The Cold-Chain Gap

Executive Summary. Bangladesh's cold-chain problem is often narrated as a shortage of warehouses. FAO's Hand-in-Hand investment plan describes a larger coordination failure. It places post-harvest losses across priority value chains in a 25-50% range and identifies roughly US$737 million in cold-storage investment within a broader US$1.5 billion immediate investment portfolio. For four selected chains, the plan proposes US$410 million for potato, US$90 million for mango, US$75 million for onion, and US$15 million for tomato. The loss estimates are at least 25% for potato, more than 30% for mango, at least 25% for onion, and about 40% for tomato. These are planning estimates, not measurements from one harmonized loss survey, and the plan is not evidence that the projects were completed. Its value is to show why a cold room alone is insufficient. Farmers need aggregation, operators need dependable electricity and skilled staff, lenders need predictable utilization, buyers need quality, and the facility needs throughput beyond a short harvest window. The narrower mechanism is missing chain governance. Bangladesh does not merely lack refrigerated capacity. It lacks enough bankable systems that keep the asset full, the produce traceable, and the operator paid.

The loss sits between harvest and sale

FAO's plan states that cold-storage facilities for fruits and vegetables other than potato are almost nonexistent and that post-harvest losses reach 25-50% across several value chains. It then gives chain-specific planning assumptions: potato loses more than 25%, mango more than 30%, onion more than 25%, and tomato about 40%.

Source: FAO Hand-in-Hand Bangladesh Investment Plan. Bars preserve the source's lower-bound and approximate qualifiers.

Those numbers should not be averaged into a national loss rate. The source does not claim that every crop, district, or season loses the same share, and the estimates use different commodity contexts. Loss can mean physical spoilage, quality deterioration, price collapse, or output that never reaches the intended market. A crate rejected for appearance may still be eaten locally. A tomato sold at distress price is not physically lost, but value has disappeared.

The common mechanism is time. A harvest arrives in a short window. Demand is spread across months. Without temperature control, sorting, packaging, processing, or rapid market information, the seller has little power to wait. The buyer knows the product is deteriorating. The price therefore reflects the farmer's clock, not only consumer demand.

Cold-chain capacity changes that bargaining condition only when it preserves a marketable product. A cold room that stores the wrong variety, loses power, lacks pre-cooling, or releases produce into the same glutted market can move the loss rather than remove it. The technical asset must be joined to standards, contracts, logistics, and buyers.

The investment need is concentrated and specific

The plan places immediate additional investment across major themes at about US$1.5 billion, including US$737 million for cold storage. Its selected value-chain package totals US$590 million: US$410 million for potato, US$90 million for mango, US$75 million for onion, and US$15 million for tomato.

Source: FAO Hand-in-Hand Bangladesh Investment Plan, selected value-chain summary.

The distribution matters. Potato absorbs much of the proposed package because it combines production volume, existing storage experience, and processing opportunity. Within the potato proposal, the plan separates about US$329 million for storage from US$79 million for processing. That split exposes an important choice. Storage postpones sale. Processing changes the product and can create a different market.

For perishables, these functions are complements. Storage can smooth supply and reduce distress selling, but it cannot absorb every grade indefinitely. Processing can turn variable raw output into paste, chips, pulp, juice, or another shelf-stable product, but it needs consistent input and utilization. A chain built around only one function leaves the other bottleneck intact.

The plan proposes cold-storage capacity equal to 25% of annual production in the selected chains. That is a planning target, not proof of optimal capacity. Utilization depends on season, crop geography, energy cost, handling practice, and market demand. Building to a national percentage without site-level throughput analysis would convert a loss problem into stranded equipment.

This is why project appraisal must begin with the chain, not the building. The relevant questions are who commits produce, who grades it, who bears shrinkage, who pays power, who has the right to sell, and what happens when market prices fall. A technically sound facility can still fail if these rights and obligations are vague.

The beneficiaries reveal a market-making project

FAO estimates that the selected potato intervention would reach about 2.9 million people, mango 2.6 million, onion 1.08 million, and tomato 164,000. These are modeled beneficiary estimates attached to a proposed investment package, not counts of people already served.

Source: FAO Hand-in-Hand Bangladesh Investment Plan, selected value-chain summary.

The scale shows that cold-chain investment is not just a private warehouse transaction. A facility changes the market faced by farmers, traders, workers, processors, transporters, and consumers. It can reduce seasonal scarcity, but it can also concentrate buying power. It can reward quality, but it can exclude smallholders who cannot meet a specification. It can lower physical loss while increasing financial risk if farmers borrow for inputs under an uncertain purchase promise.

