Select Committee on Foreign Affairs Fourth Report


Did NATO misjudge Milosevic's likely response in Kosovo to the bombing campaign?

90.  A related issue is the extent to which NATO was aware of Milosevic's plans to engage in full-scale ethnic cleansing of Kosovo. If NATO had been aware of his plans, it would perhaps have acted differently, or at least made better preparations for the large numbers of refugees who fled or were deported from Kosovo once the bombing started. If the ethnic cleansing was foreseeable, was it foreseen, and if not, should it have been?


91.  The first official Serb plan to change the ethnic character of Kosovo by reducing the number of Albanians was drafted during the late 1920s and early 1930s by the Serbian Academy of Sciences, involving the transportation of Albanians to Turkey.[213] About 100,000 left before the programme was interrupted by the outbreak of World War II. It then resumed in 1953.[214] The FCO records that "in the 1980s a draft Memorandum of the Serbian Academy of Sciences and Arts called for redressing the ethnic balance in Kosovo in favour of the Serbs and Montenegrins."[215] In 1991, the breakup of Yugoslavia coincided exactly with the collapse of the one-party state in Albania and the subsequent opening up of the country. Many Serbs then believed that the Kosovo Albanians should be encouraged to go and live in Albania. It was hoped that by practising the same intimidatory and discriminatory procedures that were used during the inter-war period, Belgrade could encourage Kosovo Albanians to leave for Albania or to join their huge diaspora. Simultaneously, Serb refugees were re-settled along the Albanian border just as Serb settlers had been in the 1930s. By 1995, however, it had become clear that no Kosovo Albanians showed any desire to go and live in Albania, and worse still, the Kosovo Albanians were planning to begin a guerilla campaign inside Kosovo itself.

92.  In 1996, as the KLA began its campaign[216], the Belgrade Academy of Sciences published a detailed report, which described the state of Yugoslavia in the year 2015 if the Kosovo Albanians remained inside Serbia/Yugoslavia. The report stated that due to the Albanians' high birthrate, the Albanian population was far younger and more fit for military service, so that 25 per cent of Serbia's conscripts were of Albanian origin. As a result, the architect of the report, Alexander Despic, suggested that it was perhaps time to let Kosovo secede from the federation. In response to this plan, Dr Dusan Batakovic, himself from a Kosovo Serb background, produced a plan which envisaged the cantonisation of Kosovo, aiming to bring together localities according to their ethnic composition. His map details the areas where Serbs would be in the majority and those with an Albanian majority. The areas form a horseshoe shape with the Serbs in control of the north, north east, Pristina and the area along the eastern Montenegrin border, and the Albanians concentrated in the south central areas and along the Albanian border. In other words, this plan amounted to partition. It was widely discussed in Serbia and elsewhere following its publication in October 1998. With the escalation of the KLA's activities in1998, Batakovic was able to sell his plan to many political and military personnel as the only way to preserve the most important Serb centres in Kosovo, i.e., the Trepca mines, Pristina and the holy monastic sites.


93.  These plans undoubtedly contributed to anti-Albanian sentiment amongst the Serbian population, and to the idea that one way of dealing with the Kosovo Albanian 'problem' was to displace large number of Albanians. The plans also gave rise to the use of the term "horseshoe" in this context. It has also been alleged that Milosevic was following a long organised plan of ethnic cleansing during the military campaign in Kosovo. On 9 April 1999, the German Ministry of Defence published a document which was alleged to describe a Serbian plan known as Operation Horseshoe.[217] This document does not in fact refer specifically to "cantonisation," but only to the "destruction and neutralisation of the UCK [KLA]." The German Foreign Minister, Joschka Fischer, was also quoted as saying that the Serbs had set in motion a plan called Operation Horseshoe, which aimed at expelling Kosovo's ethnic Albanian population.[218] The Foreign Secretary told us that "it has been reported in the press (and I can confirm it is true) that there was a plan developed in Belgrade known as Operation Horseshoe which was for the cleansing of Kosovo of its Kosovo population. That plan has been around for some time."[219] During the campaign and since, it has become established currency in the media that German intelligence had discovered this alleged plan, and that NATO's leaders should have been aware of the plan earlier.[220] Some of NATO's leaders (in particular in Germany) used the existence of the plan to illustrate Milosevic's character, and to prove that the ethnic cleansing of Kosovo was not triggered by the NATO bombing, thereby justifying the NATO campaign. A retired German brigadier-general has since alleged that the German Defence Ministry turned a vague report from Bulgarian intelligence into a "plan," coining the term "Horseshoe." Ironically the originator of the report confused the Croatian word for a horseshoe with the Serbian word. The brigadier-general also argues that German politicians subsequently misquoted the original Bulgarian report by arguing that the report demonstrated that the goal of the Serbian military was to expel the entire Kosovo Albanian population, rather than destroying the KLA.[221]

