Supported by the GIZ Promotion of Rule of Law and Judiciary in Africa (PRoLA) programme
“An indispensable part of Tanzania’s justice system.” For €159,000 a year.
That is the verdict of an independent evaluation of TanzLII — commissioned by Laws.Africa, funded by GIZ, and written by evaluators who were free to be unflattering. It set out to measure whether free access to law changes how a justice system behaves, not whether its users like it. It does. This page is the short version, with the costs, the caveats and the unfinished work left in.
“At the closing event of the programme, a law student living with visual impairment said that TanzLII had made his studies possible. That does not show up on a usage graph, but it is exactly what this work is meant to achieve.”
Muhamet Brahimi, Implementation Manager, Promotion of the Rule of Law and Judiciary in Africa, GIZ — in the foreword to Justice in DataStart with the money. Infrastructure has to justify itself on cost per result.
The entire documented investment in TanzLII between 2016 and the end of 2025 was €538,200. The intensive build years of 2023 to 2025 account for €477,200 of that — roughly €159,000 a year. The forward budget in the adopted 2026–2030 Strategic Plan is only marginally higher, at about €182,000 a year.
Cost per judgment published
4.4× moreOn the evaluation’s stricter annualised measure — counting only the judgments published in the most recent year — the figures are €16 against €63.
Cost per visit
5× moreAnnual platform spend
11× moreThe comparator’s estimated annual digital-platform spend is close to four times TanzLII’s entire recorded investment since 2016, and around thirteen times its average annual spend during the intensive 2023–25 build.
Show the underlying figuresHide the underlying figures
| Measure | TanzLII | Standalone comparator |
|---|---|---|
| Recorded cost per year | €159,000 (2023–25) €182,000 (projected) | €2,500,000 all-in €2,000,000 digital est. |
| Judgments published | ~110,000 | ~275,000 |
| Annual visits | ~1,800,000 | ~3,960,000 |
| Judgments published in most recent year | 11,400 | 31,785 |
| Cost per judgment published | €1.45–1.65 | €7.27 |
| Cost per visit | ~€0.10 | ~€0.51 |
What these numbers do not include
The evaluation is explicit about the limits of this comparison, and so are we. TanzLII’s figure is programme cash outlay only. It excludes the government-borne recurrent costs carried in kind — roughly four Judiciary staff with dedicated responsibility for uploads, around 100 court staff who upload as part of their existing duties, and office space inside the court buildings. The comparator’s €2.5 million is an all-in annual government budget. TanzLII’s judgment throughput and its cost are both inflated by a one-off historical digitisation phase. The comparator is also substantially larger, longer established, and editorially reports and headnotes its cases — real value TanzLII does not add.
Taken together, and allowing generously for the uncounted government contribution, the evaluation concludes the gap is large enough to be robust: TanzLII delivers its output at somewhere between a fifth and a tenth of the comparator’s cost, on gross spend, cost per judgment and cost per visit alike.
The comparison in plain terms
The comparator’s estimated annual digital spend is around four times TanzLII’s entire recorded investment since 2016 — and about thirteen times TanzLII’s average annual spend in the intensive build years.
Why it is this cheap
Two structural reasons, both replicable. It runs on shared regional infrastructure rather than funding a standalone institution, and the law was digitised to open, structured standards, so the same corpus now carries new services at low marginal cost.
Not a Tanzanian fluke
ULII in Uganda runs the same structural model — a Judiciary-employed team uploading from within the courts on the same shared platform — and serves 33,991 judgments to about 1.59 million visits a year. The low-cost approach is regional, not unique.
Not satisfaction scores. A measurable change in how the courts work.
The existing literature already established that users value legal information institutes. This evaluation asked the harder question — whether a platform changes how a justice system behaves — and answered it with a near-complete national corpus of citation data, a sector-wide survey, key informant interviews at every level of the courts, and a dated registry of every intervention made since the platform’s founding.
The shift begins in 2023 and accelerates through 2026, tracking full court integration, the Chief Justice’s 2023 same-day publication circular, and the March 2024 amendment to the Court of Appeal Rules recognising decisions on the official online platform as reported decisions. Neutral citations exist only for judgments from 2011 onward, so part of the rise reflects the growing share of recent cases in circulation — but not the steepness of the change after 2023.
