony by asking you to look at tab 1 of Exhibit 548. And while that is being presented on the screen before you, there have been four demographic reports authored by you submitted in this case. Can I ask you to begin your testimony by simply identifying and differentiating the four different reports?

A. The first report is the report about ethnic -- Changing Ethnic Composition and Internally Displaced Persons and Refugees in the Milosevic Case Area. The three other reports, remaining reports, are devoted to the casualties of the siege of Sarajevo. Two reports of the three, report number 2 and 3, were made at request of the trial team engaged in the General Galic case. And the fourth report, the so-called third Sarajevo report, was made specifically for the Milosevic case and will be discussed here as an addition to the two Galic reports.

Q. I'm going to first ask you questions about the first report that you've identified, and that is the Ethnic Composition of Internally Displaced Persons and Refugees From 47 Municipalities of Bosnia and Herzegovina between 1991 and 1997. Can I ask you to briefly describe the objective of that report. What was it that you set out to study?

A. In this report, we attempted at describing -- excuse me. I don't see anything on my screen. Perhaps someone can help me.

Q. Can I ask that the witness be shown tab 2 of Exhibit 548. What was it that you set out to measure statistically in this first report?

A. In the first report we described changes in the ethnic composition in 47 municipalities in Bosnia and Herzegovina. The municipalities jointly formed the Milosevic case area. This was the first objective. ²⁷⁰⁷⁰ The second objective of the report was to describe numbers of internally displaced persons and refugees from Bosnia and Herzegovina.

Q. I'm going to ask you that you look at tab 2 of 548. Is that a copy of the report that you submitted in this case?

A. Yes, this is the report.

Q. Now, how many municipalities in Bosnia did you undertake to study for this report?

A. We studied 47 pre-war municipalities that jointly covered approximately 50 per cent of the pre-war population of Bosnia and Herzegovina as measured in the 1991 census.

Q. And then those 47 municipalities which are listed in your report, were they municipalities that you yourself chose or given to you from the trial team?

A. The list was given to us by the trial team.

Q. And how do you refer to that group of 47 municipalities throughout the course of your report? What term do you use?

A. We consistently used the term "Milosevic case area."

Q. Now, before I ask you to talk about the report specifically, there are two words that are used throughout the report, or two phrases. The first is "internally displaced persons," you also refer to by the abbreviation "IDP." Can you explain what that is for the purposes of this report.

A. For the purpose of this particular report, we used the statistical definition of an internally displaced person. It is a person who between 1991 and 1997/8, changed the place of residence, so the municipality of ²⁷⁰⁷¹ residence as reported in the 1991 census is different from the municipality of residence in the 1997/8 period.

Q. And similarly, can you define the term "refugee" for us.

A. A refugee is a person - again it is a statistical definition - who left the country between 1991 and 1997/8 so in 1991 was reported in the population census and by 1997/8 moved out of the country and is reported in a country different than Bosnia.

Q. Now, the data sources that you used --

THE INTERPRETER: Could Mr. Groome and the witness, please pause between question and answer. Thank you.

MR. GROOME: My apologies to the interpreters.

Q. The sources that you used for your study can be found in table 1 on page 4. Can I ask you to now simply list the name of the source, identify the custodian of those records, and any brief description that you feel is warranted about those sources.

A. In the report we studied primary resources and used individual records of persons in the analysis. We focused on three major sources, the 1991 population census that reported the population as of the particular moment of census 31st of March, 1991. For the whole country the census included 4.4 million records. We also used voters registers, actually two voters registers from the local elections in 1997 and 1998. Merged together, the register -- merged register contained about 2.7 million records. OSCE, Organisation for Security and Cooperation in Europe, established and maintained the voters registers. The last source is the database of Displaced Persons and Refugees, ^{27072 BLANK PAGE 27073} the DPR. It is an official register of internally displaced persons and refugees in Bosnia and Herzegovina. The register was established by UNHCR together with local authorities and now it is maintained by the government of Bosnia and Herzegovina.

Q. Now, in addition to these three sources upon which you based your findings, were there other sources that you used to compare and to check some of your findings?

A. We additionally used two sources. These two sources were not used in the form of individual records. We only used summary statistics reported in these sources. These sources were two lists established by the authorities of Republika Srpska. First list is the list of citizens who moved out and into the area of Banja Luka sector, status as of 1993. And the second list is an overview of data on the size and ethnic composition of the population in the Banja Luka sector in 1991 and 1995.

Q. In comparing your findings against the findings by studies done by the Republika Srpska, did you find any material discrepancies?

A. We could compare only a limited number of municipalities with these two lists as the coverage of our report, 47 municipalities, is much broader than the municipalities contained in the Banja Luka sector. However, for the municipalities included in RS sources, we didn't find discrepancies that could be seen as significant compared to other results -- our results.

Q. Can I now ask you to describe for the Chamber briefly the methodology that you followed before I ask you to detail some of your findings. ²⁷⁰⁷⁴

A. The methodology can be described as tracing persons reported before the war through the war period and looking for them in the post-war sources. As a starting point, we used the population census 1991, and after the war we could look for persons in the merged voters register 1997, 1998. In order to compare these two sources or persons in these two sources, we had to establish individual links between records of the same persons in these two sources, and once having done this, we could compare place of residence, for instance, of one the same person in 1991 and 1997/8.

Q. Is it correct to say that what you attempted to do was to trace the whereabouts of specific individuals in 1991 and to determine, if you could, where they were in 1997 and 1998?

A. Yes, this is correct.

Q. What was the primary way in which you identified that you had linked the same person from the 1991 census and the 1997/1998 register?

A. Linking is a procedure that is well known in statistics and demography and is usually based on personal identification numbers. So if two records in two different sources have one the same personal identification number, it is very easy to conclude that these two records are of one the same person.

Personal identification numbers were available in the two sources in 1991 census and in the voters register for Bosnia as well. However, not for all citizens of Bosnia these numbers were reported. Sometimes the numbers were also deficient. Therefore, in addition to the personal identification numbers, we used a number of other items, like names, ²⁷⁰⁷⁵ father's name, date of birth, place of birth, and make -- and we made a criterion that contained a multiple number of items that were compared in order to identify persons. Actually, this procedure is quite reliable in the sense that if two records are exactly the same on a number of data items, then the probability that the person linked is the same in the census and in the voters is actually very high.

Q. What percentage of persons that you set out to link between 1991 and 1997, 1998 were you able to link in the way that you've described?

A. For the voters reported in the merged register 1997/8, it was 80 per cent, approximately, voters linked with the census.

Q. Now, before I ask you to discuss your findings in detail, can I ask you to explain to the Chamber the -- the size of the sample that you studied and how it compares to other samples that could be studied for statistical examination of a phenomenon.

A. For 1991, we studied the complete population, the census record of the 1991 population. So the coverage of the census is complete. It is not a sample, it is complete information.

And regarding voters, it is indeed a sample. First of all, it is a sample because it covers only certain ages, age groups in the population, those who are eligible to vote, that is at age 18 or higher, 18 years or higher. Secondly, not all voters eligible to vote registered in the elections in 1997 and 1998. OSCE estimated that on average for the whole country, it was 75 per cent of voters who registered of all eligible to vote. And finally, it is a sample because of the incomplete matching rate. We lost like 20 per cent of voters due to 80 per cent matching ²⁷⁰⁷⁶ rate.

Q. Did you deem it although sample a sufficient sample to carry out your statistical study of the numbers?

A. I believe as a sample it is a very large sample of the population describing the status in 1997/8. We actually worked for the whole country, for Bosnia, with a number like 2 million, more than 2 million records of voters.

Q. And you can -- can you describe for the Chamber the relationship between the size of the sample and the reliability of the findings taken from such a sample?

A. While working with samples, we cannot draw conclusive remarks about the absolute size of the population. All relative measures from large samples are reliable. The larger the size of a sample, the smaller the error we make when estimating measures like fractions, percentages.

Q. I'm going to ask you to now turn to your findings in this first report and I'm going to ask you to discuss one particular municipality in great detail and then take the remainder in a summary fashion. Could I ask you to turn your attention to Srebrenica, and could I ask you to begin by looking at Annex 6.5, table 1, on page 169 of your report.

A. Yes. I see this table on my screen.

Q. Now, this table compares the ethnic composition of the population of Srebrenica in 1991 with the ethnic composition in 1997, 1998. Can I begin by asking you, what was the overall population, people of all ethnicities, in 1991 as compared to 1997, 1998? ²⁷⁰⁷⁷

A. In 1991, in the column titled "All," we see 29.198 individuals reported. This is not the complete population of Srebrenica in 1991. It is a group, a subgroup of the population who in 1997/8 became eligible to vote. The total size of the Srebrenica population in 1991 was 36.666. So the 29 approximately thousand is clearly a subgroup. These are those who were born before 1980 and became eligible to vote in 1997/8.

Q. And did you take this subgroup of the 1991 census, that subgrouping, those people eligible to vote, so you could more accurately compare it to the registered voters in 1997, 1998?

A. Yes. This was the reason. We didn't want to compare things that were uncomparable.

Q. And what was the overall population of people eligible to vote in 1997 in Srebrenica, all ethnicities?

A. I can say what was the size of those who registered to vote and were matched with the census. The number reported here is 7.442. In Srebrenica, this is the group of those who registered to vote in this particular municipality.

Q. Can I ask you to now compare the size of the Muslim population in 1991 as compared to 1997, 1998.

A. For the Muslims, we see that for 1991 there were 21.361 Muslims there, and in 1997 we were able to identify seven registered voters of Muslim ethnicity in Srebrenica.

Q. And can I ask you to same question with respect to the Serb population.

A. For the Serb population, it is 7.205 in 1991 and 7.169 individuals ²⁷⁰⁷⁸ in 1997/8. So these two numbers is practically almost the same.

Q. And can I now ask you to speak to the same population change in Srebrenica in terms of percentages.

A. For the Muslim population we see clear -- a clear decline, even I would call it a dramatic decline from 73.2 per cent in 1991 to 0.1 per cent in 1997. It is a dramatic change that is almost hundred per cent of decline. And for the Serb population we see an increase from approximately 25 per cent to 96 per cent in 1997.

Q. Now, drawing your attention to Annex --

JUDGE KWON: Just pause there. I wonder if the doctor has the data in 1991 of the complete population of Srebrenica.

THE WITNESS: For 1991, it is a complete population of those who were born before 1980. So it is complete --

JUDGE KWON: You said that this number, 29.000, is not a number of complete population --

THE WITNESS: That's correct.

JUDGE KWON: -- rather it is a group, a subgroup. My query is whether you have a total number.

THE WITNESS: Yes.

JUDGE KWON: What is that?

THE WITNESS: The total number for Srebrenica of all population in 1991 is 36.666.

JUDGE KWON: And what was the composition like?

THE WITNESS: The composition was very close to what we see for 1991 in our figures here. Again, this subgroup of those born before 1980 ²⁷⁰⁷⁹ is a huge sample of the actual population of 1991, so if we look at fractions, percentages, only minor differences would be seen when compared for the complete population and for this particular subgroup. So we can think of these percentages for 1991 as a good description of the ethnic composition in Srebrenica, the whole population of Srebrenica in 1991.

JUDGE KWON: Thank you.

MR. GROOME:

Q. Just perhaps for additional clarification, can we subtract the number 7, the number of registered Muslim voters in Srebrenica in 1997 from the 21.361 in 1991 to find out the remaining population of Muslim ethnicity in 1997? Would we be able to do that accurately?

A. I wouldn't do that. It is incorrect from the methodological point of view. We would subtract a sample from the complete population and there are newcomers reported for 1997 among those registered voters so this is not the correct of doing this.

Q. Now, if I can draw your attention to figure 1 in Annex 6.5, that's page 169 of the report, is this a graphic representation of the same findings that you've just discussed?

A. Yes. We consistently use colours, green for the Muslims, red for the Serbs, blue for the Croats and grey for the others throughout the report. This is the illustration, so we observe a big decline, sharp decline in the green bars and sharp increase in the red bars.

Q. Can I ask you to describe to the Chamber how it was you determined that a particular individual was of Muslim ethnicity or Serb ethnicity or Croat or other? ²⁷⁰⁸⁰

A. Ethnicity has been consistently used in our report according to the definition as in the census. So for the voters reported for 1997/8 through the links we could take for them the ethnicity reported in that census. So there is no bias in all these analyses that we present related to changes in the definition of ethnicity of individuals.

Q. So the -- you collected your ethnicity or your description of ethnicity from the census and which was in turn each individual describing how they saw themselves; is that correct?

A. In that census, ethnicity was a self-reported item. It was an open-ended question. Everyone was entitled to report a self-perceived ethnicity. So this is how we used -- used that. Only that the category others is a kind of remaining ethnicities category. All those who reported themselves as Serbs or Muslims or Croats were taken as Serbs, Muslims or Croats and all remaining ethnicity, including Yugoslavs and all kinds of other ethnicities, were taken together in the others category.

Q. We can now move to table 2 of Annex 6.5. That's on page 170 of the report. This table here and it's now being displayed on the screen before you, can I ask you to describe the difference between table 1 and table 2?

A. Table 2 shows only those voters who originated from Srebrenica in 1991. They were reported in 1991 census and were identified in the voters register. So it is a subgroup of all those reported in the previous table in the number of 29.000, approximately 29.000 whom we found irrespectively were irrespectively of the place where they registered to vote. And in this table, in table 2, we show the distribution of those voters according ^{27081 BLANK PAGE 27082} to the place of registration, whether it was in Srebrenica or outside Srebrenica.

Q. Now, your table indicates that in 1991, there were 3.152 Serbs that you were able to find both in 1991 as having the same residence in 1997, 1998; is that correct?

A. Could you repeat the number.

Q. Can I ask you to explain the figure under Serbs at 1991 residence 3.152. What is the significance of that number?

A. This is those Serbs who used to live in Srebrenica in 1991 and were reported in that census and were found as registered in Srebrenica in 1997/8. This is this group of Serb voters.

Q. And how many Muslims were able to be found filling the same criteria, being there in 1991 and being also present in 1997, 1998?

A. So we see only four individuals of Muslim ethnicity that were in 1991 and 1997 in Srebrenica.

Q. Can I ask you to explain the numbers for both Serbs and Muslims with respect to individuals that were there in 1991 but were no longer there in 1997, 1998?

A. For the Serbs there were 771 such individuals. So this would be the individuals who, according to our statistical definition of internally displaced persons and refugees, must be considered internally displaced persons or refugees. And for the Muslims, the number of those internally displaced or refugees is 9.726. So these are all those Muslims who registered outside Srebrenica in other municipalities in Bosnia or outside Bosnia. ²⁷⁰⁸³

Q. And just so it's clear, that figure of 9.726 does not include any Muslims that may have died between 1991 and 1997, 1998?

A. Certainly it doesn't. Those are only those alive at the time of elections.

Q. Could I now ask you to take a look --

JUDGE KWON: Just a second. Dr. Tabeau, the total number in table 2, which is 13.891, what significance does it have?

THE WITNESS: It is more than 50 per cent of the population taken in the table 1 previously for studying.

JUDGE KWON: 1991.

THE WITNESS: Yes. So it is a sample, more than 50 per cent sample, of the 1991 complete population studied in our report that we were able to identify in the voters register in the post-war period as alive, as surviving and registered somewhere to vote.

JUDGE KWON: I think I don't follow you when you say the samples. How do you -- how did you choose them? They are not the population over the voting age, or what sample is that?

THE WITNESS: Well, generally we always, always throughout the report, always work with the population of voting age. But voting age can be seen, defined at a given moment of time. For us, it is the voting age 18 plus at the time of 1997 election. And it is always this population. Only that if we go back to that census, the same group of people is at age, like, 12 years or older because we do it cohort-wise. We still look at the same group of people to avoid any bias related to extra individuals that would be analysed incorrectly. So we still want to compare ²⁷⁰⁸⁴ statistics for one the same group of people, for voters 1997, 1998.

JUDGE KWON: Okay. What about the samples then?

THE WITNESS: The sample, we didn't draw our samples following a statistical procedure. Our samples, the size of our samples actually was determined by our success in finding people from the census in the voters register. So this number, for instance, 13.891, is a sample of those whom we were able to find from Srebrenica, from the census 1991, in the voters register. So this is how this sample was created. So this is our matching success, I would put it this way.

