From Statute to Shillings

Uganda changes its tax law on a yearly cycle. Each set of amending Acts is passed with the budget, assented to by the President, printed in the Uganda Gazette and put in force with the new financial year. The exceptions: the Excise Duty (Amendment) Act 2020 (Act 21 of 2020), in force from 27 November 2020; the Excise Duty (Amendment) (No. 2) Act 2023 (Act 24 of 2023), in force from 25 August 2023; the Stamp Duty (Amendment) Act 2023 (Act 21 of 2023), in force from 25 August 2023. In the 23 amending Acts read for this register, assent came -141 to 44 days before the Act took effect, and publication in the Gazette only -182 to 26 days before.

The principal Acts were not always so orderly. The Tax Procedures Code Act was assented to on 19 October 2014 but came into force on 1 July 2016, 621 days later. The Excise Duty Act ran the other way: it took effect on 1 July 2014, 110 days before the President assented to it on 19 October 2014.

The calendar below marks each assent, gazette publication and commencement as a separate day: 105 dated events in all, each read in a statutory text or an official report. Twelve records still wait for their statutory text and sit in a column of their own rather than in a month they do not have. Switch to the timeline to filter the records by Act, tax head, type and evidence, and to follow the threads from each change to the later ones that amended or reversed it.

57 of the 91 changes were meant to raise revenue and 21 to lower it; for the remaining 13, no source read states the intended direction. Revenue was not always the point. Exemptions, zero-ratings and reliefs give revenue up to serve health, investment or fairness, so the register's assessments weigh each change on criteria beyond the shillings it raised: clarity, cost of compliance, equity and the quality of its implementation.

14 records carry a stated reason. Eight are quoted from an official source that ties the reason to the measure; for the other six, the link between the stated aim and the measure is the register's reading, and is marked as interpretation.

Net URA collections rose from UGX 11,231 billion in FY2015/16 to UGX 31,634 billion in FY2024/25. In FY2024/25 prices that is UGX 16,041 billion to UGX 31,634 billion, a real rise of 97.2 percent. As a share of GDP, net collections moved from 11.17 percent in FY2015/16 to 13.88 percent in FY2024/25. In FY2019/20, the first year of COVID-19, nominal collections grew 0.81 percent.

These lines are the backdrop, not the verdict. Growth and prices, the exchange rate, imports, registration drives, arrears recovered, refunds paid and digital administration all move them at once. Where a year of unusual movement coincides with a change in the register, the chart notes the coincidence and claims nothing more.

Each budget forecasts what its tax measures will raise, and URA's annual revenue reports later say what they yielded. Across FY2015/16 to FY2026/27, 13 single measures have both figures. Six came in below the forecast, and the median measure delivered 112.3 percent of it. The spread is wide: excise duty of UGX 200 per litre of cooking oil reached 1,328 percent of its forecast in FY2018/19, while electronic Fiscal Devices reached 0 percent in FY2019/20.

By tax head, 17 of 31 comparisons fell short. 20 measures with a forecast never reappear with a yield in shillings; the chart draws them as missing, not as zero. The yardstick also moves: URA's outlook for FY2017/18 put the year's policy measures at UGX 184 billion, but its report for that year measured them against UGX 173 billion; URA's outlook for FY2019/20 put the year's policy measures at UGX 62 billion, but its report for that year measured them against UGX 99 billion.

A reported yield is URA's own attribution, not a causal estimate. Where URA's printed rate disagrees with its own figures (18 rows), the chart flags the row and the case file shows both.

Ask what a single change to the tax law brought in, and the public record seldom answers. Collections move for many reasons at once; budgets forecast whole packages; URA attributes revenue to some measures and not to others; and only rarely does a study separate one change from everything else around it.

So the register climbs a ladder of evidence for each change and stops where the evidence stops. The first rung is what collections did. Above it sit what the Ministry of Finance forecast, what URA later attributed to the measure, and, at the top, a study that isolates the change's own effect. Every record shows the rungs it reaches; the rungs it does not reach are drawn as missing, never filled.

The bookcase shelves each change on the highest rung it reaches. The upper shelves are nearly bare, and that is the finding. Where no rung gives a figure, the case file says so in the same words every time: Amendment-level revenue contribution not reliably quantified, followed by what is known and what is not.

A tax change does not end at commencement. Taxpayers took the mobile money levy, the digital tax stamps, electronic invoicing and the daily charge on over-the-top services to the High Court or the Tax Appeals Tribunal. Asked for interim injunctions, the High Court refused to halt the stamps and e-invoicing, holding that the public interest in collecting revenue outweighed the objecting businesses' private interests; those rulings did not decide the challenges themselves. The Tax Appeals Tribunal ruled against URA on how it applied the over-the-top charge, setting aside an assessment of UGX 33.9 billion because the Act taxed access per user per day but set no technical test of what counted as access.

All four decisions have been read in full; the register cites each holding to the judgment and links it to the change it concerns.

Each change has a case file of five papers: the law as amended, the stated reason, the revenue evidence, the consequences, and the sources. A paper is inked where the register holds evidence for it and left blank where it does not. Of the 91 records, 79 rest on wording or dates read in a statutory text and twelve on official reports alone, and say so. Every record cites at least one source confirmed on its page by the source audit.

Before the register era, the record is kept at the level of whole Acts. The long view holds 93 milestones in nine families of law, from 1987 to FY2013/14: 48 amending Acts, 22 Finance Acts and 13 statutory instruments, beside the principal Acts and the authority that collects the tax. 30 are read in the Acts on file and 63 come from ULII's catalogue of legislation; two have only a verified year.

For these years the register makes no claim about what a section said or what a change raised. It records that a law was made, under what number and on what date. Choose The long view, since 1986 on the timeline above to see them in place.

Already in force when the long view opens, and listed by ULII as since repealed: Excise Tariff Act (an Ordinance of 1953) (Ordinance 18 of 1953); Customs Tariff Act (Act 17 of 1970); Stamps Act (an Ordinance of 1915) (Ordinance 11 of 1915).