When Enemies Learn to Share Power: The Possibilities and Limits of Coalition in a Permanently Divided State

Part Three of a Comparative Series on Coalition Formation, Collective Action, and Political Survival in Fragmented Societies

By Brook Kidané

Sometimes the objective of coalition is not to create trust. It is to make distrust survivable. Afghanistan showed why actors who might otherwise destroy each other sometimes conclude they cannot afford to. Syria showed how such actors, still deeply suspicious of one another, can build limited mechanisms of coordination without ever resolving that suspicion. Lebanon poses a harder question than either case, because Lebanon did not simply produce an episode of cooperation. It produced a system—a durable constitutional architecture built on the premise that no community could be allowed to dominate another. That system has now existed, in one form or another, for more than eighty years. It has prevented any single faction from permanently ruling the others. It has also, with grim regularity, prevented Lebanon from being governed at all.

This is the paradox Lebanon forces into view: coalition can make coexistence possible without making the state effective. Afghanistan asked why enemies would ever cooperate. Syria asked how they could cooperate without trust. Lebanon asks a still more uncomfortable question—what happens when the mechanisms invented to prevent one faction’s domination become mechanisms every faction can use to prevent the state from acting at all?

I. The Problem of Coexistence

Lebanon’s founding predicament was not military but demographic and psychological. Emerging from the French Mandate in 1943, its Christian and Muslim communities carried incompatible fears rather than incompatible territories. Christians, and Maronites in particular, feared absorption into a Muslim-majority Arab world; Muslim communities, especially Sunnis, feared permanent subordination to a state tilted toward the West and toward France. Neither fear could be resolved through victory, because neither community could plausibly eliminate or fully subordinate the other without inviting regional intervention or renewed conflict.

Out of this stalemate came the unwritten National Pact of 1943, negotiated between President Bechara al-Khoury, a Maronite, and Prime Minister Riad al-Solh, a Sunni. Its terms were not a theory of justice; they were a trade of fears. Christians would abandon the search for Western protection and accept Lebanon’s “Arab face”; Muslims would abandon aspirations toward union with Syria and accept the legitimacy of Lebanon’s borders. The presidency was reserved for a Maronite, the premiership for a Sunni, the speakership of parliament for a Shia—an allocation of the state’s highest offices along confessional lines, with parliamentary seats apportioned on a six-to-five Christian-to-Muslim ratio drawn from a single, never-repeated 1932 census.

What this arrangement achieved should not be minimized. It allowed a state to exist that might otherwise never have come into being, because it substituted a division of offices for a contest over sovereignty itself. But what it assumed should not be minimized either. It assumed a static demographic and political snapshot could serve indefinitely as the foundation for representation. It assumed that once fear was addressed procedurally, the underlying suspicion between communities would gradually dissolve. Neither assumption held. As the Muslim share of the population grew relative to the fixed ratios, and as regional crises—Nasserism, the Palestinian presence, Cold War alignments—pulled Lebanese factions toward external patrons, the Pact’s silence on adjustment became a source of instability rather than a guarantee of peace. By 1975, the accumulated pressure produced not reform but civil war.

II. From Coexistence to Power-Sharing

The fifteen-year civil war that followed did not end because any faction achieved decisive victory. It ended because every major faction concluded, at devastating cost, that none of them could win outright and all of them were losing collectively. This is precisely the Afghan logic revisited in a different register: shared vulnerability eventually outweighs the appeal of continued competition. But Lebanon’s postwar settlement went further than Afghan or Syrian arrangements in one crucial respect—it did not merely coordinate the war’s factions. It rewrote the rules of the state itself.

The 1989 Taif Agreement, brokered in Saudi Arabia with Syrian, American, and Arab League involvement, replaced the six-to-five Christian-Muslim ratio with parity, transferred substantial executive authority from the Maronite presidency to a religiously mixed Council of Ministers, and formally called for the eventual “abolition of political sectarianism”—a call that has never been implemented. Taif also entrenched Syria as the postwar guarantor of the settlement, granting Damascus what amounted to tutelary authority over Lebanese politics for the following fifteen years.

