F a S T Update Ethiopia Quarterly Risk Assessment August to October

Early warning unit F A S T Update

Ethiopia

Quarterly Risk Assessment

August to October 2003 T S

A F

FAS T Update

Ethiopia | August – October 2003 | Page 2

T Contents

Country Stability and Relative Forceful Actions 3

Relative Civil Direct and Civil Violent Actions 4

Relative Civil Direct and Relative Government Direct Actions 5

Goldstein Average International Conflict and Cooperation 6

Appendix: Description of indicators used 7

The FAST Early Warning System 9 S

Contact FAST: Phone: +41 31 330 12 04 Fax: +41 31 330 12 13 [email protected] www.swisspeace.org/fast F Subscription: www.swisspeace.org/fast/subscription_form.asp

FAST Update

Ethiopia | August – October 2003 | Page 3

Country Stability and Relative Forceful Actions 1.1 ©FAST 1.0 0.9 0.8 0.7 0.6 0.5 T 0.4 0.3

0.2 Country Stability 0.1 Relative Forceful Actions 0.0 -0.1 11/02 12/02 1/03 2/03 3/03 4/03 5/03 6/03 7/03 8/03 9/03 10/03 Average number of reported events per month: 144 Indicator description: see appendix

Risk Assessment:

· During the past two months the Country Stability index has been at a constant 0.9. The proportion of S Direct Actions involving use of physical force (Relative Forceful Actions) has also fallen considerably. This chart suggests that the situation in Ethiopia is less conflictual than has it has been for quite some time. · Yet this positive picture hides a tense stalemate resulting from the government’s continued refusal to accept the Eritrea-Ethiopia Boundary Commission (EEBC) ruling which would lead to the cession of Badme and other disputed border areas to Eritrea. Bound up with this issue is the fate of the Ethiopian People’s Revolutionary Democratic Front (EPRDF) ruling coalition and the struggle for control by and within its dominant party – the Tigray People’s Liberation Front (TPLF). · This struggle has involved a continuation of moves by the regime of Prime Minister Meles Zenawi to tighten its grip on power through a range of more or less repressive – though predominantly not ‘forceful’ – means. Opposition to these moves has been growing alongside continued dissent over loss of territories to Eritrea and the possibility of EEBC compliance. Though again, this has so far mostly been without the widespread use of force despite a peak in military operations by rebel groups in August.

· The humanitarian situation remains dire, with ongoing impacts of this year’s severe drought which A threatened 13 million people with starvation, now overshadowed by a looming malaria epidemic which could affect as many as 15 million. The return of substantial rains between July and September, which greatly favoured the reproduction of malaria mosquitoes, has improved crop prospects but has done little to ameliorate the situation in pastoral zones: As in the recent past, intense competition for grazing resources by migrating herders is likely to lead to further violent clashes over the coming year unless pastoralist groups can be appropriately assisted. · Despite talk in some military circles of an impending resumption of war with Eritrea, and unless Eritrea agrees to UN suggestions that both countries enter into dialogue about local departures from the EEBC border delineation, the present status – arguably more costly to Eritrea than Ethiopia – seems set to persist as long as the present government remains in control. This may be the case until the general election, due in 2005, but the earlier removal of Meles by elements within or outside the TPLF remains a distinct possibility, especially if he tries to accept the EEBC ruling. In this event the pressure for war – and thus on Eritrea to show flexibility – may increase. F

Philip White

FAST Update

Ethiopia | August – October 2003 | Page 4

Relative Civil Direct and Civil Violent Actions 1.1 ©FAST 1.0 0.9 Relative Civil Violent Actions Relative Civil Direct Actions 0.8 0.7 0.6 0.5 T 0.4 0.3 0.2 0.1 0.0 -0.1 11/02 12/02 1/03 2/03 3/03 4/03 5/03 6/03 7/03 8/03 9/03 10/03 Average number of reported events per month: 144 Indicator description: see appendix

