In 1989, the Taif Agreement ended Lebanon’s fifteen-year civil war. Tucked into the constitutional preamble was a provision designating the abolition of political confessionalism \u2014 the system of distributing every significant government position by religious sect \u2014 as “a basic national goal.” Article 95 of the amended constitution mandated the formation of a National Committee, chaired by the president, including the speaker of parliament and the prime minister, tasked with developing and executing a phased plan to dismantle the sectarian allocation of power.<\/p>\n\n\n\n
In the thirty-six years since, that committee has never met. Not once. No president has convened it. No parliament has demanded it. No prime minister has raised its absence as a constitutional violation. The provision hasn’t been repealed, amended, or formally challenged in any court. It sits in the text of the Lebanese constitution in continuous, undisturbed non-compliance \u2014 a commitment the system makes to its own transformation while containing the mechanism to ensure the transformation never happens.<\/p>\n\n\n\n
The question most people ask about Lebanon \u2014 why did it fail? \u2014 arrives pre-loaded with its answer: too many sects, too many foreign armies, too much history packed into too small a country. Article 95 poses the better question, and the committee’s absence answers it. Why did a political system that formally committed to its own eventual abolition spend three and a half decades making certain the abolition never came?<\/p>\n\n\n\n
The answer starts with a blueprint.<\/p>\n\n\n\n
In the summer of 1943, two men negotiated the terms of a new country. President Bechara el-Khoury, a Maronite Christian, and Prime Minister Riad el-Solh, a Sunni Muslim, reached what became known as the National Pact \u2014 an unwritten agreement distributing political power along confessional lines. The presidency would go to a Maronite. The prime ministership to a Sunni. The speakership of parliament to a Shia Muslim. Parliamentary seats were allocated on a 6:5 Christian-to-Muslim ratio. And the allocation ran deep beneath the headline offices: ministries, senior civil service posts, military commands, judicial appointments \u2014 all distributed by sect, all the way down.<\/p>\n\n\n\n
The bargain was mutual and explicit. Muslims would accept Lebanon’s borders and abandon the push for union with Syria or a wider Arab state. Christians would stop relying on France for protection. Both sides agreed, in effect, to be Lebanese \u2014 on the condition that being Lebanese meant never having to share power on any basis other than communal identity.<\/p>\n\n\n\n
The entire structure was proportioned to a census conducted in 1932 under the French Mandate. It was the only official national census Lebanon has ever held. It counted Christians at approximately 51 percent of the population, though the political scientist Rania Maktabi documented in a 1999 analysis that the methodology included emigrants living in the Americas, West Africa, and elsewhere \u2014 a diaspora population that was roughly 85 percent Christian, which inflated the Christian share of the count relative to actual resident proportions.<\/p>\n\n\n\n
The Taif Agreement of 1989 restructured the formula. Parliament moved to a 50:50 Christian-Muslim split across 128 seats. The Maronite president’s executive authority was substantially reduced in favor of the Council of Ministers. The Sunni prime minister gained genuine executive power. The Shia speaker’s term was extended from one to four years, strengthening the position considerably. But Taif retained the confessional foundation. More than retained it \u2014 it codified in constitutional text, for the first time, what the 1943 Pact had left unwritten. It formalized the informal. And the provision meant to eventually abolish the whole arrangement \u2014 Article 95 \u2014 was placed in the hands of the very political actors whose positions depended on its continuation.<\/p>\n\n\n\n