94.  Regardless of whether Operation Horseshoe really existed, and what it consisted of, the evidence of the OSCE's report, Kosovo/Kosova As Seen, As Told,[222]—based on the observations of the OSCE observer team in Kosovo and then upon extensive interviews conducted with refugees after the bombing had started—is that the Serb military campaign did not appear to be oriented towards "cantonisation.". A number of factors point to this conclusion. The Serb military campaign before the start of the bombing consisted of:

    —  activity wherever there was KLA activity, and wherever it was suspected there were KLA sympathisers;

    —  efforts to control the main communication routes;

    —  with the approach of the bombing, securing Kosovo's borders.

Once the bombing had started, "expulsions took place in practically every municipality."[223] As Dr Woodward told us, "I think also that once the bombing began then it was a different objective, it was no longer a campaign against the KLA but a campaign with strong historical roots."[224] It is also the case that "over 90 per cent of the Kosovo Albanian population was displaced in 1999."[225] Almost half of the Kosovo Albanian population were refugees—as opposed to internally displaced—by 9 June 1999. According to the OSCE "the majority of refugee statements indicate that documents...were routinely taken from Kosovo Albanians by police. This indicates that it was intended that the refugees should not return. Expulsions on this scale do not appear to be part of a "cantonisation" plan, but a more ambitious effort to rid Kosovo of the majority of its Kosovo Albanian population. Of course, once Kosovo had been cleared of most of its Albanians, it may have been Milosevic's intention to implement a cantonisation plan, but we do not know this.

95.  On the other hand, the OSCE report also casts light on the question—distinct from that raised by Operation Horseshoe—of whether the campaign against the Kosovo Albanians was planned, or the result of a spontaneous outburst of anger following the launching of air strikes. The first large scale movement of refugees—69,500—left Kosovo on 23 March, one day before air strikes were actually launched, and three days after the departure of OSCE monitors (although of course there were large number of internally displaced people and refugees before then—the point here is that there was a rapid acceleration at this time, particularly in refugees). In the first eight days of the campaign, 307,500 refugees left Kosovo. As the OSCE report argues, "the arrival of such large numbers so soon after the departure of the OSCE-KVM would appear to indicate pre-planning of the operations."[226]

96.  Another indicator that the campaign was organised and planned before the launch of air strikes is the use of trains, and buses, with two policemen per bus.[227] The OSCE records an incident on 27 March when police made villagers of Randubrava of Prizren municipality board buses which took them to Zur, from where they had to walk to Albania.[228] The UNHCR record that on 31 March, two trains carrying 3,000 people reached the Macedonian border. On the next day, two trains carrying 5,000 people arrived.[229] To organise this number of people on to buses and trains requires considerable planning. The Foreign Secretary was clear that "what we have witnessed in Kosovo has not been spontaneous emotional anger by random servicemen; what we are witnessing has been a deliberate co-ordinated programme of deportation."[230] Of course, not all the refugees created by Serb action in Kosovo were actually herded onto buses or trains, or sent across the border at gun-point: many fled of their own volition, aware of the consequences of staying behind.

97.  The fact that elements of the campaign against the Kosovo Albanians were planned does not mean that, overall, the campaign was smoothly directed and well organised. Different elements of the Serb security apparatus appear to have had different agendas. One source quoted by Tim Judah states that "there were differences between the police and the army. The police were in favour of the expulsions because they could steal money from people. The intelligence guys were against it...the worst were the paramilitaries and the locals."[231] The OSCE also records that "paramilitaries appear to have meted out particularly savage treatment" and, by way of contrast, reports an incident where a young Serb soldier helped a wheelchair-bound Kosovo Albanian women, returning her documents after paramilitaries had seized them, and subsequently organising food, water and medicine for a mosque where villagers were sheltering.[232]

98.  The picture is one of generalised violence against the Kosovo Albanians, with some elements organised from Belgrade, but much of the violence was not carefully orchestrated. This picture is consistent with a confidential memorandum provided to us by the FCO. We conclude that, regardless of the accuracy of reports of "Operation Horseshoe," there were orchestrated elements to the campaign of expulsions, which could be described as a plan. Outside observers could have been aware of this plan as it would have required significant preparation. We also conclude that the withdrawal of OSCE monitors together with the international media and the start of NATO's bombing campaign encouraged Milosevic to implement this plan.


