The Tollbooth Is Gone, the Tax Remains

A truck leaves Port-au-Prince on an ordinary morning. The driver knows he will pay; he does not yet know where. The collection point will be given to him on the way, and it is rarely the same from one week to the next. No barrier across the road, no checkpoint, no visible armed presence. The tax exists; the tollbooth is nowhere to be found. This scene is not the account of a single convoy: it reconstructs a mechanism described separately, and in converging terms, by several operators in transport, trade, and logistics. Other accounts go further: companies send in their vehicles’ license plates and colors in advance, and their convoys cross the controlled zones without stopping anywhere, because the right of passage has been settled elsewhere, higher up the chain, on an agreed schedule. We must immediately guard against the mirror-image misreading: the visible checkpoints have not disappeared. Very real barriers still hold major routes, and they are still collecting. What these accounts describe is not a substitution but a superposition: the levy apparatus now has two tiers, only one of which can be seen.

Yet part of the response to the Haitian crisis rests on a very different reading of this economy: the coalition as an aggregate of predators without a model, whose taxation would boil down to its roadblocks, and which could be starved of air by destroying those visible installations. This reading has a direct operational consequence: it makes the toll checkpoint the central target of economic reconquest. It also has a virtue rare in this field: it is falsifiable. If the coalition were nothing more than that, the levy should stop wherever the barrier is absent. Yet the opposite is what we observe: on entire circuits, no checkpoint is visible and the levy persists. A levy that does not need its tollbooth in order to be collected is no longer opportunistic tolling; it is criminal governance, with a tax base, rates, ledgers, and collection.

The visible tax system

The visible layer of this economy is extensively documented. International Crisis Group traces its sequence: from mid-2021, gangs seized a stretch of the national highway linking the capital to the south of the country, then extended their grip to the other routes out of Port-au-Prince, until they controlled all the roads connecting the capital to the rest of the country. At the same time, extortion widened to larger businesses located in the coalition’s strongholds: transport companies, dealerships, power plants, factories, firms operating in the industrial parks and ports. The same report notes a detail that speaks volumes about the realm this economy is moving into: in at least four of the capital’s major cemeteries, gangs now charge for burials, taking the place of the fee the state used to levy for this service. The United Nations Panel of Experts itself, moreover, uses fiscal vocabulary: the armed groups, it writes, impose informal taxation that funds their activities. The scale is far from anecdotal: studies cited by the Panel of Experts put the revenue from extortion of container transport alone at between $60 million and $75 million a year, and estimate that the total, once public transport, gas stations, ransoms, and contributions from economic or political actors are added, is several times higher.

This taxation is also spreading from flows to assets. In Carrefour, a commune of half a million people with no police and no legitimate executive, the press documents an established order: lieutenants assigned to each neighborhood who collect tolls and extortion payments, fines handed down by a parallel justice system, general strikes decreed and observed, even a claim to install the municipal executive; the Panel of Experts also places there some of the coalition’s leaders, who fell back to the area under drone pressure. Field accounts from the same commune, corroborated by reports circulating publicly, now describe the next stage: a tax demanded on homes themselves, whether or not the dwelling houses a business. After movement, services, and transfers, residence: all this parallel tax system lacked was its property tax. This time, the tollbooth is the taxpayer’s address.

This vocabulary should be taken seriously, because it describes a structure. The toll checkpoint is to the criminal economy what the customs post is to the state: a point of contact between taxpayer and tax collector. It has the properties of a customs post: a broad base, a steady yield, assigned staff. And it has the same weakness: an address.

When the tollbooth becomes a target

Two forces are pushing this model to split in two, and both are rational. The first is military. In an environment saturated with drones, and after the terrorist designations of 2025, visibility has become a costly good; what holds for men holds for installations: anything fixed and identifiable is a target. The Panel of Experts itself notes that drone strikes have shaken the gangs and changed some of their tactics. The rational response of a threatened tax collector is not to give up the tax; it is to diversify the ways of collecting it. The second force is economic. The physical roadblock is a crude instrument: it creates congestion, exposes the collectors, and multiplies friction and opportunities for wildcat levies by lower-ranking members. A subscription system, settled upstream and collected wholesale rather than vehicle by vehicle, costs less, yields more, and cannot be attacked with a rocket launcher. If the barriers persist nonetheless, it is because they serve functions other than collection: they mark sovereignty over a territory, filter who comes in, and provide a living for the local ranks. The visible checkpoint remains; it is no longer the whole system.

