Wavell Room
Image default
Concepts and DoctrineCyber / InformationLandShort Read

UN Intelligence in the Balkans

The former UN Secretary-General, Dag Hammerskjold, believed that peace operations must have ‘clean hands’ and not use intelligence.1  For Hammerskjold, the use of intelligence would damage the UN’s ability to be seen as impartial and neutral, damaging trust in the mission.  In contrast, many UN military commanders have called for the increased use of intelligence in order to fulfil their mandate.

This article examines the intelligence cycle of the UN Protection Force (UNPROFOR) which was deployed to Balkans in the early 1990s.  It is now generally accepted that UNPROFOR did not perform well during its mission.  It watched and failed to prevent widescale ethnic cleansing, culminating in the genocide at the Srebrenica and Zepa ‘safe areas’ in July 1995.  This article doesn’t look to answer what the force knew, and purposely avoids the discussion around the mandate and arguments about mission failure.  Rather, it looks at the intelligence cycle of the mission to give an insight into the problems faced.  This article has four sections covering a simplified version of the intelligence cycle: planning, gathering, processing, and dissemination.  It draws the themes together in the conclusion to argue that modern missions face similar problems despite advances in technology and process since 1995.

Planning and Direction

The planning phase of the UNPROFOR cycle was multifaceted and undermined by strategic tensions. UNPROFOR was unable to centrally direct its requirements and was disjointed between the operational and tactical level.

At the operational level the Force Commanders had the clearest vision of the intelligence that they needed to complete their mission, making them the strongest element of this phase.  Their requests were generally, but not exclusively, fed to the Military Information Office (MIO) in Zegrab for action.  However, the effectiveness of this was often impaired by national tensions and mistrust of the UN system.  The principle that the UN is officially dedicated to ‘transparency, impartiality, and the rule of law’ filtered into mission processes, hindering the planning of even simple intelligence requirements.2  There were problems with the different caveats imposed by contributing countries.  The Russian contingent, for example, could not be tasked with collecting information without authority from Moscow.3  The multilateral makeup of UNPROFOR stopped commanders from being able to communicate their intelligence wishes to all of their subordinates, limiting the ability to centrally direct the cycle.

Lieutenant General Michael Rose

The planning of tactical intelligence was more successful because there was less dependence upon other contingents.  Secondly, the areas of intelligence interest involved were smaller.  Battalions needed intelligence about the state of the local population, where the warring factions were based and their intentions.  Each battalion had an intelligence cell within its structure working directly for the Commanding Officer as the central point of the planning phase.4  These structures have been criticised for not containing intelligence experts, hindering the effectiveness of planning.  David Cammaert argues that ‘ideally every contingent should have a few specialist teams’.5  Cammaert sees value in hosting intelligence officers with the soldiers deployed as it adds a capability that is otherwise carried out by less skilled analysts.  Yet, tactical intelligence does appear to be more effective than the operational level.  For example, the British contingent took great trouble to plan collection around local issues, historical perspectives, emotive locations, and other requirements aligned to their direct tasks.6  The lack of structural links to the operational level means that these plans were local and not aligned to the wider requirements of the force, undermining any collection plan.  Embedded intelligence staff at the tactical level would also have created a greater link to the operational level, freeing the flow of information and aiding planning.


The most common form of collection occurred at company level and consisted of low-level human intelligence. One special forces soldier recalled that a common task was to ‘map the lines and report’ the actions of the warring factions to feed into the intelligence picture.7  This was done through observation posts, patrolling, and talking to local civilians.  One Force Commander, General Rose, writes with some pride at the ability of ‘his’ teams of military observers to collect supporting information that was vital to the success of the mission.8  Moreover the deployment of specialist units allowed several contingents to collect technical and signals intelligence about the warring factions.9  It would be wrong to assume that such tasks were confined only to the special forces.  One platoon commander interviewed said that ‘gathering’ was a daily task conducted on every patrol.10  Yet this is more akin to generalist information observation than working towards a defined intelligence collection plan.  The mission relied heavily on observations instead of technical means of collection limiting its reach.

