Notes on a site visit to GiveDirectly in Western Kenya on November 6 - 8, 2012

Good Ventures, along with GiveWell, visited GiveDirectly in western Kenya in November 2012 to observe its work in the field and speak with beneficiaries of its cash transfer program.

Notes from Day 1 — Tuesday, November 6, 2012

We visited three households in Siaya. Two of these households had received their first of two $500 transfers in July 2012 and were selected in advance by GiveDirectly. The third was a non-recipient household chosen once we were in the village. We made the visits with a GiveDirectly staff member (Piali Mukhopadhyay) and the village elder accompanying us.

Notes from Day 2 — Wednesday, November 7, 2012

On Day 2 we drove to another village in Siaya district to observe GiveDirectly's enrollment process and another village to visit more households. During the drive, we had the chance to speak in great detail with Piali about the selection and auditing process GiveDirectly has used throughout its history. We then visited:

The households in the first village we visited were much more rural than households we had visited a day earlier. Where a day earlier, the village was within 100 yards of a paved road, these households were ~3 miles of very narrow dirt roads away from the paved road. Even once we had driven the 3 miles to a point near the village, we had to walk down paths for ~5-10 minutes in between each household. So, for these households the closest "major" market (Siaya town, more on this below) was about 30-60 minutes away, assuming they had a ride once they reached the paved road.

These households are part of a batch of 1,000 that GiveDirectly is currently enrolling to distribute $1 million of a $2 million grant they received in mid-to-late 2012. The other $1 million is being set aside to start up operations in a new country.

Then, we went to the next village to meet with recipients and a non-recipient. 

Notes from day 3 —Thursday, November 8, 2012

On day 3 we visited the village in Rarieda county where the RCT is currently being completed.  We observed IPA field staff members administering the endline survey to people in their homes. This village was very spread out, with at least 5 minutes walking distance between each home.  The area was very scenic and lush, again we saw many farms near the houses implying that agriculture was common in this village as well.

Innovations for Poverty Action (IPA) staff told us that they compensate surveyed individuals with in-kind gifts (e.g., with cooking fat worth roughly the wages the household could earn during that time) for the time they spend on the survey. IPA surveys both adult family members together (for about ½ day) and then surveys the male and female heads of household separately (for about 1/4 day each).

The survey asks many detailed questions —e.g., how much have you spent on meat, fish, eggs, milk, etc. in the past month — and we found ourselves wondering whether respondents could realistically provide accurate answers.

Also, given the duration of the interviews we wondered whether respondents would likely be willing to maintain sufficient energy to respond thoughtfully and carefully.

Later, we drove to a nearby village to meet with a few recipients we had chosen randomly from a list of recipients GiveDirectly gave us and the village elder.  This village had homes that were slightly more clustered together in parts. 

Note that we have mostly visited thatched-roof or previously thatched roof houses; we have only visited 2 houses from people with (pre-GiveDirectly) tin roofs.  The following are general observations of the characteristics of the homes we visited:

Overall, richer households are larger, lighter (e.g., have windows), are more likely to have radios/electronics, have more wall hangings, and are in better repair (mud homes start to crack quickly and many cracks seems correlated with the other items.)

Experience registering for and using M-PESA

GiveDirectly transfers money to recipients through the M-PESA mobile phone payment system. M-PESA is reasonably common in Kenya. We saw M-PESA outlets in virtually every shopping center we drove through both in cities and near villages, and our general impression is that it would be easy for an individual to find a place to register for, deposit funds in, and withdraw funds from M-PESA. This was reinforced by the fact that staff members of GiveDirectly and Innovations for Poverty Action both use M-PESA regularly for their personal expenses.

M-PESA charges fees on transactions which vary depending on the transaction size but are generally are low single-digit percentages of the transaction.

In order to understand better how M-PESA works, Elie registered for and used M-PESA on his phone.  This was our experience with M-PESA:

Overall, the M-PESA agent seemed very comfortable with the system and helpful. Based on our observations of every commercial area we passed through, from small markets in rural areas to city streets, M-PESA agents seem to be ubiquitous.