Ethnic Federalism in a Dominant Party State: The Ethiopian Experience 1991-2000 Lovise Aalen r 2002: 2

party and not the opposition which controls the state (Pausewang and Aalen
2001).
The centralised party structure of the EPRDF is clearly contradictory to the
provisions of the federal and regional constitutions, which give these levels the
right to self-administration. It promotes upward accountability to the party
organs above rather than downward accountability to the people of the region,
woreda and kebele. The constitutional rights for the regions to formulate and
implement plans and policies are severely diminished by the fact that the
regional governments, which are all under the EPRDF’s hegemony, follow the
centrally designed policies and five-year plans. In theory, the blurred borders
between state and party apparatus do not necessarily lead to a centralised and
unfederal division of power. If both state and party structures are decentralised,
they could strengthen regional autonomy, even though borders between the
party and the state are indistinct. But in the Ethiopian case, the party structures
are centralised, and when the state and the party are the same, this leads
inevitably to a centralised division of state power.
“Dual administration” – politics outside the legal framework
The political relationship between the regional and federal government is
regulated by both formal structures, defined in the constitution and various
proclamations, and practices outside the legal framework, that are more or less
formalised. The activity of the so-called Regional Affairs Department in the
regions is one of the semi-formalised practices that intervenes in federal-
regional relations. The Regional Affairs Department is a part of the Prime
Minister’s office in the federal government. Officially, it is established to
facilitate the communication between federal and regional government and give
advice to the Prime Minister on federal affairs (Interview Lul Seged, August
2000). But as mentioned above, it is hard to distinguish between the party and
the state administration, and in many cases the borders between the activities of
the department and the party are blurred. Although the activities of the
department are led by bureaucrats, the assistance has a relatively political
character and experts assigned to the regions play a prominent role in the
governing of regional affairs.
All the lowland regions, Benishangul-Gumuz, Gambella, Somali and Afar,
are currently being supervised and assisted by the Regional Affairs Department
(Interview Lul Seged, August 2000). The department has three sub-offices

---

C M I
86
dealing with development, human resource capacity building and political and
administrative affairs in the four regions. They implement projects at all levels
and send advisors and experts to assist the regional governments. Some
advisors are based in the regions for a two years period, while others travel to
the region frequently to monitor the situation. The assistance is generally non-
formalised and does not have any legal framework to operate within. It is given
on the basis of requests from the regions or by initiatives from the Department
itself. The assistance to the four lowland regions does not have any time limit,
but the advisors will be deployed there “until the regions are able to run their
own affairs” (Interview Lul Seged, August 2000).
In the longer run, the assistance provided by the Regional Affairs
Department might enhance the lowland regions’ capability to utilise their
constitutional rights to administer their own affairs. But currently the central
advisors virtually run the regional governments and hinder genuine self-
administration. John Young has described how the central assistance works in
Gambella and Benishangul Gumuz (Young 1999). Although the
representatives from Regional Affairs Department officially are known as
technical experts, they participate in regional council meetings and intervene
directly in regional affairs. They exceed their mandate as consultants and
assistants, and become more like managers that the regional government is
accountable to. Despite this paternalism, Young argues that the central
assistance in Gambella and Benishangul-Gumuz is generally welcomed by
regional and local governments because they lack own manpower to carry out
essential tasks. When it comes to Afar and Somali, however, the local and
regional administrations are more hostile to external assistance and less willing
to adapt traditional clan and pastoralist based authority to a more modern
bureaucracy based administration (Interview Melaku Fenta, August 2000). The
centrally appointed advisors there have been less able to carry out purely
administrative and governmental tasks than in the western lowlands of
Gambella and Benishangul-Gumuz. Due to inter-clan conflicts over cattle and
grazing land, they have been preoccupied with trying to mediate and solve
disagreements.
In addition to the semi-formalised and politico-bureaucratic assistance given
by the Regional Affairs Department, the EPRDF has deployed party officials in
the regions. These are officially known as “advisors”, monitoring and assisting
the EPRDF members or the affiliate parties in the regions, but they do not have
any formal positions or portfolios. In 1999, all of the assigned advisors were
TPLF cadres.
28
The most prominent of these is probably Bitew Belay in
SNNPRS, who was the head of Regional Affairs Department from 1997 to
2001 and member of the TPLF executive committee. Although his major area
of responsibility was the southern region, he also travelled frequently to other
regional “hotspots” like Somali region, to reconcile contending political forces
or replace inefficient or corrupt regional administrators through so-called peace
conferences (Interview Dawit Yohannes, November 2000).
28
According to the chairman of the opposition party ONC, the TPLF cadres assigned to the
regions were Solomon Tesfaye for Oromiya, Hilawi Yoseph for Amhara, Bitew Belay from
SNNPRS, Zarai Asgedom for Somali and Afar, Samuel Gebre-Mariam for Benishangul-
Gumuz and Gambella, and Mulugeta Alemseged for Addis Ababa (Merera Gudina 2001: 10).

