Goats and sheep are the cheapest way into livestock farming in Ghana, and the returns are genuine when the animals are kept alive. The harder question this pillar answers is which of the available numbers a plan is entitled to be built on, and which are a shape rather than a budget. The cost of entry is low: farm guides put the cost of starting a small flock of ten to twenty goats, with simple housing and initial feed and medicine, at about GH¢6,600 to 22,5001, with monthly feed at about GH¢20 to 40 per animal per month.
Academic and trader studies confirm the business is genuinely profitable. A profitability analysis of 100 rural keepers in three communities of the Atwima Mponua District in the Ashanti Region, measured over a full one-year production period, found a revenue-cost ratio of about 1.86, and 2.33 for sheep against 1.36 for goats2, and a survey of 284 traders found healthy net marketing margins3.
On the cost side, the biggest single item for anyone buying animals is the animals themselves. A study of small-ruminant traders in Kumasi and Tamale found that buying the live animal made up about 85 percent of total marketing cost, split further as the animal itself 84.68 percent of the batch cost, transaction cost 13.23 percent and regulatory cost 2.10 percent, with transport itself running to 7.20 percent of batch cost on 2015 field prices. When one line already commands that much of the budget, an error of 8.5 percent on it, on this cost split, already costs as much as the whole transport bill, and any larger error costs more, so buying well, not managing costs elsewhere, is the trade’s real skill. For a producer the main costs are breeding stock, housing, feed, especially in the dry season, and animal health.
On the return side, the trader study found a net marketing margin of about 24.85 percent for goats and about 17.73 percent for sheep, and a return of about GH¢0.25 on every cedi invested for goats and about GH¢0.18 on every cedi invested for sheep, again on 2015 prices; the authors set those margins against a return on savings of 22 percent at the time of the survey, which goat trading beat and sheep trading did not.4
A marketing margin is earned once on a transaction; a savings rate is earned over a year. The figure below sets them side by side with the missing term, how many times a year a trader’s capital turns over.
The left panel is the comparison the original study made, and it is not like for like: a net marketing margin is the return on one transaction, a savings rate is annual, which is why the third bar is hatched. The right panel supplies the missing term. The trader turns the same capital about 36 times a year, so roughly 25 percent a trip is not a slim premium over a deposit account, it is why the trading layer captures what it does.
The breeding model and the fattening model have different economics. A breeding flock is a slow, compounding business: the cost is the breeding stock, housing and year-round feeding and health, and the return builds as the flock multiplies and surplus animals are sold. Its enemy is mortality, because every kid or lamb lost is lost flock growth. The fattening model is a fast, repeated trade: buy lean rams in the cheap season, feed them hard for seventy-five to ninety days, a figure you should test against your own first cycle, to add weight, and sell into the festival peak. Its margin is the gap between the lean-season buying price and the festival selling price, minus feed, so its enemies are buying at the wrong price and a price crash at sale. Farm guides cite net profit margins of 35 to 60 percent for well-run goat farming, highest at the festive seasons56.
| Item | Figure | Source and caveat |
|---|---|---|
| Start-up, small flock (10 to 20 goats) | about GH¢6,600 to 22,500 | Accra Street Journal (2025) |
| Monthly feed per animal | about GH¢20 to 40 | Accra Street Journal (2025); varies by season |
| Revenue-cost ratio (rural production, one year) | about 1.86; 2.33 sheep, 1.36 goats | Agyeman et al. (2021); peer-reviewed; 100 keepers in three communities of the Atwima Mponua District, Ashanti Region |
| Net marketing margin, goats | about 24.85 percent | Kassoh et al. (2021); 284 traders, 2015 fieldwork |
| Net marketing margin, sheep | about 17.73 percent | Kassoh et al. (2021); 284 traders, 2015 fieldwork |
| Return per cedi invested (goat / sheep) | about GH¢0.25 / GH¢0.18 | Kassoh et al. (2021) |
| Live-animal purchase, share of marketing cost | about 85 percent | Kassoh et al. (2021) |
| Net profit margin, goat farming | 35 to 60 percent (festive peak) | Accra Street Journal (2025) |
| Fattening cycle | about 75 to 90 days | A planning assumption |
Sources: Accra Street Journal (2025); Agyeman, Shuai, Baion and Antwi (2021); Kassoh, Ohene-Yankyera, Nimoh, Apaassongo and Bangura (2021), whose fieldwork was carried out in 2015, so its percentages are structural and its cedi amounts are a 2015 vintage.
