The authors created a summary of the latest research in economics on Africa on the following topics: agriculture, conflict, credit, savings, risk and insurance, education, electricity access, firms, health and nutrition, households and networks, institutions, labor, political economy, poverty and inequality, and using evidence to inform policy. This summary represents the round-up of the centre for the study of African countries’ (CSAE) conference on Economic development in Africa, held in Oxford on 18 March. The full program and links to most of the papers is available here.

Agriculture

• Offering farmers harvest time loans increase grain storage to arbitrage price fluctuations. Farm profits increase, generating a 28% return on investment. General equilibrium effects show that non-recipients also benefit from smoother prices. #RCT #Kenya (Burke et al)
• Multiple equilibria driven by the intersection of supply and demand constraints help explain low adoption of a cowpea storage technology. Theory and evidence from farmer and trader survey data from #Niger. (Aker and Dillon)
• It’s not flat. It’s not inverse. The relationship between agricultural productivity and plot size is direct-inverse-direct. (Savastano) #Ethiopia
• Correlated measurement errors affecting variables used to construct ratios can bias ratios frequently used to characterize smallholders’ agricultural production severely. (Abay et al.) #Ethiopia
• Talking about agricultural labor, the listing bias and recall bias might be two sides of the same coin and can cancel each other out. (Gaddis et al.) #Ghana
• DNA fingerprinting maize samples in #Uganda shows that self-reported variety identification is highly inaccurate and that seed purity levels are very low. (Kilicet al.)

Conflict

• Sanctions targeted at specific mineral mines through the Dodd Frank Section 1502 were ineffective in decreasing sexual violence in #DRC. (Foltz & Sambo)
• Exposure to the conflict in 2013 masked substantial positive impacts of a multipronged development initiative working with adolescent girls in #SouthSudan. (Buehren et al.)
• Social capital measured by association membership increased in areas exposed to recent incidents of violent conflict in #Mali which, however, might just represent a withdrawal of Malians into their own groups’ organizations. (Calvoet al.)
• Individuals affected by armed conflict in Northern #Uganda show a higher risk and loss aversion. (Kijima et al.)

Credit, Savings, Risk & Insurance

• An unexpected regulatory reform that made it harder to buy insurance on credit reduced moral hazard. Theory and evidence from administrative data on car insurance contracts in #Ghana. (Annan)
• A labeled mobile money account for savings (together with some behavioral nudges) reduced risk-sharing among vulnerable women, but that didn’t affect their capacity to manage risk. #Kenya #RCT (Dizon et al)
• Exposure to extreme shocks (as perceived by the households) is associated with an increased probability of being poor and consuming less. Panel data evidence from #Niger. (Annan and Sanoh)
• Both flexibility and joint liability can help microfinance borrowers to cope with shocks in an experimental setting. (Czura, John & Spantig)
• Borrower performance measured by missed payments improved after a microfinance institution in #Pakistan moved from individual loans to a group-based joint liability lending scheme. (Mahmud)
• Could an endowment effect hinder savings? Not in this lab setting, but maybe in the real world? #Lab #Philippines (Spanting)
• Disbursing microfinance loans through mobile money accounts can increase profits and business assets. #RCT #Uganda (Riley)
• Expanding access to mobile money accounts among women microentrepreneurs increases mobile savings, and business training bolsters this impact. Treated women open profitable secondary businesses, have greater say, and are happier. #RCT #Tanzania (Bianchi et al)
• Investment reduces the need for informal risk sharing, weakening risk-sharing ties, thus limiting amount of borrowing that can be sustained. Theory and evidence from an #RCT in #Bangladesh. (Advani)
• Informal risk sharing deters demand for index insurance but does not affect indemnity insurance take up. Theory and #Lab evidence from rural #Uganda. (Perez-Viana et al)
• Theory informing policy: disaster risk finance instruments can mitigate commitment problems, such as the “Samaritan’s dilemma”. (Clarke and Wren-Lewis)
• Delayed “pay-after-harvest” premium increases farmers’ demand for index insurance, especially so when delivered through local informal institutions. #RCT #Ethiopia (Belissa et al)