Show the underlying figuresHide the underlying figures
| Authority | Before 2018 | After 2018 | Change |
|---|---|---|---|
| Dickson Nsamba Shapwata (2008) | 0.91 | 15.36 | ~17× |
| Nyerere Nyague (2012) | 1.26 | 12.44 | ~10× |
| Paulo Maduka & Others (2009) | 2.80 | 10.00 | ~3.6× |
| Lyamuya Construction (2011) — most-cited case | 12 | 88 | ~7× |
Figures are annual citation slopes. Citation counts before 2018 are shaped by the coverage of the corpus in those years as well as by behaviour — the evaluation flags this as a limit on the pre-launch baseline.
Of judges and magistrates now cite TanzLII decisions with the same confidence as printed law reports. 62% of all users do.
Of judgments cited at least one authority by 2025. Citation has become a near-universal feature of recent decisions.
Say legal reasoning and argumentation is noticeably more consistent — rising to over three-quarters of judges and magistrates.
Average authorities cited per judgment, up over the long run. The rise holds even after controlling for the growth of the corpus.
“I had drafted a judgment on the old position regarding proving breach of employment contract. Just before completing my draft, I found a current decision of the Court of Appeal which changed my position.”
Survey respondent, Justice in Data evaluationThree changes the interventions caused
Each runs on a short, traceable line from a dated intervention to an observed change, with the quantitative and qualitative evidence converging on the same timing and mechanism.
A step-change in the availability of Tanzanian law
The repository grew to more than 110,000 judgments, of which roughly 46,000 are superior-court decisions delivered since January 2020, alongside legislation and gazettes. Average judgments published per year rose about 323% in the 2020s and exceeded 11,000 from 2023.
Causal standard: strongest. The digitisation, the hardware and the legislative consolidation were themselves the funded interventions, so the growth in availability is caused by them rather than merely correlated with them.
A step-change in citation behaviour
Two dated interventions, each producing an observable step in a separate and independent series. The 2018 launch shows in the authority-level citation rate, where each case acts as its own control. The 2023–24 institutional measures show in the neutral-citation share, which moves from under 5% to 69%.
Causal standard: strong. Convergent quantitative and qualitative evidence, temporal precedence, and a specified mechanism.
A change in day-to-day legal practice
Users find authority faster, rely on the platform, and alter the substance of their advice and submissions. About 93% report saving time, 67% significant time savings measured in hours per research task, and judicial officers describe changing a drafted position on finding a current Court of Appeal decision.
Causal standard: strong. Direct, demand-side verification across survey and interview channels.
Four systemic outcomes they helped produce
These are the changes a justice-sector investor is ultimately buying. They are shaped by judicial reform, training and case-management change as well as by the platform, so the evaluation claims a contribution rather than a cause — deliberately, and says why in each case.
Contributed to a more transparent judiciary
Full court integration since 2023 and Chief Justice Circular No. 1 of 2023 require judgments to be typed, proofread and published the same day, with compliance monitoring. 86% of users rate publication timeliness good or excellent, rising to 91% among judges and magistrates.
Why contribution, not attribution: the evidence lacks a measure of downstream scrutiny — whether greater openness is actively used to hold courts to account.
Contributed to more consistent and predictable adjudication
68% of respondents, and over three-quarters of judges and magistrates, consider that legal reasoning and argument have become noticeably more consistent.
Why contribution, not attribution: consistency of reasoning is shaped by judicial training, case-management reform and the composition of the bench as well as by access to precedent.
Contributed to firmer legal certainty
Lower courts can align rulings with established superior-court principle; the Court of Appeal’s judgments are cited 6.86 times on average against 0.23–0.66 for all other courts, and controlling authority now propagates down into the tier that applies it day to day.
Why contribution, not attribution: a higher-order outcome shaped by many forces beyond the platform, with no available counterfactual.
Contributed to widening access beyond the professional elite
Reach has extended to students, paralegals and, through intermediaries, underserved communities — but unevenly, and dependent on the intermediary layer.
Why contribution, not attribution: on the present evidence this is the least-realised part of the theory of change, and the survey channels cannot reliably reach people who have never encountered the platform.
