Giving What We Can no longer conducts our own research into charities and cause areas. Instead, we're relying on the work of organisations including J-PAL, GiveWell, and the Open Philanthropy Project, which are in a better position to provide more comprehensive research coverage.
These research reports represent our thinking as of late 2016, and much of the information will be relevant for making decisions about how to donate as effectively as possible. However we are not updating them and the information may therefore be out of date.
Cause Area: Economic Empowerment
Among organizations that target poverty reduction, there is a long tradition of directly providing aid. Historically, this has come largely in the form of in-kind transfers, including food, asset distribution (e.g. livestock) or the supply of medical care1. In recent years, however, cash transfers have emerged as an increasingly popular strategy for poverty alleviation. Generally speaking, there exist two types of cash transfers: conditional and unconditional.
Conditional Cash Transfers (CCT) are programs that make cash disbursements to households that satisfy certain behavioural requirements, generally concerning the health and education of children resident in the household. For example, Mexico’s “Oportunidades” programme (formerly known as "Progresa"), provides cash transfers to households conditional on their children regularly attending school and making health clinic visits. CCTs originated in Latin America (viz. Brazil and Mexico) but the strategy is now widespread, and employed in both developing and developed countries2. Some of these programs reach very large numbers of beneficiaries: for instance, in 2009 the Mexican “Oportunidades” program reached 5 million households and the Brazilian “Bolsa Familia” served 11 million families3.
CCTs have two objectives: reducing poverty and increasing the human capital of future generations. The strategy achieves the first aim by targeting poor households as the recipients of the transfers, and the second by making the transfers conditional on investment in children’s health and education4.
The poverty reduction objective can target both current and future poverty. CCTs address the former by raising household purchasing power, and tackle the latter by enabling households to save and invest, thereby fostering long-term financial stability. Some programs also offer business training and support, aimed at boosting future income. The Nicaraguan program “Atencion a Crisis”, for example, required (among other things), that a member of the beneficiary household take an occupational training course and produce a business plan for approval by a technical team.
Investment in children's education is often achieved by making the transfer conditional on school enrolment and attendance. The transfer then partially covers both the direct costs of attending school, and the opportunity cost of sending children to school. With respect to the health dimension, CCTs primarily focus on preventive health care: regular check-ups for children, pregnant and breastfeeding women, and household participation in community information sessions. The value of these transfers is designed to cover the price of a basic food basket and the cost of preventive health care5.
CCTs can vary in many respects. One of the most salient distinctions is the difference between "broad" and "narrow" programmes. The former aims to achieve wider aims, like encouraging poor households to invest in human capital, especially by promoting health and education. Narrow CCTs, by contrast, focus on more specific services. India’s “Janani Suraksha Yojana”, for example, focuses on maternal and infant health. While initial CCT programmes were often broad in nature, narrower programmes are becoming increasingly common.
CCTs also vary largely with respect to scale. For example, Brazil’s “Bolsa Familia”, a program promoting children’s education and health, includes more than 46 million people. On the other hand, there are several organizations that run small scale interventions, including New Incentives, the charity we recommend below.
CCT beneficiaries are often selected by a "proxy means test", a survey that predicts household' income by gathering simple information about its assets. Beneficiaries can be either be selected for the program automatically (if evaluators visit households, conduct the survey and then automatically enroll those that qualify for the program), or receive funding through self-identification (if households apply for the program, and evaluators assess their eligibility on the basis of the survey)6.
Aside from medical trials and health policies, CCTs represent the largest and longest-running set of programmes evaluated by RCTs7. The Latin American countries that first employed the programs built in a host of mechanisms for best practice monitoring and evaluation8.
RCTs have shown that conditional cash transfers have increased consumption patterns among recipient households, with notable results in improved food consumption. Mexico's “Oportunidades”, for example, led to a 10-20% increase in food consumption. An RCT of the Mexican intervention estimated that the treatment group’s consumption was 5.6% higher than the control group's consumption, four years after the control group began to receive treatment and five and a half years after the treatment group began to receive transfers9. “Red De Proteccion Social”, a Nicaraguan CCT focusing on health and education, engendered an increase of approximately 15% in household expenditure and a 25% increase in food expenditure. Atencion a Crisis, a conditional cash transfer program in Nicaragua, led to an increase of about 30% in food consumption10. A 2009 report by the World Bank additionally found that household recipients of CCTs benefit from higher spending on food and better nutrition than comparable households that do not receive transfers11.