Risk Assessment: · The amount of Civil Direct (or conflictive) actions has remained at a low level during this analyzed period, S yet the proportion of Direct Actions that were ‘violent’ or highly conflictive has fluctuated dramatically. These trends reflect the sporadic nature of inter-ethnic violence, rebel attacks and threats set against a varying number of direct but non-violent actions including anti-government demonstrations. · The troubled Moyale woreda, located on the Kenya border and marking the split between Borena zone in Oromia region and Liben zone in Somali region, saw a continuation of clashes between Borena and other pastoralist groups during August and further inter-ethnic fighting involved Gabbra and Gerri groups in Liben in October leading up to the Somali region elections. In Bench Maji zone in SNNPR Surma tribesmen reportedly killed 17 and displaced over 300 households among Dizzi and Minit groups in October. · The August peak owes more, however, to anti-government actions. The Oromo Liberation Front (OLF) has sustained its level of military action, clashing with both government forces in Borena zone and Kenyan troops across the border in several operations, with over 30 deaths. However, it denied responsibility for a bomb explosion at Adigala on a train travelling from Djibouti to Addis Ababa, killing two civilians – this and an attack on a Save the Children (UK) vehicle in Haraghe zone in October may have been carried out by Al-Ittihad terror groups known to be targetting Ethiopia. The Ogaden National Liberation Front (ONLF) A has also been active, with attacks on army units at Degeh Bur and Degeh Madow in Somali region and the blowing up of a bridge connecting Fik zone with Harer. The United Liberation Forces of Oromia (ULFO), a coalition of several Oromo rebel organizations, mounted two further armed operations in Moyale woreda. · In September, 19 TPLF Central Committee members wrote a petition to Prime Minister Meles Zenawi in which they threaten a return to war if the EEBC decision on Badme, Irob and parts of Afar was accepted. Meanwhile the Union of Ethiopian Democratic Forces (UEDF), an umbrella grouping of 15 opposition parties formed in August to contest the 2005 elections, declared that it too rejects the EEBC ruling. · There is no reason to suppose that inter-ethnic violence and rebel action will quieten significantly in the foreseeable future: while TPLF and border community threats to fight rather than comply with any government acceptance of the EECB ruling can be expected to continue.

Philip White F

FAST Update

Ethiopia | August – October 2003 | Page 5

Relative Civil Direct and Government Direct Actions 1.1 ©FAST 1.0 0.9 0.8 Relative Government Direct Actions Relative Civil Direct Actions 0.7 0.6 0.5 T 0.4 0.3 0.2 0.1 0.0 -0.1 11/02 12/02 1/03 2/03 3/03 4/03 5/03 6/03 7/03 8/03 9/03 10/03 Average number of reported events per month: 144 Indicator description: see appendix

Risk Assessment: · As has been the case for most of the last year, government actions have been somewhat more conflictive in character than those of the civil sector, with little fluctuation during the reporting period in either case. S · According to OLF sources government operations against rebel forces included the planting of a landmine in Moyale woreda, which was to have been blamed on the OLF but detonated prematurely killing ten government agents, and a number of attacks on suspected OLF members in Oromia region in September. Other reported Direct Government Actions included the arrests of six members of the Southern Ethiopian People's Democratic Union (SEPDU) in Kefa, nine Oromo National Congress members in Ambo and 135 other Oromos in Bale zone, and several officials of the Somali regional government. There were also government raids on CAFPDE and EDUP offices and a ban on an EDUP meeting in Mekele on the border issue, the dismissal of the University of Addis Ababa Vice President, searches of Arab and Somali occupied houses in the capital by special forces on suspicion of terrorist links, and mass sackings of government employees in Addis Ababa on charges of incompetence and corruption. In ongoing trials of the Derg regime the former Minister of Public Security Colonel Tesfaye Wolde-Selassie and four others were sentenced to death in the Federal High Court. · The tide of civil and opposition party complaints over coercive and autocratic government policies, in

particular the ‘voluntary’ resettlement programme and recent moves to consolidate power at the centre at A the expense of regional governments, the armed forces and the press, has continued to rise. In East Wellega zone, some 300 resettlers demonstrated over the lack of promised services and assaulted Oromia Chief Executive Junedi Sado. With international NGOs highlighting dozens of deaths among resettlers each week due to malaria and starvation, government appeals for donor support to the programme have been received negatively. · These issues compound concerted opposition to the EEBC ruling which is mounting both within the TPLF, in Amhara, and opposition circles, leaving Meles with little room for manoeuvre and lending weight to his own assessment that his political future hangs on Ethiopia’s claim to Badme. Pressure on Meles to resist international pressure to give way on the border issue – but also on the TPLF’s ability to sustain its political dominance – can only intensify as the 2005 elections loom closer. Meles will be ineligible for re-election to a third term barring a change in the Constitution. Further claims for asylum of several diplomats, parliamentarians and other political figures while abroad have underlined these mounting stresses within the political system, which point to a potential for increased domestic if not interstate conflict during the F coming months. Philip White