JUDGE KWON: Thank you.

MR. GROOME:

Q. I'm going to now ask you to look at a series of pie charts which describe the statistical or the change in ethnic composition that you found in the municipality of Srebrenica, and I'm now asking you to look at page 171 of your report, and that's figure 2a. Can I ask you to describe what we're looking at on the screen and in this figure.

A. It is a figure summarising the place of registration or residence of the post-war population. According to it, all the residence at and alternatively not at 1991, all the residence. So we see that approximately 23 per cent of the -- of the population of voters who registered in Srebrenica were found at all 91 residence, and 70 -- approximately 77 per cent of the population of voters registered in Srebrenica were found not at 1991 residence.

Q. Now, this particular chart does not distinguish between ethnic groups. Can I ask you to look at figure 2b from the same page and ask you ²⁷⁰⁸⁵ to describe what we are looking at here.

A. This is the ethnic composition of the voters originating from Srebrenica by place of registration to vote. We have two subcharts. The left one is the ethnic composition of those who were found at 1991 residence; this means in Srebrenica. And the right chart shows the ethnic composition of those found at locations different than Srebrenica.

Q. And what was the --

A. After the war, yes.

Q. And what was the ethnic composition, what was the percentage of the ethnic composition of those that remained until 1997, 1998?

A. Excuse me, once again the question?

Q. The left-hand chart, what does that chart indicate with respect to the ethnic composition of those that remained in Srebrenica?

A. Well, 97 per cent, approximately, of them are of Serb ethnicity.

Q. And the right-hand chart, what does that say about the ethnic composition of those that left Srebrenica?

A. This is an ethnic composition dominated by the Muslim ethnicity. 91 per cent, approximately, of them were Muslims.

Q. And finally, if I could ask you to look at chart number 2c on the same page, of 171 of your report, can you please describe what these charts represent.

A. This is a presentation showing each ethnic group separately. For each ethnic group we show percentages, two percentages of those who became internally displaced or refugees. This is the light-coloured part. And the second percentage is the percentage of those who just stayed where ²⁷⁰⁸⁶ they lived, stayed after the war where they lived before the war. So for the Serbs, for instance, it is that approximately 20 per cent of the Serbs is the light-coloured part, and this is the fraction of IDPs and refugees among the Serbs. The remaining 80 per cent approximately is the domestic population who lived in 1991 in Srebrenica and in 1997 in Srebrenica as well.

Q. Can I ask you to summarise your findings for the other ethnic groups as well.

A. For the Muslim population, we see that the fraction, the percentage of IDPs and refugees is 100 per cent, it is approximately hundred per cent in our figures in tables. For Croats -- so this means simply that all Muslims here in -- almost all of this group have become refugees or displaced persons.

For Croats, the percentage of internally displaced or refugees is approximately 47 per cent, and the rest, 53 per cent, are non-displaced persons, non-refugees. And for others, the fraction of IDPs and refugees is 67.4 per cent.

Q. That concludes all the questions I want to ask you about Srebrenica specifically. Were you asked to look at six other municipalities in similar detail, six other municipalities from Bosnia?

A. Yes. We actually looked at municipalities that were located at the eastern border of Bosnia with Serbia, and these municipalities included Bijeljina, Bratunac, Srebrenica of course, Visegrad, Brcko, Foca, and Zvornik.

Q. Now, before I ask you to summarise your findings for all of these ²⁷⁰⁸⁷ municipalities, can you explain to the Chamber the phenomena where, pursuant to the Dayton Accords, three of these municipalities were actually divided and the border between the Federation of Bosnia-Herzegovina and Republika Srpska actually fell in the middle of and divided these municipalities?

A. Yes. The Dayton Peace Accords defined actually a number of municipalities that were split between two political entities. In our study, three such municipalities are included. It is Brcko, Foca, and Zvornik. So instead of just one pre-war municipality of, for instance, Foca, after the war, after November 1995, we have to deal with two municipalities, RS part of Foca, and federal, FBH part of Foca. And this is not only these three municipalities. There were many more municipalities, split municipalities in Bosnia and Herzegovina.

Q. I'm going to ask you to first deal with the four municipalities that were not divided by the Dayton Accords and draw your attention to figure 2a on page 19 of your report. Can I ask you to, using that chart, to summarise your findings for these four municipalities, understanding that you've already told us about Srebrenica.

A. This is a presentation, these charts in figure 28 -- 2a of changes in the ethnic composition in 1991 versus 1997/8. We do this for each ethnic group separately. In these charts we show percentage of a given ethnic group, like Serbs, for instance, in 1991, and next to it the percentage of Serbs in 1997/8. So the left bar is always the 1991 representation, and the right one is the 1997/8 representation. What happened to the Serb population of Bijeljina between 1991 and ²⁷⁰⁸⁸ 1997, 1998?

A. The Serb population -- the percentage of the Serbs in Bijeljina increased from approximately 60 per cent to approximately 91 per cent in 1997/8.

Q. The Muslim population?

A. For the Muslim population, we see a decline, very considerable decline, from approximately 30 per cent to approximately 3 per cent. And the changes for Croats and for others are less significant.

Q. If I can draw your attention to the municipality of Bratunac. Can you please review your findings with us.

A. It is a very similar picture. The numbers are different, though. For the Serbs, increase from approximately 36 to 97 per cent; and for the Muslims, decline from 62 per cent to almost 0 per cent. And again, others and Croats show less meaningful changes.

We discussed Srebrenica already, and Visegrad is the next municipality, last chart. Again, very similar situation; rapid increase of the Serb ethnicity group and the decline, very considerable decline of the Muslim population.

Q. There seems to be no bar representing the Muslim population in Visegrad in 1997. Does that mean that it fell at or near 0 per cent?

A. Yes. It is exactly the case, but it was the same with Srebrenica and Bratunac; it was also near 0 numbers.

Q. If I can now turn your attention to figure 2b, and these are the three municipalities that were divided by the Dayton Accords. And if I could turn your attention to the first one, the municipality of Brcko. ^{27089 BLANK PAGE 27090}

A. Yes. Municipality of Brcko, even though it was split, officially was not considered to be split into federal part and RS part, but we still show the RS and federal part because this what -- is what happened later with Brcko officially.

In the RS part of Brcko, the pattern is very similar to what we have seen already for other four municipalities, for Srebrenica, Visegrad, Bijeljina and Bratunac. The federal part of Brcko shows an entirely different perspective. In this case we see a decline of the Serb population from 13 per cent to 0.3 per cent, while the population of Muslims in the federal part of Brcko increased from 42 per cent to approximately 78 per cent.

So this is a clear picture of population movements that likely took place within the old Brcko municipality. It looks as if the Muslims would be moving from the RS part of Brcko to the federal part of Brcko and the Serbs the other way around. And it is very often the case for split municipalities that these type of results are obtained.

Q. I'm going to now ask you to deal with Foca.

A. So with Foca, the RS part is not different from the general pattern obtained for RS municipalities. We have increase of the Serb population and decline of the Muslim population. And the federal part is, as I said, we observe a decline of Serb population from 30 per cent to 0, and increase in the Muslim population from approximately 69 per cent to 99 per cent.

Q. And finally, the municipality of Zvornik.

A. And municipality of Zvornik, very similar to what we saw for Foca. ²⁷⁰⁹¹ RS part of Zvornik, as all other RS municipalities discussed here, the most affected municipalities, and the federal part of Zvornik, a decline of the Serb population from approximately 29 per cent to 0 and increase in the Muslim population from 70 to 99 per cent approximately.

Q. Now, we've been looking at a bar graph summarising your findings for these seven municipalities. Can I ask you to look at table 3M from page 16 of your report, and can I ask you to describe what it is we're looking at in this table.

A. These are the numbers that we just discussed in the form of the graphics of the charts. This table, this particular table includes fractions of the Muslim population in the 1991 and 1997/8.

Q. Can I ask you to focus on the change in percentage of Muslims, the right-hand column, and maybe describe your findings of the change of percentage of Muslims between 1991 and 1997, 1998, for these municipalities?

A. So using this table, it is easy to make an assessment of the overall change in the Muslim population between 1991, 1997/8 for these particular seven municipalities. And such a summary figure is given in the first row for all seven municipalities jointly. So the last column of this table, 1991-1997 change in the percentage of Muslims is such a summary measure of the change in the percentage of Muslims between 1991 and 1997/8, and the number reported here for seven municipalities jointly is minus 78.9 per cent, and this must be interpreted as a decline by 78.9 per cent.

Q. So does that mean that the overall percentage change in ethnic ²⁷⁰⁹² composition for these seven municipalities was a 78.9 per cent reduction of the number of Muslims that you studied?

A. Yes.

Q. Did you do a similar table with respect to the Serb population?

A. Yes, we did.

Q. Before I ask you to talk about your findings with respect to the Serb population, is it fair to say that the same sources and same methodology was used in studying the ethnic -- changes in ethnic composition of all ethnicities regardless of which particular one a table may represent?

A. Yes, of course. So we were not selective all ethnic groups, changes in the composition, IDPs, refugees were analysed on the basis of exactly the same sources for every ethnic group.

Q. What was the overall change in the Serb population in these seven municipalities that you were asked to focus on?

A. So again the summary measure in the first row of this table, last column is plus 110.2 per cent. This is an increase by 110.2 per cent in the Serb population in seven municipalities.

Q. Now, in the interests of time, I will not ask you to talk about the other tables, similar tables, but is it fair to say that there are similar tables created for the Croat population in these municipalities and for this group called other?

A. Yes. These are tables 3O and 3C for the Croats.

Q. Now, Ms. Tabeau if we can now look at some of your overall findings for the 47 municipalities you were asked to examine, and for this ²⁷⁰⁹³ if we can turn to table 2 on page 11 of the report. This table contains both absolute numbers and percentages. Could I ask you to focus for the moment on the change in ethnic population from 1991 to 1997, 1998, as measured by the change in percentage of each population?

A. We are looking now at table 2, at the last column of table 2 showing the percentage change between 1991 and 1997 for the entire Milosevic case area. So the measure we just discussed for the seven municipalities jointly is also given here in this table for the whole area, indictment area.

So actually the pattern is similar in terms of the direction of changes for the Serbs and the Muslims. For the Serbs, we see a number of 35.4, which indicates an increase in the population of Serbs in the Milosevic case area by 35.4 per cent. And for the Muslims, the number is minus 21, which is a decline by 21 per cent in the fraction of Muslims in the Milosevic case area between 1991, 1997/8.

In relative terms this change was even more substantial for Croats. For Croats, the decline is minus 62.4 per cent. So the Croat population declined in relative terms, the fraction of Croat population in the Milosevic case area declined by 62.4 per cent. And for the others, it is a decline by 24.6 per cent.

Q. Now, did you attempt to map these changes in ethnic composition on two maps of Bosnia?

A. Yes. We prepared such maps, and they are included in the report. And also I have copies I made.

Q. Are they figures 4a and 4b, found on page 26 of the report? ²⁷⁰⁹⁴

A. Yes, these are the two figures, 4a and 4b, page 26.

MR. GROOME: Your Honours, they are quite difficult to see because of their size in the report. We have included as tabs 3 and 4 in the binder larger copies of the same maps. Could I ask the usher to place tab 3 on the ELMO for the witness. Just tab 3. I'll get to that in a minute.

Q. And could I ask you just to first trace the line that is -- was the border in 1997, 1998, between the Republika Srpska and Federation of Bosnia-Herzegovina at the time you did your -- or at the time your data was collected for the study?

A. The inter-entity borderline do you want me to show?

Q. Yes, the borderline.

A. The line I am now showing is the so-called Dayton line. This is the line splitting Bosnia and Herzegovina into two political entities.

Q. And that is the blue line that you've traced?

A. Republika Srpska here and the Federation here.

Q. The municipalities that are marked in a dark green, what does that say about their ethnic composition in 1991?

A. Perhaps I should explain the principles of how this map was made. As I said, we use consistently the red colour for the Serbs, green for the Muslims, blue for the Croats, and grey for the others. In this map, we use two types of, for instance, red or green or blue or others or grey. Dark clear red colour and light red colour. The same with green and blue and others.

In addition to that, we used combinations of colours like green with red resulting in brown or other combinations. ²⁷⁰⁹⁵ The idea was to show absolute majority ethnic groups by clear dark colours so that dark red would indicate an absolute majority of the Serb population in a given territory. Absolute majority is more than 50 per cent of ethnic Serbs in a given municipality. Light red would indicate or just indicates a relative majority of the Serb population, that is, less than 50 per cent but more than -- but 5 per cent or more to the second highest share of other ethnic group.

Q. And does the same hold true for the other colours, green and grey?

A. Yes, yes.

Q. Can I ask now that tab 4 be placed on the ELMO in front you. That's not that yet. It's tab 4 in the binder. It's the map solely of 1997, 1998.

Ms. Tabeau, if you could assist the usher with that. Could I ask you now to take the pointer or that pen and just point to the municipalities where there was a significant change from 1991 in the overall case study of the 47 municipalities.

A. Well, a quick look at the previous map, this is the map for 1991, tells us that the municipalities at the border, eastern border with Serbia had a Muslim majority, and these are municipalities of Zvornik, Bratunac, Vlasenica, Srebrenica, Visegrad, Rogatica, and here is Gorazde, Srpska Gorazde. There were also a number of municipalities that had a mixed ethnic composition, like Foca. This is a mixed composition with two dominant ethnic groups, Muslims and the Serbs; Foca, Milici, and Prijedor.

Q. And how does that compare to 1997, 1998?

A. So if we now look at the map for 1997/8, we see a consistent red ²⁷⁰⁹⁶ colour at the eastern border, also in Foca, in Prijedor. So all the municipalities -- all the municipalities I indicated just a minute ago having a considerable fraction of Muslims or just even an absolute majority of Muslims, after the war had an absolute majority of the Serb population. It is generally the case that the territory of Republika Srpska of the Milosevic case area is exclusively dominated by the Serbs in terms of majority group. And the territory of the federation, as we see, all the municipalities are either with Muslim majority or Croat majority.

MR. GROOME: I'm finished with the map now. Thank you to the usher.

Q. Now, we've been talking about changes in ethnic composition in the Milosevic case. Can I now ask you to turn your attention to your findings regarding population movements that resulted in these changes to ethnic composition. And could I ask that we begin by looking at table 5 on page 29 of your report.

A. Yes. I see it on my screen.

Q. And could I ask you to concentrate on that second line of numbers, people found not to be at their 1991 residence. Can you please summarise your findings in that regard.

A. This is in fact the minimum number or at least number of internally displaced persons or refugees that we were able to identify in our study for the Milosevic case area. So in the --

Q. Can I interrupt you? When you use the word "minimum number," is that to say that this number could not be lower but could possibly be higher? ²⁷⁰⁹⁷

A. Yes, this is the meaning of this expression.

Q. Please continue.

A. For all ethnic groups jointly, in the last column of the table, under the title "All," in the second row there is a number of approximately 400.000 - specifically 399.973 - and this is the minimum number of internally displaced persons and refugees that we identified from the Milosevic case area as of 1997/8. And in the remaining columns under the titles "Serbs, Muslims, Croats, Others," we see the numbers of IDPs and refugees by ethnicity for every ethnic group.

Q. What is the minimum number of Serbs that left their 1991 residence?

A. It is, in this table, a minimum of 115.411.

Q. What is the minimum number of Muslims that left their 1991 residence?

A. It is 231.830.

Q. The same question for Croats.

A. It is 29.581.

Q. And then same for the group designated as Others.

A. It is 23.151.

Q. If I could now turn your attention to figure 5a from page 31 of your report. Can I ask you to quickly summarise what this -- these pie charts depict.

A. Well, the first pie chart, the yellow one, shows the fraction, the percentage of IDPs and refugees in the identified population, post-war population for the Milosevic case area. So four -- we have in this figure ^{27098 BLANK PAGE 27099} 39.5 per cent of IDPs and refugees as compared with 60.5 per cent of those who resided still in the same municipalities as before the war.