Taif deserves to be read on its own terms rather than as either triumph or failure. It stopped a war that had consumed more than a hundred thousand lives. It rebalanced formal representation to reflect demographic reality more honestly than the 1943 Pact had. But it also did something more ambiguous: it took a wartime “no victor, no vanquished” logic and converted it into a permanent operating principle of the state. Where the National Pact had at least imagined a temporary, expedient arrangement, Taif effectively institutionalized indefinite parity as the price of continued coexistence. Scholars examining the accord decades later have generally converged on this reading: Taif reorganized sectarian power-sharing more than it transformed it, reaffirming the very framework whose rigidities had helped produce the war it ended.

This is the moment at which the series’ argument must sharpen. Afghanistan’s cooperative arrangements were provisional and contested, built to survive a specific threat. Syria’s were bounded and transactional, designed to allow autonomous actors to coordinate without dissolving into one another. Lebanon’s arrangement was neither provisional nor bounded. It became the constitutional grammar of an entire state, embedded so deeply in the allocation of the presidency, premiership, speakership, cabinet seats, senior civil service posts, and parliamentary constituencies that no single election, however decisive, can produce a governing majority in the ordinary sense. Every government must be a coalition, because the architecture makes any other outcome constitutionally unavailable.

III. The Coalition Paradox

It is here that Lebanon’s deepest lesson emerges, and it complicates any easy inference from Afghanistan and Syria that more cooperation is always the appropriate remedy for fragmentation. A coalition built to prevent one community from dominating the others can, over time, evolve into a coalition in which everyone possesses enough leverage to prevent everyone else from governing—while no one possesses enough authority to govern at all.

Consider what a fully entrenched consociational system actually distributes. It does not merely distribute offices. It distributes veto points. Confessional quotas, communal consent requirements, and informal norms of unanimity among sectarian leaders mean that most consequential decisions require something closer to consensus among elites than a parliamentary majority. This is, in one light, a minority protection mechanism precisely of the kind political scientists have long recommended for divided societies. In another light, it is a standing invitation to obstruction, because any single confessional bloc—or any leader claiming to speak for one—can credibly threaten paralysis rather than accept an outcome it dislikes.

This produces a genuine collective-action problem, distinct from simple disagreement. Lebanese political actors are not merely divided over policy. They are individually rational to prioritize the preservation of their own bargaining position—their veto, their share of ministries, their control of a given administrative fiefdom, their relationship with an external patron—even when they recognize, often explicitly, that the aggregate result is a weaker, less capable state. No single actor benefits from unilaterally disarming this leverage, because doing so would not produce a stronger collective institution; it would simply hand that leverage to rivals who retain theirs. The rational strategy for each actor, taken alone, generates an irrational outcome for the system as a whole. Cabinet formation after Lebanese elections has repeatedly taken months, occasionally more than a year, precisely because government formation requires satisfying an intricate lattice of communal and factional claims rather than assembling a simple majority.

The presidency itself has become the starkest illustration. Because the office is constitutionally reserved for a Maronite and depends on a two-thirds parliamentary consensus that is difficult to assemble amid factional stalemate, Lebanon has repeatedly experienced prolonged presidential vacuums—most notably a gap of roughly two and a half years between 2014 and 2016, and another extended vacancy following the end of Michel Aoun’s term in 2022 that persisted until Joseph Aoun’s election in January 2025. A state can apparently function, after a fashion, without a head of state for years at a stretch. That fact alone tells you something essential about the difference between a system that prevents domination and a system capable of producing decisions.

IV. Asymmetry and the Limits of the Bargain

No discussion of Lebanese coalition politics can responsibly avoid Hezbollah, and none should attempt to resolve the question of whether the organization is legitimate or illegitimate, a resistance movement or a destabilizing force. The analytically important issue is narrower and more structural: what happens to a coalition when one of its members possesses a capability—organized, sustained, externally supplied military force—that the others fundamentally lack.

Hezbollah emerged during the civil war and the subsequent Israeli occupation of southern Lebanon as an armed Shia movement backed by Iran, and it retained its weapons after Taif’s general call for militia disarmament, justified initially by the continuing Israeli occupation that lasted until 2000 and later by the 2006 war with Israel. Over the following two decades it became simultaneously a parliamentary bloc, a service-delivery network, a media apparatus, and the most capable armed force in the country—outside, and at times more capable than, the Lebanese national army itself. This dual character means Hezbollah has participated in coalition bargaining not merely as one confessional bloc among several but as an actor whose consent has, at various moments, functioned as something closer to a precondition for stability than a negotiable input. The 2008 crisis, triggered when the government moved against Hezbollah’s private telecommunications network and was met with the group’s brief seizure of West Beirut, and its resolution through the Qatari-mediated Doha Agreement, illustrated the pattern with unusual clarity: negotiated compromise restored calm, but it did so by confirming, rather than dissolving, an effective veto for the armed actor within the coalition.

