Ending Civil Wars: A Case for Rebel Victory? — Comprehensive Study Notes (Toft, 2010)
Ending Civil Wars: A Case for Rebel Victory? — Comprehensive Study Notes
Note: These notes summarize Monica Duffy Toft’s argument and the empirical findings from International Security (Spring 2010) on civil war termination, with a focus on how termination type influences long-run peace, democratization, and development. The notes include key definitions, data/work methods, major results, conceptual mechanisms, case examples (notably Uganda), policy implications, and the author’s theoretical synthesis.
Overview and central puzzle
Post-WWII policy preference: end civil wars with negotiated settlements using third-party incentives and good offices to halt violence and preserve combatants.
Alternative view since Luttwak’s Give War a Chance: allow wars to end only if one side achieves military victory.
Empirical trend (1990s onward): negotiated settlements become the dominant termination type, but outcomes often poor in terms of durability; many end in recurrence.
Main question: Which termination type yields the most durable peace, and under what conditions? Is victory by one side more durable than negotiated peace, and why? What about democratic reforms and economic development after war?
Toft’s core claim: durable postwar order requires a balance of carrots and sticks; negotiated settlements alone (heavy on cooperation, light on punishment) have limitations. Rebel victory, despite its costs, may generate more durable peace and greater democratization when accompanied by credible security-sector reform and institution-building. The end of the Cold War context helped explain a rise in negotiated settlements, but longevity varied.
Why this matters: long-run peace, democratic consolidation, and development depend on how wars end, not just how many lives are saved in the short term. Also, the postwar environment (weak states, arms flows, regional spillovers) affects the likelihood of recurrence and regime outcomes.
Key concepts and terminology
Termination types (indicators of ending civil war):
Negotiated settlement: end of violence with a peace agreement that includes postwar power-sharing or governance terms; third-party actors may facilitate, but third-party enforcement is often limited.
Cease-fire/stalemate: violence halted without agreement on postwar governance; no comprehensive power-sharing terms.
Victory: one side explicitly defeats the other (military victory); terms of surrender may be near-unconditional.
Recurrence: re-emergence of civil war after an end to a previous war.
Independent variable (in Toft’s analysis): termination type.
Dependent variable: civil war recurrence.
Control variables considered:
Identity-based vs. non-identity wars (ethnic/religious roots).
Territorial vs. non-territorial conflicts.
War-related deaths (log of total deaths, and log of deaths per month for intensity).
Other robustness checks include regime type and third-party intervention (found not consistently significant in all models).
Postwar political regime and development indicators:
Polity score (Polity IV): ranges from −10 (most autocratic) to +10 (most democratic).
GDP growth as a proxy for postwar prosperity.
Security-sector reform (SSR): post-conflict reform of security institutions (military, police, intelligence, militias) to create a credible and controllable security apparatus after war.
Some formulas and scales used in the analysis (LaTeX):
Polity score range: -10 \, \le \; \text{polity} \; \le \; 10
For governance and conflict cost variables, the analysis uses logs (e.g., the natural log of total deaths or the log of deaths per month): \text{log}( ext{Total Deaths}),\quad \text{log}\left(\text{Deaths per Month}\right)
Effect sizes reported as first-differences in probability of recurrence (in percentage-point terms): e.g., a negotiated settlement raises recurrence by about +0.27 (27 percentage points);
a victory lowers recurrence by about -0.24 (24 percentage points).Two standard deviation (2σ) increase in violence intensity is used conceptually to benchmark effects of violence level on recurrence; specific numeric translation depends on the data distribution in the model outputs.
Data scope and endpoints:
Data set includes all civil wars fought 1940–2007; 137 wars met inclusion criteria; 118 ended with no violence for at least five years.
For the logit analysis, wars ending by 2002 were included; some decadal summaries use 1940–2000.
Key sources for prior literature include Luttwak (1999), Wagner (1993), Licklider (1995), Doyle & Sambanis (2000), Fearon & Laitin (2003), Hartzell & Hoddie (2007), and Toft’s prior work on territory and ethnicity.
Data and methods (empirical strategy)
Independent variable: termination type (three main categories: negotiated settlement, cease-fire/stalemate, victory).
Dependent variable: civil war recurrence (binary: recurrence vs. no recurrence).
Data set construction (1940–2007): inclusion criteria for wars
(1) main objective: control/governance;
(2) at least two organized combatants;
(3) one side is an internationally recognized state;
(4) average annual battle deaths ≥ 1,000;
(5) death ratio ≥ 95% to 5% (strong side must have suffered at least 5% of casualties);
(6) war begins within the borders of an internationally recognized state.