Q. In the interest of time, Ms. Tabeau, I'm going to ask you did you also graph or depict these findings on maps as well?

A. Yes, we have done this.

Q. Can I turn your attention to figure 9. It's also included as tab 5 in Exhibit 548. Can I ask you to describe what this map of Bosnia depicts.

A. This is an ethnic majority map for internally displaced persons and refugees originating from the Milosevic case area as of 1997/8. So it is a similar map to the maps we just discussed for the overall population in 1997/8.

Q. Let me ask you a hypothetical question that may illustrate what this map depicts. If we take any one of these municipalities, it is probably true that people from all ethnicities became internally displaced or became refugees; is that correct?

A. Usually for every municipality we observed IDPs and refugees practically from every ethnic group.

Q. What the dark green represents here, are those municipalities where of those displaced people or refugees more than 50 per cent of them were of Muslim ethnicity; is that correct?

A. Yes. So every green municipality indicates that the fraction of Muslims among all displaced persons and refugees was higher than 50 per cent, that this particular ethnic group was a majority group among all internally displaced and refugees. ²⁷¹⁰⁰

Q. And did you examine this phenomena and graph them on maps for each ethnic group separately?

A. We made such maps, and these are figures from 10 to 12, I believe.

Q. And they are --

A. To 13 --

Q. There they are included --

A. -- in the report.

Q. They are included as tabs 6, 7, 8 and 9. I will not go into those in detail now. I will ask the usher to place three maps that have been juxtaposed. They are figures 4a and 4b and figure 10, or tabs 3 and 4 and tab 6. They are now on the ELMO.

I apologise for the low-tech version of this, but Sanction was unable to handle this particular need.

Now, I want to draw your attention to the municipalities in the southern portion of the map, the municipalities of Trebinje, Milici, Nevesinje, Gacko, Mostar, and Kalinovik.

We can see from tab 3 and tab 4 that there is no remarkable change in ethnic composition between 1991 and 1997 that is depicted on those two figures; is that correct?

A. Which municipality did you mention?

Q. That would be Trebinje, Milici, Nevesinje, Gacko, Mostar, and Kalinovik.

A. Perhaps we should focus just on one municipality to make it easier to follow.

Q. The question I have for you: In 1991 in each of those ²⁷¹⁰¹ municipalities there was a Serb majority; is that correct?

A. Well, not in Foca, Srbinje. I see a mixed ethnic composition in 1991 in Foca.

Q. I'm not asking you to --

A. In Gacko, Kalinovik, indeed was a Serb majority in 1991. These are the two ethnic majority maps for 1991. This one for 1997/8, that one --

Q. Leaving aside Foca, just those other municipalities --

A. Yes.

Q. -- they're red in 1991. They're red in 1997.

A. Yes.

Q. The map that's marked as tab 6, figure 10, those municipalities are all green. What is the significance of the fact that those municipalities are marked green in this map depicting internally displaced persons and refugees?

A. This map shows the fraction of internally displaced persons or refugees among Muslims from a given territory and can be read together with these two maps. So for instance, in Gacko, which was dominated by Serbs in 1991 and also in 1997/8 - so ethnic composition ignored change in terms of ethnic majority group; Serbs were here in 1991 were here in 1997/8 - however, this doesn't mean that there were no internally displaced persons and refugees of -- from this particular municipality, and Muslims certainly were one of those who became refugees and displaced persons from Gacko.

Q. And what does dark green indicate? ²⁷¹⁰²

A. And dark green indicates that the fraction of Muslims among those identified after the war was extremely high. It was between 80 to hundred per cent. 80 to hundred per cent of the Muslim population from Gacko from 1991 that we traced in the voters register was not living in Gacko any more but outside this municipality, in other municipalities in Bosnia or outside Bosnia. That is the meaning of this -- of this map.

Q. That concludes all the questions I want to ask you on the first report dealing with Bosnia in general. I want to now turn your attention to your reports with respect to Sarajevo. And if I could ask you to begin by examining tabs 10, tab 11 and tab 12 of Exhibit 548. If the usher could assist in that. And can you identify them as reports regarding demographic changes in Sarajevo that you are the primary author of? That's tabs 10, 11, and 12.

A. Yes. This is the first Sarajevo report, Population Losses in the "Siege" of Sarajevo 10 September 1992 to 10 August 1994.

Q. That's tab 10. Could I ask you to quickly look at tab 11 and identify that document.

A. This is the second Sarajevo report. In fact, it is an addendum to tab 10, to the first Sarajevo report made at the request of the Chamber of the Galic case.

Q. And finally, tab 12.

A. This is the third Sarajevo report that we made recently, in August this year, specifically for the Milosevic case, and we'll discuss it later today.

Q. I want to now ask you to deal with tab 10, this first Sarajevo ²⁷¹⁰³ report that you authored. Can I again ask you -- we'll follow the same procedure as the first report. Can I ask you to describe briefly the sources of the data that you used in compiling this report.

A. We just -- before I will start speaking about sources, I want to make clear that this is an entirely different report and has a different subject. We don't speak about ethnic composition and internally displaced persons and refugees. In this case, the objective of this report was to estimate the number of casualties of the siege of Sarajevo, that part of the siege that was covered by the indictment of the General Galic case. So the subject is estimating casualties, casualties of the siege. There are several sources that theoretically could be used in such a study. However, it is difficult -- it was difficult to find a source that would cover exactly the indictment area. The indictment area in the Galic case was the area of Sarajevo located within the front lines at different moments of the siege.

Q. Did you identify a source that you determined to be the most comprehensive and the most reliable upon which you could do some statistical analysis?

A. Yes, we have identified such a source. It was a huge household survey conducted by local researchers from Sarajevo at 1994, that would be during the siege, and the coverage of this survey, household survey of Sarajevo as we will call it later, was exactly as required for the Galic indictment. Therefore, we used this source, but it is not the only reason. There were very many other reasons that this source suited the needs of this report better than any other source. ²⁷¹⁰⁴

Q. Your report describes in detail the questionnaire that was used in compiling this data. Approximately how many households altogether were interviewed in compiling this data?

A. This survey was meant or intended as a complete household survey within the front lines. It covered approximately 85.000 households living at the territory within the front lines by mid-1994. A simple extrapolation based on the average number of persons in one household resulted in a survey population of approximately 340.000 individuals, and this is actually the material that was studied in the -- in our Galic report, in the first Sarajevo report.

Q. In your study, is it fair to say that you selected any record that had evidence of a death or wounding being recorded during the time period that you were interested in examining?

A. Yes. Yes. This is exactly what we have done. We haven't computerised complete material, far too costly and far too time-consuming, but selected the questionnaires were deaths, natural or violent killings and woundings were reported and it resulted in about 40.000 events. That would be the complete coverage, this number, approximately 40.000, at mid-1994 within the front lines.

Q. Can I draw your attention to table 2, on page 4 of that report in tab 10 and can I ask you to summarise the overview of your findings?

A. This is an overview of survey statistics. We computerised, as I said, approximately 40.000 questionnaires where events, deaths and woundings were reported. We gleaned the data, recorded variables, established a database, eliminated duplicates. There were several ²⁷¹⁰⁵ thousand duplicates, approximately 3.000 duplicates found, and we ended after all this with 37.057 events in the database, events meaning killing, natural deaths and wounded persons.

Q. Now, this table doesn't distinguish between people wounded and killed as combatants and those wounded or killed as civilians; is that correct?

A. This table doesn't, but generally we were able to distinguish between civilians and combatants.

Q. If I can draw your attention to table 1 on page 3 of the same report. Does this table begin to distinguish those that were wounded or injured as a result of combat and those that were wounded or killed as civilians?

A. Table 1, please.

Q. If we could have that up on -- it's on page 3.

A. Yes. Table 1 that begins on page 3 in the report and is continued on page 4 gives an overview of major statistics obtained from this survey. In Part I of the table, we give overall numbers without making the distinction between civilians and soldiers, and on page 4 we present numbers for civilians.

Q. That is what is on the screen now so let us spend some time on that particular table. Can I ask you to go through the numbers about what you were able to determine with respect to civilians that were killed or wounded during the time period that you examined?

A. The overall number of killed civilians was 1.399, and the number of wounded civilians 5.093. ²⁷¹⁰⁶

Q. Can I ask you to break it down according to age of the victim?

A. This breakdown is available from this table. I will only discuss statistics for children and elderly, and elderly. We see the number of 267 killed children at age from 0, from birth to 17. And the number of 83 killed elderly at age 70 or more years. The same numbers for the wounded persons are for children 1.150, and for the elderly, age 70 plus, 172.

Q. Now, in the questionnaires that were done in this household survey of Sarajevo, were the people or were the people being questioned asked how the victim was either injured or killed?

A. Yes. The cause of killing or wounding was included in the records of those killed and wounded.

Q. Did you compile the numbers of how people were wounded or killed?

A. Yes. We actually prepared a standard way of reporting of causes of death or wounding and in this table we present four broad categories of causes. First category is shelling. Second is sniping. The third one are other firearms, and the last one are other casualties. All these causes are definitely war related, that because the causes were reported, we are able to attach a cause of death to every single death or wounding in the study.

So for the killed persons, 932 of them were killed by shelling. 253 by sniping, 101 by other firearms and 113 were victims or casualties of other causes, war related causes of death.

Q. Can I now draw your attention to figure 1, which can be found on page 26 of your report, and that again is tab 10. Does this graphically depict the relationship between civilian deaths and deaths of combatants ^{27107 BLANK PAGE 27108} in the siege of Sarajevo for the time period that you studied?

A. We can see this figure in this way. In this figure, the overall number of casualties, of killed persons, is distributed, first of all, by sex. The left part of the figure is showing men, and the right part women. We also show the age distribution of killed men on the left and killed women on the right.

The colours, the green colour is used to depict soldiers, militaries; combatants, in other words. The light blue colour is depicting civilians. We also had a small number of casualties of unknown status, we didn't know whether they were soldiers or civilians. They are shown marked by grey colour - there are very few - so you can see them here in this figure.

The conclusion is when looking at this figure that the vast majority of killed persons in the Galic related part of the siege are soldiers because of the green colour, and those soldiers were at age from, most of them, 19 or 18 years of age to approximately 60 -- 60 years. The rest are civilians. There is not that much difference in the number of civilian men and civilian women killed, and there are various ages at which they were killed. It is -- for women it is rather clear that these were ages from, say, approximately 20 to 60 years of age.

Q. And this report --

JUDGE MAY: Mr. Groome, the time is coming for an adjournment, when you get a to convenient moment.

MR. GROOME: I am actually at one now. I was just going to ask her to answer one more question, and then ... ²⁷¹⁰⁹

Q. This report is limited in time to the period between the 10th of September, 1992, and the 10th of August, 1994; is that correct?

A. Yes, that's correct.

Q. And after the break, I will ask you to discuss your final report which deals with the broader time period.

A. Okay.

JUDGE MAY: Dr. Tabeau, we are going to adjourn now for 20 minutes. Would you remember during that adjournment not to speak to anybody about your evidence until it's over, and that does include the members of the Prosecution team.

We will adjourn.

--- Recess taken at 10.32 a.m.

--- On resuming at 10.55 a.m.

JUDGE MAY: Yes, Mr. Kay.

MR. KAY: Your Honours, I rise at this stage to raise a matter of concern that has come to my attention that I think the Trial Chamber ought to be aware of. It's been apparent to me that Mr. Milosevic is not in the best of health this morning, and the information that's come to me is that his blood pressure at this stage this morning is 160 over 110, which I believe is a very high level. It's not something I'm inclined to sit on and do nothing about because I feel it is a matter that the Trial Chamber should know about. So I raise it at this stage because it might not be in his interests for him to continue today.

JUDGE MAY: Where do you get this information from?

MR. KAY: From Mr. Tomanovic, one of the associates who was ²⁷¹¹⁰ speaking to me outside.

JUDGE MAY: Not from a doctor.

MR. KAY: I believe there has been a blood pressure check this morning, and that was what I was told. It was taken by a nurse, I think in this building.

JUDGE MAY: But we have had no medical report to this effect. This trial is being constantly obstructed.

Mr. Milosevic, it's your health that's under discussion. How are you feeling?

THE ACCUSED: [Interpretation] Mr. May, now that we had the break, a nurse came and took my blood pressure, and that is what Mr. Kay is talking about, during the break. I didn't ask her to come. I didn't ask for my blood pressure to be taken. She simply came during the break and took my blood pressure. Those are just facts, nothing more than that. So it was not at my request that the nurse came nor have I asked for any adjournment. Probably somebody sent her to me.

JUDGE MAY: Thank you. The legal officer, please.

[Trial Chamber confers]

JUDGE MAY: My colleagues are concerned about your health. We will hear the witness to the examination-in-chief. We will then adjourn and take advice.

Yes, Mr. Groome.

MR. GROOME:

Q. Ms. Tabeau, I would now ask you to turn attention to tab 12 of Exhibit 548. That is your report entitled, "Death Toll in the Siege of ²⁷¹¹¹ Sarajevo, April 1992 to December 1995, A Study of Mortality Based on Eight Large Data Sources."

Can I ask you to first describe the difference in scope between this report on Sarajevo and the report that we've just been discussing?

A. The coverage of the third Sarajevo report, the death toll report, is broader than the coverage of the report called first Sarajevo report we have just discussed made for the General Galic case. In the death toll third Sarajevo report, we present an estimate of the number of casualties in the siege of Sarajevo from April 1992 to December 1995, and the territorial coverage of this report is also broader. We estimate casualties on the territory of six complete municipalities, entire municipalities, while in the Galic report, first Sarajevo report, only the territory within the front lines was covered. These are also the same six municipalities but taken as a part of the overall territory of those six municipalities.

Q. There is a full description of the sources you relied upon in this report in your report. Can I ask you to simply list the sources that you relied upon now?

A. I will do that, but let me finish regarding the municipalities. The municipalities included in this study are the following: Sarajevo centre, Novi Grad, Stari Grad, Novo Sarajevo, Ilidza and Vogosca. The sources that we used for the third Sarajevo report were the following. We used just a number of sources jointly not one single source as has been done in the Galic report. The sources are the following: We used the mortality database called FIS mortality database. FIS stands for ²⁷¹¹² Federal Institute for Statistics. This was a database established by the federal statistical authorities in Sarajevo through a backwards collection of death records from the territory of the Federation of Bosnia and Herzegovina and covering the period of the entire conflict.

Q. The next one?

A. The second source is the list of missing persons for Bosnia and Herzegovina completed by international committee for Red Cross and Physicians for Human Rights, frequently used source and actually the best existing on missing persons who are all believed to be dead.

Q. The next source?

A. The next is the database established by the non-governmental organisation Muslims against Genocide, covers also the entire country and the whole conflict period. Then next one is the household survey Sarajevo, the source that we used for the Galic report.

Q. And --

A. The next one is the list of death records from Bakije Funeral Home, Muslim, a Muslim funeral home. The oldest Muslim funeral home in Sarajevo. Again the coverage is the whole conflict and the territory urban part of Sarajevo practically within front lines. In addition to these five sources, we used three lists of fallen soldiers. These are official records of fallen soldiers, soldiers who died in the Bosnian war. Coverage is the entire conflict period and the territory of Bosnia and Herzegovina.

It is important to note that the lists of fallen soldiers that were provided to us by Ministries of Defence in Republika Srpska and in ²⁷¹¹³ the Federation of Bosnia and Herzegovina and the FIS mortality database are official records, death records existing for Bosnia and Herzegovina for the conflict period.

Q. Now, did these last three sources of lists of fallen soldiers enable you to distinguish between deaths caused to civilians and deaths to combatants?

A. These are just deaths of soldiers, combatants. So this is the information about -- complete information about deaths of militaries during the Bosnian war.

We used these records, after matching, of course, to distinguish -- to -- to mark the deaths of militaries, to make the distinction between civilians and militaries.

Q. Can I ask you now to briefly describe the methodology you employed in this final report.

A. It is very important in -- when estimating war related mortality related to a conflict, it is extremely important to work with more than one source. There are no sources that would be complete and that would be perfect. All sources are incomplete and deficient. All sources taken together give a much better chance to produce a reasonable number of casualties than just one single source.

