GiveDirectly - 2015 Review

We have published a more recent review of this organization. See our most recent report on GiveDirectly.

Published: November 2015

Summary

What do they do? GiveDirectly (www.givedirectly.org) transfers cash to households in developing countries via mobile phone-linked payment services. It targets extremely low-income households (more).

Does it work? We believe that this approach faces an unusually low burden of proof, and that the available evidence supports the idea that unconditional cash transfers significantly help people. It appears that GiveDirectly has been effective at delivering cash to low-income households. GiveDirectly has one major randomized controlled trial (RCT) of its impact and took the unusual step of making the details of this study public before data was collected (more).

What do you get for your dollar? The proportion of total expenses that GiveDirectly has delivered directly to recipients is approximately 85% overall (more).

Is there room for more funding? We are reasonably confident that GiveDirectly could effectively use an additional $15 million beyond what it already expects to receive and could potentially absorb up to an additional $75 million (more).

GiveDirectly is recommended because of its:

• Focus on a program with a low burden of proof and a strong track record (more).
• Strong (and evolving) process for ensuring that cash is well-targeted and consistently reaches its intended targets (more).
• Documented success in transferring a high portion of funds raised directly to recipients (more).
• Standout transparency and commitment to self-evaluation (more).
• Room for more funding - we believe that GiveDirectly can use substantial additional funding productively (more).

Major unresolved issues include:

• While GiveDirectly has one major RCT of its activities in Kenya, there is still limited evidence on the impact of the type of transfers (large, one-time transfers) that GiveDirectly generally provides, particularly the long-term impact of such transfers. There are currently several ongoing experimental evaluations of GiveDirectly's programs, including a long-term RCT.
• Through 2014 and 2015, GiveDirectly has increased the amount of time it spends networking with potential government and NGO partners. This time involves senior management taking meetings to discuss GiveDirectly's work, the impact of cash transfers, and potential shared objectives. This work has the potential to significantly increase GiveDirectly's impact, but could divert GiveDirectly's resources away from a proven program to work on one whose payoff is much more uncertain. We are continuing to monitor how GiveDirectly's resources are allocated and to think about how we should evaluate GiveDirectly's partnership work in the future.

Our full review, below, discusses our full assessment of GiveDirectly, including what we see as its strengths and weaknesses as well as issues we have yet to resolve. All content on GiveDirectly is available here.

Our review process

To date, our review process has consisted of

• Conversations with GiveDirectly staff: Paul Niehaus (Director and President), Piali Mukhopadhyay (COO, International), Joy Sun (former COO, Domestic), Carolina Toth (Manager, Finance and Operations), and Stuart Skeates (former Uganda Field Director).
• Conversations with GiveDirectly board members: Rohit Wanchoo (Director), Michael Faye (Director), and Jeremy Shapiro (former Director).
• Reviewing documents GiveDirectly sent in response to our queries.
• In November 2012, we visited GiveDirectly's operations in Kenya, where we met with beneficiaries of its work and spoke with its local field staff.
• In 2014, we retained a journalist to visit GiveDirectly in Kenya. We published his report on our blog.
• In October 2014, we visited GiveDirectly's operations in Uganda, where we met with beneficiaries of its work, spoke with local field staff, and observed a cash out day.

All content on GiveDirectly, including updates, blog posts and conversation notes, is available here.

What do they do?

GiveDirectly transfers cash to poor households in developing countries via mobile phone-linked payment services.1 It is currently active in Kenya and Uganda, and will soon be starting cash transfers in Rwanda.2

GiveDirectly's work includes:3

• Direct impact: Providing cash directly to poor households.
• Experimentation: GiveDirectly runs or participates in studies on a) the impact of cash transfers and b) the costs and benefits of various program designs, with the goal of improving its own cash transfer program, improving other cash transfer programs or encouraging the creation of new programs.4
• Partnership work: GiveDirectly pursues opportunities to partner with other organizations on cash transfer projects. Through these projects, GiveDirectly aims to encourage the evaluation of aid projects (often by using cash transfers as a standard of comparison) and ultimately influence funders to move resources from less effective aid programs to more effective ones. 5

We discuss GiveDirectly's experimentation and partnership work to some extent below, but most of our review focuses on GiveDirectly's direct impact.

GiveDirectly's standard model involves grants of approximately $1,000 (USD) per household over approximately four months to one year, after which recipients become ineligible.6 GiveDirectly aims to help the poorest households, targeting those that are in "acute poverty."7 Recipient households are currently identified by (a) selecting poor villages based on poverty data, and (b) determining household eligibility. In the past, household eligibility was based on building materials of homes. Recently, GiveDirectly started to use new, more complex eligibility criteria in Homa Bay county in Kenya (more).

Below, we discuss:

• The structure of GiveDirectly's transfers
• The status of GiveDirectly's transfer campaigns
• GiveDirectly's process for identifying recipient households and delivering cash transfers
• GiveDirectly's staff structure
• GiveDirectly's experimental work
• GiveDirectly's work on partnerships

Grant structure

GiveDirectly's standard model involves grants of approximately $1,000 (USD) over about four months to one year, after which recipients become ineligible.8 In Kenya, GiveDirectly transfers $1,040 to each enrolled household, while in Uganda, it transfers $875; these transfer amounts are based on GiveDirectly's standard transfer size, but are adjusted for purchasing power.9 This is a different approach from the approach we've seen in government cash transfer programs. One way of putting the difference (which has been reflected in GiveDirectly's communications with us) is that government programs aim for "income transfers" (small supplements to income over many years), whereas GiveDirectly aims for "wealth transfers" (large, one-off transfers that hopefully give people more flexibility to make large purchases and investments).

Note that when we reviewed household data several years ago, we found that household size varies substantially: while the mean household size was ~4.7 and the median size was 4, 16% of households had 1 or 2 people, ~20% had 6 or more, and the maximum household size was 16.10 We estimate that the average family receives $288 per capita from GiveDirectly, which is 121% of baseline annual consumption per capita.11

Frequency and sizing of transfers

In Kenya, GiveDirectly's standard transfer schedule involves a small initial transfer of about $90 (USD), followed by two larger transfers of about $475 (USD).12

In Uganda, GiveDirectly's transfer schedule until recently involved 9 monthly transfers of about $97 (USD) each.13 GiveDirectly told us that the transfer schedule in Uganda was designed to be more manageable for the mobile money providers than two lump-sums, and also that it would provide more frequent opportunities to evaluate the performance of the mobile money providers.14 In the most recent campaigns in Uganda, GiveDirectly has used a structure more like Kenya's with three transfers.15

GiveDirectly has an ongoing study of behavioral interventions that will allow some recipients the ability to choose when they receive their transfers.

Status of transfer campaigns

As of August 2015, GiveDirectly had provided partial or full cash transfers to almost 25,000 households in western Kenya and eastern Uganda, and was continuing to transfer funds to additional households in both places.16

GiveDirectly's work to-date can be grouped into 13 campaigns (note that to keep this table easy-to-read, we have included many details in footnotes):17

We have created a summary table of the campaigns noting the documentation we have for each here.

GiveDirectly's process

The steps of GiveDirectly's process are as follows:

1. Selection of a country. GiveDirectly told us that it chose to work in Kenya due to the robustness of M-Pesa as a mobile banking platform and the large population of people meeting its criteria who have access to mobile technology.37 In choosing a second country in which to work, GiveDirectly said that it considered whether there was a mobile money provider accessible to the very poor, how costly it would be to operate in the country, how politically stable the country is, and how common corruption is in government affairs. The ease of moving staff between Kenya and Uganda was also a factor, as GiveDirectly's current COO (International) oversees the work in both places.38 GiveDirectly chose to work in Rwanda to pursue a partnership with a bilateral aid donor after staff at the Rwanda office of that donor approached GiveDirectly about a possible cash transfer program there.39 Rwanda has a growing mobile money system and a large population of extremely poor potential recipients.40 GiveDirectly notes that it currently does not direct individual donor contributions to Rwanda.41
2. Selection of a region. GiveDirectly told us that it initially chose to work in western Kenya and eastern Uganda based on poverty statistics.42 We have reviewed poverty data that GiveDirectly sent us for Uganda; we have not reviewed poverty data across districts for Kenya.43 GiveDirectly has not yet selected regions within Rwanda.
3. Selection of districts or counties. For Kenya, GiveDirectly told us that its executive staff uses data on poverty, population density, security, and presence of poverty-focused NGOs (with the goal of avoiding overlapping with these) to select counties.44 We have reviewed poverty data that GiveDirectly sent us for divisions within Siaya district, but not the data used to select other counties in Kenya.45 For Uganda, GiveDirectly told us that it chose a county to target initially based on poverty statistics, logistical factors and security considerations.46 We have reviewed the poverty data for Uganda.47
4. Selection of villages. GiveDirectly states that it selects villages based on poverty level and location.48 GiveDirectly shared the full details of its village selection process for an early campaign in Kenya, including data for each village and the method for weighting the different factors used to select villages in that campaign.49 We have not reviewed recent data used to select villages in Kenya. For the pilot campaign in Uganda, GiveDirectly relied on publicly available poverty data, which we have reviewed, as well as data that it received from local officials, which we have not reviewed.50 GiveDirectly also sent field staff to test cell phone reception and measure proximity to market centers, and used these as inputs into its village selection (optimizing for good reception and longer distances from market centers).51 GiveDirectly also considered population size in the villages, so that it could enroll all eligible households in each village and not exceed the budget for the Uganda pilot campaign, which was funded by part of its Google Global Impact Award.52 For recent campaigns in Kenya and Uganda, GiveDirectly has used various methods to estimate poverty levels through census data and the proportion of thatch-roof to iron-roof households.53 For details on how GiveDirectly has targeted villages historically, see this footnote.54
5. Obtaining permission from local officials. Before beginning to work in a given area, GiveDirectly obtains permission from local officials. This process can involve officials from the national to the village level and generally requires a series of conversations to get all the relevant stakeholders on board.55 GiveDirectly signs written agreements with or obtains approval letters from local officials to formalize permissions.56
6. Village meeting. A village meeting is held "to answer questions anyone may have about the program, clarify that [GiveDirectly is not] affiliated with a political party, etc."57 Village meetings were first implemented in the Google campaign.58
7. Enrollment process.
   • Census: GiveDirectly has field staff visit the village to create a census of all households.59 In conducting the census, field staff collect data about each household and note if the household is eligible for transfers (the criteria for eligibility in a campaign depends on where the campaign is located – more).60 The census process was different in GiveDirectly's early campaigns.61
   • Registration: GiveDirectly has a separate set of field staff visit households marked as eligible in the census and register them.62 In Kenya, registration involves giving the household member a SIM card (if they do not already have an M-PESA account), which is used to transfer funds through the M-PESA system, and collecting other data that can be checked against the initial data from the census.63 Recipients are given the option of purchasing a cell phone from GiveDirectly at the time of registration, the cost of which is removed from the recipient's transfer.64 In Uganda, GiveDirectly uses a similar registration process.65 Additionally, GiveDirectly helps recipients in Uganda obtain national ID cards and arranges for mobile money agents to visit villages to register recipients in the mobile money system.66 A registered household is formally enrolled only after all phases of enrollment (census, registration, back check, and audit) have been completed and the household has obtained a mobile money account (if necessary).67 Registration was different in early transfer campaigns.68
   • Back check: GiveDirectly sends a separate team of field staff to revisit every registered household and collect data about that household that can be compared to data collected during census and registration.69 GiveDirectly field staff also ask households if they were asked to pay a bribe to register.70
   • Audits: GiveDirectly sends field staff to revisit a portion of the registered households for audits.71 In Kenya, the field staff who do audits are not involved in earlier enrollment activities.72 In Uganda, the field staff who have done audits in past campaigns were from earlier enrollment teams.73 GiveDirectly determines which households to audit based on the extent of the discrepancies between data collected at different phases in enrollment.74 GiveDirectly field staff resolve discrepancies during audits to determine whether households are eligible or ineligible. Households found to be eligible through this process are then considered formally enrolled, in addition to the households considered eligible after backcheck and not selected for audit.75 The procedure for deciding which households to audit and determining eligibility was different in prior campaigns.76

   GiveDirectly aims to enroll all eligible households.77 If eligible members of the household are not home during a phase of enrollment, GiveDirectly staff revisit the household several times until they can be found.78 GiveDirectly also checks the quality of its Field Officers' interactions with potential recipients by administering "quality audits" that test how well recipients understand GiveDirectly's program and ask how the Field Officer conducted himself or herself.79

   We have reviewed (and made public) data collected during each step of the enrollment process for most of GiveDirectly's campaigns, with deletions to preserve anonymity.80

8. Sending transfers to recipients. GiveDirectly sends transfers to recipients via mobile money providers (more).81 In Kenya, transfers are sent in an initial installment of approximately $90, then two larger installments of approximately $475.82 In Uganda, transfers have historically been sent in nine installments of approximately $100 each. Recently, GiveDirectly has begun to use a three transfer model similar to what it does in Kenya.83 See above for more on grant structure.
9. Holding "cash out" days (Uganda only): In Uganda, the mobile money agent network is less robust, so GiveDirectly coordinates cash out days for recipients to withdraw funds after each of the installments.84 Cash out days are overseen by the Uganda Field Director, who coordinates with mobile money agents to travel to the village and set up stations for recipients to withdraw cash. In each village, there are two recipients nominated by the community to assist the Uganda Field Director in monitoring cash out day activities. GiveDirectly's call center staff also conduct phone surveys with a randomly selected 10% of recipients in the village to ask if there are any issues at the payday. We have not reviewed the logs of these calls. GiveDirectly previously changed its cash out day procedure in response to a case of staff fraud in Uganda.85 GiveDirectly is experimenting with a model in Uganda that more closely matches the model in Kenya; if this model is successful, GiveDirectly may eliminate cash out days from Uganda (more).86
10. Conducting follow up calls. GiveDirectly field staff make multiple phone calls to recipients as transfers are being sent.87 There are short verification calls to confirm that the transfer was received and ask if the recipient experienced any problems after each transfer.88 There are also two longer surveys administered to a randomly selected sample of recipients after larger portions of the transfers have been sent.89 In the longer surveys, GiveDirectly staff ask recipients a number of questions including whether they received the transfers or had any trouble withdrawing funds, how they spent the funds, and whether there were any problems in their community relating to the transfers.90 The schedule of follow up calls has varied somewhat by campaign.91 We have reviewed and made public data from these calls for ongoing transfer campaigns in Kenya.92

    In addition to these calls, GiveDirectly maintains a phone "hotline" for recipients to call if they have any questions about the transfers or issues in obtaining funds.93 We have reviewed the records of calls made to GiveDirectly's Kenya hotline from May 2012 – September 2014.94 Recipients can also report issues to GiveDirectly field staff when they are in the village; GiveDirectly created a formal mechanism for recording these reports.95

Key differences in some past campaigns were (a) the lack of a "census" (instead, GiveDirectly asked village officials to take them to eligible households, and thus conducted two in-person checks of each house rather than three); and (b) the lack of a village meeting.96

Staff structure

GiveDirectly delivered its first cash transfers in 2011.97 Starting in January 2011 it had one full-time staff member.98 In early 2013 it hired a second full-time staff member to serve as COO (Domestic).99 GiveDirectly has since expanded its staff significantly. Its current organizational structure in East Africa includes:100

• Chief Operating Officer International (COO-I): The COO-I provides oversight and quality control of cash transfer programming and international operations. The COO oversees the Field Directors.
• Field Directors (FDs): FDs are in charge of overseeing field operations, as well as approving transfer schedules and rosters. GiveDirectly estimated that it would have four Field Directors by the end of 2015.101
• Field Manager: The Field Managers supervise Assistant Field Managers, focusing on quality control, management and training of Field Officers.102 GiveDirectly estimated that it would have 3 Field Managers by the end of 2015 (one each in Kenya, Uganda, and Rwanda).103
• Assistant Field Manager: Assistant Field Managers manage the logistics of transfer rounds and oversee Field Officers, as well as conduct high-level analysis of field operations and work on technology integration.104 GiveDirectly estimated that it would have 13 Assistant Field Managers by the end of 2015.105
• Field Officers (FOs): FOs implement the steps required on the ground to enroll and follow up with households. They have the most face-to-face interaction with recipients and are all hired within the country of the transfers. There is a separate group of FOs for each of the first three pre-transfer stages: census, registration, and backchecks. FOs are also hired to conduct audits and follow-up surveys with recipients post-transfers; some of the FOs hired for these roles may have previously worked on the census, registration, or backcheck phases.106 GiveDirectly estimated that it would have 69 Field Officers by the end of 2015.107

Segovia

In mid-2014, three members of GiveDirectly's board of directors began the for-profit technology company Segovia, which develops software that NGOs and developing-country governments can use to help implement their cash transfer programs.108 One other staff member who was previously working full-time at GiveDirectly switched to working part-time for each entity. We discuss potential risks from the overlap in staff in this blog post.

GiveDirectly deployed several versions of Segovia in 2015, which have automated some processes and led to slight time savings.109 GiveDirectly expects moderate efficiency gains from Segovia in the future.110

Evaluation and experimentation

GiveDirectly's goals for experimentation include increasing the evidence base for cash transfers, improving recipient returns and welfare (both in GiveDirectly's program and others), and developing capabilities necessary to implement larger-scale programs or programs in new contexts.111 GiveDirectly has told us that it has increased its experimentation to the point where every recipient is enrolled in a study or a campaign variation.112 Below, we list the studies that GiveDirectly has completed, is currently working on, or has considered.

Ongoing experimentation

• Macroeconomic effects: Based on conversations with policymakers, GiveDirectly found that a key question relevant to government cash transfer programs is the impact they have on macroeconomic factors such as inflation and job creation.113 GiveDirectly is working to conduct an RCT examining the macroeconomic effects of GiveDirectly's program in Kenya.114 This study includes a long-term follow up component that will provide information on the impacts of cash transfers several years after the transfer.115 The study is randomized at the village level, will involve 325 villages, and is expected to survey approximately 3,900 households and 4,875 enterprises.116 Baseline data collection for the study began in August 2014 and was still in progress as of September 2015.117 Endline data collection was expected to be completed by the end of 2016, although this may be delayed since baseline data collection has taken longer than expected.118 Paul Niehaus, GiveDirectly's President, is serving as one of the Principal Investigators on this study, along with Edward Miguel (UC Berkeley), Johannes Haushoffer (Princeton), and Michael Walker (UC Berkeley).119 In order to mitigate potential bias from his involvement with the research, GiveDirectly has applied a number safeguards, including preregistration of plans for measurement and analysis.120
• Behavioral interventions (Ideas42 study): GiveDirectly is conducting an RCT of two main behavioral interventions: (a) enabling recipients to decide when and how to receive their transfer payments; and (b) providing more information to recipients about spending options.121 The main outcome of interest in this study is the rate of return on spending.122 GiveDirectly is conducting the data collection for this study internally, and the analysis will be done by independent researchers.123 This study began in late October 2014 and endline results are expected mid-2016; as of May 2015 baseline results and the initial cash transfers were completed.124 It is fully funded by an anonymous donor.125
• Gender contracts: GiveDirectly ran a small pilot of informal contracts between spouses receiving cash transfers in the spring of 2015.126 External research partners are evaluating the impacts of the contracts on domestic violence and female empowerment.127 After the initial study group was completed, GiveDirectly began a second round, and expected to have results in fall 2015.128 GiveDirectly has said that if the pilot is successful it will be expanded into a larger-scale project.129
• Biometrics: GiveDirectly is testing the use of biometrics to enhance security in Uganda.130 GiveDirectly has purchased palm readers, collected palm prints during registration, and used palm scans as an additional identification measure during recipients' first cash out day; it is also collecting information on the level of comfort that recipients feel towards biometrics during its follow-up surveys.131 GiveDirectly may continue to use biometrics in contexts where national IDs are uncommon and cash out days are necessary (more).132
• Distributed cash out model in Uganda: GiveDirectly would prefer not to run cash out days in Uganda, so it is currently piloting a model of distributed cash outs there (the same model it uses in Kenya).133 More detail is below.
  
• Aspirations study: GiveDirectly is running an RCT looking at the effect of showing recipients a motivational video before their participation in GiveDirectly's program.134 The main outcome being examined is the rate of return on investments.135 Approximately 6,500 recipients are being enrolled in the study.136

Previous experimentation

• RCT of GiveDirectly's Rarieda campaign: Innovations for Poverty Action (IPA) conducted a randomized controlled trial (RCT) of GiveDirectly's program in which eligible households were selected randomly to receive cash transfers.137 These transfers were made in Rarieda in 2011-2012.138 GiveDirectly publicly provided the plan for collecting and analyzing data to determine the impact of these transfers. The RCT has been published; we discuss it in detail here.
• Small-scale RCT of cash transfers to young women: IPA conducted a RCT of GiveDirectly's Nike campaign, which provided transfers exclusively to young women ages 18-19. This campaign was funded by the Nike Foundation. GiveDirectly shared IPA's survey instrument with us prior to the study.139 We did not see an analysis plan prior to the study, as we did with the Rarieda RCT. GiveDirectly writes that "the pilot was designed at a small scale with the expectation that it would not produce statistically robust evidence, but would provide directional learnings to guide future investment and experimentation."140 The study is now complete, and GiveDirectly has shared its write-up, as well as a qualitative piece on the perspectives of the young women involved in the study, which was prepared for GiveDirectly by an independent researcher; we have reviewed these documents.141
• Extended data collection by phone: IPA received a $30,200 grant to extend data collection in a sub-sample of participants from the Rarieda RCT using mobile phone-based data collection techniques.142 The grant was from USAID via the Policy Design and Evaluation Lab at UCSD.143 The goals of the project were to generate data on longer-term effects of cash transfers (up to two years after completion of the RCT), as well as to study the potential for using mobile phones as cost-effective, easily adaptable tools for data gathering.144 We don't know when results from this study will be available.
• Broadening eligibility with more inclusive targeting: GiveDirectly conducted a small-scale study in Kenya to see whether more inclusive targeting criteria could reduce tension and conflict within villages. In treatment villages, GiveDirectly applied a "saturation" model, which included households with iron roofs and mud walls as eligible. In total, about 85% of households in saturation villages received transfers. GiveDirectly collected its standard follow up survey data, which includes questions about tension and conflict, in the 19 treatment and 18 control villages, where GiveDirectly used its standard targeting criteria. It also conducted focus groups in 3 treatment and 3 control villages to elicit opinions about targeting strategies.145 GiveDirectly found that the data on adverse events was inconclusive, and that when faced with the decision of how to allocate limited resources, focus groups preferred to prioritize thatched-roof households.146 We put limited weight on these results due to the small sample size of the study and would be interested in seeing further research on this question. Note: GiveDirectly separately piloted "community-based targeting," but is not planning to implement it more broadly.147
• Index-based crop insurance program: GiveDirectly and the Rockefeller Foundation developed a strategy for offering index-based insurance to cash transfer recipients (details on index-based insurance in footnote).148 GiveDirectly then ran a small-scale test of the program in Western Kenya, simulating a government cash transfer program.149 GiveDirectly found that the cost of the program was lower than the cost of previous index-based insurance programs and a higher rate of people bought insurance.150 Note that we have only seen a summary of the results. This study and associated cash transfers were fully funded by the Rockefeller Foundation.151
• Eligibility requirements in Homa Bay: GiveDirectly recently experimented with new eligibility requirements because a) it needs new eligibility requirements for Homa Bay county, where grass is scarce and thus thatch roofs are less common, and b) knowing how to use a number of different eligibility requirements increases GiveDirectly's ability to work in new areas.152 Recipients from approximately 55 villages were enrolled using several different targeting methods.153 GiveDirectly chose new eligibility requirements for Homa Bay in October 2015 (more).154

Possible future experimentation

Other ideas that GiveDirectly has considered for future experimentation include:

• Providing cash transfers in an urban setting155
• Providing cash transfers as humanitarian relief156
• Facilitating the pooling of recipient funds for public goods projects157
• Implementing a lifetime basic income guarantee158
• Serving as the payment provider at cash out days159

Partnerships

GiveDirectly has been exploring projects with a number of partners. The projects that GiveDirectly has partnered on or considered involve implementing cash transfers funded by an institutional partner. They have also provided informal advice to those considering cash transfer programs. For a partial list of GiveDirectly’s partnership activities, see this footnote.160

In 2015, GiveDirectly finalized an agreement for its largest partnership project to date: GiveDirectly will be implementing cash transfers in a randomized controlled trial in Rwanda; the study costs $4 million and is co-funded by an institutional funder and Google.org.161 The study will test cash transfers as a benchmark for another still-to-be-chosen aid program.162 GiveDirectly is currently in several preliminary conversations with partners for similarly large projects in the future; it currently is hoping to use the funding provided by Good Ventures earlier this year to encourage matching funds from potential partners (more).163

In 2014, GiveDirectly’s President and COO (International) spent time networking and developing potential partnerships with government officials and international aid agencies.164 In 2015, GiveDirectly's President spent approximately 25% of his time on developing partnership projects, primarily focused on Rwanda and replicating the Rwanda model.165 Although partnership projects are now taking up a significant portion of his time, GiveDirectly does not believe this has negatively affected its core operations.166 We expect partnerships to continue to take up the President's and Chairman's time and to involve a significant portion of GiveDirectly’s funding over the next few years.167

We have not yet attempted to assess the value of the partnership projects. We can imagine cases where partnership projects might be very high leverage (e.g., enabling another organization to "benchmark" its current programming against cash, perhaps ultimately directing funding away from a less effective intervention to cash transfers) and also cases that may have limited value (e.g., implementing a program that would have been implemented effectively without GiveDirectly’s involvement).