Allaying a common concern, the transfers did not lead to a large increase in spending in alcohol or tobacco: an RCT of the Mexican Programa de Apoyo Alimentario found an increase alcohol expense equivalent to 1.5% of the value of the transfer and no increase in tobacco consumption12. Similarly, an RCT of the Nicaraguan Red de Proteccion Social found that no statistically significant effect of the transfer on alcohol and tobacco consumption13. It is worth noting, however, that there exist worries concerning the possibility that consumption of these products may be under-reported14.
An RCT of the Mexican CCT 'Oportunidades' found that beneficiaries invested 26% of their transfer (while the rest was spent on consumption goods and services). The investment increased the recipients' agricultural income by 10% after 18 months of receiving benefits15. The randomised evaluation of Red de Proteccion Social (Nicaragua), by contrast, found no change in investment due to the transfer. However, this was in line with the program’s recommendations, which recommended that the transfers should be spent on education and food to improve nutritional status16.
CCTs have been consistently shown to increase school enrolment and attendance171819, positively affect whether or not children work and how much they work,2021 and lead to an higher overall use of preventive health services among mothers22 and children232425. A 2015 review of the long term effect of CCTs in Latin America found consistent positive effects on schooling, as well as moderate positive impacts on cognitive skills and learning, socioemotional skills and off-farm employment and income. However, a number of other estimates were not statistically significant. The authors noted this could have been attributable to either a lack of impact or methodological problems26.
Conditional cash transfers can be costly to administer. This is because targeting (which generally requires proxy-means tests) and monitoring programme conditions are both expensive activities. For instance, the cost associated with these two mechanisms have been estimated to account for 60% of the administrative costs of Progresa (the former name of the Mexican "Oportunidades" program), 49% of the costs for the Programa de Asignación Familiar program in Honduras, and 31% of Red De Proteccion Social in Nicaragua27.
Conditional cash transfers are extremely flexible: among other things, they can vary in size, value, target beneficiaries, and the behavioural changes sought. This means that general estimate of their costs and cost effectiveness will not be particularly valuable. Instead, programmes should be evaluated by type or on an individual basis.
Despite their many strengths, conditional cash transfer have limitations. For one, they can only be deployed when the provision of services is extensive and of good qu****ality28. Moreover, CCTs risk excluding the most vulnerable people from the distribution of benefits29. One problem is that proxy-means testing can lead to incorrect targeting. A study of targeting mechanisms in Indonesia, for example, found that standard practice was likely to incorrectly exclude 53% of poor households and include 20% of non-poor households30. Furthermore, when CCTs employ self-targeting strategies, the stringency of programme conditions may discourage the most vulnerable households from even applying to the program in the first place31.
Another issue is that individuals who are unable to meet program requirements may be negatively affected by CCTs. A study of a Malawian CCT targeted at improving the health of adolescent girls found that girls unable to meet the conditions transitioned into marriage and childbearing earlier than comparable samples of girls who received unconditional transfers32.
Yet another worry bears on perverse incentives, namely, concerns that households might intentionally worsen their condition in order to qualify for the CCT programmes33. A study of the Brazilian Bolsa Alimentação, a Brazilian CCT focused on child height and weight, suggests that the program had a significant negative impact on children’s weight and a borderline-significant negative effect on children’s height (both measured relative to age). The authors hypothesize that this was the result of perverse incentives: because a previous program made the benefit conditional on the children being underweight, households might have believed that children needed to be underweight to qualify for the Bolsa Alimentação34.
Stigmatisation is also associated with the delivery of CCT35. A study of a nationally representative survey in Argentina found that a significant majority of the population thought that CCT's beneficiaries “lie to receive a plan”, that “many of those who receive plans could find a job if they really wanted to”, and that plans lead to people “not going out to look for jobs”, “not knowing how to solve their problems”, and “becoming less prone to solidarity”36. Likewise, a study of the Peruvian CCT "Juntos" found that beneficiaries were stigmatised in their own communities: perceptions of recipients held that they “do not want to work anymore” and had “become lazy”. Communities also mistakenly believed that women would strategically get pregnant in order to qualify or remain enrolled in the programme, and that beneficiaries would spend transfer payments on alcohol37. Recent studies have focused on the relationship between CCTs and political participation: some studies show that CCTs increase voters turnout and support of incumbent parties383940, while other research suggests CCTs lead to clientelistic vote buying and threats of programme expulsion in case of non-electoral support41 decrease political participation and lower participation in demonstrations and meeting activity 42.
Thoughthere is episodic evidence concerning these risks, questions concerning these issues have not yet been systematically or widely studied.