This adaptation stopped being a hypothesis on 27 March 2026, when the Panel of Experts documented it in its interim report (S/2026/241). The gangs, the report says in substance, are adopting new methods of collection to evade police operations targeting illegal roadblocks: tighter control over money transfer offices, operations carried out in police uniforms and with police equipment, and “concealed” tolls, where the collector no longer blocks the road but watches from nearby buildings, notes who passes, and then catches up with his targets to demand what he is owed. The annex on extortion describes the logical next step: after roadblocks around the main port areas were dismantled, the groups kept their lists of targets, gave instructions to pay off the road, and deployed lookouts to identify commercial users before coming to collect the tax directly at their premises. A logistics operator interviewed by the Panel describes a relationship between gangs and large companies that has become lasting, almost contractual, in which the protection of goods is the counterpart to the racket. The report goes as far as the system’s bookkeeping: in one locality, every money transfer withdrawal is subject to a compulsory levy, recorded in a second ledger kept on the gang’s behalf and reconciled with the official ledger, with the transfer offices required to hand over the proceeds.

Operators’ accounts refine this picture, and each of the mechanisms they describe deserves a name. The first is upstream collection: the levy moves up the value chain and is collected from traders and shippers, once and wholesale, rather than at the roadside. The second is the mobile tollbooth: rotating payment points, communicated to truckers along the way, impossible for a strike to find, impossible for the payer to avoid. The third, as several operators describe it, resembles a subscription: enrollment programs that transport companies sign up for, a rate per truck, lists of pre-registered vehicles, with plates and colors sent in advance, which turn the right of passage into a status. What is sold is no longer a crossing; it is certainty. This is precisely the definition Diego Gambetta gave of the protection industry: the product was never the barrier; the barrier was merely the sales interface, and an interface, as in any payment system, can be dematerialized. Sources also describe levies carried out inside or right next to certain logistics facilities, in ways that make them far harder to neutralize than a roadblock; details that would identify the facilities concerned have been deliberately omitted here. You do not raze a legitimate facility to get at the tollbooth lodged inside it. That leaves the most sensitive matter: several accounts mention payments by check. This column was unable to examine any of the instruments concerned and therefore takes no position on either the scale or the prevalence of the practice; it will simply note that checks are not absent from the public record, since the Panel of Experts mentions the arrest in the Dominican Republic, in August 2025, of a former local elected official with undeclared foreign currency and checks in her possession, suspected of links to gangs. A check, where it exists, is a fingerprint; but a judicial and financial investigation still has to go and lift it.

Finally, the same accounts rank the two tiers. For enrolled operators, there is practically no payment on the road anymore; the visible roadblocks are still operating, but less often, and they reportedly account for only the smallest share of the yield. If these proportions are accurate, the invisible tier is no longer a supplement to the system; it may well have become its core. Measurement is still lacking, but the evidence converges, and the Panel of Experts, by its own route, reaches the same conclusion: the reduced visibility of gang activity does not mean a diminished capacity to collect, and carriers of basic necessities, in particular, remain under control.

Two older pieces of evidence complete the picture. As early as 2024, the Panel of Experts documented the case of a business executive who was financing gang members to protect his operations and secure the transport of his goods: a permanent arrangement, with neither barrier nor tollbooth. It must be said plainly: paying for the right of passage under threat does not, in itself, amount to collaboration; many operators are first and foremost the victims of a system to which the lack of any alternative forces them to submit. The Global Initiative Against Transnational Organized Crime, for its part, observed in the spring of 2025 that gangs were reducing indiscriminate violence in certain areas precisely in order to establish economic control there through extortion: the behavior of an administrator protecting his tax base. And the Panel of Experts’ final report of September 2025 (S/2025/597) had already noted that, on the corridor linking the capital to the north, transporters either pay heavy extortion fees or bypass the route by sea, with the extra cost passed on to the consumer in both cases. The sea detour is proof by the market: the land corridor has a price, a known one, built into operators’ calculations. The invisible tax casts a perfectly visible shadow: rising costs and prices along the routes it hits.

Where the rent goes

Before looking for where the surplus goes, however, we must say what the rent pays for, since this taxation covers permanent expenses, with no need to assume any central bookkeeping. The first item is military: weapons and ammunition are not manufactured locally; they are bought, and their supply chains, mainly from the United States, have been documented report after report by the Panel of Experts; the seizure at the Belladère customs post in March 2025 (weapons, ammunition, cash, and vehicles), which the press, citing departmental authorities, links to the assault subsequently launched on Mirebalais, is a reminder that resupply is a recurring expense. The second item is the payroll: paid personnel, salaried recruits, revenue-sharing agreements between groups; a coalition of several thousand men is first and foremost an employer, and extortion is how it meets its payroll. The third is luxury spending (air-conditioned residences, music studios, vehicles), flaunted by the leaders themselves and described in public reports: this consumption signals status, attests to the system’s solvency, and recruits better than any speech. To this must be added the patronage handouts documented in certain strongholds, which buy legitimacy where the state offers nothing. Only what remains then has to be put to work.

