CSAE 2026: What I Saw in Oxford This Year

What I love about CSAE is the African representation. I don’t know another economics conference where you can chat with people from across the continent over a single lunch. To measure this, I tried two approaches. First, I used AI to identify each speaker’s country of residence based on their primary affiliation. For global organizations like the World Bank, I assumed headquarters. By this measure, only about a quarter of speakers reside in Africa — though they come from 16 different African countries. The majority still come from the UK, Europe, and the US.

But residence undercounts African representation — it misses Africans in the diaspora, including PhD students at European and North American universities. So I also classified speakers by likely country of origin, using names as the primary signal.¹ By this measure, 37–41% of speakers are plausibly of African origin — but Africa-based institutions employ barely half of them.

There is also a disjuncture between where researchers are based and where they study. Papers cover 33 African countries, but speakers reside in only 16 of them. South Africa leads, with 7 speakers — and notably, this cohort looked very different from a decade ago: predominantly black, many originally from other African countries, and more women. The pipeline is diversifying, even if the overall geography hasn’t caught up.

The share of RCTs has held steady at roughly 28–30% since the 2021 conference. With funding cuts reducing the number of new trials being launched, this may shift downward over time.

Three other methodological trends caught my eye:

• Many studies examine longer-term impacts (at least 37 papers), many leveraging administrative data to track outcomes years after an intervention ended. For example, Deutschmann et al. show that a one-time offer of easier credit for water tanks gave Kenyan dairy farmers a seven-year head start that permanently boosted milk sales — and increased girls’ school attendance.
• At least 71 papers make use of administrative or non-survey data — tax records, satellite imagery, census data, transaction logs, national exam results. As primary data collection becomes harder to fund, expect this to accelerate.
• Several striking papers document impacts on outcomes not directly targeted by the program (at least 15 papers). Tamim et al. find that BRAC’s graduation program — designed to boost livelihoods — reduced mortality by 15% seventeen years later. Gobezie et al. show that Ethiopia’s school construction programme reduced women’s acceptance of wife-beating. And Cullen et al. find that phone-based mentoring during COVID-19 reduced teen pregnancy and school dropout by 90%.

One of the highlights was the graduation evidence. The headline news is good. Graduation works across multiple contexts, even when implemented by government. But design details matter.

• A graduation program for displaced Somalis in Baidoa tripled household assets and boosted consumption by 30%. Nearly all the gains came from one thing — goats (Leight et al.). #RCT
• Four coordinated RCTs across the Sahel find that graduation programs work even when run by governments in fragile states, but the full Capital+Psychosocial bundle particularly accelerates poverty exit for the not-quite-most-destitute — the very poorest may need something else (Bossuroy et al.). #RCT
• Seventeen years after BRAC’s graduation program in Bangladesh, treated households have 15% lower cumulative mortality, making this arguably the first experimental evidence that an anti-poverty program saves lives (Tamim et al.). #RCT
• A disability-inclusive graduation programme in Northern Uganda generates comparable economic returns for households with and without disabled members, with only 17% higher costs but just 7% lower benefits, debunking the equity-efficiency trade-off myth (Banks et al.). #RCT
• In Malawi, coaching adds no value, whereas capital and training do all the work (Poll). #RCT

• Combining digital climate forecasts with extension advice via SMS in Madagascar boosts smallholder yields by 45% — nearly double what either does alone — with women farmers benefiting twice as much as men (Filamant and Totohasina). #RCT
• India’s Green Revolution accidentally provided climate insurance, but the mechanism is entirely crop-specific: high-yielding rice varieties themselves are heat-tolerant, while for wheat it’s the irrigation infrastructure that came with them that matters (Jain).
• Ghana’s worst drought in 40 years reveals a cruel paradox: soybean farmers who actively sought mobile agricultural information were more likely to invest in drought-resilient strategies, but ended up worse off because no amount of better seeds can overcome a truly catastrophic shock (Wirth et al.).
• Africa’s irrigation problem isn’t just underinvestment — existing schemes are so badly placed that optimal reallocation alone could triple irrigated agricultural output across 32 countries (Fuseini).
• India’s Green Revolution rice varieties boosted yields everywhere, but high-yielding wheat initially hurt yields in the wheat belt until irrigation and fertilizer caught up decades later (Bevis and Rahman).
• GPS tracking lets Kenyan tractor owners trust their operators at a distance, expanding operating range by 160% and reducing spatial misallocation of farm machinery by 15% (Nottmeyer).
• When gold mines open in Tanzania, nearby farms shrink their planted area and shed family labor, but yields hold up fine, suggesting the trade-off is about competing for workers, not poisoning the land (Ibrahim).
• Digital agricultural advice reaches beyond the phone owner – 72-79% of recipients share with other household members in Malawi and Zambia – but men share 8-24pp more than women. (Mulungu and Ngoma). #RCT
• Fall armyworm infestation in eastern Uganda reduces maize yields and sales at high intensity levels, significantly worsening dietary diversity for mothers and young children through both reduced food production and diminished income from crop sales (Chisanga and Kassie).
• A three-day psychosocial training for 1,200 Ethiopian women farmers increased their willingness to pay for bio-fortified maize seeds by 5.4% and raised adoption by 6.2 percentage points, achieving similar effects to a price subsidy and suggesting psychological empowerment can close preference-action gaps in agricultural technology adoption (Belissa, Lensink, Saxena, and Wakoya). #RCT
• An RCT informing Kenyan informal market traders and consumers about the link between maize kernel integrity and aflatoxin contamination improved maize quality by 13.8% and changed inspection behavior, but failed to generate price premiums for safer food, suggesting information alone may be insufficient to create functioning quality markets (Kariuku, Osiemo, Magambo, and Cheruiyot). #RCT
• Social networks in the Lake Chad Basin drive adoption of climate-adaptive farming strategies that reduce output variance and food insecurity (Newman et al.).
• Land restitution boosts Colombian displaced farmers’ investments, but gains shrivel where the state itself did the displacing – trust in property rights enforcement matters as much as the title deed (Crowther).
• Peer effects on smallholder commercialization follow an inverted U across six African countries – moderate peer participation fosters market entry through social learning, but too much competition crowds farmers out (Dzanku).
• Madagascar’s vanilla price explosion made farmers dramatically richer and happier, but the windfall did nothing for their kids’ health or schooling – proof that money alone won’t buy the next generation’s human capital in extreme poverty (Boone, Kaila, and Sahn).
• Information about cooperative benefits decays sharply across space and time in rural Mexico, and while drought nudges coffee farmers toward cooperatives that offer insurance, the most extreme climate shocks overwhelm even strong social networks (Pitts, Storer, and Anttila-Hughes).
• A satellite-based credit scoring system that evaluates plot productivity rather than borrower profiles unlocks formal lending for marginalized Indian farmers (Kramer et al.). #RCT
• SMS receipts for daily milk deliveries in Ugandan dairy cooperatives help farmers catch dishonest transporters (Henning and Magana Saenz). #RCT
• A smartphone app delivering grape-growing videos in rural China boosts sweetness by 0.30 SDs – but adding aspirational content paradoxically crowds out the technical gains (Yang et al.). #RCT
• Fertilizer returns in Malawi follow an S-shaped curve, and the country’s fixed-quantity subsidy programme keeps 83% of farmers trapped in the low-return zone – redistributing the same total fertilizer at optimal doses could boost maize yields by 11% (Rodriguez-Sala).
• A livestock and agricultural input transfer programme in Southern Rwanda reduced food insecurity by 31% and improved dietary diversity by 22% among the poorest smallholder households, with female-headed households remaining most vulnerable (Yismaw and Kutela).

