29
the group at the time of the agreement or sudden changes in the preferences on either side
(such changes can occur in the wake of a successful attempt to spoil the agreement by
extremists) can prevent implementation.
37
Once a possible agreement has been identified,
both the self-determination group and the state need to gain internal consensus in their
respective populations to implement an agreement. The number and size of internal
factions, as well as their institutions, affect the ability of both self-determination groups and
states to do this. Self-determination groups with coordinating institutions and institutional
longevity are more likely to successfully implement an autonomy agreement. States with a
low number of veto players and ample ability to compensate unsatisfied veto players are
more likely to implement an autonomy agreement.
The existence of multiple factions with heterogeneous preferences is the underlying
reason that both sides need to generate some degree of internal consensus to gain
compliance for a new autonomy agreement. How much internal compliance is necessary
depends, in part, on the institutional structure of both actors. At the implementation stage,
the number of internal factions affects the likelihood of the emergence of extremist spoilers,
and the likelihood that moderates can suppress extremists within self-determination
groups.
38
Extremists (actors with preferences on the tails of the density curve) in self-
determination groups with more heterogeneous preferences are further from the average
preference of the group. These individuals’ ideal points are so far from the average, that
37
Failed implementation does not occur because the state has insufficient support at the time of the agreement
because all veto player in the state need to agree for the offer to be made. Therefore, the state does not
experience the same type of internal enforcement issues that groups do when there are no clearly defined veto
players. Self-determination group extremists are not clearly defined veto players because 1) most of the
population will not support the agreement they want and 2) they are not always successful in preventing
implementation and agreement. For example, important progress has been made in the conflict in Northern
Ireland despite continued attempts to “spoil” the peace process.
38
For more on extremist spoiling, see Walter and Kidd (2002).
Convention