Reduce actual harm
Measure land, water, toxic exposure, unsafe work and restoration. Arrests and licences are not substitutes.
Just transitionreduce harm + expand real options
Land + farmingtenure, food, future production
Mining + workcash, risk, ownership, technology
Water + healthexposure, care, restoration
Power + rulesstate, chiefs, rights, enforcement
Markets + financebuyers, credit, gold flows
Evidence synthesis for Fern
Reducing gold-mining harm without shrinking people's options
A livelihood-centred just transition for Ghana that prevents exclusion, concentrated gains and displaced harm.
livelihoodspowerpolicy package
Ghana as the main caseWest African and international responsibilitiesEvidence reviewed to 3 October 2026
Choose the route
A connected argument, with the evidence behind it
The briefing route carries the main decision story. The full route keeps every methodological, case and bibliography slide available. Nothing has been discarded.
Presentation depth
The decision
The question is not simply how to stop galamsey
Which combination of support, rules, purchasing, accountability and cooperation can reduce total harm while expanding people's real options?
Expand real options
People need adequate income, timely cash, land, credit, buyers, safety and bargaining power.
Prevent cost-shifting
Do not move extraction, pollution or exclusion to another group, district, country or trade route.
How to read the evidence
Four kinds of claim are kept separate
A comparison design estimates what an intervention changed.
Interviews, surveys or spatial data show how a process worked in a specific setting.
An institution reports what it launched, planned or procured.
Several sources support an inference that still needs testing in Ghana.
Why it mattersA programme can deliver training, equipment or licences without improving income or water.Rule used hereEvery major claim says what the evidence can show and what it cannot.
The analytical centre
The Sustainable Livelihood Approach keeps people, power and ecology in one frame
People's
real options
real options
Naturalland, water, forests, minerals
Physicalroads, tools, housing, processing
Humanhealth, skills, labour, knowledge
Financialincome, savings, credit, debt
Socialnetworks, trust, groups, influence
A livelihood is sustainable when people can withstand shocks, maintain or improve their capabilities and assets over time, and avoid damaging the resources and livelihoods that others depend on.
Context and shocksGold prices, seasons, illness, migration, land loss and environmental change.
Rules and powerLaws, institutions, tenure, gender norms and markets shape who can use each asset.
Livelihood portfolioFarming, mining, trade, services and migration are combined and changed over time.
OutcomesIncome, wellbeing, resilience, inequality and the future natural-resource base.
How the study used itIt combined oral histories with 12 indicators across the five capitals to reconstruct livelihood trajectories from 1992 to 2021.How we use itTest changes in assets, access, power, portfolios, resilience and ecology, then use social reproduction to expose unpaid work, care, time and hidden cost-shifting.
A second analytical lens
Map the production of gold and the production of life as one system
Goods and incomemining · farming · processing · trade
Maintaining lifefood · water · care · health · schooling
One connected system
“The production of goods and services and the production of life are part of one integrated process.”Luxton 2006, quoted in Buss et al. 2019
What this lens revealed in the study
Care, domestic work and farming
Less or less predictable mine time
Read as less “reliable”
Weaker access to teams and learning
Lower earnings and formalisation access
30% / 12%women / men reporting that family obligations determined their working hours. This is an observed mechanism, not a reform effect.
Use with the SLAA household income gain is not sufficient if time poverty, unpaid work, exposure or school loss increases.LimitBuss et al. studied women’s double burden in DRC, Rwanda and Uganda. It did not evaluate a reform or fully measure intergenerational effects.
The Ghana livelihood spine
Three questions keep change and inequality in view
Q1
What household livelihood portfolios existed before and after ASGM, and what influenced the changes?
Q2
What were the livelihood impacts of ASGM's emergence?
Q3
What opportunities and barriers developed for different trajectory groups, and how did they shape those trajectories?
Where the main livelihood study looked
Three farming communities in Atiwa West, Eastern Region
Ghana
Eastern Region highlighted; Atiwa West marked within it.
zoom to the district→
Atiwa West administrative outline
North-up, simplified boundary. The broad north and narrow southern extension match the study's published map.
360 householdsrandomly sampled within the three communities
30 oral historiesten purposively selected in each community
2021 fieldworkretrospective histories reconstructed back to 1992
Local evidencenot a national prevalence estimate for Ghana
CoverageCommunities were selected because ASGM was important there.LimitThe results describe these places. They are not national prevalence estimates.
How the study reached its conclusions
Several methods were combined, but the evidence is retrospective
Household surveys, sampled randomly within the three communities.
Purposively selected oral histories, ten per community.
Farmer focus groups and key-informant interviews.
Livelihood history reconstructed during fieldwork in late 2021.
StrengthSurvey patterns were interpreted alongside life histories, group discussions and observations.LimitThis was one 2021 study asking people to recall change. It was not a panel following the same households for 29 years.
Q1 · Before and after ASGM
Mining entered a portfolio of activities
Before mining expanded
Cocoa, maize, cassava, plantain, vegetables, livestock, fishing and forest products. Young people also migrated to cities or learned trades.
After about 2010
Farming remained important. Mining, trade, transport, food, rentals, repairs and construction grew around it.
Not one simple switch
Some households combined mining and farming, used mining income to fund a farm, or moved between activities as sites opened and closed.
Other changes mattered
Roads, electricity, mobile phones, cocoa support, age, skills, land and finance also shaped the available choices.
FindingASGM changed the mix of livelihoods rather than replacing farming in one uniform way.LimitThe study reconstructs sequence and experience. It does not estimate the causal effect of ASGM on each activity.
Q1 · Reported livelihood mix
Mining was important, but participation differed sharply by community
Akwabuoso
Households participating in ASGM.
Ekorso
Households participating in ASGM.
Pameng
Households participating in ASGM.
Income share
Average reported share from ASGM among participating households.
Cocoa or other cash crops
47% to 63% of households participated across the communities. Participants reported 62% to 65% of income from this activity.
Arable crops
34% to 67% participated. Participants reported 29% to 31% of income from this activity.
What this showsFor participating households, mining could dominate reported income.What it does not showSelf-reported shares do not compare net returns after capital, debt, labour time, risk or restoration costs.
Identity, assets and access
People did not enter ASGM from the same starting point
Age
Younger and physically able household heads had more access to mining and related work.
Gender
Consolidating cases were largely male-led. Women appeared in fluctuating and marginalised paths, with roles shaped by male-controlled assets and sites.
Land rights
Owners could lease or mine. Tenants could lose farms when landowners accepted mining.
Migrant status
Migrants without local networks, knowledge or secure resource rights were prominent in the marginalised group.
Education
The authors linked more education with a stronger ability to negotiate involvement.
Credit and networks
Finance, business links and influence helped people move from labour into ownership or trading.
Reported patternIdentity interacted with land, assets and networks to shape opportunity and exposure.Important cautionThe paper does not estimate a causal or statistical identity effect. These are descriptive patterns in the authors' grouped data and life histories.
Resources behind the identity patterns
Land and credit shaped who could benefit and who could be displaced
Households reporting land ownership in Akwabuoso, Ekorso and Pameng.
Households reporting that land was easy to access for ASGM.
Households reporting no access to credit.
Land could be an asset for one person and the loss of a livelihood for another.
Why this mattersA policy routed through title holders or licence holders can reinforce the same unequal starting points.Design implicationEligibility and compensation must include tenants, workers and customary users, not only formal owners.
Migration and movement
The study describes several kinds of movement, but does not count migrant arrivals
Before ASGMCocoa attracted migrant tenants and caretakers; young residents also left for cities or trades.
→
ASGM expandsWorkers, companies, equipment and services moved towards active deposits.
→
Deposits shiftSome mining companies left Akwabuoso and activity became more important in Pameng.
→
Unequal securityMigrants with weak land rights and local networks faced greater risk of marginalisation.
What is documentedDirection, reasons and consequences of movement appear in interviews and livelihood histories.Missing numberThe paper gives no migrant arrival count or percentage. It cannot tell us the scale of inflow.
Q2 · Livelihood impacts
Income gains and environmental loss happened together
Options expanded for some
- cash, jobs and new business demand
- school, health and housing expenditure
- capital for farming or another business
- skills and local market growth
Options narrowed for others
- farm, forest and water loss
- unsafe pits and unreclaimed land
- higher food costs and less farm labour
- weak compensation and insecure tenure
A household can gain income while its community loses water, land or future choices.
ObservedTransect walks found brown streams, liquid mine waste, pits, deforestation and degraded land.LimitThis project did not test water chemistry, and its 2023 and 2024 papers use the same fieldwork rather than independent samples.
Q3 · Unequal trajectories
The authors grouped 360 households into three livelihood paths
Consolidating
Assets, business, land, machinery, finance and networks supported further accumulation.
Fluctuating
Living standards or assets improved, but gains remained unstable as sites, cash and access changed.
Marginalised
Land insecurity, degradation, weak compensation and casual work narrowed choices.
How groups were madeThe authors used 12 livelihood indicators and then compared group characteristics.LimitThese are constructed analytical groups, not fixed social classes. The comparisons do not prove ASGM caused every difference.
Three published case accounts
Similar mining growth, very different lives
SourceFaithful paraphrases of Table 5, with three short excerpts.Privacy and accuracyThe portraits are fictional. Read the linked paper for the full published accounts.
Ghana case 2 · Prestea mining region
Mining widened cash options while damaging the assets that made livelihoods possible
Prestea-Huni Valley, Western Region · fieldwork from October 2013 to June 2014
2019 administrative boundary, north up and used only as a locator. Points mark communities, not mines or concessions.Authors' proposals, not tested interventionsInvest in rural jobs, education, agriculture and constructionRequire reclamation, refilling and reforestationProtect water and farms through buffers, monitoring and enforcementProvide wells, safer methods and environmental information
Method151 adults: 79 questionnaires and 72 participants in nine focus groups; three farmers also gave in-depth interviews.LimitCross-sectional and partly purposive or snowball sampled, with no comparison site or intervention test. The four driver percentages are multiple responses among 73 respondents, not causal estimates.
Prestea · DFID five-capitals application
Mining can raise cash while drawing down the assets a livelihood needs
ASGM
livelihood
strategy
livelihood
strategy
The cost test: more financial capital is not proof of a sustainable livelihood if natural and human capital decline and families or communities carry the repair.
Why this model mattersIt shows which assets mining uses, creates and depletes, and who can gain even while another group loses future options.LimitTable 5 inventories reported assets and Table 6 reports perceived vulnerabilities. There is no monetary balance sheet, environmental testing, before-and-after comparison or causal estimate.