None of the figures above prices a whole fattening cycle from purchase to sale, so this pillar builds one, set out here so a reader can audit it rather than trust it. The model is a single cycle of thirty rams: purchase at GH¢1,100 a head is GH¢33,000, feed over the cycle GH¢9,000, transport and market charges GH¢1,500, and veterinary and health GH¢900, for a total cost of GH¢44,400, about US$3,964 at about GH¢11.20 to one US dollar (Bank of Ghana, 18 June 2026). Selling at GH¢1,850 a head with 5 percent mortality gives 28.5 animals sold and revenue of GH¢52,725, so net profit is GH¢8,325 a cycle, a revenue-cost ratio of about 1.19 on that cycle. Do not read that 1.19 against the 1.86 in the table above as fattening beating or losing to breeding: one is a single eighty-two-day cycle and the other a full year.
Sensitivity analysis asks a plainer question than that model can answer on its own: if one input is wrong, how wrong does the answer become, and ranking the swings by width tells an operator where to put attention rather than where to worry out of habit. On this base case the order, widest first, is sale price, purchase price, mortality, feed cost and cycle length. Prices are swung 25 percent each way, mortality to 20 percent, feed 40 percent each way and cycle length from 60 to 120 days. Widen the price rows if you want the observed extremes: Ghana’s own 2026 Eid week moved Kumasi down about 40 percent and Accra up about 31 percent78.
Ranked by the width of the swing, and the ranking is the finding. Price moves the cycle more than anything done inside it, while feed, the line most operators worry about, swings it by GH¢7,200 against mortality’s GH¢11,100. Feed is not the narrowest bar: the cycle length is, at GH¢6,546. Purchase price is the half of the price pair a buyer controls, and 25 percent above the base the cycle barely clears its costs. Cycle length is varied rather than fixed, because it is an assumption.
Two of the top three variables are prices, and only one of them is a decision an operator actually takes. Sale price follows the festival calendar and the exchange rate, and a fattener who has already bought the animals cannot move it; purchase price is fixed on a given day for a given animal, and 25 percent above the base the cycle barely clears its own costs. Feed, the line most operators worry about most, moves the answer far less than either price and less than mortality, a swing of GH¢7,200 against mortality’s GH¢11,100, and the cycle length nobody could source is narrower still at GH¢6,546, so it matters about as much as feed and roughly three-fifths as much as mortality.
The two tools below are the 30-ram cycle above, made live. Slide the drivers a fattener can hold and watch the cost, the revenue and the net move; then read the same model as a sensitivity ranking, and see for yourself that the sale price the operator cannot move dwarfs the feed line they fret over. Every figure is indicative, so overwrite each one with your own pen’s numbers.
Slide the drivers a fattener can hold over one festival cycle of about 82.5 days: the number of rams, the lean-season buy price, the festival sell price, the feed cost per ram and the mortality rate. Vet at GH¢30 and transport at GH¢50 a ram sit behind the sliders. Every figure is indicative, so overwrite each one with your own pen’s numbers.
Base case (Festival Fattening P&L, Goats and Sheep Cost and Returns Model v2): 30 rams bought at GH¢1,100 is GH¢33,000, feed at GH¢300 a ram is GH¢9,000, vet at GH¢30 a ram is GH¢900 and transport at GH¢50 a ram is GH¢1,500, for a total cost of GH¢44,400; after 5 percent mortality 28.5 rams sell at GH¢1,850 for GH¢52,725, a net of GH¢8,325 a cycle and a revenue-cost ratio of about 1.19, on a declared working-assumption cycle of about 82.5 days. Every figure is indicative: no field-validated Ghanaian small-ruminant enterprise budget exists (gap PD-8), so re-price each line against your own pen before you rely on it. The sell price is set to GH¢1,850 to match this pillar’s own worked base case5.
The same cycle, re-run as a sensitivity read. Each bar is how far the net moves when that one driver is swung across the range the model uses, holding the others where you left them: prices 25 per cent each way, feed 40 per cent each way, mortality across the 0 to 20 per cent band, and the declared cycle length from 60 to 120 days. The order is the finding.
The festival price the survivors fetch, set by the calendar and the exchange rate, not by the fattener. It is the widest swing of all and, once the animals are bought, the one an operator cannot move. Ghana's own 2026 Eid week moved Kumasi down about 40 percent while Accra rose, so a national festival price is not safe to assume.
