Coffee and cocoa: behind the 2026 investment note,
the border that absorbs the harvest

Presented at FAO’s Hand-in-Hand Investment Forum in Rome on 13 to 16 October 2026, the 2026 investment note Investing in Haiti’s Agroforestry-based Value Chains puts a USD 35.7 million price tag on a programme to revive coffee, cocoa and roots and tubers. The technical case is sound. Its success nonetheless rests on a variable the note funds only at the margin: the ability of Haitian cooperatives to buy the harvest before Dominican intermediaries do.

Analysis · 24 October 2026 · 12 min read · Agriculture · Foreign trade · Investment

In 2024, Haiti produced 4,371 tonnes of coffee and formally exported only 161 tonnes, worth USD 1.57 million. Less than 4 percent of national output therefore passes through a formal export channel. The rest is consumed domestically or, for a share the note puts at up to 80 percent of the crop, sold as cherry across the border, where it is pulped, dried and resold under another origin. More than the USD 90 million in export earnings of the 1980s that the document recalls in its opening pages, this figure describes the problem the programme has to solve: Haiti’s coffee value chain has not disappeared, it has changed country at the processing stage.

The note, issued by the Government of Haiti through the Ministère de l’Agriculture, des Ressources naturelles et du Développement rural (MARNDR, the agriculture ministry) and the Institut national du café d’Haïti (INCAH, the national coffee institute) and prepared with the Investment Centre of the Food and Agriculture Organization of the United Nations (FAO), answers this diagnosis with a territorial investment offer. It deserves to be read for what it is: a fundraising document, built to convince development banks and climate funds, whose assumptions its own authors describe as "under validation".

A territorial portfolio rather than three projects

The programme rests on ten coffee clusters of 300 hectares each, spread across four regions: the South-East (Thiotte, Belle-Anse, La Vallée de Jacmel), the Greater South (Beaumont, Camp-Perrin, Baradères), the North and North-East (Dondon, Borgne, Vallières) and the Centre-Artibonite (Baptiste, Marmelade). Each cluster combines agroforestry plots, a wet-processing centre or solar-drying units, a small roasting business and an existing cooperative as its anchor. Coffee absorbs USD 27 million of the 35.7; cocoa, confined to 650 hectares below 900 metres of altitude in Grande-Anse, the North and the North-East, receives 3.2; roots and tubers, 1.8. The balance funds shared services: 36 nurseries, 33 bio-input units, 33 seed centres, a digital traceability platform and financing mechanisms.

The portfolio logic is the main contribution of this third version, reviewed by the Government in September 2026. The three crops share the same plots along an altitude gradient, the same nurseries and the same digital platform. Yam and cassava provide household income during the first four years, while the coffee and cocoa trees come into production. According to the note’s calculations, this combination lifts the programme’s economic rate of return from 13.4 percent for coffee alone to 14.2 percent for the whole, and brings the number of beneficiary families to 9,300.

The expected results at maturity are precise: around 700 tonnes of green coffee a year, including 140 tonnes of specialty grade, 620 tonnes of fermented and traceable cocoa, USD 4 million in formal coffee exports and USD 3 million for cocoa, an average income gain of USD 790 per family per year, and 1,520 jobs in 87 solar-powered processing units. The payback period is estimated at eleven years.

Set against the stated ambitions, these figures call for restraint. The 3,000 hectares of coffee planned represent a little more than a third of the 8,477 hectares still harvested today, and less than 4 percent of the 80,000 hectares of 1961. The USD 4 million in targeted exports amount to less than 5 percent of the 1980s earnings the note cites as its horizon. The programme does not rebuild the value chain; it installs a formal prototype at a scale large enough to be demonstrated, then replicated if partners follow.

The border as a processing plant

The heart of the problem is not agronomic. The yields of 200 kilograms per hectare recorded among smallholders can be raised to 500, as the note shows from the results of the EFOSE project of the International Fund for Agricultural Development (IFAD) and the PITAG project of the Inter-American Development Bank (IDB). The genetics are a real asset: the Typica and Bourbon mixtures of the creole gardens, documented in a 2024 study, are the heirloom origin closest to the US market. The bottleneck lies downstream, in processing: without pulping, fermentation and drying centres, the producer sells cherry, and cherry goes to whoever pays cash on the day of picking.

This is where the Dominican Republic comes in. The overland trade corridor between the two countries carried USD 776 million in trade between January and August 2025. Dominican buyers have the cash, the logistics and the processing units that Haitian cooperatives lack. Haitian coffee does exist on international markets, but under a Dominican flag and without any processing value staying in the hills. The same mechanism applies to cocoa, 70 percent of which is exported unfermented to Malaysia at a discount of around 30 percent to fermented beans.