CCTs have garnered considerable attention in the last 20 years. Fiszbein and Schady (2009) found that CCT programs had been introduced in at least twenty-eight countries by 2009, covering up to 40 percent of the population of these countries. The World Bank planned to provide US $2.4 billion to support CCTs in 2009. As CCTs have become increasingly popular, they have been employed for a wide range of behavioural objectives: they can aim to achieve wider targets, like encouraging poor households to invest in health and education in general (focusing on school attendance and preventive health checks), or focus on specific educational or health outcomes (for instance, targeting maternal and infant mortality). Many of the behavioural outcomes promoted by conditional cash transfers still require additional support.
Unconditional cash transfers consist of the transfer of funds to recipients, independent of any attached conditions.
Like CCTs, unconditional transfers vary in size and frequency. Kenya’s Cash Transfer for Orphans and Vulnerable Children program (CTOVC), for instance, provided a monthly transfer of $15 (around 21% of household spending)43 to ultra-poor families with orphans and vulnerable children aged 17 years and younger,44 while a micro-enterprise UCT in Mexico provided a one-off grant of $86 (around 4% of annual profits) to small retail firms45.
In comparison to microfinance, UCTs have the advantage of not saddling households with long term debt service, which is often the case even with relatively favourable interest rates and term lengths46. Moreover, unconditional cash transfers do not risk excluding poor, but high potential households, a danger run by microcredit institutions47.
When comparing conditional and unconditional cash transfers, the latter have a significant advantage: they do not deny a safety net to poor households simply because they may not meet certain behavioural requirements. UCTs also allow vulnerable households to identify their own needs48. Relatedly, they do not risk excluding the most vulnerable parts of the population and negatively affecting those who drop from the program.
A randomised control trial of the CTOVC unconditional cash transfer program in Kenya found that the intervention lead to a 17% increase in food consumption. The study also found a small and statistically insignificant decrease in alcohol consumption49.
An updated draft of an influential study on unconditional cash transfers was recently released by Haushofer and Shapiro5051. The study consisted in a pre-registered, randomised control trial of an unconditional cash transfer programme in Kenya, and collected data on 1372 households at endline. The study comprised randomised treatment groups: men vs. women, lump sum vs. monthly instalments, and large vs. small transfers. This study is especially significant insofar as the considered intervention was the most similar to that of GiveDirectly, the charity we recommend below. The principal dissimilarity is that while GiveDirectly transfers about $1,000 to recipients, 72% of treatment group members in the Haushofer and Shapiro evaluation received only $287 ($404 PPP), with a smaller part of the study group receiving $1,085 ($1525 PPP) in transfers. Unless otherwise specified, we focus on results for larger outcomes, because they are similar in size to GiveDirectly’s transfers. We report the outcomes in purchasing power parity (PPP) adjusted U.S. dollars.
Households that received these transfers consumed on average $51 PPP more per month than a control group, and spent about half of it on food52. Participants in the study had low levels of food security. Recipients of the transfers reported a 0.39 standard deviation increase in the index employed to measure food security53. They spent about $2.50 PPP a month more than control households on health expenditures and about 2 PPP a month more on education54. Further, it was found that regular monthly transfers are more likely than lump-sum transfers to improve food security, while lump-sum transfers are more likely to be spent on durables55. There was no increase in spending on temptation goods, such as tobacco and alcohol56.
The Haushofer and Shapiro study found that recipients of the larger transfers increased the value of non-land assets of $241- this affected in particular livestock and durable goods57. Receiving a transfer (be it large or small) increased the likelihood of owning an iron roof by 24 percentage points, and recipients of large transfers were 23 percentage points more likely to have iron roofs at endline than recipients of small transfers58. Iron roofs can be considered assets in so far as they obviate the need to periodically replace thatched roofs: a previous study by Haushofer and Shapiro suggests the related return on the investment is between 7 and 14%59. About $12 a month was dedicated to business expenses. Receipt of large transfers led to an increase of about $14 in monthly revenue, while smaller transfers lead to an increase of around $17 . With neither large nor small transfers, however, was there was an observed increase in profits. It should, though, be stressed that these results stem from only short-term analysis, and it is possible that business investments will lead to increased profits over the longer term60.
Other RCTs of unconditional cash transfers, however, show higher return in the longer term. when coupled with business training or plans. An intervention targeting young adults in a conflict-afflicted area in Uganda invited groups to submit group grant proposals for vocational training and business start-up funding. Randomly selected eligible groups were assigned $382 per participant. The funding was invested mostly in tools and materials, as well as some training. After four years, half practiced a skilled trade and, compared to the control group, the program increased business assets by 57%, work hours by 17%, and earnings by 38%61.