• Educational TV in Kenya boosts reading by 9% SD and curiosity by 12% at just $15 per learning-adjusted year of schooling, but inadvertently reinforces gender norms (Baier et al.). #RCT
• Peru’s extended school day reform boosted learning, university attainment, and labor participation seven years later, but not wages. Gains sustained for new cohorts partly because better students self-selected in (Campos et al.).
• A podcast-based career guidance program for Tanzanian secondary students raises national exam pass rates, university entry, and self-employment income – the mechanism is not information but hope (Almas et al.). #RCT
• Farmers in Cote d’Ivoire who fear losing their borrowed land hedge their bets by sending more kids to school – especially girls – making land certification a double-edged sword for education (Sory and Nguimkeu).
• Thirteen years after Ghana introduced free compulsory preschool, an RDD on census data shows 0.32 additional years of schooling and substantial literacy gains – with the biggest effects for boys and children in rural, disadvantaged regions (Lambon-Quayefio).
• South African primary school teachers systematically overestimate their learners’ reading ability – but a formative assessment intervention across 39 schools significantly improved accuracy (Lobelo and Ardington). #RCT
• Starting school too early in South Africa reduces both years of education and earnings (Samahiya).
• Uganda’s 2007 secondary school fee subsidy had no effect in public schools but significantly improved outcomes in subsidized private schools, with female students gaining nearly five times more schooling years than males, though only if they lived within five kilometers, free (Lauterbach).
• A structured literacy program in Ghanaian low-fee private schools raises reading scores by 0.5 SDs in just 9 months – at the 91st percentile of all reading interventions – with generative AI helping adapt lesson plans across countries (Andersen, Graffy, Kerwin, and Lambon-Quayefio). #RCT
• Three years after Malawi’s schools reopened, students are still learning one-fifth slower than pre-pandemic cohorts – remediation efforts have not even restored the old pace of learning, let alone closed the gap (Asim, Bashir, and Casley Gera).
• A multi-component anti-violence program in Zambian secondary schools got students talking about gender-based violence but failed to actually reduce it (Friedson-Ridenour et al.). #RCT
• Rules-based teacher allocation in Malawi was implemented very unevenly across districts, but schools where pupil-teacher ratios fell below 90 saw grade repetition drop by 3-4 percentage points (Asim et al.).
• Kenyan civil service teachers earn a causal wage premium of about 50% over identical private-sector counterparts, driven not by productivity or job amenities but by classic dual labor market segmentation (Barton, Bold, and Sandefur).
• Ethiopia’s policy mandating 70% of university seats go to STEM backfired: STEM graduates became 3pp more likely to be unemployed, with women hit hardest at 5.4pp – a cautionary tale of supply-side education policy without matching labor demand (Yigermal).
• Ghana’s school feeding programme covers only 21% of target children despite a 3.3:1 return on investment, and simulations show scaling up would significantly reduce poverty and inequality while underfunding widens the rural-urban education gap (Nunoo et al.).
• Ethiopia’s 1995 primary school construction programme gave women 0.55 extra years of schooling and reduced their acceptance of wife-beating by 4-6 percentage points across multiple justifications, causally linking education to shifting gender norms (Gobezie et al.).