Confessionalism wasn’t a purely Lebanese invention. The principle of managing divided societies through proportional power-sharing had European precedents, and Arend Lijphart later formalized it as consociational democracy in Democracy in Plural Societies<\/em> (1977), studying Lebanon among his primary cases. The idea was elegant in the seminar room: give every group a guaranteed share of power, remove the incentive for violence, manage division without resolving it. Lebanon was the oldest and most complete running instance of the template. It was subsequently exported, with modifications, to other post-conflict settings as a standard institutional mechanism.<\/p>\n\n\n\n Managing division. Not resolving it. That distinction matters.<\/p>\n\n\n\n The power formula runs on the 1932 census. Ninety-four years old. Conducted by a colonial administration for a country that didn’t yet exist in its current political form. Its data describes no one alive today and no community in its current proportions.<\/p>\n\n\n\n Current estimates \u2014 which carry no official basis, because no official survey has ever replaced the 1932 count \u2014 suggest that Shia Muslims are now the largest single community in Lebanon, with Sunni Muslims and Christians roughly comparable in size, and Maronites a minority within the Christian population. But these are estimates, not data. They can’t be data, because producing the data would require a census. And a census is precisely what the system cannot allow.<\/p>\n\n\n\n As The Nation reported in a piece titled, with characteristic directness, “In Lebanon, a Census Is Too Dangerous,” every sectarian leader in the country understands what a count would reveal: that some communities have grown and others have shrunk, and that updating the formula would mean transferring political power from the shrinking to the growing. Religious leaders across the confessional spectrum \u2014 groups that agree on virtually nothing else \u2014 have maintained a quiet consensus for nearly a century on this one point: don’t count.<\/p>\n\n\n\n The logic is perfectly circular and perfectly sealed. Reform requires consensus among the political actors. Consensus requires that no group loses. A census creates losers. Losers hold vetoes. Therefore: no census. Without a census, there’s no legitimate demographic basis for arguing that the allocation should change. Without a changed allocation, the 1943 arrangement \u2014 adjusted at the margins by Taif but structurally intact \u2014 continues to determine who governs, who profits, and who provides.<\/p>\n\n\n\n Article 95’s National Committee sits where it has sat since 1989. Required by the constitution. Never convened. Waiting for a quorum that will never assemble, because assembling it would require the very agreement the system’s architecture was built to prevent.<\/p>\n\n\n\n So how does a constitutional arrangement produce a country where the lights don’t work?<\/p>\n\n\n\n Through patronage. Every significant ministry in Lebanon is allocated by sect, which means every minister answers not to the ministry’s stated function but to the sectarian patron who secured the appointment. State services \u2014 employment, contracts, permits, utility connections, benefits \u2014 are mediated through wasta, the informal currency of communal access. You don’t get a government job because you’re qualified. You get it because your za’im \u2014 your communal leader \u2014 arranged it. And the za’im arranged it because your gratitude is his political capital.<\/p>\n\n\n\n \u00c9lectricit\u00e9 du Liban is the clearest expression of what this looks like as infrastructure. Chronically overstaffed with political appointees shielded by ministries that are themselves different sects’ patronage vehicles, EDL has carried system losses \u2014 technical faults, billing failures, and outright theft \u2014 that by various estimates consumed roughly 40 percent of the electricity it generated. In August 2024, EDL ran out of fuel entirely and the country went dark. A nationwide blackout. By early 2025, state electricity had clawed back to roughly six to eight hours per day \u2014 up from one to three hours during 2021 through 2023. The March 2026 Israeli military operations damaged even that partial recovery.<\/p>\n\n\n\n EDL isn’t a management failure. It’s a patronage system expressed as an electrical grid. No minister has an incentive to rationalize it, because rationalization means firing political appointees from multiple communities simultaneously \u2014 political cost spread across multiple sectarian constituencies at once. The rational move, for every actor in the system, is to let the grid rot.<\/p>\n\n\n\n Tax collection follows the same logic. Lebanon’s tax-to-GDP ratio collapsed from roughly 15 percent before the crisis to approximately 6 percent by 2021. Gross public debt reached 183 percent of GDP \u2014 the fourth-highest ratio in the world, per the World Bank’s Fall 2021 Lebanon Economic Monitor. The Lebanese pound lost over 98 percent of its pre-crisis value. These aren’t the figures of a state that tried and failed. They’re the figures of a state whose revenue apparatus was staffed through patronage networks that had no interest in effective collection \u2014 because effective collection would fund a state capable of providing services independently, dissolving the dependency relationships their political power runs on.