Resulting data: 137 wars qualified; 118 ended with a period of calm of at least five years.
Empirical approach:
Descriptive statistics by termination type and decade (1940–2000).
Correlation checks among key variables.
Rare-events logit model to estimate the impact of termination type on recurrence, with controls for identity/territory, death toll, and casualty rate.
Additional checks: a Cox proportional hazards model to confirm logit findings.
Summary of key data points:
Table 1: Frequency of civil wars by decade (1940–2000).
Table 2: Ended civil wars by decade; notable rise of end-of-war victories in the 1990s (vs. prior decades).
Figure 1: Decadal change in termination type shares (1950s–1990s).
Table 3: Estimated effects on recurrence by termination type (logit), with first-difference interpretation.
Table 4–5: Deaths and duration comparisons by termination type; tests for significance.
Table 6–7: Mean polity scores before and after wars; differences by termination type.
Notable patterns from the data:
Recurrence likelihood: wars ending in victory are roughly 1/2 as likely to recur as those ending in negotiated settlements; cease-fires/stalemates fall in between.
Rebel victories vs government victories: rebel wins tend to be more durable and associated with better postwar democratization outcomes than government victories.
Time horizon matters: longer-term data (beyond five years) reveals stronger patterns about durability and political trajectories than short-term snapshots.
Major empirical findings
Recurrence by termination type (1940–2000, with post-1990s emphasis):
Victory: recurrence ~12% (10 of 81 cases).
Negotiated settlements: recurrence ~22% (5 of 23).
Cease-fires/stalemates: recurrence ~31% (4 of 13).
=> Negotiated settlements are more prone to recurrence than victories, and this effect is stronger for settlements than cease-fires.
Statistical results (logit, with first-differences and controls):
A negotiated settlement increased the likelihood of war recurrence by about +27 percentage points relative to other terminations.
A military victory decreased the likelihood of recurrence by about -24 percentage points relative to other terminations.
Rebel victory reduces recurrence relative to government victory by more than 20 percentage points (e.g., moving from roughly -68% to -44% recurrence in the cited hazard range in one model specification).
These effects are statistically significant (p ≈ 0.01 in key specifications).
Durability and type of outcomes:
Rebel victory tends to generate more durable postwar outcomes and greater democratization than government victory in many cases.
Negotiated settlements often produce a longer slate of political institutions (executive, legislature, judiciary) and demobilization/integration provisions, but lack robust, credible enforcement mechanisms (the “sticks” are weak or under-implemented).
Third-party enforcement is often limited and not sufficiently credible to sustain long-term compliance.
Postwar political trajectories (democratization and regime type):
Polity scores five years before war: mean ≈ -3.63 (n ≈ 74 states involved in civil war).
Five years after war: democratization is heterogeneous; rebel victories tend to be associated with greater long-run liberalization, while negotiated settlements tend to produce higher short-run liberalization followed by a drift toward authoritarianism.
Table 7 shows that victory by rebels is associated with more democratic outcomes over the longer horizon; negotiated settlements are linked to higher authoritarianism after about five to twenty years.
Figure 2 illustrates a pattern: after rebel victories, autocracy declines over the longer term, whereas after government victories autocracy tends to rise or remain high; negotiated settlements show an uptick in democratization initially but trend toward greater authoritarianism later.
Postwar economic outcomes (GDP growth):
Figure 3 indicates GDP growth trajectories are not reliably different across termination types in the short-to-medium term; rebel-victory cases may show an initial GDP decline but converge with others over time.
Overall, economic growth trends postwar do not appear strongly tied to termination type.
Robustness and limitations discussed:
Some controls (e.g., regime type, third-party interventions) did not consistently reach significance across all models.
GDP data limitations and missing values affect some conclusions about economic outcomes; GDP results are thus more tentative than political outcomes.
The analysis emphasizes the importance of long-run horizons (beyond five years) for assessing stability, democracy, and development after civil wars.
Mechanisms and theory: why rebel victory can yield more durable outcomes
Core theoretical contrast (two camps):
Negotiated settlements: aim to preserve both sides as political actors; rely on carrots (benefits) and on threats that enforcement agencies may impose if terms are violated.
Victory by one side: eliminates the other as a political actor; the winner gains bargaining power but must address legitimacy and governance during transition.