The lean-season buy price, about 74 percent of the whole cycle cost. It is the widest swing the fattener actually decides, and 25 percent above the base the cycle barely clears its own costs. Buying well, in the lean months and in lots from several sellers, is the trade's real skill.
The share of rams lost before the sale. A dead ram is a full sale price gone, not a cost avoided, which is why it swings the net wider than feed even though it never shows as a cost line. Quarantine, arrival vaccination and a deworming calendar are what buy it down.
The line most operators worry about most, and it moves the net less than either price and less than mortality. Buying and storing crop residues and by-products in bulk at harvest trims it, but it is not where the cycle is won or lost.
The declared working-assumption cycle of about 82.5 days, carried as a swing variable and not a fact because no Ghanaian source pins it. Feed scales with days held while purchase, vet and transport do not, so it is the narrowest swing of the five, about as much as feed.
Sensitivity holds the other drivers where you left them and swings one: prices 25 per cent each way, feed 40 per cent each way, mortality across the 0 to 20 per cent band, and the declared cycle length from 60 to 120 days with feed scaled by days held. On the base case that gives swings of about GH¢26,363 for sale price, GH¢16,500 for purchase price, GH¢11,100 for mortality, GH¢7,200 for feed and GH¢6,546 for cycle length. Ranks match the Sensitivity sheet of the Goats and Sheep Cost and Returns Model. Indicative figures, to validate against your own pen.
Sensitivity says which variables move the answer; this asks the more practical question of which of them an operator can actually put a hand on, and what the move is worth, on the same cycle. Sale price is the widest swing in the chart above and is absent from the table below for a reason: a fattener who has already bought the animals can choose when to sell, but not what the market pays on the day.
| Cost line | Base | The lever | Saving | What to track |
|---|---|---|---|---|
| Animal purchase | GH¢33,000, 30 rams at GH¢1,100 | Buy in the lean months rather than three weeks before a festival, and in lots from several sellers on several days | GH¢3,300 on a 10 percent better entry | Record price, date and seller for every animal bought over a full year |
| Mortality and loss | GH¢2,775 of revenue at 5 percent | Quarantine incoming animals, vaccinate on arrival rather than on a fixed schedule, and deworm on a calendar | GH¢1,665 if mortality falls to 2 percent | Count deaths by week and age class; an annual figure hides when the loss happens |
| Veterinary and health | GH¢900 | Spend more here, not less: the quarantine, the arrival vaccination and the deworming calendar are what buy the mortality saving above | GH¢450 of extra spend, not a saving | Cost every treatment against the deaths avoided in the same batch |
| Feed over the cycle | GH¢9,000 | Buy and store crop residues and by-products in bulk at harvest rather than in the dry season (Pillar 4) | GH¢1,800 on a 20 percent reduction | Weigh what is fed and weigh the animals fortnightly |
| Transport and logistics | GH¢1,500 | Consolidate loads with neighbouring keepers and avoid the festival haulage peak | GH¢450 on a 30 percent reduction | Log every payment made on the road against the animals carried |
| Working-capital timing | GH¢44,400 tied up for the working-assumption cycle | Stagger purchases and sales so the pen is never idle and capital is never parked between cycles | About one extra cycle a year on the same capital | Date every purchase and sale, and track days held as a standing indicator |
No published Ghanaian budget prices a commercial fattening cycle from purchase to sale; Agyeman et al. (2021) and Adams et al. (2021) cost a traditional home-bred flock on decade-old data. The four costed savings are GH¢3,300, GH¢1,665, GH¢1,800 and GH¢450, which come to GH¢7,215 before the extra GH¢450 of veterinary spend that buys the mortality line, so about GH¢6,765 net on a base cost of GH¢44,400.
The same cycle asked a different question. Sensitivity says what moves profit; this says what an operator can do about it. The two bars worth acting on are the top and the middle: a 10 percent better purchase saves GH¢3,300, more than any other single lever on the chart though a little less than the GH¢3,915 the other three raise between them, and cutting mortality from 5 percent to 2 percent recovers GH¢1,665 that never appeared as a cost line. The veterinary bar rises deliberately, because the GH¢450 of extra spend there is what buys the saving above it, which is why the net of all four levers is about GH¢6,765 rather than GH¢7,215.