Does it work?

This section discusses the following questions:

• Generally speaking, are unconditional cash transfers a promising approach to helping people? We believe that this approach faces an unusually low burden of proof and that the available evidence is consistent with the idea that unconditional cash transfers help people.
• How effective and well-founded are GiveDirectly's criteria? The evidence we have suggests that GiveDirectly targets low-income recipients. We have reservations about the approach GiveDirecly has used in the past (and continues to use in Uganda) of targeting people based on the materials their houses are made of, and we do not yet have a good sense of how effective GiveDirectly's new targeting criteria are.
• Is GiveDirectly effectively targeting people who meet its criteria? We believe GiveDirectly's enrollment process is a relatively effective way of targeting people who meet its criteria.
• Does GiveDirectly have an effective process for getting cash to recipients? GiveDirectly's process seems to have been relatively successful so far, with one notable exception.
• How do recipients spend their cash, and how does this spending impact their lives? We present a variety of evidence, including findings from a randomized controlled trial of GiveDirectly's work.
• Do the cash transfers cause problems and complications that offset their positive impact? GiveDirectly has taken some measures to address this question, and we believe that the evidence so far suggests that while the cash transfers do raise some problems, these problems are relatively minor.
• Are the size and structure of the cash transfers well-thought-through and appropriate? We find GiveDirectly's approach to be defensible, but we look forward to seeing the results of GiveDirectly's experimentation with different approaches in the future.
• Does GiveDirectly divert skilled labor away from other areas? Our best guess is that this is not a serious concern.

Generally speaking, are unconditional cash transfers a promising approach to helping people?

We discuss this question more extensively at our report on cash transfers. In brief:

• The evidence most relevant to GiveDirectly comes from an RCT of a GiveDirectly campaign (available here). We discuss the findings of this RCT in our cash intervention report.
• Cash transfers are among the best-studied development interventions, though questions remain. These studies generally show substantial increases in short-term consumption, especially food, and little evidence of negative impacts (e.g., increases in alcohol or tobacco consumption). It is important to note that most of these studies are of income transfers; there is more limited evidence for programs with wealth transfer models like GiveDirectly's. This is a potential cause for concern and one of the reasons that we are particularly interested in GiveDirectly experimenting with and evaluating different approaches.
• There is also some evidence that recipients are able to invest cash transfers at high rates of return (e.g., ~20% per year), leading to long-term increases in consumption.
• We feel that this intervention faces an unusually low burden of proof, given that poverty reduction is an outcome by definition, though donors' intuitive reactions to it may vary widely.

How effective and well-founded are GiveDirectly's eligibility criteria?

GiveDirectly currently uses two different sets of eligibility criteria for its standard campaigns:

• Assets and vulnerability status: In its campaign in Homa Bay county, Kenya, GiveDirectly uses an algorithm to determine eligibility; this algorithm uses a number of inputs related to the household assets and the vulnerability of recipients.168 GiveDirectly recently developed this algorithm after testing a number of new potential criteria, and expects to use similar algorithms for its other campaigns in the near future.169
• Thatched roofs: To date, GiveDirectly has used housing materials to select recipients in all of its standard campaigns, enrolling households who live in a house made of organic materials (thatched roof, mud walls, and a mud floor) and excluding households with iron roofs, cement walls, or cement floors.170 In GiveDirectly's campaigns in Kenya, about 35-45% of households have been eligible based on these criteria, while in Uganda about 80% of households have been found to be eligible.171 GiveDirectly still uses this criteria in Uganda, but expects to soon adjust the Uganda eligibility requirements based on some of its learnings from recent experimentation with requirements in Homa Bay county, Kenya.172

The assets and vulnerability status criteria

GiveDirectly recently started to work in Homa Bay County in Kenya, where families are less likely to have thatch-roofed houses due to a scarcity of grass.173 For these reasons, GiveDirectly has changed its eligibility criteria for Homa Bay county to better capture the poorest households.174 The new criteria algorithm takes into account a range of factors including a household's assets and vulnerability status; we are unable to elaborate because GiveDirectly would prefer to keep the new criteria confidential so as to prevent households from gaming the system.175

To test possible proxies for poverty to use as its new criteria, GiveDirectly told us that it attempted to determine the validity and replicability for each metric.176 For example, GiveDirectly would test the same criterion on the same group of people at different times to see if respondents gave consistent answers that led to that criterion producing the same results each time about who should be considered eligible.177 We have not yet had an opportunity to fully review how GiveDirectly ultimately analyzed all of its options and arrived at its new criteria, as the new criteria were finalized as we were finishing this report. We expect to discuss this more with GiveDirectly next year. Note that GiveDirectly expects to adjust its eligibility criteria for other campaigns based on its experience in Homa Bay and GiveDirectly told us that it is currently enrolling most of its recipients in Homa Bay, so we expect these eligibility criteria to be the main ones GiveDirectly uses for the foreseeable future.178

GiveDirectly tried to choose criteria that (a) included recipients who would benefit the most from the transfer, (b) were difficult to fake, (c) were low cost to implement, and (d) were perceived as fair both by community members and by GiveDirectly staff.179 GiveDirectly believes that the new criteria will be more difficult to fake, somewhat more expensive to administer, and more difficult to explain (which might lead to people believing the criteria are not fair).180 Note that recipients will not be made aware of the full criteria (as a measure to prevent cheating), which may also contribute to decreased perceptions of fairness. Because the criteria explicitly put weight on vulnerability, they could also increase perceptions of fairness, or at least offset other fairness concerns.

GiveDirectly's development of new eligibility criteria may help GiveDirectly expand to new areas more easily and could provide valuable guidance for other cash transfer programs (although we are unsure if GiveDirectly will be able to share learnings from this project since it hopes to keep its new criteria confidential). However, our evidence for GiveDirectly's impact and for low rates of conflict within villages is based on previous campaigns in which GiveDirectly used different eligibility criteria, and it is possible that the new eligibility criteria will change these outcomes.

The thatched roof and mud house criteria

As part of the baseline survey for the RCT of its program, researchers collected more in-depth information on poverty levels of recipients. GiveDirectly has shared the full survey form used to interview participants, as well as its own summary of the data collected as of March 2012:181

GiveDirectly also provided charts that show a clear difference in the consumption, expenditures, and assets of households in mud and thatch homes compared to those in cement homes, but fairly small differences between those living in mud and thatch homes and those living in mud and iron roof homes.182

End-line data from the RCT on food consumption among control group recipients also suggests that eligible households are extremely poor.183 This data shows that "20% of the control group reports that not all household members usually eat until they are content, 23% of respondents report sleeping hungry in the last week, and only 36% report having enough food in the house for the next day."184 Other results related to food consumption are measured as well which are, in our view, consistent with the notion that recipients are extremely poor.

GiveDirectly has received feedback from field staff and recipients that using housing materials as the targeting criteria systematically misses some households that are viewed within communities as comparably poor to those in thatch-roof houses.185

We have reservations about the approach of targeting people based on the materials their houses are made of:

• It is not clear to us that people in thatched-roof homes (eligible for transfers) are substantially and consistently poorer than people in iron-sheet-roofed homes with mud walls and floors (not eligible for transfers in a standard campaign). In community-based targeting pilots, GiveDirectly recipients identified households that did not meet GiveDirectly's standard targeting criteria but seemed comparably poor.186 GiveDirectly still feels that housing materials are an effective means of targeting the poorest of the poor, on average, in areas where it has worked to date. It is considering modifying its targeting criteria to catch some of the outlier cases in these regions.187
• To the extent that there are differences in income or wealth between residents of thatched-roof homes and those who live in iron-sheet-roofed homes, it seems possible that these differences come down to fortune/luck (i.e., people in iron-sheet homes have been more fortunate and thus able to afford iron sheets), but we also think it may come down to differences in choices regarding financial management (i.e., people in iron-sheet homes may have demonstrated better financial management and planning, thus allowing them to acquire iron sheets). If the latter is the case, there is a potential risk that GiveDirectly is systematically targeting the people who are less likely to use additional money well. GiveDirectly comments: "The most informative data available on this point are the differential impacts we’re seeing within the set of eligible households – specifically, poorer families seeing bigger impacts on nutrition while richer households see bigger impacts on tangible investment."188
• GiveDirectly's follow up surveys demonstrate that cash transfers can lead to tension between recipients and non-recipients.189 Though follow up surveys report low levels of tension and conflict, we would expect these to be underreported by recipients to GiveDirectly staff, a dynamic that GiveDirectly has seen play out in past cases.190 GiveDirectly conducted a small-scale study in Kenya to see whether more inclusive targeting criteria could reduce tension and conflict within villages. We find the results inconclusive (more). In addition, when we spoke with three field staff in Uganda, two of them suggested that it would be better for GiveDirectly's transfers to reach more people in a village, even if it meant reducing the size of a standard transfer. According to the Assistant Field Manager, the current targeting model causes bragging and unrest in the communities, potentially motivating those who don't benefit to steal from those who do. He said it would be better for GiveDirectly to provide transfers to everyone in a village, even if some transfers were small.191

Anecdotal evidence from GiveWell's site visit to Kenya

In November 2012, GiveWell staff visited Kenya to view GiveDirectly's program in the field. See our notes and photographs from the site visit. We visited 5 locations (3 in Siaya and 2 in Rarieda) where GiveDirectly had transferred funds or was in the process of enrolling recipients to receive funds. We visited approximately 15 households across the 5 locations (including 2 non-recipient households with metal roofs and cement walls and floors that did not qualify for GiveDirectly's program). For details on how homes we visited were selected, see this footnote.192 Note that when we visited, GiveDirectly was using thatched roofs and mud building materials as its criteria.

We would characterize the ~15 households we visited (as well as other households we saw while walking but did not speak with directly) as extremely poor. We summarize characteristics of these households as follows:

• Most homes are made up of three rooms.
• The main room is a sitting area. In the homes we visited, this room varied in size from approximately 10'x12' to 12'x20'. This room has 3 doors: one to the outside; one to what appeared to be a storage room; one to a bedroom. The husband and wife sleep in the bedroom and the children sleep in the storage room or in the kitchen.
• The kitchen is often a separate structure, most often thatched-roof (even for homes that have tin roofs). Some households have no kitchen structure and cook outside. Others have a small kitchen in place of a storage room.
• There are no doors in between rooms in the house, just hung curtains.
• The living room has many chairs and couches for sitting. They often almost fully cover the wall area (aside from doors). There are also coffee tables in the middle of the room. Poorer homes had less furniture; one home we saw had a single chair and a single broken table.
• People have wall hangings for decoration. The most common hanging we saw was old calendars (e.g., from 2003) that have pictures and can be used for decoration.
• Most houses had 1-2 kerosene lamps that provide light since they don't have electricity. One home (a non-recipient we visited) had two electric lights, which were powered with a solar panel.
• People we spoke with reported walking 5-20 minutes daily to obtain drinking water, which one recipient reported doing 1-2x per day.
• Most households owned a bicycle.
• Some households have radios (both of the non-recipients we saw had radios; one had a TV). Of the others, 3/14 had radios. These were most often powered with what looks like a car battery.
• Most people owned one cell phone pre-GiveDirectly.
• Households tend to own some livestock. Most commonly, we saw 1-2 cows and 4-5 chickens each. They reported selling the milk or saving it for personal use, and many mentioned being able to sell their cows in the future to pay for their kids to attend secondary school.

Note that among the households we visited, many had already received part or all of their transfer from GiveDirectly, so these characterizations are based on a selection of households that include some newly-built or renovated structures in addition to older structures. Given that some of the recipients we met used transfers to build larger houses or buy livestock, our observations would likely over-estimate the assets of each household pre-transfer.

In addition, the homes we saw from afar in villages we visited and homes we passed while driving in the area appeared to be at a similar level of extreme poverty.

Is GiveDirectly effectively targeting people who meet its criteria?

GiveDirectly's process for identifying and enrolling households is described above. It involves multiple unannounced visits by different staff to each recipient home in order to confirm that recipients meet the criteria. (That is, if someone were to temporarily occupy a mud and thatch home in order to be enrolled, they would be unlikely to be sure of being present for future re-checks.) We have examined data collected by GiveDirectly from its enrollment process (registration, backchecks, remote checks and audits) for most transfer campaigns; we have only spot-checked the data GiveDirectly shared with us in 2015.193

If the information collected about a household at different stages of enrollment is inconsistent, GiveDirectly staff revisit the household for an audit.194 GiveDirectly tracks the percentage of households found to be ineligible at the backcheck and audit stages on its website; between September 2013 and July 2015, 3.5% of recipients initially eligible after registration were found to be ineligible after the backcheck or audit stage.195 We believe this process to be generally effective at identifying and enrolling households that meet its criteria.

Does GiveDirectly have an effective process for getting cash to recipients?

Mobile money providers and distribution models

GiveDirectly transfers funds to recipients through mobile money providers. In Kenya, the mobile money provider, M-PESA, allows users to receive, send, deposit, and withdraw funds on their mobile phones. When withdrawing funds, recipients must present ID along with their mobile phone number and a user-specified M-PESA PIN number to an M-PESA agent.196 Users enter the amount they want to withdraw on their own phone, and after each transaction, they can see their remaining balance, reducing the ability of agents to defraud clients of funds.197 GiveDirectly has told us that recipients are generally able to withdraw cash from mobile money agents located in or near their villages.198

In Uganda, the agent network is less robust, so GiveDirectly has coordinated cash out days for recipients to withdraw funds after each payment.199 On cash out days, the mobile money provider that GiveDirectly works with will send an armoured vehicle with large amounts of cash, security personnel, and multiple agents to a location close to recipients' villages, so that recipients can easily come and withdraw their funds.200

GiveDirectly worked with two different mobile money providers in its pilot campaign in Uganda: EZEE Money and MTN (745 recipients were assigned to EZEE Money, 215 recipients to MTN).201 After assessing the relative performance of these two providers, GiveDirectly chose to work exclusively with MTN in the next campaign.202 MTN also has the advantage of having a more robust agent network than EZEE Money, so MTN recipients are somewhat less dependent on cash out days.203 MTN has similar security measures as M-PESA: a user must present ID to an agent before making withdrawals, provide their phone or SIM card, and enter their PIN number. Confirmation messages are sent after withdrawals.204 GiveDirectly is also testing the use of biometrics to increase security during cash out days.205

GiveDirectly recently tested working with a different payment provider (Centenary Bank) in Uganda and experienced difficulties.206 While GiveDirectly will continue to work with MTN in Uganda, it is also piloting a distributed cash out model (the same model used in Kenya).207 GiveDirectly hopes that communicating intensively to recipients about where the nearest MTN mobile money agents are will cause the distributed cashout model to work in Uganda.208

Staff fraud

The most significant issue that GiveDirectly has had in making sure that cash gets to recipients is the case of staff fraud in the Uganda pilot campaign. In mid-2014, GiveDirectly experienced a case of large-scale crime, when two of its field staff colluded with mobile money agents to defraud recipients of funds. The staff and mobile money agents were able to steal a total of $20,500 in the form of $20 deductions from 85% of recipients and $100 deductions from 15% of recipients.209 GiveDirectly found out about the fraud through follow-up calls to recipients, which were accelerated after a separate issue had been reported to GiveDirectly's hotline.210 GiveDirectly has taken multiple measures to address the vulnerabilities exposed by this case of fraud (see footnote for details).211 We consider fraud to be an ongoing risk to the success of GiveDirectly's programs, but feel that the risk is mitigated by these measures as well as by GiveDirectly's monitoring.

As GiveDirectly scales, we would expect greater awareness of its program and more attention to be paid by people outside of the villages in which it works. This could increase the risk of large-scale crime. The cash out days that GiveDirectly administers in Uganda seem to be particularly easy targets for large-scale theft, as there is a substantial amount of cash in one location (although it is our understanding that GiveDirectly's partners send security personnel to the cash out days to mitigate this risk).212 We are not aware of security measures that GiveDirectly has implemented to mitigate the risk of large-scale crime beyond its response to the staff fraud, although it has piloted a few measures.213 In addition to harming recipients, crime would likely cause delays for GiveDirectly's work.

Other issues

Despite the mobile money security measures, Lydia Tala, an Assistant Field Manager who has been responsible for making post-transfer phone calls to recipients in Kenya, reports that one of the most common client complaints is the belief that M-PESA agents are overcharging or stealing funds.214 Ms. Tala believes that these reports are incorrect and that approximately 10% of recipients are not fully aware of how to use M-PESA, so they may withdraw funds without checking their balance, ultimately being surprised when they have drawn down their account.215 GiveDirectly told us that it recognizes this issue and maintains a hotline to provide recipients with assistance in navigating the M-PESA system.216

In Uganda, some recipients have experienced delays in accessing their funds due to MTN not activating them immediately.217

Another issue that GiveDirectly has noted is that recipients who have not previously had a mobile phone or mobile money account are often less familiar with how to use them and how to keep their account information secure. GiveDirectly field staff explained that they provide training to recipients in how to use their phones and accounts, and reminders are also given at cash out days.218 Some of the recipients that GiveDirectly serves are not able to fully understand how to use the mobile money payments system on their own, or do not have the mobility to go to agents or cash out days to withdraw their funds.219 For these recipients, GiveDirectly finds a trustee or helper who aids them with their cash transfers; GiveDirectly tries to ensure that this person is someone the recipient trusts.220

After the transfers are sent, GiveDirectly also administers follow up surveys that ask recipients if they have collected their funds and if they had any trouble doing so.221 The percentage of recipients who report issues withdrawing funds is consistently low (<5%) across campaigns. See the table below for details. Follow up surveys also ask recipients what size of transfer they received. These amounts generally appear to vary slightly among cohorts of recipients. For example, in follow-up surveys of recipients in Kenya from 2014, recipients reported receiving various amounts between 37,000 KES – 40,000 KES.222 Other than the mobile phone purchase deduction, we do not know the causes of this variance.

How do recipients spend their cash, and how does this spending impact their lives?

Findings from the RCT

72% of treatment group households in the evaluation received just $287; the rest received $1,085.223 In the sections below, we use the outcomes from the larger transfer group unless otherwise specified, because GiveDirectly typically gives transfers of roughly $1,000. (More above. For every outcome, the larger transfer led to more spending compared to the smaller transfer with a few exceptions, where we have noted the outcome for the smaller transfer group below, and for tobacco and alcohol and indices of health and education, where the effects were not statistically different from zero. Though we report transfer sizes in exchange-rate adjusted terms, we report the outcomes in purchasing power parity (PPP) adjusted U.S. dollars.224

How GiveDirectly transfers were spent
Researchers collected data by surveying members of the treatment and control groups about their recent spending. All data that follows comes from participant self-reports. GiveDirectly recipients increased the value of their non-land assets and their monthly consumption.225 Their spending is broken down in more detail below.

• Total non-land assets.226 Receipt of large transfers increased households’ non-land assets by an average of $463 (95% CI: $378 to $549).227 The largest categories of asset increases were livestock ($131, 95% CI: $79 to $183), durable goods ($100, 95% CI: $71 to $129; primarily furniture), and savings ($18, 95% CI: $9 to $27).228 Households receiving transfers (small or large) were 23 percentage points (95% CI: 17% to 29%) more likely to have an iron roof than the control households.229 Though Haushofer and Shapiro 2013 doesn't report the change in likelihood for recipients of large transfers alone, recipients of large transfers were 23 percentage points (95% CI: 13% to 33%) more likely to have iron roofs at end-line than recipients of small transfers.230 Haushofer and Shapiro 2013 estimated that iron roofs cost about $564 USD PPP based on a survey of one respondent in each of 20 villages.231 GiveDirectly ran a survey that sampled a respondent from each of 20 villages and found that iron roofs cost $418 USD PPP on average.232 We do not know what explains this discrepancy.
• Business expenses. Households receiving large transfers spent about $13 per month (95% CI: $1 to $25) more than control households on business expenses, which were primarily made up of non-durable expenses on non-agricultural businesses.233 Recipients of small transfers also spent about $13 more per month (95% CI: $4 to $22).234
• Health expenditures. Recipients of large transfers spent about $3 (95% CI: -$1 to $6) per month more than control households on health expenditures.235 Recipients of small transfers also spent about $3 (95% CI: $1 to $5) more.236 This spending was also included within the estimate of spending on consumption, below.
• Education expenditures. Haushofer and Shapiro 2013 reports that treatment households receiving large transfers spent $1.89 (95% CI: $0.20 to $3.58) more than the control households on education expenditures and treatment households receiving small transfers spent $0.79 (95% CI: -$0.31 to $1.89) more.237 We're not sure of the time period over which this estimate is calculated. Haushofer and Shapiro 2013 also reports that treatment households receiving large transfers spent $16.26 (95% CI: -$6.50 to $39.02) more than control households on education expenditures in the past month and treatment households receiving small transfers spent $19.41 (95% CI: -$12.22 to $44.74) more.238 We're not sure if the difference between the two estimates is due to the difference in the samples used to calculate them (they have different sample sizes) or the different time periods over which they might be calculated or some other explanation.239 Education expenditures were also included within the estimate of spending on consumption, below.
• Consumption. Treatment households consumed about $51 more per month (95% CI: $32 to $70) than control households.240 About half of this additional consumption was on food.241 This additional consumption also included increased spending on social expenditures and various other expenditures.242
• Alcohol and tobacco. Treatment households did not increase their spending on alcohol or on tobacco.243

Impacts of GiveDirectly transfers on recipients

• Food security. At baseline, food security was low among participants.244 Program participants reported a 0.37 standard deviation (95% CI: 0.17 to 0.57) increase in a food security index over controls.245
• Health and education. The study did not detect an effect on indices of health and educational outcomes.246
• Revenue and profits. Receipt of large transfers lead to a $15 per month (95% CI: -$1 to $32) increase in total revenues and receipt of small transfers lead to a $17 (95% CI: $4 to $30) increase but neither resulted in a detectable increase in profits.247 We emphasize that these are very short-run effects and we do not know whether participants’ business investments might lead to profits in the longer run.

Researchers also considered more subjective measures of impact on recipients' quality of life:

• Psychological well-being. Treatment improved an index of psychological wellbeing by 0.45 standard deviations (95% CI: 0.25 to 0.65).248 There was no observable effect on cortisol for the treatment group as a whole although cortisol, an indicator of stress, was slightly lower in the large transfer group than the small transfer group, a difference that was statistically significant at the 10% level when controls were included in the model.249
• Female empowerment. Control households in treatment villages measure 0.23 standard deviations (95% CI: 0.05 to 0.41) higher on an index of female empowerment than control households in control villages.250 This suggests that cash transfers to a village unexpectedly empowered females in both recipient and non-recipient households. The researchers propose potential mechanisms for this effect, but are explicit that these measured results are surprising and warrant further investigation.251 Note that we report this result for the sake of comprehensiveness but would guess that it is more likely to be random than real.

Data from follow-up surveys

GiveDirectly staff surveyed recipients on how they used their cash transfers. This data was collected at different points in the transfer cycle of each campaign.252 We summarize the data from recent campaigns in Kenya and Uganda below. The spending data from Kenya covers portions of the Kenya 2M, Kena 1.2M, and Kenya rolling campaigns, and covers dates from February 2014 to September 2015. The spending data from Uganda covers some of the Uganda pilot campaign from October 2013 to April 2014; we have not seen more recent data from Uganda.253

Number of recipients who reported spending, by category

Amount of reported funds spent, by category

Anecdotal evidence from our site visit

In our site visit to GiveDirectly recipients in Kenya, we asked about the value of items commonly purchased with transfer funds.256 Recipients reported that their thatched-roofs frequently leak when it rains and require replacement every 3-4 months at a cost of 1,000 Kenyan shillings ($11.68 based on the exchange rate as of November 15, 2012257 ) as well as time/labor. One recipient also reported that when it rains, she moves her family and their belongings into other structures to stay dry. Recipients reported buying livestock as an investment/savings device, hoping that they could (a) use the milk from the cow or goat for additional income and (b) sell the cow or goat and any offspring in the future if/when they needed additional funds (for e.g., secondary school fees for their children which are approximately 15,000 Kenyan shillings per year258 ($175.13 based on the exchange rate as of November 15, 2012259 )).

