In Chapter 2, I provide a review of the literature on commitment problems and fac- tionalization in order to position my theory within the context of this wider literature. What follows here is a more limited accounting of the alternative explanations for suc- cessful implementation of a negotiated settlement. These explanations are included here for the purpose of establishing the hypotheses against which I will evaluate those listed in the previous section. Table 4.1 lists independent variables along with the proposed direction of their effect according to previous literature. This Table is a modified and extended version of Table 1.1 in Barbara Walter’sCommitting to Peace:
the Successful Settlement of Civil Wars (2002).
Hyp. Variable Hypothesized Effect on Probability
of Negotiation Success
A1 Costs of War (High) Increases
A2 Govt. and Challenger Parity Increases
A3 Democracy Increases
A4 Ethnic/Identity Issues Decreases
A5 Divisibility of Stakes Increases
A6 Mediator Increases
A7 Third Party Security Guarantee Increases
A8 Factionalization Increases
A9 Extremism Decreases
Table 4.1: Alternative Hypotheses for the Causes of Success and Failure in Conflict Resolution
The effects of the cost of war are fairly straightforward - higher costs of con- flict widen the bargaining range between the state and its opposition because when the continuation of conflict is very costly, each side will have a lower utility for war (Wittman 1979). Since it is true that the costs of war are non-trivial, then states and their challengers should always be able to find a settlement they prefer to war (Fearon 1995). In the event that locating an acceptable agreement is a challenge, increasing costs of war should mitigate this challenge. As Walter (2002) notes, the hypothesized effect of parity between a government and challenger reflects the effect of balance in the international system proposed by Organski (1968) - that is to say that parity is hypothesized to increase the chances of a negotiated resolution by in- creasing the uncertainty that surrounds the decision to go to war. Democracy in the target state is hypothesized to increase the chances of a successful resolution because
of the domestic constraints that are proposed to increase the costs of using force on a domestic population, the increased ability of democratic leaders to make credible commitments, and a tradition of sharing power in government. This argument fol- lows in the tradition of the democratic peace theory more generally, as espoused by Russett (1993) and others. Ethnic identity is hypothesized to decrease the perceived divisibility of stakes in civil conflict, increase the potential for identification and pun- ishment of outgroup members if a settlement fails, and create disincentives for those seeking compromise (Horowitz 2000, 2003; Kirschner 2010) All of this is expected to decrease the probability that any compromise will be struck. In turn, the divisibility of stakes increases the likelihood of negotiated settlement, as the high costs of war mean that when dividing the contested issue is possible, there should always be some division that is preferred by both sides to continued conflict (Fearon 1995). Media- tors are argued to promote compromise by their ability to observe and communicate the degree to which each side is cooperating and to provide external incentives for cooperation (Regan and Aydin 2006). Walter adds that without third-party security guarantees successful implementation of negotiated settlements is very rare due to the vulnerability to defection that implementation produces on both sides (2002).
Alternative Hypotheses A8-A10 are all specifically related to the issue of faction- alization within an opposition. Hypotheses A8 and A9 propose opposite effects of the same variable - factionalization. Specifically, Nilsson proposes that as the number of factions increase, the chances of weaker factions making peace deals with the gov- ernment increases because when the opposition is very divided the government can make very small offers of concessions to multiple small factions, effectively peeling these factions away from the opposition (2010). In particular, Nilsson proposes that when there are more than five factions in the opposition, deals with small factions are more likely than when there are fewer factions. Nilsson makes no specific pre- dictions regarding the chances for a broader peace settlement, but does argue that
states negotiate with smaller factions in an effort to reduce the size and effectiveness of the insurgency as a whole. In a related argument, Cunningham proposes states use a divide and concede approach to factionalized insurgencies whereby the state uses concessions to learn more about the nature of the insurgency and its factions or to advantage moderate factions over their extreme counterparts (2011). When the state offers a concession it can determine based on which groups accept that concession which are the true moderates and which the true extremists by observing what fac- tions accept the offer. Finally, the more traditional theories regarding the effect of factionalization, suggest that increased factionalization may lead to barriers to peace as a result of extremist violence (Stedman 1997; Kydd and Walter 2002; Bueno de Mesquita 2005). I contend that it is not the mere presence of factions that is signif- icant for peace prospects. Rather, it is the relative strength of the factions and how much (or little) they have in common.