Another program offering women in Uganda a $150 grant and business training, led to a doubling of monthly cash earnings, tripling of cash savings, and an increase of 30 to 50% of short-term expenditures and durable assets relative to the control group62. The difference between the results of these programs and the Haushofer and Shapiro study could be explained by a variety of factors, including the lack of a business plan requirement, lack of business skills training, broader age range of participants and shorter time horizon of the latter study63.
Participants in the Haushofer and Shapiro study showed a significant improvement in psychological well-being for recipients of unconditional cash transfers (0.45 standard deviations)64. There were no overall effect on levels of the stress hormone cortisol, although cortisol levels are significantly lower when transfers are made to the wife rather than the husband. This is especially interesting since other outcomes do not differ between male and female recipients65. The study found a large improvement in female empowerment in control households for treated villages. This suggests there might be a positive spillover effects of unconditional cash transfers, though the authors note that the result warrants further investigation66.
As noted above (3.1), two principal reasons motivate the usage of conditions on cash transfer recipients: firstly, households might lack information on the long term benefits of health and education and, secondly, the interest of children and parents may in some cases diverge67. Unconditional cash transfers, by contrast, are rooted in an alternative explanation of the limited use of public services: that it is simply limited by cash constraints. With this in mind, proponents of UCTs believe that distributing cash without any conditions attached will suffice to increase the use of the relevant social provisions68.
At the moment, evidence seems to suggest that, at least in some cases, unconditional cash transfers improve investment in human capital, but not as much as conditional ones. A randomised experiment in Burkina Faso found that conditionalities did not have significant effects on schooling attendance of children who are traditionally favored by parents for school participation - such as boys, older children, and higher ability children. However, conditional transfers were considerably more effective than unconditional ones in improving the enrollment of children who were not traditionally favoured for school participation - such as girls, younger children, and lower ability children69.
A recent review of the effectiveness of conditional and unconditional cash transfers for schooling outcomes in developing countries showed that interventions with no conditions or unmonitored conditions lead to a 18-25% improvement in odds of being enrolled in school, while monitored and enforced conditions lead to a 60% improvement in odds of enrollment. However, in both cases there was little effect on test scores70. With respect to health outcomes, a recent study estimated that preventative visits for children were 50% lower in the absence of conditionalities71.
It should be stressed, however, that the comparative assessment of UCTs and CCTs constitutes a new area of research and it warrants further study.
UCTs tend to have lower costs than CCTs, generally because they do not require the same degree of monitoring and follow up72. As technology advances, the administrative costs of unconditional cash transfers will likely further decrease. An RCT conducted in 2011 compared manual distribution of cash transfers with distribution via mobile technologies, finding that mobile delivery reduced both distribution costs for the implementing agency and collection costs for the recipients73.
Just as for CCT, one might worry that UCT could lead to an increase in prices (inflation). However, two RCTs of unconditional cash transfers have found no evidence of inflation74. A recent study of a Mexican program showed that cash transfers lead to a positive but negligible increase in prices, though the effect was larger in remote villages, where cash transfers lead to a 5 percent price increase in the price of certain foodstuffs75. Similarly, there is no evidence that UCT reduce wage-earning work by adults76 and cash transfers coupled with business encouragement increased hours worked7778. The intervention working with Ugandan women found low levels of hostility from the community, though the level was 38% higher than one found in control villages79. A 2015 study on the intervention analysed by Haushofer and Shapiro finds that a US$100 increase in the mean wealth of the village causes a 0.11 standard deviation decrease in life satisfaction among individuals in households that did not receive transfers. They also find the negative effect on neighbours' life satisfaction decreases over time80.
As it was mentioned at the beginning of this report, the World Bank estimates that 900 million people were living in extreme poverty (less than $1.90 PPP a day)in 201281. There is therefore wide scope for extending UCTs programmes. These types of programs have started to receive attention in public debates only relatively recently. Moreover, aid agencies often find unconditional cash transfers politically unpalatable, since they distribute money without strings attached outside of their country. This means that at the moment UCTs do not yet receive adequate levels of funding. For instance, UCTs currently make up only 6% of total spending on humanitarian aid82.
We have identified two effective charities working on economic empowerment through cash transfers.
New Incentives operates in the Akwa Ibom State in Nigeria. They offer conditional cash transfers to mothers who are HIV-positive or have at-risk pregnancies. They provide transfers up to about $150 over the course of the program. The transfers are conditional on the mother registering her pregnancy and being tested for HIV, delivering in a clinic, and (if HIV-positive) having their newborns tested for HIV.83
GiveDirectly operates in Kenya and Uganda, making unconditional cash transfers to very low-income households. It uses an electronic payment system to transfer roughly $1,000. It will soon start working in Rwanda84.