Accounts and public sources trace three destinations for this balance. First, disposal: the proceeds of raids on provincial towns, stolen livestock included, become merchandise again within days on physical markets well known to local actors, and melt into ordinary trade. Next, embedding, which calls for weighing every word. Converging observations describe businesses whose apparent activity seems persistently insufficient to cover their costs and which nonetheless endure: gas stations with modest volumes, nearly empty shops that survive year after year, nightspots whose upscale standing bears no relation to their observable clientele. Low visible activity does not prove an absence of revenue; the persistence of these profiles is a warning sign, not proof, and only an analysis of accounts, flows, and beneficial owners could establish whether such establishments serve a conversion function. But that is precisely what a signal is for: to show where to look.

Finally, exit, which calls for the same rigor. The Panel of Experts’ final report documents a sharp asymmetry in seizures of undeclared cash between Haiti and the United States: one hundred nine interceptions totaling nearly $3.5 million between January 2022 and June 2025, only three of which involved funds entering Haiti. Nothing allows these sums to be attributed to the gangs; they establish only the existence of massive channels through which undeclared cash leaves the country, in which criminal proceeds can blend in among other flows. What is documented, on the other hand, is the search for a legal identity for what has been accumulated: requests for amnesty noted by the Panel of Experts as early as 2024; an announced transformation into a political party, in January 2025, that was never followed up in any substantive way; a desire to influence the composition of the next administration, noted by Crisis Group in late 2025; and even the bloody bringing to heel, in late 2025, of an alliance leader who kept on kidnapping in defiance of the narrative the coalition seeks to project about itself. An organization that disciplines its members to protect its political image, and that asks for amnesty, is not asking for permission to go on taking; it is asking for the right to keep.

What breaking it would mean

If this model is correct, then the toll theory leads to an operational misreading: destroying the visible barriers cuts off the apparatus’s most exposed tier, not its most productive one. One could raze every visible installation along a corridor and reduce only the fraction of the levy that still needed an address. The decisive targets lie elsewhere, and there are three of them. First, the tax base: as long as passage through controlled territory remains the only route between the capital and the rest of the country, willingness to pay is total; the effective and lasting securing of the roads, or the existence of alternatives, is the only policy that can reduce it. Next, the channels: informal fees are recorded somewhere in the books of the companies that pay them, traces exist when payments go through formal instruments, and atypical business profiles can be detected in tax data as well as in bank flows; the Panel of Experts itself recommends making the specialized judicial units fully operational, including the one devoted to financial crimes, and it is on this austere, slow ground that a fiscal war is won. Finally, the conversion interface: any discussion of amnesty that does not deal with the question of assets, their origin, and their fate is not negotiating the end of a criminal economy; it is organizing its exit through the front door.

That leaves the test, for a thesis that cannot be refuted is worthless, and it must be framed honestly: prices alone are not enough, since they also move with fuel, the exchange rate, shortages, and risk premiums. The test is therefore composite. Once the visible installations along a corridor have been durably neutralized, one would need to track together freight rates, turnaround times, volumes carried, risk premiums, payments reported by operators, and the price gap for the same products between the markets served by that route and comparable markets. If payments persist although the roadblocks have disappeared, the thesis of displaced, dematerialized collection will come out strengthened; if all these indicators ease durably, the toll theory will have been right, and this column will have been wrong. Either way, we will at last know what we are fighting. For a tax is not abolished by burning its tollbooths: it is abolished by drying up its base, closing its channels, flushing out its fronts, and making its conversion impossible. It is on this last front, that of assets and their legitimation, that the decisive game is now being played. It is also where the coalition is probably most vulnerable, because that is where it now has the most to lose.

Methodological note. This column draws on two distinct bodies of evidence, identified as such. On the one hand, interviews conducted separately, on condition of anonymity, with operators and managers in the transport, trade, and logistics sectors; only mechanisms described independently and consistently by several sources have been retained, and no amount, place, identity, or detail that could identify a company or a facility has been drawn from them. On the other hand, public sources against which these elements were checked: the interim report of the United Nations Panel of Experts of 27 March 2026 (S/2026/241), in particular its section on illicit financial flows and its annex on extortion; the final report of 25 September 2025 (S/2025/597) on extortion fees, maritime detours, cash seizures, and specialized judicial units; report S/2024/704; the Global Initiative’s April 2025 bulletin and its study on criminal governance; the International Crisis Group report of 15 December 2025; as well as reporting by the Haitian and international press, including coverage of the commune of Carrefour, and the reporting cited in these documents. The tax on homes mentioned in Carrefour rests on field accounts corroborated by public reports; no amounts have been taken from them. Figures whose primary source could not be established have been omitted.