Prestea signal + wider Ghana evidence
Migration status does not tell us a person's role, power or impact
Short-stay migrants
23 of 79 survey respondents lived in the community for a short period or purpose. The study gives no nationality, work-role, legal-status or power breakdown.
Long-term settlers
Another 24 of 79 had moved there permanently. Length of residence can change local ties and land access, so settlers should not be merged with mobile workers.
Mobile labour and services
Ghanaian internal migrants, West African workers, migrant women and mobile traders can face different land rights, wages, policing and eligibility for support.
Capital and equipment-linked actors
In Upper Denkyira, Chinese-linked miners, financiers and equipment networks used pooled finance, machinery and Ghanaian partnerships to accelerate extraction and concentrate some gains.
Prestea evidence boundaryThe paper's discussion mentions unlawful participation by non-Ghanaians, especially Chinese immigrants. It does not identify a Chinese subgroup or measure its prevalence or effects in Prestea.What the focused study addsIts 15 interviews, three focus groups and site visits link mechanisation and Ghanaian partnerships to unequal gains, land and water harm, and women being pushed towards tailings work in Upper Denkyira East.
Method and lawThe Upper Denkyira evidence is retrospective and local. Act 703 makes only citizens eligible for small-scale licences; it does not make every migrant a foreigner, operator or offender, and other roles need separate legal analysis.Actor-mapping ruleRecord origin, nationality, residence history, work role, gender, tenure, documentation, finance, equipment control and partnerships separately.
Ghana case 3 · Tarkwa Nsuaem · 2013 data
Mining growth did not translate into secure local livelihood options
What this supportsA plausible livelihood squeeze: farmland loss, weak local access to formal mine jobs and unemployment can make ASGM hard to refuse.What it does not proveThe survey is cross-sectional and retrospective. It does not observe a before-and-after cohort, distinguish ASGM from all surface mining, or establish that mining caused each economic change.
Tarkwa Nsuaem · Linked livelihood and exposure evidence
Land, forest and drinking water sit in the same livelihood system
250resident adults surveyed
200filtered water samples
10sites: nine surface waters and one borehole
10 monthsbiweekly sampling, January to October 2013
Forest-dependent livelihoods
The municipality contained the Bonsa, Ekumfi and Neung forest reserves, covering 440.15 km². The paper says rural women in particular relied on oil palm, snails, herbs, spices and firewood.
Boundary: the study describes this dependency but does not measure mining inside the reserves or changes in forest cover.
Measured water quality
At all ten sites, mean arsenic, lead, cadmium and mercury concentrations exceeded the guideline values listed by the study. Manganese exceeded its listed value at one site.
Method: atomic absorption analysis, blanks and certified-reference recovery reported at 95% to 100%.
Experienced land and water loss
In the narrative, 87% reported lost farmland, 83% rated drinking water bad, 82% said traditional water sources were no longer drinkable and 93% attributed water pollution to mining.
Boundary: these are residents' reports, not exposure-linked health outcomes.
Source attribution limitNo unmined control watershed, pre-mining baseline or mine-level tracing was used. Local geology, ASGM, larger surface mines and tailings could contribute; the paper often moves between these categories.Detection limit problemThe 10 μg/L laboratory detection limit equals the listed arsenic and mercury guideline and exceeds the cadmium guideline. The claim that 100% of individual samples exceeded those values is therefore not established for results below detection; the ten site means are the safer comparison.
Tarkwa Nsuaem · Stepwise logistic regression
The regression is useful as a question, not a causal answer
Reported signals for perceived pollution
Deteriorated water qualityOR 6.92 · p=.02
Belief that continued mining would pollute waterOR 5.97 · p=.010
These variables partly restate the perception being predicted, so the association is not an independent test of pollution.
Reported signals for stopping mining
Household income below US$1 a dayOR 5.11 · p=.003
Junior High School educationOR 3.25 · p=.009
The paper does not explain the second outcome in its methods or report how the reference groups were coded.
The published table fails basic consistency checksFor no formal education in the first model it reports B=3.12, p=1.21 and OR=4.46. A p-value cannot exceed 1, exp(3.12) is not 4.46, and the prose instead says “about six times.” Confidence intervals, model fit, missingness, validation and full reference coding are absent. Stepwise selection adds instability.
Use the model forA hypothesis: material position and direct experience can shape environmental and livelihood judgements; low formal education does not invalidate local knowledge.Do not use it forCausal claims, targeting groups, reliable effect sizes, or proof that age and gender do not matter. Their omission from the final table is not a reported null result.
Incentives and constraints
Mining can be attractive, or hard to refuse, for six different reasons
Cash timing
Frequent payments meet food, health, school and debt costs before crops pay.
Possible high returns
A successful strike can look better than years of uncertain farm income.
Farm pressure
Input prices, weather, labour loss, weak credit and degraded land make farming uncertain.
Nearby buyers
Gold is liquid. Traders may also provide advance finance.
Low entry, unequal advancement
Casual work may be accessible, while ownership requires land, capital and networks.
Local demand
Mining supports food, transport, rent, repairs, trading and construction.
A credible alternative must replace the functions people rely on, not only the activity.
What makes impacts grow
Six conditions can turn one livelihood change into a reinforcing system
Land and tenure
Owners authorise use and capture value; tenants can lose land and compensation.
Finance and equipment
Capital enables scale and mobility. Debt can bind gold to a buyer.
Political protection
Influence can shape licences, enforcement and access.
Fragmented government
Mining, water, farming, labour, health and purchasing can pursue different goals.
Weak alternatives
Training without land, childcare, finance, buyers or bridging income is not a viable option.
No remedy
Unrestored land and health costs are transferred to neighbours and future users.
SynthesisThese mechanisms recur across the Ghana studies.ImplicationChanging one element can leave the rest of the system intact or make power more concentrated.
Gendered effects
A woman can do mining work and still be made invisible as a miner
Buss et al.: 878 survey participants, 60 focus groups with more than 400 participants, 28 life histories and repeated observation, 2015 to 2017. National boundaries are Natural Earth 1:110m; colours show country scope, not mine sites.
14% / 64%women / men reporting digging
30% / 12%women / men saying family obligations set their hours
8.7% / 17.2%women / men receiving training
Almost nonewomen among licence holders
Use in GhanaTreat these as warning mechanisms to test, not as Ghanaian prevalence estimates.LimitThe sites were purposively selected and the study examined likely barriers; it did not evaluate a formalisation reform.
Prestea · Gender and child labour
Adults described women as integral and children as present, but the study did not count workers
83.6% · 61 of 73selected a response category that included women among people involved in ASM.
35.6% · 26 of 73selected a response category that included children among people involved in ASM.
Derived from the mutually exclusive categories in Table 3. These are respondents' descriptions, not observed workforce shares or prevalence estimates.
Just-transition testRecognise workers beyond digging, protect wage and decision rights, and replace the income and care functions households depend on before restrictions.Child protection testProhibit hazardous child labour while tracking household income, schooling and displacement so enforcement does not push children into less visible harm.
Direct Ghana evidence
Women do mining work, but men often control the work, pay and rules
What the Ghana studies did
Prestea and Bondaye, 2017. One 93-person qualitative dataset included 49 women workers, 19 male miners, site committees, local and traditional leaders, an association official and migrant-women focus groups.
- A connected study interviewed the same 49 women and observed eight sites over ten days.
- The shared sample means the papers are complementary analyses, not independent confirmation.
- Both are local, non-probability studies and cannot estimate national prevalence.
Why exclusion still harmsMining income could strengthen household provisioning and bargaining power. Unequal work is not disposable work; removing it without a viable alternative can deepen hardship.Policy testRecognition, wage rights, childcare, safety and decision power must cover processors and carriers, not only diggers or licence holders.
Direct Ghana evidence
Better jobs for a few can coexist with fewer ways for women to enter
Conditions improved for entrants
Observed jobs had regular pay, protective equipment, health arrangements, injury insurance, water, toilets and shelter.
Some women held skilled roles
Women included a supervisor, equipment operator, production worker, nurse or health assistant and office or cleaning staff.
Machines removed common entry jobs
Carrying, washing and amalgamation work, often performed by women at labour-intensive sites, had been replaced by machines.
Skills became a gate
Mechanised firms recruited for experience and competence. Women with less access to technical training could be excluded before job quality is counted.
What to measureCount all women working before formalisation, then track who enters, who loses work, redeployment, net income, care and control over assets.LimitTwo firms and seven women cannot establish a national or causal effect. The study did not follow displaced workers.
Gendered value of mining
Mining can matter more to women even when women earn less
Reported mine income above outside income
Study countryWomenMen
Rwanda
+223%mine versus other income
+30%mine versus other income
Uganda
+335%mine versus other income
+64%mine versus other income
These are within-gender comparisons reported by the authors, not a women-versus-men pay comparison.
Who is modernisation for, and whose livelihood is treated as expendable?
InterpretationRemoving mining access may eliminate one of women's most remunerative options even where their roles remain unequal.LimitIncome was self-reported in 2015 and highly dispersed. The authors could not conclude that women's and men's mean mine earnings differed statistically, and no reform effect was tested.
What Fern should keep track of
Do not count “women in mining”; trace positions, power and losses
Measure before, during and after the policy. Publish outcomes by role and social position, not only participant totals.
FieldworkUse private, role-sensitive interviews and do not recruit only through male leaders, licence holders or association heads.StatusThis is a synthesis-based accountability framework, not a validated scale.
Affected communities
People receiving little mining income can carry large cumulative costs
- neighbouring and downstream residents
- tenant farmers and farm workers
- fishers, livestock keepers and forest users
- children, caregivers and injured former workers
- future users of land and water
Measure exposure, not only perception
Obiri et al. sampled ten water sites over ten months and surveyed 250 residents. Many sampled-site metal levels exceeded drinking-water guidelines; residents also reported water and farmland loss.
The study had no control watershed and could not trace pollution to individual mines or separate all surface-mining sources.
Mining policy is also water, food, health, land and intergenerational policy.
Define success before choosing a measure
Three different outcomes must not be treated as the same thing
Less harmful mining
Lower damage or exposure per unit.
Less mining
A smaller total footprint where extraction must fall.
A secure exit
People can stop without losing basic security or power.
Technology or formalisation may improve the first while increasing output. A ban may reduce visible activity locally while failing on livelihoods or displacement.