FAST Update

Ethiopia | August – October 2003 | Page 6

Goldstein Average International Conflict and Cooperation 10 ©FAST 9

8 Average International Conflict 7 Average International Cooperation 6

5 T

0 11/02 12/02 1/03 2/03 3/03 4/03 5/03 6/03 7/03 8/03 9/03 10/03 Average number of reported events per month: 144 Indicator description: see appendix Risk Assessment: · As this chart suggests, events have been neither strongly conflictive nor cooperative during the reporting period. This reflects a stalemate situation in which international relations have continued to be driven primarily by Ethiopia’s continued rejection of the EEBC ruling while maintaining that it does not intend to return to war with Eritrea. This has necessitated another postponement of the demarcation exercise and S extension of the UN peacekeeping mission (UNMEE) mandate. · In early August incursions by groups of armed and uniformed Ethiopians into the Temporary Security Zone (TSZ) raised border tensions, even though there was no aggression and their purpose was ostensibly to find flat ground to play football. More serious was an incident in October in which a group of armed, uniformed Ethiopians allegedly entered the TSZ and opened fire, killing an Eritrean militiaman. The Ethiopian government denied that its troops were involved, and both governments assured UNMEE of cooperation to avoid any recurrence. Occasional casualties from landmines – some newly-laid – and unexploded ordnance have continued. · On other fronts, repeated clashes between OLF cadres, other pastoralist groups, and Kenyan security forces across the southern border with Kenya led to several deaths and subsequent Kenyan restrictions on cross- border travel and trade from Moyale woreda. Ethiopian government soldiers were involved in some of the incursions, which related to OLF operations as well as cattle raiding and competition for pastures and

water. Djibouti’s move to expel all illegal immigrants by 15 September affected some 30,000 Ethiopians A whom the government was obliged to assist to enter the country. The government has accused Egypt of frustrating its efforts to secure international loans to develop Nile water resources, but has continued to forge links with Sudan with an agreement on a rail link into Ethiopia from Port Sudan which would expand maritime access while the boycott on Eritrean ports is maintained. · Ethiopia continued to receive international assistance for development and to combat the humanitarian crisis, as well as military aid. Alongside diplomatic exhortations donors have continued to use the threat of aid withdrawal to try to lever the government into accepting the EEBC decision, but to little effect. Steps in this direction are included in a bill currently under consideration by the US Senate Foreign Relations Committee, though this would exclude humanitarian assistance – and, significantly, aid related to the War on Terror. · In the absence either of broad domestic support for accommodation with the EEBC decision, an alternative

peace proposal, or any real prospect of war being reignited from the Eritrean side, the government seems F likely to seek to maintain the status quo for as long as possible. Depending mainly on domestic political developments, this is a difficult and uncertain balancing act. Philip White

FAST Update

Appendix: Description of indicators used | Page 7

Variable Name Description / Definition Direct Actions are conflictive events that can be assigned to the following event types: threaten, demonstrate, reduce relationships, expel, seize and force. These categories encompass direct actions Direct Actions regardless of origin or target. Proportion of Direct Actions compared to All Actions (WEIS-Categories). Direct Actions are conflictive events that can be assigned to the following event types: threaten, Relative Direct demonstrate, reduce relationships, expel, seize and force. These categories encompass direct actions T Actions regardless of origin or target. Proportion of Violent Action events compared to all direct actions. The indicator Violent Actions is a subcategory of Direct Actions that includes highly conflictive events Relative Violent especially threats with force specified, ultimatums with time limit specified, military mobilisation, Actions exercise, or display, and all reported uses of physical force by any actor. Proportion of Forceful Action events compared to All Direct Actions. Relative The indicator Forceful Actions depicts all reported uses of physical force by any actor. This includes Forceful Actions non-injury destructive acts, non-military injury-destruction, and military engagement.

Civil Sector Relative Civil Actions Proportion of all civil actions to all civil events (with non-governmental, or civil sector actors). Proportion of Civil Direct Actions compared to All Civil Actions. S Civil Direct Actions are conflictive events that can be assigned to the following event types: threaten, Relative Civil demonstrate, reduce relationships, expel, seize and force. These categories encompass direct action Direct Actions limited to non-governmental, or civil sector actors. Proportion of Civil Violent Action events compared to All Civil Direct Actions. The Indicator Civil Violent Actions is a subcategory of Civil Direct Actions that includes highly conflictive events especially threats with force specified, ultimatums with time limit specified, Relative Civil military mobilisation, exercise, or display, and all reported uses of physical force. These categories Violent Actions encompass violent action limited to non-governmental, or civil sector actors.

