Rooms, bar and kitchen: the guesthouse accounts problem

The owner of a fourteen-room guesthouse in Masindi does the accounts after the last guest has gone up. Three books lie on the table and they do not speak to one another. Reception keeps the room register, a ruled exercise book with names and what was paid. The bar keeps a tally sheet, written up by whoever closed it, and the kitchen writes orders on a pad. Against these she has a cash tin, a phone showing the day's mobile money, and two guests who said they would settle in the morning. The three books produce three figures, and she writes down the one that balances.

Multiply her across the country and the first thing that becomes clear is that nobody can say how many of her there are. The Uganda Tourism Satellite Account Report, published by the Uganda Bureau of Statistics and the Ministry of Tourism, Wildlife and Antiquities in March 2025, puts short-term accommodation establishments at 22,616, and states plainly that the figure has not moved since 2019 for want of up-to-date data. Room and bed capacity, at 350,550 and 371,221, are carried forward the same way, leaving the official count of the country's guesthouses six years old and standing still while the sector it describes grows underneath it.

What sits on top of that base is a genuine success and it deserves to be said plainly first. The Ministry's Statistical Abstract published in April 2026 records foreign exchange earnings from tourism of USD 1.62 billion in 2025, against USD 1.025 billion in 2023, on arrivals of 1,642,215 and an average stay of 8.8 nights. Accommodation is also the largest single thing those visitors buy, at 37.0 per cent of inbound tourism expenditure in 2023, with food and beverage services taking a further 20.4 per cent. Rooms, bar and kitchen — the three books in Masindi — take well over half of every shilling a foreign visitor spends here, which is why what happens to them at closing is not a bookkeeping detail.

A SECTOR THE STATE CANNOT SEE

Against 22,616 establishments, the Uganda Tourism Board had graded and classified 117 accommodation facilities by the end of 2025, 77 of them town hotels and 23 safari lodges. The Abstract is explicit that grading happens on request from interested facilities, a voluntary act by operators who already see something in it. Sixty-five per cent of graded facilities sit in the Central region, Kampala and Wakiso alone hold 66, and the whole Eastern region contains three. Everything outside that list is invisible to the institutions planning for the sector, and invisible more immediately to the owner who has to decide tonight whether the day made money.

THE LICENCE EXPLAINS THE LEDGER

The separate books are not carelessness. They are the shape the state itself imposed. The World Bank's Uganda Economic Update of June 2023 reports that industry operators count as many as 25 licences and taxes on tourism firms, and that an accommodation provider must hold a separate licence for each amenity it runs. The licensing regime meets a guesthouse not as one business but as four, so keeping four sets of records to answer four sets of demands is the rational response. The same report notes the Tourism Board is responsible for inspecting accommodation but lacks the capacity, so the task often passes to district tourism officers with limited training.

Leakage happens in the gaps between those records, and it is mostly ordinary rather than dramatic. A booking is taken by phone and written on one page, the guest checks in and is written on another, and drinks are charged to the room at a bar whose tally sheet is closed by a different person at a different hour, before the bill is settled partly in cash and partly by mobile money. Each handover is a point where a number is copied by hand, and every copied number is somewhere the total can quietly change without anybody having decided to steal a shilling.

WHAT A LEGIBLE OPERATION IS WORTH

The best measurement of what changes when those handovers stop being manual comes from outside the region, and it matters precisely because it is not a vendor's claim. Pierce, Snow and McAfee, writing in Management Science in 2015, studied 392 restaurant locations that installed theft-monitoring software on their point-of-sale systems, using staggered adoption dates to isolate the effect. They found reduced theft and improved productivity, driven primarily by changed worker behaviour rather than staff turnover, and concluded the gains benefited both the firms and the legitimate tip-based earnings of the workers. Catching theft recovered far less than it was worth to run an operation that could be read at a glance.

For a Ugandan guesthouse there is a second prize and it may matter more. The same update observes that the sector is largely small and medium enterprises with limited collateral, that land is the security lenders demand while only about 21 per cent of land in Uganda holds title, and that banks treat tourism as high risk and price it accordingly. An owner who cannot pledge a title can pledge a record instead, but only where one exists, and three exercise books that disagree are not a record any credit officer will lend against.

RWANDA MADE EVERY SALE LEAVE A TRACE

Rwanda ran this at national scale, and a peer-reviewed study documents the result. Kotsogiannis, Salvadori, Karangwa and Murasi, publishing in the Journal of Development Economics in 2025 and working from the universe of Rwandan tax filings between 2012 and 2019, describe a staggered rollout of electronic invoicing that took large firms and selected sectors first. In January 2014 the revenue authority set a deadline of that March for every VAT-registered firm, and by September 2014 more than 3,943 firms were issuing electronic invoices, 77.8 per cent of those registered, and that first phase raised VAT payments by 6.5 per cent between March 2013 and September 2014.

Two findings there deserve more attention than the headline. The first is that the compliance gain came from audits becoming more efficient rather than from firms turning honest once they held a machine, in a sample where audits uncovered underreporting worth about 28 per cent of the potential tax base audited. A machine record does not manufacture integrity; it makes a claim checkable, which is the service it also performs for an owner who was not there when the bar closed. The second is cost, since the same paper records that the upfront installation outlay took four full years of VAT payments for a small firm to recover, which is why an obligation designed around large hotels will never reach a fourteen-room guesthouse.

WHAT IS THE WAY FORWARD?

Several parties hold a piece of this. The Bureau of Statistics and the Ministry should fund a current count of accommodation establishments, because a 2019 figure carried forward six years running cannot carry a plan. The Tourism Board should pair voluntary star grading with a registration tier a rural lodge can complete, and resource inspection rather than devolve it to untrained officers. The revenue authority should read Rwanda's cost finding before any record-keeping duty reaches the smallest operators. Lenders and the trade associations should settle what an operating record must contain to count as evidence of turnover, so an owner without land title has something to bring. Providers should treat one reconciled ledger across rooms, bar and kitchen as the whole product, priced in shillings and working when the power drops, rather than selling three modules to a guesthouse that already owns three books.

None of this is about catching staff. It is about the moment at the end of the night in Masindi when three figures disagree, and the only way to close the day is to pick one of them and hope.