Core metricTrack absolute environmental harm and livelihood security together.DistributionReport outcomes separately by work role, gender, age, land tenure, migrant status, wealth and place.
Ghana policy status · checked 3 October 2026
Four policy tracks are moving, but none yet proves a just transition
Mining-law overhaul
Parliament still lists the Minerals and Mining Bill, 2026 as a bill. This deck keeps Act 703 as the working legal baseline until an enacted replacement is confirmed.
NAIMOS enforcement
The Ministry reports operations, arrests and equipment seizures. These are activity counts, not evidence that total harm fell or livelihoods became safer.
Livelihoods and cooperatives
NAELP remains listed in Ministry RTI material available in 2026. rCOMSDEP launched its first cooperative scheme in June 2026. Independent outcomes are not yet established.
GoldBod
Act 1140 centralised gold-trade regulation. On 3 September 2026, GoldBod still described the national traceability system as being procured.
A programme architecture is not an outcome.
Official sources establishLegal status, stated design, launches, procurement and reported activity.Independent evidence must establishWater, land, income, distribution, rights, displacement and long-term effects.
The Ghana evidence gap
Reported activity answers only the first question
| Reported indicator | What it tells us | What remains unknown |
|---|---|---|
| people trained | delivery and reach | income, job durability, return to mining |
| equipment or plants | inputs deployed | use, maintenance, rebound, who benefits |
| licences or cooperative members | formal enrolment | conduct, worker rights, exclusion, capture |
| gold purchased or exported | formal-channel volume | verified origin and lower mine-level harm |
| sites closed or arrests | enforcement activity | net extraction, livelihood cost, displacement |
No retained independent evaluation measures Ghana's current package across environment, livelihoods, distribution, rights and displacement.
The next section · Five measure families
We will test each measure, then ask what only a package can do
The sequence moves from restricting harmful activity to changing livelihoods, legal access, technology and markets. Measure 3 is the longest because formalisation changes several kinds of power at once.
Same tests each timeTotal harmLivelihood adequacyDistributionPower and captureDisplacement and durability
After Measure 5, the deck compares lessons from other sectors and assembles the policy package.
Measure 1 · Restrictions
A restriction can protect a place and still create a harmful transition
When it may help
Credible no-go rules can stop activity in a bounded place, especially when enforcement reaches organisers, finance and equipment.
When used alone
Income disappears while demand and productive capacity remain. Activity may hide, move, become more dangerous or rely on more powerful intermediaries.
Interviews with 58 selected former young miners in Ghana describe income loss, debt, family strain and coping during the ban.
What the Ghana study showsHow selected affected young people experienced the restriction.What it cannot showThe ban's national net effect, environmental benefit or total displacement. There was no baseline or comparison group.
Measure 2 · Livelihood support
An alternative must work as a livelihood, not only exist as an activity
Incomeadequate net earnings
Timingcash when bills arrive
Inputsland, credit, assets, care
Marketreliable buyer and price
Riskprotection from weather, debt and failure
Without a clear path that makes mining unnecessary, farming, training or a small business may simply be added while mining continues.
Review findingA systematic review screened 22,146 records, but only 21 studies met its criteria.ConclusionEvidence was too sparse and varied to name one universally effective livelihood alternative.
Measure 3 · Formalisation
A licence can open a path or concentrate control
Possible pathway
Visibility, secure tenure, investment, inspection, safer schedules, wages, insurance and environmental capacity.
Possible exclusion
High cost, delay and unsuitable land favour connected operators. Workers can remain informal under a formal licence, while new machinery expands extraction.
Formal status is an input. Conduct and outcomes are the test.
Ghana evidenceInterview and process-map studies identify licensing bottlenecks and unequal influence.LimitThey do not estimate the causal effect of receiving a licence. Officials and licensed actors are more represented than unlicensed miners and farmers.
Regulate each form of power differently
The alternative to exclusion is not deregulation
1. Entry
Make the miner route attainable: suitable land, proportionate fees, clear decisions, time limits and shared technical services.
2. Conduct
Keep firm limits on location, water, mercury and cyanide, labour, restoration and community rights, regardless of formal status.
3. Market power
Require buyers, aggregators, financiers and equipment owners to verify origin, disclose ownership, reconcile volumes and report anomalies.
4. Oversight and remedy
Cross-check licences, production, purchases and payments. Publish audits and provide an appeal route independent of sellers, buyers and chiefs.
Brazil's 2013 rule kept transaction paperwork but weakened first-buyer liability for false origin claims. The final peer-reviewed study estimates almost 10 additional homicides per 100,000 people in more-exposed municipalities, about a 30% increase.
Evaluated effectBrazil supports hard accountability at the first purchase. The design compares places by geological and protected-area exposure; it does not observe every transaction.Ghana design inferenceLower barriers for low-capital miners without lowering buyer, financier, environmental or labour duties.
Kumah 2022 · What the headline number can and cannot say
The 85% estimate is not proof that miners rejected an accessible system
85%+reported as informal
The paper repeats an older estimate for Ghana's ASM sector. Its 58 interviews did not calculate this percentage, and the denominator, reference year and present-day precision remain uncertain.
Issue a legal title
Unlock credit and investment
Create enforceable duties
Reduce informality and harm
What the study establishesWhy selected miners and stakeholders experienced the legal route as misaligned with local realities.What it does not establishA current national informality rate, a causal effect of licensing, or that every unlicensed miner faces the same barriers.
Kumah 2022 · Official guidance checked 3 October 2026
Eight stages remain online, but current official sources do not align cleanly
Kumah's hidden workhire a geological surveyor and prepare a site planuse legal support for gazette publicationregister an enterpriseresolve valuation or compensation disputestravel and follow up with central offices
Current conflictAct 703 section 85 and the Commission's rights table say a small-scale licence may last up to five years; its procedure page says renew every three. Environmental permitting under Act 1124 and L.I. 2504 also sits outside this diagram.Reform statusOfficial 2026 statements say district committees now review applications, while Parliament still lists a Bill that would make them the entry point. Treat the eight boxes as published guidance, not verified practice. Kumah did not observe completion times or costs.
Kumah 2022 · Current legal check: 3 October 2026
This split still applies: surface interests and mineral ownership are distinct
AI-generated explanatory image, not a mapped place
Mining crosses both systems
Minerals in their natural stateOwned by the Republic and vested in the President in trust for the people.
Still currentArticle 257(6) makes minerals property of the Republic, vested in the President in trust. Article 267 and the Land Act protect distinct stool, skin, clan, family and user interests; the government does not privately “own” all land or minerals.Current safeguardsLand Act sections 11 and 13 make discriminatory customary-tenure decisions void and treat chiefs, tendanas and other managers as accountable fiduciaries. These rules still need enforcement and direct representation.
Kumah 2022 · One label, different operations
A pan and a trommel should not disappear inside the same regulatory category
Shallow “dig and wash”
Alluvial work described as a survival activity along riverbeds.
panspickschiselsshovels
Mechanised alluvial work
Deeper and more capitalised extraction can disturb larger areas and waterways.
excavatorsdredgerspumps
Mechanised processing and hard rock
A trommel is a rotating screen used to wash and separate material by size.
shaftscrusherstrommels
Accounts from southern rudimentary minersThey attributed worsening water damage to people using heavy machines and argued that those operators needed licences.
↔
Accounts from Talensi hard-rock minersThey attributed environmental damage to southern mining and described their own pits as less harmful.
What the “blame game” revealsMiners distinguish activities that the formal category collapses, and regulation can lose legitimacy when it ignores those differences.What it cannot proveThe paper collected perceptions, not water tests, land-cover data or comparative site audits. Self-description cannot determine exemption.
Field visit · A methodological prompt, not a proven recipe
Psychological safety is designed around the participant, not guaranteed by the interview mode
What Kumah reports
Telephone interviews, full anonymity and local recruitment support appeared to encourage openness about sensitive and potentially illegal activity.
There was no in-person comparison. The study cannot isolate whether the phone, anonymity, trust, recruitment or pandemic context produced greater disclosure.
Recommendations · Keep proposal and evidence separate
The paper proposes a different formalisation model, but it did not test one
What Kumah proposes
- Bottom-up, collaborative governance
- Meaningful roles for miners, chiefs and local government
- Different procedures and costs for different operations
- Support that helps survivalist miners meet safety and environmental duties
- Formalisation beyond legal documentation
Challenges the paper identifies
- Some customary authorities may capture money or access
- The right degree of authority for chiefs is unresolved
- Centralisation creates delay and distance
- Local legitimacy does not prove safe environmental conduct
- Blanket enforcement may deepen livelihood insecurity
What a responsible trial must test
- Who enters, who is excluded and who abandons the process
- Total time, cost, travel and specialist burden
- Changes in income, unpaid work, worker status and control
- Measured water, land, mercury and safety outcomes
- Elite capture, complaints, displacement and durability
No intervention trialThe study is an exploratory qualitative investigation with 58 interviews and document analysis. It evaluates neither a differentiated licence nor collaborative governance.Implication for FernAdvocate for co-designed pilots with prospective evaluation and stopping rules, not immediate national rollout based on this paper alone.
Measure 4 · Cleaner technology
Efficiency can lower exposure per unit and raise total extraction
Adoption
Can miners afford and access the equipment?
Use
Is it used correctly and consistently?
Maintenance
Who pays for parts, repair and oversight?
System effect
Does recovery increase output or remove women's work?
The evidence includes abandoned equipment, maintenance failures, weak gender results and toxin substitution, alongside some sustained improvements.
Systematic reviewTen eligible occupational-health interventions were found.Evidence gapThe technology studies did not demonstrate a human-health benefit, and the broader GEF evaluation did not estimate one common causal effect.
Independent evaluation · 2022
When pilots end, what survives?
Five countries. Three completed mercury-reduction pilots. A post-completion test of whether equipment, institutions and behaviour lasted.
Where and how the pilots were assessed
Three pilot regions connected project records to later experience
West Africa
Burkina Faso and Senegal
Equipment sites in Sud-Ouest and Kédougou. Mali was planned but received no activities because of insecurity.
Latin America
Ecuador and Peru
Border project around the Puyango-Tumbes watershed, linking mining and downstream contamination.
Southeast Asia
Philippines
Several pilot areas, including Diwalwal, where later enforcement differed from other sites.
3Completed GEF-5 pilot projects, all implemented by UNIDO and completed in 2016 or 2017.
149Named interviewees across governments, implementers, miners, civil society, industry and experts.