Public policy therefore has a role even when the asset is privately owned. That role is not necessarily to build and operate the store. It is to create transparent grades, standard contracts, metering, food-safety rules, reliable energy arrangements, and competition among buyers. These institutions determine whether the value of lower loss reaches farmers and consumers or remains with one gatekeeper.

The beneficiary estimates also create a verification duty. A proposal that claims 2.9 million people will benefit must define direct and indirect reach, avoid double counting, and measure actual use after construction. Capacity installed is not the same as produce saved. Produce stored is not the same as farmer income gained. Each link needs its own outcome.

The asset fails when the chain is incomplete

FAO's commodity cases repeatedly identify power risk and shortages of skilled staff. Those are not side notes. Refrigeration is an operating service. The asset must maintain temperature, manage humidity and airflow, monitor product condition, and turn inventory before quality falls. A power interruption can destroy the value the facility was built to protect.

Utilization risk is equally important. A single-crop store may be full for a harvest period and empty for the rest of the year. Debt service and maintenance continue. Multipurpose design can improve utilization, but different products require different temperature, humidity, packaging, and hygiene protocols. Calling a store multipurpose does not make every crop technically compatible.

Finance must therefore underwrite operations, not just construction. A lender needs evidence of supply agreements, buyer contracts, energy cost, management capability, and expected occupancy. A concessional loan cannot rescue a facility with no throughput plan. Conversely, a credible chain may justify blended finance because some benefits, including lower food loss, smoother seasonal supply, and better safety, extend beyond the operator's revenue.

Digital tracking can help, but it is not the chain itself. Sensors can record temperature. They cannot resolve who bears a loss when a buyer rejects a batch. Platforms can display prices. They cannot create a purchase obligation. The institutional work remains contractual.

What would change this conclusion

The conclusion would change if audited operating data showed that stand-alone cold stores consistently achieved high utilization, preserved quality, repaid investment, and passed gains to farmers without coordinated procurement, processing, or buyer contracts. The FAO plan instead treats storage and agro-processing as connected interventions and repeatedly lists chain enablers.

Three moves would test a governed cold-chain model.

1. Require a throughput contract before public finance supports capacity. Every proposal should identify committed suppliers, buyers, crop calendar, power plan, grading rules, and responsibility for loss. Owner: Ministry of Agriculture and participating financiers. Success signal: supported facilities report utilization and verified loss reduction against the source's 25-50% problem range.
2. Procure shared chain services competitively. Metered refrigeration, testing, sorting, and traceability should be open to multiple farmers and buyers under published fees. Owner: agricultural marketing authorities and local project entities. Success signal: the proposed US$590 million selected-chain package produces contestable services rather than closed local monopolies.
3. Publish outcome accounts, not construction accounts. Track tonnes received, quality retained, energy used, prices paid, rejected batches, and beneficiaries reached. Owner: FAO counterpart agencies and project sponsors. Success signal: reported reach can be reconciled with the current 2.9 million, 2.6 million, 1.08 million, and 164,000 planning estimates.

The counterargument

The strongest objection is that FAO's document is an investment pitch. Its loss rates, returns, and beneficiary estimates are designed to identify opportunities, not to provide an audited national baseline. Building policy around approximate figures could produce oversized facilities or optimistic social returns.

That objection is correct and should govern implementation. The plan's numbers must not be converted into guaranteed outcomes. The loss bars preserve qualifiers. The beneficiary chart labels estimates as estimates. The investment totals describe proposals, not commitments or completed assets.

But uncertainty about the exact national loss rate does not justify another uncoordinated construction program. It strengthens the case for staged investment tied to observed throughput and verified performance. A small facility with real contracts can generate the evidence for expansion. A large facility built from a national ratio cannot.

Bangladesh's possibility is not cold storage as an object. It is cold-chain service as a governed market: power stays on, quality is measured, contracts allocate risk, and produce reaches a buyer before time erases its value.

---

Data source: FAO Hand-in-Hand Bangladesh country material and investment plan, retrieved on the publication date. Approximate and lower-bound values retain their original qualification. No value is presented as completed investment or observed project impact.

Sources

1. FAO Hand-in-Hand, Bangladesh country page: https://www.fao.org/hand-in-hand/previous-editions/hih-IF-2023/bangladesh/en
2. FAO Hand-in-Hand, Bangladesh Investment Plan: https://www.fao.org/media/docs/handinhandlibraries/default-document-library/bangladesh_investmentplan_221006.pdf?sfvrsn=c862c940_1