<\/p>\n\n\n\n And the loop closes. Confessionalism produces patronage. Patronage produces service failure. Service failure produces dependency on sectarian providers \u2014 Hezbollah’s hospitals, schools, and welfare networks in Shia areas; Christian parties’ parallel services; the Future Movement’s Sunni social programs. Dependency reinforces confessional electoral loyalty. Loyalty reproduces the patronage system.<\/p>\n\n\n\n The loop is closed. The lights stay off.<\/p>\n\n\n\n A reader following the patronage mechanism might still hold one refuge: if the right reformers won, if political will existed, if the international community pushed hard enough, the system could be made to work.<\/p>\n\n\n\n It can’t. Every actor inside the Lebanese system has rational incentives to prevent the accumulation of state capacity, because centralized capacity would shift the power balance between sectarian patrons. State weakness isn’t a failure to achieve an objective. It’s the objective.<\/p>\n\n\n\n A genuinely capable Lebanese state \u2014 one that taxes effectively, delivers services independent of sect, commands a military monopoly on violence, enforces law through an independent judiciary \u2014 would redistribute power between sectarian communities. If the state delivered electricity reliably, Hezbollah’s parallel power infrastructure loses its value as a patronage asset. If the civil service were staffed on merit, sectarian patrons lose their primary vehicle for distributing community benefits. If the judiciary were independent, the patronage apparatus becomes prosecutable.<\/p>\n\n\n\n Every reform that produces genuine state capacity threatens the power base of every actor controlling a piece of the state.<\/p>\n\n\n\n The clearest test: the IMF. After Lebanon defaulted on its sovereign debt in March 2020 \u2014 the first default in the country’s history \u2014 the Fund conditioned any rescue program on specific structural reforms: banking sector restructuring, loss allocation legislation, capital controls, an independent audit of the Banque du Liban. A staff-level agreement was reached in April 2022. As of 2026, no comprehensive reform package has been enacted. Not because Lebanese officials didn’t understand the conditions \u2014 the conditionality was specific, public, and repeated at every IMF mission. Because each reform distributed losses in ways that would devastate a specific sectarian patron’s financial or political base. And every actor capable of blocking that particular reform had both the incentive and the institutional veto to do it.<\/p>\n\n\n\n Hezbollah’s parallel state isn’t, in this framework, a unique pathology. It’s the extreme expression of a logic every sectarian actor follows. Hezbollah built what the state wouldn’t provide, to build the dependency relationships its political power runs on. Every other major sectarian actor has done the same, at smaller scale. The difference between Hezbollah and its rivals isn’t the logic of what they do. It’s the scale \u2014 and the military capacity that protects it.<\/p>\n\n\n\n The Nawaf Salam government, formed on February 8, 2025, with explicit reform ambitions, operates inside this same architecture. The architecture has absorbed every reform attempt of the past eight decades without yielding.<\/p>\n\n\n\n If the dysfunction is internally self-sustaining, why does it matter what Iran or Saudi Arabia or France do?<\/p>\n\n\n\n Because external patrons don’t merely tolerate Lebanese state weakness. They invest in it \u2014 through three distinct mechanisms, each calibrated to a different patron’s strategic interests, but all requiring the same precondition: a state too weak to govern itself.<\/p>\n\n\n\n Iran’s mechanism is capacity substitution. In the vacuum left by state absence, Iran built \u2014 through Hezbollah \u2014 a parallel military and social service infrastructure that made the vacuum politically productive rather than merely harmful. Before the 2024 war, Hezbollah’s military capacity exceeded the Lebanese Armed Forces by every meaningful measure: an estimated 40,000 to 50,000 active fighters and an arsenal of 120,000 to 200,000 rockets and missiles, depending on the source. The 2024 war degraded Hezbollah substantially. Hassan Nasrallah was killed on September 27, 2024. The arsenal was reduced to an estimated 25,000 rockets and missiles. The organization suffered prolonged leadership disruption and significant command losses at every level. But by early 2026, reporting indicated Hezbollah’s rehabilitation pace was outstripping Israeli disruption capacity. A functioning Lebanese state with genuine sovereignty would close the vacuum Hezbollah fills. Iran’s thirty-year investment was a bet on the permanence of that vacuum.