---

C M I
87
The activities of the TPLF cadres in the regions are confirmed by
representatives of the EPRDF and the central bureaucracy, but they present it
as if the party advisors are “helping to introduce the federal system on the
ground” (Interview Lul Seged, August 2000). As my studies in SNNPRS will
show, however, the TPLF cadres constitute a second administrative structure in
the regions. The regional officials in the regular and formal administration are
to a large extent accountable to the TPFL cadres and dependent on them to
remain in their positions. Other studies of the Ethiopian federal system have
also pointed at the existence of a second administrative structure in the regions.
Aklilu Amraham (2000: 21) states that “[…] in each regional state, political
“advisors” are placed to guide and supervise the decisions of top executives of
the regional government. Virtually every critical decision has to be considered
by these ‘watchdogs’.” Jon Abbink (1995: 196) affirms that “EPRDF cadres
and army units in most of the country form a kind of dual administration.
They are acting behind the locally elected administrators, and the local officials
are ultimately accountable to and dependent (for their jobs) on the EPRDF
representatives.”
One of the mechanisms that the ruling party is using to maintain its control
in the regions all down to kebele level is the gimgemma, which implies public
evaluation of politicians and bureaucrats, plans and programmes. This practice
was originally initiated by the TPLF during the armed struggle to evaluate the
mistakes and progresses of their military strategies (Interview Solomon Enqui,
August 2000). It was legitimised by the claim that it made the leaders more
efficient and accountable to the people. Today, the system of evaluation is
introduced at all levels of government, both in political and bureaucratic
organisations (Young 2000: 10). There are different kinds of evaluation of
political leaders. On the higher levels in the regional administrations, the
sessions are mostly conducted in meetings were only the evaluated person (s),
witnesses and the chairman of the session are present. On lower levels, in the
woreda and in the kebele, many evaluation sessions are open to the public
where the audience can directly contribute with their views. Some gimgemma
are pre-planned and conducted regularly, while others have a more immediate
character and are arranged as a result of a request from the public or the party
discovering an administrator’s “misconduct”. In all the different types of
evaluations, the evaluated person is dismissed if s/he fails to justify his/her
performance. In all the four lowland regions, the assigned TPLF cadres have
frequently used gimgemma to replace inefficient and corrupt regional officials.
The deputy ministers and governments of Somali, Gambella and Benishangul-
Gumuz have been dismissed and some imprisoned as a result of gimgemma
conferences that exposed mal-administration and corruption (Young 2000: 10).
Little research has been done on the implications of gimgemma on the power
structures in the Ethiopian political system. Young (2000) claims that the
Ethiopian evaluation measures have the potential of enhancing democracy and
encourage transparency. They give the people a voice in their administration
and make the leaders responsible to the citizens. He admits, however, that
when major political programs and leading figures in government are subject
of evaluation, there is a significant measure of orchestration to ensure that the
political objectives of the central leadership of the TPLF are achieved.

---

C M I
88
According to Jon Abbink (2000: 160), gimgemma is the most widespread
method to remove people from their jobs, and this explains the extraordinary
rapid turnover of administrators and officials. Aklilu Ambraham (2000: 27)
also links the high turn over of manpower with the evaluation practices. He
argues that gimgemma has become more punitive than corrective and
educating. My interviews from Tigray show that even in this region, where the
TPLF’s ideas have substantial popular support, people fear gimgemma. Many
TPLF officials claimed that those who criticise the evaluation practice are
essentially intellectuals or civil servants who cannot stand to be evaluated by
peers or subordinates (Interview Hassen Shiffa, July 2000). My experience,
however, was that not only intellectuals, but also people on the ground were
critical to gimgemma. Although many people defended the practice, some saw
it as the “soap of the TPLF”, efficient for cleaning the hands of the party cadres
and the decision making process look transparent (Informants interviewed in
Tigray in May 2000 during the elections and July/August 2000). In SNNPRS,
the scepticism to gimgemma was more outspoken than in Tigray. Since the
evaluation idea derives from the TPLF, people in the south viewed it as a
practice imposed from outside or above (Interviews of people, bureaucrats and