One virus decides whether a Ghanaian cocoa block is an asset, and nothing on sale resists it. The climate answer costs the farmer money. European compliance runs on a date most published documents state wrongly. And the sustainability reporting an underwriter would rely on does not survive arithmetic.
Cocoa swollen shoot virus disease is the only biological loss quantified in this evidence base. The measurement is 21.17 per cent in a two-season trial across 41 clones, from a mean 1.29 tonnes of dry beans a hectare a year on symptomless trees to 1.01 on symptomatic ones1: bean weight falls 2.5 per cent and beans per pod 5.5, so the loss is pods per tree rather than bean size. Grade barely moves either, since 92 per cent of symptomless trees against 82 per cent of symptomatic ones met the industry’s one-gram minimum bean weight. The loss is invisible in the bean and shows only as tonnage that never arrives. Prevalence is over 30 per cent, on the third country-wide surveillance programme2, and a little over 300 million diseased trees cut since the eradication programme launched in 1946.
No cocoa variety on sale escapes the virus, because none of the cacao varieties tested has been reported immune or resistant3. Ghana plants tolerance, not resistance, and the tolerance has degraded. Upper Amazon material was introduced in the early 1950s and later found to have low tolerance, and the 1978 Inter-Upper Amazon hybrids were judged not adequate for long-term protection. Of the sixteen clones the trial names as high-yielding with either no disease incidence or low yield reduction, only three are parental clones in Ghana’s seed gardens4. The other thirteen can be specified by name; Pillar 1 sets out what a nursery can issue.
Tolerance, not resistance
A CRIG on-station trial at Tafo: 41 clones, two seasons, one strain, managed shade. Not a farm-representative loss estimate, and no on-farm CSSVD yield-loss measurement is held for this playbook. The wider 30 to 50 per cent figure in circulation is a cited secondary estimate, not a measurement in this trial.
Cutting out infected trees fails on farmer opposition, limited funding and frequent discontinuities in implementation5. Rehabilitation is priced at about US$64.49 million in Bank financing and US$27.5 million in COCOBOD counterpart funding to rehabilitate 25,000 hectares of infested farms, roughly US$3,680 a hectare all in6; read any replanting quote against it. Chemical control of the mealybug vector was abandoned because nicotine, parathion and DDT were discontinued because they tainted the beans and cut quality and marketability on residue levels.
The European market now enforces the same failure mode through residue rules, and the exposure is live: about 70 per cent of insecticides used were COCOBOD-approved, and 9.9 per cent of farmers used active ingredients that were unknown and unapproved7, while CRIG tests and approves cocoa pesticides against the EU’s current list of approved active ingredients8. No source held here measures residues in Ghanaian beans, so this is documented practice, not a measured exceedance. Black pod appears in this evidence base only as a spraying-programme target: no Ghana loss figure, no pathogen species and no fungicide efficacy is held here, and this playbook will not supply them. Symptoms lag infection by two years in some genotypes before an infected tree shows symptoms, so a walk-over inspection dates the disease at best two years late. Underwrite on graft-indexed status, or price at the infected yield.9
Climate change lowers Ghana’s cocoa suitability without removing it. The model moves the country from 7,755,000 hectares above 50 per cent climatic suitability to 4,604,000 by the 2050s, a projected loss of 41 per cent10. Location decides: the projected losses fall on the northern fringes, while most of the Ghanaian cocoa belt is projected to stay highly suitable. And on the model, maximum dry-season temperature is projected to become as or more limiting for cocoa as dry-season water availability. Inaction is costed at about US$410 million a year by the 2050s, roughly 1 per cent of current real GDP, on a 90 per cent range of US$270 million to US$660 million11.
The northern fringe loses first
Modelled, not observed: 19 climate models on an intermediate pathway, 2040 to 2069 against a 1950 to 2000 baseline, published 2016. The classes sum to the same national area in both columns, so this is land moving down a class, not land leaving production. Schroth et al. map the suitability change across the belt in their Figure 5; charted here is the Ghana national row from their Table 2.