Education

• In India, high temperatures reduce both math and reading test scores. But why? Hot days during the growing season reduce agricultural yields and test scores with comparatively modest effects of hot days in the non-growing season. (Garg, Jagnani, & Taraz)
• Can #cashtransfers improve education for refugees? For displaced Syrian children in Lebanon, school attendance increased by more half a day each week. (De Hoop, Morey, & Seidenfeld)
• During #SouthSudan’s civil war, schools located in the war zones lost on average 85 children per year, which represents 18.5 percent of total enrollment. (Mayai)
• Mother-tongue instruction in early grades increases longer-term labor outcomes, but the size of effects starts to decrease after grade 4 (Seid). #Ethiopia
• Compulsory high schooling slightly reduced youth crime through incapacitation, but not effective for poor kids due to over-crowding in classes (Nishijima & Pal). #Brazil
• Learning can be more intuitive! 1 SD learning gains is about 6.8 equivalent years of schooling under business-as-usual conditions in developing countries (Evans & Yuan).
• Mobile money facilitated intra-household transactions on educational expenditure even in the event of a negative shock (Tabetando & Matsumoto). #Uganda
• The opportunity cost of schooling expenditure is the major constraint on secondary schooling, while the opportunity cost of time significantly impedes tertiary schooling (Afoakwah, Deng & Onur).#Ghana
• Allowing women to have a second child without having to pay a fine increased schooling by 0.7 years (Raiber). #China
• Exposure to PROGRESA during lower secondary school reduced the probability to transition to upper secondary school” (Wiegand). #Mexico
• Female students are more likely to enroll in and perform better in universities than their male peers even controlling for prior performance. Although females have less access to fields such as engineering and computer science, they are less likely to drop out in any field once they are enrolled (Broekhuizen&Spaull). #South Africa
• Six years after allocating low-cost private school vouchers, initial results show that the program had a negative impact on students’ performance in Hindi and no impact on math or English (Crawfurd, Patel and Sandefur). #India
• Providing interactive smartboards in grades 1&2 classrooms had a positive impact on student scores of ONLY urban students, and the impact sustained when students progressed to the 3^rd grade (Lehrer & Mbaye). #Senegal
• Using Pasec data, the papers shows that students from the Francophone side performed better in math in grade 2 but worse in grade 5, and the difference cannot be explained by selection or diff in repetition rates (Bekkouche & Dupraz). #Cameroon
• The introduction of a rural hardship allowance (20% of salary) in 2010 increased teacher mobility but no significant effects on teacher characteristics, teacher attrition or student grades (Chelwa et al.). #Zambia
• Teacher performance pay based on students’ proficiency of grade-specific skills benefited a wider range of students than the pay based on students’ rank, which led teachers to focus on the best students (Mbiti et al.). #Tanzania #RCT
• Low teacher content knowledge can account for a substantial portion of the large learning gap and moving students from the worst performing teachers to teacher with relatively better knowledge, will raise average student learning (Bold et al.). #Sub-Saharan Africa
• Like mother, like child. A national school building program in India in the 1990s increased school opportunities for women. Children whose mothers were impacted by this program had higher cognitive abilities as measured by scores on standardized math, reading and English tests… The impacts on test scores for girls is between 13-32 percentage points higher than the corresponding effect for boys. (Sunder)
• What are the benefits of preschool attendance in Africa? Consistently smarter kids in later grades in both #Kenya and #Tanzania. (Bietenbeck, Ericsson, & Wamalwa)
• All I really need to know I learned in…second grade math class. In #Madagascar and #Senegal, cognitive skills as measured using test scores in the second grade, are strong predictors of school attainment and cognitive skills in young adulthood in both countries. The authors also find that second grade math scores more strongly predict adult cognitive skills in comparison to French scores. (Kaila, Sahn, & Sunder)
• Drawing on an education reform in #Zimbabwe, Agüero & Ramachandran find large and robust intergenerational transmissions of education across generations. Evidence of assortative mating suggests that the marriage, rather than the labor, market is a key mechanism for these transmissions.

