Records, Carbon and Rebuilding a Broken Budget

#The bull that was never written down

The four-step sequence in the last lesson — breeding season, pregnancy diagnosis, cull the empties, condition at bull-in — costs nothing but attention. It also collapses if you cannot say, in August, on which day the bull went in and on which day he came out. Without those two dates there is no calving percentage and no inter-calving period, and the 608-day cow that quietly costs about R36 700 over her working life — at July 2026 weaner prices — stays in the herd because nobody can prove she is the problem.

That is the whole argument for records. Not compliance, not neatness. A KPI is arithmetic performed on a record that already exists; without the record there is no KPI at any price. The same discipline runs outward: every number handed to you — an aggregator's rand per hectare, a published budget — is a claim until you rebuild it.

#Seven fields, and nothing else

Most farm record-keeping fails by being too ambitious: somebody buys an app, records forty fields for four months, then stops. The minimum set that mechanically produces the whole KPI list is short enough to survive a bad year.

Read them as trends, not scores: one season's calving percentage tells you about last season's rain, five seasons tell you about your management. And do not skip field 7: it turns "the veld was good this year" into a grazing-days-per-100 mm figure per camp that you can rank and act on.

There is an eighth record, and it decides whether an animal can be sold at all.

#What a hectare of carbon is actually worth

Carbon income arrives on Highveld farms as a phone call, and the pitch is a rand per hectare with no denominator attached. Take the most optimistic SA claim at face value and do the arithmetic. AgriCarbon's programme terms, as advertised in July 2026, promise 1 to 2 tonnes of credits per hectare per year with payments starting from US$10 per verified carbon credit (AgriCarbon). At the top of that range and at roughly R16.50 to the dollar, that is about R330 per hectare per year, gross.

If it were net, reliable and annual, R330/ha would change a Highveld farm. Whether it is any of those three is the actual question.

#Four reasons the gross is not the net

The aggregator's share. Registry fees, validation and verification come out of the gross, and on a small farm they are largely fixed costs spread over few hectares. That is why aggregation exists and why the aggregator takes a cut: GRASS bundles verification cost across thousands of farmers and hundreds of thousands of hectares. For a 100 to 500 ha farm an aggregator is the only realistic route in; a solo project is not economic.

MRV cost. Baseline and repeat soil sampling, measurement, reporting and verification are yours. International benchmarks for direct soil-carbon sampling run at US$15–30 per acre — an undated overseas figure with no confirmed South African equivalent, but on low-rate rangeland a bill of that order can exceed the credit revenue outright.

Issuance timing. Credits are issued only after a monitoring period is verified, and GRASS's first issuance covered 2021 to 2023. That is a lumpy payment years after you changed how you graze, not an annual cheque.

The buffer pool. Registries withhold a share of credits against reversal — a drought plus one overgrazing episode can undo semi-arid soil carbon — so permanence discounts sit between the tonne in your soil and the rand in your account.

Under the Carbon Tax Act 15 of 2019 agriculture is not itself a taxed sector, so a farmer takes part as an offset project developer selling into somebody else's compliance obligation. That is market access, not a liability — and markets carry pricing risk.

The clearest guide to priority is Meat Naturally's own tally: over ten years, R134 million to rural communities through formal livestock and wool sales, against R2.7 million from carbon credit sales by 15 communities in 2026 (African Farming, 11 July 2026). Different periods, so treat roughly fifty to one as an order of magnitude, not a ratio. Never buy the carbon story before the market-access story.

#Rebuild the broken budget

The Land Bank's Livestock Enterprise Budget 2023/24 (v2) is the closest thing South Africa has to a public standardised enterprise budget set, and this course has used its structure throughout. It also contains internal arithmetic that does not reconcile.

The book is not worthless. A published budget is a line-item structure, not a source of numbers. Which cost lines exist, in what order, at what unit of account, is hard to build from scratch and worth having. The values are 2023/24, belong to somebody else's farm, and at least three of them are wrong.

Step 1 alone settles the Dorper page. R896.92 less R566.08 is R330.84 per small stock unit, not R161.46 — so a farmer who accepted the printed figure would judge Dorpers to earn roughly half what that budget actually shows, and might put cattle on those hectares for the wrong reason. One subtraction, two minutes, a different enterprise decision.

#What to do on Monday

Buy a hard-backed notebook and a rain gauge, and start with field 2: the day the bull goes in and the day he comes out. Then take the last budget anybody sent you — a carbon term sheet, a feed rep's projection, a published gross margin — and subtract the costs from the income yourself before you read the conclusion at the bottom.