Why rebel victory can outperform negotiated settlements in durability and democratization:
Rebels often begin with legitimacy and capacity-building advantages when their leadership can frame reforms and respond to grievances; postwar legitimacy can be built by promising and delivering reforms that appeal to domestic publics and international audiences.
Rebels’ governance capacity is often endogenous to their victory; successful rebels build institutions during conflict (military, political networks, mobilization mechanisms) that can be repurposed for postwar governance.
Rebel victories tend to be associated with the emergence of new or reformed political orders that can incorporate liberalizing reforms, attracting external legitimacy and internal support.
In contrast, negotiated settlements frequently emphasize institutional reform and power-sharing but may neglect security-sector reform (SSR) and credible mechanisms to punish noncompliance; this weakens the long-run stability of settlements.
The Uganda case as an illustrative example:
Uganda’s NRA/NRM victory (1986) produced a set of durable institutions: local-level Resistance Councils, expanded representative government, broader social inclusion, and a capable security apparatus.
Economic reforms and liberalization contributed to growth: per capita GDP rose from 58% of its independence level in 1986 to 69% by 1994 and 80% by 1997.
Democratic reforms, while not perfect (Museveni’s tenure and term-limit debates persisted), nonetheless produced significant democratization relative to the prewar regime.
The Uganda example is used to argue that rebel victory can be associated with durable peace and substantive reforms when accompanied by inclusive institutions, broad-based governance, and credible security-sector reform.
Why negotiated settlements often fail to sustain peace:
They provide credible benefits (no violence, participation in elections, governance promises) but weak enforcement; credible “sticks” (punishments for defection) are often limited by the credibility of third-party enforcement.
Over time, settlements can generate political openings that are re-closed as elites reassert power, leading to renewed conflict or authoritarian rollback.
The combination of weak SSR and insufficient postwar political integration can lead to recurrence and rising repression after initial liberalization.
The broader policy takeaway: successful peace requires design that links benefits with credible consequences for defection (cs: carrots and sticks), with particular emphasis on robust SSR and postwar governance reforms.
The Uganda case in detail (illustrative narrative linked to theory)
Context prior to 1986: long period of repression and violence under Idi Amin; GDP decline (roughly 25% 1970–1980) and export decline (≈60%), polity score as low as -7.
Postwar reforms under Museveni and NRA/NRM:
Establishment and expansion of governance institutions (Resistance Councils at village to district level).
Broadening of political inclusion, absorption of former opponents, and integrated security forces that included soldiers from opposing factions.
Emphasis on economic liberalization and growth; strong focus on development and stabilization.
Outcomes:
Significant improvements in GDP per capita relative to independence baseline; stabilization of political order; ongoing but managed political tensions.
Museveni’s regime maintained power while implementing liberalization, with continued debates about term limits and multiparty politics.
Implications for the theory:
The Uganda case provides a concrete example of rebel victory generating major postwar reforms and credible governance capacity, yielding durable peace even as political liberalization evolves over time.
Policy implications and recommendations
The two extremes of policy (all carrots vs. all sticks) are counterproductive when treated as universal prescriptions for civil war termination.
A more effective approach combines elements of both termination types, with emphasis on:
Security-sector reform (SSR): integrate and reform defense and internal security institutions as part of the peace deal; ensure demobilization, demilitarization, reintegration (DDR) is complemented by credible SSR to avoid postwar militarization or fragmentation.
Credible enforcement: designs for enforcement and monitoring by third parties or international actors that are credible and sustainable beyond the initial bargaining phase.
Postwar governance design: robust, inclusive constitutions, power-sharing provisions where appropriate, and credible timelines for democratic reform and elections; ensure that postwar governance structures can absorb former combatants while maintaining state authority.
Economic reconstruction: link peace accords to tangible development programs and investments that create a stable economic base and incentivize compliance.
Colombia and El Salvador as cautionary/cautionary examples:
Colombia’s long peace processes showed the risk of militarized, disjointed implementation post-conflict; later Plan Colombia and consolidation efforts aimed to defeat rebels and unify state control.
El Salvador’s peace agreement included strong security-sector transformation provisions; the case shows how integrated carrots-and-sticks provisions can help sustain peace but require committed implementation.
Practical takeaway for international actors:
Enduring peace requires credible incentives and credible penalties for noncompliance, embedded in security-sector reform and strong governance institutions.
External actors should commit to sustained engagement beyond the initial peace agreement, with clear governance and SSR roadmaps, to prevent relapse into conflict.