99.  It is well established that Belgrade was planning an offensive in spring 1999: an offensive had been conducted during the spring of 1998, and in the three week hiatus between Rambouillet and Kleber, "Belgrade had just about deployed double the forces permitted under the Holbrooke terms."[233] The FCO records that "by 19 March there were reports that the offensive against the KLA and ethnic Albanian civilians had intensified significantly, driving thousands of Kosovo Albanians from their homes, carrying out summary executions and destroying property."[234]

100.  An FCO spokesman was quoted in The Sunday Telegraph of 4 April 1999 as saying that "We did not expect Slobodan Milosevic to move the levels of population that he is moving—perhaps that was a failure of imagination."[235] The Foreign Secretary told us that this statement "was not one which he would support"[236] but that "we anticipated...the spring offensive [and] that it would be accompanied by ethnic cleansing; we did not have any intelligence to suggest he was going to load up whole trains and run a shuttle train deportation from Pristina to the Macedonian border and I think, in fairness, for all decent people, that would be beyond imagination."[237] Dr Jones Parry gave us a slightly different perspective from that of the Foreign Secretary that "there was one piece of information that alluded to this sort of activity [organised ethnic cleansing]. There was no corroboration for it and we had masses of conflicting information which conformed more to what we had seen and what we reasonably expected."[238] He also told us that "what we expected...was more of the same, people being driven out of their homes and on to the hill tops. We had no grounds for believing in making a judgment that we could have expected the sorts of events that then took place over Easter"1999.[239] However, the director of the CIA told the US Congress in early February 1999 that the Serbs were preparing a spring offensive that probably would produce "huge" refugee flows.[240] Thus even if the FCO had not been aware of plans for an acceleration of ethnic cleansing, it should have been aware that the Serb offensive would produce large numbers of refugees.

101.  The UNHCR records that "just before the air strikes started on 24 March, the office of population and refugee affairs in the Department of State started asking persistent questions regarding UNHCR's preparedness in case of large refugee outflows. The communications had an urgent tone and a formal classification that suggested this was not routine information gathering."[241] It does seem, therefore, that there were pieces of information which were available to the USA—and if United Kingdom-US intelligence co-operation remains strong, which should have been available to the United Kingdom—that indicated that the internally displaced in Kosovo were about to become refugees. If only a few pieces of information indicated that there were likely to be large numbers of refugees created, amongst a raft of information indicating something else, it is unreasonable to expect these pieces of information to have been given particular weight. On the basis of the information available to us, it appears that there are some grounds for criticising the assessment of intelligence in this case. A separate question is why there were so few pieces of intelligence which pointed in the right direction—that is, why there was a failure of intelligence collection. In mitigation, it is presumably the case that penetrating Milosevic's inner circle is, to say the least, a challenge. We do not know the answer to these questions because, as we refer to above, we were prevented from taking evidence from the Chairman of the JIC and CDI.[242]

102.  Leaving aside intelligence questions, there is the matter of whether specific intelligence was required to predict how Milosevic would react to the withdrawal of OSCE monitors and the start of bombing. Dr Jones Parry told us that "perhaps we are open to the charge that we addressed this through a view which said that that sort of activity was not something that in the late 1990s we could have expected. I am afraid we were wrong on that."[243] While we appreciate Dr Jones Parry's candour, we believe that the FCO should have been able to anticipate Belgrade's reaction to the bombing campaign. In the light of the widespread ethnic cleansing which occurred in Bosnia, orchestrated and aided by Milosevic, it is not clear why it was so difficult to imagine him planning mass ethnic cleansing in Kosovo, particularly in view of the long-standing enmity between the Serbs and Kosovo Albanians, and of what we note above about the tendency of war to create the conditions for ethnic cleansing.[244] Dr Gow, otherwise very supportive of the United Kingdom's role in the Kosovo crisis, wrote that "it is hard to follow how the FCO, with the Secretary of State in the forefront, and HMG as a whole could have been taken by surprise by Belgrade's ethnic cleansing campaign when so much energy and so many resources had been put into trying to avert that campaign...of particular pertinence here is the fact that the logistical means for wholesale and rapid ethnic cleansing were ready in situ—as with other aspects of military preparation, this takes weeks and months and does not come spontaneously over night."[245]

103.  The Secretary of State for International Development, Clare Short, told the House on 31 March 1999 that: "I reject absolutely suggestions that we should have been prepared in advance for a movement of population on this scale. It would have been an appalling act of complicity in ethnic cleansing to set up in advance a network of camps to await the Albanian population of Kosovo. That would only have assisted Milosevic's objectives."[246] We do not accept this. Dame Pauline Neville-Jones told us that the bombing campaign presented the " get a greater grip on [Kosovo], i.e. to drive the KLA out definitively, and if that meant actually that you drove the populations out and drove the villagers out, because you had not got the time to distinguish, in any case why should you bother, they were all in it ultimately, he went ahead and did that."[247] On this basis, it is difficult to believe that Milosevic was interested in whether there were camps waiting to receive the expelled Kosovo Albanians, or not. And given that one of the aims of the bombing campaign was to permit the refugees to return home, it should have been clear that the early provision of camps would have been a significant humanitarian measure, rather than an "appalling act of complicity in ethnic cleansing."