The note says so plainly: working capital is "the first constraint" faced by cooperatives of the type of the Fédération des coopératives cacaoyères du Nord (FECCANO, the northern cocoa cooperatives’ federation). Its answer is a revolving fund of USD 150,000 per cluster, USD 1.5 million in all, topped up by a national guarantee facility of USD 1 million to unlock credit at 3 to 5 percent from microfinance institutions.

Comparing this with expected output gives the measure of the bet. At maturity, a 300-hectare cluster is meant to deliver around 70 tonnes of green coffee, worth, at the conservative price of USD 6.5 per kilogram used in the financial analysis, some USD 455,000 per season. The revolving fund therefore covers about a third of that value, and more in the early years when volumes are low. It is enough if the cooperative manages to turn its cash over several times a season, that is, to sell quickly what it has processed. It falls short if the exporter pays at sixty or ninety days, a common practice in the coffee trade. In that second case, the Dominican buyer keeps his decisive advantage: he pays the same day.

Returns calculated at the top of the cycle

The note was written in an exceptional price environment, and it knows it. The composite indicator of the International Coffee Organization (ICO) stood at 248.9 US cents per pound in June 2026, a high but volatile level, with average intraday swings of 8.6 percent over the month. Cocoa traded at around USD 6,300 per tonne on 2 September 2026, two to three times its 2020-2023 average. The authors took the precaution of using USD 6.5 per kilogram for green coffee, against the USD 8.5 that Haitian exporters actually obtained in 2024, and of presenting coffee and cocoa as medium-term investments whose quality and traceability premiums would cushion reversals.

The market has already begun to test that assumption. On 1 October 2026, a few days before the Rome forum, New York cocoa futures fell back towards USD 5,100 per tonne, their lowest since late July, on the back of a global surplus in the 2025-2026 season and certified stocks at their highest in more than two years. In one month, the reference price for the cocoa module lost close to a fifth. The outlook for 2026-2027 remains uncertain, between a 400,000-tonne deficit and a small surplus depending on weather scenarios in West Africa, and it is precisely that uncertainty which rules out building a programme on a price level.

The programme’s safety margin is thin. A benefit-cost ratio of 1.10 in the base scenario means that a fall of about 9 percent in discounted benefits, or a 10 percent cost overrun, brings the net present value to zero. The models for densifying existing systems (AFS 2 and AFS 3), which carry two thirds of the coffee area, show ratios of 1.11 to 1.15 and rates of return of 14 to 17 percent over fifteen to twenty years. The 21 to 38 percent rates highlighted in the summary concern the processing centres and newly planted plots, that is, the components most dependent on the selling price and on market access. The 18.6 and 22.6 percent scenarios include a valuation of sequestered carbon at World Bank shadow prices; they describe a collective benefit, not a cash flow for the cooperatives.

The European regulatory calendar calls for the same caution. The note presents the European Union Deforestation Regulation (EUDR) as a first-mover advantage, applicable from December 2025 and from mid-2026 for small operators. In the meantime, the European Union has postponed application to 30 December 2026 for large and medium-sized companies and to 30 June 2027 for micro and small enterprises, having already pushed the deadline back once. The traceability argument remains valid in substance, since specialty buyers already ask for it, but the regulatory premium will arrive later than planned, and possibly diluted if the simplifications under way are confirmed.

The blind spots: financing, institutions, security

The first blind spot is the financing itself. Of the ten clusters, only the South-East has an identified anchor, the ninth replenishment of the Global Environment Facility (GEF). The other nine are presented as "open" to development banks, climate funds and bilateral partners, even if ongoing GEF projects in the North and the Greater South may contribute, at USD 2.7 million per coffee cluster and USD 550,000 per cocoa module. The USD 35.7 million is therefore an envelope to be raised, not a secured budget. The start-up schedule, staggered between 2026 and 2028, will depend on the pace of commitments, and the record of previous programmes in Haiti shows that project-aid disbursements rarely reach half of forecasts: for fiscal year 2024-2025, the Cour supérieure des comptes et du contentieux administratif (CSCCA, the supreme audit institution) recorded a realisation rate of 47.7 percent for project grants.

The second blind spot concerns institutions. INCAH is expected to steer value-chain governance, varietal research, statistics and the innovation platform, with USD 1 million for research and a share of the USD 7.6 million for capacity building. The note says nothing about its current staffing, its operating budget or its actual presence in the four regions. The cooperatives cited as anchors have uneven track records: FECCANO, with twelve cooperatives and 8,000 members, has exported fermented cocoa to Ethiquable at a 45 percent premium for several years; other coffee networks have no demonstrated commercial capacity, or no longer have one. The programme nonetheless entrusts these structures with the ownership or co-management of 87 processing units.