• In Malawi, foreign aid to family planning, HIV/AIDS, and education reduces teenage childbearing by 15-25%, while other aid types do essentially nothing (Congdon Fors et al.).
• $1,000 cash transfers in Kenya cut infant mortality by 48% and child mortality by 45%, with the biggest gains where cash meets actual health facilities (Walker et al.). #RCT
• Chinese investment projects in Ethiopia improve children’s dietary diversity suggesting that the downstream welfare effects of infrastructure investment extend to nutrition (Terefe).
• Phone-based mentoring during COVID-19 reduced teen pregnancy and school dropout by 90% among 2,000 Botswanan adolescents, one of the largest mHealth effects ever documented (Cullen et al.). #RCT
• Group-based psychosocial programs in war-torn Myanmar did nothing for mental health — a precisely estimated zero — but both the skills curriculum and a placebo arm of cooking and gardening boosted social cohesion equally, suggesting it’s the group, not the therapy, doing the work (Cerkez et al.). #RCT
• A new multidimensional Period Poverty Scale reveals that menstrual deprivation among Nigerian schoolgirls independently reduces mental well-being by half a standard deviation and slashes happiness by 20 percentage points, even after controlling for household poverty (Nduka, Eze, and Mgbemena).
• An RCT in 140 Malagasy schools finds that tackling menstrual stigma with sanitary products, infrastructure, and peer ‘Young Girl Leaders’ boosts girls’ test scores by 0.15 SD through reduced stress and better peer networks rather than improved attendance (Macours, Vera Rueda, and Webb). #RCT
• World Bank aid cuts infant mortality by 5% regardless of local politics, but Chinese aid only works where the mother shares the president’s ethnicity – revealing how ethnic favoritism determines who benefits from foreign assistance in Africa (Jalloh).
• Replacing sugary date bars with healthier meals in Jordan’s school feeding program reduced children’s snack spending by 8%, increased physical activity, and improved attendance by one extra school day per year (Lombardini et al.). #RCT

• Exporting digitally-delivered services helps Latin American manufacturers innovate, but only when paired with R&D and employee training. Digital exposure alone does nothing (Achmad).
• Ghana’s firms could be 125% more productive if all frictions were removed, but the diagnosis depends on the sector: manufacturing suffers from poor roads and connectivity, while services suffer from too little local competition (Mensah and Van Biesebroeck).
• African SEZs attract more FDI when they’re large, multi-sectoral, and run as public-private partnerships — but without reliable electricity and port access, establishing a zone is just an expensive gesture (Tamno Tekam and Keneck-Massil).
• A $3-per-head soft skills program for Bangladeshi university students boosted employment by 4-5 percentage points, with women benefiting most. Proof that the binding constraint might be negotiation skills, not technical ones (Khan and Sen). #RCT
• Digital payments in Senegal’s taxi drivers cut cash costs by half and boost driver effort by 34% through monitoring – but half of drivers, especially the poorest, refuse to adopt precisely because they fear being watched (Houeix). #RCT
• Fortnightly business performance reports boost female microenterprise profits by 20% and sales by 35% in Uganda – but simply encouraging record-keeping does nothing, proving it is the analytics, not the data entry, that matters (Kempis, Ogden and Riley). #RCT
• Tanzania’s 2010 minimum wage drew workers to higher-wage locations but pushed them into less productive jobs. A spatial model shows a national rather than local minimum wage would halve the efficiency loss (Marshall and Martinez Leyva).
• Mining companies’ CSR across 14 African countries has zero effect on education, wealth, or child nutrition – and mines with higher exports do MORE CSR but have WORSE outcomes (Lefoll et al.).
• A global panel of firms across 67 countries shows competition crushes innovation everywhere, but the blow lands hardest in developing economies where financial constraints and informal rivals amplify the damage (Duodu).
• A small conditional cash incentive gets Kenyan micro-traders to cross the Uganda border more often, boosting revenue and profits that persist months after payments stop – the nudge matters as much as the money (Falco, Stein, and Wiseman). #RCT
• Dar es Salaam restaurants burn 11 tonnes of charcoal yearly and would recoup solar heater costs in under two years, yet their WTP is only 25% of market price even with flexible credit – information and payment design alone cannot close this gap (Alem). #RCT
• Business training programs consistently fail women, but a ten-factor readiness scorecard tested across Uganda, South Africa, and Mexico can close the gender gap entirely by targeting women with the highest program readiness (Anderson, Grover, et al.). #RCT
• An Indian textile firm would be better off hiring far more migrants – they work more consistently and have higher marginal productivity – but social frictions and recruitment costs keep the workforce suboptimally local (Ramachandran et al.).
• A field experiment with 500 Egyptian managers shows that introducing market competition significantly increases hiring of women, especially mothers, by reducing taste-based gender discrimination (Falco et al.). #RCT
• Adding soft skills to hiring algorithms at South Africa’s largest recruitment platform increased job offers by 22% and favored younger workers without formal credentials, with no negative spillovers on displaced applicants (Orkin et al.). #RCT
• Rural-urban migration in Ethiopia boosts agricultural output per worker by 30-37% in sending households, as remaining family members cultivate more land and reduce disguised unemployment rather than depleting farm labor (Yonis et al.).
• An RCT with 17,000 Kenyan households finds that individual migration information delivery raises incomes 7-9% by reaching first-time migrants, while group sessions crowd them out with zero net economic gain – removing all information frictions could shrink the rural-urban income gap by 25% (Barnett-Howell et al.). #RCT
• An RCT in rural Senegal finds that information about the U.S. Diversity Visa Lottery shifts preferences toward legal migration but backfires for ineligible individuals, who increase their irregular migration intentions instead (Beber et al.). #RCT
• Contrary to fears of rural hollowing-out, migrant households in 145 Senegalese villages are MORE socially connected than non-migrants – driven by others seeking their information networks, not their money (Beber et al.).
• Colombia’s Free Economic Zones boosted top firms’ profits and wages but pushed low-productivity formal firms out of business, increasing informality rather than creating net new employment (Contreras-Portela).
• South African manufacturing firms that automated saw 96% higher productivity after five years, and workers benefited with 25% more employment and 31% higher wages rather than displacement (Kilumelume et al.).