4Evidence streams: project and terminal-evaluation records, literature, interviews and focus groups, plus geospatial analysis.
1Limited central-team field mission, to Ecuador. COVID restrictions meant local consultants conducted the other visits and interviews.
Why these casesThey were the only regions with both a recently completed GEF pilot and a current GOLD project.Main limitationPurposive cases, project monitoring and stakeholder recall support implementation lessons, not a shared causal effect.
First: what was actually implemented
Ecuador banned mercury. Peru did not.
The evaluation grouped countries into three regional projects, but governments and project teams used different measures.
EcuadorBan + pilot
Government policyA complete ban on mercury in ASGM was adopted during the project. The pilot did not advocate the ban.
Pilot actionPromoted gravity processing, trained miners and monitored mercury and cyanide in the Puyango-Tumbes watershed.
Immediate tensionThe ban made mercury users illegal, which evaluators said could make them harder to assist and monitor.
PeruNo full ban
Government policyMercury remained legal with restrictions on trade, permits and safe handling. Peru did not adopt Ecuador's complete ban.
Pilot actionTrained miners, monitored the shared watershed and helped mining organisations enter the formalisation process.
At closureNo supported miner had completed formalisation, although some organisations had begun the process.
Burkina Faso + SenegalEquipment pilot
Pilot actionInstalled one non-mercury processing system in each country and trained miners and health workers.
Policy actionProduced guidance for national mercury action plans, but the plans were not completed within the project.
At closureMiners used the machinery for some stages but still used mercury in other processing steps.
PhilippinesTechnology + health
Pilot actionInstalled non-mercury technology, trained miners and tested participants' blood and hair before and after intervention.
OrganisationHelped establish a national association of small-scale miners to support representation and formalisation.
EnforcementGovernment enforcement differed by site. Diwalwal later had stronger enforcement than the other pilot areas.
Second: what worked, what failed and why
No country was simply a success or a failure
Ecuador Harm shifted
What improved: The Ecuador-Peru project reported a 40% reduction after recalculation, and stakeholders later described further decline.
What failed or remained unsafe: Promoted gravity methods were not widely adopted. Cyanidation and flotation drove much of the decline, while poorly managed cyanide tailings remained a pollution problem. Long-term monitoring stopped.
Why: Cyanide was accessible and efficient. The mercury ban criminalised remaining users but did not supply a complete safer system.
Peru Partial, delayed
What improved: Some assisted miners reported completing formalisation after the project.
What failed: None had completed it by project closure; most miners remained informal nationally. Sediment and biological monitoring did not continue, and promoted gravity methods were not widely adopted.
Why: Formalisation was slow, institutional memory weakened through staff turnover, and cyanidation was easier to access. Peru did not impose a complete mercury ban.
Burkina Faso + Senegal Technology failed to last
What improved: Equipment initially reduced mercury use for some processing stages. Mercury use remained below its earlier level.
What failed: Some machinery stopped working, so miners returned to mercury for some steps.
Why: Spare parts, local maintenance skills and continued finance were missing. National action plans were completed only later with additional GEF funding.
Philippines Mixed, site-dependent
What improved: The project reported 368 kg less mercury use per year, lower blood and hair levels among participants, and a miners' association that remained active and grew.
What failed: Biomonitoring stopped. Mercury use kept falling in some sites but showed no improvement in others.
Why: Diwalwal was both the only site with further decline and the only one with evidence of stronger enforcement. The evaluation could not isolate enforcement's causal effect.
Do not compare the numbers as effect sizesThe 40% estimate covers the shared Ecuador-Peru project, not either country alone.Evidence limitClosure figures relied on project records; later findings relied heavily on interviews and sparse quantitative data.
The uncomfortable findings
The same outcome can look different once we ask how it happened and whether it lasted
A project claim changed materially after the terminal evaluation recalculated it.
Miners praised new equipment mainly for better recovery, not for health protection.
The durable substitute was often a different toxin that the pilots had not promoted.
Without continued biomonitoring, the reported health improvement could not be checked after closure.
A fall in mercury use is not enough evidence that total harm fell.
A mercury project is not yet a just transition
The evaluation also shows who and what a narrow target can miss
Women
Projects recognised women’s low-paid and highly exposed roles. Most results frameworks still counted participation, with few indicators of power or economic improvement.
Downstream communities
People consuming contaminated water or fish were not directly served by GOLD, even though they can bear harm without mining income.
Local buyers
Buyers can be exploitative, but they also provide proximity, supplies and credit. Only one newer project specified livelihood mitigation if buyers were removed.
Whole landscapes
Most work centred on mercury and project facilities. Water, land, forests, biodiversity and damage beyond selected sites received less attention.
Distribution testCount changes in income, exposure, bargaining power and access within groups, not only participant totals.Boundary testTrack harm outside the supported mine, including downstream and displaced effects.
What Fern should carry into Ghana
Judge the package after support ends, not only while the pilot is staffed
Baseline
Measure mercury, cyanide, tailings, water, health, recovery, income and roles before intervention.
Adoption package
Match yield and cash needs. Fund equipment, credit, parts, operators and maintenance together.
Whole-toxicity test
Verify that one hazardous input is not replaced by another unmanaged process.
Distribution test
Track women, workers, owners, buyers, migrants, farmers and downstream residents separately.
Exit and follow-up
Assign a funded institution to monitor, repair, enforce and publish results after donor support ends.
The transferable lesson is a set of design conditions, not a technology shopping list.
What is well supportedMaintenance, finance, local capacity, monitoring and system-wide pollution checks are necessary for durability.What remains unknownThis evaluation does not identify one sufficient package or estimate what it would achieve in Ghana.
Law£TechTrade
planetGOLD
Mercury is treated as a system problem
The programme connects formalisation, finance, technology, formal markets and shared learning. That is a stronger starting point than an equipment-only pilot.
The just-transition question is whether the package also expands real options, reduces total harm and shares power.
A system design, not one intervention
planetGOLD connects barriers that earlier pilots treated separately
Knowledge and communication across all fourGender and safeguards across project activities
The causal chainFormal status is expected to unlock finance; finance enables technology; responsible production opens markets; shared learning supports replication. Every arrow is an assumption to test.
Why this mattersFinance was the largest component, a major change from older technology-led projects.Evidence limitThese are programme design shares. They do not show which component caused an outcome.
Ghana entered Phase 2
Ghana’s project turns the model into a five-year national test
Jurisdictional formalisation
Coordinate policy and practice within defined places rather than treat each site in isolation.
Finance and supply chain
Expand financial inclusion and build responsible routes to market.
Mercury-free technology
Support uptake of processing methods that avoid mercury.
Knowledge and capacity
Build local skills and share lessons within Ghana and the global programme.
StatusThe project launched in 2023. These figures are stated resources and targets, not verified results.Phase 2 changeThe wider programme says it will add land, biodiversity, climate and integrated landscape approaches.
Why the design is promising
It responds to several reasons earlier pilots did not last
Earlier problemEquipment was unaffordable or abandoned.
Design responseFinance, hands-on training, local equipment and service providers.
Earlier problemCleaner methods lost on yield or convenience.
Design responseTest recovery efficiency and business viability, not health messages alone.
Earlier problemInformality blocked loans and markets.
Design responseConnect policy reform, legal recognition, finance and formal buyers.
Earlier problemPilots repeated mistakes in isolation.
Design responseA global hub, peer calls, regional exchange and shared reporting.
Earlier problemBuyers could not verify responsible production.
Design responseCommon social and environmental criteria plus supply-chain links.
Concrete learning exampleEcuador shared why a mobile plant failed so projects in Peru and Colombia could reconsider that approach.Remaining uncertaintyThe 2022 evaluation could not yet judge training quality or final adoption.
Do not confuse programme activity with impact
Two major reports answer different questions
Independent evaluation, 2022
- Reviewed the design and early implementation of Phase 1.
- Compared the design with lessons from completed GEF pilots.
- Used portfolio records, interviews, focus groups, field work and geospatial analysis.
- Could assess coherence and monitoring plans, not the final effect of the full package.
Programme assessment, 2025
- Documents the Global Project’s finance tools, criteria, knowledge hub, shared indicators and exchange.
- Reports technical-report pilots in six countries and an online reporting system.
- Describes country teams’ experience of common tools and peer learning.
- Programme-authored, with no separate methods section or counterfactual outcome design.
What can be saidThe integrated machinery exists and has produced tools, common standards and learning channels. These sources do not establish that the complete model caused durable gains in health, livelihood security, gender equality or total environmental harm.
Gaps identified inside the GEF evaluation
A strong mercury programme is not yet a complete just-transition programme
Attribution
Only one-third of the 369-ton mercury target was expected from child projects. The rest relied on later replication and knowledge diffusion that was hard to monitor or attribute.
Water and health
GOLD did not fund water or sediment monitoring and did not include activities to improve long-term monitoring of health outcomes.
Whole landscapes
Land, forests and biodiversity received little funding in Phase 1. Safeguards mostly applied to supported facilities, not damage beyond them.
Gendered power
Projects planned gender analysis, representation and sex-disaggregated counts, but few indicators measured empowerment or equality.
People outside the mine
Downstream communities exposed through water and fish were not directly served. Indigenous conflicts were often avoided by locating projects elsewhere.
Displaced livelihoods
Shorter supply chains could remove local buyers who also provide credit and supplies. Only one project specified alternative-livelihood mitigation.
Important updatePhase 2 says it will add land, biodiversity, climate and jurisdictional approaches.Still to testWhether those additions change outcomes beyond supported sites and across social groups.
Formalisation risk · Working law checked 3 October 2026
A licence can protect miners, or make an operation easier to capture
What the working legal baseline already doesAct 703 separates land from minerals. A small-scale licence may be transferred only to a citizen and with ministerial consent. Other mineral rights need ministerial approval for transfer, assignment, mortgage or encumbrance. Parliament still lists the 2026 replacement as a Bill. These checkpoints do not cover every route to practical control.
Documented mechanisms
Ghana studies report concession overlap, attempted enclosure, temporary corporate access agreements followed by eviction when deposits became attractive, and gains concentrating among actors with land, finance and networks. Adranyi et al. found entrepreneurs operating ASGM companies and employing residents as casual workers.
Not yet measured
No retained evaluation follows newly formalised Ghanaian miners and estimates how often a company or financier later acquires legal or practical control. A formalisation-driven corporate grab is therefore a serious risk to test, not a demonstrated programme effect.