Electricity Access

• What do electricity shortages in Africa do? They adversely impact employment – through creation of new businesses, employment in existing firms, and reducing export competitiveness. Employment impacts are particularly bad for women. (Mensah)
• Why are electricity connection charges so high in Africa? Regulation on usage price (cost-per-kilowatt) means that utilities actually lose money on new customers. Why the regulation? Visible benefit to households with an existing connection. Hidden costs to households with no connection. (Blimpo, McRae, & Steinbuks)
• There is vast evidence that economic well-being must precede high take up of electricity. At the same time, to increase take-up among the poorest, electrification efforts may be more successful if bundled with programs to help households acquire appliances for productive enterprises. (Blimpo & Postepska)
• Access to electricity has a strong positive impact on tax morale across Sub-Saharan Africa, based on Afrobarometer data and an instrumental variables approach. Grid extension to a community improves the willingness to pay taxes even for unconnected households. (Blimpo et al.)

Firms

• Labor and material inputs are core drivers of output in the Ethiopian garment and textile industry while capital’s contribution is negligible. (Tekleselassie et al.)
• Misallocations in input markets are more severe than in output markets in #Ghana and removing these distortions could increase TFP by up to 59%. (Ackah, Asuming & Abudu)
• What is the impact of a simple Kaizen business training for firms in Tanzania? Improvements in management capacity and profits for groups that had both class and onsite training in the more medium run. #RCT in #Tanzania and #Vietnam (Higuchi)
• Business training improves outcomes for more established agribusinesses in Ethiopia.#RCT #Ethiopia (Bastian)
• Video with a nudge to change aspirations leads to higher business profits for firms, but adding goal setting to the video leads to negative impacts on profits. Business skills training don’t have any impact. #RCT #Mozambique (Seither)
• Management practices are associated with higher firm productivity in #Vietnam. (Hoang)

Health and nutrition

• Training to improve the management of health centers in Nigeria didn’t lead to enduring effects, but supportive supervision improved practice. (Dunsch et al.) #RCT
• “In 2015, Ethiopia experienced one of its worst droughts in decades.” Child malnutrition grew worse, but only in areas with “limited road” access. Infrastructure saves lives! (Hirvonen, Sohnesen, & Bundervoet)
• When floods hit Malawi, crop yields fell. Social safety nets protected some households, but higher coverage could have done much more. Other risk-coping strategies – such as financial account ownership, access to off-farm income sources, and adult children living away from home – didn’t help much. (McCarthy et al.)
• Why are kids getting fatter? In Indonesia, exposure to television during childhood increases consumption of soft drinks and snacks during adulthood. Results in higher body-mass index for boys (but not girls). (Oberlander)
• Adoption of modern crop seed varieties led to a reduction in infant mortality and severe stunting but impact concentrated among boys. Effect could be linked to a biological effect since boys are less robust in the womb? #Diff-in-diff geospatial and DHS data 36 countries. (McCord et al.)
• Adverse effect on child health through increased maternal stress and reduction in prenatal care use during a time of increased criminal violence. Exploiting a spike in the homicide rate from a crackdown on drug trafficking in #Mexico. (Nasir)
• Gains in decision making power for women through the receipt of old age pensions in #South Africa leads to lower child deprivation. Panel data analysis using the actual take-up of pensions. #RDD (van Biljon)
• No impact on nutrient intake from cash transfer programs in #Kenya and #Lesotho #Diff-in-diff RCT (Schmerzeck)
• What intra-class correlation (ICC) should you use in power calculations for new studies? Referencing cash transfer program datasets in #SSA #methods. (Seidenfeld et al.)
• Children have poorer nutritional outcomes (height-for-age) in polygamous households than in monogamous households in #Nigeria. (Owoo)
• A number of countries have legalized sex work but also provided certification for sex workers which verifies health. In a free certification experiment in Senegal, take-up was extremely low. Why? Health status seems to be partially observable, so the gains to certification are lower. (Manian)
• How much do “psychic” costs of vaccination reduce take-up? 13 percentage points, in rural Nigeria. (Psychic costs are “psychological barriers, such as the fear of needles or the concern for vaccine safety.) (Sato)
• How well can social incentives (e.g., a colorful bracelet) help individuals to undertake health investments with big spillovers? Some incentives work, some don’t, but a nudge (in the form of a text message) made a big difference in uptake of deworming medication in Kenya. (Karing & Naguib)