The 2023–2024 war with Israel and its aftermath have altered this balance more than any development since 2006, and the change is still unresolved as this is written. Hezbollah’s military capacity was severely degraded, and the November 2024 ceasefire, followed by the formation of Nawaf Salam’s government in February 2025 and the election of army commander Joseph Aoun to the presidency, produced the first Lebanese cabinet in years to commit explicitly to the state’s exclusive monopoly on arms—a formulation that pointedly abandoned Hezbollah’s longstanding “army, people, resistance” framing. Yet by late 2025, when the cabinet approved an army plan to bring weapons fully under state control, every Shia minister walked out in protest, and Hezbollah’s leadership rejected the imposed disarmament timeline outright, citing continued Israeli strikes and a lingering Israeli military presence at several points inside Lebanese territory. Whether this moment produces a genuine renegotiation of the coalition’s underlying asymmetry, or merely a new equilibrium in which reduced capability still translates into disproportionate leverage, remains an open and contested question, not a settled one.

The theoretical point, independent of how this particular episode resolves, connects directly back to Syria and Afghanistan. A coalition among actors of roughly comparable capacity can plausibly develop reciprocal restraint, because each actor can impose real costs on the others and each has correspondingly real incentives to honor bargains. A coalition containing one radically more capable member operates under a different logic. Formal parity in cabinet seats coexists with informal disparity in coercive power, and the two do not cancel out. This is a structural condition worth naming precisely, without moralizing about which side is at fault: coalitions are not simply about who sits at the table, but about what each participant can do if the table is overturned.

External patronage compounds this asymmetry without being reducible to it. Syria’s military and intelligence tutelage over Lebanon from 1990 until the 2005 withdrawal—triggered by the assassination of former Prime Minister Rafik Hariri and the mass protests that followed—demonstrated how an external guarantor can simultaneously stabilize a fragile settlement and prevent that settlement from becoming genuinely domestic. Iranian support for Hezbollah, Gulf and Western support for rival blocs, and periodic American and Saudi financial leverage over Lebanese institutions have each, at different moments, done the same double duty: making local compromises possible while making Lebanese sovereignty partially contingent on regional bargains the Lebanese themselves do not fully control. External involvement stabilizes a coalition when it constrains escalation and enforces agreements neither side alone can guarantee; it prevents the coalition from becoming national when domestic factions come to depend more on their foreign patrons’ calculations than on their compatriots’ consent.

V. What Lebanon Adds to Afghanistan and Syria

Read together, the three cases trace a progression rather than three isolated illustrations of the same idea. Afghanistan established the foundational proposition: actors who cannot destroy one another without destroying themselves may eventually recognize that separate competition produces collective vulnerability, and that recognition alone can generate a fragile willingness to cooperate. Syria showed what such cooperation requires operationally once distrust is assumed to be permanent rather than temporary—bounded objectives, defined spheres of autonomy, channels of communication that do not require goodwill, and mechanisms that prevent any single participant from converting cooperation into domination.

Lebanon shows what happens when this logic is not merely practiced but constitutionalized written into the structure of the state itself, sustained across generations, and treated as the permanent operating system of government rather than a transitional arrangement. And it shows, with unusual clarity, that this move solves one problem while creating another. Institutionalized power-sharing can indeed prevent the renewed civil war that a return to majoritarian competition might invite; Lebanon has not relapsed into full-scale civil conflict since 1990, a nontrivial achievement given the surrounding region’s trajectory. But the same veto-laden architecture that prevents domination also diffuses accountability so thoroughly that responsibility for governance becomes nearly impossible to locate. When the state fails—as it failed catastrophically in the 2019–2020 financial collapse, one of the more severe economic crises recorded anywhere in the world in the past century and a half according to World Bank assessments, and again in the August 2020 Beirut port explosion, whose investigation has been repeatedly obstructed by political interference—no single faction can be cleanly blamed, because no single faction was ever fully in charge. Fragmented sovereignty produces fragmented culpability.