We have not visited Homa Bay county. GiveDirectly has informed us that most potential recipients in Homa Bay county already have iron roofs.260 To date, our estimate of investment returns from GiveDirectly's cash transfers has been based on the return to buying an iron roof (due to this being a particularly common purchase). The fact that iron roofs are already common in Homa Bay raises questions about how recipients will spend transfers and what returns on their investments they will get.

Do the cash transfers cause problems and complications that offset their positive impact?

Below, we discuss questions about the possible negative effects of cash transfers. We also spoke with recipients and non-recipients about potential problems during our site visit to GiveDirectly's operations in Kenya in November 2012. For more, see our site visit notes.

Does distribution to some community members and not others result in jealousy, conflict, or related issues?

The RCT that Innovations for Poverty Action conducted of GiveDirectly's transfers in Rarieda found no significant effects of transfers on the rate of crime in treatment villages or on instances of physical, sexual, or emotional violence in treatment households as compared to control households in treatment villages.261

GiveDirectly surveys recipients (post-transfer) on the following questions:262

• Have you heard complaints about GD in your community? What complaints are you hearing? Who is upset/complaining? Who are they upset with?
• Has there been any shouting or angry arguments among people in your village about these transfers? If yes, describe.
• Has there been any violence, theft, or other crime in your village related to these transfers? If yes, describe.

GiveDirectly has sent us results from follow-up surveys conducted in multiple transfer campaigns. Below, we summarize the survey data from recent campaigns in Kenya and the pilot campaign in Uganda for some of the questions included in these surveys.

This table includes follow up survey data primarily from the Kenya 2M, Kenya 1.2M, Kenya rolling enrollment, and Kenya behavioral optimization campaigns (survey results are from 2014 and 2015) ) and from the Uganda pilot campaign, the Uganda 2M campaign, and the Uganda model variations campaign (survey results are from 2013, 2014, and 2015). Note that recipients may have been surveyed more than once and would therefore be included more than once in the data presented.263 Percentages reported in this table represent the number of recipients who are marked as having responded "yes" (that they had the issue) out of those for whom a response is recorded in the data.264

Note that GiveDirectly surveys only cash recipients, not non-recipients, and all data is self-reported.

We have found very limited information about jealousy and conflict related to other cash transfer programs, but one study that found small levels of hostility towards recipients of an unconditional wealth transfer in Uganda is discussed in our cash transfer intervention report.

We have also reviewed records of calls made to GiveDirectly's hotline from May 2012 – August 2015, which provide anecdotal evidence of tension and conflict caused by the cash transfers, according to recipient reports, including marital disputes, fraud committed by helpers, trustees, or family members, and Village Elders requesting funds from recipients.267 In the most recent complete hotline call data that we have seen (from October 2014; in 2015 we asked for sample data only), the most common type of adverse event recorded is household conflict, followed by theft.268 The number of issues reported was about 6% of the total households in the campaigns (though it is possible that single households account for more than one issue recorded).269

Do the cash transfers lead to more frequent or more serious criminal activity?

The RCT that Innovations for Poverty Action conducted of GiveDirectly's transfers in Rarieda found no significant effects of transfers on the rate of crime in treatment villages.270 It is possible that cash transfers cause more serious crimes (in terms of damages) even if they do not cause more crimes; this seems plausible given that cash transfers create an influx of resources into villages. GiveDirectly notes that crime could become a more serious problem as its program becomes larger and more well-known, but GiveDirectly does not expect to see significantly higher rates of crime in the near future.271

Examples of attempted and/or successful criminal activity relating to GiveDirectly cash transfers include:

• People stealing cash and cell phones from recipient households272
• People contacting recipients and posing as GiveDirectly staff to defraud recipients of funds273
• Mobile money agents defrauding recipients of funds274
• GiveDirectly staff defrauding recipients of funds (we discuss one particularly large case of this above).

To mitigate the risk of small-scale crime, GiveDirectly emphasizes ways that recipients can keep their mobile money accounts and phones secure.275 It does not communicate with recipients via text message and tells recipients of this policy in order to protect against mass attempts at fraud, and it follows up with recipients who report crimes to try to resolve the issues.276

Do grants distort incentives and decision-making?

We have not seen information on the question of whether individuals who live in the areas served by GiveDirectly change their behavior in order to increase their chances of receiving transfers – for example, by spending more time at home to increase their chances of being at home when GiveDirectly staff visit, or by choosing to live in poorer quality housing in hopes of receiving transfers. The one-off nature of transfers (recipients are not eligible for a second round of transfers) may help to mitigate these effects among past and current recipients, though there is information to suggest that some recipients believe transfers could be given again in the future.277

Another way in which grants may distort decisionmaking is if they are promised and not delivered in time (causing people to make plans that cannot be executed). We do not have data directly addressing this issue, but GiveDirectly provides some statistics on the speed with which transfers are received.278 In the Rarieda campaign, 67% (359 of 536) of recipients waited less than a month, 84% (448 of 536) waited 3 months or less, and 6% (34 of 536) waited 6 months or more. In the Siaya campaign (a later campaign), 188 of 193 recipients waited less than a month, and the remaining 5 waited 2-3 months.279

GiveDirectly told us that in its Kenya campaigns the key factor determining when a recipient receives funds is when he or she registers for M-PESA; recipients are told that they will not receive transfers until they have registered.280 In GiveDirectly's Kenya campaigns, approximately 49% of recipients have received their first transfer on time (within 9 weeks of being enrolled) and the average time for recipients between the census survey and their first payment was 71 days.281 GiveDirectly's records of calls to its Kenya hotline demonstrate that some recipients are delayed in registering for M-PESA or collecting transfers due to issues outside of their control (e.g., a recipient's SIM number was already registered to someone else's M-PESA account; another recipient reported that an agent mistakenly claimed that the recipient's account had expired).282

In Uganda, the agent networks of mobile money providers are not as robust, which means that recipients must either travel farther, on average, to reach an agent or must withdraw cash only on days when GiveDirectly has arranged for agents to visit the villages (which GiveDirectly announces in advance).283 These challenges and the lack of flexibility may hamper recipients' ability to execute on plans for how and when to use funds. GiveDirectly told us that so far, the vast majority of recipients have been able to collect their transfers, with a few delays of up to a few hours on days when transfers are scheduled due to agents needing to replenish their cash stocks.284 In the most recent Uganda campaign (Uganda model variations), 81% of recipients have received their transfers on time (within 15 weeks of enrollment) and 14% have so far had registration problems.285

Do grants distort local markets?

It seems possible to us that a large infusion of cash into an area could alter economic opportunities for both recipients and non-recipients. Such effects could be positive (for example by spurring investment and job creation or by increasing the availability of retail goods) or negative (for example, by leading primarily to local inflation). The limited evidence addressing this issue in the RCT of GiveDirectly's program in Rarieda and the broader literature on cash transfers points to no distortion. There is an ongoing RCT of GiveDirectly's program that is testing for macroeconomic effects.

Do cash transfers lead to large increases in spending on alcohol and tobacco?

The RCT of GiveDirectly's program in Rarieda did not find an increase in spending on alcohol or tobacco. As discussed in our intervention report on cash transfers, RCTs of other programs that report spending on alcohol or tobacco find no impact on spending or decreased spending on these goods.

Are the size and structure of the cash transfers well-thought-through and appropriate?

GiveDirectly’s standard model is to grant about $1,000 (USD) to households over approximately four months to one year, after which recipients become ineligible for future transfers.286 GiveDirectly has also experimented with different transfer sizes and structures, plans to continue doing so in the future.287 In the past, GiveDirectly has given the following rationale for the size of its transfers:288

We have reservations about the above reasoning:

• Regarding "fair." Pre-cash-transfer wealth/income differences between eligible and ineligible recipients may exist for a number of reasons; we don't believe it's warranted to assume that a fair world would see the two groups with the same wealth/income due to an equalizing transfer, and more to the point, we don't believe that the ineligible households are likely to see the situation as fair. In addition, we are concerned that by aiming to equalize eligible and ineligible households, GiveDirectly takes on a substantial risk of its calculations being off in a way that leads to eligible households becoming systematically better off than ineligible households, which could distort incentives and lead to conflict.
• Regarding "well-understood." GiveDirectly notes that its transfers are similar — in dollar terms — to those of government programs, but that they are likely much larger in "percentage of income" terms. We note that the cash transfer programs that have been studied to date seem to be in the range of 9-27% of recipients' annual consumption; by contrast, if GiveDirectly's recipients average $0.65 in daily per capita consumption and receive an average of $288 per person over the course of a year (see above), this implies that people receive an average of 121% of their annual consumption in the year in which they receive the transfer.289 The quote above states that the lower level of initial income is an argument for making the cash transfer larger, but to us, it also means that the risks of distorting incentives, causing conflict, etc. are likely to be greater than those of previously-studied programs, since the transfers are a substantially greater percentage of consumption. This issue is somewhat mitigated by the fact that GiveDirectly's transfers are designed as "wealth transfers" rather than as "income transfers": recipients receive funds over the course of a year and then become ineligible, whereas the government programs it alludes to have longer periods of eligibility. GiveDirectly has also told us that its decision to make larger transfers over a shorter period of time is based on recipients' reported preferences.290

Perspectives of recipients and field staff

During our site visit to Uganda in 2014, we spoke with a small number of recipients and field staff about the size of transfers. We asked whether people felt it would be better for GiveDirectly to keep the transfer size the same or reduce the transfer size but provide transfers to more people.291 4 out of 4 recipients said that the transfer size should stay the same (or be increased).292 One GiveDirectly Field Officer also held this view, saying that $1000 is enough to help someone advance but is not so much that it would distort incentives to work. Two other field officers suggested that it would be better for GiveDirectly's transfers to reach more people in a village, even if it meant reducing the size of a standard transfer.293

Merits of further research

GiveDirectly has considered experimenting with transfer size but does not view this as a high priority, in part because it feels that although further research on this question may improve GiveDirectly's program, it would be unlikely to influence other cash transfer programs.294 GiveDirectly is not concerned that people will run out of good uses of funds from $1000 transfers.295 The Rarieda RCT included both a $300 transfer treatment group and a $1000 transfer treatment group, but did not provide strong evidence on what the best transfer size would be, because of small sample sizes.296

Does GiveDirectly divert skilled labor away from other areas?

GiveDirectly predicted that by the end of 2015 it would have 89 total field staff members across Kenya, Uganda, and Rwanda: 4 Field Directors, 3 Field Managers, 13 Assistant Field Managers, and 69 Field Officers.297 GiveDirectly recruits Field Officers through referrals from peer organizations, postings at universities, and job advertisements. The application process involves an interview with a Field Director and a language competency exam. GiveDirectly reports that it receives approximately six times the number of resumes as openings for Field Officer positions.298 Regarding its field staff in Kenya, GiveDirectly explained that successful candidates generally have a college education and are paid approximately $12 per day, in addition to expenses for travel and lodging while working.299 GiveDirectly reported greater language heterogeneity in the areas in which it works in Uganda, which made it harder to hire qualified field staff who also had the necessary language skills.300 Based on GiveDirectly's current staffing situation, we do not see diversion of skilled labor as a serious concern.

What do you get for your dollar?

What percentage of GiveDirectly's expenses end up in the hands of recipients?

Cash grants make up 84.5% of GiveDirectly's all-time incurred expenses.301 This figure includes some fundraising costs that are expected to generate revenue in the future and excludes some of the costs of following up with recent recipients.302 Note that this an improvement from last year: by October 2014, 80.4% of GiveDirectly's total expenses had been transferred directly to recipients.303

When including projected future costs for current campaigns (transfers for still-to-be-enrolled recipients and the costs of sending and following those transfers) and core operations for 2016, the ratio of direct grants to total spending is 80.8%.304 Note that this analysis is based on financial documents GiveDirectly sent us before its transfer sizes were updated for inflation.305

Over the course of its existence, GiveDirectly has spent a total of $24.9 million (through June 2015).307 Below we break down GiveDirectly's total spending through June 2015, and its future budgeted costs through December 2016.308 Costs not included in GiveDirectly's total spending are at least some of the research costs of the independently-run studies of GiveDirectly's program (these costs are not funded by GiveDirectly)309 and the roughly $2.4 million that GiveDirectly has set aside as reserves to cover staff salaries in the event that GiveDirectly has a funding shortfall.310

Breakdown of GiveDirectly's total spending by category311

Does GiveDirectly offer a large amount of humanitarian impact per dollar?

We have not conducted a cost-effectiveness analysis that attempts to quantify the benefits of cash transfers in humanitarian terms. Instead, in comparing cash transfers to the interventions conducted by our other top charities, we have attempted to monetize some of the benefits of the latter, in particular the “developmental effects” of deworming and bednets. (In the case of the comparison with bednets, for instance, this means quantifying the estimated impact of bednets on later-in-life income of children through a comparison with the effects of deworming, and then subjectively comparing the cost per life saved with the value of that amount of money as a cash transfer.)

In practice, these calculations are highly sensitive to assumptions, especially regarding:

• the investment returns to cash transfers;
• how much confidence one places in the developmental impacts of deworming; and
• the subjective assessment of the relative value of averting child mortality and improving incomes.

We guess that in purely programmatic terms, and given our values, bednet distributions are more cost-effective than deworming, which is more cost-effective than cash transfers. However, we think there are plausible values for these assumptions that would permit any ordering of these three programs.

We encourage readers who find formal cost-effectiveness analysis important to examine the details of our calculations and assumptions, and to try putting in their own. To the extent that we have intuitive preferences and biases, these could easily be creeping into the assumption- and judgment-call-laden work we’ve done in generating our cost-effectiveness figures, and we’re not entirely confident that the figures themselves are adding substantial information beyond the intuitions we have from examining the details of them.

Our 2015 cost-effectiveness spreadsheet is available here. See also, our 2012 discussion of the cost-effectiveness of cash transfers and other interventions.

Is there room for more funding?

We are reasonably confident that GiveDirectly could effectively use an additional $15 million beyond what it already expects to receive and could potentially absorb up to an additional $75 million. In short:

• Estimated maximum: In the next year, GiveDirectly estimates that it can scale up so that it can enroll recipients at a distribution rate of $94 million in cash transfers per year. It estimates that it will cost up to an additional 10% ($9.4 million, total $103.4 million) to support the costs of enrollment, transferring funds, and follow-up.
• Cash on hand: GiveDirectly expects to have at least $30-35 million available for cash transfers in its 2016 budget year (March 2016 to February 2017). About half of this are funds GiveDirectly would like to hold until 2017 in order to attract matching funds.
• Other sources of funds: GiveDirectly expects to receive a $6 million grant at the end of 2015.
• Past spending: In recent months, GiveDirectly has enrolled recipients at a rate corresponding to transferring $16 million per year. Funds transferred to recipients has generally kept pace with commitments.
• Additional considerations: GiveDirectly has a track record of success in scaling its operations quickly. Recently, it grew its capacity for cash transfers by a factor of two in a year. It is not clear whether it will be able to continue this trend.

Details follow.

GiveDirectly's projections for 2016

GiveDirectly estimates that it can enroll recipients for a total of $94 million in cash transfers in its 2016 budget year (March 2016 to February 2017).313 This assumes that each Field Director can oversee disbursement of $17.25 million per year in Kenya and Uganda and $8 million in Rwanda.314 This represents the amount each Field Director could commit in the next year (i.e., specific recipients enrolled) rather than the amount they could transfer in that time. (We discuss below how this compares to GiveDirectly's past enrollment rate.) It also assumes that GiveDirectly will be able to hire two additional Field Directors (going from 4 to 6).315 GiveDirectly told us that it does not expect hiring Field Directors, or more junior staff, to be a challenge.316 GiveDirectly plans to first use additional funding for its campaigns in Kenya and Uganda, and only enroll additional recipients in Rwanda once its capacity in those two countries has been filled.317

Note that the above estimates refer only to GiveDirectly's room for more funding for cash transfers (the estimates exclude, for example, the cost of GiveDirectly's staff next year). GiveDirectly estimates that the costs of delivering the transfers will be approximately an additional 5-10% of what it transfers.

Available and expected funds

Absent additional funding due to a GiveWell recommendation, GiveDirectly expects to have $30-35 million available for cash transfers in its next budget year (March 2016 to February 2017):

• $16-19 million of a $25 million grant from Good Ventures. The grant is unrestricted, but GiveDirectly expects to use a portion to support fundraising (more below). GiveDirectly has told us that it expects to hold these funds at least through 2016 because it hopes to use the grant funds to attract matching funds from other funders, and such partnership agreements can take a significant amount of time to negotiate.318
• $0.6 million raised in July and August 2015.319
• $4 million GiveDirectly expects to raise between September 2015 and February 2016.320
• $6 million from a grant GiveDirectly expects to receive before the end of 2015.321
• $4 million from a grant for cash transfers in Rwanda.322

In addition, GiveDirectly held, as of June 2015:

• About $6 million committed to cash transfers for enrolled recipients.323
• $16 million designated for ongoing campaigns in Kenya and Uganda. GiveDirectly expected to commit all of these funds to recipients enrolled between June 2015 and February 2016.324
• $6-9 million of a $25 million grant from Good Ventures. This grant was unrestricted. GiveDirectly plans to use this portion of the grant to expand its fundraising capacity over the next few years.325
• Approximately $2.4 million that has been officially set aside for salary reserves and $1.1 million that has been officially designated for fundraising.326

Past enrollment rate

GiveDirectly's past rate of committing funds to recipients is much lower than its projected rate for 2015. Its enrollment rate in March 2015 to June 2015 implies a transfer rate of about $16 million per year.327 GiveDirectly has until recently been operating with two Field Directors,328 implying a commitment rate of $8 million per year per Field Director.

Rate of funds committed329

With a lag of about four months, distributed transfers have generally kept pace with committed transfers.330

Bottom line

There is significant uncertainty about the maximum funding GiveDirectly could use in 2016. GiveDirectly estimates that it can scale up to several times its current enrollment rate in the next year implying the ability to utilize an additional ~$75 million.331 Note that GiveDirectly has successfully scaled up over time, recently increasing its rate of transfers by about a factor of two in a year,332 but it is unclear if it will be able to continue this trend. If we assume a doubling of GiveDirectly's current rate of commitments, this implies that GiveDirectly would be able to commit $32 million in cash transfers in its 2016 budget year; excluding funds it may hold through 2016, GiveDirectly expects to have at least $15 million available for cash transfers in that year.

Risks to room for more funding

GiveDirectly believes it can grow extremely quickly. GiveDirectly has previously identified the following risks, which might impede its ability to grow as fast as it believes it can. We do not find any of the items below particularly concerning now given GiveDirectly's progress:

• Crime: Incidences of large-scale crime could cause delays and reduce GiveDirectly’s ability to transfer funds to recipients. The risk of crime could increase as GiveDirectly becomes better known in the regions in which it works. We discuss this risk more above. We consider this a low to moderate risk and plan to continue to check up on it.
• Government permissions: In order to expand into new areas, GiveDirectly must obtain permission from government officials at many levels. This process could be held up by an official who refused to grant permission, causing delays and possibly preventing GiveDirectly from expanding into an area indefinitely. GiveDirectly has attempted to mitigate this risk by networking with people with expertise in navigating such government relationships and could intervene if there were a problem.333 GiveDirectly feels that it now has a good understanding of the process for seeking government approvals and does not see this as a major risk.334 We do not consider this to be a limiting factor for FY 2016, as GiveDirectly has already obtained permissions to enroll a cumulative capacity of about 100,000 households across Kenya and Uganda.335
• Security: GiveDirectly notes that political violence and terrorism could hamper its ability to work in an area, and while these are risks in Kenya, they have not impacted Western Kenya (where GiveDirectly works) since 2008. GiveDirectly has attempted to mitigate this risk by working in multiple locations, so that it could shift its operations to Uganda if there were an issue.336 We know very little about security risks in Kenya and Uganda, but would guess based on GiveDirectly's assessment that it presents a low risk. As GiveDirectly continues to expand to other countries (e.g., Rwanda), we think this risk will be reduced by giving GiveDirectly more areas to redirect its work if necessary.
• Payment provider: Relying on one payment provider in each country introduces a risk that problems with the payment provider could cause delays. GiveDirectly feels that this risk is low, because if there were problems, it could switch to alternative providers.337 We would guess that this risk is low, as the mobile money providers that GiveDirectly uses are national networks, and GiveDirectly has identified alternatives. However, we note that GiveDirectly recently tried working with an alternative provider in Uganda (EZEE Money), and was not impressed by the partnership (more).338

Unrestricted vs. restricted funds

We prefer that GiveDirectly spend funds in the way that it believes will maximize its potential and, accordingly, do not recommend that GiveWell donors restrict their donations in any way. We plan to grant funds to GiveDirectly unrestricted (such that GiveDirectly may use funds for all purposes, including experimenting with its model and process and organizational capacity building). Donations made directly via GiveDirectly’s website can be designated for any county, Kenya, Uganda, or experimental.339

GiveDirectly as an organization

GiveDirectly is a relatively young organization. It was founded in 2009 when its founders were graduate students in economic development; Paul Niehaus, President and co-founder of GiveDirectly, is also an Assistant Professor of Economics at the University of California, San Diego.340 Professor Niehaus was on sabbatical from his teaching position and working full time on GiveDirectly in 2014-2015.341 He returned to his professorship in fall 2015.342

We believe GiveDirectly to be an exceptionally strong and effective organization:

• Self-evaluation: GiveDirectly has invested heavily in self-evaluation from the start, and furthermore, the study design of its Rarieda RCT was pre-registered for additional accountability and credibility. It continues to demonstrate a strong commitment to rigorous analysis of its work.
• Track record: Although it is relatively young, we feel that GiveDirectly's first few years have gone well: GiveDirectly has successfully accomplished its goal of transferring cash to extremely low-income people at roughly a 90% ratio. We have also seen GiveDirectly refine its process over the years and take thoughtful measures in response to problems that arise, demonstrating a commitment to continuous improvement.
• Communication: GiveDirectly has always communicated extremely clearly and directly with us and given thoughtful answers to our critical questions. Generally, GiveDirectly seems to come to conclusions that we find reasonable on key questions.
• Transparency: GiveDirectly appears to value transparency as much as any organization we’ve encountered. We have not seen it hesitate to share information publicly (unless it had what we consider a good reason).

More on how we think about evaluating the leadership of organizations at our 2012 blog post.

Sources

• 1
  • In Kenya: "We transfer your donation electronically to a recipient's cell phone

    We move the money from our US bank to our account with Safaricom's M-Pesa mobile payment system using a foreign exchange broker. We then transfer money from our M-Pesa account to the recipient's M-Pesa account. As a security measure we only transfer funds to a recipient if the name in our records matches the name on the national ID document he or she used to register for M-Pesa. The recipient gets an SMS text message reminding him or her of the transfer and then collects the transfer from a local M-Pesa agent, who is typically a shopkeeper in the recipient's village or in the nearest town. The recipient transfers his or her electronic balance to the agent's phone in return for cash." GiveDirectly, How it works 2013

  • "In Uganda, we use MTN's mobile payment system to send recipients their transfers. The recipient can collect cash from an MTN agent, who organizes a “payday” at the village level each month. Community-nominated “monitors” help us oversee the payday." @GiveDirectly, What we do - Operating Model@, Uganda tab.

• 2

  Paul Niehaus and Carolina Toth, conversation with GiveWell, September 7, 2015.

• 3
  • GiveDirectly, Update for GiveWell, May 2015
  • GiveDirectly, Update for GiveWell, September 2015, pg. 3
  • Paul Niehaus and Carolina Toth, conversation with GiveWell, September 7, 2015

• 4
  • "Design Lab Nature of impact: In transferring funds, GD generates knowledge that expands or improves existing $150B+ cash market… Related org. priorities: Research studies" GiveDirectly, Update for GiveWell, September 2015, pg. 3
  • " Rigorous, experimental evaluation of impacts is rare among nonprofits. GiveDirectly collaborates with third-party researchers to measure the impacts of cash transfers and answer complex design questions. Researchers are fully independent and independently-funded. We report the results of our evaluations and also announce studies in progress before the data are in, so that we can be held accountable for the results." GiveDirectly - Evidence - Research at GiveDirectly
  • This understanding is from many conversations with GiveDirectly and following GiveDirectly's progress over time.

