Source: Review of African Political Economy (ROAPE) | Published: 2026-10-21
Category: 정권·선거 변동 | Keywords: democracy, election, elections, parliament
Somalia’s parliamentary election is not one election. It is 275 simultaneous, separate auctions. Each of Somalia’s 275 House of the People seats belongs to a specific sub-clan. Each is contested by a separate and distinct pool of delegates — assembled only for that seat, voting only on that seat, with no overlap with any other. What looks from outside like a single national election is, on the ground, 275 parallel closed markets operating at the same time — each with its own elder-broker, its own candidate payments, its own hotel rooms full of men who have been paid, sequestered, and made to swear on the Quran before they are allowed to vote. This is not electoral dysfunction. It is the operational logic of a system designed to convert clan identity and foreign financing into a self-reproducing governing class — one that controls parliament, elects the president, and uses both to block accountability for those who commit abuses in its service.
This investigation draws on primary source interviews with more than a dozen electoral delegates, parliamentary candidates, and traditional clan elders who participated directly in Somalia’s 2016–2017 and 2021–2022 parliamentary election cycles. It is corroborated by field research conducted by Marqaati — Somalia’s leading independent accountability organisation — across five federal states, and by an academic study of Somalia’s electoral corruption published in the International Journal of Multicultural and Multireligious Understanding ( Bincof, 2023 ). The companion piece to this investigation, “The Debt That Governs,” documents what happens next: how the parliament assembled through this process feeds directly into the presidential election — and how the president who emerges repays his debt through Somalia’s ports, procurement budget, and foreign policy.
The hotel room
The invitation arrives as a courtesy. A candidate’s representative calls to say there will be a dinner — a chance to meet the person seeking the parliamentary seat, to hear their plans, to share a meal. The delegate, one of those chosen by his sub-clan elder to cast a vote for their specific seat, accepts. It seems reasonable. It seems normal. It is neither.
After the dinner — plates of rice and camel meat, sweet tea, pleasantries — the delegates are not taken home. They are taken to a hotel. Or a large private residence. Somewhere with rooms, with guards, with doors that are not locked but might as well be. They are fed again the next morning. And the morning after that. Their phones are watched. Their movements are noted. And at some point, over tea or in a quiet room, they are asked to place their right hand on the Quran and swear that their vote, when election day finally comes, belongs to the candidate who is feeding them.
“They take you to a hotel. They feed you. They watch you. And then they make you swear on the Quran. After that, you are not a delegate anymore. You are a captive.” — A former delegate, speaking on condition of anonymity
This is not an isolated incident. The same sequence emerged with striking consistency across both the 2016–2017 and 2021–2022 election cycles: the invitation, the dinner, the hotel, the oath, the vote. And it is happening not once but 275 times simultaneously — one operation for each of Somalia’s 275 parliamentary seats, across every contested race in the country, at the same time.
How the system is built
On paper, Somalia’s indirect election system is a carefully constructed compromise. In a country with no functioning voter registration infrastructure and a political landscape organised primarily along clan lines, each of the 275 seats in the House of the People belongs to a specific sub-clan. The four major clans — Hawiye, Darod, Dir, and Digil and Mirifle — receive 61 seats each, and every smaller and minority clan is folded into a single combined share of 31 — a structure known simply as the “4.5” system: four equal shares for the big clans, half a share for everyone else. Each seat belongs to a precise lineage group within those larger blocs, and competition for it is always internal — always within the sub-clan, always intimate.
In the 2016–2017 cycle, each seat had 51 delegates, giving a total delegate universe of approximately 14,025 people across all 275 seats. By the 2021–2022 cycle, that pool had grown to 101 delegates per seat — approximately 27,775 people in total. Each delegate pool is its own closed market: its own elder, its own candidates, its own corruption dynamics. Across both cycles combined, tens of thousands of individual delegates have been subjected to the bribery, confinement, and vote-capture described in this investigation.
The system includes formal safeguards: at least 30 percent of delegates must be women; delegates must clear background checks, carry QR-coded IDs, and pass a four-stage verification process before voting. A 21-member Electoral Dispute Resolution Committee can investigate complaints and rule within seven days. Every one of these safeguards, across two consecutive election cycles, has been systematically bypassed.
One bloc’s seats follow the same process as every other seat, just in a different city. Dir clan members and the registered elders who broker their delegate lists already reside in Mogadishu. So the same thing happens as with every other seat: the elders assemble the delegate list, the delegates are gathered, and the vote is held — all in Mogadishu, the same process used everywhere else in the country. The only reason it happens there instead of in Somaliland is that Somaliland — a self-declared independent state — will not allow Somalia’s federal elections to take place on its territory. The seat, the delegates, the elders, the vote: all standard. Only the location is different.