Read the two columns together. The left-hand column is what the money directly bought. The right-hand column is what that produced in the justice system — more transparent courts, more consistent adjudication, firmer legal certainty, wider access. No credible evaluation can isolate those from every other reform running at the same time, and this one does not pretend to. That restraint is what makes the whole set worth relying on.
The cheapest line in the budget is also the highest-return one.
Availability does not produce use on its own. The evaluation tested this directly, comparing trained and untrained users on outcome measures rather than on how the training felt. Trained users reach the top of every scale more often — and 95% of those trained said it changed how they use the platform.
It is already institutional, not consultant-led
Where training has been delivered, the organisers were universities and law faculties (29%), the Institute of Judicial Administration (17%), Laws.Africa and AfricanLII directly (11%), the GIZ PRoLA programme (10%) and the Tanganyika Law Society (6%). The delivery capacity already sits inside Tanzanian institutions.
Authority rests where the law is made
The Judiciary owns the case law; the Office of the Attorney General and Chief Parliamentary Draftsman took responsibility for legislation in 2023. Only the technical layer is shared. Coherence is designed in rather than negotiated later.
A national archive, handed over
A scanning team worked through the paper archive from May to August 2024. The Tanzanian government left the September 2024 Sustainability Conference with more than 650,000 digitised pages — the Gazette, and judgments of the Court of Appeal and the High Court.
The digitisation was paid for once. It keeps returning.
Because the law was digitised to open, structured, machine-readable standards, the corpus behaves like a data foundation rather than a website. Each new service built on it is a further return on the original investment rather than a new cost — the dividend characteristic of infrastructure spending, which accrues over time and in forms nobody specified at the outset.
Services already built on the same corpus
An automated listing of statutory provisions the superior courts have declared unconstitutional, each linked to the judgment and to any subsequent parliamentary response. Listings of provisions enacted but not yet in force. A machine-generated glossary of defined legislative terms, served through an open API. Each is impossible without interlinked, structured judgment-and-legislation data — and each cost a fraction of the original build.
The trust layer in an AI-mediated legal system
The UK’s national legislation platform receives on the order of a billion bot requests a month from AI systems. One UK study found that large language model accuracy on questions about welfare schemes falls by 42% when the underlying public data is removed. Authoritative, timely, structured legal data is exactly what these systems require — which makes a well-built LII the layer against which AI-generated legal output is verified. Around 60% of TanzLII users have already tried its AI-assisted research tools.
The same logic reaches people a citation count never will. With programme support, TanzLII was upgraded to a world-class accessibility standard — and that was not a trade-off between reach and cost, because digital text that works with a screen reader is both more useful and far cheaper to produce than converting material into Braille. Structured data does not only lower the cost of the next service. It lowers the cost of including the next reader.
“Digitising the law is the costly ‘heavy lifting’ that must be done before private innovators can build upon it.”
Justice entrepreneur, South Africa — five-country study for GIZ PRoLAOpen legal data is precondition infrastructure. This is what it precedes.
A parallel five-country study — Tanzania, Kenya, South Africa, France and the United Kingdom, drawing on more than 50 expert interviews and a validation workshop in Dar es Salaam — traced the pathways from open-access legal information to economic and governance outcomes. It found a moderately strong positive correlation between open digital access to law and GDP per capita across 33 African countries, and was careful to say that correlation is not causation.
01Access to justice and early self-resolution
When laws and procedures are freely available, people can identify the provision that applies, understand which forum has jurisdiction, and resolve a problem before it becomes litigation. Between October 2024 and August 2025, 637,269 visits to TanzLII originated from Google searches alone, and fewer than 1% of those users were logged in — consistent with a large population of infrequent, non-professional users arriving from a search engine while trying to solve a problem.
The qualification matters: raw legal text is dense, and in Tanzania 90% of laws are in English while more than 80% of the population speaks Swahili. Availability is necessary but not sufficient.
02Micro, small and medium enterprises
Open legal information lowers the cost of setting up and doing business, reduces compliance burdens, improves contract security, drives formalisation, and protects against extortion. Small firms cannot carry commercial database subscriptions and are rarely repeat users of specialised legal information, so the free layer is the only layer they have.
