Machine learning and phone data can improve targeting of humanitarian aid.

The COVID-19 pandemic has devastated many low- and middle-income countries, causing widespread food insecurity and a sharp decline in living standards1. In response to this crisis, governments and humanitarian organizations worldwide have distributed social assistance to more than 1.5 billion people2. Targeting is a central challenge in administering these programmes: it remains a difficult task to rapidly identify those with the greatest need given available data3,4. Here we show that data from mobile phone networks can improve the targeting of humanitarian assistance. Our approach uses traditional survey data to train machine-learning algorithms to recognize patterns of poverty in mobile phone data; the trained algorithms can then prioritize aid to the poorest mobile subscribers. We evaluate this approach by studying a flagship emergency cash transfer program in Togo, which used these algorithms to disburse millions of US dollars worth of COVID-19 relief aid. Our analysis compares outcomes-including exclusion errors, total social welfare and measures of fairness-under different targeting regimes. Relative to the geographic targeting options considered by the Government of Togo, the machine-learning approach reduces errors of exclusion by 4-21%. Relative to methods requiring a comprehensive social registry (a hypothetical exercise; no such registry exists in Togo), the machine-learning approach increases exclusion errors by 9-35%. These results highlight the potential for new data sources to complement traditional methods for targeting humanitarian assistance, particularly in crisis settings in which traditional data are missing or out of date.

Tackling psychosocial and capital constraints to alleviate poverty.

Many policies attempt to help extremely poor households build sustainable sources of income. Although economic interventions have predominated historically1,2, psychosocial support has attracted substantial interest3-5, particularly for its potential cost-effectiveness. Recent evidence has shown that multi-faceted 'graduation' programmes can succeed in generating sustained changes6,7. Here we show that a multi-faceted intervention can open pathways out of extreme poverty by relaxing capital and psychosocial constraints. We conducted a four-arm randomized evaluation among extremely poor female beneficiaries already enrolled in a national cash transfer government programme in Niger. The three treatment arms included group savings promotion, coaching and entrepreneurship training, and then added either a lump-sum cash grant, psychosocial interventions, or both the cash grant and psychosocial interventions. All three arms generated positive effects on economic outcomes and psychosocial well-being, but there were notable differences in the pathways and the timing of effects. Overall, the arms with psychosocial interventions were the most cost-effective, highlighting the value of including well-designed psychosocial components in government-led multi-faceted interventions for the extreme poor.

Megastudies improve the impact of applied behavioural science.

Policy-makers are increasingly turning to behavioural science for insights about how to improve citizens' decisions and outcomes1. Typically, different scientists test different intervention ideas in different samples using different outcomes over different time intervals2. The lack of comparability of such individual investigations limits their potential to inform policy. Here, to address this limitation and accelerate the pace of discovery, we introduce the megastudy-a massive field experiment in which the effects of many different interventions are compared in the same population on the same objectively measured outcome for the same duration. In a megastudy targeting physical exercise among 61,293 members of an American fitness chain, 30 scientists from 15 different US universities worked in small independent teams to design a total of 54 different four-week digital programmes (or interventions) encouraging exercise. We show that 45% of these interventions significantly increased weekly gym visits by 9% to 27%; the top-performing intervention offered microrewards for returning to the gym after a missed workout. Only 8% of interventions induced behaviour change that was significant and measurable after the four-week intervention. Conditioning on the 45% of interventions that increased exercise during the intervention, we detected carry-over effects that were proportionally similar to those measured in previous research3-6. Forecasts by impartial judges failed to predict which interventions would be most effective, underscoring the value of testing many ideas at once and, therefore, the potential for megastudies to improve the evidentiary value of behavioural science.

A 680,000-person megastudy of nudges to encourage vaccination in pharmacies.

Encouraging vaccination is a pressing policy problem. To assess whether text-based reminders can encourage pharmacy vaccination and what kinds of messages work best, we conducted a megastudy. We randomly assigned 689,693 Walmart pharmacy patients to receive one of 22 different text reminders using a variety of different behavioral science principles to nudge flu vaccination or to a business-as-usual control condition that received no messages. We found that the reminder texts that we tested increased pharmacy vaccination rates by an average of 2.0 percentage points, or 6.8%, over a 3-mo follow-up period. The most-effective messages reminded patients that a flu shot was waiting for them and delivered reminders on multiple days. The top-performing intervention included two texts delivered 3 d apart and communicated to patients that a vaccine was "waiting for you." Neither experts nor lay people anticipated that this would be the best-performing treatment, underscoring the value of simultaneously testing many different nudges in a highly powered megastudy.

