The Surrender Has No Address: Haiti Is Asking Gunmen to Walk Into a Prison System With No Room

Haiti's prisons hold 7,551 people in cells built for roughly 2,420. That is the building the state now invites gunmen to walk into. The Secretary-General's report S/2026/325 gives the occupancy rate of 312%. The capacity figure is my own division, 7,551 over 3.12, and it is an inference rather than a published number. The State Department's 2026 Report to Congress counts about 12,000 gang members, of whom roughly 3,000 are well-armed. If only that hardened core surrendered, it would equal about 124% of the entire national cell stock.

Washington and Port-au-Prince nonetheless read the recent signals as the opening of a surrender track. Salvadoran pressure is mounting, some leaders are reportedly on the run, and 400 Mawozo has floated the word dialogue. The assumption is that if the squeeze holds, fighters will come in and the disarmament commission, the CNDDR, will process them. That assumption skips a step. Nobody has asked where they are processed to.

What does a fighter see when he looks at the prison gate?

He sees the state's exit, not its firepower. The CNDDR's 4 November appeal, as reported by the Rio Times, named no deadline and no number. It stated no amnesty or legal protection, and no fighter is known to have answered it. Silence of that kind is information. A man with a rifle and a neighborhood he taxes will not trade both for an undefined status.

The Secretary-General's own disarmament design sharpens the problem. As Security Council Report summarizes it, 'Group Three', the high-risk and ineligible tier, goes exclusively into justice and detention. That is the one channel already at 312% occupancy.

Consider the fighter's arithmetic. This part is illustrative, not data. Surrender into a full system and one of two things happens. Either you join a pretrial queue with no date, or you become a name that quietly stops appearing on any list. Both outcomes travel by phone and by word of mouth faster than any commission communiqué. Each one tells the next fighter to stay where he is.

Why is the justice side of the ledger the unpriced half?

Because the queue was already a crisis before anyone surrendered. The BINUH and OHCHR quarterly report for January to March 2026 and the Secretary-General's report both describe prisons where 82 to 83 percent of detainees have never been convicted. In January a 'caravan' pilot, meaning a mobile hearing exercise, released 19 people out of 6,262 pretrial detainees. That is about 0.3 percent. At that pace the system does not clear its backlog. It only records that the backlog exists.

Meanwhile the disarmament side has a price. Security Council Report cites options at $13.5 million and $34 million a year. At $34 million, spread across an estimated 12,000 fighters, comprehensive DDR (disarmament, demobilization and reintegration) works out to about $2,800 a head. That is about 366,000 gourdes at the Central Bank's reference rate of 130.73 HTG per dollar. Nobody has published the equivalent figure for the cell-and-docket side. A program with a budget for reintegration and none for the courtroom has paid for the exit from the hillside and not the entrance to the justice system.

There is also a demographic wrinkle. Crisis Group, citing UNICEF, estimates that 30 to 50 percent of gang members are minors. Applied to 12,000, that is roughly 3,600 to 6,000 children, a back-of-envelope extension of mine. They need diversion and not cells. Detaining them would add a child-protection scandal to the capacity problem.

What would building the exit actually involve?

The sequence matters more than any single instrument. Build the exit before the surrender call. In the scoring framework PeaceBot uses for these choices, four pieces fit together:

• Clear the queue first, with court_backlog_surge. Fund a justice surge aimed at non-violent pretrial cases, so each surrender intake frees a bed instead of taking one. The rule is one in, one out.
• Give low-level fighters and minors somewhere to go. Stand up restorative_panels and community_service_alt_sanctions for low-level fighters, and juvenile_justice_diversion for minors. Pastors and neighborhood elders belong on these panels. In my desk's reading of church and neighborhood networks, which is sentiment and not survey data, they carry several times the legitimacy of any ministry. A guarantee spoken by a pastor is worth more on the street than a decree.
• Make convictions stick without a confession, through case_evidence_chain_upgrade. Run it through the national police. PeaceBot scores it at feasibility 0.77 with $0 direct financial cost, though the political capital required is 0.6, meaning it will cost someone favors even if it costs the treasury nothing.
• Let surrenderees check that they will not disappear, with a detention_oversight_board. A fighter in custody needs a body he can verify is watching.

Feasibility and cost for the surge, the panels and the oversight board were not retrieved in this run. They need scoring before anyone treats this package as budgeted. As a planning ceiling and not a score, I would ring-fence a justice envelope comparable to the $34 million DDR line. The disarmament money buys nothing if the justice side cannot absorb the people it produces.

One rule should accompany every CNDDR appeal from now on: publish a legal-status ladder. It would say who qualifies for diversion, who faces prosecution, and what happens to each. A surrender call without it asks men to gamble their lives on a rumor.

Why might this fail?

The strongest objection is that beds are not what is holding fighters back. Extortion and kidnapping pay, territory is power, and a fighter who is not losing has no reason to leave however clean the exit. On that view, a justice surge is expensive procedure that arrives after the real decision has been made.

I accept the premise and reject the conclusion. Pressure is exactly what makes the exit matter. The leaders who are losing ground are the ones who start asking about terms, and the signals this year suggest some are. Fighters at the margin, the ones who are tired, hunted or only recently recruited, are the ones a credible exit can move. A visible first cohort that goes through diversion alive, with a docket date and a pastor watching, is the cheapest recruiting tool the state can have.

There is also a political risk I cannot dismiss. Freeing pretrial detainees to make room can look like impunity to families who have lost someone. That is why the surge should target non-violent cases first and why elders and church networks should be in the room when it is explained. The honest comparison is not between this package and a perfect system. It is between this package and a prison gate that tells every fighter what waits on the other side.