The recommended answer and the farm arithmetic then disagree. Shade is the only practical way of protecting cocoa trees from high temperatures, cutting leaf temperature by up to 4 °C12. Priced to the farmer who would plant it, it returns minus 10 per cent on internal rate of return and minus 10 per cent on net present value against business as usual, with fear of higher disease incidence a named barrier13. No smallholder carries shade establishment on those returns, so a supply chain that needs shade must buy it: a shade payment, a timber right or a contracted agroforestry service.
The regulation applies from 30 December 2026 for all companies except most micro and small operators, and 30 June 2027 for those14. Compliance requires geolocation coordinates for every plot, GPS polygons for plots larger than 4 hectares, the production date and the last supplier, against a cut-off that disqualifies any land converted from forest after 31 December 2020. COCOBOD holds the trade monopoly and will supply that data with the beans, on stated terms: companies will be given the traceability data they need once they buy cocoa from COCOBOD and pay the corresponding service fee. No source held here rates the fee, so no Ghanaian actor can yet cost compliance. Two gaps sit underneath it: Ghana does not have an operational forest monitoring system, and lack of funding prevents putting a national one in place, and certification cannot replace an operator’s due diligence obligation, nor does it cover the full production volume of Ghana.15
One scope question is open and should not be closed either way. COCOBOD’s interpretation is that child labour and forced labour fall outside the scope of the EUDR legality criterion, on guidance the same assessment records as not legally binding and interpretable only by national judges16. Separately, Ghana’s hazardous work list does not cover all occupations where child labour is known to occur, including in cocoa production17, so Ghanaian legal compliance alone cannot establish that a task performed by a child was lawful. Contract for child-labour evidence regardless.
Symptom onset lags infection, so a visual inspection describes the farm as it was two years earlier, and the loss shows up in pods per tree rather than in the bean.
Mean dry bean yield falls 21.17 per cent, 1.29 to 1.01 tonnes a hectare a year, while bean weight falls 2.5 per cent. Prevalence is put at over 30 per cent nationally.
Anyone buying, leasing or lending against an existing block; offtakers underwriting a volume from a named farm.
Require graft-indexed status block by block, not a walk-over. Where it cannot be produced, model at the infected yield and put the rehabilitation cost, about US$2,580 a hectare on the World Bank's own budget, into the entry price.
Require graft-indexed status, not a walk-over. Symptom onset lags infection by up to two years in some genotypes, so a visual inspection describes the farm as it was two years earlier. Require graft-indexed status block by block. Where it cannot be produced, model at the infected yield and put the rehabilitation cost, about US$2,580 a hectare on the World Bank's own budget, into the entry price.
Sell shade as a paid intervention, never as advice. Shade cuts leaf temperature by up to 4°C and is the only practical protection against the heat the models project, but it prices at minus 10 per cent on both IRR and NPV to the farmer who would plant it, with fear of higher disease incidence a named barrier. No smallholder carries that on its own returns, so a supply chain that needs shade must buy it: a shade payment, a timber right or a contracted agroforestry service, written into the contract.
CSSVD cuts dry bean yield 21.17 per cent in the one measured trial, and the loss is pods per tree, not bean size or grade. Prevalence is put at over 30 per cent.
No cacao variety tested is immune or resistant. Ghana plants tolerance, it has degraded, and only three of the sixteen best-performing clones are in the seed gardens.
Shade is the only practical protection against the projected heat, and it prices at minus 10 per cent on both IRR and NPV to the farmer. If your chain needs it, buy it.
EUDR applies from 30 December 2026, and 30 June 2027 for most micro and small operators. The traceability data comes from COCOBOD at a fee no source states.
Where this connects. Pillar 1 owns the planting material tested for tolerance here. Pillar 2 owns the European market share that makes compliance a condition of access. Pillar 4 owns the shade regime this pillar prices, and Pillar 5 turns disease loss into a margin.