• Pakistani garment manufacturers want to hire women and don’t fear breaking social norms — what stops them is the 15% hit to the hiring probability from the cost of providing safe transportation (Shibuya et al.).
• Submarine cable arrivals in Africa boost women’s employment by 5.8pp – but only where roads and electricity already exist, and device access barriers mean fast internet alone cannot close the gender gap (Moses and Minasyan).
• Earnings info temporarily closed the gender gap in vocational training returns in Congo, but women’s raised aspirations collided with structural barriers and the gains evaporated – empowerment without opportunity is a double-edged sword (Chang et al.). #RCT
• Training women as primary care nurses in ultra-conservative Upper Egypt to see if these ‘pioneer’ workers shift not just their own employment but their entire community’s attitudes toward women working (Elsayed et al.). #RCT
• Free primary education cuts fertility across 17 African countries, but only where cultural norms already give women a voice – in high male-dominance ethnic groups, the same policy backfires (Collins et al.).
• Every extra year of schooling from Nigeria’s 1976 education reform cuts women’s support for patriarchal norms by 5 percentage points – and the effect is strongest in rural areas, not cities (Mpuure).
• Unconditional grants to Ghanaian women entrepreneurs do nothing – but tie the grant to a savings goal and sales jump 15% and profits 10%, with the biggest gains for women whose households expropriate their earnings, showing that protecting autonomy matters more than the cash itself (Campos et al.). #RCT
• Mobile money empowers Ugandan women economically (31% income boost) but cash handed publicly to couples cuts IPV by 12 percentage points – revealing a painful trade-off between women’s economic independence and their physical safety (Greco et al.). #RCT
• A Family Dialogue program in Mauritania improved couple communication and reduced IPV risk not by challenging male authority but by amplifying women’s voice within existing power structures – incremental but meaningful openings for change (Bedi et al.). #RCT
• Co-wives in Benin are not rivals – in public goods games they team up against their husband’s interests, especially when they have low bargaining power, flipping the standard narrative about polygamous household inefficiency (Baraka Kusinza and Guirkinger).
• Standard marriage ratio measures miss the real squeeze – once you account for polygyny intensity, Sub-Saharan Africa’s marriage market imbalance is far worse than conventionally measured, persisting at high levels since 2006 (De Swardt and Hartwig).
• Spousal cooperation and diet quality in Cameroon follow a surprising U-shape – couples who barely cooperate or cooperate deeply eat well, but the mushy middle of half-hearted cooperation produces the worst diets (Ndip, Saleemi, and Qaim).
• Ethiopia’s legal marriage age reform delayed marriage by only 0.6 months on average, and for the youngest girls it backfired – they ended up marrying partners 4.5 years older due to asymmetric constraints on brides versus grooms (Nkansah).

• Chinese-owned critical mineral mines in sub-Saharan Africa increase local protest likelihood by 15 percentage points relative to non-Chinese operations, driven primarily by environmental degradation rather than economic harm (Tsang).
• Rising agricultural potential across Sub-Saharan Africa paradoxically fuels more conflict (Murken et al.).
• Climate variability drives conflict in Sub-Saharan Africa through agriculture and markets, but the link is deeply nonlinear and context-dependent (Gattone, Romano, and Tiberti).
• UN peacekeepers in Sierra Leone boosted women’s schooling by 0.8 years and cut child marriage by 22% — not by rebuilding the economy or constructing schools, but apparently by making it safe enough for girls to walk to class (Perez-Parra et al.).
• Showing displaced Nigerian farmers a video about the climate crisis that forced herders onto their land reduced anti-herder prejudice across every dimension – compassion, trust, stereotypes, and policy support – even among people who rated their trauma a 9 out of 10 (Adejumo et al.). #RCT
• Protests do get you more money from Nigeria’s federal government – but only if your state governor is from the president’s party, and the extra cash comes through opaque VAT transfers where nobody’s watching (Archibong, Okafor, Osabuohien, and Moerenhout).
• A short scientific-reasoning primer makes Egyptian X users less likely to believe AND share fake news, while warning about social harm only reduces belief – so teach people to think, not just to fear consequences (Hassan). #RCT
• Conflict-hit communes in Burkina Faso budget ambitiously for reconstruction but the promised central government support never arrives, trapping local governments in a low-capacity equilibrium for years (Apeti, Lee, and Vincent).
• Ethiopia’s cash-plus graduation programme shields conflict-affected Amhara households from food insecurity but cannot protect their assets – social protection is a consumption lifeline, not a resilience silver bullet, in wartime (Caeiro, Sabates-Wheeler, and Justino). #RCT
• Ethnic violence in Sub-Saharan Africa makes people feel more ethnic and paradoxically more trusting of strangers, but the effect flips depending on whether your group was attacker or victim (Vysotskaya).
• Machine learning predicts African conflict from climate data with 96% recall, and causal methods show the same rainfall shock can pacify croplands or inflame forests depending on the landscape – context is everything (Foltz and Jing).
• A new Resource Volatility Index covering 193 countries shows commodity price swings erode governance and corruption control, especially in developing economies with weak institutions (Salihu, Ohiare, and Isa).
• Mourfou unpacks how conflict, military budgets, and foreign aid interact in the volatile Sahel – where the security crisis is reshaping both government spending priorities and donor aid flows.
• Tunisia’s 2016 suspension of drought relief declarations cost the incumbent 10 percentage points in drought-affected districts in the 2019 presidential election, with the mechanism running through increased riots and public discontent (Hashish).
• Central government transfers to Cameroonian municipalities spike during election years, with ruling-party-allied municipalities receiving disproportionately more grants, confirming that intergovernmental transfers serve as electoral instruments (Manga Amougou).
• A vignette experiment in Angola finds that vote-buying works only through tangible enforcement – monitoring turnout or questioning ballot secrecy – while promises of future clientelistic benefits are ineffective, with about 40% of respondents having experienced vote-buying (Smirnov et al.). #RCT