<\/p>\n\n\n\n Saudi Arabia’s mechanism was financial patronage dependency. Rafik Hariri built his fortune through Saudi Oger, a construction company he founded in Saudi Arabia that grew into one of the largest in the Middle East through extensive royal patronage contracts. He received Saudi citizenship in 1978. He became Lebanon’s most powerful Sunni politician, serving as prime minister from 1992 to 1998 and 2000 to 2004, channeling Gulf investment into postwar reconstruction while leaving the confessional architecture untouched. His assassination on February 14, 2005, disrupted the patronage chain but didn’t sever it: his son Saad inherited both the political franchise and the Saudi relationship, serving as prime minister in 2009\u20132011 and 2016\u20132020. The nature of the dependency was exposed with brutal clarity in November 2017, when Crown Prince Mohammed bin Salman detained Saad Hariri during a visit to Riyadh and forced him to read a pre-written resignation speech on Saudi television. The client’s political existence depended entirely on the patron \u2014 not on any Lebanese state institution. Hariri was released only after French diplomatic intervention. Saudi Oger subsequently went bankrupt. Saad Hariri suspended political activity in January 2022. The Saudi-Sunni patronage channel is largely defunct. But its logic stands: the patron’s influence depended on state weakness, not despite it.<\/p>\n\n\n\n France’s mechanism is perpetual crisis stewardship. Protector of Levantine Christians since the sixteenth century, administrator of Greater Lebanon under the Mandate from 1920, architect of the constitutional framework from which confessionalism grew \u2014 France co-sponsored the November 2024 ceasefire, deployed Macron to Beirut after the 2020 explosion, backed the 2025 political transition. Each intervention addresses a symptom. None touches the architecture. A Lebanon capable of governing itself wouldn’t generate the recurring crises that make French engagement necessary \u2014 or welcome.<\/p>\n\n\n\n The three mechanisms are distinct: capacity substitution, financial dependency, crisis stewardship. But they converge. A Lebanon that taxes at scale, delivers services independent of sect, and commands a military monopoly on violence would close the vacuum Iran needs, end the dependency Saudi Arabia exploited, and resolve the crises France manages. External interference isn’t the cause of institutional weakness. It’s the preference \u2014 each patron’s separate preference \u2014 for a specific kind of weakness that serves their interests.<\/p>\n\n\n\n Three counterarguments each name something real. None survives as an alternative explanation once you trace it through the confessional architecture.<\/p>\n\n\n\n The Palestinian presence. Jordan and Syria also hosted large Palestinian refugee populations in the same period. Jordan faced Black September in 1970; the Hashemite monarchy survived because the state could mount a unified security response. Syria absorbed refugees without civil war because the Assad regime exercised centralized authority. Lebanon couldn’t formulate a unified response because the confessional system made unified state action structurally impossible. Christian militias opposed the Palestinian armed presence as a threat to Maronite demographic standing. Sunni leaders accommodated it on pan-Arab solidarity grounds. Shia communities organized against it through the Amal movement as Palestinian factions competed for the same southern terrain. The Palestinian question didn’t cause the civil war. It was the pressure the confessional architecture couldn’t contain.<\/p>\n\n\n\n The civil war itself, 1975 to 1990. Sometimes treated as the event that broke the country. But the war was triggered by confessional dynamics already fully operational. And Taif, which ended it, reproduced rather than resolved those dynamics \u2014 entrenching wartime militia leaders as postwar political actors and codifying their sectarian constituencies in the settlement. The war deepened the dysfunction. It didn’t create the condition that made the war possible.<\/p>\n\n\n\n The Syrian occupation, 1990 to 2005. Hafez al-Assad managed Lebanese politics for fifteen years not by replacing the confessional system but by working through it, treating sectarian patronage networks as leverage points rather than obstacles. When Syria withdrew after the Cedar Revolution following Rafik Hariri’s assassination, the confessional architecture was entirely intact. Syria hadn’t imposed the dysfunction. It had rented it, and left the infrastructure undamaged when it pulled out.