Paying for It: People, Tenure and the Money

#Two workers and six hundred and twenty-five cows

The Land Bank's 18-month beef budget runs two full-time workers. It also runs 625 cows, so the fixed labour line comes to R185.25 per cow at 2023/24 input prices (Land Bank). At that scale an employee is a small line on a long list.

Now put one of those employees on a hundred hectares of Highveld veld. From 1 March 2026 the national minimum wage is R30,23 per ordinary hour, up 5% from R28,79, and farm workers are on the full general rate (Labour Guide) — parity phased in in the early 2020s, not a 2026 novelty. At 45 ordinary hours a week that is roughly R5 895 a month, about R70 700 a year, before UIF and COIDA. The Highveld base case this module works at the end returns about R470 a hectare of gross margin — some R47 000 on a hundred hectares. One lawful employee costs half again as much as the whole enterprise earns.

That is not an argument against employing anybody. It is why a plan that quietly assumes "more labour, less machinery" must be costed before it is admired.

Labour also decides enterprise choice. The Land Bank's 100-doe Boer goat budget is the closest thing in that book to a veld-only enterprise, and labour is 71% of its costs — which works on family labour and stops working the day two minimum-wage employees are needed for a hundred does.

#The one place the herding wage bill has been measured

Herding gets offered as three things at once: a grazing tool, an answer to theft and predation, and a substitute for fencing you cannot afford. One South African cost-benefit study has actually priced it, and it is worth more than a shelf of enthusiasm.

Herding still won — 1,55:1 is a real return, and off-take value doubled in the same sample. But the honest number is the smaller one, and it is the one to carry into your budget: a saving that depends on someone in the household working unpaid has been moved, not made.

There is a second re-costing to do. Hired herders working alone managed 24% of the herds in that study and earned about R800 a month in 2016 rands — a line that cannot be carried forward now that the statutory floor is R30,23 an ordinary hour. The structure of the finding survives; the rand figures are a 2016 cost base and must be rebuilt at today's wage.

One further result belongs in a first-year plan. The reported causes of loss were tick-borne disease 33%, theft and disputes 23%, drought 19%, predation 13% — animal health, not grazing management, was the largest single loss line.

#Who already lives on your land

Tenure is where transition plans die quietly. An occupier under ESTA is someone residing on another person's land with that person's consent — a residence relationship, not an employment one. An employee living in town is not an occupier; an occupier who never worked for you still holds the Act's rights.

Work s8(4) on a real farm. A man has lived in a house on the property for twelve years and turned 61 last year. His right of residence is not something you negotiate away with notice, however carefully drafted. That belongs in the purchase price, the lease, the labour plan and the cash flow, before signature. If you are buying or leasing, ask who lives there and for how long, and get the answer in writing — not to keep people off the land, but to know which rights on it are not yours to move.

#What a lender will lend against

The financing problem a transition creates is a timing problem, not a size problem. The spend is front-loaded — water, fencing, handling, a destocked season, a year of learning — and the return sits in years three to seven. Ordinary production credit is structured for a season and secured against land the bank values for its existing use. That mismatch, not the interest rate, breaks transitions.

Name what the bank is looking at. Your security is livestock — moveable, stealable, and in a foot-and-mouth controlled area temporarily unsaleable. Your improvement is in the veld, which does not appear in a valuation and cannot be repossessed. A lender is not being obtuse when it declines to fund better basal cover.

Arrange the credit before the cash trough. On the Highveld that trough is spring: the winter lick spend is still running and no income has arrived. A lender's answer changes when you are visibly desperate.

#The risks that belong in the budget, and the income that does not

Two risks get hoped about instead of budgeted, and one income line gets counted before it exists.

Theft is a cost line. An Eastern Cape study of police records across 2018 to 2024 recorded 209 235 animals stolen worth R715 million, with per-head recorded values of R12 876,35 for cattle, R1 675,10 for sheep and R2 021,90 for goats. Only 17,2% of animals were recovered, and 93,1% of the value never came back. Read the structure: sheep were the majority of animals stolen, but cattle — 14,7% of them — accounted for R399 million of the loss (Ndzungu, Ngoshe & Jaja 2025). Those recorded values sit well below 2026 replacement cost, so official statistics understate what a theft costs you today. Budget expected annual loss as herd × local incidence × replacement value, and weight prevention towards cattle.

Carbon is real money and still cannot carry the plan. AgriCarbon's upper claim is 1–2 tonnes of credits per hectare per year at payments "starting from US$10 per verified carbon credit" (AgriCarbon). At the top of that range and an assumed R16,50 to the dollar — an assumption, not a quote, and a 10% rand move shifts it by R33 — that is about R330 a hectare a year, gross. GRASS's issued volumes, 266 255 VCUs over the 2021–2023 monitoring period across more than 95 000 hectares, imply roughly 0,93 t/ha/yr, half that claim (CarbonCredits.com).

Before the next planting rains: write your roster and multiply it by R30,23; list every person residing on the property and the year they arrived; and get the exclusions on your livestock mortality policy from your broker in writing.