There is a part of the return that does not show up as cash, and it explains why so many households keep goats and sheep even when they do not run them as a business. A study in northern Ghana found that at least 60 percent of the net benefit from sheep and goat production came in non-cash forms: the animal as a savings account that can be sold in an emergency, as insurance against a bad harvest, as a source of manure that saves on fertiliser, and as a store of wealth that holds value better than cash during inflation9. For a commercial entrant this is useful context: it explains why owners hold animals rather than selling, which keeps market supply tight, and it points to manure as a real secondary product on a mixed crop and livestock farm.10
paying too much for the animal at purchase, which on this cost structure is the largest and least recoverable error a fattener can make.
Buying the live animal is about 85 percent of a trader’s marketing cost11, and on that split an error of 8.5 percent at purchase already costs as much as the entire transport bill for the batch. In this pillar’s own sensitivity analysis, purchase price is the second widest swing in the result after sale price.
the fattener and the trader hardest, and the new entrant hardest of all, because valuing an animal on the hoof is a skill learned by repetition and paid for in mistakes.
buy in small lots until the eye is trained, weigh or measure rather than estimate where a scale or a girth tape is available, walk away from a market day rather than fill a truck at a bad price, and keep a written record of purchase price against final sale weight for every animal, which is the only way the skill compounds.
bringing finished animals to market when prices are falling, which the 2026 record shows can happen even in a festival week.
Sale price is the widest single swing in this pillar’s sensitivity analysis. Pillar 2 records Kumasi ram prices falling to about GH¢1,500 from about GH¢2,500 a year earlier in the same Eid window that saw Accra rise, so a national assumption about festival prices is not safe12.
anyone holding stock to a date rather than to a price, and worst of all a borrower whose repayment is timed to a single sale.
spread sales across more than one market and more than one week, agree a price before trucking where a buyer will commit, and hold enough feed to carry animals past a bad week rather than selling into it.
Buying agent, paid on purchase discipline. Purchasing is about 85 percent of a trader's cost and the widest controllable swing in this pillar's sensitivity analysis, yet almost nobody in the chain is paid to buy well. An agent who values animals on the hoof, works a circuit of markets and takes a share of the saving is selling a skill, not capital.
Contract fattening on someone else's animals. The fattener's biggest exposure is the purchase price of stock. Fattening animals owned by a trader or an institutional buyer removes that exposure entirely and converts the business into a fee on weight gained, which is a far easier proposition to finance.
Festival-window finishing. Both the sensitivity analysis and Pillar 2's price record say sale price moves the answer more than any cost line. A finishing operation timed to the Eid and Christmas windows, buying thin animals in the cheap months, is the single highest-leverage use of the same capital.
A record-keeping service, then a costing service. No Ghanaian operator publishes a real cost of production. Whoever first keeps clean per-cycle records across a group of fatteners owns the only cost dataset in the sector, and that dataset is itself sellable to lenders.
The digital and AI angle. Simple AI is already within reach here, and it is not the exotic kind. A phone-photo weight estimate from a handful of images, of the sort now common in cattle and pig systems elsewhere, would let a buying agent value an animal without a scale, which is the exact judgement this pillar shows decides the profit. Nothing in Ghana offers it for small ruminants today. Build the record set first; the model is worthless without it.
Goats and sheep are the cheapest entry into livestock farming; a small flock can start for roughly GH¢6,600 to 22,500.
The business is genuinely profitable where it is measured: a revenue-cost ratio of about 1.86 in production, and trader net margins of about 24.85 percent for goats and 17.73 percent for sheep.
Buying the live animal is about 85 percent of a trader's cost, so buying well, not managing costs elsewhere, is the key skill in any trading or fattening venture, and an error of 8.5 percent at purchase already costs as much as the whole transport bill.
On a fattening cycle, profit swings most on sale price, then purchase price, then mortality, then feed, then cycle length. Sale price is the one an operator who has already bought cannot move; purchase price, mortality, feed and transport all carry a lever. The purchase lever is worth more than any other single one, GH¢3,300 against the GH¢3,915 the other three raise between them.
The fattening cycle of about 75 to 90 days is a planning figure. Measure your own from the first cycle.
At least 60 percent of the benefit to traditional keepers is non-cash, savings, insurance and manure, which explains low off-take and tight market supply.
Where this connects. The mortality that ranks third in the sensitivity ranking above is bought down by the vaccination and biosecurity regime costed in Pillar 3. The feed cost that moves the answer less than either price, but decides whether an animal reaches weight inside the cycle, is a housing and dry-season storage decision taken in Pillar 4. And the margin structure measured here in the trading layer runs along the whole chain in Pillar 6, where the question of who captures the value is settled.