The third blind spot is security, handled by avoidance. All sites lie outside the metropolitan area, on the northern and southern logistics corridors, and the Centre-Artibonite is subject to "security-sensitive targeting". The choice is reasonable. It leaves untouched the question of transport between the hills and the ports of Cap-Haïtien or Les Cayes, and on to US markets, in a country where the Institut haïtien de statistique et d’informatique (IHSI, the national statistics office) attributes the 2024-2025 contraction of transport and trade to extortion on the main roads. Specialty coffee sells on the regularity of deliveries as much as on the quality of the cup.

A few internal inconsistencies are worth flagging, because they betray a document assembled quickly from sources of uneven quality. Cocoa exports for 2024 are given as USD 13.5 million on one page and USD 14.9 million on another. Cocoa production ranges from 3,000 tonnes according to FAO to 5,000 according to sector surveys. Average coffee yield is 516 kilograms per hectare in official statistics and 200 in field surveys. These gaps are not trivial: they measure the absence of a reliable agricultural statistics system, that is, one of the capacities the programme asks INCAH to build while relying on it to justify its forecasts.

An offer worth taking, provided it is completed

For Haitian actors, the note matters less for its rates of return than for what it reveals about where value is created. Across the whole programme, only USD 3 million goes to processing and energy equipment, against USD 13.6 million for production and inputs and USD 19.1 million for services, digital tools and capacity building. The 87 processing units are designed as cooperative enterprises or joint ventures open to public-private partnership, with matching grants of 70 percent for solar dryers and 50 percent for roasters. For an established exporter or roaster, that is the entry point: bringing the cash and the commercial outlet the programme funds only halfway, in exchange for a stake in units whose initial investment is largely subsidised.

The nurseries, bio-input units and seed centres, endowed with USD 5.1 million, open a second market. The note specifies that the business model of a bio-input unit yields a rate of return of about 11 percent over twenty years, provided the initial investment is grant-funded. It is a viable service activity for small rural businesses, not a private-equity opportunity.

For public decision-makers, the stakes shift to three decisions the programme assumes without taking them. The first concerns cross-border trade: as long as selling cherry to Dominican intermediaries remains safer and faster than delivering to a cooperative, no pulping centre will run at full capacity. The USD 150,000 of working capital per cluster is a start; the USD 1 million guarantee facility proposed by the Government will have to be topped up well beyond that to carry weight against a trade corridor worth several hundred million. The second concerns INCAH, which cannot simultaneously produce statistics, lead research, arbitrate governance and monitor ten clusters without an operating budget of its own, written into the finance law rather than into a project. The third concerns the consistency of the customs exemptions extended in April 2025 to agricultural enterprises and export activities: a programme that bets on formalising flows would benefit from tax advantages being conditioned on the traceability it funds elsewhere.

The note has the merit of framing the problem at the right scale, that of territories, and of recognising that Haitian coffee lacks neither a market nor quality. What it lacks is a national buyer able to pay on the day of harvest. Thirty-five million dollars fund trees, dryers and software. The buyer remains to be built.

References

  1. Government of the Republic of Haiti (MARNDR, INCAH) and FAO Investment Centre, Investing in Haiti’s Agroforestry-based Value Chains. Investment Note 2026, version 3, presented at the Hand-in-Hand Investment Forum, Rome. Source of all programme, production and export figures cited without further attribution.
  2. FAO, Hand-in-Hand Initiative Investment Forum 2026, forum programme, special event “Investing in Haiti’s coffee value chain”.
  3. European Commission, Access2Markets, Delay until December 2026 and other developments in the implementation of the EUDR Regulation.
  4. Tea & Coffee Trade Journal, The ICO reports coffee prices rebound from lows in early June, based on the ICO monthly report.
  5. Trading Economics, Cocoa Futures at Over 2-Month Low.
  6. Cour supérieure des comptes et du contentieux administratif, Rapport sur la situation financière du pays et sur l’efficacité des dépenses publiques, exercice 2024-2025, for the realisation rate of project grants, the April 2025 tax measures and the IHSI diagnosis on transport.

The comparisons between areas, volumes and values (the programme’s share of harvested area, the weight of working capital in a cluster’s output value, the change in cocoa prices between September and October 2026) are the author’s calculations from the note’s figures and the sources cited.

Also read

Stay in touch

Your name, your email, and we write to you at each new publication: a report, a case study, a brief. No set frequency, no promotional content.