• 5
  • "Benchmark Nature of impact: Success of GD and cash transfers generally creates pressure for transparency, evidence,and for other approaches to prove they outperform cash… Related org. priorities: Institutional partnerships, real time transparency." GiveDirectly, Update for GiveWell, September 2015, pg. 3
  • This understanding is from many conversations with GiveDirectly and following GiveDirectly's progress over time.

• 6
  • "We transfer recipient households roughly $1,000, or around one year's budget for a typical household." GiveDirectly, What We Do - Operating Model, Overview tab
  • GiveDirectly recently adjusted the size of its transfers for inflation, so that it now transfers ~$1,040 to recipients in Kenya and ~$875 to recipients in Uganda.
    • Paul Niehaus and Carolina Toth, conversation with GiveWell, September 7, 2015
    • "In Kenyan Shillings, they are 10K, 50K, 50K (assuming no phone purchase). Ugandan shillings is 3.2M but division into different transfers hasn't been decided yet." Carolina Toth, email to GiveWell, September 25, 2015
    • XE currency converter, Kenya shillings to US dollars, September 25, 2015
    • XE currency converter, Uganda shillings to US dollars, September 25, 2015
  • In recent campaigns, the enrollment process has taken slightly more than 2 months on average, and in past campaigns transfers have been sent over a period of 8 months in Kenya and a period of 10 months in Uganda for the vast majority of recipients. This is the basic schedule followed in standard campaigns including Google, Kenya 2M, Kenya 1.2M, and Kenya rolling enrollment. (Example of transfer schedule: GiveDirectly, Google transfer schedule, July 2013) Conversation with Carolina Toth, GiveDirectly, November 20, 2014
  • The amount of time that it takes for total transfers to be sent has varied between campaigns. Campaigns that involved an experiment (like Rarieda and Nike) and older campaigns like Siaya have differed from the standard schedule. Conversation with Carolina Toth, GiveDirectly, November 20, 2014
  • "We've now changed the distribution schedule to be token transfer, two months until first lump sum, and then two more months until the final lump sum. This puts current households on a 4 month schedule." Carolina Toth, email to GiveWell, November 10, 2015

• 7
  • "GiveDirectly uses objective criteria to determine which households are eligible for transfers. We target households which do not have solid (cement or iron) walls, floors or roofs in their houses. Our research shows this criterion is highly correlated with acute poverty." GiveDirectly, FAQs 2012
  • "How do you decide to whom to give cash transfers? We aim to find the poorest possible recipients while using criteria that are objective and easy to understand." GiveDirectly, FAQs 2015

• 8
  • "We transfer recipient households roughly $1,000, or around one year's budget for a typical household." GiveDirectly, What We Do - Operating Model, Overview tab
  • GiveDirectly recently adjusted the size of its transfers for inflation, so that it now transfers ~$1,040 (USD) to recipients in Kenya and ~$875 (USD) to recipients in Uganda. Paul Niehaus and Carolina Toth, conversation with GiveWell, September 7, 2015
    • "In Kenyan Shillings, they are 10K, 50K, 50K (assuming no phone purchase). Ugandan shillings is 3.2M but division into different transfers hasn't been decided yet." Carolina Toth, email to GiveWell, September 25, 2015
    • XE currency converter, Kenya shillings to US dollars, September 25, 2015
    • XE currency converter, Uganda shillings to US dollars, September 25, 2015
  • In recent campaigns, the enrollment process has taken slightly more than 2 months on average, and transfers have been sent over a period of 8 months in Kenya and a period of 10 months in Uganda for the vast majority of recipients. This is the basic schedule followed in standard campaigns including Google, Kenya 2M, Kenya 1.2M, and Kenya rolling enrollment. (Example of transfer schedule: GiveDirectly, Google transfer schedule, July 2013) Conversation with Carolina Toth, GiveDirectly, November 20, 2014
  • The amount of time that it takes for total transfers to be sent has varied between campaigns. Campaigns that involved an experiment (like Rarieda and Nike) and older campaigns like Siaya have differed from the standard schedule. Conversation with Carolina Toth, GiveDirectly, November 20, 2014
  • In November 2015, GiveDirectly informed us that it now aims to send transfers within 4 months. Carolina Toth, email to GiveWell, November 10, 2015

• 9
  • GiveDirectly plans to continue to adjust its transfer sizes for inflation in the future, possibly reassessing every 6 months. Paul Niehaus and Carolina Toth, conversation with GiveWell, September 7, 2015
  • "In Kenyan Shillings, they are 10K, 50K, 50K (assuming no phone purchase). Ugandan shillings is 3.2M but division into different transfers hasn't been decided yet." Carolina Toth, email to GiveWell, September 25, 2015
  • XE currency converter, Kenya shillings to US dollars, September 25, 2015
  • XE currency converter, Uganda shillings to US dollars, September 25, 2015
  • GiveDirectly, Inflation analysis - Kenya. Note that the inflation analysis does not include information on how the transfer size should be adjusted in Uganda, so we are not sure how GiveDirectly settled on the Uganda amount.

• 10

  GiveWell Household size analysis. Note that this data is based on a small sample from one of GiveDirectly's first campaigns (the Siaya campaign).

• 11
  • GiveWell Household size analysis
  • Mean daily per capita consumption among eligible households = $0.65 (GiveDirectly, Offering Memorandum (January 2012) Pg 24.)
  • Note that GiveDirectly estimates that it transfers approximately $200 (USD) per person, although we have not asked where this estimate comes from: "How much do recipients get? We send each recipient household roughly $1,000 over the course of one year, or $200 per household member for the average household." GiveDirectly, FAQs 2015

• 12
  • "In Kenyan Shillings, they are 10K, 50K, 50K (assuming no phone purchase). Ugandan shillings is 3.2M but division into different transfers hasn't been decided yet." Carolina Toth, email to GiveWell, September 25, 2015
  • "1 KES = 0.00947200 USD" XE currency converter, Kenya shillings to US dollars, September 25, 2015. So 10K KES = $94.72, and 50K KES= $473.6.

  These transfers are sent over a period of approximately 4 months (previously over about 8 months).

  • In recent campaigns, the enrollment process has taken slightly more than 2 months on average, and transfers have been sent over a period of 8 months in Kenya and a period of 10 months in Uganda for the vast majority of recipients. This is the basic schedule followed in standard campaigns including Google, Kenya 2M, Kenya 1.2M, and Kenya rolling enrollment. (Example of transfer schedule: GiveDirectly, Google transfer schedule, July 2013) Conversation with Carolina Toth, GiveDirectly, November 20, 2014
  • The amount of time that it takes for total transfers to be sent has varied between campaigns. Campaigns that involved an experiment (like Rarieda and Nike) and older campaigns like Siaya have differed from the standard schedule. Conversation with Carolina Toth, GiveDirectly, November 20, 2014
  • In November 2015, GiveDirectly informed us that it now aims to send transfers within 4 months. Carolina Toth, email to GiveWell, November 10, 2015

• 13
  • "In Kenyan Shillings, they are 10K, 50K, 50K (assuming no phone purchase). Ugandan shillings is 3.2M but division into different transfers hasn't been decided yet." Carolina Toth, email to GiveWell, September 25, 2015
  • Before transfer sizes were adjusted for inflation, recipients in Uganda received $100 each month. GiveDirectly, Uganda transfer schedule
  • 3.2M Ugandan shillings = $874. XE currency converter, Uganda shillings to US dollars, September 25, 2015. Divided into 9 transfers is about $97 per transfer.

• 14

  Conversation with Paul Niehaus, President, and Joy Sun, COO, Domestic, GiveDirectly, October 16, 2013

• 15

  Carolina Toth, email to GiveWell, November 10, 2015

• 16
  • GiveDirectly, Monthly operations report, August 2015
  • GiveDirectly works in Bukedia district in Uganda, and Siaya and Homa Bay counties in Kenya. GiveDirectly, Update for GiveWell, May 2015, pg. 3.

• 17
  • GiveWell, GiveDirectly financials 2015 Sheets: "2015-Total spend + efficiency" and "Campaign efficiency FY 2014"
  • GiveDirectly, Monthly operations report, August 2015
  • Carolina Toth, email to GiveWell, September 25, 2015

• 18As of August 31, 2015
• 19As a % of total funds committed as of June 2015
• 20This campaign is also labelled as Ke-RCT on GiveDirectly documents (for example, see the Campaign efficiency FY 2014 tab of GiveWell, GiveDirectly financials 2015
• 21This campaign is also labelled as Ke-200K on GiveDirectly documents (for example, see the Campaign efficiency FY 2014 tab of GiveWell, GiveDirectly financials 2015
• 22This campaign is also labelled as Ke-Nike on GiveDirectly documents (for example, see the Campaign efficiency FY 2014 tab of GiveWell, GiveDirectly financials 2015. It was funded by the Nike Foundation.
• 23This campaign is also labelled as Ke-Google on GiveDirectly documents (for example, see the Campaign efficiency FY 2014 tab of GiveWell, GiveDirectly financials 2015. It was funded by Google.
• 24This campaign is also labelled as Ke-201307 on GiveDirectly documents (for example, see the Campaign efficiency FY 2014 tab of GiveWell, GiveDirectly financials 2015
• 25This campaign is also labelled as Ug-201305 on GiveDirectly documents (for example, see the Campaign efficiency FY 2014 tab of GiveWell, GiveDirectly financials 2015
• 26This campaign is also labelled as Ke-201311 on GiveDirectly documents (for example, see the Campaign efficiency FY 2014 tab of GiveWell, GiveDirectly financials 2015
• 27This campaign is also labelled as Ke-201403 on GiveDirectly documents (for example, see the Campaign efficiency FY 2014 tab of GiveWell, GiveDirectly financials 2015.
• 28This is an approximation. We've taken the sum of all households registered for the Ke-201403 campaign through August 31, 2015 according to GiveDirectly, Monthly operations report, August 2015 (15,328 households) and subtracted the number of households that Carolina Toth told us were enrolled in the Ke-201407 campaign (585 households), because it is our understanding that the households for the Ke-201407 campaign are enrolled with the households from the Ke-201403 campaign (Carolina Toth, email to GiveWell, September 25, 2015). However, the number of households for the Ke-201407 campaign was given to us in late September, and therefore this number may not accurately reflect what was true as of August 31, 2015.
• 29In discrete campaigns, GiveDirectly had clear start and stop dates for the campaign, as a set amount of funding was allocated to the campaign. In a rolling model, there is no set end date; enrollment continues month-to-month. GiveDirectly can continue enrolling recipients on its rolling model as long as it has funds to transfer. Studies that differ significantly from GiveDirectly's standard model (more below are still implemented separately in discrete campaigns. Paul Niehaus and Carolina Toth, conversation with GiveWell, September 7, 2015
• 30This campaign is also labelled as Ug-201404 on GiveDirectly documents (for example, see GiveDirectly, Monthly operations report, August 2015. We've also included the small extension of this campaign, "Ug-201404(ext)" in our numbers for the Uganda 2M campaign in the "Summary of campaigns" table.
• 31This campaign is also labelled as Ke-201407 on GiveDirectly documents (for example, see GiveWell, GiveDirectly financials 2015
• 32This number reflects what we knew as of late September, and may not be directly comparable to the other numbers in this column which are from the end of August. Carolina Toth, email to GiveWell, September 25, 2015
• 33 This campaign is also labelled Us-201411. Carolina Toth, email to GiveWell, September 22, 2015. See also GiveDirectly, Rockefeller index insurance update, July 2015 and Paul Niehaus and Carolina Toth, conversation with GiveWell, September 7, 2015. It was funded by the Rockefeller Foundation.
• 34This campaign is also labelled as Ug-201503 on GiveDirectly documents. GiveWell, GiveDirectly financials 2015. See also Carolina Toth, email to GiveWell, September 25, 2015. Variations being tested: The use of biometrics (e.g. fingerprint or eye scanning) for enhanced security; a new mobile bank partner (Centenary instead of MTN); and distributed cash out model (instead of the payday model). Paul Niehaus and Carolina Toth, conversation with GiveWell, September 7, 2015.
• 35This campaign is also labelled Ug-201509. Carolina Toth, email to GiveWell, September 22, 2015.
• 36In discrete campaigns, GiveDirectly had clear start and stop dates for the campaign, as a set amount of funding was allocated to the campaign. In a rolling model, there is no set end date; enrollment continues month-to-month. GiveDirectly can continue enrolling recipients on its rolling model as long as it has funds to transfer. Studies that differ significantly from GiveDirectly's standard model (more below are still implemented separately in discrete campaigns. Paul Niehaus and Carolina Toth, conversation with GiveWell, September 7, 2015
• 37

  "Kenya was selected due to 1) the robustness of M-Pesa as a mobile banking platform 2) large population of target poor with access to mobile tech" Email from Paul Niehaus, President, GiveDirectly, and Joy Sun, COO, Domestic, GiveDirectly, November 18, 2013

• 38

  Conversation with Paul Niehaus, President, and Rohit Wanchoo, Director, GiveDirectly, March 18, 2013

• 39

  The employee heard about GiveDirectly through an NPR news piece and became excited about the prospect of GiveDirectly working in Rwanda. GiveDirectly and the donor were then able to secure interest from Google in co-funding such a project through mutual connections. Paul Niehaus and Carolina Toth, conversation with GiveWell, May 28, 2015

• 40

  "While the results of this project will be applicable in many places, Rwanda is an ideal setting in which to conduct it. Rwanda features a) a burgeoning mobile money landscape, with 17% mobile money penetration in mid-2013 as per National Bank of Rwanda and cell phone penetration at 72% in September 2014, b) a sizeable population of poor households (45% below the poverty line), and c) a track record of prioritizing transparent and innovative aid programming." GiveDirectly, Rwanda technical application

• 41

  Carolina Toth, email to GiveWell, November 10, 2015

• 42
  • Paul Niehaus, GiveDirectly Founder, email to GiveWell, November 20, 2012
  • "GiveDirectly selected a region of the [second] country to set up initial operations in based on poverty data." Conversation with Paul Niehaus, President, and Joy Sun, COO, Domestic, GiveDirectly, July 18, 2013

• 43

  GiveDirectly sent us poverty data, which we internally reviewed. We have not written up the process we used to review this data.

• 44
  • Paul Niehaus, GiveDirectly Founder, conversation with GiveWell, October 22 2012
  • "County level – we look at data on poverty, population density, presence of poverty-focused NGOs (which we try to avoid), and security."
    @GiveDirectly, Operational Process Overview@ Pg 1.

• 45

  GiveDirectly, Siaya poverty data by location

• 46

  "Factors that informed decision to locate initial campaign in […] County: poverty rate, […] logistical ease for set up activities and cross-country management, […] minimum security for staff" GiveDirectly, Uganda targeting data, July 22, 2013

• 47

  GiveDirectly sent us poverty data, which we internally reviewed. We have not written up the process we used to review this data.

• 48

  "Recently we have been relying on sublocation-level poverty data and looking at village names and locations to omit villages that would be too urban for our current model. We have also been working with satellite data to approximate village-level poverty. " Conversation with Carolina Toth, GiveDirectly, November 20, 2014

• 49

  See @GiveDirectly, Siaya Village Index@ for the calculations that GiveDirectly did to create "poverty scores" for different villages in Siaya. The weights placed on each indicator (in constructing the index) were determined using the process described in @GiveDirectly, Village Targeting Regression@: the more detailed "poverty scores" from GiveDirectly's Rarieda study were regressed on indicators such as "village population," "number of boreholes," etc.

• 50

  Conversation with Piali Mukhopadhyay, COO, International, GiveDirectly, October 22, 2013

• 51

  Conversation with Piali Mukhopadhyay, COO, International, GiveDirectly, October 22, 2013

• 52
  • Conversation with Piali Mukhopadhyay, COO, International, GiveDirectly, October 22, 2013
  • GiveDirectly prefers not to enroll only part of a village. GiveDirectly slightly extended its Uganda 2M campaign after realizing that its target enrollment numbers would have caused it to stop enrollment halfway through a village. See the Ug-201404 (ext) campaign in GiveDirectly, Monthly operations report, August 2015. Paul Niehaus and Carolina Toth, conversation with GiveWell, September 7, 2015

• 53

  In the Kenya 1.2M campaign, GiveDirectly selected villages by manually estimating the proportion of thatch- to iron-roof homes with satellite imagery. In the Kenya rolling enrollment campaign, GiveDirectly used a machine learning algorithm to estimate thatch-iron proportions at the village level based on satellite imagery. In the Uganda 2M campaign, GiveDirectly relied on parish-level census data with poverty measures, as well as mobile money coverage. GiveDirectly, Update for GiveWell, April 2014, Pg 3.

• 54
  • In Rarieda (the site of the RCT and the first transfers GiveDirectly provided), GiveDirectly sought poorer districts (based on 2005 census data) that were in places with sufficient accessibility, M-PESA usage, population density to make it more convenient, proximity to Innovations for Poverty Action (the RCT implementer) offices and where there would be a sufficient number of potential recipients in this district (# of thatched roof houses). GiveDirectly chose to work in Rarieda District because it was slightly poorer than the nearby Siaya district. In Rarieda, there are slightly more than 300 villages, and GiveDirectly conducted a census in each village to determine the number of eligible (thatched-roof) and ineligible households in each. It then selected the 100 with the highest proportion of thatched-roof to non-thatched-roof households, and randomly selected 60 of those to serve as the treatment and control groups in its trial. Faizan Diwan, Innovations for Poverty Project Associate, conversation with GiveWell, November 8, 2012
  • GiveDirectly chose to work in Siaya District, the location of the three other sets of transfers GiveDirectly has initiated, because it shared local administration with the Rarieda District, making expansion easier; it chose not to remain in Rarieda because it did not want to overlap in areas in which the RCT was being conducted. Using administrative data, it chose 3 locations within Siaya that it believed had the highest poverty levels. It then carried out the process described in the main text to rank the 100 villages in these locations. GiveDirectly was not able to contact all Village Elders to obtain data (staff estimate they reached 85 out of 100) and it excluded villages whose Village Elders it was not able to reach. In June 2012, it selected the 7 villages which its model ranked as highest poverty to receive the Siaya transfers. For the project funded by the Nike Foundation, GiveDirectly selected the next 36 villages in its ranked list of 100 villages in Siaya District, as this number of villages provided a sample size sufficient to meet their target size. In the Google transfers, GiveDirectly is continuing to work down the list of 100 villages to target those not already targeted in the Siaya or Nike Foundation transfers. Piali Mukhopadhyay, COO, International, GiveDirectly, conversation with GiveWell, November 8, 2012

• 55
  • Seeking government approvals for GiveDirectly cash transfer campaigns

    By now, GiveDirectly understands well the process for seeking government approvals in Kenya and does not see acquiring approvals as a major risk. GiveDirectly said that in Kenya it is important to maintain relationships with government officials at the county and district levels; district commissioners introduce GiveDirectly to chiefs, and chiefs introduce GiveDirectly to Village Elders. In Uganda, there are no counties, so GiveDirectly coordinates with a few people at the district level to acquire approvals, and from there connect with officials at the local level […] As part of its networking in Kenya, GiveDirectly staff have met with the Permanent Secretary for Devolution and Planning. This is someone who could help GiveDirectly acquire permission to work in new counties in Kenya. Conversation with Piali Mukhopadhyay, GiveDirectly, October 20-21, 2014

  • "When entering a new area, the COO meets with a series of officials to explain the project, obtain permission, and establish a relationship in case any problems arise:
    • District Commissioner
    • Chief
    • Assistant chiefs
    • Village elders"

    @GiveDirectly, Operational Process Overview@, Pg 1.

  • In Uganda, GiveDirectly had to spend more time meeting with officials early on in the process, because there is a greater bureaucratic structure than in Kenya. This engagement tapered off after the early stage, though GiveDirectly remains in touch with officials by phone to let them know when GiveDirectly has started conducting field activities. Conversation with Piali Mukhopadhyay, COO, International, GiveDirectly, October 22, 2013

• 56
  • "Govn’t relations: Signed MOUs with local officials to maximize buy-in and formalize relationship" GiveDirectly, Update on process changes, August 28, 2013
  • Typical approval process
    Kenya:
    • "Seek buy-in from County and District Commissioner and sign written agreement w/district
    • Ensure Governor’s office and relevant Country admin officials informed of expansion activities"

    Uganda:

    • "Attain approval letter from Resident District Commissioner for natl renewal
    • Attain approval letters from RDC, District Security Officer, District Intelligence Officer, and District Development Officer for local renewal"

    GiveDirectly, Update for GiveWell, September 2014, Pg 9.

• 57

  @GiveDirectly, Operational Process Overview@, Pg 2.

• 58

  Village meetings were implemented for the first time as part of the Google campaign. GiveDirectly, GW Q&A, April 2013

• 59

  Data collected during census can be found in enrollment databases, for example: GiveDirectly, Kenya 1.2M enrollment database

• 60
  • "Enumerators enter villages, engage a local to serve as guide for the day, and
    enumerate all households living in the village, noting which homes are eligible." @GiveDirectly, Operational Process Overview@, Pg 2.
  • Data collected during census can be found in enrollment databases, for example: GiveDirectly, Kenya 1.2M enrollment database
  • "Household level – we enroll households living in mud and thatch homes." @GiveDirectly, Operational Process Overview@, Pg 2.
  • "Dropped mud walls as eligibility requirement" GiveDirectly, Update for GiveWell, September 2014
  • "We've now completed our targeting pilots in Kenya, and have selected a new targeting criteria for Homa Bay. We looked at a range of alternatives from simple proxy means tests to more complicated rules to subjective scoring to community-based methods, and evaluated them along dimensions of accuracy, perceived fairness, gameability, and cost. The results were useful in the near term for Homa Bay, and also provide a framework and starting point for criteria evaluation in new geographies going forward." Carolina Toth, email to GiveWell, October 20, 2015

• 61
  • In the Rarieda campaign, a census of all households was completed before enrollment. However, in this case, the census process was implemented by Innovations for Poverty Action, the organization conducting the randomized controlled trial of GiveDirectly's program, as opposed to GiveDirectly staff.
  • In the Siaya campaign, GiveDirectly did not complete a full census of all households in the village. Instead, staff went to the Village Elder and asked him or her to take them to each thatched-roof household in the village to verify that it was eligible for the transfer. GiveDirectly discovered that some Village Elders were assisting friends or family members in pretending that they live in thatched-roof houses so that they could receive transfers, so it now has as a policy to work with a village member who can serve as a guide rather than to rely on the Village Elder. Piali Mukhopadhyay, COO, International, GiveDirectly, conversation with GiveWell, November 7, 2012; Conversation with Piali Mukhopadhyay, COO, International, GiveDirectly, October 22, 2013
  • In the Nike campaign, GiveDirectly staff requested that the Village Elder lead them to eligible households: women aged 18-19, living in thatched-roof homes. Piali Mukhopadhyay, COO, International, GiveDirectly, conversation with GiveWell, November 7, 2012

• 62

  "Enrollment. A second, distinct enumerator returns to enroll households identified as eligible, give them a SIM card and instructions on how to register if needed." @GiveDirectly, Operational Process Overview@ Pg 1.

• 63
  • "Enrollment. A second, distinct enumerator returns to enroll households identified as eligible, give them a SIM card and instructions on how to register if needed." @GiveDirectly, Operational Process Overview@ Pg 1.
  • See process description in @GiveDirectly, Operational Process Overview@ Pg 3.
  • Data collected during registration can be found in enrollment databases, for example: GiveDirectly, Kenya 1.2M enrollment database

• 64

  GiveDirectly gave cash transfer recipients the option of spending some of the money that they receive to buy a phone provided by GiveDirectly. Conversation with Paul Niehaus, President, and Michael Faye, Director, GiveDirectly, October 6, 2012

• 65
  • "Do recipients need to have a mobile phone to participate?
    No. Households need at least a SIM card to participate, and we give SIM cards to households that do not already have one. We also give recipients the option of purchasing a phone from us at bulk rates in order to make it easier to communicate with them. When recipients choose this option we deduct the value of the phone from their transfer. Historically the large majority of recipients in both Kenya and Uganda have chosen to buy a phone." GiveDirectly, FAQs 2015
  • "How do you prevent corruption?
    The two main corruption risks that typically arise in transfer programs involve (a) manipulation of the list of eligible recipients and (b) diversion of transfers sent to eligible recipients. We address the first through a comprehensive audit process, using multiple independent checks to ensure that recipients are eligible and have not been charged bribes to get on the list. These checks include in-person visits by different staff members, in-person audits by senior management, remote audits of image and satellite image data, and phone calls with each recipient, all prioritized using modern analytics. We address the second through identity-matching between our records and those of our payment providers, through comprehensive follow-up calls to ensure money is reaching the intended recipients, and in some cases through direct staff monitoring of cash-out points." GiveDirectly, FAQs 2015

• 66
  • GiveDirectly told us that in Uganda, it is possible to purchase ID cards that can be formally approved by the signature of one's Local Councilperson. GiveDirectly helped recipients obtain ID cards by purchasing the cards, sending field staff to villages to take photographs of recipients, printing the photographs for the ID cards at a local printer, working with Local Councilpeople to approve the cards, and arranging for recipients to collect their cards.