One of the hallmarks of the mission was that the warring parties could restrict soldiers’ freedom of movement.  It was not uncommon for patrols to be stopped at checkpoints and refused entry into sensitive areas, hindering the collection of information.11  The ease with which this could happen in the mission means that it was possible to easily deceive the UNPROFOR.  With limited technical intelligence collection assets, the force was dependent on sighting and patrol reports.  One of the risks of this collection method was that it could be deemed as spying and potentially endanger the neutrality of the force if any side chose to take issue with it.12  Worse, a Canadian colonel recalls that the UN became the subject of many attacks in retaliation to their attempts to gather information.13  This was not a universal picture and lots depended on local relationships; however during times of greater friction UNPROFOR would often find it was denied access and therefore unable to collect information about key events at relevant times.

Building an OP on the Kosovo-Macedonia border 1993. By Paalso Paal Sørensen 1993 – Own work, CC BY-SA 3.0, https://commons.wikimedia.org/w/index.php?curid=11073382

There was considerable tension between elements of the mission which marred the collection of intelligence.  One notable example is between the British Joint Commission Officers (JCOs) and the Dutch battalion deployed to Srebrenica.  Three British SAS JCOs deployed to Srebrenica to facilitate communication and to monitor the situation for General Rose.  The JCOs were deployed in small groups and successfully started communications between the Dutch battalion and the warring factions in Srebrenica.14  Nevertheless, the Dutch commander, Lieutenant Colonel Kerrammans, banned further talks between the JCOs and the warring factions for fear of breaching his neutrality and because of the threat of Serb attacks.15  This undermined a key intelligence collection asset in a critical area.  The Dutch battalion was also stopped from sharing information with the JCOs further hindering effective collection.16  This could be justified by the need to remain impartial but it shows the problems of national caveats on UN operations.

The use of Non-Governmental Organisations (NGOs) partly mitigated these restrictions and filled gaps in the intelligence picture.  For various reasons (including to steal the contents) humanitarian convoys were often allowed through otherwise blocked check points when UNPROFOR was not.  A UN military information corporal suggested that these were a good source of human intelligence.  Although, he also noted, that this was on an informal basis.  There were not formal links and often happened with soldiers ‘chatting up’ NGOs ‘over a brew’.17  These relationships were also not universally positive.  The peace worker, Margaret Preston, noted that a group of NGOs refused to help UNPROFOR because of a lack of trust and a need to be seen to be impartial.18  Despite this tension, it created a low-level of human intelligence which assisted the mission.  It was an NGO who informed UNPROFOR that Serbian troops were advancing into the ‘safe area’ of Gorazde, generating information otherwise denied to the force.19

UNPROFOR personnel were also used in more secretive operations.  Canadian special forces deployed ‘well camouflaged’ observation posts behind the front lines of the warring factions and collected images and established patterns of life.20  Such moves carried significant operational risk and potentially undermined the neutrality of the force.  If discovered, it is likely that the soldiers would have been captured and used to bargain with the operational commander of the mission.21


The UNPROFOR structures for processing and analysis were complex and dislocated nodes which hindered effectiveness.  No single element of the UNPROFOR intelligence staff ever had access to the whole force’s collated information.  Whilst some elements were very good, the evaluation of the intelligence at UNPROFOR was generally poor and disjointed.

The Military Information Office (MIO) was not capable of evaluating the information that it had access to because many of the staff deployed to it were not suitable.  The problem that the staff faced was the need to provide daily briefings for both the Force Commander and also various political customers, each requiring different analysis.22 This split their effort.  The complexity of generating the political and economic intelligence that was needed to support the politics of the mission was beyond the ability of staff posted to it.23   There was a standing joke around the MIO that ‘if you understand the situation, you must have been poorly briefed’ suggesting a culture that had failed to fundamentally grasp the issues it was tasked with evaluating.24  One Deputy Force Commander, General Ashton, complained that the end product of the MIO was often ‘so poor’ that it could not represent the work of a professional intelligence staff.25

1st Battalion Royal Highland Fusiliers on a combined foot patrol in Novi Travnik © IWM UKLF-1994-004-19-31

The MIO was also not structured to allow the efficient processing of information.  The MIO was split into three sections dealing with the orders of battle, analysing the situation on the ground, and a section for administrative functions.  The core problem lay not in the structure itself but limited interaction between them. The lack of an internal telephone network, coupled with limited computing power, also stopped effective information sharing and all source analysis.  These difficulties were further complicated by the MIO using 13 different languages.26

To some extent this was by design and because of the need for counterintelligence procedures.  The hiring of local cleaners brought the fear of bugging, driving the compartmentalisation of information.  This was combined with more traditional threats such as state sponsored spying mirroring the operational environment.  As the mission matured some of these problems were solved.  The British contingent deployed a ‘category 3’, equivalent to UK Secret, safe box from which to secure intelligence allowing its more effective evaluation.27  The arrival of an American intelligence contingent improved the situation by bringing secure internal communications.  Whilst the infrastructure improved, the processes did not.