Important caveat: rebel victories are not a universal panacea; the Ugandan example is illustrative but not universally representative. The key is implementing credible SSR and inclusive governance to convert victory into durable peace and development.
Postwar regime change, democratization, and development: nuanced patterns
Democracy and termination type:
Negotiated settlements do not guarantee durable democracy; rebel victories tend to be associated with greater long-run democratization (though not guaranteed) relative to government victories.
In the long run, rebel victories can facilitate liberalization and reduced autocracy, whereas negotiated settlements may initially liberalize but trend toward authoritarian outcomes after the first election cycles.
Economic development patterns after civil wars:
GDP growth appears generally similar across termination types in the medium term; rebel-victory cases may exhibit early declines but often rebound to levels comparable to other endings after about a decade.
Economic trajectories depend on a broader set of factors (institution-building, external investment, macroeconomic policy) beyond the immediate termination type.
Theoretical synthesis: a theory of civil war termination
Central insight: there is no one-size-fits-all; stable peace requires a balance of incentives (carrots) and credible punishment (sticks).
Two broad camps, with endogenous dynamics:
Negotiated settlements work best when they include credible security-sector reforms, strong institutions, and credible enforcement mechanisms, thereby reducing incentives for future defections.
Rebel victories work better when they create legitimate governance structures that can be reformed through inclusive institutions and liberalization, thus generating a durable peace and potentially more democratization.
A general theory that reconciles the empirical patterns:
Durable peace emerges when the postwar settlement (whether negotiated or victory-based) embeds incentives for compliance, credible punishment for defection, and robust SSR to prevent the relapse of violence.
Negotiated settlements need robust postwar enforcement and SSR to avoid recurrence; pure carrot-based agreements without credible sticks may be unstable.
Rebel victories can be associated with stronger postwar reform trajectories if the victor can build legitimate governance and inclusive institutions; this tends to reduce autocracy and improve democratic outcomes in the long run.
Policy synthesis: promote settlements that mix benefits with credible threats of punishment, and emphasize SSR as a core part of peace implementation to raise the probability of enduring peace, democracy, and development. If credible, strategic backing for rebel-led transitions can be considered when accompanied by strong SSR and inclusive governance provisions.
Additional notes on sources, tables, and figures (quick reference)
Tables and figures:
Table 1: Frequency of civil wars by decade (1940–2000).
Table 2: Ended civil wars by decade with termination type shares (notably 1990s rise of negotiated settlements).
Figure 1: Trends in termination types across decades.
Table 3: Logistic regression results (logit) for recurrence by termination type, with first-differences (e.g., +0.27 for negotiated settlements; −0.24 for victories; rebel victories relative to government victories).
Table 4–5: Deaths and duration analyses by termination type; mixed results on mortality and recurrence significance.
Table 6–7: Mean polity scores before and during war, and by termination type; rebel victories linked to greater long-term democratization; negotiated settlements linked to higher short-term openness but more long-run authoritarian drift.
Figure 2: Change in polity score before/after civil war by termination type (5, 10, 20 years after).
Figure 3: GDP growth trajectories before/after civil war by termination type.
Case example: Uganda (1986–1997+), rebel victory under NRA/NRM, governance reforms, and rapid GDP growth recovery; the World Bank cited Uganda as a major turnaround due to farm/firm/government reform balance and capacity-building outcomes.
Policy relevance: emphasize SSR, credible enforcement, and sustained external support to sustain peace after settlements; balance carrots with credible sticks; consider rebel victories as a vehicle for reform when conditions support credible governance and inclusion.
Takeaway messages for exam-ready notes
Ending civil wars is not simply a choice between “negotiate now” and “let them fight until victory”; the durability of peace depends on credible enforcement and strong postwar institutions, especially SSR.
Rebel victory, often viewed as risky, can yield more durable peace and greater democratization in the long run when coupled with credible governance reforms and security-sector reform.
Negotiated settlements, while appealing for immediate mercy and inclusion, need robust enforcement mechanisms and postwar SSR to avoid recurrence and authoritarian drift.
The time horizon matters: five-year snapshots can mislead; long-run analysis (10–20 years) reveals that rebel victories often outperform negotiated settlements in durability and democratization, though not universally.
Uganda serves as an instructive case where rebel victory was associated with strong postwar reform and development progress, illustrating the potential advantages of victory when accompanied by institutional rebuilding and broad-based governance.
If you want, I can convert these notes into a tighter 1-page outline or a quick-reference table of key findings by termination type (with all the exact percentages and p-values cited in the article).