    GiveDirectly told us that the mobile money agents did not charge a fee to visit villages to register recipients, and that GiveDirectly field staff were present to supervise the process. Conversation with Piali Mukhopadhyay, COO, International, GiveDirectly, October 22, 2013

  • The logistics are significantly harder in Uganda than in Kenya. For example, when GiveDirectly enters a new village in Uganda, over 90% of recipients need SIM cards because they did not previously have cell phones, and about 70-80% of recipients need national IDs. GiveDirectly coordinates registration drives for people to get national IDs - they buy national ID booklets, print a photo of each recipient to put in the booklets, and have the Local Councilperson stamp the booklets to approve them. GiveDirectly was able to reach 85-90% of people through these registration drives, returning IDs within about 1 week of visiting eligible households. In the Uganda 2M campaign, there are 9 villages, and GiveDirectly was able to put them all through the national ID registration process within 1 month, so that 90% of eligible households were ready to receive transfers when payments started. (The remaining households will receive their transfers on a delayed schedule, once they complete registration.) Conversation with GiveDirectly field staff, October 20-21, 2014
  • GiveDirectly also facilitates recipients signing up for a mobile money account with MTN by having an agent visit the villages. Once recipients have signed up for an account, MTN generally activates their line within 2-3 weeks. By the time the backcheck team visits villages, most recipients’ lines are active. Conversation with GiveDirectly field staff, October 20-21, 2014

• 67

  GiveWell site visit to GiveDirectly, October 2014
  For example, see GiveDirectly, Uganda pilot enrollment database - Akumure. Note that even households that have passed the "Registration" phase (starts in column AB) must still go through the "Back check" phase (the "Census-registration comparison" (column EF), "Photograph check" (column EP), and "Remote Check" (column ES)), and a possible "Audit" (column FH) before their final enrollment status is determined (columns FK and FL).

• 68
  • As noted above, in the Siaya campaign, GiveDirectly's staff initially engaged the Village Elder to lead them to eligible households. Once they had been led to an eligible home, GiveDirectly staff provided the household with a SIM card and enrolled them in the program. Thus, in Siaya, there was one fewer backcheck than exists in the current process. Piali Mukhopadhyay, COO, International, GiveDirectly, conversation with GiveWell, November 7, 2012
  • In the Nike campaign, GiveDirectly staff requested that the Village Elder lead them to eligible households, so there was one fewer backcheck in this process as well. (Piali Mukhopadhyay, COO, International, GiveDirectly, conversation with GiveWell, November 7, 2012)

• 69

  Data collected during backchecks can be found in enrollment databases, for example: GiveDirectly, Kenya 1.2M enrollment database

• 70

  "Back-checking. Another enumerator, distinct from the census and enrollment workers, revisits each enrolled household to check that they eligible, didn’t have to pay a bribe to enroll, etc." @GiveDirectly, Operational Process Overview@, Pg 3.

• 71

  "How do you prevent corruption?
  The two main corruption risks that typically arise in transfer programs involve (a) manipulation of the list of eligible recipients and (b) diversion of transfers sent to eligible recipients. We address the first through a comprehensive audit process, using multiple independent checks to ensure that recipients are eligible and have not been charged bribes to get on the list. These checks include in-person visits by different staff members, in-person audits by senior management, remote audits of image and satellite image data, and phone calls with each recipient, all prioritized using modern analytics." GiveDirectly, FAQs 2015

• 72

  "[In Kenya] the follow up team sends some of its members to do audits and staff are not pulled from prior enrollment teams." Conversation with Carolina Toth, GiveDirectly, November 20, 2014.

• 73

  Conversation with GiveDirectly field staff, October 20-21, 2014 (Note that this point was not included in the notes from this conversation). We are not sure if this means that GiveDirectly staff are auditing their own work. It is possible that staff members were assigned to houses that they did not initially register, or that staff members enroll so many houses that it's difficult to remember the household's details from census or registration. We are not sure who is currently chosen to conduct the audits in Uganda.

• 74

  GiveDirectly's current procedure for identifying households to audit:

  • GiveDirectly collects information about recipients during the first three stages of a campaign: census, registration, and backcheck. Some information, such as recipient name, GPS location, housing materials, and identifying photograph, is collected at more than one stage and then checked for mismatches. (These checks are currently conducted using Excel but will eventually be automated through Segovia technology. One exception is comparing identifying photographs of recipients, which is done using a crowdsourced work platform called Mechanical Turk.)
  • Each mismatch in recipient information is assigned a certain weight depending on how likely it is to be an indication of gaming. (GiveDirectly said that it determined the likelihoods of various mismatches indicating gaming by conducting an analysis of the mismatches present in past cases of gaming. We did not review this analysis.) GiveDirectly said that the mismatches with the highest weights are mismatches in identifying photographs and housing materials.
  • Each recipient is assigned a total mismatch "score" (the composite of all their weighted mismatches). Recipients with scores above a certain level are audited.

  Conversation with GiveDirectly field staff, October 20-21, 2014

• 75

  Conversation with GiveDirectly field staff, October 20-21, 2014
  GiveWell site visit to GiveDirectly, October 2014

• 76
  • In prior campaigns, GiveDirectly audited all households for which there was a discrepancy in enrollment data collected, and tended to exclude recipients if there was reason to believe that the potential recipient did not meet GiveDirectly's eligibility criteria. (Ms. Mukhopadhyay told us that she had reviewed cases in which some members of the village told GiveDirectly's staff that enrollees were not eligible because they did not live in thatched-roof homes. In these cases, Ms. Mukhopadhyay decided to exclude these potential recipients.) Piali Mukhopadhyay, COO, International, GiveDirectly, conversation with GiveWell, November 7, 2012.
  • In enrollment rounds completed before November 2012, GiveDirectly did not use all of these steps [data comparisons] as "hard checks." Recipients were still visited by two independent field teams who verified eligibility, but potential recipients could remain eligible even if they failed one of these steps. Paul Niehaus, GiveDirectly Founder, email to GiveWell, November 20, 2012.
  • More information in our November 2013 review.

• 77
  • Conversation with Carolina Toth, GiveDirectly, November 20, 2014
  • For an example of GiveDirectly's commitment to potential recipients, see GiveDirectly, Blog post, August 21, 2015

• 78

  People have to be at home at registration and backcheck so that they can be given phones, have safety information explained etc. If they are not at home during the first attempt to visit, we re-visit them several times until they can be found." Conversation with Carolina Toth, GiveDirectly, November 20, 2014

• 79
  • Paul Niehaus and Carolina Toth, conversation with GiveWell, September 7, 2015
  • The summary results of these quality audits for Kenya can be seen in some monthly operations reports (e.g. GiveDirectly, Monthly operations report, August 2015 GiveDirectly has sent us results from these audits for the Ug-201503 campaign as well, which we have briefly reviewed but have not published. GiveDirectly, Uganda model variations quality audits - census and GiveDirectly, Uganda model variations quality audits - registration

• 80

  GiveDirectly, Siaya enrollment database
  GiveDirectly, Google enrollment database
  GiveDirectly, Nike enrollment database
  GiveDirectly, RCT Enrollment Database
  GiveDirectly, Kenya 2M enrollment database, September 2013
  GiveDirectly, Uganda pilot enrollment database - Kosile
  GiveDirectly, Uganda pilot enrollment database - Kawo
  GiveDirectly, Uganda pilot enrollment database - Kanyamutamu
  GiveDirectly, Uganda pilot enrollment database - Akumure
  GiveDirectly, Kenya 1.2M enrollment database
  GiveDirectly, Kenya 2M census results, July 2013
  GiveDirectly, Kenya rolling campaign enrollment database - Homa Bay
  GiveDirectly, Kenya rolling campaign enrollment database - Siaya
  GiveDirectly, Uganda 2M campaign enrollment database
  

• 81

  "We use electronic payment systems; typically, recipients receive an SMS alert and then collect cash from a mobile money agent in their village or nearest town." GiveDirectly, What We Do - Operating Model, Overview tab

• 82
  • "In Kenyan Shillings, they are 10K, 50K, 50K (assuming no phone purchase). Ugandan shillings is 3.2M but division into different transfers hasn't been decided yet." Carolina Toth, email to GiveWell, September 25, 2015
  • "1 KES = 0.00947200 USD" XE currency converter, Kenya shillings to US dollars, September 25, 2015. So 10K KES = $94.72, and 50K KES= $473.6.
  • Before transfers were adjusted for inflation, recipients in Kenya would receive an initial transfer of $50, then two transfers of $475. Example of campaign with GiveDirectly's past transfer structure in Kenya: GiveDirectly, Google transfer schedule, July 2013

• 83
  • In Uganda, transfers are sent in ten equal installments. Conversation with GiveDirectly, April 8, 2014
  • Carolina Toth, email to GiveWell, November 10, 2015

• 84
  • In Uganda, transfers are sent in ten equal installments, and GiveDirectly works with the mobile money providers to coordinate pay out days, as their agent networks are less robust than those in Kenya. Conversation with GiveDirectly, April 8, 2014
  • In November 2015, GiveDirectly clarified that the transfers in Uganda had been split into nine installments, not ten, and that the new campaign in Uganda was split into three installments, like the campaigns in Kenya. Carolina Toth, email to GiveWell, November 10, 2015

• 85
  • GiveDirectly is making the following changes to its structure and procedures in Uganda:
    • Separating jobs that were previously done by one person (e.g., GiveDirectly has moved the complaint hotline and followup calls to an office in Kamapala, so that the employees in charge of these are in a different part of the country and do not know the field staff). GiveDirectly’s larger network in Kenya is already structured this way.
    • Increasing payday audits by the Field Director from about 25% to 100% of paydays (2 or 3 per month).
    • Conducting real-time phone spot-checks, i.e. calling recipients during payday to make sure the event is going smoothly and that recipients are receiving the correct amounts.
    • Using MTN Mobile Money (MTN) instead of EZEE Money for more transfers. The network of EZEE Money agents is very limited, so the only feasible option for recipients was to withdraw funds on paydays. MTN has a larger network of agents, so while it is still more convenient for recipients to withdraw on paydays, recipients can seek other options if they prefer.
    • Building a network of local, English-speaking informants (e.g., journalists, well-respected farmers). Several people in civil service roles have told Dr. Niehaus that it is important to build such a network. Having English-speaking informants may have helped prevent the fraudulent translation that occurred in this case.

    Conversation with GiveDirectly, September 5, 2014

  • Our observations of a cash out day in Uganda: GiveWell site visit to GiveDirectly, October 2014

• 86

  Paul Niehaus and Carolina Toth, conversation with GiveWell, September 7, 2015

• 87

  Conversation with Piali Mukhopadhyay, COO, International, GiveDirectly, October 22, 2013

• 88

  Conversation with Piali Mukhopadhyay, COO, International, GiveDirectly, October 22, 2013

• 89
  • Piali Mukhopadhyay, COO, International, GiveDirectly, conversation with GiveWell, November 7, 2012
  • "For Uganda 201404, the plan is to do monthly short surveys, and then long follow up after payday 2 and 9 or 10." Email from Carolina Toth, GiveDirectly, November 14, 2014
  • Carolina Toth, email to GiveWell, November 10, 2015

• 90
  • GiveDirectly, Uganda pilot follow up data, April 2014
  • GiveDirectly, Kenya follow up data, November 2014

• 91
  • "For Uganda 201305, we did long follow up surveys on a rolling basis and most recipients were followed up with twice-- this is the data you've already received. At the end of 201305, there was going to be one short survey, and this survey was accelerated as part of the fraud investigation. These are labelled _short and _short_crosscheck and are attached. There were no monthly short surveys. We've not yet decided when/if to fill in the rest of the short surveys for those not reached during the fraud investigation." Email from Carolina Toth, GiveDirectly, November 14, 2014
  • "For Uganda 201404, the plan is to do monthly short surveys, and then long follow up after payday 2 and 9 or 10." Email from Carolina Toth, GiveDirectly, November 14, 2014
  • In early transfer campaigns, GiveDirectly transferred a first, full installment to recipients before its first call; in its Google campaign, GiveDirectly implemented the initial small transfer to enable it to identify problems before transferring the larger amount. Piali Mukhopadhyay, COO, International, GiveDirectly, conversation with GiveWell, November 7, 2012.

• 92
  • GiveDirectly, Rarieda verification stats
  • GiveDirectly, Siaya verification stats
  • GiveDirectly, Siaya follow-up data - disaggregated
  • GiveDirectly, Nike follow-up data - disaggregated
  • GiveDirectly, Google follow-up data - disaggregated (LS - long)
  • @GiveDirectly, Rarieda Top-up verification (short)@
  • GiveDirectly, Rarieda verification (top ups), May 26, 2013
  • GiveDirectly, Siaya verification, June 15, 2013
  • GiveDirectly, Google verification, September 2013
  • GiveDirectly, Nike verification (final), September 2013 (based on full survey given to recipients after whole transfer was sent)
  • GiveDirectly, Nike verification (combined), May 2013 (based on full survey given to recipients early in campaign, which GiveDirectly later stopped administering because it was too time-consuming with monthly transfers Conversation with Carolina Toth, Field Director, GiveDirectly, October 24, 2013)
  • GiveDirectly, Nike verification (short version), June 2013 (based on abbreviated survey given to recipients after each part of the transfer was sent, in place of "combined" survey Conversation with Carolina Toth, Field Director, GiveDirectly, October 24, 2013)
  • GiveDirectly, Kenya follow up data, November 2014
  • GiveDirectly, Uganda pilot follow up data, April 2014
  • GiveWell, GiveDirectly follow up surveys summary - Uganda, September 2015
  • GiveWell, GiveDirectly follow up surveys summary - Kenya, September 2015

• 93

  Piali Mukhopadhyay, COO, International, GiveDirectly, email to GiveWell, November 23, 2012.

• 94
  • GiveDirectly, Kenya hotline log, July 2013 (May 2012 – July 2013)
  • GiveDirectly, Follow-up tracker, July 2013 (July – October 2013)
  • GiveDirectly, Follow-up tracker, October 2014 (January 2013-September 2014). This file also contains issues reported through other channels, such as to field staff during field visits and during follow up surveys.

• 95
  • GiveDirectly told us that for the Kenya 2M campaign, its enrollment field staff conducted a short survey with anyone who approached the field staff to complain that they had been unfairly or mistakenly skipped during the census. Though [in campaigns prior to and including Kenya 2M] GiveDirectly [did] not add recipients after the census [had] been conducted, it intends to continue carrying out the surveys for future campaigns, as a way of tracking complaints, recognizing potential issues with the census, and assessing changes intended to improve GiveDirectly's census process. Conversation with Carolina Toth, Field Director, GiveDirectly, October 24, 2013. We have not reviewed the results of this survey. As of November 2015, GiveDirectly informed us that it still conducts these surveys, and that people who are skipped in error have an opportunity to be enrolled. Carolina Toth, email to GiveWell, November 10, 2015
  • Starting with the Kenya rolling enrollment campaign in early 2014, GiveDirectly now adds into the enrollment process households that complain about having been skipped at census. Conversation with Carolina Toth, GiveDirectly, November 20, 2014

• 96
  • In the Rarieda campaign, a census of all households was completed before enrollment. However, in this case, the census process was implemented by Innovations for Poverty Action, the organization conducting the randomized controlled trial of GiveDirectly's program, as opposed to GiveDirectly staff.
  • In the Siaya campaign, GiveDirectly did not complete a full census of all households in the village. Instead, staff went to the Village Elder and asked him or her to take them to each thatched-roof household in the village to verify that it was eligible for the transfer. GiveDirectly discovered that some Village Elders were assisting friends or family members in pretending that they live in thatched-roof houses so that they could receive transfers, so it now has as a policy to work with a village member who can serve as a guide rather than to rely on the Village Elder. Piali Mukhopadhyay, COO, International, GiveDirectly, conversation with GiveWell, November 7, 2012; Conversation with Piali Mukhopadhyay, COO, International, GiveDirectly, October 22, 2013
  • In the Nike campaign, GiveDirectly staff requested that the Village Elder lead them to eligible households: women aged 18-19, living in thatched-roof homes. Piali Mukhopadhyay, COO, International, GiveDirectly, conversation with GiveWell, November 7, 2012
  • Village meetings were implemented for the first time as part of the Google campaign. GiveDirectly, GW Q&A, April 2013

• 97

  GiveDirectly FY 2011 Form 990

• 98

  "GD has had one person working full-time since January of 2011. Jeremy Shapiro was full time during 2011 and Piali Mukyopadhyay is full-time now." GiveDirectly, clarifications on GiveWell's draft review of GiveDirectly

• 99

  Conversation with Paul Niehaus, President, and Rohit Wanchoo, Director, GiveDirectly, March 18, 2013

• 100
  • Conversation with Paul Niehaus, President, and Joy Sun, COO, Domestic, GiveDirectly, July 18, 2013 (unpublished)
  • GiveDirectly, Update for GiveWell, July 2013
  • Conversation with Paul Niehaus, President, and Joy Sun, COO, Domestic, GiveDirectly, October 16, 2013
  • Conversation with Piali Mukhopadhyay, COO, International, GiveDirectly, October 22, 2013

• 101

  GiveDirectly, Estimate of personnel 2015

• 102
  • "Introduced Project Associate role to strengthen focus on productivity, QC and professional development and increase FD leverage (Ke)" GiveDirectly, Update for GiveWell, April 2014
  • GiveDirectly changed the name of Project Associates to Field Managers in 2015. GiveDirectly, Estimate of personnel 2015

• 103

  GiveDirectly, Estimate of personnel 2015

• 104
  • "Greater leverage for FD to work on tech integration, high-level analysis (e.g. “smart” audits)" GiveDirectly, Update for GiveWell, September 2014, Pg 6.
  • GiveDirectly changed the name of Senior Field Officer to Assistant Field Manager in 2015. GiveDirectly, Estimate of personnel 2015

• 105

  GiveDirectly, Estimate of personnel 2015

• 106
  • Email from Paul Niehaus, President, GiveDirectly, and Joy Sun, COO, Domestic, GiveDirectly, November 18, 2013
  • In Uganda, GiveDirectly hired some of the highest performing FOs from the pre-transfer FO staffs to stay on as FOs who would conduct follow-up surveys. (Conversation with Piali Mukhopadhyay, COO, International, GiveDirectly, October 22, 2013). It is our impression that FOs who conduct follow-up surveys are often hired out of the pre-transfer FO staffs.

• 107

  GiveDirectly, Estimate of personnel 2015

• 108

  "Three members of GiveDirectly‘s board of directors (Paul Niehaus, Michael Faye, and Chris Hughes) are planning to start a for-profit technology company, Segovia, aimed at improving the efficiency of cash transfer distributions in the developing world. Segovia plans to sell software to developing-country governments for use in implementing their cash transfer programs." (From GiveWell's update on GiveDirectly, June 20, 2014).

• 109

  Paul Niehaus and Carolina Toth, conversation with GiveWell, May 28, 2015

• 110

  GiveDirectly expects efficiency and transparency gains from Segovia to be roughly on par with the gains from other types of organizational software (e.g. new budgeting software). Paul Niehaus and Carolina Toth, conversation with GiveWell, September 7, 2015

• 111
  • GiveDirectly, Update for GiveWell, July 2014, Pg 5.
  • Paul Niehaus and Carolina Toth, conversation with GiveWell, September 7, 2015

• 112

  Paul Niehaus and Carolina Toth, conversation with GiveWell, September 7, 2015

• 113

  "Based on conversations with policymakers, GiveDirectly has identified gaps in the evidence-base for cash transfers that currently limit policymakers’ ability to implement cash transfer programs or to do so effectively. GiveDirectly has spoken with policymakers in Kenya and Indonesia, as well as representatives of the UK’s Department for International Development (DFID). The leading question that came out of these conversations was about the macroeconomic impacts, or “general equilibrium effects,” of cash transfers when conducted at a national scale. This includes factors such as enterprise structure, prices, local public finance, and local government activities. This question is the primary motivation for GiveDirectly’s top research priority: a study of general equilibrium effects." Conversation with GiveDirectly, July 7, 2014, pg. 2.

• 114

  "Objective:

  • Understand macro-economics impacts of transfers at scale (in-flation, job creation, etc.)
  • Measure impacts over a long time horizon (e.g., [less than[sic]] 5 years)

  Status:

  • Started baseline, with long term follow up mechanisms in place
  • Not fully funded– facing a gap of ~8M

  Partners:

  • Edward Miguel, Berkeley
  • Johannes Haushofer, Princeton

  Potential impact:

  • Increase government use of CT programs
  • Increase support for our particular model in proving LT impact"

  GiveDirectly, Update for GiveWell, September 2014, Pg 13

• 115

  "This study will potentially include long-term follow up as well, to address a separate question raised by policymakers about the long-term impacts of cash transfers. Professor Miguel previously worked on a study of the impacts of deworming (Miguel and Kremer 2004) that involved follow-up over a period of ten years and obtained a high response rate, so he has experience in setting up effective systems for tracking study participants over time." Conversation with GiveDirectly, July 7, 2014, pg. 3

• 116

  GiveDirectly, GE study design, Pgs. 4-5

• 117

  Baseline data collection was slower than expected. This forced GiveDirectly to delay its initial token transfers to recipients (the baseline surveys must be conducted before recipients receive any funds). Paul Niehaus and Carolina Toth, conversation with GiveWell, September 7, 2015

• 118
  • GiveDirectly, GE research and measurement plan, pg. 6.
  • Baseline data collection was slightly slower than expected, which meant that GiveDirectly had to delay some transfers (so that researchers could complete the baseline survey before recipients had received cash). Paul Niehaus and Carolina Toth, conversation with GiveWell, September 7, 2015
  • Midline data was scheduled to be collected from November 2014 to early 2016. GiveDirectly, GE research and measurement plan, pg. 6.

• 119

  GiveDirectly, GE study design, Pg. 2.
  Carolina Toth, email to GiveWell, November 10, 2015

• 120

  "Paul Niehaus, GiveDirectly’s Co-Founder and President, will serve as a Principal Investigator on the general equilibrium effects study. To mitigate potential bias when GiveDirectly staff are involved in research on the impacts of its programs, GiveDirectly has decided on a few safeguards: the research team conducting the study will 1) preregister their plans for measurement and analysis 2) use a (non-GiveDirectly staff) third party for measurement, and 3) include academic PIs who are not involved in GiveDirectly and have a reputation for honesty." Conversation with GiveDirectly, July 7, 2014, Pgs 2-3.

• 121

  "Objective:

  • Measure impact of providing information on spending options
  • Measure impact of getting to choose when and how to receive cash"

  GiveDirectly, Update for GiveWell, September 2014 Pg 13

• 122

  Conversation with Stuart Skeates, GiveDirectly, October 20-21, 2014 (unpublished)

• 123
  • GiveDirectly notes:
    • "We have replaced IPA on this project with an RA hired by ideas42 + GD field staff, for cost reasons
    • This represents our first in-house data collection, and a change in policy: whenever treatments are variations on how to do cash (rather than: does cash work at all?) we do not face a conflict of interest and can be involved in data collection"

    GiveDirectly, Update for GiveWell, July 2014

  • "Partners:
    • Anandi Mani, Warwick
    • Sendhil Mullainathan, Harvard
    • Anuj Shah, Chicago Booth"

    GiveDirectly, Update for GiveWell, September 2014 Pg 13

• 124
  • Conversation with Stuart Skeates, GiveDirectly, October 20-21, 2014 (unpublished)
  • GiveDirectly, Update for GiveWell, May 2015, pg. 5.

• 125
  • GiveDirectly, Update for GiveWell, April 2014 Pg 6
  • Paul Niehaus and Carolina Toth, conversation with GiveWell, September 7, 2015

• 126

  "Objective:

  • Test if informal contracts can help further reduce domestic violence and improve female empowerment

  Status:

  • Small pilot, spring 2015
  • If successful, grow into a more large-scale project

  Partners:

  • Simone Schaner, Dartmouth
  • Jessica Leight, Williams"

  GiveDirectly, Update for GiveWell, September 2014, Pg 13.