UN Military Observers (UNMO) established a separate evaluation centre near the MIO.  A typical UNMO report would contain information about troop locations and who fired what and when.28  This separate processing node watered down any information collected by the force.  They did not routinely share raw information issuing only verified reports in slower time.  The information collected by the UNMOs in the field was sent to this office and not the central MIO before being distributed.  Conversely, one UNMO wrote in frustration at their inability to get messages to the UN.29  This split structure meant that it was often the UNMO office and not the MIO who would have the fastest access to this data, slowing the process of effective analysis.

A major problem with the evaluation of information was that some nations did not want the truth to be revealed for fear of damaging the political prospects of their favoured side.  For example, the Americans were concerned that evidence of their flights supplying the Croatian Army would be uncovered.30  The most famous case of this is the so-called ‘black flights’ to Tuzla air base over February-March 1995.  The flights were recorded by UNMOs and reported across the intelligence cycle publicly.31  Nevertheless, the American influence in the MIO stopped widespread publication of what the UNPROFOR intelligence cycle was confident had happened.  At a NATO press conference after one of the flights, the Alliance denied the reports were true, despite them being confirmed by a number of sources.32  Indeed, the final UNPROFOR report about the event was, unusually, a joint UNPROFOR/NATO report which made only limited references to the flights at all. National interest played a role in how the mission intelligence cycle was able to evaluate information.

Casualties arrive by United Nations helicopter at Tuzla airfield following their evacuation from Srebrenitza.© IWM BOS 86


The key problem for the dissemination of intelligence was restrictive sharing arrangements that undermined who could read what and when.  Mark Urban points out that ‘channelling of SIGINT from the Government Communications Headquarters and (HUMINT) from MI6…to troops… was constrained by strict rules about dissemination’.33  This high-level intelligence could not be directly shared across the force.  This meant that tactical commanders were often reliant on collated patrol reports instead of the full intelligence product available.

That a majority of the troops came from a shared NATO background in theory helped this distribution.  Shared doctrines and experiences theoretically result in a high level of trust between both nations and commanders, allowing a greater flexibility of intelligence sharing.  Urban argues that the British and French contingents regularly ‘swapped information’ in an effort to collate data.34  But across the force, the flow of this support was more limited.  In contrast an Indian commander, General Nambiar, ‘could not receive intelligence from NATO sources’.35  Likewise, there is no evidence of sharing formal intelligence products, such as UK Top Secret, with mission partners.  There is a difference between an intelligence sharing process and the relationship between individuals involved in the mission.  National caveats were a limiting factor in the UNPROFOR intelligence cycle; that the mission did not have a coherent intelligence structure and was the product of multiple national systems bolted together limited its effectiveness.  These problems manifested themselves at the tactical level with less sensitive forms of intelligence distribution.  The network was dependent on good will over formal structures to share products.


This article has provided a snapshot of the problems faced by UNPROFOR.  Concerns of national integrity and trust between contributing nations were a key reason for the ineffectiveness of the intelligence cycle.  More fundamentally, the mission intelligence design could never have been effective because it was not resourced adequately.  The intelligence cycle of UNPROFOR was deeply flawed and no single nation had the will or resource to solve the issues.

A British Army Warrior vehicle in the Bosnian hills, c1993 NAM. 1996-02-238-16

Since UNPROFOR, the UN has attempted to improve its intelligence processes.36  The addition of sensors and other technical means now augment low level collection efforts.  At a strategic level, the UN has also professionalised its intelligence staff to some extent.37  In contrast to the experience of UNPROFOR, many nations now also provide tactical intelligence at times of crises.38  Despite this, the lack of standing UN forces and professional intelligence staff at the operational and tactical level means that many of the problems that UNPROFOR faced are likely to remain similar today.  Missions remain dependent on the assets contributed to them and are still subject to the same national tensions and the frictions they create.