Where regulatory information is scattered across ministries, regulators and revenue authorities, even a well-resourced business association reports reconstructing the picture piece by piece — and monitoring ministerial speeches to learn of regulatory change.
03Foreign direct investment and legal certainty
Reliable, up-to-date access to legislation and case law signals legal certainty, transparency and good governance — the conditions investors examine before committing capital. It enables remote due diligence and cross-jurisdictional comparison, clarifies setup costs and timelines, and helps investors anticipate how courts handle disputes and enforce judgments.
The effect is strongest for mid-sized investors and start-ups. Large multinationals can afford specialised counsel and negotiate terms directly; it is the smaller entrant for whom legal opacity is a deterrent.
04Judicial efficiency and the quality of adjudication
When judges and magistrates in lower courts have reliable access to superior-court precedent, they can align rulings with established principle, reducing appeals and easing backlogs. A senior judicial interviewee put the share of Tanzanian cases filed in subordinate rather than high courts at around 70%.
Without accessible repositories, judicial officers risk ruling on obsolete or incomplete sources. Participants in the validation workshop were blunt about the counterfactual: if these portals shut down, court operations would immediately slow.
05Gender equality
Three mechanisms. Women are direct users and beneficiaries. They are indirect beneficiaries through paralegals and intermediaries — cross-border traders, for instance, who are vulnerable to extortion by officials trading on their ignorance of tariff exemptions. And practising women lawyers gain professional ground as digital access dismantles gatekeeping.
A participant in the validation workshop called it “breaking the boys’ club”: where regulatory updates were shared in male-centric social settings and files were hoarded by senior colleagues, open access lets women research independently at any hour.
Comprehensive legislative datasets also let law reform commissions and attorneys-general identify discriminatory provisions through systematic analysis.
06An enabling layer for justice innovation
Before any entrepreneur can build a legal service, the law must be digitised, structured and consolidated — years of labour that innovators call undifferentiated heavy lifting. When a public platform absorbs that cost, it removes a large barrier to entry.
Services built on this layer include Sheria Kiganjani, which turns English statutes into more than 10,000 plain-Swahili questions and answers delivered over app, USSD and SMS; Sauti East Africa, which pushes simplified trade and border rules to informal cross-border traders on feature phones; Wakili AI; Afriwise; and M-Haki.
07Legal education
Law students are among the heaviest users of open-access platforms — one expert called the resource “the bible” of legal training. In Tanzania, 38.7% of TanzLII users were first introduced to it through a university or law faculty.
The reach extends further down: the judiciary uses the platform during the annual Law Day to show primary school students how digital law reports and the justice system work.
On the strength of the economic evidence
The correlation across 33 countries rests on a cross-sectional dataset with substantial missing data, and cannot establish direction — wealthier states can also afford better legal infrastructure. The study says so plainly, and proposes the research that would settle it: a multi-year panel of the African Law Index, a proper comparative cost study, and stated-preference valuation of the public good. We would rather point funders at that agenda than overstate what has been shown.
The same evidence shows where the next money goes furthest.
The return so far is concentrated in the professional core of the justice sector, where it is reliable and well evidenced. It reaches the wider public unevenly. That is not a reason to discount the investment — it is the clearest available map of where additional investment would compound fastest, because in each case the expensive part is already built and paid for.
have never been trained — and training is the strongest lever in the data
Trained users are 23 points more likely to almost always find what they are looking for, 14 points more likely to use the platform weekly, and 28 points more likely to use its analytical tools. The evaluation names training and communications as the single most cost-effective way to extend reach.
did not know the offline edition exists
Pocket Law was built precisely for low-connectivity users, and it works. The barrier is not engineering, it is distribution: the product is finished, the awareness is not. In rural and underserved areas, lack of awareness (37%) outranks connectivity (34%) as the leading barrier.
of practitioners say their communities can use the platform directly
For everyone else, value arrives through an intermediary — a paralegal, a legal-aid worker, a law clinic. That layer is doing the translation work, and it is not funded as part of the platform. Ninety per cent of Tanzanian law is in English; more than 80% of the population speaks Swahili.
of users are women, and the base skews to the urban legal hubs
The evaluation is candid that its own survey channels reach people who already use the platform, so this describes who currently uses it rather than who needs it. The five-country study is clear that open access dismantles professional gatekeeping for women — that benefit is available and under-collected.