• 127

  "Test if informal contracts can help further reduce domestic violence and improve female empowerment" GiveDirectly, Update for GiveWell, September 2014, Pg 13.

• 128

  GiveDirectly, Update for GiveWell, May 2015, pg. 5.

• 129

  "Objective:

  • Test if informal contracts can help further reduce domestic violence and improve female empowerment

  Status:

  • Small pilot, spring 2015
  • If successful, grow into a more large-scale project

  Partners:

  • Simone Schaner, Dartmouth
  • Jessica Leight, Williams"

  GiveDirectly, Update for GiveWell, September 2014, Pg 13.
  

• 130
  • GiveDirectly, Update for GiveWell, September 2015, Pg. 10.
  • The government of Uganda started a large cash transfer program called Social Assistance Grants for Empowerment (SAGE), which provides $20 monthly transfers to eligible people in Uganda. The program currently serves 100,000 households in 17 districts and has plans to scale up; it is not currently active in Bukudea, where GiveDirectly operates. The government of Uganda is working with the mobile money provider MTN to build the capability to use biometric authentication (fingerprinting) for transactions and account access. GiveDirectly is interested in running a pilot of biometric authentication with its own cash transfer recipients who are serviced by MTN. Conversation with GiveDirectly, April 8, 2014, Pg 5.

• 131
  • GiveDirectly, Update for GiveWell, September 2015, Pg. 10
  • GiveWell, GiveDirectly follow up surveys summary - Uganda, September 2015

• 132

  GiveDirectly, Update for GiveWell, September 2015, Pg. 10

• 133

  Paul Niehaus and Carolina Toth, conversation with GiveWell, September 7, 2015

• 134

  GiveDirectly, Aspirations study proposal

• 135

  GiveDirectly, Aspirations study proposal

• 136

  Carolina Toth, email to GiveWell, November 10, 2015

• 137
  • "We conducted a randomized controlled trial (RCT) of the unconditional cash transfer program implemented by the NGO GiveDirectly in Western Kenya between 2011 and 2012, in which poor rural households received unconditional cash transfers through the mobile money system M-Pesa." pg. 1
  • "In each chosen village, GD conducted a census, usually with the help of the village
    elder, which identified all households in the village that met this targeting criterion. Among the
    eligible households, treatment households were chosen randomly (details are described in Section 2). Households were aware that recipients would be chosen by lottery, but the actual selection was done privately by means of random number generation." pg. 4
  • "We are grateful to the study participants for generously giving their time; to Marie Collins, Faizan Diwan, Conor Hughes, Chaning Jang, Bena Mwongeli, Joseph Njoroge, Kenneth Okumu, James Vancel, and Matthew White for excellent research assistance; to Innovations for Poverty Action for implementation." pg. 1

  Haushofer and Shapiro 2013 Policy Brief

• 138

  GiveDirectly, Rarieda transfer schedule, August 2013

• 139

  GiveDirectly, Nike instrument

• 140

  GiveDirectly, Final report Nike girls study, Pg 3.

• 141
  • GiveDirectly, Final report Nike girls study
  • Jean Junior, The Perspectives of Young Women in Siaya County, Kenya: Their Lives and Their Thoughts on Cash Transfer Programs

• 142

  Johannes Haushofer and Paul Niehaus, DIL Demonstration Proposal
  UCSD, Policy Design and Evaluation Lab, "Tracking the Impact of GiveDirectly Transfers with Mobile Surveys in Kenya"

• 143

  UCSD, Policy Design and Evaluation Lab, "Tracking the Impact of GiveDirectly Transfers with Mobile Surveys in Kenya"

• 144

  Johannes Haushofer and Paul Niehaus, DIL Demonstration Proposal

• 145
  • In its standard model, GiveDirectly provides cash transfers only to the households that have thatch roofs. GiveDirectly experimented with more inclusive targeting in 19 randomly selected villages, in which nearly all households received transfers (all except those made from fully permanent materials such as cement walls and iron roofs). GiveDirectly compared these villages to 18 villages in the same region where standard targeting was applied. The factors being compared were cases of conflict/tension reported in follow-up surveys and focus groups, and instances of gaming that were discovered by GiveDirectly field staff throughout the cash transfer process. Conversation with GiveDirectly, April 8, 2014
  • “GiveDirectly: As of March 2013, GiveDirectly had received $790,000 from GiveWell donors designated as “flexible funds.” This includes a $500,000 gift from Good Ventures.

    The research question we are most interested in is whether providing cash transfers to all households in a village, rather than targeting the poorest households, could reduce tension and improve social outcomes of the transfer campaigns.

    In order to address this question, we’ve created 3 groups of randomly assigned villages for GiveDirectly’s most recent campaign in Kenya:

    • Villages in which no households will receive transfers
    • Villages in which only mud-wall and thatch-roof households will receive transfers
    • Villages in which nearly all households will receive transfers (all households with mud walls and thatch or metal roofs will receive transfers, only households with cement walls and metal roofs will be excluded)

    We are currently finishing up enrollment for this campaign, so transfers will be sent soon. We plan to collect data by administering our standard phone surveys, which include questions about tension, disagreements with neighbors, etc. We expect to receive the first round of data within the next month or two.

    The “flexible funds” received from GiveWell donors are going to be used for transfers to villages in group 3, including households with mud walls and either thatch or metal roofs.”

    Conversation with Paul Niehaus, President, and Joy Sun, COO, Domestic, GiveDirectly, August 27, 2013

  • In GiveDirectly's Kenya 2M campaign, "saturation villages" are villages in which all households with mud walls and thatch or iron roofs will receive transfers. These households make up about 85% of the households in saturation villages. The other 15% of households are built with all man-made materials (e.g., cement walls and iron roof) and will not receive transfers. Conversation with Carolina Toth, Field Director, GiveDirectly, October 24, 2013

• 146
  • "Quantitative analysis is inconclusive: some metrics favorable in saturation villages (e.g. fewer overall complaints), others less favorable (e.g. larger % asked to pay a bribe)
  • FGDs suggests that conflict levels were low across both categories of villages (mainly rumors and awkwardness) and that when faced with the choice we have to make, people prioritize the poorest, who they feel are more deserving"

  GiveDirectly, Saturation analysis Pg 1

• 147
  • GiveDirectly, Update for GiveWell, July 2014, Pg 11
  • In Kenya, GiveDirectly experimented with a community-based targeting process, whereby residents gave input on households that they felt were deserving of transfers but had been excluded by GiveDirectly’s criteria. GiveDirectly felt that to do this process well required significant resources (staff time) and that the benefits were not worth the costs. In addition, some of the villages involved in this experiment gave feedback that they would prefer for GiveDirectly to make the decisions about targeting. Conversation with GiveDirectly field staff, October 20-21, 2014, Pg 4.

• 148
  • "Index insurance for individuals builds resilience and increases investment for
    stallholder farmers in a more effective way than traditional insurance
    • — Remote triggers for payout (e.g., measurements at rain stations) prevent moral hazard and decrease cost of determining payout vs. traditional insurance
    • — Index-insured households in Ghana significantly increased investment in
      agricultural inputs vs. control (Karlan et al 2013)" GiveDirectly, Rockefeller index insurance update, July 2015, pg. 3.
  • "GiveDirectly & Rockefeller co-created an innovative approach to index insurance delivery, embedding it into social protection programs already at massive scale… Governments could embed index insurance into these programs by allowing beneficiaries to choose to exchange part of their status-quo benefits for insurance" GiveDirectly, Rockefeller index insurance update, July 2015, pg. 4.

• 149

  "GiveDirectly constructed a cash transfer program simulating Gov’t of Kenya’s flagship Hunger Safety Net Program (HSNP)

  • ~$100 in two tranches, two months apart
  • GiveDirectly negotiated a 100-day index-based crop insurance policy with a local commercial insurer (APA) based on standard, pre-existing offerings
  • Beneficiaries of the cash transfer program were given a choice of how much of benefit to receive as cash v.s. insurance premiums
  • Neutrally framed as a choice of benefits (v.s. a purchase decision)
  • Priced in 0.1 acre increments
  • Half offered approximately actuarial price (200 Ksh / 0.1 acre), half offered 50% subsidy (100 Ksh / 0.1 acre)

  GiveDirectly, Rockefeller index insurance update, July 2015, pg. 5.

• 150
  • "Take-up rates, even without subsidy, were higher than in other subsidized schemes." GiveDirectly, Rockefeller index insurance update, July 2015, pg. 8.
  • "Average acquisition costs were 75% lower than commercial insurer’s,
    and incremental costs were 94% lower" GiveDirectly, Rockefeller index insurance update, July 2015, pg. 7.

• 151

  Paul Niehaus and Carolina Toth, conversation with GiveWell, September 7, 2015

• 152

  "Objective: Build expertise in a set of targeting competencies to rapidly identify poor locations and households across diverse settings, and structure an evaluation framework for targeting effectiveness
  Approach: Desk research followed by field piloting for most promising household targeting methods
  Geography: Homa Bay, where thatch is uncommon and an overly-restrictive proxy means test, and so change is already needed" GiveDirectly, Update for GiveWell, September 2015, Pg. 9

• 153
  • Criteria being tested include being a widow, being an orphan, community based targeting, and subjective judgments about poverty level. GiveDirectly, Update for GiveWell, September 2015, Pg. 9
  • GiveDirectly, Uganda 2M campaign enrollment database

• 154

  Paul Niehaus and Carolina Toth, conversation with GiveWell, September 7, 2015

• 155
  • GiveDirectly, Update for GiveWell, July 2014, Pg 5.
  • This was deprioritized for 2015: "Urban pilot –limited interest expressed by inst. donors, govt, private donors in urban" GiveDirectly, Update for GiveWell, February 2015, pg. 20

• 156

  GiveDirectly, Update for GiveWell, July 2014, Pg 5.

• 157

  Cash transfers currently fund private goods. GiveDirectly is considering experimenting with a system whereby transfer recipients could propose public goods projects and individuals could pool their resources to fund projects they consider worthwhile. Conversation with GiveDirectly, September 5, 2014, Pg 5.

• 158

  GiveDirectly, Update for GiveWell, July 2014, Pg 5.

  GiveDirectly would like to conduct a separate study on the effects of a basic income guarantee. The annual cash transfers involved would be smaller than those in GiveDirectly’s standard program but granted consistently throughout the recipient’s life. GiveDirectly is interested in whether people would take greater risks if they know that their basic needs are met. This study could use the same control-group villages as the GE study. Conversation with GiveDirectly, September 5, 2014, Pg 5.

• 159
  • GiveDirectly has begun discussing a pilot project in Bukedea District as part of this focus on new operational challenges. GiveDirectly is considering managing cash withdrawals for recipients in the pilot rather than relying on an independent mobile money network, pending board approval. Conversation with GiveDirectly, October 6, 2014, Pg 3.
  • As of September 2015, GiveDirectly will continue to use its partners to manage cash withdrawals. Managing the withdrawals in-house would be expensive for GiveDirectly. GiveDirectly tested working with a new partner in Uganda this year (Centenary Bank) and decided that working with MTN was more efficient (more below), which may negate the need for cash out days. Paul Niehaus and Carolina Toth, conversation with GiveWell, September 7, 2015

• 160

  Some of the GiveDirectly's potential partnership projects (note that these projects are small compared to the projects GiveDirectly is currently interested in):

  • Uganda District: "Request from an MP to discuss working in his district" GiveDirectly, Check in with GiveWell, September 2014 Update: "GD declined the request to work in the MP's particular district - no obvious opportunity for policy impact to offset relocation costs." GiveDirectly, Update for GiveWell, September 2015, Pg. 6.
  • GiveDirectly has spoken with researchers in a multilateral-funded program to which the partners were considering adding unconditional cash transfers. Conversation with Piali Mukhopadhyay, GiveDirectly, October 20-21, 2014 (unpublished). Update: GiveDirectly declined to participate in this program despite high potential policy impact, because a) the program was located in an area GiveDirectly had never worked and GiveDirectly did not have enough time to run pilots of cash transfers before making its decision and b) GiveDirectly was worried that study results would not be rigorous due to low sample size for the cash transfer arm. Paul Niehaus and Carolina Toth, conversation with GiveWell, September 7, 2015
  • Multilateral agency: GiveDirectly has also spoken with another agency, which is considering using cash transfers as a benchmark to assess the impacts of one of its nutrition programs. Conversation with Piali Mukhopadhyay, GiveDirectly, October 20-21, 2014 (unpublished). Update: "GD declined to pursue implementation of nutrition benchmarking study, but will provide advice." GiveDirectly, Update for GiveWell, September 2015, Pg. 6.
  • Haiti evaluation: GiveDirectly is discussing experimental impact evaluation in Haiti to
    assess how cash performs relative to other reconstruction projects there. GiveDirectly, Update for GiveWell, September 2015, Pg. 6.
  • Evaluation non-profit benchmarking initiative: GiveDirectly spoke with a representative at an evaluation non-profit about adding cash transfer arms to some of the non-profit's studies. A survey of 31 ongoing studies found that 6 have a cash arm currently and 20 would like to add one. GiveDirectly does not plan to be the implementing partner for these studies because there are often implementing partners that are already set up who could implement the cash transfers if they had enough funding. But GiveDirectly is spreading awareness of this opportunity through its network. Paul Niehaus and Carolina Toth, conversation with GiveWell, September 7, 2015

• 161
  • Funding for the project came from an institutional funder and Google.org; each provided $2 million. Paul Niehaus and Carolina Toth, conversation with GiveWell, May 28, 2015
  • GiveDirectly, Rwanda technical application

• 162
  • The project partners are still in the process of designing the project; they expect to have plans for the RCT ready by mid-November 2015. Paul Niehaus and Carolina Toth, conversation with GiveWell, September 7, 2015
  • GiveDirectly, Rwanda technical application

• 163

  For example, GiveDirectly is in conversations with an institutional funder about testing other humanitarian aid interventions against cash transfers. Paul Niehaus and Carolina Toth, conversation with GiveWell, September 7, 2015

• 164

  Paul Niehaus, GiveDirectly's President, transitioned this year from working part time to full time. Ms. Mukhopadhyay said that it has been good to have Prof. Niehaus working full time, and that it has enabled them to spend a lot more time thinking about partnerships and in-country networking. Prof. Niehaus and Ms. Mukhopadhyay have also worked together to plan for GiveDirectly's potential expansion into Rwanda. Conversation with Piali Mukhopadhyay, GiveDirectly, October 20-21, 2014

• 165

  Paul Niehaus and Carolina Toth, conversation with GiveWell, September 7, 2015
  Carolina Toth, email to GiveWell, November 10, 2015

• 166
  • "In practice, we do not think there is much scope for additional senior leadership time devoted to fieldwork to materially improve the quality or efficiency of execution there." GiveDirectly, Update for GiveWell, September 2015, pg. 14.
  • Paul Niehaus and Carolina Toth, conversation with GiveWell, September 7, 2015

• 167

  For example, see GiveDirectly's intended use of Good Venture's $25 million grant here and here.

• 168
  • We are unable to be more specific because GiveDirectly would prefer to keep the particular inputs the algorithm uses confidential, to make it harder for people to game the eligibility requirements.
  • Carolina Toth, email to GiveWell, October 20, 2015

• 169

  Carolina Toth, conversation with GiveWell, November 12, 2015

• 170
  • "We typically use building materials as eligibility criterion—organic materials like a thatched roof, mud walls, or mud floor have the advantage of being (a) a strong predictor of poverty, (b) easy for community members to understand, and (c) relatively easy to audit in a number of ways, including both digital imagery captured by our field staff and satellite imagery captured remotely." GiveDirectly, FAQs 2015
  • GiveDirectly has tweaked these criteria in the past, e.g., "Dropped mud walls as eligibility requirement" GiveDirectly, Update for GiveWell, September 2014
  • Carolina Toth, email to GiveWell, November 10, 2015

• 171

  Conversation with Piali Mukhopadhyay, GiveDirectly, October 20-21, 2014

• 172

  Carolina Toth, conversation with GiveWell, November 12, 2015

• 173
  • "In Siaya, we also honed our process of targeting households that are most in need. In a new region, our old ways of targeting may not work. For example, in Homa Bay there is a scarcity of grass, so identifying the poorest families using thatched roofs may be less effective." GiveDirectly, Blog post, August 25, 2015
  • "Geography: Homa Bay, where thatch is uncommon and an overly-restrictive proxy means test, and so change is already needed" GiveDirectly, Update for GiveWell, September 2015, pg. 9.
  • "The "thatch roof, mud walls, mud floor" eligibility criteria was not going to work in Homa Bay, as <3% households had thatch roofs" Carolina Toth, email to GiveWell, October 20, 2015

• 174
  • [GiveDirectly] is also considering modifying its targeting criteria to include certain types of people who may be especially vulnerable whether or not they live under an iron roof.
    Conversation with GiveDirectly field staff, October 20-21, 2014
  • GiveDirectly is considering expanding its eligibility criteria to include:
    • Widows living in iron-roofed houses
    • Houses with iron roofs that are severely corroded
    • Households with partially cemented floors

    Conversation with GiveDirectly, October 6, 2014

  • "Methods being piloted
    • Community based targeting (multiple variants)
    • Subjective rankings (by staff, external parties)
    • Proxy means tests
      • PPI (Progress out of Poverty Index)
      • MPI (Multidimensional Poverty Index)
      • Other additional proxies (e.g., widows)"

    GiveDirectly, Update for GiveWell, September 2015, pg. 6.

  • "We've now completed our targeting pilots in Kenya, and have selected a new targeting criteria for Homa Bay." Carolina Toth, email to GiveWell, October 20, 2015

• 175

  Carolina Toth, email to GiveWell, October 20, 2015

• 176

  Paul Niehaus and Carolina Toth, conversation with GiveWell, September 7, 2015

• 177
  • Paul Niehaus and Carolina Toth, conversation with GiveWell, September 7, 2015
  • GiveDirectly gave an example of one inconsistent criteria model that it tested: some versions of community-based-targeting were inconsistent. GiveDirectly found that in some community-based targeting models, unique community groups would rate the same family's level of poverty quite differently; while one community group would claim the family was extremely poor, another would tell GiveDirectly that the family was comparatively wealthy. These discrepancies led to GiveDirectly discarding several community targeting models as eligibility models. Paul Niehaus and Carolina Toth, conversation with GiveWell, September 7, 2015

• 178

  Carolina Toth, conversation with GiveWell, November 12, 2015

• 179

  GiveDirectly, Eligibility criteria presentation

• 180

  "Major differences between this criteria in Kendu vs thatch criteria in Siaya
  Pro: harder for all parties to game, as algorithm is not known
  Pro: includes vulnerable groups important to fairness perceptions
  Con: more expensive to administer as there are more questions to answer. Cost savings from reduced gaming may offset this.
  Con: difficult to explain to communities why someone was or was not included, decreasing fairness perception
  Taking all of these: we think cost and fairness differences are probably going to be about the same, and so the major gain is the fact that it is difficult to game" Carolina Toth, email to GiveWell, October 20, 2015

• 181
  • @GiveDirectly, Survey for Randomized Controlled Trial@
  • GiveDirectly, Offering Memorandum (January 2012) Pgs 23-24.

• 182

  GiveDirectly, Offering Memorandum (January 2012) Pg 25.
  

• 183

  Haushofer and Shapiro 2013 Policy Brief

• 184
  • The data reported for this variable in Table 4 of Haushofer and Shapiro 2013 Policy Brief, Pg 32, appears to demonstrate that 21% (not 20%) of the control group reports that not all household members usually eat until they are content.
  • Haushofer and Shapiro 2013 Policy Brief, Pg 18, corresponding data in Table 4, Pg 32.

• 185
  • "Some (but not all) mabati [iron-roofed] or even permanent HH are as deserving as thatched HH […] 6/6 groups mentioned deserving special cases." GiveDirectly, Saturation analysis, Pg 4.
  • Mr. Ekeu thinks that roofs are too rough a way to target poverty because some people may live under an iron roof but actually be very poor (e.g., someone who inherited an iron-roofed house from his grandfather, or a widow whose late husband built her an iron-roofed house long ago). Ms. Mukhopadhyay said that GiveDirectly hears this kind of feedback from a lot of its field staff, but believes that building materials are still a good criteria on average.
    Conversation with GiveDirectly field staff, October 20-21, 2014

• 186
  • "Majority of households are widows, almost half have one family member that is disabled or very sick.
  • Households seem to be as or more needy than typical GD recipients"

  GiveDirectly, Update for GiveWell, July 2014 Pg 11

• 187

  Mr. Ekeu [the Senior Field Officer in Uganda] thinks that roofs are too rough a way to target poverty because some people may live under an iron roof but actually be very poor (e.g., someone who inherited an iron-roofed house from his grandfather, or a widow whose late husband built her an iron-roofed house long ago). Ms. Mukhopadhyay said that GiveDirectly hears this kind of feedback from a lot of its field staff, but believes that building materials are still a good criteria on average. It is also considering modifying its targeting criteria to include certain types of people who may be especially vulnerable whether or not they live under an iron roof. Conversation with GiveDirectly field staff, October 20-21, 2014, Pg 4.

• 188

  Paul Niehaus, GiveDirectly Founder, email to GiveWell, November 20, 2012.

• 189
  • GiveDirectly, Kenya hotline log, July 2013
  • GiveDirectly, Follow-up tracker, July 2013
  • GiveDirectly, Follow-up tracker, October 2014

• 190
  • One adverse event reported in the Kenya follow-up tracker involved a case of domestic abuse resulting in the death of the mother and child, where the particular instance of conflict may have been related to the use of transfer funds. When GiveDirectly investigated the event, the parents of the deceased expressed that there was not anything GiveDirectly could have done to prevent this adverse event. People in the community, including the parents of the deceased, also said that they did not want to report the incident because they were afraid that it would cause GiveDirectly to cancel the cash transfer program, which they did not want to happen. Conversation with GiveDirectly, July 7, 2014 Pg 5
  • "During this process there were some reports of problems during paydays - recipients were hesitant to come forward initially"
    Email from Carolina Toth, GiveDirectly, September 12, 2014

• 191

  We asked the field officers what they think about the current transfer size ($1000), and whether they’d choose to keep it at that level, increase it, or decrease it, given the effects that an adjustment would have on how many people GiveDirectly would be able to serve.

  • Mr. Okello: Typically there are multiple households on one compound, each inhabited by relatives of the same family, and any household that meets the targeting criteria can receive transfers. Mr. Okello said that it may make more sense for GiveDirectly to group some households on a compound together so that transfers are shared across them, rather than each eligible household receiving the full $1000.

    Mr. Okello also said that if GiveDirectly increased the size of the transfers, that could create a high level of dependency. One of the messages that field officers send is that people should use the $1000 transfers to develop themselves as much as possible, but if someone knew they were getting $2000, they may stop farming, for example. With $1000 people can get some things but not everything; it is the right amount.

  • Mr. Ekeu: Mr. Ekeu prefers reducing the amount of money in each transfer and expanding the recipient base to reach everyone in the village. He said that the current targeting model causes bragging and unrest in the communities. The people who don’t benefit may be brought to use force to get some of the money, such as by breaking into recipients’ homes. Mr. Ekeu suggested that it would be better for GiveDirectly to provide all households in a village with some amount of money, even if it was less for households that are currently deemed ineligible (e.g., $100). This way, each of the households would be busy figuring out how they would spend their own money rather than how to get money from another.
  • Mr. Olinga: Mr. Olinga said that to reduce extreme poverty the bigger transfer is better, but he didn't have a strong opinion on $1000 transfers to some people versus $500 transfers to twice as many. [This is how we posed the question to Mr. Olinga.]

  Conversation with GiveDirectly field staff, October 20-21, 2014 Pg 5

• 192

  Several recipient households had been selected by GiveDirectly for our visit as representative of how recipients use funds. 2 locations (and the households within them) were selected as a function of GiveDirectly's activities that day -- an area of Rarieda where the end-line survey for the RCT was being implemented and an area of Siaya where enrollment was being undertaken. In both cases, we don't know whether enrollment and surveying activities were taking place elsewhere that day which would have given GiveDirectly discretion in choosing these areas.

  On the final day of our visit, we asked GiveDirectly whether we could randomly select 3 households to visit. GiveDirectly sent us a list of 15 households in a location in Rarieda where the end-line survey for the RCT was complete and therefore we could question recipients without interfering with the RCT. We don't know whether GiveDirectly had discretion in choosing these 15 households. We selected 5 households from the list using Excel's RAND() function and visited 3 of them. (GiveDirectly made appointments with the households in advance and could not reach 2 of them.)

  We would characterize all the households we visited -- those that GiveDirectly fully selected for us, those over which GiveDirectly had less discretion, and those we selected randomly -- as extremely poor. We did not see any significant differences in wealth between them.