The UN should not forget the importance of this when putting together future mandates and force packages. As Britain looks to increase its commitment to UN operations, specialist intelligence capabilities could be a force multiplier.  Especially if Britain was prepared to loosen access to intelligence to serve wider peace objectives.  With a 2021 perspective, many of these issues feel like they should be easily solved.  In a UN context, nations who often oppose each other (such as Britain and Argentina working together in Cyprus, or Russian and Chinese contingents) work together but counterintelligence and political concerns will always remain.  The reality is that the problems faced by UNPROFOR are common to modern missions and it therefore provides a case study that remains relevant for those deploying on peace operations.

Cover photo: Not intelligence related.  British Army warriors in Bosnia 1992. © IWM BOS 32


Steve Maguire

Steve Maguire is a British Army Officer serving with The Royal Irish Regiment.  He has served at regimental duty, with an armoured infantry brigade, and with the Army Headquarters.  He is also the Wavell Room Senior Land Editor.


The views expressed in his writing are his and do not represent the views of the Ministry of Defence.


  1. Quoted from Conor Cruise O’Brien, To Katanga and Back, (New York, Grosset and Dunlop, 1962) p. 76
  2. Walter Dorn, ‘The Cloak and the Blue Beret’ (International Journal of Intelligence and Counter Intelligence, Vol. 12, No.4, pp 414 – 447) p.415
  3. Steven Burg & Paul Shoup, The War in Bosnia-Herzegovina (M. E. Sharpe, New York, 2000) p.145 see also Jacob Kipp & Tarn Warren, ‘The Russian Separate Airborne Brigade – Peacekeeping in Bosnia-Herzegovina’ in John Mackinlay & Peter Cross (ed.), Regional Peacekeepers, The Paradox of Russian Peacekeeping (United Nations University Press, New York, 2003) p.40
  4. See US Army manual FM 100-5 Operations (1993) (available from https://cgsc.contentdm.oclc.org/digital/collection/p4013coll9/id/49/f accessed 2 April 2021) and see also Robert Cassidy, Peacekeeping in the Abyss (Praeger, London 2004)
  5. Patrick Cammaert, ‘Intelligence in Peacekeeping Operations: Lessons for the Future’ in  Wies Platje et al (ed.) Peacekeeping Intelligence: Emerging Concepts for the Future, (OSS International Press, 2003) p.21
  6. Major Roger Marshall, ‘Operation GRAPPLE: British Armed Forces in UN Protection Force’ (http://www.fas.org/irp/agency/army/mipb/1996-4/marshall.htm accessed 10 April 2021) See also the un-official forums of the British and Canadian military: www.arrse.co.uk and http://forums.milnet.ca/ respectively. Particularly these threads: ‘UN and Intelligence’ http://www.arrse.co.uk/Forums/viewtopic/t=132246.html and ‘HUMINT Assessment Program’ http://forums.milnet.ca/forums/index.php?topic=29480.0 which demonstrate a good understanding of the requirements of intelligence.
  7. Mike Curtis, CQB: Close Quarter Battle (Transworld, London, 1997) p.334
  8. General Sir Michael Rose, Fighting for Peace,(Clays Ltd, St Ives, 1998) pp 90 – 91
  9. Mark Urban, UK Eyes Alpha (Faber and Faber, London, 1996) p.216 & see also Royal Signals Archive Blandford, ‘Report on Squadron support for British Forces in the Former Yugoslavia’. The author has also spoken, but can provide no evidence for, several service personnel engaged with collecting signals intelligence.
  10. Correspondence with a British Officer deployed to UNPROFOR 1995 (1)
  11. James Tow, Triumph of the lack of will (Columbia, New York 1997) pp129 – 131
  12. Robert Cassidy, Peacekeeping in the Abyss (Praeger, London 2004) p.178