So the idea of this work was to merge the sources together, to eliminate duplicates, to make a list of unique records, death records related to the siege of Sarajevo, and to present this number as the overall number of casualties of the siege.

Q. And were you able to combine these sources, perform the analysis ²⁷¹¹⁴ that you say you did and result -- resulting in a single, reliable and comprehensive source of information for your study?

A. We were able to achieve this goal. First of all, we worked with each source separately, cleaned, structured, recorded, eliminated duplicates within each source. Then we merged the sources and matched them with the population census through the links with using the same principles for the matching as already discussed on the example of internally displaced persons and refugees report.

Q. Now --

A. Yes.

Q. This report is contained in tab 12 of Exhibit 548. Can I draw your attention to tables 1a and 1b, both to be found on page 2, and can I ask you to describe your findings or summarise your findings for us.

A. The table 1a gives the numbers of matched or linked records, records linked with the census that at the same time were unique each and were all related to the siege of Sarajevo as defined by six municipalities in the period from April 1992 to December 1995. These are only linked records. This means that a number of records are left aside, not included in this table. We just were unable to match all records. Under the table, the average matching rated is included. It is 76.6 per cent. So the related loss of information is like 23.4 per cent. In table 1a and 1b as well, we see four categories of death listed. We have soldiers and three categories of civilians. The most important for our purposes are civilians war related. These were deaths -- these are the deaths that we are most interested in. ²⁷¹¹⁵ Next to those civilians are two other categories of civilians not related to war. This means natural deaths, of causes like aging or disease. And civilians, other deaths, of which we didn't know exactly what was the cause of death and we were unable, therefore, to decide whether a death was related to the conflict or resulted of the conflict or not.

Q. Can we focus on your figure of war related civilian deaths in 1a, the figure of 4.015. Can you explain the difference between that number and the figure of 4.954 for the same category in table 1b.

A. This number, 4.015 is a minimum number, at least number, obtained on the basis of linked records. So with losing 23 per cent, approximately, of records. The other number, 4.954, is a corrected number by increasing the number 4.015 by 23.4 per cent just to compensate for the unmatched records that we still have in our sources but because unmatched couldn't be included in this table.

Q. I'm going to ask you now to address the point of how the Chamber should use the first Sarajevo report and this Sarajevo report, and my question is: Are the numbers of civilian casualties in the first report, the report originally submitted in the Galic case, reflected in these figures here on tables 1a and 1b?

A. The number of civilian casualties, killed civilians from the Galic report, first Sarajevo report, is of course included in the number in tables 1a and 1b. It was 1.399 deaths, as we remember. So the 4.015 deaths do include those 1.400 deaths from the other report. And of course the number 4.954 does also include the number reported in the first big ^{27116 BLANK PAGE 27117} Galic report.

So these two reports should be read together. So the big Galic report is just a case study in which we showed how number of war related casualties are produced. We studied the survey data, large survey, complete survey, and on the basis of this survey produced certain numbers related to the Galic indictment, and however, it is not enough to present a number based on one source and on a part of a conflict. We made the third Sarajevo report to produce a complete, more complete statistic, and these are the numbers in table 1a and 1b. These are actually the numbers, especially the number in table 1b, 4.954, almost 5.000 civilian deaths, war related deaths. This is the number. This is the closest to the reality.

MR. GROOME: Thank you. I have no further questions.

JUDGE MAY: We will now consider the position. We have a note explaining the nurse's view.

[Trial Chamber confers]

JUDGE MAY: We've got the answer, Mr. Milosevic, to your health. The position is this, that the test which you didn't ask for, as you say, it was a routine matter which the nurse is carrying out. In the opinion we have, you can go on today here, and the view is that the blood pressure reading is not dangerous, but we will ask the doctor to come and have a look at you to make sure you're all right this afternoon. Now, it's for you to cross-examine the witness.

Cross-examined by Mr. Milosevic:

Q. [Interpretation] Ms. Tabeau, you entitled your paper "Ethnic ²⁷¹¹⁸ Composition, Internally Displaced Persons and Refugees from 47 Municipalities" from 1991 to 1997.

Is it correct to say that you didn't have the figures from March 1991 until December 1999 [as interpreted], that is, until the end of the war, and that is why you had no sources, because they don't exist for that period?

A. Well, I had the figures from March 1991. It was the census. The Bosnian war, as far as I know, ended in 1995, in the end of 1995, not in December 1999. So I think I did have sources for the post-war period, the voters register that could be used in studying changes in ethnic composition.

Q. Perhaps the translation was not correct. I was talking about March 1991 until December 1995 and not 1999, that is, up until the end of the war you didn't have sources because they don't exist. Is that right?

A. For the war period, between the year 1991 and 1995, I didn't have sources to study changes in between.

Q. That is precisely what I am referring to, because you say on page 5 that for the period from March 1991 to December 1995, we do not have sources which could be used for complex analysis such as those that are being reviewed in this report. Such sources do not exist. Is that right?

A. Yes, that is correct. That is what I wrote.

Q. And then you analyse seven municipalities and not 47 as stated in the title; is that right?

A. This is not correct. I studied 47 municipalities, and in addition to that, I took seven municipalities that were shown as case studies. The ²⁷¹¹⁹ seven municipalities are included in 47 municipalities of the Milosevic case area.

Q. On page 17, you say in paragraph 22: "We continued examining changes in the ethnic composition by focusing on seven selected municipalities all located in the Milosevic case area." This is a new discovery of the opposite side. Those municipalities are Bijeljina, Bratunac, Brcko, Foca, Srebrenica, Visegrad, and Zvornik.

A. Well, it is -- I can only repeat what I said. I studied all 47 municipalities. Seven municipalities, Bijeljina, Bratunac, Brcko, Foca, Srebrenica, Visegrad, and Zvornik are just part of the 47. It is a subgroup of municipalities that I wanted to show more specifically.

Q. Very well. But will you please answer my question, because I really find it quite unclear. How did you choose --

JUDGE MAY: For the moment, it sounded as though you were going to repeat the question, but you can go on if you're going to ask a different question.

How did you choose --

THE ACCUSED: [Interpretation] They're different certainly, Mr. May.

MR. MILOSEVIC: [Interpretation]

Q. How did you select those particular seven different municipalities if you wanted a sample of migrations or persecutions, whatever you want to call it, why didn't you take the largest towns, for instance, in Bosnia and Herzegovina instead of those seven municipalities? For instance, in ²⁷¹²⁰ Sarajevo, you had 175.000 Serbs, and in 1995, you know yourself how many were left. How did you select those particular seven municipalities?

A. These municipalities are called a representation of the most affected municipalities in the Milosevic case area and were selected according to their percentage change in the ethnic composition between 1991 and 1997/8. So these municipalities showed the highest change in the ethnic composition. And this is just a representation. Not all of them with the highest rate of change.

Q. So they represent the highest ethnic changes in the whole of Bosnia-Herzegovina. Is that right, madam, Madam Tabeau?

A. It is in the Milosevic case area, not the whole country. But I believe if we would look at the whole country, these municipalities still would belong to the group of most affected municipalities. But the subject of our report is the Milosevic case area, 47 municipalities.

Q. So that those municipalities were selected by Mr. Groome's team; is that right?

A. At a certain point we certainly discussed the selection, and I believe there were also other reasons to discuss in our report these municipalities, but demographically speaking, without any problems I included these municipalities as case studies due to the change in the ethnic composition.

Q. Very well. Since you studied the whole of Bosnia and Herzegovina, I assume, Madam Tabeau, what would the picture be like if you took the municipalities, for example, of Grahovo, Drvar, Glamoc, Petrovac. In Grahovo and Drvar, more than 99 per cent of the population were Serbs were ²⁷¹²¹ before the war, and after 1995, not a single one remained. What would your conclusions be then?

A. I would like to draw your attention to the fact that I showed a number of municipalities in the Milosevic case area where the Serb population between 1991 and 1997 declined considerably. So it is not that I avoided in my report to give statistics on the changes in the Serb population, but the Milosevic case area comprises mainly municipalities from Republika Srpska. Many of these municipalities were just dominated by Serbs in 1991 and remained dominated by Serbs in 1997/8. So it is -- you asked me to study municipalities located in the Federation that perhaps had a Serb majority in 1991, and in 1997/8 became Muslim dominated but which are not part of the Milosevic case area. Why would I study those municipalities?

Q. I understand. So from the demographic point of view as a science on population, you are narrowing your focus to several municipalities assigned to you by this so-called Prosecution. Does that seem to you to be a proper scientific approach for drawing conclusions?

A. I indeed studied 47 municipalities indicated to me by the Prosecution team, and I said so in the beginning of my testimony, and I have no problems with making a report that is meant for this particular case with the Milosevic case area of the study area. If I would be requested to make a report for the whole country Bosnia and Herzegovina, I would do that, I would make such a report. It is not -- it has nothing to do with scientific or non-scientific approach that I use in my work. This report, 47 municipalities, was needed for this case. I made it, and today ²⁷¹²² we are discussing the figures.

Q. But that is the point, because as you know, if you take from a whole a particular segment and review it alone, then you lose the full picture of the reality of the whole. Isn't that so, madam?

A. It is that a context, a broader context is always useful in comparison, sir. We have made a limited number of comparisons of our figures with the figures for Bosnia and Herzegovina, a limited number of comparisons, sir. I don't think that because of the limited number of comparisons we lost the reality of the whole. I don't think so. So if anything more is needed, I am open to provide any analysis, any comparisons that would be requested, that will be requested.

Q. Very well. And is it in dispute, since you are a specialist in figures, according to official data and the 1991 census data in Bosnia and Herzegovina, there was 1.905.274 inhabitants of Muslim ethnicity? Is that correct? And furthermore, in the category of Yugoslavs, there were 239.857 recorded as Yugoslavs.

Madam Tabeau, is it disputed that in the seven mentioned municipalities, there were 12 per cent --

JUDGE MAY: The witness must deal with these matters one by one. The first suggestion is to the census data for 1991. Can you help us about that, Dr. Tabeau, or not?

THE WITNESS: Well, I can help, but I don't have the figures with me. I have a publication in my office, official figures from the census, that we can use to check the figures. It is probably that these are the numbers that Mr. Milosevic is quoting here. ²⁷¹²³

JUDGE MAY: Very well.

THE WITNESS: But I cannot confirm with hundred per cent certainty because I don't have anything on paper in front of me.

JUDGE MAY: Yes. Now, go on, Mr. Milosevic, and deal with the seven municipalities.

MR. MILOSEVIC: [Interpretation]

Q. Let me just ask you, what kind of demography do you specialise in?

A. At this moment I am working for the Office of the Prosecutor, and I am responsible for estimating demographic consequences of the conflicts in the former Yugoslavia, in particular Bosnia and Croatia.

Q. All right. Before you became a specialist for The Hague and demography in that sense, you took part in several projects, international conferences, et cetera. So tell me, which field of demography were you particularly dealing with in those papers?

A. It was analysis of mortality, and my first research domain was modelling and prediction of mortality by cause of death in European countries, in the Netherlands in latest years before I came here and also in countries in Europe, including countries in Central European region like Poland, Czech Republic, or Hungary.

Q. All right. So most of your papers deal with the field of mortality. Until you became an expert for The Hague, did you ever deal with census statistics, and were you ever in a position to study Yugoslav censuses?

A. Until I came to The Hague, my experience with censuses was indeed limited. I participated at one time in a census as an interviewer in ²⁷¹²⁴ Poland, as a student, but this cannot be taken as a professional experience relevant to the work I have been doing here. I don't have -- I didn't have any previous experience with Yugoslav censuses.

Q. All right. So in terms of your scientific work before you became an expert here, what was primary, sources from vital statistics, that is to say deaths, births, divorces, marriages, et cetera? So you dealt with this kind of information coming from vital statistics; isn't that right, Mrs. Tabeau?

A. It is indeed the case that statistics come from vital events registration. Indeed, you are correct when saying that my experience, previous experience, was with other sources and related but not necessarily exactly the same methods.

Q. All right. At that time you did not have to be familiar with the methodology that is used in censuses, population censuses. Quite simply, you didn't deal with these matters.

A. But that I didn't deal personally with that census or other survey data doesn't mean that I don't understand the survey data, individual level analysis. Yes, it was part of my education to study these issues. The very first information we received at demography lectures in the university was about sources, statistical sources used in demography, and the population census was mentioned as the first place. And since then, systematically throughout all the years of my professional career, I referred to census at different occasions. So I cannot entirely agree with you that I am not aware of censuses, other large population surveys, ^{27125 BLANK PAGE 27126} specificity of survey data and methods used in the analysis of survey data.

Q. All right. But did you publish a single paper or take part in a single scientific project that required a detailed knowledge of such statistics?

A. Well, if you think -- if I am thinking about what I have written in my life, last 13 analytical reports produced here were all made on the basis of individual level data. Some of them based on census and other large population sources. Other reports were based on smaller surveys, on mortality. So I think even though the reports have not been published outside this Tribunal, but many of them were submitted as evidence in several cases, and I think it was -- it is quite a portion of experience that I can refer to today.

Q. My question was -- my question was about before you became an expert here. Did you deal with that before you became an expert here? No, you didn't. Isn't that the answer?

A. The publications from the period before I came here were based indeed on other sources than censuses and slightly different methods, different methods, analytical methods than the methods I have been using here.

Q. All right. That is what I wished to clarify. But before The Hague, did you ever deal with ethnodemographic matters?

A. I didn't personally, but demography of ethnic group is a dynamically growing field of demography and many colleagues from the institute where I worked before I came to the Tribunal did this type of ²⁷¹²⁷ research. So it is not a new subject to me. And after all, what is demography of ethnic group? It is just studying the population by ethnicity and making comparisons of different distributions for different areas and different time periods or looking at historical trends. It is not such a complicated issue that you wouldn't be able to learn within a relatively short time.

Q. All right. Well, tell me then, did you ever publish a single paper before The Hague which looks into demographic matters in relation to ethnic characteristics?

A. No, I didn't.

Q. Very well. And do you know which questions in Yugoslav censuses, including Bosnia and Herzegovina censuses, are considered to be significant for determining ethnicity, ethnic affiliation?

A. Well, it is just one single question that is asked in censuses about ethnicity, and this is an open-ended question in which respondents are expected to identify themselves in terms of an ethnic affiliation. But I am aware of studies where ethnicity was measured in a different way, still based on census date. Ethnicity was a constructed variable for which not only self-reported census response was taken as part of the variable but also the language and religion, for instance. Well, for me personally, and there are many people like me, ethnicity is a self-perception of a person. So the open-ended question from the census in which people identify themselves in terms of one ethnic group is the best way of measuring ethnicity of a person. Any other variable, constructed variable, is not appealing to me. It is just an ²⁷¹²⁸ artificial variable that is measuring something, but I don't understand exactly what is the variable measuring.

Q. All right. In view of the fact that you mentioned yourself that in addition to this ethnic affiliation that is put as a question and then the person completing the questionnaire says himself what his perception is, that you also refer to mother tongue, religion, et cetera. And according to the information I know, this was part of the Yugoslav censuses in 1948, 1953, 1961, 1971, 1981, and religion was in 1953 and in 1981. So did you do your expertise taking into consideration a combination of all three of these characteristics that determine ethnic affiliation?

A. I didn't, and I did it on purpose. This is because I believe that the best measure of ethnicity is just the self-perception. The single question about ethnicity, not a variable, constructed variable that would be based on other items like religion and mother tongue, next to ethnicity, self-reported ethnicity.

Q. All right. Tell me, Ms. Tabeau, how did you deal with Yugoslavs? How did you categorise them? There were 239.857 of them according to the official figures published in the census from 1991.

A. I already explained this today, but I will repeat. The Yugoslavs were taken under the term "Others," and not only Yugoslavs were covered by the term "Others" but also all ethnic groups reported in that census that were different than the three specific groups, Serbs, Muslims, and Croats. So it is a kind of reminder. All other ethnic groups are taken together under the name "Others," including the Yugoslavs. ²⁷¹²⁹

Q. So you included Yugoslavs among "Others."

JUDGE MAY: She just said that. You just said that. Yes.

THE WITNESS: Yes, I did.

