Food security is a supply portfolio decision

Most supply risk is concentration risk, and concentration is frequently invisible in a cost comparison. A business may have four suppliers for a critical input, all sourcing from the same origin and shipping through the same route. On paper that is diversification; in practice a single disruption affects all four simultaneously.

The diagnostic question is not how many suppliers exist but how many independent failure modes the portfolio has. Sources that share an origin, a shipping route, a port, a currency or a regulatory treatment are correlated. Correlated sources should be counted as one for resilience purposes, however many contracts exist.

This is where an honest matrix beats a score. Assigning countries numerical risk scores invites false precision and tends to encode judgements the analysis cannot support. Describing the concrete dependencies instead — which origins, routes and processing steps a product depends on — produces a picture a business can act on.

The size of contingency stock is determined by two variables that interact: how variable the lead time is, and how long the product keeps. Neither alone tells you what to hold.

A stable long lead time can be managed with modest safety stock, because the variability is low and the order can be planned. A short but highly variable lead time demands more cover, because the business cannot predict when replenishment arrives. The relevant measure is therefore lead-time variability, not lead time itself.

Shelf life caps the strategy. Where a product has a short shelf life, holding deep contingency stock is not possible — the cover expires before it can be used, and the holding cost becomes a write-off. For those products, resilience has to come from elsewhere: multiple sources, contractual priority, or reformulation to an ingredient with greater tolerance.

This is also where substitution earns its place. A qualified alternate specification, already tested and approved, converts a disruption from a production stoppage into a changeover. It is usually cheaper to qualify a substitute in advance than to find one during an event.

For domestically produced food, security of supply depends on utilities as much as on ingredients. Controlled-environment production and processing both depend on continuous power and, frequently, on water of a specific quality.

Where a business is considering domestic production, the utility question is not whether a connection exists but what the reliability is, what redundancy is available, and what an interruption costs. For some processes the tolerance for interruption is very low, which makes standby generation or storage a capital requirement rather than an operating preference.

Water deserves separate attention. Availability, quality and treatment requirement vary substantially by location and by process, and the cost of treatment to the required standard belongs in the comparison rather than in a later surprise.

The comparison that decides most sourcing questions is total landed cost to the point of use: purchase price, freight, duty, insurance, handling, and the cost of the inventory carried across the lead time. Domestic production replaces several of those lines while introducing others — energy, labour, utilities and capital recovery.

Comparing a domestic unit cost against a foreign ex-works price is the most common error. It overstates the domestic case by omitting the inventory and logistics the import actually carries, and understates it by assuming the domestic plant will run at the utilisation the model assumes.

The honest comparison runs both options at realistic utilisation, includes inventory cost across the actual lead time, and states what has been assumed about volume and mix. Where the answer is close, the decision usually turns on characteristics the cost comparison cannot capture: responsiveness, lead time, quality consistency and the strategic value of having a domestic source at all.