A call for structured ethics appendices in social science papers.

Ethics in social science experimentation and data collection are often discussed but rarely articulated in writing as part of research outputs. Although papers typically reference human subjects research approvals from relevant institutional review boards, most recognize that such boards do not carry out comprehensive ethical assessments. We propose a structured ethics appendix to provide details on the following: policy equipoise, role of the researcher, potential harms to participants and nonparticipants, conflicts of interest, intellectual freedom, feedback to participants, and foreseeable misuse of research results. We discuss each of these and some of the norms and challenging situations of each. We believe that discussing such issues explicitly in appendices of papers, even if briefly, will serve two purposes: more complete communication of ethics can improve discussions of papers and can clarify and improve the norms themselves.

A megastudy of text-based nudges encouraging patients to get vaccinated at an upcoming doctor's appointment.

Many Americans fail to get life-saving vaccines each year, and the availability of a vaccine for COVID-19 makes the challenge of encouraging vaccination more urgent than ever. We present a large field experiment (N = 47,306) testing 19 nudges delivered to patients via text message and designed to boost adoption of the influenza vaccine. Our findings suggest that text messages sent prior to a primary care visit can boost vaccination rates by an average of 5%. Overall, interventions performed better when they were 1) framed as reminders to get flu shots that were already reserved for the patient and 2) congruent with the sort of communications patients expected to receive from their healthcare provider (i.e., not surprising, casual, or interactive). The best-performing intervention in our study reminded patients twice to get their flu shot at their upcoming doctor's appointment and indicated it was reserved for them. This successful script could be used as a template for campaigns to encourage the adoption of life-saving vaccines, including against COVID-19.

Impact of savings groups on the lives of the poor.

Savings-led microfinance programs operate in poor rural communities in developing countries to establish groups that save and then lend out the accumulated savings to each other. Nonprofit organizations train villagers to create and lead these groups. In a clustered randomized evaluation spanning three African countries (Ghana, Malawi, and Uganda), we find that the promotion of these community-based microfinance groups leads to an improvement in household business outcomes and women's empowerment. However, we do not find evidence of impacts on average consumption or other livelihoods.

How can Bill and Melinda Gates increase other people's donations to fund public goods?

We conducted a fundraising experiment with an international development nonprofit organization in which a matching grant offered by the Bill and Melinda Gates Foundation raised more funds than one from an anonymous donor. The effect is strongest for solicitees who previously gave to other BMGF-supported, poverty charities. With supporting evidence from two other fundraising experiments as well as a survey experiment, we argue this is consistent with a quality signal mechanism. Alternative mechanisms are discussed, and not ruled out. The results help inform theories about charitable giving decision-making, and provide guidance to organizations and large donors on how to overcome information asymmetries hindering fundraising.

Being surveyed can change later behavior and related parameter estimates.

Does completing a household survey change the later behavior of those surveyed? In three field studies of health and two of microlending, we randomly assigned subjects to be surveyed about health and/or household finances and then measured subsequent use of a related product with data that does not rely on subjects' self-reports. In the three health experiments, we find that being surveyed increases use of water treatment products and take-up of medical insurance. Frequent surveys on reported diarrhea also led to biased estimates of the impact of improved source water quality. In two microlending studies, we do not find an effect of being surveyed on borrowing behavior. The results suggest that limited attention could play an important but context-dependent role in consumer choice, with the implication that researchers should reconsider whether, how, and how much to survey their subjects.

A Randomized Trial of Behavioral Nudges Delivered Through Text Messages to Increase Influenza Vaccination Among Patients With an Upcoming Primary Care Visit.