MR. MILOSEVIC: [Interpretation]

Q. All right. We'll go back to that later. And tell me, what did you use as your main source of information for your analysis?

A. I used population census from 1991. I used two voters registers from 1997 and 1998 merged together as just one source. And we also used the official Registration of Displaced Persons and Refugees that was obtained from the government of Bosnia and Herzegovina. I believe the last source also exists in Serbia, Montenegro, Croatia. These are just official registers that are used for the purpose of various benefits the internally displaced and refugees are entitled to.

Q. If I understand this correctly, your main sources were the census of 1991, which was a comprehensive one, and it is no doubt reliable data as such, and you also have the registers of the OSCE from 1997 and 1998. What was the main methodological problem for using jointly, in a way, these rather unequal sources in terms of reliability?

A. I don't know exactly what you mean. If you could reformulate your question. Do you --

Q. On the one hand you have a population census, the one from 1991, and it is a comprehensive one, and it was done in accordance with -- it was done lege artis, if we can put it that way. On the other hand, you have, by way of comparison, voters registers of the OSCE, including the citizens of age who voted, who voted in the elections. This is, no doubt, ²⁷¹³⁰ a piece of information that cannot be nearly as reliable as the census can. So that's why I'm asking you, what was your main methodological problem in terms of relying on these two sources that are so different in terms of comprehensiveness and reliability?

A. I believe that these two sources are indeed very different, but at the same time, there are many similarities of these two sources. The census is a complete, say, comprehensive source of information about the 1991 population. The voters register is just a sample, but the voters register was established on the basis of census records. The complete lists of all eligible voters came from the 1991 census. The personal identification numbers included in the voters register are exactly same as in the census. So it is, after all, not so difficult to work with these two sources together, because part of information -- of the information in these two sources was just overlapping. That part related to personal details, including the identification number, names, father name, date of birth. And the main methodological problem was of course to link these two sources once matching have been achieved, satisfactory matching rate has been achieved. Then many analyses became possible.

Q. Well, that is precisely why I put the question, because this matching seems quite troublesome to me if one bears in mind the fact that the census deals with the permanent population throughout the years, including workers temporarily employed abroad and their family members staying with them and so on and so forth; whereas, the OSCE registers include only persons of age who volunteered to vote in the elections. Is there any data concerning ethnic affiliation in the voters ²⁷¹³¹ registers of the OSCE?

A. No. There is no information about ethnicity in the voters register, but the ethnicity of voters can be taken for every record matched with the census from the census record of the person. And this is what we have done and what was actually the best way to avoid any bias in ethnicity as reported after the war.

Q. Did I understand you correctly, Ms. Tabeau, that actually this matching, as you refer to it, means to take a person from the census of 1991 and according to definite criteria find a matching person in 1997, 1998? Would that be right?

A. In principle that's the way to do it, but I would think of taking a person from the voters register and looking for the person from the census as well. It can be done both ways. But in principle, it is that thing.

Q. Yes. But for the most part, you said here that your main source was the personal identification number; is that right?

A. I said that the personal identification numbers were included in the criteria for matching, and this is indeed the most important part of criterion -- of a criterion, the personal identification number.

Q. That is what I wished to hear, that the personal identification number is the most important part of a criterion. And do you know what the percentage number was of wrong personal identification numbers in the population census of 1991 and in the voters registers of 1997, 1998 of the OSCE? Do you know what the margin was?

A. Well, I don't remember exactly the number. We can provide you ²⁷¹³² with specific statistics on that. But I never said that the personal identification numbers were perfect. On the contrary, I mentioned during this testimony that personal identification numbers were frequently deficient, and this is why while matching we had to use many more items as part of every criterion than the personal identification number only. So we used also names, first name, father's name, family name, and other items in matching criteria.

Q. Well, according to the information our demographers have, the percentage of wrong personal identification numbers in the census of 1991 and in the voters registers of 1997, 1998 ranges between 30 to 40 per cent. Is that also your estimate, Ms. Tabeau, or not?

A. It depends how you look at the personal identification number. This number contains several components. One of them is just a record for date of birth and place of residence at the moment of census in terms of a region of residence. There is also a check number included in this personal identification number, et cetera.

The most important part of the number is the date of birth and region of residence, and this part of JMBG was certainly correct for than more than what you mentioned in your statistics here.

Q. Tell me, as you apply this method of matching information from the census of 1991 and the OSCE voters registers, did you do any other pieces of expertise?

A. I don't exactly understand the question. If you could reformulate, please.

Q. I mean apart from this matching, as you call it, and comparisons ²⁷¹³³ between data such as personal identification numbers from 1991 and 1997/8, that is to say, the census and the voters registers, did you -- did you work on any other expert opinions?

A. Well, it is that you are asking about other sources, external sources, figures produced by others on the ethnic composition and displaced persons and refugees?

Q. I saw some information somewhere, for example, that you worked in the cases of Samac, Odzak, et cetera, that you looked at other data as well and produced expert opinions.

A. Well, it is also in this report that we included an additional source, additional to the census and voters analysis. We analysed data from the official register of displaced persons and refugees received from the Bosnian government. So my answer is yes, I did use additional sources in my report, additional to the census/voters analysis, but not as many as you would expect. The principle we followed in this report and we try to follow also in other projects is to work with primary sources and includes statistics that can be assessed from the level of individual records that can be studied. All the sources I included were the primary sources, individual records that could be assessed through analysis.

Q. All right. As you applied these same sources, that is to say, the census, the OSCE register, and the matching method, did you reach such results on the basis of which you came to the conclusion that this was ethnic cleansing, that this involved ethic cleansing?

A. Well, I showed the scale of changes in the ethnic composition and presented numbers of internally displaced persons and refugees and tried ^{27134 BLANK PAGE 27135} to estimate the real extent of the displacement process and process of becoming a refugee. So I -- in my view, in many instances you can think of ethnic cleansing because of the dramatic changes in ethnic composition, but I don't think I put it on paper in my report as one of the major conclusions.

Q. And how is the notion of ethnic cleansing defined in demography? I mean, in demography as a science related to the population.

A. I wouldn't think of a formal definition of ethnic cleansing. Such a thing doesn't exist in demography. I don't know, probably not in other science disciplines.

Q. That's what I think too. I think it doesn't exist in demography. As far as I know, in demography there are some changes of ethnic structure, things like that, but --

JUDGE MAY: We're not going anywhere at the moment. There's no argument about this. It's not a term of demography, it's a general term which is used. What you can ask the witness about is what's in her report, and she doesn't use that term, as I recollect.

THE ACCUSED: [Interpretation] All right.

MR. MILOSEVIC: [Interpretation]

Q. As you used the census from 1991 and the OSCE registers from 1997, 1998, were you able to receive data about the causes that led to changes in ethnic structure and pattern?

A. I didn't study the causes that led to the changes presented in the report, and I never mentioned that the intention of this report was studying the causes. ²⁷¹³⁶

Q. All right. That is very important. This should be borne in mind. And tell me, on the basis of available sources, are you in a position at all to give very accurate information about the changes taking place between 1991 and 1995 as the war went on or, rather, to establish that these changes did not take place between 1991 and 1992 or between 1991 and 1997, 1998? I'm trying to say before or after the war.

A. I think during the war or after the war.

Q. Yes. On the basis of the sources you have available, can you place these changes within a time frame, between 1992 and 1995 when the war was on, or, rather, can you claim that they did not happen, say, between 1991, 1992 or between 1995, 1997, 1998, for instance?

A. Well, these sources I studied summarised the status as of 1991 on one hand and 1997/8 on the other hand. And indeed, no data were used from the period in between to show changes year by year or month by month even, not at all. However, I would be very, very -- it would be very unlikely that such huge changes in ethnic composition would occur in the years after the war as all written reports I saw regarding the years 1996/7 and 8 were related to the problem of returns of the huge numbers of refugees and displaced persons home, not to mention the reports of UNHCR on refugees and internally displaced persons during the war. There are many reports of UNHCR presenting summary statistics. I don't refer to these reports because I couldn't study the individual data underlying the statistics presented in these reports. This was not our purpose to refer to summary statistics that I cannot assess at all. But I believe that there is a lot of truth in these UNHCR ²⁷¹³⁷ statistics. Our figures compared to the figures presented by UNHCR for 1997 are very extremely small, but this is what we can prove. These are minimum at least numbers that we can present in terms of individual records. So we can be sure that our numbers are the numbers of real people who became displaced or refugees and were still displaced and refugees as of 1997/8.

We didn't count returns, for instance. There were not many returns after the war until the elections in 1997/8, but if there were any returns, there were a few, a few, especially of minority groups. The returns are not shown in our statistics of displaced -- as displaced or refugees because those who returned were found in our data at home, simply.

So I wouldn't believe, to answer your question, that such dramatic changes in ethnic composition and in becoming a displaced person or refugee could take simply place after the war.

Q. Well, let me give you example that speaks to the contrary. You certainly know the dramatic change that occurred with the moving out of the Serb population from Sarajevo after Dayton. You are familiar with that, I'm sure.

A. Well, according to the data on Sarajevo, several municipalities of Sarajevo were split between Republika Srpska and the Federation and there was a flow of Serb population into the RS parts of Sarajevo.

Q. According to your criteria, can that be considered a radical dramatic change or not? And it occurred after Dayton.

A. Probably it would be shown as a radical change indeed, but even ²⁷¹³⁸ though it occurred after the Dayton peace agreement, it was a result of this conflict and what happened during the conflict that those Serbs actually moved out of federal part of Sarajevo and moved into the RS parts of the country.

Q. A similar event, similar development occurred among the Muslims who moved to the territory of the Federation. And let me digress for a moment. On page 16 of your report, having analysed some data, it's at the very end. There's a paragraph which says: "Summing up. Even though in relative terms the changes in Republika Srpska and the Federation of Bosnia-Herzegovina, parts of the Milosevic case area were not different for the Muslims." You say, "Minus 95.5 per cent in Republika Srpska and for the Serbs, minus 88.2 per cent in the Federation." So the figures are quite close. The same order of magnitude. Isn't that right, madam?

A. If measures are indeed are, this is what I indeed wrote. But I further concluded that in absolute terms, the losses of the Muslim population were much higher than the losses of the Serb population in the Milosevic case area.

Q. Very well. I understand that, Madam Tabeau. But as a demographer and a statistician, you know very well that the relative figures are decisive. The Muslims were more numerous generally. So I assume that could be one of the explanations why there were more victims on their side. I am talking about the relative figures which are close, even for this very narrow focused area of seven municipalities selected by the Prosecution. Isn't that right, Madam Tabeau?

A. In relative terms, there is quite some similarity between these ²⁷¹³⁹ two numbers you referred to. But at the same time, I --

Q. What you referred to.

A. I preferred to the two numbers 95.5 per cent decline of Muslims in Republika Srpska and 88.2 per cent decline of Serbs in the Federation area of the Milosevic case area. It is all about, you know, the two parts, two political entities of the Milosevic case area. And indeed these two numbers are mentioned in my report on page, I think 15, not 16.

JUDGE KWON: English.

THE WITNESS: In the English version on page 15, indeed, perhaps in B/C/S version, page 16. As I said in the beginning, we analysed changes of ethnic composition and showed all ethnic groups, Muslims, Serbs, Croats, and others. It is a report that is fair in the sense that we used the same data sources and the same methodology to estimate consequences for every ethnic group. I am not saying that the Serbs didn't suffer during the war or didn't become displaced or refugees or in certain municipalities Serbs -- fraction of Serbs declined and fraction of Muslims increased instead. I'm not saying this. I have shown all these changes for all ethnic group. But in my view, it is the Muslim population who suffered most, and it is both in absolute terms and in relative terms too. This is the major conclusion of this report.

Q. Very well, Ms. Tabeau. Let us leave aside the Milosevic case area, because I really do not know what these seven or 47 municipalities have to do with the Milosevic case. But let's see what this small segment that you have shown and that was assigned to you by the other side, we can find almost the same level of casualties in relative terms, and what would ²⁷¹⁴⁰ the situation be if you had taken municipality with a purely Serbian population such as Grahovo, Drvar, Glamoc, Petrovac, where the population was 99 per cent Serbs. Wouldn't the situation be far more dramatic than what you have shown in this case, applying the statistics and methods that you applied for this report?

A. I believe it is a highly hypothetical conversation. You please just ask me for statistics for a municipality or number of municipalities, and I will just present you with the figures.

Q. This is not hypothetical, Madam Tabeau. If in Grahovo 99 per cent were Serbs and now there's not a single one left, then those figures must surely be dramatic.

Let us look at your 167th page, attachments A63, the Federation of Bosnia-Herzegovina, Brcko. On the tables, the changes in the Muslims, population plus 83.5 per cent, 1991 to 1997. And Serbs, minus 97.8 per cent. So that's the whole trick in the choice of the points being presented here.

JUDGE MAY: I'm not following the question. You're putting, are you, that this is a selective use of statistics? Is that what you're trying to put?

THE ACCUSED: [Interpretation] Absolutely so. I'm not trying to say that it's selective, but what I'm saying is that the opposite side has made a malicious manipulation of statistics.

JUDGE MAY: Yes. We've heard these kind of allegations before. That's not for the witness to answer, allegations of that sort against the Prosecution. The witness is here to give evidence about the study and the ²⁷¹⁴¹ report which she's made.

Now, you can ask questions about that, and indeed you can call evidence in due course, if you want, to support what you say happened, but we're not going to get much further with this sort of argument.

THE ACCUSED: [Interpretation] Please don't worry that I will not produce such evidence.

MR. MILOSEVIC: [Interpretation]

Q. Look, please, at page 167 of your report, ERN 0308081.

JUDGE MAY: We'll find it in the English. What's it supposed to be showing? Which figure is it?

THE ACCUSED: [Interpretation] It is showing an opposite instance, but I want to use it as an illustration. It is annex A6.3. A6.3, Brcko-Rahic, the Federation of Bosnia and Herzegovina. So let's look and see what happens, Your Honour.

JUDGE MAY: Yes, Mr. Groome?

MR. GROOME: That's page 149.

THE ACCUSED: [Interpretation] In my version it's 167.

JUDGE MAY: We've got different versions, that's why we need to get them right.

JUDGE KWON: If you can give the number of figure or table.

THE ACCUSED: [Interpretation] The table is A6.3, figure 1. Table 1, I'm sorry. A6.3, table 1, Ethnic Composition in Rahic in the Federation of Bosnia and Herzegovina in 1991 versus 1997.

JUDGE KWON: 154, page.

MR. MILOSEVIC: [Interpretation] ²⁷¹⁴²

Q. Please look at the change. The last horizontal column, "Change." For the Muslims, it's plus 83.5, Croats minus 58.1, for the Serbs minus 97.8.

THE WITNESS: It is page 149, table --

THE INTERPRETER: Microphone.

JUDGE MAY: Yes, Mr. Milosevic, what is the point?

THE ACCUSED: [Interpretation] The point is, Mr. May, that these phantom seven municipalities were taken out of context. And even in them we see that the level of suffering in relative terms is roughly the same, even using such highly selected cases for the demographer to study, selected by the opposite side.

JUDGE MAY: I misspoke: What is the question for the witness?

MR. MILOSEVIC: [Interpretation]

Q. Is it clear that a selection has been done in such a way so as to satanise the Serb side, Madam Tabeau, to demonise them?

A. I --

JUDGE MAY: If the answer is no, just say no.

THE WITNESS: No. The answer is no. No intention of demonising any ethnic group.

MR. MILOSEVIC: [Interpretation]

Q. To go back to the fact that in Bosnia and Herzegovina, according to the official census, there were 1.905.274 Muslims, and in the seven municipalities of yours, assigned to you by the side opposite for analysis, you have 212.811 members of Muslim ethnicity, which means 12 per cent of the total Muslim population in Bosnia and Herzegovina. ^{27143 BLANK PAGE 27144} Now, tell me, please, what do these seven selectively chosen municipalities in which before the war and according to the 1991 census 12 per cent of the Muslim population lived proves in relation to the total population and the 109 municipalities that existed in Bosnia and Herzegovina?

Do you consider this to be a methodologically permissible extraction of elements out of context?