Households and networks

• Men who believe in the supernatural powers of menopausal women in voodoo ethnic groups could afford her with significant decision-making power. #Diff-in-diff in #West Africa. (Alidou, Verpoorten)
• Differences in sibling pairs’ relative time and consumption allocations are driven by their preferences rather than differences in parents’ altruism. #Panel data from #Ethiopia/India/Peru/Vietnam. (Kirchberger)
• Could communication improve intra-household allocative efficiency? Evidence from an in-the-field lab experiment show couples do not behave optimally even with full information. #Kenya (Castilla, Nasuda & Zhang)
• Does insurance motivate some of the large monetary contributions to the church? #RCT evidence that the church acts as an insurer (materially and spiritually) in #Ghana. (Panin et al.)
• Why do community reporting mobile platforms in #Uganda get adopted by some villagers and not others? Networks help but not always. Higher externalities mean lower network effects on adoption unless communities can secure truthful communication. #RCT (Grossman)
• Reduced form estimates are biased, even with valid IVs, in the presence of strategic interactions (if compliance is imperfect for example). So you need to understand and model strategic interactions. [read this if you are doing random saturation]. #Methods (O’Keefe-O’Donovan)
• If your lab in the field experiment participants can communicate beforehand, they’ll change their behavior. Communication across villages on a lab in the field experiment in #Rwanda led to later participants changing behavior. Villages with pre-treated neighbors contribute 10-15% more! #Lab in the field. (Coutts)
• Does rivalry affect diffusion? Yes. How: central people share less when rivalry binds. #Lab in the field. (Viela)
• Business training increases bargaining power and income in Vietnam. And leads to more domestic violence. #RCT # Vietnam (Lensink)
• Women in Pakistan prefer to run a business from home, household members agree, and access to finance doesn’t change these preferences. #RCT #Pakistan (Said)
• Bride price is driven by characteristics of the wife but don’t correlate with her access to household resources. On the other hand, the gift the husband gives her at marriage is. #Senegal (Hotte)
• A secondary school stipend (with a condition on not being married) in #Bangladesh led to higher education, no change in fertility or female labor force participation but higher earning husbands. (Tanaka) #RD
• How is smallholder competitiveness affected by rapid market integration via competitive large-scale farms in #Zambia? Smallholders near large farms play more competitively, with no gender gap in villages close to large farms but a significant gap for villages farther away. (Sipangule) #lab-in-the-field
• Do social networks promote mobile money usage for financial inclusion in sub-Saharan Africa? Yes. (Lwanga)
• Is it better to get cash transfers quarterly or monthly? Turns out it doesn’t matter at all for nutrition, consumption or investment outcomes. And cash transfers increase women’s labor force participation and income. #RCT #Nigeria (Papineni)
• A cash transfer microenterprise program led to an increase in consumption and assets. #Uganda #RCT (Sedlymayr)
• Shocks from conflict lead to reduced household food consumption by 17 percent and they cope through remittances and borrowing. #Nigeria (Nwokolo)
• The same people have demand for savings and credit programs simultaneously, but they really don’t want commitment products. And the finance has no effect on household welfare, but a small effect on empowerment. #RCT #lab-in-the-field #Pakistan (Quinn)

Institutions

• How do colonial legacies affect African cities? Based on a sample of 318 cities across 28 countries using satellite data on built cover over time, Anglophone origin cities sprawl compared to Francophone ones. (Baruah, Henderson, & Peng)
• Across Sub-Saharan Africa, ethnic diversity is associated with lower provision of most public goods (schools, roads), but not all: The electrical grid and cellphone coverage are linked to MORE diversity. (Clochard & Hollard)
• In order to improve record-keeping in health facilities in #Nigeria, inspectors rated record-keeping and posted stars on the wall. It worked in one state, not in another. Differences aren’t explained by observed characteristics across states. (Gauri et al.) #RCT
• Previous work has suggested that Christian missions positively affected later economic development. New, richer data from #Ghana suggests that these results are spurious, as early missions targeted “more pacified and non-malarial areas.” (Meier zu Selhausen, Moradi, & Jedwab)
• That survey question about trust in general? It doesn’t reflect trust in institutions. The good news is a strong correlation between survey responses on trust in institutions and the results of games. Evidence from Ethiopia. (Carlsson et al.)
• Among youth groups in Ethiopia, more risk tolerant groups are also more trustworthy. (Holden & Tilahun)
• How much is your intervention affected by the fact that an NGO implemented it? In India, communities with previous exposure to an NGO adopted a new cooking technology at much higher rates. (Jeuland, Pattanayak, & Usmani)