PURPOSE: To evaluate if nudges delivered by text message prior to an upcoming primary care visit can increase influenza vaccination rates. DESIGN: Randomized, controlled trial. SETTING: Two health systems in the Northeastern US between September 2020 and March 2021. SUBJECTS: 74,811 adults. INTERVENTIONS: Patients in the 19 intervention arms received 1-2 text messages in the 3 days preceding their appointment that varied in their format, interactivity, and content. MEASURES: Influenza vaccination. ANALYSIS: Intention-to-treat. RESULTS: Participants had a mean (SD) age of 50.7 (16.2) years; 55.8% (41,771) were female, 70.6% (52,826) were White, and 19.0% (14,222) were Black. Among the interventions, 5 of 19 (26.3%) had a significantly greater vaccination rate than control. On average, the 19 interventions increased vaccination relative to control by 1.8 percentage points or 6.1% (P = .005). The top performing text message described the vaccine to the patient as "reserved for you" and led to a 3.1 percentage point increase (95% CI, 1.3 to 4.9; P < .001) in vaccination relative to control. Three of the top five performing messages described the vaccine as "reserved for you." None of the interventions performed worse than control. CONCLUSIONS: Text messages encouraging vaccination and delivered prior to an upcoming appointment significantly increased influenza vaccination rates and could be a scalable approach to increase vaccination more broadly.

Health education for microcredit clients in Peru: a randomized controlled trial.

BACKGROUND: Poverty, lack of female empowerment, and lack of education are major risk factors for childhood illness worldwide. Microcredit programs, by offering small loans to poor individuals, attempt to address the first two of these risk factors, poverty and gender disparity. They provide clients, usually women, with a means to invest in their businesses and support their families. This study investigates the health effects of also addressing the remaining risk factor, lack of knowledge about important health issues, through randomization of members of a microcredit organization to receive a health education module based on the World Health Organization's Integrated Management of Childhood Illness (IMCI) community intervention. METHODS: Baseline data were collected in February 2007 from clients of a microcredit organization in Pucallpa, Peru (n = 1,855) and their children (n = 598). Loan groups, consisting of 15 to 20 clients, were then randomly assigned to receive a health education intervention involving eight monthly 30-minute sessions given by the organization's loan officers at monthly loan group meetings. In February 2008, follow-up data were collected, and included assessments of sociodemographic information, knowledge of child health issues, and child health status (including child height, weight, and blood hemoglobin levels). To explore the effects of treatment (i.e., participation in the health education sessions) on the key outcome variables, multivariate regressions were implemented using ordinary least squares. RESULTS: Individuals in the IMCI treatment arm demonstrated more knowledge about a variety of issues related to child health, but there were no changes in anthropometric measures or reported child health status. CONCLUSIONS: Microcredit clients randomized to an IMCI educational intervention showed greater knowledge about child health, but no differences in child health outcomes compared to controls. These results imply that the intervention did not have sufficient intensity to change behavior, or that microcredit organizations may not be an appropriate setting for the administration of child health educational interventions of this type. TRIAL REGISTRATION: This study is registered with ClinicalTrials.gov, NCT01047033.

Evaluation of Combined Financial Incentives and Deposit Contract Intervention for Smoking Cessation: A Randomized Controlled Trial.

INTRODUCTION: We evaluate whether a combination of financial incentives and deposit contracts improves cessation rates among low- to moderate-income smokers. METHODS: We randomly assigned 311 smokers covered by Medicaid at 12 health clinics in Connecticut to usual care or one of the three treatment arms. Each treatment arm received financial incentives for two months and either (i) nothing further ("incentives only"), (ii) the option to start a deposit contract with incentive earnings after the incentives ended ("commitment"), or (iii) the option to precommit any earned incentives into a deposit contract starting after the incentives ended ("precommitment"). Smoking cessation was confirmed biochemically at two, six, and twelve months. RESULTS: At two, six, and twelve months after baseline, our estimated treatment effects on cessation are positive but imprecise, with confidence intervals containing effect sizes estimated by prior studies of financial incentives alone and deposit contracts alone. At two months, the odds ratio for quitting was 1.4 in the incentive-only condition (95% CI: 0.5 to 3.5), 2.0 for incentives followed by commitment (95% CI: 0.6 to 6.1), and 1.9 for incentives and precommitment (95% CI: 0.7 to 5.3). CONCLUSIONS: A combined incentive and deposit contract program for Medicaid enrollees, with incentives offering up to $300 for smoking cessation and use of support services, produced a positive but imprecisely estimated effect on biochemically verified cessation relative to usual care and with no detectable difference in cessation rates between the different treatment arms.

Falling living standards during the COVID-19 crisis: Quantitative evidence from nine developing countries.