A. Well, the answer to the first question, per cent of the Muslim population in seven selected versus percentage of the Muslim -- versus the entire Muslim population in 109 municipalities, it wasn't studied in our report, so it is just a different story, I would say. Methodologically --

Q. I think it's a different story too.

A. A methodologically permissible extraction of elements out of context, I don't think we have done this. We just made a case study for seven selected municipalities as an illustration of most affected municipalities, and that's it. But at the same time, we presented complete statistics for the whole area, 47 municipalities constituting the Milosevic case area. And this is certainly permissible.

JUDGE MAY: It's now time to adjourn. Mr. Milosevic, you can have three-quarters of an hour more to cross-examine this witness and then we'll hear from the amicus.

We will adjourn now.

--- Recess taken at 12.18 p.m.

--- On resuming at 12.42 p.m. ²⁷¹⁴⁵

JUDGE MAY: I should make it plain there has been consultation with the medical officer, the point about the accused's health having been raised, and there is no reason why the hearings cannot continue. Yes, Mr. Milosevic.

THE ACCUSED: [Interpretation] Mr. May, I first have to make an objection regarding the time. You have given me as much time as Mr. Groome had, and this is an expert witness who is not testifying only viva voce but also through the rather voluminous material given through her so that the two cannot compare.

JUDGE MAY: We have that in mind. In fact, you will have longer than the Prosecution. We've timed that. You will have longer than them. Now, you should make the best use that you can of the time which remains.

THE ACCUSED: [Interpretation] Very well.

MR. MILOSEVIC: [Interpretation]

Q. Madam Tabeau, a moment ago you said that these were the most affected municipalities, and that is how you studied them or, rather, the side opposite selected them.

According to your data, Bijeljina, Bratunac, Brcko, Foca, et cetera, these municipalities, were they not more affected in terms of the consequences that you are analysing than, for instance, Grahovo, Petrovac, Drvar and other municipalities in which 99 per cent of the population used to be Serb? On the basis of your own information, is that what you are claiming, that these were worse affected?

A. I said the seven municipalities selected for the case studies are representation of the most affected municipalities. I didn't say that ²⁷¹⁴⁶ these are the most affected municipalities. It is just an illustration of what happened in the worst case scenario.

Q. You haven't answered my question. If we were to compare them with Grahovo, Drvar, Petrovac, for instance, what would your finding then be in terms of the figures that you have at your disposal?

A. We would have to take a closer look at these municipalities, and the data are available in tables -- in Annex 1, I guess. So -- so for Bosanski Petrovac, I think it is only Bosanski Petrovac that is located in the Milosevic case area. I don't see Grahovo and Drvar as part of the indictment area.

Q. That's where the problem lies.

A. You are saying that I didn't include --

JUDGE MAY: That's a comment, Mr. Milosevic. We know what it is that the witness has sought to set out to do and what she has done. Now, you can ask her about her evidence and her report, but I don't think there is very much point in going on to ask about speculation about other municipalities. I made the point and I make it again that if you wish to bring evidence about those other municipalities, of course you can in due course, but there is no point arguing with the witness about it when she hasn't made a study.

Mr. Groome.

MR. GROOME: Your Honour, I'm going to point out that I'm not sure whether it is a misunderstanding on Mr. Milosevic's part or a misrepresentation, but the report wasn't a generalised demographic study of Bosnia. Mr. Milosevic was charged with crimes originally in 47 ²⁷¹⁴⁷ municipalities and Mrs. Tabeau was charged with studying the demographic effects of those 47 municipalities, not on unrelated municipalities to the indictment.

THE ACCUSED: [Interpretation] Very well.

MR. MILOSEVIC: [Interpretation]

Q. Towards the end of the examination-in-chief, Mr. Groome put several tables or reviews to you. There were several towns, the municipality of Kalinovik, Gacko, Mostar, Foca. He showed you those. Tell me, please, what about Mostar? Where are the Serbs from Mostar? This was on the ELMO.

A. Yes, I know. I know. But we didn't study Mostar, not in this report.

Q. But Mr. Groome mentioned it a moment ago. He mentioned Mostar when he placed it on the ELMO. Maybe he misspoke.

A. Well, we can look again at the maps, but I think that Mostar is not part of the indictment area, at least in terms of 47 municipalities. It is not. If you look at the copies of the maps that you have in front of you, then you will see it.

Q. Very well. So that only illustrates the selective nature of this review. You mentioned Gacko. Now, tell me, please, what was the percentage of Muslims living in the Gacko municipality according to the 1991 census?

A. We can check this in the publication of census results made in Croatia. In this publication, the results of that census are mentioned for each municipality. This is how it is done for Gacko. ²⁷¹⁴⁸ We don't see the percentage, but we see the number of Muslims, 3.858 in Gacko in 1991. And now we can go to table 1M in our annex. In our study, 3.014 Muslims are reported for 1991. It is page 70 of our report. Page 70 of our report, table 1M. The 1991 population of Muslims in Gacko is 3.014.

Q. And what was the total population of Gacko in 1991?

A. The total population as reported in that census, based on this publication, is 10.788.

Q. So the Muslims --

A. So the Muslims would be approximately estimated in our report 34 per cent. But this is a percentage from our report that was made for those eligible to vote in 1978, so born before 1980. But it would be more or less the same percentage. Thirty-four per cent in 1991.

Q. So 34 per cent in all in 1991. I have a lower figure, but never mind. And then you say out of those 34 per cent, a large percentage left Gacko; is that right?

A. Probably, yes, but in order to make sure that I said so, we go to table 1.2 -- sorry, 2M in our report. 2M, page 86 in the English version of the report. For Gacko, the number of internally displaced and refugees as of 1978 is 2.034. The minimum number. This is the minimum at least number of Muslim IDPs and refugees from Gacko.

Q. And what is, for you, a refugee? Does that include everyone who left his home?

A. A refugee is a person who left the country, Bosnia and Herzegovina, and moved into another country in the region of the former ²⁷¹⁴⁹ Yugoslavia or an entirely different region.

I must note here that the refugee component is relatively weakly represented in our analysis. This means that in the sources that we used, only 300 -- approximately 325.000 records for 1997/8 were the refugees, so people who were registered to vote in other countries, most of them in the region of the former Yugoslavia, in Croatia and Serbia. We did not study records of refugees registered in major receiving countries like Austria, Switzerland, Sweden, Germany, of course. These are hundreds of thousands records that we were unable to collect and to study at the individual level. This is perhaps why our numbers of refugees and IDPs shown jointly are so relatively low.

Q. Did you, when talking about refugees, you're talking about refugees of all ethnicities or only about Muslim refugees?

A. I'm talking about refugees of all ethnicities, but the records that we had at our disposal were mainly the refugees of Serbian and Croatian ethnicity, Bosnian Serbs and Bosnian Croats, I would say. Because those two ethnic groups moved to countries in the region, Serbs -- Bosnian Serbs to Serbia and Montenegro, Bosnian Croats to Croatia. Muslims actually moved to Western countries, countries of Western Europe, like Germany, Switzerland, Austria. And the records that we had on refugees, these were out-of-country voters, 325.000 in total, 325.000, it's a very small representation of the universe of all refugees from Bosnia and Herzegovina, but the Serbs, Bosnian Serbs, are included in those records.

Q. Very well. But before 1997, do you know that, for instance, in ²⁷¹⁵⁰ Serbia there were more than 70.000 registered Muslim refugees from Bosnia-Herzegovina?

A. I don't -- I'm not aware of this particular number, but I believe it is perhaps the case as Serbia was a receiving country for very many refugees from Bosnia and Herzegovina, but mainly Bosnian Serbs.

Q. I'm mentioning 70.000 Muslims, and there were half a million refugees from Bosnia-Herzegovina in Serbia. Are you aware of that figure?

A. Well, I am aware of the existence of the registration of internally displaced persons and refugees in Serbia and in Montenegro. I know that Serbia and Montenegro established registers, official registers of persons internally displaced and refugees in Serbia and Montenegro already in 1992, and I know that there are official databases that contain these type of records.

Q. I have here, due to circumstances -- this is nothing to do with Serbia, it's a questionnaire. It says Republic of Bosnia and Herzegovina, ERN 03032112. It's a questionnaire for registering households in the free areas of the city of Sarajevo: The present address Novi Grad; the local commune, Dobrinja 1; there's a street number, a street and a house number; and then it says, question number 4: "Members of the family household who have fled or who have moved outside Sarajevo or who have remained within the territory of the Republic of Bosnia and Herzegovina which is under the control of the aggressor." And then there is a lady here. Where is she now, and the answer is: "With her son." When a woman goes to stay with her son, do you treat her as a refugee as well? You can have a look at this questionnaire. I told you what the ERN number is. It's your ^{27151 BLANK PAGE 27152} document, 030321102. So I don't have to give it to you, I think.

A. Well, in order to answer this question, I would have to search for this woman and her son and check in our databases, especially voters, for details regarding this particular person.

But the questionnaire you mentioned, it's the questionnaire from household survey Sarajevo 1994 that we used for entirely different purposes than counting the displaced -- internally displaced persons and refugees. There was such a question in this questionnaire, but from this questionnaire we only have taken killed, natural deaths, and wounded persons and didn't collect information on internally displaced and refugees.

It is -- I'm aware of the fact that there are more sources that could have been used for presenting a better estimate of internally displaced persons and refugees. I am aware of this. But due to time constraints and resources, limited resources that we had at our disposal, we have done what we have done.

I would love, for instance, to analyse the statistics from the Serbian register of internally displaced persons and refugees, and I requested this database in March this year, but so far we haven't received any response yet from Belgrade. So it is a matter that collecting certain information is a very time-consuming and difficult process. If I'm thinking of collecting records of refugees from Western countries, Western Europe, at least five, seven major countries would have to be approached regarding these records, and I would have to apply for permission to obtain individual records with every single country separately and this ²⁷¹⁵³ would have to go through every government of these countries and it all costs a lot of time. So certain things in principle can be done, but cannot be done within a limited period and having the limited resources at our disposal. So these are -- these are the reasons that not all existing sources have been studied in our report.

Q. Tell me, please, from the territorial point of view, the municipalities you looked at in 1997, 1998, do they coincide with the territorial levels of the municipalities from 1991?

A. Well, in fact in our analysis, they -- they do. We statistically analysed the same type of municipalities as defined after the Dayton peace agreement. These are the so-called new municipalities. But the new and the pre-war municipalities are not all identical. There is a number of municipalities, new municipalities that were introduced after the war, and some municipalities have a different territory before and after the war.

Q. All right. But since the territories of the municipalities do not coincide when one compares 1991 and 1997, what was the percentage of the population of 1991 that was no longer there?

A. I think I should clarify here that in order to be able to study exactly the same areas before and after the war, we had to restructure the records from the census according to the new administrative division of Bosnia and Herzegovina. The new division introduced after the Dayton peace agreement.

This means that for every individual reported in that census, a new code was created reporting the place of residence of this person as a new post-Dayton municipality. ²⁷¹⁵⁴ In principle, for the vast majority of individuals, this was easily possible. However, it -- there was also a number of individuals for which we couldn't do that because the individuals used to have place of residence in 1991 census as reported as a settlement. Settlement is just a smaller area within a bigger area that is municipality. So each municipality consists of a number of settlements. A number of settlements have been split by the Dayton line between the Federation and the RS, and there are no statistical records of recalculating for split settlements the place of residence in terms of either of the split municipalities, RS or FBH.

So simply speaking, for those individuals who were reported in split settlements in that census, we were unable to present a post-Dayton municipality. The number of such individuals is not large for Bosnia and Herzegovina. If I recall it correctly, it was approximately 150.000 individuals from split settlements, but I -- I would have to be -- in order to be sure, I would have to check this, the exact figure. These individuals had to be excluded from our study, and we have done so.

Does that answer your question, Mr. Milosevic?

Q. Well, my question was what the percentage was of the population that was practically expelled from this comparison because of the differences between the territories of the respective municipalities in 1991 and 1997.

A. For me it would be the answer. The individuals excluded from our analysis due to difficulties in assigning them post-Dayton municipality ²⁷¹⁵⁵ code. Perhaps I misunderstood the question. I don't know.

Q. Since you say that this has to do with about 150.000 persons because of these shifts in borders, is it correct that in Sarajevo, according to the census in 1991, that is to say before the war, there was a total of 185.000 Serbs? Is that piece of information correct?

A. We can check this by looking into the big Sarajevo report, Galic report. There is a table there. On page, I believe, 25 in the English version of the report, there is table 4.

So here in this table we have numbers of individuals for each ethnic group for the territory of six urban municipalities of Sarajevo. The number of Serbs here is 126.113.

Q. That pertains to four municipalities only; is that right?

A. Six municipalities. To six municipalities. But if you take ten municipalities, then of course the number can be as you mentioned. I don't have statistics for ten municipalities in front of me.

Q. But since now in these municipalities, this includes Pale, Trnovo, there is less than 30.000. That means that these 150.000 could have been accounted for by Sarajevo only. What about the rest of Bosnia-Herzegovina and the said municipalities that you referred to who are outside the comparisons made between 1991 and 1997 therefore?

A. I must say I'm confused now. I would like you to reformulate your question.

Q. I am going to reformulate it. You mentioned the figure of 150.000 who somehow got excluded from the comparison. As for this 150.000, there are only 150.000 Serbs that are missing from Sarajevo, and where are the ²⁷¹⁵⁶ remaining tens of thousands of people from the seven municipalities or the 47 municipalities that you could not compare?

A. What do you mean by missing from Sarajevo? In what table are the Serbs missing from Sarajevo in my report? You know, reading statistics is --

Q. I'm not saying -- your report includes the census. You mentioned the 1991 census, which was done lege artis. According to that census in the entirety of the city of Sarajevo there was a total of 185.000 Serbs and now there are only 30.000 left or less than 30.000. Where are the remaining 150.000? Is that right?

JUDGE MAY: Just see if the witness can deal with the question. It may be she can't. That's the question, Dr. Tabeau.

THE WITNESS: I'm afraid I don't understand the question. In my view, it is good to remember that our report deals with the population eligible to vote in 1997, 1998. This means that from that census, we only included records of those born before 1980. These statistics for -- cannot be taken literally as the numbers of the complete population in the municipalities studied. It is just a smaller population. So if there is any Serbs missing or Muslims or Croats or whatever. So first of all, we have to think of those not included, or those excluded due to the design of this study. And you cannot compare our numbers for 1991 with official statistics published by statistical authorities because these numbers will be different then.

MR. MILOSEVIC: [Interpretation]

Q. All right, Ms. Tabeau. Let's not waste any time. You gave a map, ²⁷¹⁵⁷ and on that map you provided municipalities in their entirety, and then, depending on the proportion of the population involved, you used the colours red, green, et cetera; is that right? And then that conveys a certain picture, doesn't it? I hope you will agree that a more accurate picture would have been given by a survey of municipalities also using these colours if it were done according to individual neighbourhoods, settlements, not municipalities as a whole, because municipalities as a whole can give a somewhat different picture. So if it were done in greater detail according to individual settlements or neighbourhoods; is that right? So I have here a map that was made at the time of the census in 1991. It has a legend in both the Serbian and English languages. It was at the faculty of geography. According to settlements, Serbs are blue, Muslims green, the rest are red. Croats are orange, then grey is without an overwhelming ethnicity, and then the inaccessible areas or whatever are yellow but that's a negligible amount. So could you please have this map put on the ELMO and please note how different it is from the picture that you portray on the basis of the administrative borders of the municipalities.

THE ACCUSED: [Interpretation] On my monitor, one cannot see a thing. Well, now perhaps. But could I please ask the usher to put the map sort of -- to flatten it out. That would make it easier because there is a shadow in the middle. Perhaps you could use a ruler to flatten it out so there wouldn't be this shadow so that the map can be seen properly. Now it can be seen properly.

MR. MILOSEVIC: [Interpretation]. ²⁷¹⁵⁸

Q. So can you see how different this picture is, the one that was done by the faculty of geography? It looks like a mosaic. It was done according to settlements. See how different it is from the map you made according to municipalities where a large concentration in individual town or big city can completely change or distort the actual picture that can be seen here.