104.  The Foreign Secretary put a more convincing argument to us that in order to plan for the arrival of refugees, "we would have needed the co-operation of neighbouring countries and those neighbouring countries had difficulty contemplating the scale of the population's displacement which came their way."[248] Macedonia's reluctance to admit Kosovo Albanian refugees stemmed from the Macedonian government's view that the ethnic balance within Macedonia was delicate (up to a third of its population may be ethnic Albanians—see below).[249] On 3 April, the Macedonian government closed the border, so that 65,000 refugees were stuck in appalling conditions in a makeshift camp in the no-man's land between Kosovo and Macedonia for several days.

105.  Macedonia was eventually persuaded to allow large numbers of refugees on to their territory, and NATO's 8,000 troops, already deployed in Macedonia, conducted a highly effective operation to build camps.[250] However, had the effort to win over the Macedonian government started earlier, it is just possible that Macedonia would have been willing to open its border earlier. Of course, we cannot know this, as the alliance did not start to try to convince them until the exodus had actually started—because there was a failure to anticipate that there would be an exodus. UNHCR too was unprepared, having "received no advance warning from any government or other source."[251] The failure to predict Milosevic's reaction meant that adequate preparations were not made for the greatest movement of refugees in Europe since the Second World War. The UNHCR estimates that 862,979 refugees left Kosovo from 23 March to 9 June 1999.[252] We believe a very serious misjudgement was made when it was assumed that the bombing would not lead to the dramatic escalation in the displacement and expulsion of the Kosovo Albanian population. Although we accept that the government could not have established refugee camps before NATO action started, for fear of giving tacit encouragement to expulsion of refugees, equipment and supplies could have been stockpiled so that the refugees could have been housed more speedily once the exodus occurred. We are confident that NATO has undertaken an assessment of the reasons for its failure to predict Milosevic's response. We believe that this issue is of such over-riding public interest that the Government should make its conclusions available to Parliament for scrutiny.

213   This section draws upon Vickers, pp.127-129, as well as an FCO memorandum, Ev. p. 178. Back

214   Ev. p. 178. Back

215   Ev. p. 178. Back

216   See para 25. Back

217   HC Deb 8 March 2000, col 686W. Back

218   Judah, p. 240, quoting RFE/RL Newsline, 7 April 1999,­see/see­070499.html Back

219   QB109. Back

220   See for example, RFE/RL Newsline, 14 May 1999, Back

221   The Times, 2 April 2000. See also Le Monde 11 April 2000. Back

222   OSCE Report. Back

223   OSCE Report, p. 100. Back

224   QC244. Back

225   OSCE Report, p. 98. Back

226   OSCE Report, p. 98. Back

227   OSCE Report, p. 110. Back

228   OSCE Report, p. 110. Back

229 Back

230   QB110. Back

231   Judah, p. 241, quoting Braca Grubacic. Back

232   OSCE Report, p. 105. Back

233   Weller, p. 291. Back

234   Ev. pp. 9-10. Back

235 Back

236   QB106. Back

237   QB107. Back

238   QC33. Back

239   QC32. Back

240   Cited in The New York Times, 18 April 1999. Back

241   Para 87, report available on Back

242   See para 7. Back

243   QC33. Back

244   See para 91, and para 89. Back

245   Ev. p. 368. Back

246   HC Deb, 31 March 1999, col 1089. Back

247   QC243. Back

248   QC386. Back

249   See para 242. Back

250   Judah, p. 252. Back

251   UNHCR submission to inquiry by International Development Committee, Fourth Special Report, 1998-99, 27 July 1999 (HC 795), para 13. Available on A further assessment of UNHCR's response is available on its web site: Back

252   The second largest exodus was that of 150,000 Serbs from Krajina in Croatia in 1995. See, for example, Glenny, p. 650. Back

previous page contents next page

House of Commons home page Parliament home page House of Lords home page search page enquiries index

© Parliamentary copyright 2000
Prepared 7 June 2000