Labor

• 39 percent of the world’s population speaks a “gendered” language (i.e., nouns are masculine or feminine). This is strongly associated with lower women’s labor force participation, more sexist attitudes, less schooling for girls. (Jakiela & Ozier)
• Do women matter more for intergenerational mobility of children? Model and empirical results from Nigeria. Gender inequality in human capital investments comes from parental psychic costs/pessimism about the state of the world. (Azomahou)
• Bank cards for cash grants in #South Africa increase women’s autonomy and their LFP. Effects are bigger for women in male-dominated HHs. #Diff-in-diff. (Pasha)
• Does discrimination from subordinates make women leaders less effective? Folks are 10% less likely to follow the advice of a female leader, but this is more than erased when women can signal ability. # Lab in the field #Ethiopia .(Sheth)
• Non-cognitive skills found to matter for labor market participation and wages in a developing country setting. #latent factor model. #Latin America (Leighton)
• Evidence that a combination of cash and skills training was successful at stimulating employment and entrepreneurship for urban youth in a labor market weakened by the Ebola crisis in #Sierra Leone. #RCT (Acavedo, Rosas & Zaldivar)
• Young people’s time allocation around ages 19 is entrenched in the past choices and economic behavior at age 12 and earlier. An analysis of the dynamic nature of children’s gendered time allocation in #Ethiopia. (Bruckauf & Chzhen)
• It may be time for a fundamental reassessment of approaches for addressing youth employment in low-income countries. A review of the evidence so far. (Fox & Kaur)

Political Economy

• In #Kenya, a choice experiment reveals that “politicians value each person served in their home village more than twice as much as each person served outside their home village.” An underlying model suggests that constitutional action to control home bias fares poorly in the presence of corruption. (Hoffman et al.)
• In #Nigeria, elected local government councils are better at promoting development than appointed caretaker committees. (Kyburz)
• Information from local polls improves voter turnout, standard get-out-the-vote message doesn’t. Voters update beliefs about party quality when their preferred parties lead polls. #RCT #SouthAfrica (Orkin)
• Reducing logistical costs improve voter registration, especially when combined with information and SMS messaging. Impacts stronger in poorer areas, among men and the youth. #Kenya #RCT (Harris)
• Investments in infrastructural public goods reduce public expenditure in neighboring jurisdictions. Consistent with idea that local governments free-ride in allocation of public resources. #Kenya #Diff-in-diff (Mbate)
• Children of ethnically distant mothers experience higher mortality rates and lower height-for-age scores. Household survey and spatial evidence from 14 African countries. (Gomes)

Poverty and Inequality

• Evidence suggests that the 2010 earthquake had long-lasting disruptive impact on households’ livelihood system and the assistance programs’ coverage was imperfect. (Saint-Macary & Zanuso). #Haiti
• Rainfall shocks in early life partly explain lower labor participation and lower expenditures in adulthood. Rainfall anomalies affected crop performance and had negative and persistent effects on human growth (0.5-0.7cm shorter height in adulthood) (Baez, Caruso & Niu). #Mozambique
• Examined the determinants of household resilience and used machine learning algorithms to identify predictors of food insecurity (Knippenberg, Jensen & Constas). #Malawi
• The poor continue to suffer more welfare losses because of drought but access to PSNP helps to mitigate (Fuje) #Ethiopia
• Relative wage matters more than absolute wage for the effort level exerted during a low skill task in an experimental setting. (Sseruyange & Bulte) #RCT #Uganda
• Wage differentials between firms versus within firms is an important source of inequality in #SA. (Chatterjee & Kerr)
• A relatively larger share of the service sector decreases gender inequality through increased female labor force participation in Africa. (Bittencourt, Clance & Zawaria)
• Watching a two-minute video on legal advances of women’s rights is not sufficient to change attitudes towards gender in #Tunisia. (Goedhuys et al.) #RCT

Using Evidence to Inform Policy

• Among other things, Rachel Glennerster (keynote) propose three style tips for engaging with policymakers on research evidence: (1) Make it easy and cheap for policy makers to apply it. (2) Salience matters: Hearing it once isn’t enough. (3) Framing matters: Use their language, not yours. Furthermore – in terms of substance – she underlines the importance of drawing on the full tool box: theory, descriptive data, contextual knowledge, and well-identified causal evidence. You can watch the video of her talk. We recommend it!