• The hardest part of targeting food aid in Egypt isn’t choosing between the poorest and the most impacted households — it’s that calorie goals, dietary diversity goals, and food security goals each call for reaching entirely different people (Mahmoud and Kurdi). #RCT
• South Africa’s old-age pension cuts energy poverty by 10 percentage points in urban areas — where the grid exists but is unaffordable — but does nothing in rural areas where the infrastructure simply isn’t there (Nkosi et al.).
• Ethiopia’s Productive Safety Net Programme boosts non-farm income by up to 75% and increases business ownership, making it a de facto climate adaptation tool for eight million rural participants (Weldegebriel et al.).
• Losing a parent reduces children’s self-reported well-being by 8.4% on average across Ethiopia, India, and Vietnam, with effects that intensify over time and hit rural and younger children hardest (Bahru et al.).
• Nigerian households that stay poor are not trapped by their past poverty but by enduring characteristics like low education and large family size – meaning cash transfers alone will not break the cycle without addressing root causes (Eigbiremolen).
• Large unconditional cash transfers in rural Kenya don’t just make households richer – they eliminate market friction distortions, pushing recipients into higher-capital activities like cattle and enterprises and actually increasing labour supply (Brimble, Garlick, and Orkin). #RCT
• Matta and Zizzamia apply portfolio theory to humanitarian aid and find the biggest bang for the buck in anticipatory action comes from the first improvements in forecasting accuracy – after which returns diminish sharply.
• A $150 bicycle delivers a 14.7x ROI in rural Zambia: recipients see 43% higher income, 24% higher consumption, and women gain even more – turns out the binding constraint was just getting there (Ahmed et al.). #RCT
• Combining cash transfers with business training for at-risk youth in Cote d’Ivoire during COVID-19 boosted food consumption by 39pp and microbusiness operation by 19pp (Kimou et al.). #RCT
• Peru’s JUNTOS CCT gets kids through secondary school and into formal jobs by age 22, but does not improve reading comprehension or boost earnings – with men pushed toward technical education and women delaying marriage (Woodman Deza et al.).
• A 20% aid cut at Kenya’s Kakuma refugee camp triggers a miniature financial crisis – collapsing credit, deflation, and welfare losses that dwarf the donor savings, showing aid economies are more fragile than we thought (Bruni and Sterck).
• In Mozambique, zero targeting effort is optimal for over 60% of social pension areas – simply expanding coverage in high-poverty rural areas would raise welfare by 13% more than refining who gets benefits (Jones et al.).
• An RCT in 180 Ethiopian villages shows community leaders facing scarce resources spread transfers thinner to minimize exclusion errors, and discretion helps them include conflict-affected households – but real stakes increase favoritism (Abay et al.). #RCT

• Drought drives up risky sex across Sub-Saharan Africa, but your ancestors’ kinship system matters – matrilineal communities show stronger behavioral responses yet paradoxically lower HIV prevalence, suggesting cultural institutions quietly shape health outcomes (Atta Aikins).
• Drought across 18 African countries doesn’t just destroy crops — it increases unintended pregnancies, with children from these births receiving less prenatal care and falling sick more often, and the burden falls hardest on women with no education (Ahmed et al.).
• Flooding in Zambia increases malaria and cholera in children under five by striking amounts, and it is not just a rural problem – urban wards with their poor drainage are equally hit (Ng’andu).
• Drought in Nigeria hits girls hardest, reducing their weight-for-age by 0.5 SD – nearly double the effect on boys – with maize prices spiking 66% and subsistence households bearing the brunt while livelihood diversification offers no buffer (Okoror).
• Unregulated pumping from Morocco’s Haouz-Mejjate aquifer destroys 60% of potential welfare, but a smart stock-dependent water tax can close nearly all the gap while costing farmers only 10% of profits (El Mansoum et al.).
• Giving Dakar shop owners vouchers for flood barriers works – but only when they coordinate: uncoordinated cement barriers divert water onto downhill neighbors, increasing THEIR flood closures (Doruska). #RCT
• Cyclone Ana hit Malawian farmers hard – 12-13pp increase in reported income losses – and pre-existing NGO and government transfers did nothing to buffer the shock (Grigolli Cesar).
• Pakistan’s 2022 floods cut wholesale agricultural supply by 65% and spiked prices 49%, but the bottleneck was NOT transport – truckers rerouted around flooded roads, while government price ceilings took months to catch up (Asad et al.).
• Republic of Congo’s biggest-ever road expansion boosted economic activity but at the cost of significant forest loss in the world’s second-largest rainforest (Martorano et al.).
• When the FSC tightened rules to protect intact forests, logging firms simply stopped certifying rather than complying – and deforestation continued unabated, a cautionary tale about the limits of voluntary environmental certification (Houngbedji et al.).
• Radio news clips in rural Tanzania dramatically boost climate change awareness, but here is the twist: telling people climate change is caused by China and the US actually DECREASES their willingness to protect the local environment compared to emphasizing domestic causes (Emmanuel et al.). #RCT
• A community-driven approach to flood adaptation in a Mozambican city – meetings, expert info, and seed grants – reduced water levels and improved walkability at flood-prone points after just one rainy season (Caetano et al.). #RCT
• Protected areas across 44 developing countries boost household wealth and cut child labour by pushing families out of farming into less resource-intensive livelihoods, but the gains depend heavily on how rich you were to start (Nikiema).
• Migrants in Madagascar are scapegoated for deforestation, but census and satellite data show they are no more likely to clear forests than locals – the blame narrative does not survive empirical scrutiny (Rakotonarivo et al.).
• Mineral export bans do boost basic metal exports in developing countries, but the industrial upgrading stops there – no evidence they help countries climb into complex manufacturing, deflating a popular policy dream (Chen).
• Across 13 West African countries, individual characteristics and rural livelihoods – not just poverty traps – explain why most climate-vulnerable people prefer to stay rather than migrate, challenging the conventional ’trapped populations’ narrative (Bekaert et al.).

