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

parties are organised might reinforce or corrode the federal division of power.
Riker claimed that the structure of the party system should be regarded as the
main variable in intervening between the background social conditions and the

---

C M I
21
specific nature of federations. Whether the party organisations are centralised
or decentralised have crucial effects on the relationship between central and
regional level. “The federal relationship is centralised according to the degree
to which the parties organised to operate the central government control the
parties organised to operate the constituent governments. This amounts to the
assertion that the proximate cause of variations in the degree of centralisation
(or peripheralisation) in the constitutional structure of a federalism is the
variation in degree of party centralisation” (Riker 1964:129). Although the
degree of centralisation and decentralisation of political parties and its effects
on central-regional relations do not tell us anything about whether a state is
federal or distinguish federations from other kinds of states (King 1982: 124-
126), they might help us to identify causes of why the federal systems of the
world work so differently from one another.
Constitutional and institutional guarantees are meant to prevent political
leaders at the central level of the federation from breaking the rules or changing
the federal bargain unilaterally. Riker argues that these guarantees were
ineffectual when the political leaders at the centre also control the party on the
constituent level, because then there will be no opposition against the change.
Thus, the degree of unity between the constituent and central governments is
closely linked to changes in the federal relationship (Riker 1964:130). He
showed how the structures of the party systems affect the operation of various
federal systems. In the United States, the political parties have a highly
decentralised organisation and there is a lack of unity on the national level of
the organisation. Despite the fact that the Democrats and the Republicans
control both the regional and central level of government, decentralisation of
the two parties is sufficient to prevent the national leaders (e.g. the president)
from controlling the lower level of the organisation either by organisational or
ideological devices. In Canada, the federal government is controlled by
nationally based parties, while the regional governments are controlled by
provincial parties without any national base. Elections to the two levels are
held on two different dates. These circumstances maintain the Canadian federal
system as relatively peripheralised. The Soviet Union is an example of how the
structure of the only legal party made the federal state as centralised as
dictatorships in fully unitary states. The Communist party was highly
centralised and hierarchical, and its position as the unchallenged authority
made the federal system extremely centralised and the autonomy of the
regional states only a paper provision.
More recent studies of federal systems in one-party states illustrate how
important the organisation of the political party is for the operation of a
federation. Smith (1995b) shows how the communist party employed
totalitarian techniques in order to ensure the continuing imposition of
centralised communist party rule over the nationality-based union republics.
This was done through retaining the control of appointments, promotions and
dismissals of local party cadres and state officials in the ethno-regions, and
appointing Russians to certain regional key positions. This made them able to
ensure the loyalty of the local party leadership and further reinforce the
centralised control and the expedition of the centre’s policies. Popovski’s study
shows similar traits in the Yugoslav federation: “Despite being referred to as

---

C M I
22
federal, the new [Yugoslav] state was organised according to the communist