• 193

  @GiveDirectly, Siaya Enrollment Database@
  GiveDirectly, Google enrollment database
  GiveDirectly, Nike enrollment database
  GiveDirectly, RCT Enrollment Database
  GiveDirectly, Kenya 2M enrollment database, September 2013
  GiveDirectly, Uganda pilot enrollment database - Kosile
  GiveDirectly, Uganda pilot enrollment database - Kawo
  GiveDirectly, Uganda pilot enrollment database - Kanyamutamu
  GiveDirectly, Uganda pilot enrollment database - Akumure
  GiveDirectly, Kenya 1.2M enrollment database
  GiveDirectly, Kenya rolling campaign enrollment database - Homa Bay
  GiveDirectly, Kenya rolling campaign enrollment database - Siaya
  GiveDirectly, Uganda 2M campaign enrollment database
  

• 194

  GiveDirectly's current procedure for identifying households to audit:

  • GiveDirectly collects information about recipients during the first three stages of a campaign: census, registration, and backcheck. Some information, such as recipient name, GPS location, housing materials, and identifying photograph, is collected at more than one stage and then checked for mismatches. (These checks are currently conducted using Excel but will eventually be automated through Segovia technology. One exception is comparing identifying photographs of recipients, which is done using Mechanical Turk.)
  • Each mismatch in recipient information is assigned a certain weight depending on how likely it is to be an indication of gaming. (GiveDirectly said that it determined the likelihoods of various mismatches indicating gaming by conducting an analysis of the mismatches present in past cases of gaming. We did not review this analysis.) GiveDirectly said that the mismatches with the highest weights are mismatches in identifying photographs and housing materials.
  • Each recipient is assigned a total mismatch "score" (the composite of all their weighted mismatches). Recipients with scores above a certain level are audited.

  Conversation with GiveDirectly field staff, October 20-21, 2014

• 195
  • See GiveDirectly, Eligibility check, which we downloaded from GiveDirectly's website on September 14, 2015. GiveDirectly, What We Do - Operating Model, Overview tab Note that the document does not include all households that GiveDirectly has ever enrolled (e.g. there are almost no houses enrolled from 2012).
  • GiveDirectly used to track these metrics in its monthly operations reports. For example, see "Gaming detection" in GiveDirectly, Monthly operations report, October 2014

• 196

  Piali Mukhopadhyay, COO, International, GiveDirectly, conversation with GiveWell, November 8, 2012 and GiveWell visit to M-PESA agent, November 8, 2012.

• 197

  GiveWell visit to M-PESA agent, November 8, 2012.

• 198

  "We then transfer money directly to the recipient's account […] The recipient collects the transfer from a local M-Pesa agent, who is typically a shopkeeper in the village or the nearest town. The recipient transfers his or her electronic M-Pesa balance to the agent using his or her mobile phone in return for cash." GiveDirectly, How it works 2014

• 199

  Conversation with GiveDirectly, April 8, 2014, Pg 4.
  Carolina Toth, email to GiveWell, November 10, 2015

• 200

  Paul Niehaus and Carolina Toth, conversation with GiveWell, September 7, 2015

• 201

  Conversation with Piali Mukhopadhyay, COO, International, GiveDirectly, October 22, 2013

• 202

  "Selected MTN as preferred provider in Uganda after assessing performance of Ezee/MTN (building relationship with Airtel so as to have an additional hedge)"
  GiveDirectly, Update for GiveWell, July 2014 Pg 9.

• 203

  The network of EZEE Money agents is very limited, so the only feasible option for recipients was to withdraw funds on paydays. MTN has a larger network of agents, so while it is still more convenient for recipients to withdraw on paydays, recipients can seek other options if they prefer. Conversation with GiveDirectly, September 5, 2014

• 204

  Observations from GiveWell site visit to GiveDirectly, October 2014, Pg 5.

• 205

  "Biometrics

  • Palm-vein readers purchased and working well
  • Scans have been taken twice: once for initial registration, once for token payment
  • Helpful compliment to identification measures taken by banking partner at paydays
  • Skills built on how to explain to recipients, how to incorporate into process
  • GD likely to use in contexts where national IDs are not present, paydays are necessary
  • May not use in Uganda in favor of distributed cash out
  • Potential to incorporate biometrics into enrollment process, but not high priority"

  GiveDirectly, Update for GiveWell, September 2015, pg. 10.
  

• 206

  "Bank as payments provider

  • Partnered with Centenary Bank, which offered lower transaction fees than MTN
  • Difficulty scheduling cash delivery logistics (e.g., reserving vans) has pushed back first lump sum payment by a month
  • Significant amount of FD time spent managing weak counterparts at bank"

  GiveDirectly, Update for GiveWell, September 2015, pg. 10.

• 207

  "Pilot of distributed cashout with MTN in lieu of paydays underway as well" GiveDirectly, Update for GiveWell, September 2015, pg. 10

• 208
  • Paul Niehaus and Carolina Toth, conversation with GiveWell, September 7, 2015
  • Note that GiveDirectly has not previously tested the distributed cashout model in Uganda. It originally developed and implemented the cashout day model for Uganda because it thought that some recipients would be too far away from MTN mobile money agents to (e.g., some recipients would have to make a three hour round trip to withdraw their funds). Now GiveDirectly thinks that with targeted communication about where the nearest MTN agents are, GiveDirectly might be able to spare recipients these longer trips under the distributed cashout model. Paul Niehaus and Carolina Toth, conversation with GiveWell, September 7, 2015

• 209

  GiveDirectly 's blog post on the staff fraud: GiveDirectly blog, Fighting fraud in Uganda
  "Two GD staff members stole 2% of transfers on Google campaign ($20,500 in total) in form of 50,000 UGX deductions from penultimate or final transfers (85%) or theft of entire final transfers (15%)"
  GiveDirectly, Check in with GiveWell, September 2014

• 210

  "Discovery of the fraud

  • A GiveDirectly recipient had given their SIM card to the SFO [Senior Field Officer] (whose contract had been terminated in April due to an unrelated issue involving a fraudulent receipt he brought to GiveDirectly for reimbursement). The recipient asked the SFO to replace his SIM card (recipients have to travel about 4 hours round trip to get this done) and the SFO had not returned it. This report was made to the hotline that the OM [Office Manager] was answering three months after the recipient had given over their SIM card. Mr. Skeates audits the logs of these hotline calls.
  • GiveDirectly had been planning a full round of follow-up surveys as part of its normal process, but accelerated the follow-up surveys in response to this issue. GiveDirectly’s backcheck team paused their work on enrollment for the Uganda 2M campaign and called all the recipients in that village (Kosile) to ask whether they had received all of their transfers, had any problems withdrawing, and whether GiveDirectly currently had any of their documents (e.g., SIM cards, IDs).
  • During this process, there were some reports of problems during paydays. Recipients were initially hesitant to come forward.
  • Because of the reports of payday problems, GiveDirectly began calling another village, Kawo, the following day to gather more information. Recipients in Kawo were far more forthcoming with information when asked specific questions about payday problems.
  • GiveDirectly continued follow-ups (conducted by a new Field Officer brought on after the SFO's and OM’s dismissals) until it had spoken to about 92% of its recipients across all villages. GiveDirectly also conducted in-person visits to villages."

  Conversation with GiveDirectly, September 5, 2014

• 211

  Changes implemented in response to staff fraud
  In the Uganda pilot campaign, GiveDirectly cash out days were managed by the Uganda Senior Field Officer, the Uganda Office Manager who also managed the GiveDirectly hotline, and mobile money agents. After these people fraudulently diverted funds from recipients, GiveDirectly implemented a series of changes:

  • Terminated the GiveDirectly staff who had been involved in the fraud; started working with new mobile money agents.
  • Removed all of its staff from the cash out day process except the Uganda Field Director. The Uganda Field Director had previously been making planned visits to oversee some of the cash out days; he now actively manages all of them along with new mobile money agents.
  • Appointed community-nominated monitors to assist the Uganda Field Director on the cash out day with translation, observe transactions between recipients and mobile money agents, and report any issues they see. GiveDirectly compensates the monitors with 10,000 UGX (~$4) for their time during a cash out day.
  • Developed networks of English-speaking informants who are not formally announced within the villages, but are tasked with also reporting any issues they see regarding transfers. To date, 4 of the 9 informants have provided GiveDirectly with helpful information, such as identifying that households in the enrollment process were actually ineligible, and telling GiveDirectly that someone had taken a recipient's phone after the recipient passed away.
  • Moved the GiveDirectly call center (hotline) to Kampala, to increase the separation of call center staff from field staff, who are based in Mbale.
  • Tasked the call center with calling a randomly selected 10% of the village during a cash out day to see if it is going smoothly.
  • Changed the contractual agreement GiveDirectly has with mobile money agents to include an indemnity clause, so that in the case of stolen funds, GiveDirectly could remove funds directly from a mobile money agent's account.

  Conversation with Stuart Skeates, GiveDirectly, October 20-21, 2014

• 212

  Paul Niehaus and Carolina Toth, conversation with GiveWell, September 7, 2015

• 213
  • GiveDirectly has piloted a few changes that would increase security, including the use of biometrics (more) and partnering with a banking partner for cash out days (more).
  • GiveDirectly is piloting a distributed cashout model in Uganda and it is possible that a distributed cashout model is more secure from large-scale crime than a payday model because a) without the cashout days, funds are not as concentrated in one location and b) it's easier to obscure who is a recipient when recipients withdraw their funds at different times from different places. Paul Niehaus and Carolina Toth, conversation with GiveWell, September 7, 2015. GiveDirectly has not yet implemented such a model in Uganda.

• 214

  Lydia Tala, GiveDirectly Field Assistant, conversation with GiveWell, November 7, 2012.

• 215

  Lydia Tala, GiveDirectly Field Assistant, conversation with GiveWell, November 7, 2012.

• 216

  GiveDirectly helps recipients with issues using M-PESA via our hotline. Staff sometimes contact M-PESA agents to notify them one of our clients will be visiting and may need assistance. Piali Mukhopadhyay, COO, International, GiveDirectly, email to GiveWell, November 23, 2012.

• 217

  "There were four types of issues that were responsible for most of the people who were unable to withdraw funds at the cash out day: […] MTN (the payment provider) had not yet activated the funds in the person's account. Mr. Skeates said that this was not a common problem in the previous campaign in Uganda, but it affected many people at this cash out day. He estimated that people affected by this issue would receive their first installment of funds in another 2-3 weeks. […]" GiveWell site visit to GiveDirectly, October 2014, Pg 6.

• 218
  • A challenge of backchecks is that field officers often end up teaching recipients how to use their cell phones and mobile money accounts, so that they can access their money and are less likely to be scammed. Field officers will teach recipients how to check their balance and distinguish messages that say they received money from other messages. Field officers will sometimes write out instructions for recipients in the local language that describe step-by-step how to operate the phones and mobile money accounts. Recipients often do not understand the importance of keeping their PIN numbers secure. Some elderly recipients do not want a trustee to manage their transfers, but they are unable to remember their PIN numbers or read the messages on their phone, so they are more likely to have issues receiving transfers. Some people are still skeptical that the money will actually come, even after they have received messages on their phone, so they don’t pay attention to the instructions about how to use the mobile money account. Conversation with GiveDirectly field staff, October 20-21, 2014, Pg 3.
  • Mr. Skeates, the Uganda Field Director, made announcements at the start of the event, translated by the 2 community monitors. The announcements included reminders about how to keep account information secure (e.g., after entering your PIN number, make sure to press "send" before handing your phone back to the agent; make sure you have received a confirmation message after withdrawing and that it states the correct amount; count the cash immediately after receiving it.) GiveWell site visit to GiveDirectly, October 2014 Pg 5.

• 219

  Paul Niehaus and Carolina Toth, conversation with GiveWell, September 7, 2015

• 220
  • A "trustee" is someone who is registered for the mobile money payments on behalf of the recipient. A "helper" is someone who is not registered for the payments, but who helps the recipient with the process (e.g. assisting with transportation to the locations one can withdraw cash or helping to use the phone properly).
  • Roughly, the process for choosing a trustee or helper is to get the recipient alone (out of earshot of family) and ask who that recipient trusts the most. This choice is typically validated with some neighbors (ensuring that that person is regarded as trustworthy). Generally, GiveDirectly prefers to choose trustees and helpers who are already recipients themselves, so that they have less of an incentive to steal the transfer and so that GiveDirectly can stop transfers to them if they are not performing their role appropriately.

  Paul Niehaus and Carolina Toth, conversation with GiveWell, September 7, 2015 Note that GiveDirectly has offered to send us the protocol used to determine helpers and trustees; we have not yet reviewed this protocol.
  

• 221

  @GiveDirectly, Siaya Verification Stats@
  GiveDirectly, Rarieda verification stats
  GiveDirectly, Siaya follow-up data - disaggregated
  GiveDirectly, Nike follow-up data - disaggregated
  GiveDirectly, Google follow-up data - disaggregated (LS - long)
  @GiveDirectly, Rarieda Top-up verification (short)@
  GiveDirectly, Rarieda verification (top ups), May 26, 2013
  GiveDirectly, Siaya verification, June 15, 2013
  GiveDirectly, Google verification, September 2013
  GiveDirectly, Nike verification (final), September 2013
  GiveDirectly, Kenya follow up data, November 2014
  GiveDirectly, Uganda pilot follow up data, April 2014
  GiveWell, GiveDirectly follow up surveys summary - Uganda, September 2015
  GiveWell, GiveDirectly follow up surveys summary - Kenya, September 2015

• 222

  This is based on data GiveWell reviewed in 2014. GiveDirectly, Kenya follow up data, November 2014

• 223

  72% = 348/(137+348). "Large vs. small transfers Finally, a third treatment arm was created to study the relative impact of large compared to small transfers. To this end, 137 households in the treatment group were randomly chosen and informed in January 2012 that they would receive an additional transfer of KES 70,000 (USD 798, PPP 1,112), paid in seven monthly installments of KES 10,000 (USD 114, PPP 160) each, beginning in February 2012. Thus, the transfers previously assigned to these households, whether monthly or lump-sum, were augmented by KES 10,000 from February 2012 to August 20128, and therefore the total transfer amount received by these households was KES 95,200 (USD 1,085, PPP 1,525). The remaining 348 treatment households constitute the 'small' transfer group, and received transfers totaling KES 25,200 (USD 287, PPP 404) per household. These three treatment arms were fully cross-randomized, except that, as noted above, the 'large' transfers were made to existing recipients of KES 25,200 transfers in the form of a KES 70,000 top-up that was delivered as a stream of payments after respondents had already been told that they would receive KES 25,200 transfers. Section 3.1 outlines how this issue is dealt with in the analysis." Haushofer and Shapiro 2013 Policy Brief, Pgs. 7-8

• 224
  • The PPP adjusted values of the small, mean, and large transfer are $404, $721, and $1,525 respectively. “28% of the treatment group received a transfer of KES 95,200 (USD 1,085, PPP 1,525), while the remaining 72% received KES 25,200 (USD 287, PPP 404); the average transfer was thus KES 45,016 (USD 513, PPP 721).” Haushofer and Shapiro 2013 Policy Brief, pg. 12.
  • "All USD values are calculated at purchasing power parity, using is the 2012 World Bank PPP estimate for private consumption in Kenya: 0.016." Haushofer and Shapiro 2013, Pg. 2

• 225

  Table 28, Haushofer and Shapiro 2013 Appendix, Pg. 54

• 226
  • Non-land assets includes “the value, in USD PPP, of all moveable assets the household owns, including savings, plus the value of the roof of the home.” Haushofer and Shapiro 2013 Policy Brief, pg. 10.
  • Haushofer and Shapiro 2013 Appendix, Pgs. 1-2

• 227

  Small transfers: $210 (95% CI: $158 to $263). Table 28, Haushofer and Shapiro 2013 Appendix, Pg. 54

• 228
  • Table 32, Haushofer and Shapiro 2013 Appendix, Pg. 58
  • Small transfers, livestock: $68 (95% CI: $35 to $100)
  • Small transfers, durable goods: $36 (95% CI: $18 to $54)
  • Small transfers, savings: $7 (95% CI: $2 to $13)

• 229

  Table 5, Haushofer and Shapiro 2013, Pg. 53

• 230

  Table 5, Haushofer and Shapiro 2013, Pg. 53

• 231

  "To this end, we conducted a separate survey of one respondent from each of 20 villages to obtain estimates for the costs of purchasing and maintaining metal and thatch roofs. The purchase of a metal roof represents an expenditure of on average USD 564, or 75 percent of the average transfer value." Haushofer and Shapiro 2013, Pg. 34

• 232

  "Based on the anonymized individual-level survey data, an iron roof costs $418 on average, thatch roof replacement (including the cost of grass for making the roof and the labor) costs $95 on average, and thatch roof repair (including the cost of grass for making the roof and the labor) costs $107 on average. These numbers appear to conflict with the full paper and the policy brief. It may be that the results were from a different survey. Haushofer and Shapiro have not yet finished verifying which data were used." GiveWell's non-verbatim summary of a conversation with Carolina Toth, GiveDirectly, October 1, 2014

• 233

  Table 44, Haushofer and Shapiro 2013 Appendix, Pg. 70

• 234

  Table 44, Haushofer and Shapiro 2013 Appendix, Pg. 70

• 235

  Table 36, Haushofer and Shapiro 2013 Appendix, Pg. 62

• 236

  Table 36, Haushofer and Shapiro 2013 Appendix, Pg. 62

• 237

  Table 36, Haushofer and Shapiro 2013 Appendix, Pg. 62

• 238

  Table 52, Haushofer and Shapiro 2013 Appendix, Pg. 78

• 239
  • N = 1372. Table 36, Haushofer and Shapiro 2013 Appendix, Pg. 62
  • N = 1174. Table 52, Haushofer and Shapiro 2013 Appendix, Pg. 78

• 240

  Small transfers: $31 (95% CI: $18 to $43). Table 28, Haushofer and Shapiro 2013 Appendix, Pg. 54

• 241
  • Large transfers: $25 (95% CI: $11 to $39). $25/$51 = about 50%
  • Small transfers: $18 (95% CI: $9 to $27). $18/$31 = about 60%
  • Table 36, Haushofer and Shapiro 2013 Appendix, Pg. 62

• 242
  • Large transfers, social: $3 (95% CI: $1 to $5)
  • Small transfers, social: $2 (95% CI: $1 to $3)
  • Large transfers, other: $19 (95% CI: $13 to $24)
  • Small transfers, other: $7 (95% CI: $3 to $11)
  • Table 36, Haushofer and Shapiro 2013 Appendix, Pg. 62

• 243
  • Large transfers, alcohol: -$2.07 (95% CI: -$4.6 to $0.5)
  • Small transfers, alcohol: -$0.51 (95% CI: -$2.7 to $1.7)
  • Large transfers, tobacco: -$0.38 (95% CI: -$1.0 to $0.2)
  • Small transfers, tobacco: -$0.08 (95% CI: -$0.6 to $0.4)
  • Table 36, Haushofer and Shapiro 2013 Appendix, Pg. 62

• 244

  “Food security is low in this population. Though instance of skipped meals are not extreme, 20% of the control group reports that not all household members usually eat until they are content, 23% of respondents report sleeping hungry in the last week, and only 36% report having enough food in the house for the next day.” Haushofer and Shapiro 2013 Policy Brief, pg. 18.

• 245

  Small transfers: 0.21 (95% CI: 0.07 to 0.35). Table 28, Haushofer and Shapiro 2013, Pg. 54

• 246
  • Large transfers, health index: -0.09 (95% CI: -0.27 to 0.09)
  • Small transfers, health index: -0.02 (95% CI: -0.16 to 0.12)
  • Large transfers, education index: 0.11 (95% CI: -0.05 to 0.27)
  • Small transfers, education index: 0.07 (95% CI: -0.07 to 0.21)

• 247
  • Large transfers, profits: -$1 (-$10 to $8)
  • Small transfers, profits: $0.26 (-$8 to $9)
  • Table 44, Haushofer and Shapiro 2013 Appendix, Pg. 70

• 248

  Small transfers: 0.11 (95% CI: -0.01 to .23). Table 28, Haushofer and Shapiro 2013, Pg. 54

• 249
  • Table 56, Haushofer and Shapiro 2013, Pg. 82
  • Table 7, Haushofer and Shapiro 2013, Pg. 55

• 250
  • "These are generally small and not significant, with one exception: we observe an increase of 0.23 SD in the female empowerment index among the control group in treatment villages. This increase is significant at the 5 percent level using conventional p-values. Together with a non-significant direct treatment effect of SD −0.01 on this measure, this spillover effect suggests that the treatment group shows a significant increase in female empowerment relative to the pure control group, which we confirm in the Online Appendix. However, since this is the only outcome that shows any spillover effect and we do not have a good theory for why spillover effects might occur in female empowerment, we do not offer an interpretation of this result at this stage, and instead note that it needs to be replicated." Haushofer and Shapiro 2013, Pg. 23
  • Table 1, Haushofer and Shapiro 2013, Pg. 49

• 251

  “As changes in domestic violence were hypothesized to arise through mechanisms directly associated with cash transfers (such as a change in women’s bargaining power, or a reduction in domestic tension over economic hardships), these spillover effects are somewhat surprising. One possible explanation is that the results are simply an artifact of reporting bias, where the spillover sample believed that a different answer was desired from them than the control group. However, given that we do not find spillover effects in other measures that target unobservable outcomes, we find this explanation implausible. Another possibility is that the presence of the cash transfer program in the village motivated the husbands in untreated households to change their behavior in the hope of receiving transfers in the future. For instance, knowing that the primary female in the household was equally likely to receive the transfer as the primary male, men may have shifted their behavior to establish better relationships with their spouse. Alternatively, the spillover effect may operate through changes in attitudes among either or both husbands and wives in non-treated households. Our data does not distinguish between these possibilities; we find these unexpected results intriguing and believe they warrant further investigation.” Haushofer and Shapiro 2013 Policy Brief, pg. 21.

• 252
  • GiveDirectly, Kenya follow up data, November 2014
  • GiveDirectly, Uganda pilot follow up data, April 2014
  • GiveDirectly, Rarieda verification (top ups), May 26, 2013
  • GiveDirectly, Siaya verification, June 15, 2013
  • GiveDirectly, Google verification, September 2013
  • GiveDirectly, Nike verification (final), September 2013
  • GiveWell, GiveDirectly follow up surveys summary - Uganda, September 2015
  • GiveWell, GiveDirectly follow up surveys summary - Kenya, September 2015

  Note that in some cases, we have cleaned this data (where it was obvious what the error was in the data entry). These cases have been marked with cell comments. In other cases, when it was not clear what the error was (for example, when an amount of funds was mistakenly placed in a spending category column, and it was not clear whether the amount was intended to be placed in the former or latter column), we deleted the data and commented on the cell to indicate that data had been deleted. Deletions were made no more than 5 times in any given spreadsheet.

• 253
  • The tables include follow up survey data from the Kenya 2M, Kenya 1.2M, Kenya rolling enrollment, and Kenya behavioral optimization campaigns (GiveWell, GiveDirectly follow up surveys summary - Kenya, September 2015) and from the Uganda pilot campaign (GiveWell, GiveDirectly follow up surveys summary - Uganda, September 2015). Note that recipients may have been surveyed more than once and would therefore be included more than once in the data presented.
  • We still have not seen spending data for the Ug-201404 or Ug-201503 campaigns (the Uganda 2M and Uganda model variations campaigns, respectively). The Uganda spending data in the tables below is the same as the data we presented in 2014.

• 254In follow up surveys administered in Uganda, recipients were asked about spending on large household items and small household items. The figure reported here is the combination of those two categories.
• 255In follow up surveys administered in Uganda, recipients were asked about spending on large household items and small household items. The figure reported here is the combination of those two categories.
• 256

  For full details of our interviews with recipients, see GiveWell Site visit notes.

• 257

  Provided by Google via Citibank N.A. on November 15, 2012.

• 258

  Lydia Tala, GiveDirectly Field Assistant, conversation with GiveWell, November 7, 2012.

• 259

  Provided by Google via Citibank N.A. on November 15, 2012.

• 260

  "The 'thatch roof, mud walls, mud floor' eligibility criteria was not going to work in Homa Bay, as 3% households had thatch roofs." Carolina Toth, email to GiveWell, October 20, 2015

• 261
  • Haushofer and Shapiro 2013 Policy Brief, Table 10, Pg 38.
  • "Additional variables in table 9 show the frequency of any episode of physical, sexual or emotional violence in the last six months, and the percentage of respondents who believe that domestic violence is justified in some instances. The point estimates for these variables suggest a reduction in domestic violence, although none are individually different from zero at conventional significance levels." Haushofer and Shapiro 2013 Policy Brief, Pg 21.