• Zambia has no subnational resource curse – instead, the colonial railway built over a century ago still determines who is wealthy, with the Copperbelt’s mines and rail creating an irreversible spatial equilibrium (Harrison).
• The rollout of 3G/4G broadband across Ethiopia’s maize markets cuts price dispersion by 2.6%, with the biggest gains between distant markets where information was scarcest – internet does what roads alone could not (Kassa et al.).
• Using 21 million firm-to-firm transactions in Zambia, a 1% increase in copper production boosts domestic procurement by 0.5%, though Chinese-owned mines generate significantly fewer local linkages (Benshaul-Tolonen and Fernandez Musso).
• A cheap currency gets Egypt into global value chains, but climbing up them requires something devaluation can’t provide — access to the foreign knowledge embedded in imported inputs (Abdou).
• Africa’s leapfrog from agriculture to services — bypassing industrialization — is actively worsening trade balances, because only the industrial sector generates the export competitiveness and import substitution needed to close the gap (Sileshi and Salisu).
• Two decades of African urbanisation have delivered essentially flat public goods provision, with network infrastructure like electricity and piped water lagging furthest where cities sprawl most (Harman).
• Living near aid projects in Mali helps borrowers repay loans on time but does not boost how much they borrow — and only grant-financed projects (not concessional loans) deliver consistently positive effects (Mader).
• Forget private adoption rates — when Zambia’s grid reaches a village, electrified schools attract younger, better-qualified, more female teachers and enrollment jumps, while electrified clinics vaccinate more children, suggesting the social value of rural electrification through public infrastructure alone may justify the World Bank’s $30bn ‘Mission 300’ (Figueiredo Walter et al.).
• Doubling Aid for Trade to African countries reduces trade costs by 2-3.6%, with infrastructure and trade policy interventions driving the effect while productive capacity aid shows no impact, and lower trade costs in turn correlate with increased trade and reduced poverty (Ayele, Mendez-Parra, and te Velde).
• Support for free trade across 35 African countries is shaped more by national political and economic context than by regional classification, with only 12% of citizens aware of the AfCFTA, suggesting its success depends on country-level communication rather than continent-wide messaging (Beecham and Yakubu).
• The Chinese New Year causes significant but heterogeneous trade contractions across 26 African economies, hitting low-income coastal non-resource states hardest and peaking during the 2022 post-COVID recovery, suggesting African policymakers should treat this predictable calendar shock as a meaningful macroeconomic event (Odu, Akadiri, and Okafor).
• Broader and more balanced trade linkages sustain economic expansions while deeper geoeconomic connections speed recovery from recessions globally, but these relationships weaken for Sub-Saharan Africa alone, suggesting trade’s growth-sustaining role depends on complementary structural factors (Qu, Dabalen, and Calderon).
• Water infrastructure matters most for African economic performance – more than electricity, ICT, or transport – but the returns vary dramatically by region, with East Africa gaining from ICT and Central Africa from water (Boukhaled and Al-Mouksit).
• FDI clustering raises green productivity in African manufacturing, but the devil is in the origin: Chinese FDI boosts efficiency in less industrialised countries while non-Chinese FDI reduces it in more industrialised ones (Gonhi and N’Drin).
• Paved roads – not just improved ones – doubled non-agricultural employment in rural Mali, but the structural transformation happened in remote villages catching up, while near-city areas just got bigger without changing what people do (Mesplé-Somps et al.).
• Morocco’s Special Economic Zones do boost regional industrial growth – especially exports – but the payoff varies widely across zones, underscoring that the zone itself is only as good as the complementary policies around it (Znagui).
• Non-tariff measures reduce African firms’ export values by up to 0.17% per additional standard, hitting small firms hardest, but high-quality manufacturing firms actually benefit from compliance (Fiankor, Kassa, and Lartey).
• Younger, smaller, and foreign-owned firms in Vietnam are systematically less likely to use letters of credit for trade, and firm-level variation dominates country-product variation in explaining trade finance use (de Nicola et al.).
• Vietnam’s Greater Mekong Subregion highway investment sparked polycentric industrial development, with peripheral rural areas gaining more firm entry and employment than urban cores once market access crossed critical thresholds (Nose and Sawada).
• Across Nigerian states, a 1% rise in fiscal capacity increases solar projects by 7.5%, but oil-producing and older states adopt far less solar energy, revealing how resource wealth and institutional rigidity create compound governance barriers to the energy transition (Akeredolu).
• Mangrove restoration in the Philippines draws poor households into fishing but also shifts diets from fish toward meat and eggs, improving protein diversity for both fishing and non-fishing households in treated villages (Pagel et al.).
• Tariff reductions across 20 Sub-Saharan African countries from 2000-2020 significantly increased access to clean cooking fuels and technologies, with renewable energy, remittances, and institutional quality also helping to alleviate energy poverty (Zahonogo).
• The value-added maximising rules of origin for African exports to the EU is just 21% domestic content – half the level for all developing countries – and the EU’s current requirements are too restrictive for Uganda and Kenya, reducing rather than increasing local value-added (Ticku et al.).