<\/p>\n\n\n\n These aren’t alternative explanations. They’re events the confessional system processed. And each time it processed one \u2014 absorbing the crisis, surviving it, emerging from it \u2014 the system came out more entrenched. Because each crisis that required sectarian actors to serve as the primary providers of security or services deepened the dependency relationships that keep the system self-perpetuating.<\/p>\n\n\n\n If the confessional architecture survived a civil war, a foreign occupation, an assassination that triggered a revolution, a financial collapse, a mass-casualty explosion, and repeated Israeli military campaigns \u2014 with its logic intact every time \u2014 then the question is no longer about Lebanon. It’s about the logic itself. And the logic has been exported.<\/p>\n\n\n\n The 1995 Dayton Agreement created a governance structure for Bosnia-Herzegovina built on the same post-conflict power-sharing template: a three-member rotating presidency \u2014 one Bosniak, one Croat, one Serb \u2014 ethnically allocated positions at every level of government, two semi-autonomous entities with separate administrations and laws. The result, in Bosnia’s High Representative’s own 2008 count: five presidents, four vice presidents, thirteen prime ministers, fourteen parliaments, and seven hundred members of parliament serving a population of just under four million.<\/p>\n\n\n\n The governance formula was derived from wartime demographics, not from any principle of representation \u2014 the same design logic as the 1943 National Pact and the 1989 Taif Agreement. The reform mechanism self-seals by the same logic: in 2009, the European Court of Human Rights ruled in Sejdic and Finci v. Bosnia that the ethnic allocation of the presidency and the upper house violated the European Convention on Human Rights by excluding minorities from the highest offices. Seventeen years later, the ruling remains unimplemented \u2014 because implementation requires the consent of the actors whose positions depend on the allocation continuing. Sejdic-Finci is Bosnia’s Article 95: an authoritative finding that the system is illegitimate, sitting unexecuted because the system’s architecture prevents the execution. The system can’t generate internal order on its own; external management isn’t a supplement but a structural necessity, and it degrades when that management weakens or turns adversarial.<\/p>\n\n\n\n Bosnia does have what Lebanon never had: a European legal framework wrapping the dysfunction \u2014 ECHR jurisdiction, EU accession conditionality, NATO security presence. The difference slows the timeline. Bosnia’s dysfunction matures at a slower tempo because the scaffolding around it is heavier and better resourced. But it doesn’t change the trajectory. The self-sealing mechanism is internal to the power-sharing design and operates identically regardless of the external framework. The ECHR ruled. The ruling sits unimplemented. The mechanism preventing its execution in Sarajevo is the same mechanism preventing Article 95’s committee from convening in Beirut.<\/p>\n\n\n\n Government formation in Bosnia routinely takes months as no ethnic group accepts an outcome that reduces its share. EU accession \u2014 conditioned on reforms that would redistribute ethnic power \u2014 has been blocked for two decades. Republika Srpska under Milorad Dodik operates separatist politics structurally analogous to Hezbollah’s relationship with the Lebanese state: building parallel institutions, leveraging central-state weakness as a political base, backed by external patrons. When Saudi Arabia withdrew from active Lebanese patronage management and the Syrian occupation ended, Lebanon’s dysfunction accelerated toward collapse. The inference is direct: the mechanism is the same, and Lebanon shows what it produces at institutional maturity.<\/p>\n\n\n\n Israel’s current military operations in Lebanon \u2014 its sixth major intervention since 1978, with five divisions deployed across southern Lebanon as of spring 2026, pushing toward the Litani River, more than 3,000 killed and over a million displaced \u2014 are the latest iteration of external management of a system that has always required it. The architecture made this predictable. Not the specific actor. Not the specific year. The structural need for external intervention in a system that can’t produce internal order.<\/p>\n\n\n\n Article 95 committed Lebanon to abolishing political confessionalism thirty-six years ago. The committee tasked with planning the abolition has never met. That isn’t a political failure. It’s the most precise expression of what confessionalism is \u2014 a system that can accommodate any formal commitment to its own transformation, because it contains the mechanism to ensure the commitment is never acted on.