Formalisation risk · Safeguards before finance
Legal status needs rules that keep value and decision power local
Map every right
Record land users, customary claims, tenants, licence holders, workers, operating companies, financiers and beneficial owners. A licence holder cannot represent them all.
Consent and valuation
Require accessible information, independent advice and fair valuation before transfer, lease, partnership or loss of use. Include women, migrants and tenants directly.
Limit concentration
Publish proposed transfers and encumbrances, disclose beneficial owners and related parties, review cumulative holdings, and give affected groups a route to object.
Finance without surrender
Disclose interest, collateral, default, equipment and offtake terms. Offer patient capital and independent legal support so compliance finance does not become control.
Share value and power
Protect wages, safety, bargaining and profit participation. Do not count a one-off payment or rehiring as a successful transition without durable income and voice.
Follow what happens next
Track ownership, contracts, defaults, transfers, roles, earnings, land access and decision power for several years, broken down by gender, wealth, tenure and migrant status.
The post-formalisation testWho owns each right?Who controls decisions?Who receives profit?Who carries debt and damage?
StatusThese are proposed safeguards derived from the documented risk mechanisms. The retained studies do not evaluate this complete package.Design implicationFinance and formalisation should not proceed before the ownership, transfer, labour and monitoring rules are clear.
Collective organisation · A contested institution
A cooperative can build miners' power, or become a new gatekeeper
What collective organisation can unlock
- Bargaining with buyers, licence holders and government
- Shared equipment, compliance services, savings and credit
- Safer access to work, mutual support and public recognition
- A platform to challenge exclusion and demand rights
The result depends on who creates it and who controls it
How the same structure can exclude
- Fees, documents or role definitions keep poorer workers out
- Founders, financiers or customary elites choose the leaders
- Leaders control machinery, contracts, prices and payments
- Gender, class, ethnicity and migrant hierarchies enter the organisation
Seven self-organised women's groups supported savings, welfare, work access and resistance to male exclusion. Women also reported favouritism, class divisions and occupational exclusions. Qualitative evidence from one district, not a causal evaluation.
In two mandated cooperatives, leaders came through political, customary and trading networks; miners reported little voice and weak accountability. The study used 43 interviews and six focus groups, but its purposive conflict-setting cases cannot establish prevalence or direct transfer to Ghana.
Across six mine areas, existing male control of land, capital and authority shaped access to associations. Some women still gained referrals, standing or mutual support. This large mixed-method study described mechanisms; it did not evaluate a cooperative reform.
DRC, Rwanda and Uganda
Women's presence in leadership did not reliably redistribute power
Uganda gold siteAssociation controlled by founders
Five men supplied most capital and, respondents said, made all decisions. A woman processor reported that members were contacted for money, but not useful information.
Mechanism: formal organisation consolidated founder, investor and landowner power.
South Kivu, DRCA woman at the top was not enough
A women's committee dealt with literacy and sensitisation but was described as peripheral to mining decisions. In a cooperative headed by the chief's wife, members reported centralised decisions and “no redistribution of wealth.”
Mechanism: descriptive representation without member control.
Rwanda wolframite siteA women-only cooperative had disbanded
Former members recalled loans, health insurance, advocacy and a private space to discuss family planning. They wanted it restored. The study does not establish why it ended or that the state closed it.
Mechanism: autonomous space and material support, with uncertain durability.
Woman-led is not automatically feminist governanceLeadership identity matters, but the institutional test is whether power and value move.
Members set the agendaAssets and income are redistributedCare and safety shape decisionsLeaders can be questioned or removed
MethodThe cases come from interviews and focus groups inside a wider mixed-method study; they identify mechanisms, not comparative treatment effects.Authors' warningPolitics, nostalgia or grievance may have shaped the contrasting accounts. The paper did not evaluate a cooperative reform.
Grassroots organising · state-led inclusion · state administration
A cooperative can organise from below, or administer from above
Talensi women's groups
Women organised around work access, savings, welfare and resistance to eviction. Their groups created bargaining power but still contained favouritism, class divisions and occupational exclusions.
Question: how can support strengthen capacity without taking over the agenda?
Ghana's rCOMSDEP design
The official programme promotes locally owned cooperatives, legal concessions, training and shared processing. Official descriptions do not yet establish who selects members, controls facilities or receives value.
Feminist question: does recognition resource women's own agenda, or replace it with a state-designed form?
Egypt as a historical warning
Roger Owen describes post-1952 compulsory and later nationwide agricultural cooperatives coordinating crops, inputs and marketing. His wider account treats supervised cooperatives as one route through which postcolonial states extended rural administration.
Question: is the collective accountable downward to members, upward to the state, or both?
Trace who definesmembershipleadershipthe agendaasset controlbenefit sharingexit and remedy
Do not infer equivalenceEgyptian agricultural history does not predict Ghanaian mining outcomes. It exposes a question about institutional purpose and accountability.Ghana evidence statusrCOMSDEP sources document programme design and rollout, not livelihood, gender or governance effects.
Formalisation risk · Formalise whom?
A legal operation can still rest on an informal workforce
Entrepreneur or rights holder
The actor with land access, capital, networks and administrative capacity is best placed to obtain the permit and new support.
licencefinancemachinerycontractsmarket access
formal asset
≠formal labour
≠formal labour
Workforce underneath
Diggers, haulers, processors and service workers may remain casual, paid by piece or share, unregistered and unable to influence the formal entity.
womenmigrantsyoung workersancillary rolescasual labour
Seven months of fieldwork and more than 200 open interviews linked outside capital to greater differentiation. Permits often recognised financiers and landowners while workers remained informal. Strong process evidence, no population estimate or intervention counterfactual.
Eight months, 89 recorded interviews and site observation mapped semi-formal production around tributors, sponsors and layered labour. It shows why a licence-holder count can miss who works and captures value. Two purposive sites, so the pattern is analytically rather than statistically generalisable.
Do not ask only whether an operation or cooperative is legal. Ask which people obtained rights, who remains a worker, whose work is recognised, and whether wages, safety, bargaining power and profit shares changed.
What is establishedFormal status can be selective within one production network.What remains openNeither study estimates how often this occurs under Ghana's current gold reforms, or effects by gender and migrant status.
Design implication · Representation is not automatic
Test the rules that turn a group into democratic power
Follow the distribution chaineligiblejoinsattends and votesholds office and controls assetsreceives work, income and remedy
Disaggregate every stageGender, age, migrant or place-of-birth status, ethnicity, disability, wealth, tenure, work role and owner or worker status.Evidence gapGendered and class exclusion are documented. Migrant exclusion inside mining cooperatives is a critical question, but it was not directly tested in these cooperative studies.
A cooperative membership count is not an inclusion result. The test is whether different groups can enter, influence decisions, control assets and share value.
What wider evidence adds
Each lever needs a distribution and displacement test
Lever
Risk if applied alone
What Ghana should measure
Formalisation
Costs and land access can exclude poorer miners. A documented right can also be transferred, pledged, aggregated or controlled through contracts.
Who gets, keeps or loses a right; every transfer, beneficial owner, debt claim, role and income by gender, migrant status, tenure and wealth.
Cooperatives
A mandated group can centralise licences, machinery, contracts and payments in founders or local elites while workers remain informal.
Eligibility, active membership, votes, offices, assets, contracts, work, income and complaints by gender, migrant status, role, tenure and wealth.
Finance
Capital can improve practices, but can also increase debt, extraction capacity or control by financiers.
Loan terms, ownership, production scale, land conversion, defaults and who carries risk.
Technology
Equipment can fail after support or replace mercury with unmanaged cyanide.
Total toxicity, water and health, recovery, maintenance, spare parts and use after project exit.
Formal markets
Compliance costs or buyer withdrawal can divert gold to less visible routes.
Mine-level origin, price, prompt payment, credit replacement, excluded sellers and trade diversion.
Jurisdictional pilots
Harm can move beyond the selected district, watershed or border.
Mining, equipment, workers, pollution and gold routes inside and outside the intervention area.
InferenceThe finance row states a risk to test, not a measured planetGOLD effect.Core lessonTrack household trajectories and affected non-miners, not only supported operations and mercury totals.
Recommendation for Fern
Use planetGOLD as a backbone, then add a just-transition contract
Keep the integrated core
Formalisation + finance + technology + markets + shared learning is more credible than any one measure alone.
+
Make outcomes binding
Finance and formal status are means. The tests are harm, livelihoods, power, durability and spillover.
Fern does not need to deliver every function. It can press for the network, guardrails and public evidence that make the package accountable.
Measure 5 · Ghana status checked 3 October 2026
Regulating buyers is necessary, but a formal sale does not prove lawful origin
Tier 1
May buy from a miner and sell to a Tier 2 buyer. Published terms require seller identity, receipts, monthly reporting, due diligence and suspicious-transaction reporting.
Tier 2
May buy from miners or Tier 1 buyers and sell to an aggregator. The same public record, ownership and anti-money-laundering duties apply.
Aggregator
May buy from miners and lower tiers for supply to GoldBod. Entry requires much greater capital and fees, making concentration an outcome to measure.
On 16 February 2026 GoldBod suspended new Tier 1, Tier 2 and self-financing aggregator applications; only new aggregator applications remained open. Existing applications continued. This can strengthen coordination and restrict market entry at the same time.
Official limitsGoldBod says a substantial share of smuggling is linked to buyers and that gold was not yet fully traceable to concessions. On 3 September 2026 the national system was still being procured.Test in practiceTrack prices, credit, excluded sellers, ownership concentration, trade diversion and whether declared mine output matches purchases. Official statements establish rules and plans, not effects.
Violence and the first buyer
Brazil shows the risk. Ghana shows warning signs, not a causal estimate.
Brazil: evaluated effect
After first buyers could presume origin declarations were true, the final study estimates almost 10 additional homicides per 100,000 in more-exposed municipalities, about 30% above the pre-change mean.
Ghana: observed warning signs
Qualitative research documents armed robbery, extortion, self-defence groups, site conflict and violent police or military raids around transient mining economies.
Ghana: current official concern
GoldBod attributes a substantial share of smuggling to buyers and says concession-level traceability was incomplete. This is an institutional account, not an independent causal study.
For Ghana, monitor violent deaths and injuries, robbery, extortion, firearms, threats to women, migrants and whistleblowers, enforcement violence, suspicious transactions and production-to-purchase mismatches.
EstablishedWeakening first-buyer liability increased homicide exposure in the Brazilian setting; Ghanaian studies record serious violence around gold economies.Not establishedNo retained Ghana study identifies a causal chain from buyer opacity or laundering to a rise in homicide. That relationship must be investigated, not assumed.
