Counting the herd: the records behind an animal-health forecast

Ask how many head of cattle a country holds and three official figures will be available, differing by a fifth. Each is correct for its own purpose. A tax register counts holdings that claim a rural rate. A movement database counts animals with an identifier that crossed a boundary. A herd survey counts what a sample of farmers said in a particular month. A vaccine forecast built on the wrong one of the three will be wrong in a direction nobody can trace afterwards.

Our animal-health research therefore begins with the obligation that produced the record. Where identification and movement recording are compulsory and enforced, the movement database approaches a census and can carry a market model. Where identification is compulsory but enforcement is light, the same database counts compliance rather than animals. Where recording attaches to a subsidy, the count rises and falls with the subsidy. The obligation explains the number better than the number does.

Species differ sharply in how well they are counted. Cattle are individually identified in most markets we cover, because disease control and trade both require traceability of a single animal. Pigs and poultry are counted by batch or by house, which suits vaccination economics but not treatment economics. Companion animals are barely counted outside registration schemes and insurance books, which is why forecasts for that segment carry the widest error and attract the most confident claims.

This matters commercially because the addressable market for a veterinary product is a population multiplied by a treatment rate, and both terms are estimated. A producer reporting strong growth in a market where the animal count is drifting upward for administrative reasons has not necessarily sold more product. Our desk reconciles reported unit sales against doses implied by the population and the label, and a persistent gap between the two is a diligence finding rather than a rounding difference.

Data provenance is a firm-wide requirement and applies to research inputs as strictly as to client data. Every series this desk uses is recorded with its origin, its collection basis, its update frequency and the lawful basis on which we hold it. Where a series is licensed, its permitted use is recorded beside it. The work is administratively dull and it prevents the commonest failure in sector research, which is a number carried forward for years with no traceable beginning.

The practical output is narrower forecasts held with more confidence. Where the underlying records are strong, we will size a market and support a position. Where they are weak, we say so, and the Investment Committee treats the position as one that must be justified without a market size. Several animal-health commitments have been made on that basis, underwritten on customer relationships and the registration portfolio alone. A confident number from an unexamined source is worse than none.

Published 2022-03-16 by the Information & Data division. Research is prepared for eligible counterparties and does not constitute advice.