• Your neighbor’s war is your fiscal problem: conflict in one African country significantly increases debt burdens in neighboring states through military spending spillovers, with the strongest cross-border effects in West Africa (Adeosun).
• The much-discussed ‘African risk premium’ on sovereign bonds is actually quite modest once you control for fundamentals — Africa’s real financing challenge is underdeveloped domestic debt markets as countries try to shift away from risky foreign-currency borrowing (Alter et al.).
• Nigeria’s 2017 diaspora bond wasn’t just a fundraising gimmick — it lowered sovereign yields by 1.2 percentage points and tightened spreads by 148 basis points, signaling to markets that the diaspora’s willingness to invest is a credible vote of confidence (Musselwhite).
• Monetary policy works for household credit in Uganda, but banks stuffed with government bonds pass rate hikes on more aggressively — and it’s low-income borrowers who bear the brunt of this sovereign-bank nexus (Namwanja et al.).
• A one-time offer of easier credit for water tanks gave Kenyan dairy farmers a seven-year head start that permanently boosted milk sales by 10% and increased girls’ school attendance, even after control farmers caught up in tank ownership (Deutschmann et al.). #RCT
• High inflation spells in West Africa’s WAEMU zone typically burn out in under two years, vindicating the BCEAO’s credibility – except when Covid hit (Gammadigbe).
• Uganda’s adoption of inflation targeting coincided with GDP growth slowing from 7.4% to 4.9% – Kiberu et al. use DSGE and SVAR models to assess whether the sacrifice ratio increased, providing rare evidence on IT’s costs in a low-income economy.
• Printing money in South Africa temporarily shrinks inequality – seigniorage transfers help the young while inflation taxes the old and wealthy – but push it too far and you get instability, especially if the economy becomes more equal (Getachew, Matshaka, and Viegi).
• Central banks that target inflation also shield their countries from energy vulnerability – mainly by boosting GDP and taming price volatility – with the biggest payoffs in emerging economies (Nomo Beyala and De Soko).
• Production shocks cascade through guarantor networks in a Kenyan financial cooperative, with each shock triggering more than one additional default among connected borrowers, though these contagion effects are short-lived (Brimble, Carney, and Deutschmann).
• When countries sign China’s Belt and Road Initiative, developed-market investors cut equity holdings by 11% and reallocate toward US-aligned low-income economies – the BRI acts as a geopolitical shock that reprices risk along alignment lines (Lochner).
• When commodity prices spike, 20 cents of every dollar in extractive MNE windfall profits gets booked in tax havens – meaning resource-rich developing countries are losing a big chunk of the gains that are supposed to fund their development (Chiocchetti and Moreau-Kastler).
• When China’s 2014 Budget Law reform tightened local fiscal constraints, cities pivoted from subsidies to government venture capital – and the shift actually improved entrant quality, boosted innovation, and crowded in private investment (Cheng, Xu, and Yang).
• Noah examines how state fragility in Africa undermines both inclusive growth and financial inclusion, exploring the vicious cycle where weak institutions lock out the populations that need financial services most.
• A macro model calibrated to South Africa shows a 5 percentage point tax cut would boost output by 11% and reduce income inequality by 3%, but increase consumption inequality as government transfers decline (Getachew).
• Zimbabwe’s monetary policy distortions – money printing, overvalued exchange rates, and resulting informalization – cost the Treasury over $3 billion in lost tax revenue between 2020 and 2023 (Steenbergen et al.).

• After using AI to catalog how 355 papers use rainfall as an instrument, Michler et al. test 14 common rainfall metrics across six African countries and find that most become weak instruments once you add household fixed effects — a quiet indictment of a workhorse strategy in development economics.
• Malawian farmers report ‘drought’ for both too little and too much rain – and the standard full-season rainfall measure economists love explains maize yields worse than just asking communities what happened (Stern).
• Current global multidimensional poverty measures are theoretically promising but practically flawed – Decerf and Fonton develop a preference-consistent alternative that, applied to Nigeria, yields substantially different state-level well-being rankings than monetary poverty alone.
• Global inequality fell since 1990, but who drove it? Moramarco and Sterck develop a new decomposable measure and find the answer is large middle-income countries (India, China) and poor countries (DR Congo, Ethiopia) – while in the US, rising inequality is a within-state phenomenon.
• Standard cross-sectional surveys understate food insecurity by 24 percentage points – tracking 17,000 households across six African countries, Wollburg et al. find 38% cycle in and out of hunger, and these transient cases are geographically dispersed, making them nearly impossible to target.
• Phone surveys – the backbone of pandemic-era data collection – systematically distort responses by up to 63% compared to in-person interviews in rural Nigeria, with a median mode effect of 17-18% driven by lower cognitive engagement (Castaing et al.). #RCT
• LLMs outperform human enumerators at classifying Zambia’s Labour Force Survey responses into occupation and industry codes by 10-17 percentage points, with GPT-5 Nano doing it for under $10 per annual national survey (Rossow et al.).
• Want to ask people about sensitive stuff like irregular migration? The ‘ballot-bag’ – respondents drop anonymous answers in a transparent bag – is as unbiased as a list experiment but way more precise, letting you actually detect treatment effects in Egypt (Crépon, Elsayed, and Gazeaud). #RCT
• The balance tests used in 90% of published RCTs systematically over-reject the null – Kerwin et al. show that switching to randomization inference fixes the problem and reveals that two prominent RCTs flagged as imbalanced were actually fine.
• Re-analyzing 780 treatment effects from published RCTs, PDS Lasso reduces standard errors by less than 1% versus standard ANCOVA and selects no treatment-predicting variables in over half the cases (Cilliers, Elashmawy, and McKenzie).
• Multiple imputation methods using small benchmark soil surveys of 350-750 plots can accurately reproduce soil quality distributions in Ethiopia and Uganda at a fraction of the cost of full-scale objective measurement (Dang et al.).
• Low-cost in situ weather sensors in Uganda match professional stations far better than satellite products for measuring rainfall, and the choice of weather data source meaningfully changes estimated crop yield effects (Paolantonio et al.).