New evidence block · Cross-sector studies
Now we leave gold mining briefly
The purpose is not to import ready-made solutions. It is to identify recurring mechanisms that can be tested against Ghana's land, labour, market and political conditions.
ImportantThe two cross-sector papers are connected, not independent confirmations. Gather, Prowse and Seussler reanalyse cases assembled by the Yeung et al. realist review.
Learning from other sectors
Other settings reveal conditions for success, not ready-made templates
Afghanistan · drug crops
Replacement fails when a crop also provides credit, land access, cash timing and risk management.
Uganda · forest payments
Payments cut short-run tree loss, but enrolment was limited and clearing resumed after payment.
Cash-plus programmes
Income protection plus livelihood support can improve average welfare, without proving exit from harmful activity.
Labour programmes
Training takes time and assumes jobs exist. It cannot replace income lost tomorrow.
Fisheries
Rights, diverse user participation and adaptation are associated with better outcomes; much evidence is observational.
Coal transitions
Packages for formal workers can miss informal work, women, care and place-based loss.
Transfer ruleBorrow mechanisms only after checking land, labour, market, state and time differences.Use for GhanaThese cases identify questions and design conditions. They do not predict Ghanaian effect sizes.
Study 1 of 2 · Yeung et al. realist review
Study 1 asks how change happens, but lasting outcomes are thinly reported
Stated climate objective, 99 interventions
Mitigation
61
Adaptation
26
Both
12
4interventions recorded respect for human rights as an explicit output, one in each detailed sector or sector combination.
Interventions with no reported evidence in the category
Climate output
30
Social-equity output
28
Climate outcome
64
Social outcome or gain
58
These are documentation gaps, not proof that nothing changed. Many interventions were ongoing, and reporting quality varied.
What can be visualisedCoverage, reported activities, barriers, enablers and outcomes across the review records.What cannotNo pooled effect size, counterfactual or ranking of interventions. Incidence in the literature is not causal impact.
Study 1 findings · Recurring enablers and barriers
Projects need material capacity and social legitimacy
Hard enablers
- Reliable, long-term finance
- Infrastructure, technology and market links
- Technical expertise, knowledge and usable data
- Clusters of reinforcing activities
Soft enablers
- Political ownership and cross-government coordination
- Fit with local needs and public policy
- Trust, awareness and credible benefits
- Social dialogue, community power and partnerships
Hard or structural barriers
- High technology and system costs
- Credit, land-tenure and administrative barriers
- Skills, labour and supplier shortages
- Unstable finance and dominant incumbent investment
Relational and political barriers
- Exclusion and unequal benefit distribution
- Patriarchal norms and unrecognised work
- Weak trust, buy-in and clarity about benefits
- Fragmented institutions and poor local fit
Hard without softAssets may be delivered but rejected, captured or poorly maintained.Soft without hardParticipation cannot supply finance, land, jobs, infrastructure or enforcement on its own.
Reading rule“Soft” describes relational and institutional conditions. It does not mean optional, inexpensive or secondary.Method noteThe review explicitly uses hard and soft for enablers. The barrier grouping here is our transparent synthesis of its sector findings.
Study 1 cases · Recognition and power
Consultation matters when it changes design and control
Bihar, IndiaSolar microgrid and recognition-based justice
External actors installed a village microgrid
Residents distrusted its quality and status
Grid connection reduced use of the microgrid
Mechanism: a technical solution that did not match residents' aspirations or recognise their knowledge lost legitimacy. Participation after key choices could not repair the original mismatch.
Peer-reviewed qualitative case studyPanamaCommunity-Based REDD+ and Indigenous governance
Indigenous authorities rejected an exclusionary plan
Programme was redesigned through joint planning
Funds and governance shifted toward Indigenous projects
15pilot projects
881participants, 54% women
2,887 hareported as supported
Common lessonEngagement is consequential when communities can reject, redesign, govern and control resources, not merely attend meetings.Transfer to GhanaAsk which decisions affected communities can change, who chooses representatives, and whether refusal triggers redesign or punishment.
Study 1 cases · Local institutions and livelihoods
Co-design works through concrete authority, labour and services
CambodiaUplands irrigation and Farmer Water User Communities
Water-user groups joined design and implementation
Irrigation systems and local management were built together
Completion reporting linked both to production and income gains
31,140 hareported irrigated
257,292unskilled person-days
26.2%of those days by women
IndiaMGNREGA, local public works and active citizenship
Local knowledge shaped small public works
Employment and assets were delivered together
Households reported labour and water-access gains
1,400households in 35 villages
90%benefited from at least one project
90 vs 75median workdays, women vs men
Mechanism to testParticipation can improve local fit when it carries knowledge and authority into the design of assets, work and services.BoundaryNeither case proves that participation alone caused the reported outcomes, and neither involves gold mining.
Comparison · After both studies
Same records, different questions
Realist synthesis
Yeung et al. reconstruct how interventions were intended or reported to work, for whom and in what context.
FindsEngagement, trust, local fit and collaboration recur as mechanisms associated with climate and social outcomes.Cannot establishAn average causal effect of participation.76 studies
99 interventions
99 interventions
Fuzzy-set QCA
Gather et al. retain 80 interventions and test combinations linked to a composite output proxy.
FindsDocumented engagement is absent or low in most labelled solution terms, alongside funding and planning.Cannot establishThat removing participation improves results.No contradiction“Absent” in the QCA means low or unreported calibrated membership, not a controlled test of exclusion. Its proxy only partly captures voice, agency, recognition and historical injustice. Read the QCA as a warning that engagement is not sufficient on its own; read the realist cases as evidence that power-sensitive engagement can shape fit, legitimacy and distribution.
Fern positionCombine material capacity with chosen representation, decision rights, transparent resources and remedy. Measure whether engagement changes decisions and who gains, not how many meetings occur.Independence ruleDo not cite these papers as two independent confirmations. Gather is a re-analysis of the Yeung review cases.
Study 2 of 2 · Gather, Prowse and Seussler 2025
Study 2 tests combinations, not isolated policy measures
How the conclusion was reachedResearchers calibrated documentary evidence into fuzzy-set scores, built a truth table and searched for recurring configurations linked to the composite output.What this isA comparative, configurational association study. It is not a pooled effect estimate and does not evaluate a Ghana mining package.
Before the findings · The shared evidence base
One evidence map supports both cross-sector studies
Regional distribution of the 99-intervention review
Asia
39
Africa
38
Latin America
17
Europe
2
Caribbean
1
Pacific Islands
1
These counts cover 98 interventions. One continent-scale Africa intervention was omitted from the review's regional figures. The English-only search produced little small-island evidence.
Sector distribution of the 99 interventions
Energy only
46
Cross-sector*
23
Agriculture only
15
Ecosystems only
13
Infrastructure only
2
14India
13South Africa
7Indonesia
5China
5Ethiopia
*Cross-sector is the derived remainder after the four published single-sector counts. These are review-wide counts; the paper does not publish this breakdown for the 80-case QCA subset.
Denominator ruleThis is 76 studies, 99 extracted interventions and 80 QCA cases, not 80 studies.Data accessThe calibrated 80-case dataset is not published with the article. Its data-availability statement says data are available on request.
Study 2 results · Published complex-solution metrics
The seven paths overlap and explain less than half of positive-outcome membership
PathRaw coverage barRawUniqueConsistency
A
15.9%0.1%99.9%
B
15.8%0.5%99.9%
C
23.0%6.7%96.4%
D
24.2%5.5%94.2%
E
8.4%2.6%93.5%
F
11.5%2.2%90.4%
G
15.9%0.1%99.9%
How to read thisRaw coverage is the share of positive-outcome membership associated with a path. Unique coverage is the share associated only with that path. The paths overlap, so raw percentages cannot be added.Published-table problemTable 1 prints A and G identically. Table 3 describes them differently. These values reproduce Table 1 rather than repairing the article.
Study 2 diagnostics · The appendix and regression
No single condition passes the paper's necessity threshold
Necessity consistency, Appendix Table 5
The vertical marker is the study's stated 0.90 threshold for calling a condition necessary.
Policy alignment
0.80
Robust funding
0.79
Community participation
0.75
Strategic planning
0.71
Technical knowledge
0.68
Clustering
0.55
Stakeholder engagement
0.35
Resource barriers
0.32
Uncertainty
0.30
Structural challenges
0.30
Implementation constraints
0.10
Just-transition regression, Table 4
Fractional-logit coefficient with standard error below. Green is positive; clay is negative. Only funding is statistically significant in this composite model.
Robust funding1.15***SE 0.30
Strategic planning0.40SE 0.26
Structural challenges0.37SE 0.34
Clustering0.26SE 0.30
Community participation0.24SE 0.24
Resource barriers0.12SE 0.31
Policy alignment-0.21SE 0.27
Stakeholder engagement-0.22SE 0.30
Technical knowledge-0.26SE 0.31
Uncertainty-0.29SE 0.30
Implementation constraints-0.45SE 0.28
Safest conclusionFunding is the strongest individual association and funding plus planning appear in all seven labelled paths. The appendix does not establish either as individually necessary under its own threshold.LimitRegression n=79, adjusted R²=0.22. These are associational coefficients, not causal effects.
Study 2 findings · Seven reported configurations
One shared floor, seven reported routes
Robust fundingFinance judged sufficient and sustained enough for the intervention.Strategic planningSocial policy and resilience measures structured around goals, resources and timelines.
Read the paths togetherThey overlap and are not a ranking. Most conditions matter only in some combinations.Do not treat absence as advice“Absent” means low or unreported membership in the coded case material. It does not prove that removing a condition improves outcomes.
Study 2 interpretation · A justice safeguard
“Not detected as necessary” does not mean “not needed”
configurations in Table 1 code stakeholder engagement as absent
The paper also reports policy alignment absent in more than half. It suggests central coordination or institutional flexibility may sometimes support short-run outputs. Table 3 conflicts with Table 1 for configuration D.
Authors' warningTransitions without procedural justice may face weaker long-term trust, legitimacy and social stability.Ghana implicationDo not let finance or state capacity substitute for the chosen voices of miners, women, migrants, customary users and communities living with the damage.
A stronger comparison than participatory versus state-led
Capacity and participation solve different problems
State capacity
×
social power
×
social power
Power narrowly held
Power broadly distributed
Strong state
Weak state
What Gather et al. detectSome documented outputs occurred where broad engagement was not reported.What justice still asksWho set the goals, who could refuse, whose knowledge counted, and who could correct the state?