And then there is permanence. The platform is still grant-funded. Completing the governance transition to the TanzLII Trust and securing a domestic budget line is the decisive task ahead, and it is already well advanced — the adopted 2026–2030 Strategic Plan, the September 2024 Dar es Salaam Declaration on Open Access to Law, an interim secretariat, and early funding diversification are all in place. What is being asked for is a bridge across a transition that has been designed, costed and started, not an open-ended subsidy.
Put plainly: the expensive, uncertain part of this investment is behind it. The corpus exists, the courts have adopted it, and the behaviour change is measurable. What remains is comparatively cheap, well specified, and is where each additional euro now buys the most.
“The hardest part is the last part.”
Justice in Data, conclusionFour conditions, in the order they have to happen.
The most transferable finding in the evaluation is not a number. It is a sequence — content and coverage first, then editorial standards and formal authority, then the features, training and outreach that drive adoption, with governance and a sustainability plan running underneath from the start. The order is the deliberate point.
Build the content foundation first
And resist pressure to compress it. Institutional ownership cannot be retrofitted once a platform has launched without it.
Insist on open, structured standards
With content authority resting on the institutions that make the law, so the investment interoperates and compounds rather than locking value in a proprietary product.
Plan the exit from the outset
Keep primary law free, layer value-added services and tiered API access on top, and finance the permanent institutional home as part of the programme — not in a closing phase.
Treat non-elite reach as a core cost
Mobile-first and offline access, local-language summaries, and the paralegal and legal-aid intermediaries through whom most citizens are reached. Not an afterthought — a budget line.
Blended, publicly anchored
Core funding embedded in the national justice-sector budget, primary legal materials free and non-negotiable, with value-added services, tiered API access, a Law Society membership contribution and institutional subscriptions layered on top. Development partners bridge the transition rather than sustaining it.
€182,000 a year, falling
About two-fifths of the five-year plan is the salaried core — a small coordination and editorial team. The recurrent cost is smaller still once the one-off scanning tapers after 2028. This is a running cost a national institution can reasonably carry.
Everything on this page is drawn from two bodies of work.
Laws.Africa commissioned and co-funded the TanzLII evaluation but did not write it. ALT Advisory wrote the report; Parallax Information Consulting ran the citation analysis on which the quantitative findings rest, and ran it independently of us; an independent field coordinator managed the fieldwork in Tanzania. Where our involvement limits what can properly be claimed, the report says so.
Justice in Data — Report 2
Measuring the Impact of Digital Access to Law in Tanzania: an impact evaluation of assistance for legal information infrastructure, assessed against the OECD-DAC criteria. Prepared by ALT Advisory, commissioned by Laws.Africa, supported by GIZ under the PRoLA programme. July 2026.
Impact of Digital Legal Infrastructure on Economic Development and Governance
A five-country comparative study — Tanzania, Kenya, South Africa, France and the United Kingdom — prepared by The Hague Institute for Innovation of Law (HiiL) and commissioned by GIZ under the PRoLA programme. It is published as one product in three parts: the main study, the country case studies, and a policy brief setting out four institutional pathways, financing scenarios and ten recommendations.
ALT Advisory prepared the TanzLII impact evaluation. HiiL prepared the five-country study and its policy brief. The citation analysis underpinning the quantitative findings was run independently by Parallax Information Consulting.
Laws.Africa and AfricanLII, the technical partners to TanzLII and the network of African Legal Information Institutes.
The German Development Cooperation’s Promotion of the Rule of Law and Judiciary in Africa (PRoLA) programme, implemented by GIZ.
Figures on this page are reproduced from the sources above without adjustment. Two are worth flagging for readers holding the reports: cost per judgment appears as €1.45 in the evaluation’s executive summary and €1.65 in its efficiency chapter — the first divides average annual spend in 2023–25 by the corpus, the second divides the projected forward budget; both are correct and this page shows the range. And the dependency series (56% in 2021, 66% in 2024, 93.5% in 2026) spans a change in question framing between the 2024 and 2026 survey waves, so the direction is reliable while the levels are not strictly comparable. Justice in Data is published by Laws.Africa and AfricanLII under CC BY 4.0.