Despite numerous journalistic accounts, systematic quantitative evidence on economic conditions during the ongoing COVID-19 pandemic remains scarce for most low- and middle-income countries, partly due to limitations of official economic statistics in environments with large informal sectors and subsistence agriculture. We assemble evidence from over 30,000 respondents in 16 original household surveys from nine countries in Africa (Burkina Faso, Ghana, Kenya, Rwanda, Sierra Leone), Asia (Bangladesh, Nepal, Philippines), and Latin America (Colombia). We document declines in employment and income in all settings beginning March 2020. The share of households experiencing an income drop ranges from 8 to 87% (median, 68%). Household coping strategies and government assistance were insufficient to sustain precrisis living standards, resulting in widespread food insecurity and dire economic conditions even 3 months into the crisis. We discuss promising policy responses and speculate about the risk of persistent adverse effects, especially among children and other vulnerable groups.

Small individual loans and mental health: a randomized controlled trial among South African adults.

BACKGROUND: In the developing world, access to small, individual loans has been variously hailed as a poverty-alleviation tool - in the context of "microcredit" - but has also been criticized as "usury" and harmful to vulnerable borrowers. Prior studies have assessed effects of access to credit on traditional economic outcomes for poor borrowers, but effects on mental health have been largely ignored. METHODS: Applicants who had previously been rejected (n = 257) for a loan (200% annual percentage rate - APR) from a lender in South Africa were randomly assigned to a "second-look" that encouraged loan officers to approve their applications. This randomized encouragement resulted in 53% of applicants receiving a loan they otherwise would not have received. All subjects were assessed 6-12 months later with questions about demographics, socio-economic status, and two indicators of mental health: the Center for Epidemiologic Studies - Depression Scale (CES-D) and Cohen's Perceived Stress scale. Intent-to-treat analyses were calculated using multinomial probit regressions. RESULTS: Randomization into receiving a "second look" for access to credit increased perceived stress in the combined sample of women and men; the findings were stronger among men. Credit access was associated with reduced depressive symptoms in men, but not women. CONCLUSION: Our findings suggest that a mechanism used to reduce the economic stress of extremely poor individuals can have mixed effects on their experiences of psychological stress and depressive symptomatology. Our data support the notion that mental health should be included as a measure of success (or failure) when examining potential tools for poverty alleviation. Further longitudinal research is needed in South Africa and other settings to understand how borrowing at high interest rates affects gender roles and daily life activities. CCT: ISRCTN 10734925.

The effect of images of Michelle Obama's face on trick-or-treaters' dietary choices: A randomized control trial.

OBJECTIVE: To evaluate the microfoundations of a personality-inspired public health campaign's influence on minors. DESIGN: Multi-year randomized control trial. SETTING: Economics professor's front porch in New Haven, CT. PARTICIPANTS: 1223 trick-or-treaters in New Haven over three years; on average, 8.5 years old and 53% male (among children whose gender was identifiable). ELIGIBILITY: Trick-or-treaters over the age of three that approached the house. INTERVENTION: Random assignment to the Michelle Obama side of the porch or the Comparison side of the porch. MAIN OUTCOME MEASURE: Selection of fruit over candy. METHODS: Difference-in-means estimates. RESULTS: We estimate that viewing a photograph of Michelle Obama's face relative to control conditions caused children to be 19% more likely to choose fruit over candy. CONCLUSIONS: Michelle Obama's initiative to reduce childhood obesity has influenced children's dietary preferences. Whether this influence extends beyond Halloween trick-or-treating in New Haven, CT on the porch of an economics professor requires further research.

Unconditional cash transfers for clinical and economic outcomes among HIV-affected Ugandan households.

BACKGROUND: HIV infection has profound clinical and economic costs at the household level. This is particularly important in low-income settings, where access to additional sources of income or loans may be limited. While several microfinance interventions have been proposed, unconditional cash grants, a strategy to allow participants to choose how to use finances that may improve household security and health, has not previously been evaluated. METHODS: We examined the effect of an unconditional cash transfer to HIV-infected individuals using a 2 × 2 factorial randomized trial in two rural districts in Uganda. Our primary outcomes were changes in CD4 cell count, sexual behaviors, and adherence to ART. Secondary outcomes were changes in household food security and adult mental health. We applied a Bayesian approach for our primary analysis. RESULTS: We randomized 2170 patients as participants, with 1081 receiving a cash grant. We found no important intervention effects on CD4 T-cell counts between groups [mean difference 35.48, 95% credible interval (CrI) -59.9 to 1131.6], food security [odds ratio (OR) 1.22, 95% CrI: 0.47 to 3.02], medication adherence (OR 3.15, 95% CrI: 0.58 to 18.15), or sexual behavior (OR 0.45 95% CrI: 0.12 to 1.55), or health expenditure in the previous 3 weeks (mean difference $2.65, 95% CrI: -9.30 to 15.69). In secondary analysis, we detected an effect of mental planning on CD4 cell count change between groups (104.2 cells, 9% CrI: 5.99 to 202.16). We did not have data on viral load outcomes. CONCLUSION: Although all outcomes were associated with favorable point estimates, our trial did not demonstrate important effects of unconditional cash grants on health outcomes of HIV-positive patients receiving treatment.