Institutional lensAcemoglu and Robinson define inclusive political institutions as the combination of a capable state and broadly distributed power.Use in GhanaDo not trade participation for capacity. Build the public capacity to act and the countervailing power to keep action answerable.
Historical lenses, not transferable models
Japan and Egypt show why delivery and justice must be assessed separately
Meiji JapanCentralised state-building with limited popular participation
Why Nations FailTreats the Meiji Restoration as a critical break toward more inclusive institutions.Yamada's challengeInstitutional upgrading occurred under an authoritarian elite revolution. Old elites with capital or skills were selectively redeployed into new state, education, finance and commercial institutions.
Blind spot: rapid institution-building was elitist. Many former samurai lost income, tenant farmers bore costs, and the later imperial trajectory makes this a poor justice template.
Egypt after 1952Cooperative services and an expanded administrative state
Owen's evidenceCompulsory and later nationwide cooperatives coordinated rotations, inputs and marketing, but also enabled compulsory delivery, state deductions and unequal burdens on smaller farmers.Why Nations FailInterprets the Nasser-to-Mubarak trajectory as an old extractive order reconstituted under new rulers because power did not become broadly distributed.
Blind spot: neither book provides a matched Japan-Egypt evaluation. Owen documents institutional functions and tensions; Acemoglu and Robinson offer a broad historical theory.
Ghana governance testWho sets goals?Who controls assets?Who receives value?Who can refuse or exit?Who can obtain remedy?
InferenceA capable state may deliver visible outputs quickly. That does not establish fair distribution, durable legitimacy or freedom from capture.TargetPublic capacity plus organised counter-power, enforceable rights, transparent value flows and institutions that affected groups can challenge.
Study 2 limits · Blind spots and transfer boundaries
Use the seven pathways as design hypotheses, not a recipe
Our source checkThe appendix reports necessity-consistency values of 0.79 for funding and 0.71 for planning, below the paper's stated 0.90 necessity threshold. Table 1 prints configurations A and G identically, while Table 3 describes them differently; the tables also conflict over whether engagement is present in configuration D. Safest wording: funding and planning appear in all seven labelled complex solution terms; individual necessity and the exact seven-fold typology are not cleanly established by the published tables.
Study Tables 1, 3 and Appendix 2configurational evidence · Ghana transfer untested
New section · From evidence to policy design
Now we assemble the package
The next slides stop reviewing interventions one by one. They combine the findings into policy functions, sequencing rules and safeguards for a just transition in Ghana.
Question for this sectionWhat has to operate together, in what order, and with which protections so that reducing mining does not reduce people's real options?
The package logic
Four jobs must be done together
Protectincome, food, health, debt and rights during disruption
Createviable paths with land, finance, buyers, skills and care
Constrainecological limits and duties for organisers, buyers and financiers
Repairland, water, health and lost livelihoods through remedy
Creation without constraint can add another activity. Constraint without protection can destroy options.
Evidence statusCross-sector reviews support combined, context-sensitive pathways.LimitNo reviewed study evaluates this complete package in Ghana. It is an architecture to test, not a proven recipe.
General levers, locally designed
Nine functions define the package
Ecological limits
Places, thresholds and deadlines.
Transition floor
Income, food, health and debt protection.
Real capabilities
Land, finance, buyers, care and viable work.
Attainable legal route
Suitable land, proportionate rules and shared services.
Power and value
Rights, organisation, prices and joint decisions.
System duties
Obligations for organisers, equipment, finance and buyers.
Repair and remedy
Restoration, health, compensation and appeal.
Spatial coordination
Detect relocation, diversion and origin laundering.
Durable governance
Multi-year finance, evaluation and correction.
Sequence matters
Support must become usable before most restrictions remove income
Beforemap roles and rights; set limits; establish income and debt protection; secure land, finance and buyers
→
Duringphase restrictions; enable compliance; protect workers; verify purchasing; fund restoration
→
Afterfollow exits and returns; measure water, land and income; remedy loss; monitor destination areas
ExceptionAn imminent danger to life or an irreplaceable ecosystem can require emergency action.Continuing dutyEmergency action still requires protection, fair process and remedy for affected people.
Do not ask people to cross a bridge before it exists.
What can go wrong
Each measure can cause harm when it acts alone
| Measure alone | What it misses | Possible harm |
|---|---|---|
| enforcement | income functions and demand | debt, concealment, selective policing, relocation |
| licensing | land, capital and worker status | rights and profit concentrate in connected operators |
| training | jobs, cash timing, care and buyers | attendance without a viable livelihood |
| cash | productive assets and markets | temporary relief followed by return |
| clean technology | scale, ownership and rebound | more extraction or displaced women's work |
| traceability | mine practice and excluded miners | clean paperwork, dirty production or diversion |
| buyer exit | corrective finance and other routes | risk moves to weaker producers and buyers |
Do not mistake movement for reduction
Four kinds of displacement need different evidence
Mining relocation
Workers, equipment and extraction move.
Trade diversion
Gold takes another route while mining stays.
Origin laundering
Documents or intermediaries conceal the mine.
Environmental spillover
Water, sediment or exposure crosses boundaries.
ECOWAS cooperation should link customs and finance with labour rights, migrant protection, watershed monitoring and livelihood support.
Evidence gapNo retained study evaluates an ECOWAS just-transition package.Do not overclaimHigher exports from a neighbouring country do not by themselves prove that mining relocated there.
New section · From policy design to delivery
Now we ask who must make it work
A policy package is also a distribution of authority, money, duties and risk. This section maps who must participate, who must be regulated and who must be able to challenge decisions.
Reading rulePresence is not power. For every actor, ask what information, resources and decisions they can control, and who can hold them accountable.
Actor map 1 of 5 · Rights-holders
Start with the people whose rights, income and environment change
Inside mining
diggerswasherscarriersprocessorsteam leaderscasual workerslicensed workersformer miners
Land and farming
smallholder farmerslandownerstenant farmerssharecroppersfarm workerscustomary userscocoa caretakers
Daily life and care
women inside and outside miningpartners and dependantsunpaid caregiverswater and fuel collectorschildrenyoung peopleolder peoplepeople with disabilities
Migration and mobility
Ghanaian internal migrantslong-term settlersWest African cross-border workersmigrant women and service workersforeign financiers, operators and equipment suppliers
Exposed non-miners
downstream residentsforest-edge communitiesfisherslivestock keeperswater usersfuture land users
Participation ruleEach group must speak through people it chooses.SafeguardDo not treat chiefs, licence holders, household heads or cooperative leaders as substitutes for everyone else.
Actor map 2 of 5 · Current law checked 3 October 2026
Use distributed checks, not a chief-or-state binary
Chiefs, tendanas, family heads, queen mothers and councils can map interests, convene negotiations and contribute dispute knowledge. They must disclose conflicts, payments and beneficiaries.
Women, tenants, migrants, farmers, workers, downstream users and other customary users need separately selected seats. A chief cannot substitute for them or consent on everyone's behalf.
Public agencies set mineral, environmental and labour rules. Independent review must publish reasons and payments, audit compliance and hear appeals outside the licensing and purchasing chain.
Surface land and customary interestsCustomary rights, usufructs, tenancies, farming, residence and compensation claims shape practical access. Affected users have interests that cannot be reduced to a chief's position.
A just transition must reconcile both systems
Minerals and statutory authorityMinerals in their natural state belong to the Republic and are vested in the President in trust. Customary permission is not a mining licence, while a mineral right does not erase surface interests.
Three rulesNo chief grants a mineral right alone.No state licence erases surface rights.No buyer's paperwork proves origin by itself.Every decision needs an independent challenge route.
EvidenceKumah finds that chiefs are influential and locally legitimate, while also recording opaque deals, personal appropriation and resistance to accountability. The study does not test a governance model.Design responseDivide authority, record all interests, publish decisions and payments, and protect direct refusal, compensation and remedy rights. Land Act section 11 voids discriminatory customary-tenure decisions.
Actor map 3 of 5 · Ghanaian duty-bearers
Mining policy crosses mandates that must act together
Land and mining
Ministry of Lands and Natural ResourcesMinerals CommissionLands Commissiontraditional authoritiesdistrict assemblies
Environment and water
EPA GhanaWater Resources CommissionGhana Water CompanyCommunity Water and Sanitation AgencyForestry Commission
Livelihood and welfare
Ministry of Food and AgricultureCOCOBODLabour DepartmentMinistry of HealthGhana Health Servicesocial protection agencies
Law and accountability
ParliamentcourtspoliceCHRAJAuditor-GeneralOffice of the Special Prosecutor
Coordination questionWho has the budget, authority and duty for each outcome?Accountability questionWho can challenge a decision, obtain data and secure a remedy?
Actor map 4 of 5 · Ghanaian power and scrutiny
Follow control of land, money, equipment, gold and evidence
Production power
land and mineral-right holdersmine operatorscooperative founders and leaderscooperative members and excluded workerslarge-scale minesparent companies and beneficial ownersequipment ownersfuel suppliers
Money and trade
informal financiersbanks and mobile-money providersTier 1 buyersTier 2 buyerslicensed aggregatorsGoldBodexporters, refiners and off-takers
Collective voice
worker unionsfarmer organisationswomen's organisationsmigrant associationscommunity-based organisationsfaith groups
Financial crime, safety and scrutiny
Financial Intelligence CentreEOCOGhana Police ServiceWACAM and A Rocha GhanaGhanaian universities and laboratoriesjournalists and legal aid
Power testAsk who sets price, credit, eligibility, site access and acceptable evidence.Conflict testSeparate programme delivery, purchasing, verification, violence monitoring and grievance handling.
Actor map 5 of 5 · Beyond Ghana
Responsibility follows flows of people, money, equipment and gold
West Africa
ECOWASneighbouring mining regulatorscustoms and border agenciesGIABAriver basin bodiesmigrant-worker networks
Trade chain
UAE trading hubsSwiss refinersUK and EU tradersLBMAjewellery and technology buyers
Rules and finance
OECDEU regulatorsbanksinsurersassurance providersdevelopment finance institutions
Support and evidence
ILOUNEPUNDPWorld BankGEF and planetGOLDinternational NGOs and researchers
Fern's leverageEU buyers, regulators and finance should fund correction and remedy, not only demand paperwork.Regional safeguardMonitor new sites and trade routes so improvement in one place is not financed by harm elsewhere.