The elder: franchisor of the seat
The 135 traditional elders registered with the federal government are the formal gatekeepers of the entire process. Their signatures authorise delegate lists. Their physical presence at voting halls is required to verify identity. They are, in theory, the last line of accountability in a system that has deliberately avoided relying on formal institutions.
In practice, many are active participants in the corruption. Marqaati’s field research documented this with hard evidence: one elder was paid $150,000 to create a favourable delegate list, which was changed four times as delegates suspected of disloyalty were replaced. In a separate documented case, a presidential campaign paid $100,000 to an elder to produce a favourable delegate list — effectively pre-purchasing the parliamentary seat before a single delegate had voted.
The elder’s financial model works both ends of the transaction. Below him, aspiring delegates pay — in money, gifts, or sustained support — for a place on his list. Bloodline helps but doesn’t guarantee entry: the elder must be cultivated continuously between elections, not just at election time. Those who have invested over years get on the list; those who haven’t are passed over, regardless of standing. The delegate pool this produces is not representative of the community. It is representative of who can pay for access to it.
Above him, candidates pay the elder directly to assemble a favourable list. Sources describe delegates who received payments from parliamentary candidates passing a portion back up to the elder — making the elder’s preference for relatives as delegates not merely nepotism but financial architecture. A relative is more reliable as a collection agent. By populating the list with family members and close dependents, the elder builds a collection mechanism that pays him at both ends: upfront from candidates, and again from the delegates’ payments on election day.
The delegate list is, in economic terms, a franchise. The elder is the franchisor. The parliamentary seat is the territory. The election is the revenue event. And across Somalia’s 275 seats, 275 versions of this franchise operate simultaneously — each invisible to the others, each running the same extraction model.
The federal member state president’s reach
The elder does not operate alone. Above him sits a layer of authority that shapes, expands, or blocks his reach entirely: the federal member state’s own president — the leader of one of Somalia’s several semi-autonomous regional states, a separate figure from the national president in Mogadishu.
A federal member state president cannot command delegates or elders. He cannot build a delegate list himself. What he can do is pay the same elders who build lists for every other seat, using his office’s money and connections to push them toward his preferred candidate.
This never gives him full control. The elders still work for whoever pays the most. A rival candidate, a rival sub-clan, or an opponent inside the federal government can simply offer more and win the same elder over. And if the president’s own grip on the state is shaky, or a rival family holds real power there, he may not be able to reach certain elders at all.
He has one more option that needs no elder’s help at all: he can simply stop a rival candidate from entering the state to compete. No elder is paid. No delegate is contacted. The candidate is just kept out before the race begins.
None of this is a chain of command. It is money, relationships, and control over territory — and how much of each a president has decides how much he can actually do.
The clearest example of this is real, not hypothetical. In the 2021–2022 cycle, then-South West State President Abdiaziz Laftagareen simply kept Abdulkadir Mohamed Nur Jaamac — Hassan Sheikh Mohamud’s own candidate — out of Baidoa, so he could never compete for the seat at all. This has happened more than once in South West State, where federal-aligned candidates have used the same trick to clear their rivals out before voting even starts. It hasn’t worked the same way in Galmudug or Hirshabelle — those states have competing power centres and no single figure controls them completely, so no one there can shut a rival out the same way. The tool is real. It just depends on one thing: whether a single person actually controls the ground the vote will happen on.
The delegate: commodity in the market
To understand what happens inside the hotel room, it is first necessary to understand how the delegate got onto the list — and what he expects to happen when he gets there. The answer, in most cases, is that he paid to be there. And he expects to be paid again.
The primary transaction is always money. A candidate does not pay every delegate in his pool. He targets only the number needed to secure a majority — in the 2016–2017 cycle, that meant 26 votes plus a buffer; in the 2021–2022 cycle, 51 votes plus a buffer as insurance against defections. On standard seats, the baseline payment to each targeted delegate is $500. On hotly contested seats — involving well-resourced candidates or sitting MPs defending their positions — that figure can escalate as high as $30,000 per delegate, documented by Marqaati’s field research and corroborated separately by Somalia’s former Auditor General, who told VOA Somali that delegate votes went for anywhere from $5,000 up to $30,000 depending on how contested the race was. Every delegate in the hall is likely receiving money from multiple campaigns simultaneously.
The phrase that governs this system is understood by every participant. In Somali: lacagi waa qaadasho, codna ciddaad doonto sii — the money is to be taken, and the vote is given to whoever you wish. The shorter, blunter version: ka cun oo ka codee — take from it and vote against him. These are not street expressions; they are recognised terms in Somali political culture, and the fact that this corruption has its own formal vocabulary is itself a measure of how deeply it has settled into Somali elections.
Payment alone does not guarantee loyalty, which is why the hotel and the