What payroll records cannot tell an employer

It is the twenty-sixth of the month in an office above a hardware shop in Nakawa, and the woman who keeps both the accounts and the personnel files is closing the payroll. Before she locks up she can state to the shilling what thirty-eight people will be paid on Friday, what the Uganda Revenue Authority gets as PAYE, and what the National Social Security Fund is owed by the fifteenth. The managing director, on his way out, asks how many staff the firm lost this year and why. All she has to answer with is a payroll file, a drawer of appointment letters, and her own memory.

Multiply that Friday across the country. The hotel in Fort Portal that knows its wage bill exactly and cannot say how many housekeepers have lasted a year. The contractor running four sites off four casual registers nobody has added up. None of them is careless.

The payment record deserves credit first, because it is neither small nor universal. Uganda's employers moved UGX 2.13 trillion into the National Social Security Fund in the year to June 2025, up from UGX 1.93 trillion, and 14,000 registered with the Fund for the first time, according to its 2025 Annual Integrated Report. In each of those firms somebody computed a deduction and filed it, monthly, on time.

WHAT THE PAYROLL REGISTER CANNOT ANSWER

The arithmetic and the record are not the same thing, and the gap is wider than most managers assume. The National Labour Force Survey 2021 found only 33 per cent of employed Ugandans working under a written agreement, the rest on an oral one. Of people in paid employment, 76 per cent had an employer contributing to no social protection scheme, and a further 4 per cent did not know whether theirs contributed. Twenty-seven per cent received any workplace benefit — paid leave, sick leave, maternity leave, a pension or medical cover. Those are national figures, but where there is no written agreement there is usually no start date on file, no job title that means anything, and no note of why anybody left.

THE RECORD THE LAW ALREADY ASSUMES EXISTS

The law was drafted on the assumption that the record exists. Section 19 of the Employment Act, 2006 allows the Commissioner for Labour to require any employer to furnish, in writing, returns and statistics of the number of employees, the rates of remuneration and any other conditions of service, and the Ministry of Gender, Labour and Social Development has invoked it to ask every employer for those returns, down to each affected worker's name, salary and national identification number. Failing that is a data problem before it is a willingness problem, and the National Social Security Fund reported employer compliance of 52 per cent at 30 June 2025, down from 57 per cent.

The Uganda Bureau of Statistics went and counted inside the registered firms. Its Labour Market Survey 2025 covered 2,635 formal establishments employing 2,373,109 people, of whom 1,028,630 held formal jobs and 1,344,479 — the majority — held informal ones: no social security contribution, no written terms, no paid leave, inside businesses registered and paying tax.

RESIGNATIONS, NOT GROWTH, ARE DRIVING THE HIRING

The same survey asked those establishments why their vacancies were open. Of 155,240 vacancies in Uganda's formal businesses, 35.9 per cent — 55,684 posts — existed because somebody had resigned, against 30.0 per cent opened by expansion, and a further 46,217 jobs were lost over the same period. More hiring was being done to replace people than to add them.

A firm keeping only a payroll sees none of that in its own numbers, because a leaver stops appearing and a joiner starts, with no departure date, no reason and no length of service to set against last year or the branch down the road. Turnover is the most expensive thing most Ugandan employers are not measuring.

THE COST OF GUESSING AT SKILLS

The second cost lands on hiring. Of those 155,240 vacancies, 43,900 went unfilled, for two reasons in almost equal measure: inadequate financial resources, at 43.6 per cent, and no workers available with the required skills, at 43.0 per cent. The training that would close the second gap sits with the largest firms: the World Bank's Enterprise Survey of 605 Ugandan firms in 2025 found 24.5 per cent offering formal training, from 17.2 per cent of small firms to 50.8 per cent of medium ones and 72.0 per cent of the large.

Uganda has done this once, well. Spero, McQuide and Matte, writing in Human Resources for Health in 2011, described the Uganda Nurses and Midwives Council replacing lost and outdated paper files with a human resources information system covering the 26,046 people who had entered nursing or midwifery training. A senior official told them: "I used to feel guilty when requested to talk about the total number of qualified nurses and midwives in the country because I knew that we did not have accurate data." Once it existed the record could be asked questions, and one answer was a 2009 public-hospital nursing vacancy rate of 53 per cent.

RWANDA COUNTS ITS WORKFORCE FOUR TIMES A YEAR

Rwanda treated workforce data as infrastructure and funded it accordingly. The National Institute of Statistics of Rwanda runs its Labour Force Survey quarterly, collecting in February, May, August and November, and publishes an annual report with a continuous series back to 2017. Its 2024 report states that 68.8 per cent of Rwandan employees were on short-term or casual contracts that year, down 3.3 points on 2023, and 25.8 per cent permanent, up 1.7 points. Those are not flattering numbers, but they move, and a ministry can act on a figure that moves.

Rwanda then built the demand side to match, its Ministry of Public Service and Labour launching a labour market information system and a graduate tracer survey in August 2024, and its National Employment and Skills Strategy for 2024 to 2029 committing to a planning and reporting module on the finding that job creation was being held back by reporting delays. Uganda is moving the same way: its National Labour Force Survey ran every five years until 2021, and the survey replacing it will publish quarterly.

WHAT IS THE WAY FORWARD?

Several things would move this. The Bureau of Statistics should hold the quarterly cadence it has promised, since a figure two years old cannot inform this month's hiring decision. The Ministry of Gender, Labour and Social Development should make the section 19 labour return a routine electronic filing and publish what it collects, so that an employer who files gets a sector benchmark back and not only an obligation. The National Social Security Fund tracks contributions member by member — its compliance measure is the share of active members contributing consistently — which makes it the holder of the country's most complete formal workforce dataset, and it should be resourced to publish it in aggregate. Employers can start with four fields that cost nothing: start date, role, leaving date, reason.

None of this requires a large system. What the woman above the hardware shop needed on Monday morning was four columns nobody had given her a reason to keep, and a payroll that wrote to them each time it ran.