Credit with Health Education in Benin: A Cluster Randomized Trial Examining Impacts on Knowledge and Behavior.

We evaluate whether health education integrated into microcredit lending groups reduces health risks by improving health knowledge and self-reported behaviors among urban and rural borrowers in eastern Benin. In 2007, we randomly assigned 138 villages in the Plateau region of Benin to one of four variations of a group liability credit product, varying lending groups' gender composition and/or inclusion of health education using a 2 × 2 design. Women in villages receiving health education, regardless of gender composition of the groups, showed improved knowledge of malaria and of human immunodeficiency virus/acquired immunodeficiency syndrome (HIV/AIDS), but not of childhood illness danger signs. No significant changes in health behavior were observed except an increase in HIV/AIDS prevention behavior, a result predominantly driven by an increase in respondents' self-reported ability to procure a condom, likely an indicator of increased perceived access rather than improved preventative behavior. Women in villages assigned to mixed-gender groups had significantly lower levels of social capital, compared with villages assigned to female-only groups. This suggests there may be an important trade-off to consider for interventions seeking improved health outcomes and social capital through provision of services to mixed-gender groups. Although bundling health education with microcredit can expand health education coverage and lower service-delivery costs, the approach may not be sufficient to improve health behaviors.

Development economics. A multifaceted program causes lasting progress for the very poor: evidence from six countries.

We present results from six randomized control trials of an integrated approach to improve livelihoods among the very poor. The approach combines the transfer of a productive asset with consumption support, training, and coaching plus savings encouragement and health education and/or services. Results from the implementation of the same basic program, adapted to a wide variety of geographic and institutional contexts and with multiple implementing partners, show statistically significant cost-effective impacts on consumption (fueled mostly by increases in self-employment income) and psychosocial status of the targeted households. The impact on the poor households lasted at least a year after all implementation ended. It is possible to make sustainable improvements in the economic status of the poor with a relatively short-term intervention.

SAVINGS BY AND FOR THE POOR: A RESEARCH REVIEW AND AGENDA.

The poor can and do save, but often use formal or informal instruments that have high risk, high cost, and limited functionality. This could lead to undersaving compared to a world without market or behavioral frictions. Undersaving can have important welfare consequences: variable consumption, low resilience to shocks, and foregone profitable investments. We lay out five sets of constraints that may hinder the adoption and effective usage of savings products and services by the poor: transaction costs, lack of trust and regulatory barriers, information and knowledge gaps, social constraints, and behavioral biases. We discuss each in theory, and then summarize related empirical evidence, with a focus on recent field experiments. We then put forward key open areas for research and practice. JEL Codes: D12, D91, G21, O16.

Microcredit in theory and practice: using randomized credit scoring for impact evaluation.

Microcredit institutions spend billions of dollars fighting poverty by making small loans primarily to female entrepreneurs. Proponents argue that microcredit mitigates market failures, spurs micro-enterprise growth, and boosts borrowers' well-being. We tested these hypotheses with the use of an innovative, replicable experimental design that randomly assigned individual liability microloans (of $225 on average) to 1601 individuals in the Philippines through credit scoring. After 11 to 22 months, we found evidence consistent with unmet demand at the current price (a roughly 60% annualized interest rate): Net borrowing increased in the treatment group relative to controls. However, the number of business activities and employees in the treatment group decreased relative to controls, and subjective well-being declined slightly. We also found little evidence that treatment effects were more pronounced for women. However, we did find that microloans increase ability to cope with risk, strengthen community ties, and increase access to informal credit. Thus, microcredit here may work, but through channels different from those often hypothesized by its proponents.