Product costing in Uganda's small manufacturing firms

A metal fabricator in Katwe is asked to quote for forty school desks. He works it out on the back of a delivery note: so many lengths of angle iron at the price he paid in March, a tin of paint, something for the welder, something for the boy who sands. He leaves out the yard rent, the power, the grinding discs he will burn through, the two desks that will come back rejected, and four days of his own time. He names a price, the customer agrees to it without arguing, and he takes that as good news. It's the only signal he will ever get, and it is pointing the wrong way.

Multiply him by thirty-eight thousand. That is roughly the number of manufacturing businesses the Uganda Bureau of Statistics counted in its 2019/20 Census of Business Establishments, employing 226,456 people between them at an average of six people per business. The bakery in Mbale pricing bread against the loaf next door. The grain mill in Lira charging by the sack. The tailoring workshop in a Kampala arcade, the furniture yard in Ndeeba, the plastics moulder on the Jinja road. More than nine in ten, 90.7 per cent, are sole proprietorships, so the man quoting is the man buying the inputs and absorbing the loss.

Uganda measures this sector well from above, and that deserves saying before anything critical. The Bureau's Index of Industrial Production, rebased to 2016/17, reports manufacturing output quarterly, and recorded production up 3.2 per cent in the year ended the second quarter of 2025/26. Its Uganda Business Inquiry for 2019/20 put manufacturing's gross output at UGX 19.2 trillion, the largest of any sector, against value added of UGX 10.8 trillion. The country can say roughly what its factories produce and what they consume in producing it.

WHAT THE NATIONAL FIGURES ARE BUILT FROM

The difficulty is in how that picture is assembled. The Bureau's methodology note for the index states that manufacturing sales data comes from monthly VAT returns, validated against a selection of establishments, and that the establishments covered contribute about 80 per cent of total value added in their sectors. Nothing in that is improper; it is the standard way to build a short-term production index. It does mean the statistic is assembled out of firms already keeping records good enough to file with. The Bureau's own census classifies 69.4 per cent of other manufacturing businesses and 84.5 per cent of food processors as informal, and their costs are imputed rather than measured, because in those workshops there is nothing to measure.

THE RECORD STOPS AT THE CASH BOX

What the census did ask, plainly, was whether a business keeps books of accounts at all, and the answer is among the most uncomfortable numbers in Ugandan business statistics. Not one micro-business kept them. Among small businesses 5.8 per cent did, and among medium-sized ones 42.6 per cent. Every large business did, which is rather the point: the record appears at the size where somebody else — a bank, an auditor, a revenue office — starts asking for it, and very rarely before.

Manufacturing is no exception. The census found 43.2 per cent of manufacturing businesses owning a computer, an improvement on the 5 per cent recorded in 2010/11 and still a minority. An owner who cannot state a cost per unit is not being careless; he is being rational about a working day in which nothing produces that number. The timber was bought in cash at three different prices, the welder was paid for the week rather than the job, and nothing in that sequence asks anybody to record which desk consumed what.

A PRICE SET WITHOUT A FLOOR

Three costs follow, and the first is the price. A firm that cannot compute a cost floor cannot know whether an order is worth taking, so it prices by reading the customer instead: the size of the order, how badly the buyer wants it, whether he arrived in a car. Bad orders aren't visible as losses; they surface much later, as a year of hard work that ended with nothing in the account. Waste hides for the same reason: scrap, rework and an idle machine only become expensive once something converts them into shillings per unit. The third cost is credit. The World Bank's 2025 Enterprise Survey of Uganda found 26.0 per cent of firms holding a bank loan or line of credit, 6.0 per cent fully credit constrained and a further 34.0 per cent partially so. In 2013 the same survey found 86.7 per cent of loans requiring collateral averaging 161.8 per cent of the loan. Lending here is secured against land largely because it cannot be secured against a cash flow nobody can evidence.

TANZANIA WENT AND LOOKED

The nearest careful measurement of this problem anywhere in the region is Tanzanian, and Uganda has no equivalent, which is itself part of the argument. Bigambo, Marandu and Elias, writing in the Tanzania Journal of Engineering and Technology in 2023, interviewed twenty metalwork enterprises in Dar es Salaam about how they arrive at a selling price. Eighty-five per cent charged different prices for the same product depending on the size of the order and the importance of the customer, and while margins ranged from 10 to 40 per cent, not one firm could state a basis for the margin it used. The researchers collected fifteen distinct costing formulae across those twenty workshops: some added materials, labour and a margin but no machine charge, another multiplied materials and inputs by 1.4 without specifying the inputs. Seventy per cent of the owners had learned costing from experience rather than instruction, and 85 per cent wanted training. The study then built a costing system and tested it on those firms, and that's the step nobody here has taken.

WHAT IS THE WAY FORWARD?

Several things would move this. The Bureau should put the costing question into the next business census — whether a firm can state a cost per unit, and how it arrives at one — because the sector is presently managed against an output index and a books-of-accounts tick box, and neither answers it. The Ministry of Trade, Industry and Cooperatives should treat costing as an industrial skill rather than a bookkeeping one, and fund it through institutions that already reach workshops. The Uganda Manufacturers Association is the natural home for costing templates by trade, since a bakery's formula is not a fabrication yard's. Lenders should learn to read twelve months of costing records as evidence, which would reach small manufacturers faster than another guarantee fund would. Providers building for this market should capture cost at the job rather than at the month, work offline, and price in shillings. Owners can begin with one product and one month.

The fabricator will deliver the forty desks. He will be paid, the money will cover the angle iron and the welder and leave something over, and he will quote the next order the same way. What he will never learn, because nothing in that workshop is built to tell him, is whether the job just finished paid him for his time or charged him for the privilege of doing it.