<\/p>\n\n\n\n Israel is in Lebanon in 2026 as it was in 1978, 1982, 2006, and 2024. Managing the latest symptom of a system that every actor, external and internal, has found preferable to the alternative. The Nawaf Salam government’s reform ambitions are real. They operate inside a system whose self-sealing logic has survived every reformist government, every popular uprising, every foreign invasion, every financial collapse, and every mass-casualty event that preceded them.<\/p>\n\n\n\n The question isn’t whether Lebanon can be fixed. It’s whether anyone with the power to fix it has an incentive to do so.<\/p>\n\n\n\n Lebanon’s answer, across eighty-three years of the confessional system, is that they don’t.<\/p>\n\n\n\n Bosnia is the next data point.<\/p>\n\n\n\nThe census that runs the country<\/h3>\n\n\n\n
The patronage engine<\/h3>\n\n\n\n
The Beirut port explosion as structural evidence<\/strong>\n\nOn August 4, 2020, 2,750 tonnes of ammonium nitrate stored in Warehouse 12 of Beirut's port detonated. The blast killed at least 218 people, wounded 7,000, and destroyed large sections of the city. The ammonium nitrate had been confiscated from the cargo vessel MV Rhosus in 2014 and stored unsecured for six years. Multiple government agencies \u2014 customs, military, judicial authorities \u2014 knew it was there. None acted, because acting required coordination across institutions controlled by different sectarian patrons, none of whom had an incentive to absorb the political cost of dealing with hazardous material stored under another sect's jurisdictional authority. The criminal investigation that followed was blocked repeatedly: investigating judges recused or replaced, political figures using their institutional positions to challenge proceedings, Lebanon's chief public prosecutor eventually halting the inquiry after the lead judge named him a person of interest. As of mid-2026, no senior official has been convicted. The explosion is the most concentrated illustration of what confessional patronage produces when its failures become physical \u2014 not broken services but uninvestigated mass death.<\/code><\/pre>\n\n\n\nState weakness is the product<\/h3>\n\n\n\n
The 2019 thawra and why it couldn't win<\/strong>\n\nOn October 17, 2019, the largest cross-sectarian civic mobilization in Lebanese history erupted. The slogan was \"kullun yaani kullun\" \u2014 all of them means all of them. The protests targeted every political party regardless of sect, constructing a national grievance in a system designed to recognize only communal ones. Hundreds of thousands participated across the sectarian map. The movement sustained itself for months. It failed \u2014 not because it lacked legitimacy or popular mass, but because the confessional system contained no mechanism through which cross-sectarian pressure could translate into governance change. Every institution the thawra needed to move was held by sectarian patrons whose power depended on the sectarian loyalty the protests were attacking. Then the financial collapse deepened. COVID arrived. The port exploded. Each crisis drained the movement and redirected attention. The thawra proved that confessionalism resists not just elite-level reform but mass popular pressure \u2014 because the pressure has nowhere to land when every institutional gate is controlled by a sectarian actor.<\/code><\/pre>\n\n\n\nThe patrons who need it broken<\/h3>\n\n\n\n
The Banque du Liban and who protected it<\/strong>\n\nFrom roughly 2016 onward, the Banque du Liban under Governor Riad Salameh conducted a series of \"financial engineering\" operations: borrowing dollars from commercial banks at above-market rates to defend the pound's peg to the dollar, creating a circular system that required constant new dollar inflows and generated outsized returns for Lebanon's politically connected banking sector. When the arrangement collapsed in October 2019, ordinary depositors lost access to their savings. The pound lost over 98 percent of its value. No comprehensive depositor-protection legislation has been enacted. Why didn't any politician intervene while it was happening? Because the banking sector was the primary vehicle through which sectarian elites stored and managed their wealth. Threatening the peg meant threatening the financial base of your own patron network. The Banque du Liban didn't fail its mandate. It succeeded at protecting the right people until it ran out of other people's money.<\/code><\/pre>\n\n\n\nThe rival explanations don’t survive<\/h3>\n\n\n\n
Bosnia is running twenty years behind<\/h3>\n\n\n\n