  13. Colonel K Hague, ‘UNPROFOR: A Perspective the Field’ (available from https://apps.dtic.mil/dtic/tr/fulltext/u2/a295010.pdf accessed 03 February 2021)
  14. See Peter MacDonald,The SAS (London, Pan Books, 1998) p.153
  15. Cees Weiss, Intelligence and the War in Bosnia 1992 – 1995 (Lit Verlag, London, 2003) p.129
  16. Confidential Interview (49) cited in Cees Weiss, Intelligence and the War in Bosnia 1992 – 1995 (Lit Verlag, London, 2003) p.138
  17. Author correspondence with a British soldier deployed to UNPROFOR 1995
  18. Imperial War Museum Archive, interview with the peace worker Margaret Preston.
  19. Cees Weiss, Intelligence and the War in Bosnia 1992 – 1995 (Lit Verlag, London, 2003) p.41
  20. David Pugliese, Canada’s Secret Commandos (Esprit de Corps Books, Ontario, 2002) pp41 – 42
  21. Steven Burg & Paul Shoup, The War in Bosnia-Herzegovina (M. E. Sharpe, New York, 2000) p.135
  22. Thomas Quiggin, ‘Response to ‘No Cloak or Dagger Required’’ (Intelligence and National Security, Vol.13 No.2 pp 203 – 207) p.203
  23. Thomas Quiggin, ‘Response to ‘No Cloak or Dagger Required’(Intelligence and National Security, Vol.13 No.2 pp 203 – 207) p.205
  24. Michael Smith, New Cloaks, Old Daggers (Gollanca, London, 1996) p.210
  25. End of Tour Report by Major General Barry Ashton, Deputy Force Commander, United Nations Peace Forces, C.1996
  26. Cees Weiss, Intelligence and the War in Bosnia 1992 – 1995 (Lit Verlag, London, 2003) p.45
  27. Correspondence with a British Officer deployed to UNPROFOR 1995 (2)
  28. See ‘Incident study report regarding mortar shelling Dobrinja, Sarajevo on 1 June 1993: investigation’ in Final Report of the United Nations Commission of Experts Established Persuant to the Security Council Resolution 780, annex VI.A
  29. Lieutenant Colonel J D Deverell, The Role of UNMOs in a Peace Support Operation, British Army Review, Spring 2020, pp108-124.  specific reference is pg 118. Available from https://chacr.org.uk/2020/04/02/british-army-review-special-report-spring-2020-the-yugoslav-wars/
  30. Whilst this paragraph deals with the Americans, the Iranians were importing significantly more arms into the country via a variety of channels. President Izetbegovic, the Bosnia leader, had strong ties with the Iranians. See John Schindler, Unholy Terror, Bosnia, Al-Qaida, And The Rise Of Global Jihad (Zenith Press, USA, 2007)
  31. Serbian Network, Dutch Government Srebrenica Study: Secret Arms Supplies to the Bosnian Muslims: The Black Flights to Tuzla, (available from http://www.srpska-mreza.com/Bosnia/Srebrenica/Tuzla-delivery.html accessed 10 April 2021)
  32. NATO Press Conference 16 February 1995. The flights were also denied in a report to the US Senate. See US Actions Regarding Iranian And Other Arms Transfers To The Bosnia Army, 1994- 1995 (available from http://www.globalsecurity.org/intell/library/congress/1996_rpt/bosnia.htm accessed 10 April 2021)
  33. Mark Urban,UK Eyes Alpha (Faber and Faber, London, 1996) p.214
  34. Mark Urban, UK Eyes Alpha (Faber and Faber, London, 1996) p.214
  35. Hugh Smith, ‘Intelligence and UN Peacekeeping’ (Survival, Vol.36 No.3 pp 174-192) p.177
  36. Walter Down, ‘UN Peacekeeping Intelligence’, 2010, https://www.walterdorn.net/79-united-nations-peacekeeping-intelligence, accessed 25 April 2021
  37. For example https://careers.un.org/lbw/jobdetail.aspx?id=149284
  38. Walter Down, ‘UN Peacekeeping Intelligence’, 2010, https://www.walterdorn.net/79-united-nations-peacekeeping-intelligence, accessed 25 April 2021

Related posts

#WavellReviews “Armies of Sand” by Kenneth Pollack

The Wavell Room Team

Uncrewed Drones in Naval Warfare: A Force Multiplier?

Andy Wong


Commander Tom G Sharpe OBE RN (Retd)

Leave a Comment