A. Well, first of all, I agree that such maps can be made at different levels. We use municipalities, and this is how ethnic majority maps are made in many, many other organisations active in the area of the Bosnian conflict, Yugoslav conflict and other conflicts. UNHCR, for instance, shows municipalities; international monitoring group that is looking at property and damaged -- damages of property; German government who is particularly interested and active in the population issues related to the conflict in Bosnia and Herzegovina; they all prepared maps of ethnic majority at a certain point and used municipalities to show changes in the ethnic composition.

The -- first of all, I don't agree that the picture is so different in this map and our map. If you look carefully at municipalities like Prijedor, for instance, [indicates], that is -- yes. Thank you. This area. Then you see two colours there, not only blue that depicts the Serbs but also green that depicts the Muslims. And in our map, the municipality of Prijedor was shown as a brown municipality, a municipality with a mixed ethnic composition of Muslims and Serbs as two dominant ethnic groups.

If you look carefully at the municipalities located at the eastern ²⁷¹⁵⁹ border of Bosnia with Serbia, then again you see not only one colour, the blue colour depicting the Serbs, but a mixture of the two, blue and green, and green is the colour of Muslims. So it is not necessarily that the map you showed us is so different from the map we presented at the municipal level, it is a matter of how you read the map, and this is the key issue for drawing conclusions, I think.

Q. Well, this, Ms. Tabeau, shows the ethnic pattern according to territory regardless of administrative borders of municipalities. So administrative borders of municipalities were not taken into account when calculating how many people there were but, rather, the distribution of Serbs, Croats, Muslims, et cetera, in various territories. And as you can see, this blue territory that we can see on the screens, these are the Serbs who carried out an aggression against their own land. That's the picture according to the 1991 census.

So they carried out an aggression against their own country, and that is what is referred to in your report, in these lists, aggression and so on.

JUDGE MAY: This is your interpretation of what the witness said. She cannot possibly answer any questions of this sort.

THE WITNESS: I only see that this map is depicting the ethnic composition in 1991. This would be the population census in 1991. It was published in 1992 in Belgrade, and the principles of this map are similar to the principle we used in our mapping projects. The difference is that colours are different and the settlements are used as a basic unit for the mapping, not municipalities. But for the rest, in my view, this map is ^{27160 BLANK PAGE 27161} very much in line with all maps I know for Bosnia, all maps of ethnic majority in that census.

MR. MILOSEVIC: [Interpretation]

Q. Very well. Now, you wrote your expert opinion entitled "Human Losses During the Siege of Sarajevo"; is that right?

A. We wrote several reports related to the losses -- population losses of the siege, casualties of the siege, yes.

Q. And what were the sources you used then?

A. I'm trying to get transcript. Yes.

Q. What were the sources, the data sources you used?

A. We used several sources. The number of sources we used altogether is eight. Three of the sources were lists of fallen soldiers, official records of fallen soldiers in the Bosnian war, and five of the sources were sources such as household survey Sarajevo, like official records of the Federal Institute for Statistics in Sarajevo, like ICRC list of missing persons, MAG - Muslims Against Genocide - mortality database, Bakije Funeral Home records.

Q. All right. This house hold survey that you refer to and that is very often mentioned in your report, this household survey, was it ever officially recognised and published anywhere else but in your report?

A. No, it wasn't, and I clearly state this in my report. This was a huge survey. Eighty-five thousand questionnaires were distributed and 85.000 interviews were taken during the census, or say this survey, don't call it a census, a population census, it wasn't a population census. And in total, information of about 340 approximately thousand individuals was ²⁷¹⁶² collected. It was such an incredible quantity of material, statistical material, that this research institute who conducted this survey has never been able to computerise this material and to analyse the data statistically.

We requested the material and computerised the records that we needed for the Galic report at OTP.

Q. All right. Now how this was done is what you were told. You don't actually know how this was done. You were told how this material was compiled; right? You did not take part in compiling this material.

A. We -- me personally and other colleagues from the demographic unit did not participate in the survey itself. It was in 1994, mid-1994 when the survey took place. It was rather impossible for us to participate. We learned about the survey after several visits to the research institute who conducted the survey, and we learned from the researchers from this institute that such a survey took place.

Q. All right. But in your opinion, since you're a demographer and since you are involved in scientific work, do you think that this approach is partial?

A. Well, I can make an assessment of the scientific quality of the questionnaire and how the survey was conducted on the basis of what I know about the survey and the questionnaire, and indeed this is one thing. I cannot say whether it was partial or impartial. I can say that the material was probably biased a little bit towards more frequently counting Muslim casualties than casualties from other ethnic groups. Why? Because people who conducted the survey and were engaged in the survey were mainly ²⁷¹⁶³ the Bosnian Muslims from Sarajevo. So therefore they could have had a little bit easier access to the respondents of the same ethnicity than as to respondents from other ethnic groups.

But the Sarajevo case is not a case of ethnic cleansing or anything like -- you know, we don't say anything about genocide in the ethnic composition, disappearance of certain ethnic groups from the territory of Sarajevo in our report. It is a report about casualties, casualties of the siege of Sarajevo. And the casualties were from all ethnic groups, Muslims, Serbs, Croats, others. All ethnic groups are there amongst the casualties.

So if you want me to make an assessment of the scientific quality of the survey and the questionnaire, I can do that. And it is, on the other hand, also done in the report itself.

Q. All right. But I assume you've noticed. I mean, I have 03031234 here, a questionnaire. Please take a look at this questionnaire. Most of the questions -- look at number 5. Most of the questions refer to some kind of aggression. So no impartiality can be referred to in this context. Those wounded during the aggression, those who went missing during the aggression, who were in prison during the aggression, newly born babies during the course of the aggression, then also stillborn babies during the course of the aggression.

I imagine that what the target group of this questionnaire is is quite obvious and what the approach is of those who organised this questionnaire and this poll, generally speaking. How can you rely on this as a reliable source? What is this ²⁷¹⁶⁴ aggression?

A. Well, the aggression, how would you feel being in Sarajevo with the Serb forces, armed forces standing there on the hills and shooting people, with snipers shooting people? It's enough to understand that it was war going there.

So first of all, to me, this means -- this term "aggression" refers to the war circumstances in Sarajevo at the time of the survey. This is one.

Secondly, by mid-1994, the war in Bosnia had been going already for a long time, as we all know, and it was -- it was clear that there were certain forces fighting with each other. Perhaps the term "aggressor" had a different meaning for Muslims and Serbs and for Croats, depending on the time, in addition, and location, but I don't think there is any doubt about that there was a conflict there by mid-1994.

Q. You saw the colours and the ethnic distribution and where people were. As far as I can see, the Serbs were around Sarajevo even according to the 1991 census. They didn't come there in 1992. That is quite obvious from the map that you looked at a moment ago.

THE ACCUSED: [Interpretation] Could the usher please give me back that map. I'll need it again.

MR. MILOSEVIC: [Interpretation]

Q. So is it your position that such a source of data such as this household survey can be considered a reliable source of data?

A. Well, we are speaking about a source that was used to estimate the number of casualties, that is deaths, in the siege of Sarajevo. ²⁷¹⁶⁵ I said earlier today that all sources, mortality sources, all lists of deaths, in other words, from the war period, and I am speaking about not only about the war in Bosnia but generally, all these sources are incomplete and deficient and are uncomparable with official records from death registration that normally takes place in every country in normal circumstances. But the normal registration system doesn't function during conflicts.

The war in Bosnia is an example of a human emergency situation which caused that normal structures, organisations, did not function in Bosnia. So you cannot expect that you will have, in order to estimate the number of casualties, that you will have at your disposal just regular statistics from vital events registration, that you just take them and you look at that records, you just add them up and then you come up with this one number with the estimate. This is not that simple. So the way we look and assess the sources for use in estimating war casualties in conflict times and under peace must be different. The sources are simply must be different because normal sources don't exist. Sarajevo household survey is an example of a source that certainly can be used as a source to measure casualties in a human emergency situation. There is no doubt about that. It is a large survey, conducted in a relatively thorough manner, prepared starting with the questionnaire, quite clearly designed with clearly formulated questions that can be later easily used in data processing and producing statistics. It is a survey that was prepared by the means of organising a training for interviewers. The interviewers were instructed how to deal with questions, and the ²⁷¹⁶⁶ respondents, and what was missing in the survey was that final step, the computerisation of the material, cleaning of the material, and analysis. And that is what we have done starting from the computerisation, and as such I believe that this source can be used to produce statistics as to ones we show for Sarajevo for General Galic case.

Q. Very well. Ms. Tabeau, could you please give briefer answers, because the length of answers cannot compensate for the absence of validity of these data.

Tell me, in the expert report for Sarajevo, was there an assessment of the population figures, the number of inhabitants?

A. Yes. We included an assessment, 340.000 individuals covered by the survey.

Q. Fine. In relation to your assessment regarding the total population, the mentioned 340.000, what is the share of registered Muslims as opposed to all the other ethnicities together?

A. Well, it is that we made an assessment of ethnicity in the survey population, but ethnicity is not available from the survey for every event like that or wounding. We actually assessed ethnicity through linking the records from the survey with the census. So ethnicity that we studied, we showed in the Galic report is just ethnicity as reported in that census. It is again the same table, table 4, on page 25, English version of the report, that shows our assessment of ethnicity for the casualties, for killed and wounded persons -- actually, only killed persons.

Q. I'm not asking you about the killed persons. As you yourself said, you made an assessment of the population at 340.000, will you tell ²⁷¹⁶⁷ me in relation to that total figure what was the share of registered Muslims as compared to all the other groups or ethnicities?

A. And my answer was that we -- we don't have --

Q. Your answer was that you don't know.

A. Yes. The answer is that we don't have the ethnic breakdown of the survey population from the household survey Sarajevo, 1994. We haven't used these numbers.

Q. Very well. I understand, Ms. Tabeau. But is it in dispute that the targeted group of this survey of households in Sarajevo was primarily Muslim households?

A. Not at all. Not at all. Ethnic targeting is not the subject of this report. It is -- the subject of this report is estimating the number of casualties irrespective of ethnicity. That we included ethnicity --

Q. No, no, Mrs. Tabeau. You haven't heard my question. I am asking you -- I'm not asking you about your report. I'm talking about the household survey for Sarajevo done in 1994. Is it disputed that the targeted group of that survey was Muslim households?

A. I would disagree, because I am not aware of any procedures or any activities conducted during the survey preparation and the survey itself that only the Muslim households would be primarily meant as the respondents for this survey. I don't think so, especially the ethnicity is available for every household as ethnicity of the head of household and this can be easily checked at least for the records that we have in our database. And I can assure you that all ethnic groups are represented there amongst the records that we have, the records of casualties. ²⁷¹⁶⁸

Q. Ms. Tabeau, how can you say that you can guarantee that they were represented when a moment ago in answer to my question as to the proportion of Muslims in the 340.000 inhabitants, you've said that you didn't know. I'm not denying that there may be a Serb or a Croat name among them.

JUDGE MAY: This is all argument. You've heard what the witness has said.

Now, you've had nearly ten minutes more, Mr. Milosevic, than you would have otherwise had. You've gone beyond your time by ten minutes. We'll give you another two minutes and then you must wind up.

THE ACCUSED: [Interpretation] I can't finish in two minutes, Mr. May, but that has become the customary practice here. It's not the first time.

MR. MILOSEVIC: [Interpretation]

Q. Tell me, please, Madam Tabeau, were all mortality rates in the report for Sarajevo expressed in relation to the figure 100.000 inhabitants?

A. Yes. This is how we show --

Q. Why? Now, why? Why were they shown in that way?

A. In order to make it possible to recalculate the rates. If you would be interest in daily rates, for instance. In order to make this possible relatively exactly.

Normally the rates in statistics, official statistics, are shown per 1.000 population. Rates for mortality by cause of death, especially for less frequent causes of death are shown in epidemiology per 100.000 of ^{27169 BLANK PAGE 27170} population, but the same rate can be easily recalculated from the number per 100 population to 1.000 population just by dividing by 100 every number, and this is still the same intensity, that intensity. So I don't think it is really so essential to have the rates shown. Per 10.000 population as the daily rates are shown or per 1.000 population as the annual rates are shown or per 100.000 population as the rates that we have shown in our report. These are still the same relative measures. The same intensity is shown by the rates.

Q. Those are the same relative rates. I'm sure you will agree, Mrs. Tabeau, for experts, but for a layman in statistics, it's not the same when you see the rate 548 to -- per 100.000 or 54.8 per 1.000. But judging by the rate, he has an impression of the scope. So unlike the customary practice, you used a hundred thousand because you get a bigger figure and it sounds better for the media. Wouldn't you agree with that, Ms. Tabeau?

A. That would be very naive of myself, I think, to believe that media don't read statistics.

In table 17 of page 42, we show ratios of the event rates for Sarajevo, September to August, September 1992 to August 1994. Ratios are just the rates compared with a baseline standard. The baseline standard used for our rates is the empirically observed deaths rates for the six Sarajevo municipalities in the period of 1990, 1991. This is exactly the same period two years long as the period of the siege studied in Galic, two years long. So we took a baseline and every rate was recalculated as a ratio that can be read immediately in terms of how much higher, how much ²⁷¹⁷¹ lower given that rate was compared with the baseline standard. And this is all in our report. So I don't think I wanted more publicity about the rates.

Q. Tell me, please, just a few points of explanation. On the basis of which assumptions do you reach the conclusion -- and this is on page 6, paragraph 3 of your report. On what grounds do you come to the conclusion that the pre-census emigration, the emigration would return? What -- on what grounds do you make such a conclusion?

A. To what are you pointing now?

Q. Your basic report --

JUDGE MAY: Which one?

MR. MILOSEVIC: [Interpretation]

Q. -- submitted here in this case.

JUDGE MAY: Which one? There are two reports. Which one are you referring to?

THE ACCUSED: [Interpretation] I think it's the first one.

MR. GROOME: Your Honour, I believe he's referring to the first report, and it's page 6, top of the page of that report.

THE WITNESS: The displacement report. The displacement report.

MR. MILOSEVIC: [Interpretation]

Q. Yes. You say they may perhaps return. The emigration -- pre-census emigration, they may or they may not return. They were not influenced by the war. They left before the war. They started in the 1960s onwards. People from all over Yugoslavia went to work abroad, and on what grounds do you come to the conclusion that perhaps they might now ²⁷¹⁷² return? For them the war was not a factor that prompted them to leave.

A. Well, I believe you are referring to the data included in that census, data on the population residing temporarily abroad in 1991 and still enumerated in that census and included in our study as well. Well -- and indeed I say that we include these records of the pre-census emigration in our statistics of displaced -- of refugees in this case. It is all people in other countries. And I believe that perhaps they would have returned if there was no war in Bosnia. Perhaps not. We don't know.

Q. That's what I'm saying. Perhaps they wouldn't. You say perhaps they would. But they didn't leave because of the war; they left to find jobs, to make a living. How many people from other countries, not from Yugoslavia, are working in Germany, Austria, France, or in other countries of the European Union? "Gastarbeiters," as they're known, "guest workers."

A. Well, in official statistics on the census published in the former Yugoslavia, the population temporarily residing abroad was always included as the population of a given territory in a given municipality or even in a settlement.

So you yourself, your statisticians include -- consider this population as part of the population of Bosnia and Herzegovina or Croatia, and you ask me now to exclude them from the census statistics and not to trace them in the post-war sources. So I think I would be, first of all, inconsistent compared with official sources; and secondly, what I said about the intention or the actual return of those who temporarily stayed, ²⁷¹⁷³ these are people who still have their houses in Bosnia and Herzegovina whose often family members live there in Bosnia and Herzegovina.

Q. Ms. Tabeau, wait a moment, please. I am not arguing with you at all. Of course people at the time, citizens of Yugoslavia who have their houses, who have Yugoslav passports who are working abroad should be included in the census of citizens of Yugoslavia. My question was why do you assume that they might return? And then you link that to the war. And why are you counting those people who left before the war to work all over Europe, do you consider them refugees? How can you qualify them as refugees? Somebody who may have left from Serbia or Bosnia, Slovenia, Croatia, anywhere. They may have gone to Australia. Why would he be considered a refugee, for instance, if he went to Germany or Austria to work there?