Government Sector Relative Government Proportion of all Government Actions to all Government Events (with political, or government sector Actions actors). Proportion of Government Direct Actions compared to All Government Actions. Relative Government Direct Actions are conflictive events that can be assigned to the following event types: A Government threaten, demonstrate, reduce relationships, expel, seize and force. These categories encompass Direct Actions direct action limited to the political sector, or government actors. Proportion of Government Violent Action events compared to All Government Direct Actions. The Indicator Government Violent Actions is a subcategory of Government Direct Actions that Relative includes highly conflictive events especially threats with force specified, ultimatums with time limit Government specified, military mobilisation, exercise, or display, and all reported uses of physical force. These Violent Actions categories encompass violent action limited to political sector, or government actors. The Conflict Carrying Capacity (or CCC) is a composed index that depicts the overall stability of the country or region of interest. The CCC is operationalized in terms of the multiplicative interaction among three Proportional measures: (1) civil contentiousness or the Proportion of civil actions that are reported as contentious or "direct" and thus challenge (at least implicitly) the state's monopoly on conflict regulation; (2) state repression or the Proportion of state actions that are reported as extra-institutional or "direct" both in response to direct challenges from the civil sector and those

Conflict initiated by the state to repress and control opposition; and (3) violent contention or the Proportion F Carrying of actions entailing physical damage to persons or property. the index is scaled between 0 and 1, Capacity where 1 means high and 0 low stability.

FAST Update

Appendix: Description of indicators used | Page 8

Variable Name Description / Definition Country The country stability index is another version of the CCC measure with minor changes in order to Stability improve the responsiveness of the index to events that influence the stability of a country.

Goldstein

Goldstein indicators are used to display Proportions of conflict and cooperation events in time. Each T event category is assigned a Goldstein rating (value), ranging from -10 (extreme conflict) to 10 Goldstein (extreme cooperation). Zero value events are excluded from these calculations. The Goldstein Average indicator is a cumulative average of all events that are classified with a Goldstein positive or negative value by Goldstein's conflict/cooperation scale. The indicator displays the mean Average of the conflict/cooperation event values, excluding all zero value events. The Goldstein Average International Conflict indicator displays the cumulative average of the Goldstein negative (Goldstein) values of all conflictive interstate or international events in a specific time Average period (means the sum of the negative Goldstein values divided by the total number of conflictive International international events). For interpretation purposes we take the absolute values (means positive Conflict values). Goldstein The Goldstein Average International Cooperation indicator displays the cumulative average of the Average positive (Goldstein) values of all cooperative interstate or international events in a specific time International period (means the sum of the positive Goldstein values divided by the total number of cooperative Cooperation international events). S

A F

FAST Update

The FAST Early Warning System | Page 9

Who are we? FAST (German acronym for “Early Analysis of Tensions and Fact-finding”) is the early warning project of swisspeace, based in Berne, Switzerland. In 1998 the Swiss Agency for Development and Cooperation (SDC) assigned swisspeace to set up a political early warning system for early identification of impending armed conflict and political crisis situations. T

What do we want? FAST aims to enhance political decision makers’ ability to identify critical developments in a timely manner, so that coherent political strategies can be formulated either to prevent or limit destructive effects of violent conflict or to identify opportunities for peacebuilding.

How do we work? FAST uses both quantitative and qualitative methods for its analysis. The centerpiece in the quantitative analysis is based on event data analysis and the respective tools developed in the framework of the Program

on Nonviolent Sanctions and Cultural Survival (PONSACS) at Harvard University. The logic of event data S analysis is fairly simple: all events considered relevant to conflict escalation and de-escalation are assigned a certain numeric value according to a distinct conflict scale. These values can then be added up for specific time intervals and graphically displayed in a curve over time. The quality and quantity of data input is crucial for the success of such a method. In order to gather the quality and quantity of data required to suit early warning purposes, FAST sets up its own local information networks (LINs) and thus overcomes shortcomings of existing information sources (e.g., international news wires). A unique set of data is collected for each country completely independently from Western news-media coverage. For qualitative data analysis, FAST contracts internationally renowned country experts.

What are our products? FAST products are risk assessments tailored to individual customers' needs. The standard product (“FAST A Update”) consists of three to five charts depicting the latest conflict related trends and a concise expert interpretation. As the time-series of collected data grow, FAST analysts will apply statistical methods to go beyond retrospective description and forecast future developments. FAST Updates are available in either hard copy or electronic form, covering time intervals as chosen by the customer.

Which countries do we monitor? Africa: Angola, Burundi, Ethiopia, Kivu region, Madagascar, Mozambique, Rwanda, Somalia Asia: Afghanistan, India/Kashmir, Kazakhstan, Kyrgyzstan, Nepal, North Caucasus region, Pakistan, Tajikistan, Uzbekistan

Europe: Albania, Bosnia and Herzegovina, Kosovo, Macedonia, Serbia and Montenegro F Middle East: Palestine