Representation
No actor can stand in for everyone affected
Direct representation is required
Workers by role; women inside and outside mining; tenants and landowners; migrants; former miners; downstream and forest-edge residents; fishers and other resource users.
Important, but not substitutes
Chiefs, licence holders, cooperative leaders, household heads, professional women's groups, district officials, buyers and NGOs. A cooperative leader does not represent every member or worker.
A chief is not every customary user. A licence holder is not a worker. A miner is not a downstream resident.
Minimum ruleSeparate seats, accessible information, safe dissent and paid participation.Remedy ruleComplaints must go somewhere independent of the body delivering or buying from the programme.
No one organisation can deliver the system
Build a network with four centres of responsibility
Community authority
Define acceptable outcomes, select representatives, monitor harm and use independent remedy.
Public coordination
Align land, mining, water, farming, labour, health, purchasing and social protection.
Market obligation
Buyers, refiners, financiers and equipment owners fund verified improvement and repair.
Independent evidence
Ghana-led research, laboratories, audits and public reporting track distribution and displacement.
Fern does not need to do everything. It can help make the whole system accountable.
Recommendations to Fern
Advocate for conditions, sequence and accountability
Fund Ghanaian leadership
Affected communities and Ghanaian researchers define outcomes and evidence.
Set clear red lines
No criminalisation-first package and no success claim based only on arrests or exports.
Demand the right sequence
Income protection and tested options before restrictions, except immediate danger.
Move costs upward
Buyers and finance support correction, verification and remedy without automatic exit.
Build independent checks
Community monitoring, water science, worker evidence, open contracts and safe complaints.
Track displacement
Monitor districts, borders and trade routes with ECOWAS partners.
Before advocacy or scale-up
Six questions for every proposed initiative
Coverage
Who is missing because they lack title, licence, time, membership or safety?
Adequacy
Does support match net income, timing, debt, risk and duration?
Power
Who can join, vote and hold office? Who controls land, licences, finance, contracts, price, equipment and complaints after formalisation?
Environment
Will total harm fall after rebound and toxin substitution?
Displacement
Where could people, extraction, finance, gold or pollution move?
Correction
What evidence triggers expansion, redesign, compensation or stopping?
Next-stage research
Evaluate the package without averaging away injustice
Follow people and places
Track owners, workers, tenants, women, migrants and affected non-miners before, during and after intervention. Include destination districts and watersheds.
Measure seven domains together
Absolute harm; livelihood security; social reproduction and future wellbeing; distribution; power; displacement and durability; implementation and cost.
Match method to claim
Use counterfactual evaluation where possible, realist synthesis for mechanisms, and meta-analysis only for genuinely comparable interventions.
Protect evidence quality
Use Ghana-led protocols, open methods, independent laboratories, conflict disclosure, community data governance and publication of harmful or null results.
Bottom line
The evidence supports a common architecture, not one universal blueprint
Protect people. Create viable options. Constrain the harmful system. Repair damage. Test every package for unequal effects and displacement.
Fern's strategic role is to help build the relationships, rules, finance and accountability that stop any one measure from becoming another source of harm.
recommendation from the full synthesisClickable bibliography 1 of 5
Ghana: livelihoods, gender, environment and affected people
Adranyi, Stringer and Altink (2023). The impacts of artisanal and small-scale gold mining on rural livelihood trajectories. Mixed-method study in Atiwa West.
Adranyi et al. (2024). Joined-up governance for livelihood vulnerability and mining impacts. Same underlying Ghana field project as the 2023 paper.
Arthur et al. (2016). Mixed-method livelihood study in Prestea, Bondaye and Himan, including reported participation by women and children.
Yakovleva et al. (2022). Gender equality, women’s authority and the likely effects of bans and formalisation in Prestea and Bondaye.
Arthur-Holmes and Busia (2021). Women’s work, care, safety, wage rights and inter-ethnic discrimination in Prestea and Bondaye.
Ofosu, Torbor and Sarpong (2022). Women’s employment and occupational conditions in two formalised, mechanised Ghanaian ASM firms.
Baddianaah, Tuu and Baatuuwie (2021). Livelihood implications of artisanal gold mining in farming communities, Wa East.
Obiri et al. (2016). Environmental and socio-economic impacts in Tarkwa Nsuaem, including repeated water sampling.
Osei et al. (2021). Ghana's ASM ban, youth livelihoods and imagined futures. University of Ghana and KNUST.
Hilson and Banchirigah (2009). Are alternative livelihood projects alleviating poverty in mining communities?
Clickable bibliography 2 of 5
Ghana: governance and current policy
Adu-Baffour, Daum and Birner (2021). Governance challenges of small-scale gold mining in Ghana: a Process Net-Map study.
Crawford and Botchwey (2017). Conflict, collusion and corruption in small-scale gold mining in Ghana.
Kumah (2022). Formalisation barriers and customary legitimacy in Ghanaian small-scale mining.
Mensah (2021). Legal pluralism, customary tenure and large-scale mining bias in Ghanaian ASM formalisation.
Parliament of Ghana (2020). Land Act, 2020 (Act 1036), including fiduciary duties and Customary Land Secretariats.
Constitution of Ghana, arts 257 and 267. Distinguishes minerals vested in the Republic from stool land held in trust for stool subjects.
Office of the Administrator of Stool Lands (2026). Current implementation account for Land Act 2020 and expansion of Customary Land Secretariats.
Arthur-Holmes and Mengba (2024). Informal women's mining cooperatives in Talensi, Northern Ghana.
McQuilken and Hilson (2018). Production networks and layered labour in Ghana's alluvial diamond sector. Open thesis with full methods.
Parliament of Ghana (2006). Minerals and Mining Act, Act 703, including mineral ownership and transfer rules.
Parliament of Ghana (2026). Current Bills catalogue, listing the Minerals and Mining Bill, 2026 as laid and gazetted on 26 May 2026.
Minerals Commission (checked 3 October 2026). Published eight-stage small-scale-mining licence guidance; its three-year renewal statement conflicts with Act 703 section 85 and the Commission's mineral-rights table.
Environmental Protection Authority (2026). Current permitting notice under Act 1124 and L.I. 2504, including small-scale mining.
Ministry of Lands and Natural Resources. National Alternative Employment and Livelihood Programme.
Ministry of Lands and Natural Resources (2026). First rCOMSDEP cooperative scheme launch at Akyem Kotoku.
Parliament of Ghana (2025). Ghana Gold Board Act, 2025 (Act 1140).
Ghana Gold Board (3 September 2026). Latest located official status still describes the national traceability and assurance system as under procurement.
Clickable bibliography 3 of 5
Mining comparisons, supply chains and displacement
Tsang et al. (2019). Systematic review of occupational-health interventions in artisanal and small-scale gold mining.
GEF Independent Evaluation Office (2022). Evaluation of GEF interventions addressing artisanal and small-scale gold mining.
planetGOLD (2025). Global Project Assessment, Phase 1. Programme-authored account of global tools, coordination and lessons.
planetGOLD Ghana (2023). Inception workshop, project components, finance and stated targets.
planetGOLD Phase 2. Programme scope and integrated formalisation approach.
Álvarez-Berríos et al. (2021). Formalisation, land cover and mining expansion in Peru.
IMPACT (2021). Just Gold lessons from incentive and traceability work.
BGR (2019). Traceability in artisanal gold in Kampene, DRC.
De Haan and Geenen (2016). Mandated mining cooperatives, elite capture and revenue distribution in eastern DRC.
Verbrugge (2015). Formal recognition of ASM entrepreneurs alongside persistent informal labour in the southern Philippines.
Buss et al. (2019). Social reproduction, gendered livelihoods, authority and likely formalisation effects in DRC, Rwanda and Uganda.
Pereira and Pucci (2024 working paper; 2026 article). A Tale of Gold and Blood, Brazil.
Brugger et al. (2024). The state and the legalisation of illicit financial flows: trading gold in Bolivia.
Stoop et al. (2018). Mining ban, mineral substitution and conflict in eastern DRC.
OECD (2018). Gold at the crossroads: assessment of supply chains in the Liptako-Gourma region.
Clickable bibliography 4 of 5
Cross-sector evidence and just-transition design
DFID (1999). Sustainable Livelihoods Guidance Sheets, including the five-capitals asset framework.
Scoones (1998). Sustainable Rural Livelihoods: A Framework for Analysis. IDS Working Paper 72.
Roe et al. (2015). Systematic review of alternative livelihood projects and conservation.
Mansfield (2020). Alternative development and opium-poppy livelihoods in Afghanistan.
Jayachandran et al. (2017). Cash for carbon and forest conservation in Uganda.
Leight et al. (2024). Cash-plus programming and economic outcomes.
Walk et al. (2021). Gender and justice in coal transitions.
Kivimaa and Kern (2016). Policy mixes for sustainability transitions.
Gather, Prowse and Seussler (2025), full text. Pathways towards just transitions in the Global South. Publisher record.
Yeung et al. (2026). Peer-reviewed realist review of just-transition interventions in developing countries. Read the complete 2024 IEU and ILO report.
Mejía, Restrepo and Rozo. Quasi-experimental evidence on enforcement and illegal coca markets in Colombia.
Toledo Orozco (2022). Gold-mining displacement and environmental governance.
Clickable bibliography 5 of 5
Participation, institutional power and the state
Acemoglu and Robinson (2012). Why Nations Fail. Used for its distinction between inclusive and extractive institutions and its historical interpretation of Egypt and Meiji Japan, not as an intervention evaluation.
Acemoglu and Robinson (2019), full accepted manuscript. Rents and Economic Development: The Perspective of Why Nations Fail. Defines inclusive political institutions as a capable state combined with broadly distributed power.
Yamada (2022), open article. Making Reform and Stability Compatible with Each Other: Elite Redeployment in Meiji Japan. Challenges the participatory reading of Meiji institutional change and documents its authoritarian, elitist mechanisms and uneven effects.
Owen (2004). State, Power and Politics in the Making of the Modern Middle East, used as a historical lens on rural state expansion.
Owen (1999), full paper. A Long Look at Nearly Two Centuries of Long Staple Cotton. Documents the productive, administrative and distributive functions of Egypt's supervised cooperative system.
Comparison ruleThese sources explain competing institutional mechanisms. They do not constitute a controlled Japan-Egypt comparison or predict Ghana's outcomes.Transfer ruleUse them to formulate tests about capacity, power, rents, accountability and remedy, then answer those tests with Ghanaian evidence.