JUDGE MAY: The time is very nearly up. The witness can answer this question so far as she can, so far as it's relevant, then you have one more question, Mr. Milosevic.

Yes, Ms. Tabeau.

THE WITNESS: We dealt with this problem. We included these persons in our refugee statistics. At the same time we made an assessment of the error that we possibly made by including them, and the error is 4.2 per cent, which is less than widely accepted 5 per cent error in statistics.

So because it is unclear whether they would return -- have returned or not but they still were abroad in the time of elections 1997/8, according to our statistical definition, they had to be included ²⁷¹⁷⁴ in the refugees statistics. So -- but even you think that it is an error. The error is fairly acceptable.

MR. MILOSEVIC: [Interpretation]

Q. Very well, Ms. Tabeau. I think that we don't understand one another, because my belief is that they cannot be considered refugees. It would be the same as if you were to count tourists as refugees. Now, is it clear from the data at your disposal that the increase of the Serb population in Republika Srpska is a direct consequence of the persecution and reduction of the Serbian population in the territory of the Federation of Bosnia-Herzegovina?

A. I don't think that there is such a conclusion in my report, so I don't think we have shown something like that.

JUDGE MAY: Yes. Amicus, please. We'll sit on to try and finish this witness.

MR. KAY: Let us just turn to the report at tab 2 of the bundle, page 5.

Questioned by Mr. Kay:

Q. And I want to ask you a question about a matter that Milosevic has just asked you concerning the percentage of population temporarily residing overseas. You described it as 4.2 per cent.

A. This was a different per cent I mentioned. It was the possible error that we make by including those who were pre-census emigration and were still abroad by 1997/8. But the percentage of population overseas in the 1991 census is 5.4 per cent.

Q. Yes. ²⁷¹⁷⁵

A. This is what is mentioned on page 5.

Q. Yes. Is that a different statistic to the one you gave a moment ago or the same statistic but you got it wrong?

A. It is a correct statistic, and it is an entirely different statistic.

Q. Thank you. In your report, the base population statistic, is that derived from the population census of 1991? Is that right?

A. Yes.

Q. And in that census, it included those people working abroad who would be known before 1991 as economic migrants?

A. Well, it is the census, and it includes the population working and staying abroad.

Q. You're familiar with the phrase "economic migrants"?

A. Yes, I am, but in statistics, official statistics published in the region of the former Yugoslavia on the population temporarily residing abroad, working or staying with family members, this term is never used. It is just the population overseas, temporarily abroad.

Q. I'm asking you, though, about your report and about the terminology. Working abroad is something that is known as "economic migrants."

A. Yes, that's correct.

Q. And in considering the statistics that were later produced, these were not considered as economic migrants but as refugees.

A. That's correct. We included this subgroup in the statistics of refugees. ²⁷¹⁷⁶

Q. That is an inconsistency in their status, isn't it?

A. Well, it is questionable. Well, first of all regarding the population working abroad, I must make clear that among the records of -- that we had at our disposal, we in total are speaking about 325.000 records. 325.000 records of those who were found in 1997/8 in total as refugees abroad. So among those records approximately 110.000 persons, 110.000 persons, roughly one-third of the persons were persons who worked in neighbouring countries like Croatia or Serbia. So it is one-third of records that have a different meaning, the economic migration. It is not that simple that all these records can be considered economic migration from Bosnia and Herzegovina to Western countries. This were people who in 1991, one-third of them, one-third of them, roughly 110.000 people at 1991 census were in Bosnia. And probably already at that time worked in neighbouring countries like Croatia and Serbia. So would you consider this an economic migration?

Q. You don't know where they were, though, do you, where they were working?

A. I do know in terms of a country. I do know.

Q. How do you know that?

A. Because the country of residence was reported in the records that we have.

Q. In the 1991 census, was it?

A. In 1991 census and also in 1997/8 records of voters, out-of-country voters.

Q. In the 1997 census, was there a category there that dealt with ^{27177 BLANK PAGE 27178} economic migrants?

A. Well, we had a category of out-of-country voters, and we didn't know how many of them were the economic migration and how many of them were just people who were working abroad because they couldn't return to their own homes in Bosnia.

Q. The basis of the census then chose the address, the building in which a person lived as the place at which they were to be determined for the purpose of the census; is that right?

A. Address is indeed available from the census records for every resident of Bosnia and Herzegovina. Indeed it is information that is available. But for those who at the time of census were not in Bosnia and Herzegovina but resided in a different country, the address is unavailable. Only the country is available.

Q. In the 1997 census, did it deal at all with the reason why someone might be living abroad?

A. Not at all. The 1997/8 census, as you call it, it was just voters register which didn't deal at all with matters as the reasons for staying abroad at the time of the elections.

Q. So if someone had been an economic migrant before 1991, in the 1997 census, he wouldn't be shown as an economic migrant. He would be shown as a refugee.

A. He would be shown as a voter abroad, as an out-of-country voter. And statistically speaking, we would consider this person a refugee due to the fact that the place of residence of this person in 1991 and 1997/8 is different. ²⁷¹⁷⁹

JUDGE ROBINSON: Dr. Tabeau, the -- are you familiar with the definition of refugees in the refugees convention?

THE WITNESS: Yes. This is the legal definition that we, I said, haven't used in our report. We use the statistical definition of an internally displaced person as refugee as well.

JUDGE ROBINSON: Because the most important element of that definition is the fear of persecution. You don't take account of that is correct the motivation.

THE WITNESS: No, we don't take the motivation into account because it is impossible on the basis of the sources we used.

JUDGE MAY: Yes, Mr. Kay.

MR. KAY:

Q. That is the point I was making, that it was your interpretation of them as refugees rather than whether they were in fact refugees according to any legal definition.

A. May I say something?

Q. Yes. You can answer the question. That's why I asked it.

A. This is not my interpretation. It is the way we have done this using the statistical approach, and I wrote how -- what -- who was a refugee in our study in the first section of the report. I never pretended to use the legal definition of a refugee in this report, because it was impossible. From the beginning we used the statistical definition of a refugee who is defined in our report in the first section, I believe. Yes. It is on page 7 in the English version of the report. This is the third last paragraph in the bottom. Starting "Refugees were the persons, ²⁷¹⁸⁰ were persons who in 1991 were reported in the population of Bosnia and Herzegovina, including those temporarily residing abroad and who in 1997/8 registered to vote in countries different than Bosnia." So I am -- I have been clear about this from the very beginning of this report. The refugee statistics in the sense of the legal definition or the motivational definition, that would be the records of refugees from Western countries. Not only Western countries but also from Croatia or Serbia, Montenegro, because these type of definitions were used when giving people, when granting people the status of a refugee, the certain legal status that brings with it certain material benefits and certain aid to which a person is then entitled.

Q. Just as another matter on this particular subject, again in the 1997 census, had there been any distinction made in relation to those who were abroad but registered to vote who were there for economic reasons, who were economic migrants?

A. No, I don't think so.

Q. In your report at page 5, you said that there were deficiencies in the population census of 1991 because of the unstable political situation. Could you tell the Court what impact that would have had for the collection and accuracy of that data in the 1991 census?

A. I think the census as such has been conducted -- had been conducted following the procedures and methodological guidelines as required for the census. So I don't have any doubts about that the census was conducted correctly and the coverage of the census was complete, but because it was such a huge population survey, a few individuals could not ²⁷¹⁸¹ be covered, but this is -- we are speaking about numbers that can be safely neglected. So it is a complete population survey reporting the status of the population as of the critical moment of that census, March 31, 1991.

After the census was completed, the data processing issue started, and the deficiencies of the census are related to this phase of that census statistics, of producing that census statistics. First of all, the census -- but I should add that there is a report in annexes in our report made by a person, by the head of population section from the statistical office in Sarajevo who was involved in that census from 1984/5, from the very first preparations, and was part of the Belgrade commission, methodological commission on the census since the very beginning of preparations. So all details about how census was conducted and what were the problems met at that time, how the data computing processing was done, how the data were computerised, what has not been finished, is in this report in detail.

What I want to say, there were deficiencies related to the data processing phase of that census. First of all, the census material had to be scanned. Optical scanning was the method of getting the data computerised. The optical scanning was not perfect. There were mistakes in the names of the listed in the census questionnaires, spelling mistakes. That is a problem, a big problem because it cost us a lot of time to improve, to eliminate the mistakes, at least largely if not all of them, in order to -- and with correct names that were used in matching. Secondly, the data cleaning phase was not entirely finished due to ²⁷¹⁸² the conflict. So this simply means that there is one portion of the census material, and this is the portion related mainly to dwelling census and agricultural part of the census, that have never been finished in terms of data quality control and recording of variables. But we haven't used any information from the dwelling and agricultural censuses. So the items that we used were actually cleaned largely, and recorded if there were inconsistent cards reported, and could be used in the analysis as our study.

So this would be, in brief, what I can say about deficiencies. So that would be mistakes in the names and unfinished data quality control.

JUDGE ROBINSON: Essentially, Dr. Tabeau, essentially I mean, you're looking at people who were displaced by the war either internally or, I suppose, externally. Those displaced externally, you call them refugees irrespective of their motivation for leaving. But a person who leaves for economic reasons may not have left because of the war conditions, and wouldn't that lead to a mistake in the analysis?

THE WITNESS: Well, this is what -- what we were afraid of, and we investigated this issue in a very great detail, and there is one annex in our report in which we thoroughly discuss consequences of inclusion of population temporarily abroad into our study. I am speaking about Annex C, I believe.

JUDGE ROBINSON: And you said earlier that it is within acceptable limits.

THE WITNESS: Yes. I already summarised the error, possible error that we are making by including those people in the statistics of ²⁷¹⁸³ internally displaced and refugees, and the error is 4.4 per cent. This is a very small bias. If -- I don't think it is a problem in our report, not at all. Not at all. Especially, as I said earlier today, the refugee component of our report, of our statistics, is very limited. It is heavily underestimated because of lack of sources that report on refugees. I am speaking about sources like the Serbian and Montenegrin registers of refugees in these countries and also about records of refugees from Western European countries, mainly receiving countries. So these are huge numbers that are not included here. So this error I mentioned would become close to nothing if these records would have been included in our refugee statistics. Well -- that I didn't include any references by others like UNHCR on refugees and internally displaced persons from Bosnia is because it was impossible for me to work with individual records underlying the statistics compiled by UNHCR, and UNHCR themselves always concluded that their statistics are very approximate.

So it is not a good reference to be included in reports like this one that we are presenting today, but these statistics are available, and for your information we can compile an overview of UNHCR statistics on internally displaced persons and refugees from Bosnia starting in 1992, throughout the war - we have a few reports from the war period - and then until 1997. That is for your information if it is -- if you would request it.

MR. KAY: I note the time, so I'll ask one more question, Your Honour. ²⁷¹⁸⁴

Q. In relation to the household survey of 1994, in answer to Mr. Milosevic's questions, you said that the research institute involved was properly biased a little bit. Who was the research institute that was probably biased a little bit?

A. This is the Institute for Research of War Crimes and International Law in Sarajevo, led by Professor Smail Cehic. To make things absolutely correct, I didn't say that the institute was biased towards the Muslims but generally how the survey was conducted, that the people who did the survey were mainly Muslims who had perhaps, and this is my opinion, personal opinion, easier access to the Muslim population than to other ethnic groups.

Q. And if so, of course, then the statistics can be mistaken.

A. This is a misunderstanding. Ethnicity doesn't really play any role in the Sarajevo reports, any of them at all, because it is not that we discuss ethnic cleansing or whatever. We discussed casualties, casualties irrespective of ethnicity, the number of how many people died, were killed or wounded in the siege of Sarajevo.

Q. But it's casualties expressed by ethnicity. The report clearly has that.

A. We have such an information in our databases, and this is included for the readers, for those who will look at the numbers, but I myself wouldn't use these figures as an example of targeting a particular ethnic group, like Muslims. It was a siege. It was -- I don't know how is it possible to be selective, you know, in shooting from a hill a few hundred metres below and to know whether you are shooting a Muslim or a Serb. It ^{27185 BLANK PAGE 27186} is absolutely impossible if my view.

Q. So if the figures are used in that way, it would be a misleading use of the statistic, would it?

A. I don't think it is misleading because this is the correct ethnicity, the numbers reported under ethnicities are correct in the sense, were obtained from a certain source in a certain way. So they are correct statistically speaking. But you must be careful with your interpretation. I never interpreted these numbers as an example of ethnic cleansing in any of the Sarajevo reports.

JUDGE MAY: Very briefly.

MR. GROOME: Just one question, Your Honour Re-examined by Mr. Groome:

Q. Both the accused and Mr. Kay have asked you about the reliability of this household survey of Sarajevo. My question to you is, in the third Sarajevo report, you worked with seven other sources of data and you talked about the process of comparing them and putting them together, and my question to you is: In the course of that process, did you form a view as with respect to the reliability of the household survey of Sarajevo when it was worked with in -- with these other seven independent sources?

A. Well, the analysis of the Sarajevo household survey in the context of other sources just confirmed that this source can be used as a source for estimating casualties. There was a large overlap between this source and other sources, good sources like the official records, death records of the Federal Institution for Statistics. This confirms that after all, these responses in Sarajevo household survey made a lot of sense and I ²⁷¹⁸⁷ definitely believe it is a reliable source to be used for this purpose.

MR. GROOME: Thank you. No further questions.

JUDGE MAY: Dr. Tabeau, that concludes your evidence. Thank you for giving it. You are free to go.

We will adjourn now unless Mr. Nice has an announcement.

[The witness withdrew]

MR. NICE: No. Only that the witness tomorrow is one whom the accused may want to know is the subject of the recent application, and he may wish to ensure that he reads both the witness statement and the draft proofing summary ahead of tomorrow morning's session. I forecast that by then the witness will have signed the proofing summary and it may be that I will make an application that we should deal with the evidence in some more expeditious way than --

JUDGE MAY: Just a moment. So that he follows -- just clarify that it is witness -- he has a number. 225.

MR. NICE: 225. He's not going to be protected once he starts giving evidence, but it's probably better --

JUDGE MAY: For the moment he is.

MR. NICE: For the moment he is.

JUDGE MAY: So the accused has it, it's B-225 next up.

JUDGE KWON: Was the proof summary translated?

MR. NICE: No, and I don't imagine it's going to be possible for it to be translated by tomorrow. I can discuss those matters generally tomorrow, but in any event, he will have signed it. I'm not necessarily suggesting that I'm going to be seeking to rely on that, but he will have ²⁷¹⁸⁸ signed it and we'll see how we go from there.

JUDGE MAY: Is it a case in which there is much added in the proof summary which wasn't in the statement?

MR. NICE: There's quite a lot. I have just been checking it. I think it's at least 20 per cent.

JUDGE MAY: We may have that indicated to us tomorrow.

MR. NICE: You will find on most paragraphs at the end of the paragraph there's a reference to either -- that is the paragraphs of the proofing summary, there is a reference to either a paragraph in the statement or to proofing or to both.

JUDGE MAY: We're not going to go into it now, Mr. Milosevic, but there will be an application tomorrow morning by the Prosecution to adduce those, either the summary or the statement instead of the examination-in-chief. We will rule on it, that application, tomorrow morning.

THE ACCUSED: [Interpretation] Mr. May.

JUDGE MAY: Yes.

THE ACCUSED: [Interpretation] Mr. May.

JUDGE MAY: Yes.

THE ACCUSED: [Interpretation] This is some kind of new practice now, and in effect, this abolishes the examination-in-chief as such. What is presented here is some kind of a summary that is not written by the witness but by Mr. Nice and his team, and then the witness serves the purpose of cross-examination only here. This has gone even further than that Rule of yours, 92 bis. I think this is unacceptable. The ²⁷¹⁸⁹ examination-in-chief has to be heard.

JUDGE MAY: Tomorrow morning. We're going to adjourn now. You're quite right, but that is ruling of the Appeals Chamber by which this Chamber is bound. Perhaps you would like to have a copy of it. No doubt Mr. Kay could give you a copy of it if you haven't got it. We will adjourn now.

--- Whereupon the hearing adjourned at 2.21 p.m., to be reconvened on Wednesday,

the 8th day of October, 2003, at 9.00 a.m.