• 262
  • GiveDirectly sent 8 versions of the survey forms it used, as the form was modified over time. For the most recent version we have seen, see GiveDirectly, Kenya verification template, August 2013
  • For earlier versions, see @GiveDirectly, Verification Template (October 1, 2012)@ and @GiveDirectly, Verification Template (November 7, 2011)@ (corresponding data for the latter: GiveDirectly, Verification data (November 17, 2011))

• 263

  GiveWell, GiveDirectly follow up surveys summary - Kenya, September 2015
  GiveWell, GiveDirectly follow up surveys summary - Uganda, September 2015
  For a list of follow-up survey questions from 2013, see GiveDirectly, Kenya verification template, August 2013. Note that this template may now be out of date.

• 264

  Note that not every follow-up survey asks about precisely the same issues (e.g. some do not ask about domestic violence or trouble collecting), which is part of what accounts for the different sizes of each group of respondents.

• 265In the Uganda follow up data, this issue is denoted "stole_item"
• 266In the Uganda follow up data, we identified 4 issues that we believe all asked about bribes ("bribe," "pay to collect," "others_bribes," "agent_bribe"). We only include the "bribe" issue in this table for Uganda (which has the highest rate of complaints out of the four).
• 267
  • GiveDirectly, Kenya hotline log, July 2013 (May 2012 – July 2013)
  • GiveDirectly, Follow-up tracker, July 2013 (July – October 2013)
  • GiveDirectly, Follow-up tracker, October 2014 (January 2013-September 2014). This file also contains issues reported through other channels, such as to field staff during field visits and during follow up surveys.
  • GiveDirectly, Kenya top 10 adverse events 2015 (Note: we are unsure if all of the issues reported in this file were reported through the hotline)
  • GiveDirectly, Uganda top 10 adverse events 2015 (Note: we are unsure if all of the issues reported in this file were reported through the hotline)
  • GiveDirectly, Kenya randomized sample of adverse events, 2014-2015
  • GiveDirectly, Uganda randomized sample of adverse events, 2014-2015
  • We have not checked to see whether the issues reported in these files are overlapping, though note that the time ranges are overlapping.

• 268

  GiveDirectly, Follow-up tracker, October 2014 Sheet: "Summary" In 2015 we did not ask GiveDirectly to send us a follow-up tracker because it is our understanding that sharing the follow-up tracker databases take a significant amount of effort on GiveDirectly's part (much of the effort goes into anonymizing the entries). Instead, we requested a random sample of adverse events and a sample of some of the most serious adverse events from the previous year. We have reviewed these samples, but have not made them public (they are not anonymized); the issues they cover are broadly similar to the types of issues we have seen in previous years.

• 269

  GiveDirectly, Follow-up tracker, October 2014 Sheets: Summary; GiveWell notes

• 270

  Haushofer and Shapiro 2013 Policy Brief, Table 10, Pg 38.

• 271
  • The RCT of GiveDirectly’s program in Rarieda did not find an increase in crime, so at that scale it does not seem to be an issue. It’s possible that crime would be a more serious problem if GiveDirectly became a substantially larger and better-known organization. Conversation with GiveDirectly, December 7, 2013
  • GiveDirectly has become very well-known in Siaya County, Kenya, but has not seen a significant increase in crime rates there. As GiveDirectly begins to work in Homa Bay County, it expects crime rates to be lower, because the context is similar to Siaya but fewer people know about GiveDirectly in Homa Bay. If GiveDirectly were to start working in more urban areas, where crime rates tend to be higher, GiveDirectly would put more time into strategizing about crime and security. Paul Niehaus and Carolina Toth, conversation with GiveWell, September 7, 2015

• 272
  Example: "The recipient was given all the cash withdrawn as she requested.. then as she {deidentified} she was just outside the house since she's blind and her door is not lockable, she came to find her money missing but she doesn't know who might have stolen the KES."
  GiveDirectly, Follow-up tracker, October 2014 Sheet: "Tracker" (text removed in deidentification)

• 273
  Example: "He was phoned by unknown person who posed as GD staff and requested for 500/= bribe to hasten the processing of his transfer."
  GiveDirectly, Follow-up tracker, October 2014 Sheet: "Tracker" (text removed in deidentification)

• 274
  Example: "She lost the phone, and in the process of renewing the line the Agent transfer the money to another line in order to withdraw later."
  GiveDirectly, Follow-up tracker, October 2014 Sheet: "Tracker" (text removed in deidentification)
  
• 275

  Some recipients, especially elderly ones, have to learn how to use cell phones for the first time in order to manage the GiveDirectly transfers in mobile money accounts. These people have a more difficult time understanding how to keep their phones secure; for example, they often keep the phone in its original packaging and do not conceal it. Another problem with security is that some recipients will share the PIN numbers for their mobile money accounts, either intentionally or unintentionally by handing the phone to a mobile money agent before pressing "Send" (so the PIN number is still apparent on the screen of the phone.) This makes recipients more vulnerable to people who wanted to steal money from their accounts. Teaching PIN safety has long been a priority, and GiveDirectly has added additional emphasis on the topic (e.g., emphasis during village meetings, additional trainings given by the mobile provider) Improved security is a reason why GiveDirectly is interested in piloting biometric authentication for mobile money accounts, though it does not currently have plans to do so. Conversation with Stuart Skeates, GiveDirectly, October 20-21, 2014, Pgs 2-3.

• 276
  • Conversation with Stuart Skeates, GiveDirectly, October 20-21, 2014 (unpublished)
  • GiveDirectly, Follow-up tracker, October 2014;GiveDirectly, Kenya hotline log, July 2013 (May 2012 – July 2013); GiveDirectly, Follow-up tracker, July 2013 (July – October 2013)

• 277
  • "GiveDirectly stops issuing transfers after two years and clearly informs recipients that they should expect this." GiveDirectly, Offering Memorandum (January 2012) Pg 33.
  • Update [November 2014]: the two year timeline mentioned above is now outdated; transfers are generally sent over a period of 8 months in Kenya and 10 months in Uganda. It is still the case that recipients can only receive one full transfer, and that they are ineligible for additional transfers thereafter. Conversation with Carolina Toth, GiveDirectly, November 20, 2014
  • "Attrition was not a significant concern in this study because it became evident early on in GD’s work in Kenya that respondents were highly interested in maintaining relations with GiveDirectly in the hope of receiving future transfers (although these are never promised)." Johannes Haushofer and Jeremy Shapiro, Welfare Effects of Unconditional Cash Transfers: Pre-Analysis Plan, June 27, 2013 Pgs 5-6.

• 278

  E.g., GiveDirectly, Enrollment speed of distributions - Siaya and Rarieda

• 279
  • GiveDirectly, Enrollment speed of distributions - Siaya and Rarieda
  • GiveDirectly commented: "We were able to accelerate [the time it took for recipients to register for M-PESA] significantly for two reasons: (a) we gave clearer instructions, and (b) we let recipients designate which household member they wanted to receive the transfers, which gives them flexibility to choose someone who already has a National ID; in the Rarieda round we could not do this as we were randomizing recipient gender. I expect the Nike cohort will take longer to register as that project focuses on 18-19 year old women, many of whom will not yet have IDs." GiveDirectly, Updated data (March 31, 2012)

• 280
  • Paul Niehaus, GiveDirectly Founder, email to GiveWell, November 20, 2012.
  • Email from Paul Niehaus, President, GiveDirectly, and Joy Sun, COO, Domestic, GiveDirectly, November 18, 2013

• 281
  • GiveDirectly, Monthly operations report, August 2015
  • Note that these numbers are slightly skewed to be later than average because GiveDirectly has had to delay token transfers for the last two months so as not to bias the General Equilibrium study. Paul Niehaus and Carolina Toth, conversation with GiveWell, September 7, 2015

• 282
  • The hotline records also document the steps that GiveDirectly has taken to follow up on recipients' concerns; we have reviewed some of these cases and find the steps taken to resolve issues to be reasonable and appropriate.
  • GiveDirectly, Kenya hotline log, July 2013
  • GiveDirectly, Follow-up tracker, July 2013
  • GiveDirectly, Follow-up tracker, October 2014
  • Data from GiveDirectly indicates that on average 11% of recipients in Kenya campaigns are delayed due to registration issues. GiveDirectly, Monthly operations report, August 2015

• 283
  • "The main downside to both of the mobile money services in [Uganda] as compared to MPESA in Kenya is that there are fewer mobile money agents in the rural areas that GiveDirectly is targeting. In response, GiveDirectly has been more proactive in coordinating with the mobile money service for the transfers that have begun, for example, by giving the mobile money service advanced notice before sending the funds so that agents could be prepared. In some cases, agents traveled to the villages in which recipients live to reduce recipient travel time." Conversation with Paul Niehaus, President, and Joy Sun, COO, Domestic, GiveDirectly, July 18, 2013
  • "We do inform recipients in advance when pay-days will be happening" Email from Paul Niehaus, President, GiveDirectly, and Joy Sun, COO, Domestic, GiveDirectly, November 18, 2013

• 284

  Conversation with Piali Mukhopadhyay, COO, International, GiveDirectly, October 22, 2013

• 285

  Carolina Toth, email to GiveWell, September 14, 2015

• 286
  • The amount of time that it takes for total transfers to be sent has varied between campaigns. The Google transfer campaign likely gives the best approximation of timing for a standard campaign, because it does not involve an experiment (like Rarieda and Nike), and it is more recent than the Siaya campaign. In the Google campaign, the planned schedule for transfers lasts approximately eight months for any given recipient, not including the time for census and enrollment. GiveDirectly, Google transfer schedule, July 2013
  • "GiveDirectly stops issuing transfers after two years and clearly informs recipients that they should expect this." GiveDirectly, Offering Memorandum (January 2012) Pg 33.
  • Update [November 2014]: the two year timeline mentioned above is now outdated; transfers are generally sent over a period of 8 months in Kenya and 10 months in Uganda. It is still the case that recipients can only receive one full transfer, and that they are ineligible for additional transfers thereafter. Conversation with Carolina Toth, GiveDirectly, November 20, 2014
  • In November 2015, GiveDirectly informed us that it now aims to send transfers within 4 months. Carolina Toth, email to GiveWell, November 10, 2015
  • Paul Niehaus and Carolina Toth, conversation with GiveWell, September 7, 2015
  • "In Kenyan Shillings, they are 10K, 50K, 50K (assuming no phone purchase). Ugandan shillings is 3.2M but division into different transfers hasn't been decided yet." Carolina Toth, email to GiveWell, September 25, 2015
  • XE currency converter, Kenya shillings to US dollars, September 25, 2015
  • XE currency converter, Uganda shillings to US dollars, September 25, 2015

• 287

  Carolina Toth, email to GiveWell, November 10, 2015

• 288

  @GiveDirectly, Contextualizing Transfer Size@

• 289

  $0.65 in pre-cash-transfer income per person per day implies (365*$0.65) = $237.25 per person per year. If each person receives $288 in a year from GiveDirectly, that's (288/237.25) = 121%.

• 290

  Paul Niehaus, GiveDirectly Founder, email to GiveWell, November 20, 2012. We have not reviewed the data GiveDirectly used to reach this conclusion.

• 291

  In our conversations with recipients and field staff, we phrased this question two different ways:

  • Do you think it would be better for GiveDirectly to provide $1000 transfers to households in one village or $500 transfers to households in two villages?
  • Do you think that GiveDirectly should keep the transfer size the same or reduce the transfer size but provide transfers to a greater number of people?

  GiveWell site visit to GiveDirectly, October 2014
  Conversation with GiveDirectly field staff, October 20-21, 2014, Pgs 5-6.

• 292

  GiveWell site visit to GiveDirectly, October 2014

• 293

  We asked the field officers what they think about the current transfer size ($1000), and whether they’d choose to keep it at that level, increase it, or decrease it, given the effects that an adjustment would have on how many people GiveDirectly would be able to serve.

  • Mr. Okello: Typically there are multiple households on one compound, each inhabited by relatives of the same family, and any household that meets the targeting criteria can receive transfers. Mr. Okello said that it may make more sense for GiveDirectly to group some households on a compound together so that transfers are shared across them, rather than each eligible household receiving the full $1000.

    Mr. Okello also said that if GiveDirectly increased the size of the transfers, that could create a high level of dependency. One of the messages that field officers send is that people should use the $1000 transfers to develop themselves as much as possible, but if someone knew they were getting $2000, they may stop farming, for example. With $1000 people can get some things but not everything; it is the right amount.

  • Mr. Ekeu: Mr. Ekeu prefers reducing the amount of money in each transfer and expanding the recipient base to reach everyone in the village. He said that the current targeting model causes bragging and unrest in the communities. The people who don’t benefit may be brought to use force to get some of the money, such as by breaking into recipients’ homes. Mr. Ekeu suggested that it would be better for GiveDirectly to provide all households in a village with some amount of money, even if it was less for households that are currently deemed ineligible (e.g., $100). This way, each of the households would be busy figuring out how they would spend their own money rather than how to get money from another.
  • Mr. Olinga: Mr. Olinga said that to reduce extreme poverty the bigger transfer is better, but he didn't have a strong opinion on $1000 transfers to some people versus $500 transfers to twice as many. [This is how we posed the question to Mr. Olinga.]

  Conversation with GiveDirectly field staff, October 20-21, 2014, Pgs 5-6.

• 294

  Conversation with GiveDirectly, September 5, 2014

• 295

  Conversation with GiveDirectly, September 5, 2014

• 296

  Conversation with GiveDirectly, September 5, 2014

• 297
  • GiveDirectly, Estimate of personnel 2015
  • As of August 2015, GiveDirectly had 60 staff: 3 Field Directors, 3 Field/Finance Managers, 5 Associate Field Managers, 46 Field Officers, and 3 other staff. @GiveDirectly, Monthly Operations report, August 2015@

• 298

  GiveDirectly has so far received about six applications for every one FO [Field Officer] position it has open, which it sees as an indicator that the necessary talent is available for it to scale its operations. Conversation with Paul Niehaus, President, and Joy Sun, COO, Domestic, GiveDirectly, July 18, 2013 (unpublished)

• 299
  • Paul Niehaus, GiveDirectly Founder, conversation with GiveWell, October 22 2012.
  • GiveDirectly, Budget summary, July 2013

• 300

  Conversation with Piali Mukhopadhyay, COO, International, GiveDirectly, October 22, 2013

• 301

  GiveWell, GiveDirectly financials 2015, Sheet: "2015-Total spend + efficiency"

• 302

  Carolina Toth, conversation with GiveWell, November 12, 2015

• 303
  • GiveWell, GiveDirectly financials 2015 Sheet: "2014-Total expenses"
  • Note: GiveDirectly used to exclude marketing (i.e. fundraising) AND set-up costs from its calculation:
    • "First, as a general principle we'll always aim to publish spending clearly enough so that anyone can calculate efficiency however they want. Second, when we analyze our efficiency for management and forecasting purposes we distinguish between (1) operating costs that are ongoing but shared across campaigns and (2) one-time setup costs. (1) is included in the efficiency calculation and (2) is not. Examples of (1) are Field Director time (which is "fixed" for the duration of his/her contract) and office space (which is "fixed" with respect to number of ongoing campaigns). Examples of (2) are costs of registering as an NGO in a new country and costs of developing code which can then be re-used indefinitely. We accept donations from the public for (1) but not (2)." Email from Joy Sun, COO, Domestic, GiveDirectly, August 9, 2013
    • "Set-up costs are excluded because one-time costs (e.g. NGO registration) can be re-used indefinitely, as opposed to marginal costs." Email from Paul Niehaus, President, GiveDirectly, and Joy Sun, COO, Domestic, GiveDirectly, November 18, 2013
  • However, GiveDirectly now includes set-up costs in its campaign costs, as it no longer thinks that separating out set-up costs provides additional useful information. Paul Niehaus and Carolina Toth, conversation with GiveWell, September 7, 2015

• 304

  GiveWell, GiveDirectly financials 2015 Sheet: "2015-Total spend + efficiency"

• 305

  Carolina Toth, conversation with GiveWell, November 12, 2015

• 306

  Carolina Toth, email to GiveWell, November 10, 2015

• 307

  GiveWell, GiveDirectly financials 2015 Sheet: "2015-Total spend + efficiency"

• 308

  GiveWell, GiveDirectly financials 2015 Sheet: "2015-Total spend + efficiency" Note that the incurred "direct grants to households" includes transfers that GiveDirectly has already officially committed to, but not yet sent. The budget for the "Future" column may not represent the total amount that GiveDirectly is planning to spend in 2016, because GiveDirectly told us they expected to run out of their unallocated funds by February 2016. Paul Niehaus and Carolina Toth, conversation with GiveWell, September 7, 2015. Note that this prediction was before GiveDirectly received news that it should soon expect a $6 million grant from another funder, and that it did not include the $25 million in funding that GiveDirectly received from Good Ventures in mid-2015.

• 309

  For example, the National Institutes of Health funded Innovations for Poverty Action to conduct the RCT of GiveDirectly's program: "GiveDirectly is conducting a longer-term evaluation to provide more detailed, context-specific evidence on how its recipients use cash transfers.The study is coordinated by an external research organization, Innovations for Poverty Action, led by Dr.Johannes Haushofer of Harvard University, and funded by the National Institutes of Health." GiveDirectly, Offering Memorandum (January 2012).

• 310
  • Conversation with Paul Niehaus, President, and Joy Sun, COO, Domestic, GiveDirectly, October 16, 2013
  • GiveDirectly, GW scratch sheet
  • GiveWell, GiveDirectly financials 2015 Sheet: "2015-Balance Sheet"

• 311

  GiveWell, GiveDirectly financials 2015 Sheet: "2015-Total spend + efficiency"

• 312Includes Core Operations and Core Operations-general. Excludes fundraising.
• 313

  GiveWell, GiveDirectly financials 2015 Sheet: "RFMF Scenarios"

• 314

  GiveWell, GiveDirectly financials 2015 Sheet: "RFMF Scenarios"

• 315

  GiveWell, GiveDirectly financials 2015 Sheet: "RFMF Scenarios"

• 316
  • Paul Niehaus, Michael Faye, and Piali Mukhopadhyay, conversation with GiveDirectly supporters, August 11, 2015
  • GiveDirectly has a pipeline of several Field Directors whom it could hire if additional Field Directors were needed. It is easy to find additional Field Managers, Assistant Field Managers, and Field Officers in Kenya and Uganda because unemployment rates are high, and many qualified candidates are looking for jobs. Paul Niehaus and Carolina Toth, conversation with GiveWell, September 7, 2015
  • GiveDirectly has had more difficulty building up its domestic capacity; it is currently increasing its search efforts for several key leadership positions in the U.S. Paul Niehaus and Carolina Toth, conversation with GiveWell, September 7, 2015

• 317

  Carolina Toth, conversation with GiveWell, November 12, 2015. GiveDirectly currently has a $4 million grant to conduct a benchmarking study Rwanda (more below), but could use the staff it is setting up in Rwanda to start a standard campaign that would enroll recipients for up to $4 million in transfers.

• 318

  Paul Niehaus and Carolina Toth, conversation with GiveWell, September 7, 2015

• 319

  GiveWell, GiveDirectly financials 2015 Sheet: "RFMF Scenarios"

• 320

  GiveWell, GiveDirectly financials 2015 Sheet: "RFMF Scenarios"

• 321

  Carolina Toth, email to GiveWell, October 9, 2015

• 322

  GiveDirectly, Update for GiveWell, September 2015

• 323

  GiveWell, GiveDirectly financials 2015 Sheet: "2015-Balance Sheet"

• 324

  Paul Niehaus and Carolina Toth, conversation with GiveWell, September 7, 2015

• 325

  See our blog post on the grant.

• 326

  "Officially" in this context means that GiveDirectly's board has approved the use of funds for that purpose. GiveWell, GiveDirectly financials 2015 Sheet: "2015-Balance Sheet"

• 327

  $1.34m/month in March - June 2015 x 12 months = $16.08m/year. GiveWell, GiveDirectly financials 2015, Sheet: "2015 - Rate of money moved." Time periods are first and second halves of GiveDirectly's current fiscal year, with the exception of the last period, for which we only have data through June 2015.

• 328

  GiveDirectly, Update for GiveWell, September 2015 states that, without any additional hiring, GiveDirectly will have 3 Field Directors involved in cash transfers in 2016. In Paul Niehaus and Carolina Toth, conversation with GiveWell, September 7, 2015, GiveDirectly told us that one of these Field Directors had been recently hired and was expected to start in September 2015.

• 329

  GiveWell, GiveDirectly financials 2015 Sheet: "2015-Rate of money moved"

• 330

  GiveWell, GiveDirectly financials 2015, Sheet: "2015 - Rate of money moved." See chart for "Committed and distributed transfers to date."

• 331

  See above where we note that GiveDirectly would need around $103.4 million for transfer and operation costs, and we estimate that GiveDirectly will have $30-35 million on hand by the start of its next budget year (sans GiveWell funding).

• 332
  • GiveWell, GiveDirectly financials 2015 Sheet: "2015-Commitments by month"
  • If GiveDirectly's Field Directors are currently operating at a capacity of $6 million per year, then $17.5 million per year would require that GiveDirectly increase its rate by 3x ($6 million x 3 = $18 million).

• 333

  GiveDirectly is also meeting with senior or retired government officials who can provide guidance on navigating the government and connect GiveDirectly to allies on the public sector side. Conversation with GiveDirectly, April 8, 2014, Pg 11

• 334

  By now, GiveDirectly understands well the process for seeking government approvals in Kenya and does not see acquiring approvals as a major risk. Conversation with Piali Mukhopadhyay, GiveDirectly, October 20-21, 2014, Pg 3.

• 335
  • "Bukedia district [Uganda] still has 27,000 un-visited, eligible households, and country-wide registration is in process that will provide approval for all 7.3m households in Uganda." GiveDirectly, Update for GiveWell, September 2015, Pg 3
  • County-level government approval equivalent to ~70 K additional eligible households in Siaya and Homa Bay counties" GiveDirectly, Update for GiveWell, September 2015, Pg 3
  • GiveDirectly already has approval to work in Bukedia district in Uganda, so even if it did not obtain country-wide permission in Uganda, it would still be able to work in Uganda for a long time. Most of the households that GiveDirectly has permission to enroll in Kenya are in Homa Bay county, although GiveDirectly still needs to speak to some of the districts in Homa Bay to get approval at a more local level. GiveDirectly could also go back and enroll the houses that are controls in the General Equilibrium study in Siaya county once that study is complete. Paul Niehaus and Carolina Toth, conversation with GiveWell, September 7, 2015
  • At its current rate of enrollment (~1,000 households/month), with one Field Director in Uganda, GiveDirectly could continue to enroll households in Uganda for the next 2 years with no new permissions (1000 households per months x 12 months per year x 2 years = 24,000 households, which is still less than 27,000). In Kenya, GiveDirectly could enroll households for approximately 6 years with no new permissions, assuming it obtains the district permissions successfully (1000 households per months x 12 months per year x 6 years = 72,000 households, which is slightly more than the 70,000 that are available).

• 336
  • Political violence and terrorism are both risks in Kenya. Western Kenya has not been impacted since 2008 election violence
  • Operations in Uganda provide an alternative, and funds could be shifted more heavily toward UG

  GiveDirectly, Update for GiveWell, September 2014, Pg 16.
  

• 337
  • "Dependency on a single provider: low risk. In Uganda, we have Ezee Money and Airtel as a back-up, and in Kenya could use Equity Bank."
    GiveDirectly, Update for GiveWell, July 2014 Pg 12
  • "Selected MTN as preferred provider in Uganda after assessing performance of Ezee/MTN (building relationship with Airtel so as to have an additional hedge)"
    GiveDirectly, Update for GiveWell, July 2014 Pg 9
  • GiveDirectly continues to seek other mobile money providers who it might want to partner with for future campaigns in Uganda (e.g., Airtel).
    Conversation with GiveDirectly, April 8, 2014 Pg 5

• 338

  Paul Niehaus and Carolina Toth, conversation with GiveWell, September 7, 2015

• 339GiveDirectly, Give now
• 340

  "They created GiveDirectly as a private giving circle in 2009 and opened it to the public in 2011 after two years of operational testing." GiveDirectly, FAQs 2015
  GiveDirectly, Team
  Paul Niehaus, GiveDirectly Founder, email to GiveWell, November 20, 2012

• 341

  Conversation with Paul Niehaus, November 14, 2014

• 342

  Paul Niehaus and Carolina Toth, conversation with GiveWell, September 7, 2015