• Women in sub-Saharan Africa flock to Evangelical churches not for charity but for protection — when economic opportunities arrive, churches that shield successful women from witchcraft accusations become engines of female empowerment in Benin and beyond (Alvarez-Aragon et al.).
• Each year a Bangladeshi girl’s marriage is delayed increases her food consumption quality by 6-12% decades later, with gains driven by better dietary diversity and stronger bargaining power within the household (Schmidt).
• In China, children’s own aspirations at age 14-15 predict their educational attainment two years later more strongly than their parents’ aspirations – and the relationship is non-linear, with very high aspirations potentially backfiring (He, Janssens, and Lanjouw).
• Rwanda’s stunning contraceptive uptake wasn’t bottom-up empowerment – community meetings only boosted adoption after leader performance contracts kicked in, and some women got pregnant faster to dodge the pressure (Linek and Stryjan).
• Enlisting religious leaders in Ghana’s water utility campaigns nearly doubled the reduction in bill arrears and tripled water conservation effects compared to utility-only outreach, leveraging the ‘gospel water’ approach (Amoah et al.). #RCT
• In Tanzania, mothers allocate 10% more to children than fathers at baseline, but cash transfers and parenting programs shift fathers’ spending enough to close the gap entirely – suggesting fathers are the more policy-elastic margin (Almas et al.). #RCT
• A health information campaign in rural Guinea-Bissau that integrated traditional health practitioners and acknowledged witchcraft beliefs reduced child malnutrition and improved health knowledge, especially among those with the strongest traditional beliefs (Alvarez-Pereira et al.). #RCT
• A bibliometric analysis of 168,000 African publications finds women scientists represent just 31% of the continent’s researchers, with persistent underrepresentation in STEM fields despite growth in environmental science and sustainability research (Ibourk et al.).

• Women business owners across five African countries file and pay taxes at similar rates to men, but comply for entirely different reasons — fear of reputational harm and enforcement rather than trust in institutions — suggesting tax systems need gender-sensitive reform even when headline compliance numbers look equal (Mascagni and Iyengar Srivatsa).
• A two-day stats training for civil servants in the DRC boosts test scores by 1.1 SDs and the gains stick four months later – but trained officials are actually less likely to read follow-up materials, showing knowledge alone does not change behavior (Megha-Bongnkar, Metzger, and Gunther). #RCT
• Zambian firms wait 700 days for VAT refunds – 2,300% longer than required – and when the money arrives they double investment, revealing how the tax system itself throttles firm growth (Overbeck et al.).
• Tax nudges among 60,000 Kampala firms produce precisely zero effect – because when only 16% comply and enforcement probability is near zero over a decade, a threatening letter is just another piece of paper (Dietrich et al.). #RCT
• Corruption kills tax morale across Africa, but citizens in resource-rich countries care less about graft because they bear less of the tax burden – one-size-fits-all revenue strategies will not work (Iddrisu et al.).
• Firms in South African municipalities narrowly won by the ANC face weaker tax audits and report lower tax liabilities despite higher sales – political alignment quietly hollows out enforcement during the state capture era (Chingwere, Nsabimana, and Sen).
• Kenyans place six times more social welfare weight on formal taxpayers than identical informal workers, and incorporating this horizontal inequality aversion into optimal tax models explains the large exemptions developing countries actually use – challenging calls to broaden the tax base (Eldrup).
• Comprehensive tax audits in Uganda cause 11% of audited firms to shut down and surviving firms to cut sales by 56%, costing the revenue authority $3.6 million annually as the economic distortions exceed any revenue gains (Henning and Okello Ayo).
• Cameroon’s mobile money tax cuts agent commissions, especially for high-volume operators in Yaounde, pushing agents to charge extra fees and diversify activities – threatening the agent network critical to financial inclusion in sub-Saharan Africa (Noah and Tacneng).
• Voluntary digital tax payments in Eswatini slash late payments by 19-20pp, but mandating everyone to go cashless produces a more modest 2.1% revenue bump – early adopters and reluctant converts are very different taxpayers (Santoro, Masuku, and Magongo).

That is ~200 papers in three days. The graduation model got its most thorough stress-test yet – and survived, though with some uncomfortable questions about which components actually matter. Climate papers dominated, from floods to droughts to forests. And the methods papers reminded us that how we measure things matters as much as what we find.