This yields the proposition toward which the series has been building: coalitions can solve the problem of mutual destruction without solving the problem of collective action. Preventing any one faction from ruling the others is a genuine achievement in a society that has experienced civil war. It is not the same achievement as producing a state capable of maintaining its currency, supplying electricity, prosecuting corruption, or investigating its own catastrophes. Lebanon has, for eight decades with one calamitous interruption, largely accomplished the first. It has struggled, especially in the past decade, to accomplish the second at all.

VI. The Ethiopian Question

The relevance of this progression to Ethiopia should be approached carefully, because the temptation to extract a template rather than a set of questions is exactly the error the Lebanese case warns against. Lebanon’s specific confessional architecture—fixed offices reserved by religious community, parliamentary seats apportioned by sect—is a product of a particular colonial and demographic history and should not be treated as an exportable design. What might travel is not the institutional form but a small number of underlying propositions.

The first is that agreement need not precede coalition. Lebanese Christian and Muslim elites in 1943 did not reconcile; they did not share a common national vision; many barely trusted one another. What they shared was a narrower, more modest proposition—that none of them could safely eliminate the others, and that a state built on this recognition was preferable to a contest that might destroy everyone’s position simultaneously. For deeply divided Ethiopian political actors, this suggests that the sequencing question—must reconciliation come before cooperation—may be posed in an unnecessarily demanding way. A narrower foundation, resting on mutual security rather than mutual affection, may be sufficient to begin.

The second is that the design of veto points matters enormously, and that more protection is not costlessly better. Minority guarantees that prevent domination are valuable; minority guarantees so extensive that they prevent the state from making any decision at all are a different thing entirely, and the line between the two is not obvious in advance. Ethiopia’s own federal structure already distributes power along ethnic and regional lines in ways that raise a version of this same question. The Lebanese experience suggests that whatever arrangement is chosen should be built with an explicit mechanism for revisiting and adjusting the balance over time—something Lebanon’s 1943 ratio conspicuously lacked, with consequences that took three decades to fully manifest.

The third is that formal institutions rarely tell the whole story in a fragmented polity, and Ethiopian actors and outside observers alike should expect the real political system to consist of formal rules plus informal elite bargains plus credible external guarantees plus the ever-present calculation of relative coercive capacity. Pretending that a constitutional text alone will govern behavior, when armed or economically dominant actors exist outside or alongside it, invites the same asymmetry problem Lebanon has never fully resolved with Hezbollah.

The fourth is the sharpest and least comfortable lesson: coalition arrangements should be designed, from the outset, with some mechanism for evolving from negative cooperation—coexistence that merely prevents catastrophe—toward positive cooperation capable of producing functioning institutions and public goods. Lebanon’s tragedy is not that its elites once made a pact to avoid destroying one another. It is that the pact was never systematically upgraded into a state that could also govern well, and that by the time this deficiency became undeniable, the coalition’s own veto structure made upgrading it extraordinarily difficult. An Ethiopian settlement that solves only the problem of mutual destruction, without building in the capacity to later solve the problem of collective governance, risks inheriting Lebanon’s second, quieter crisis alongside its first, more visible success.

Conclusion

The ultimate test of coalition is not whether enemies can sit at the same table. Afghanistan showed that even that is not guaranteed, and that its achievement can be significant. Syria showed that sitting at the table durably requires real mechanisms, not just goodwill. Lebanon shows the test beyond both: whether the rules built to let enemies sit together without needing to destroy one another still leave the table capable of deciding anything once they are seated. A coalition that only prevents domination, without ever building the capacity for collective action, has not failed—but it has not fully succeeded either. It has produced, at great and lasting cost, a state that survives without quite governing, and a peace that endures without quite becoming secure. That is the distinction Lebanon, more than any other case in this series, forces its observers to hold in mind—and it is the distinction that any society now attempting to convert coexistence into a lasting political settlement cannot afford to lose sight of.

Editor’s Note: Views in the article do not necessarily reflect the views of borkena.com  

__

Join our Telegram Channel: t.me/borkena

Like borkena on Facebook

To submit a press release, send the submission to info@borkena.com

Add your business to Ethiopian Business Listing / Ethiopian Business Directory  

Join the conversation. Follow us on X (formerly Twitter)  @zborkena to get the latest Ethiopian news updates regularly.